QE | A399 Volume 91 FERN nite RP L January—March 2001 QUARTERLY JOURNAL OF THE AMERICAN FERN SOCIETY The Group of Adiantum gracile in Brazil and Environs David B. Lellinger and Jefferson Prado 1 Dryopteris X correllii hyb. nov. (D. sepsis x goldiana), a Rare Woodfern Hybrid from Vermont Warren H. Wagner, Jr. and Arthur V. Gilman 9 The Relictual Fern Genus Loxsomop Marcus Lehnert, Maren Ménnich, Thekla Pia Alexander Schmidt-Lebuhn, and Michael Kessler 13 Growth, Leaf Cl teristics, and Spore Prod in Native and Invasive Tree Ferns Hawaii Leilani Z. ova and Guillermo Goldstein 25 Shorter Notes Binomial for Dryopteris clintoniana X goldiana James H. Peck 36 Cryopreservation of Shoot Tips of Selaginella uncinata Valerie C. Pence 37 The American Fern Society Council for 2001 BARBARA JOE HOSHIZAKI, 557 N. Westmoreland Ave., Los Angeles, CA 90004-2210. President CHRISTOPHER H. HAUFLER, Dept. of Botany, University of Kansas, Lawrence, KS 66045-2016. Vice-President W. CARL TAYLOR, 800 W. Wells St., Milwaukee Public Museum, Milwaukee, WI 53233-1478. ecretary JAMES D. CAPONETTI, Dept. of Botany, University of Tennessee, Knoxville, TN 37916-1110. Trea. DAVID B. LELLINGER, 326 West St. NW., Vienna, VA 22180-4151. robin Gece JAMES D. MONTGOMERY, Ecology III, R.D. 1, Box 1795, Berwick, PA 1860 RA ack Issues Curator GEORGE YATSKIEVYCH, Missouri Botanical Garden, P.O. Box 299, St. Louis, 0 63166-0299. Journal Editor DAVID B. LELLINGER, U.S. National Herbarium MRC-166, Smithsonian Institution, Washington, DC 20560-0166. Memoir Editor CINDY JOHNSON-GROH, Dept. of Biology, Gustavus Adolphus College, 800 W. College Ave., St. Peter, MN 56082-1498. Bulletin Editor American Fern Journal EDITOR ' R. JAMES HICKEY y Departme ie Geet OH 45056 ph. (513) oa e-mail: hickeyrj @ muohio.edu ASSOCIATE EDITORS GERALD I. GASTONY ..3. i050 Dept. of Biology, Indiana egies Bloomington, IN 47405-6801 CHRISTOPHER H. HAUFLER .... Dept. of Botany, University of Kansas, Lawrence, 66045-2106 ROBBIN C. MORAN New York Botanical Garden, Bronx, NY 10458-5126 JAMES H. PECK Dept. of Biology, University of Ar Arkansas—Little Rock, 801 S. University Ave., Little Rock, AR 72204 “American Fern Journal” (ISSN 0002-8444) is an illustrated quarterly devoted to the Sosa seat of ferns. ii is owned by the American Fern Society, and ares at 326 West St. NW., VA 22180-415 . Periodicals postage paid at Vienna, VA, and additional entry. aims adie’ - missing issues, made 6 nestic) to 12 sip pies ae after the date of issue, r back issues should be addressed to Dr. James D. Montgomery, Ecology III, R.D. 1, wrest Sg ay 18603-9801. Changes of address, dues, and applications for membership should be sent to the Membership General inquiries concerning fe the Secretary. Subscriptions $20.00 gross, $19.50 net if paid thro sol an Bg bs eg fee $0.50); sent gti - members of the Somer lage! en (annual dues, $15.00 + feng ailing surcharge beyo U.S.A.; life membershi + $140.00 raed a beyon Back volumes are cachet a most years as issues or on ce Please contact the Back Issues Curator for prices and availability. STMASTER: Send address changes to AMERICAN FERN JOURNAL, 326 West St. NW., Vienna, VA 22180-4151. FIDDLEHEAD FORUM The editor of the Bulletin of the American Fern Society wel contributions from members and non-members, including miscellaneous notes, spi to — or purchase materials, personalia, horticultural notes, and reviews of non-technical books on fi SPORE EXCHANGE ae Stephen McDaniel, 1716 Piermont Dr., Hacienda Hts., CA 91745-3678, is Director. Spores changed and lists of available spores sent on request. _http://www.amerfernsoc -org/sporexy.html to m b and to others interested sifts qi the ety in ferns. Back issues of the Journal and cash or other gift J] it d are tax-deductible. MISSOURI BOTANICAL JUL 1 2 2001 GARDEN LIBRARY American Fern Journal 91(1):1—8 (2001) The Group of Adiantum gracil in Brazil and Environs DAvip B. LELLINGER Department of Botany, National Museum of Natural History, Smithsonian Institution, Washington, DC 20560-0166 JEFFERSON PRADO Segao de Briologia e — Instituto de Botanica, Caixa Postal 4005, 061-970 Sao Paulo, SP, Brasil ABSTRACT.—The Adiantum gracile group of Brazil and adjacent Bolivia is a natural group distin- guishable from A. tetraphyllum and related species. We provide a key to the group a describe two new Brazilian species belonging to the group, A. cinnamomeum and A. dawson Adiantum is a large and widespread genus in the tropics, numbering ap- proximately 200 species in the Neotropics, 65 to 70 of them growing in South America. Brazil has about 50 species, about 75% of the total known for the continent. No recent monograph or revision exists for Adiantum in the Neotropics, nor has a comprehensive subgeneric classification of the genus ever been devel- oped. In preparing keys for floristic works, most authors have used as principal characters lamina venation (anastomosing vs. free) and shape of the ultimate divisions (roughly rounded and not dimidiate vs. more or less oblong and usually dimidiate). Secondary characters include lamina division (simple, pin- nate, bipinnate, or decompound), pinnule attachment (sessile vs. pedicellate), ultimate division articulation (articulate or not), and lamina apex shape (con- form vs. gradually attenuate). The species with bipinnate laminae, a conform terminal pinna, and oblong and usually dimidiate pinnules may be placed together conveniently as the A. tetraphyllum group. Delimited in this broad manner, the group includes about a third of the Neotropical species of Adiantum. Most, if not all, of the species of this group lack evenly and finely serrate sterile lamina margins that seem to predominate in other groups. Within the A. tetraphyllum Willd. group sensu Jato, four groups can be fairly readily distinguished, principally on the basis of their rachis and lamina in- dument. These groups are not yet fully characterized, nor are their constituent species known with certainty due to several names for which we have yet seen any specimens. Three of these groups are large and widespread in the Neo- tropics, but one small group is nearly restricted to Brazil: A. gracile Fée and two new species. The species of the A. gracile group differ from the other three groups in having more pinna pairs (5-7), more pinnules on a pinna (35—40 pairs), and pinnules that often are approximate or nearly so, narrower (2—4 mm wide) but 2 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) relatively longer, usually at least slightly lobed, and in lacking an obvious midrib, although they could be obscurely dimidiate. In addition, the larger scales on the abaxial surface of the pinnules, when present, are several cells wide and long-ciliate throughout, much like those of the rachises and costae. These scales are in contrast to those of A. tetraphyllum and its other allies, which are narrowed and sometimes even uniseriate distally. In the A. gracile group, the rachises are conspicuously covered with cinnamomeous, linear- lanceolate scales with long-ciliate margins and base, at least when young; these scales are most persistent in A. cinnamomeum. KEY TO THE SPECIES OF THE ADIANTUM GRACILE GROUP 1. Pinnules abaxially glaucous, the scales few or none, but rather inconspicuous, yellow- ish, sessile glands often present; pseudoindusia erose-ciliate at maturity, adaxially gla- brous. Rhizome scales narrowly lanceate, dark to medium brown, sparingly denticulate. Medial pinnules ca. 2 times longer than wide, usually acute at the sterile apex, nearl wounistl ot tee Fertile ays ee oe ed iy kag 3. A. dawsonii 1. Pinnules abaxially not glaucous or glandular, always sparsely scaly; pseudoindusia entire to erose, scarcely ciliate, adaxially glabrous or bearing toothed scales ........ 2(1). Medial pinnules ca. 2 times longer than wide, obtuse to acute at the sterile apex, angular at the fertile apex; pseudoindusia glabrous; rhizome scales medium to Ct AUTEN, WORE OS ee ee oe ya te ec ess a Fy 1. A. gracile 2(1). Medial pinnules ca. 3 times longer than wide, acuminate to acute at the always sterile apex; pseudoindusia bearing hairlike scales that are toothed at the base; thizome scales dark brown to blackish, the margins sometimes slightly paler, alwys sparsely dentioniate cc. 3. 5 Ske Ww eww ok 2, A. cinnamomeum 1. Adiantum gracile Fée, Gen. Fil. 116. 1852. Fig. 1A, B. Plants terrestrial. Rhizomes stout, short-creeping, ca. 4-5 mm in diam., scaly, the scales somewhat shiny, essentially concolorous, medium to dark brown, narrowly lanceate, entire. Fronds monomorphic, 2-pinnate, (30)60-100 cm long, the laminae (7)20—30 cm wide; stipes approximate, black, adaxially sul- cate, scaly, the scales appressed throughout or sometimes distally patent, con- colorous, cinnamomeous, 2~3 mm long, narrowly lanceate with a filiform apex, strongly denticulate proximally; rachises similar to the stipes and their indument similar; pinnae oblong-lanceate, slightly decreasing at the base, ta- pering at the apex, (7)15—22 cm long, 1.5—2.2 cm wide, the lateral pinnae (3)7- 10 pairs, patent, alternate, the terminal pinna conform, 1—1.5 times longer than the subtending pinnae 0.67—1 times as long as the medial pinnae; indument of the costae like that of the stipes and rachises; pinnules 27—43 pairs, ca. 2 times longer than wide, chartaceous, free-veined, without an evident midrib, the proximal pairs reduced, somewhat rounded or triangular, the medial pairs dimidiate, oblong to somewhat tapering, the acroscopic base truncate, the ster- ile apex obtuse to acute, the sterile margin irregularly and distantly serrate, except at the pinnule apex, the fertile apex angular, the distal pinnules ca. 1/2 as long as the medial pinnules, the adaxial surface of the pinnules glabrous or sparsely scaly near the sori, the veins slightly prominulous, the idioblasts LELLINGER & PRADO: ADIANTUM GRACILE GROUP 3 Acc¥ Spot Magn = wo 69. Roo TF a2 AccV¥ SpotMagn Det WD Ep F- 1WOkKV40 3027 105 6 IBUA. grac WOkV 40 3112x E 1069 IBA. grac Spot Magn Det WD Exp -————————1_ 50 um Spot Magn a wD Ep -— § 1366x SE 16.0 6 IBUA. cinnam. 10 OnvV40 1316x 160 3 IBA. cinnam 100 jm Spot Magn Det WD a ————————|, 2 Acc V agn «= Det WD Ep hooKy 40 2636x SE 105 (BUA. daws 100kV 40 7x SE 1054 BUA. daws Fic. 1. Scanning electron micrographs of Adiantum. Fics. 1A—B. Proximal eh mRaie views of A. gracile spores (Salino 393, UEC). Fics, 1C-D. Overall and distal v A a ype US). Fics. 1E-F. Suprabasal nee on ofa ance scale and an pseudoindusial scale of A. cinnamomeum (MacFarland et al. 283, holotype US). G-H. Proximal and distal views of A. dawsonii spores (Dawson 14868, holotype US). 4 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) inconspicuous or nearly so, the abaxial surface of the pinnules glabrescent, with patent, sparse, toothed scales ca. 0.5 mm long but otherwise similar to those of the stipes, glands absent, the veins slightly prominulous, the idioblasts conspicuous; sori oblong, up to 7 or 8 per pinnule; pseudoindusia glabrous, entire to erose; spores trilete, 25-30 4m in equatorial diameter, tetrahedral- globose without prolonged angles, the surface rugulate with a thin, fragment- ing outer layer, and so sometimes appearing finely cristate. TYPE: Brazil: sin. loc., Claussen s.n. (P, Morton photo 2596, photos GH, SP; possible isotype US1148302 [Claussen 177 ex C]). DISTRIBUTION: Endemic to central Brazil (Mato Grosso, Distrito Federal, and Goias). HABITAT: Slopes and stream banks in cerrado and campo vegetation, at ca. 50-850 m elevation. SPECIMENS EXAMINED: Brazit: Mato Grosso: Sararé, RADAMBRASIL, Serra de Pedra (S. Aguapei), 59°25'W, 16°10'S, Pires & Santos 16572 (UEC); Chapada dos Guimaraes, Véu-da-Noiva, Verardo 23668 (SP, UEC), same locality, Salino 393, 500 m (UEC), same locality, Hatschbach 37628 (MBM, UC). Goids: Ca. 1 km S of S. Joao de Alianga, ca. 850 m, Irwin et al. 31972 (NY not seen, SP, US); Serra do Caiap6, 50 km S of Caiaponia, Prance & Silva 59588 (GH, K, NY not seen, US). Distrito Federal: Brasilia, R. Windisch & Ghilldny 253 (HB). 2. Adiantum cinnamomeum Lellinger & Prado, sp. nov. Figs. 1C-F, 2. A specie A. gracili Fée paleis rhizomatis integris (vs. sparse denticulatis), pinnis 5—7 (vs. 8-10) jugatis patulis, pinnulis oblongis acuminatis, 3-plo (vs. 2-plo) longioribus quam latioribus differt. TYPE: Brazil: Rond6nia: 120 km SW of Pérto Velho-Highway BR 364, 15 km W of Mibrasa, 9°10’ S, 63°07’W, 29 May 1982, McFarland et al. 283 (holotype US photo SP; isotypes BM, GH). Plants terrestrial. Rhizomes stout, short-creeping, ca. 3 mm in diam. exclud- ing the protuberances, scaly, the scales shiny, slightly bicolorous, dark brown to blackish, narrowly lanceate, the margins lighter and sparsely denticulate. Fronds erect, monomorphic, 2-pinnate, 35-65 cm long, the laminae 20-25 cm wide; stipes ca. 5 mm distant, black, adaxially sulcate, scaly, the scales ap- pressed throughout or distally patent, concolorous, cinnamomeous, 4—5 mm long, linear-lanceate with a filiform apex, sometimes long-ciliate at the base, medially copiously toothed-ciliate; rachises more densely covered by scales than the stipes, the indument similar to that of the stipes; pinnae oblong- lanceolate, slightly decreasing at the base, tapering at the apex, 8-19 cm long, 2.0—3.5 cm wide, the lateral pinnae 5-7 pairs, patent, alternate, the terminal pinnae conform, 1—1.25 times longer than the subtending pinnae, ca. 1.33 times longer and wider than the medial pinnae, the indument of the costae = Fic. 2. Adiantum cinnamomeum. Fic. 2A. Frond (Steward et al. P20394, US). Fic. 2B. Long- creeping rhizome (MacFarland et al. 283, US). Fic. 2C. Rachis scales (MacFarland et al. 283, US). Fics. 2D-E. Abaxial surface of pinna and pseudoindusia showing scales (Irwin et al. 10180, US). LELLINGER & PRADO: ADIANTUM GRACILE GROUP SAAN aM ~*~. 2% a Wy ) Wh lee mil Wh ESS V/ 4 f\\ AS VES LS Ns \\ « WS . = CRS GN E Fe ae a Se Se Sl g a \\ SA Sif a SR 6 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) similar to that of the rachises; pinnules 35—40 pairs, ca. 3 times longer than wide, chartaceous, free-veined, without an evident midrib, the proximal pairs reduced, flabellate to subrhombic, the medial pairs dimidiate, oblong to some- what tapering, 0.5—-2 cm long, 2-4 mm wide, subarticulate (the dark color of the stalk terminating almost at the base of the pinnule), patent to erect, the acroscopic base truncate, roundish to straight, the apex subacute to acute, the sterile margins serrate to biserrate, except the basioscopic margin entire to the middle of the pinnule and dentate towards the apex, the distal pinnules ca. 1/4—1/5 as long as the medial pinnules, the adaxial surface of the pinnules glabrous, the veins not prominulous, the idioblasts conspicuous, the abaxial surface glabrescent, with sparse, cinnamomeous scales, the scales filiform, ca. 1 mm long, pectinate at the base, glands absent, the veins prominulous, the idioblasts inconspicuous; sori oblong, 7 or 8 per pinnule; pseudoindusia with filiform scales, bearing short basal processes, the margins entire to erose; spores trilete, 25-30 ym in equatorial diameter, tetrahedral-globose with pro- longed angles, the surface sparsely rugulate with a thin, fragmenting outer layer, and so sometimes appearing finely and sparsely cristate. DISTRIBUTION: North-central Brazil (Rondénia, Amazonas, Pard, Mato Grosso, Distrito Federal). HABITAT: On clay soil, in open canopy and dense understory, at 50-1000 m elevation. PARATYPES.—BRAZIL: Rondénia: Pérto Velho to Cuiabé Highway, vicinity of Santa Barbara, 15 km E of Km. 1117, Prance & Ramos 7157 (NY, UC). Amazonas: Manaus-Caracaraf Highway, Km. 160, Steward et al. P20394 (BM, MO, NY, US); Manaus-Caracarai Highway, Km. 185, Prance et al. 22691 (MO, NY, US); Rio Javari between Estirao do Equador and Rio Javarizinho, Prance et al. 24046 (NY, UC); Manaus, Rio Araras, SIDERAMA, Loueiro et al. s.n. (GH); Manaus, Pivetta 254 (HRCB); Barra, Prov. Rio Negro, Spruce s.n. (GH, US). Para: Municipio de Oriximind, Rio Trombetas, es- trada da Mineragao Santa Patricia, ramal 22, Cid et al. 1441 (NY, US); Rio Jamunda, Municipio de Faro, Sao Jorge, Black & Ledroux 50-10761 (HB). Mato Grosso: Vila Bela da Santissima Trindade, Faz. Cabixi, junto ao Rio Cabixi, 13°S, 60°10’W, ca. 12 km da divisa com Rondénia, Prado & Salino 8 (HB, SPF, UEC). Distrito Federal: Parque Municipal do Gama, Irwin & Soderstrom 5884 (NY, US); Irwin et al. 10180 (GH, K, MO, NY, US); Brasflia, Reserva Ecolégica do IBGE, Heringer et al. 3767 (MBM). Adiantum cinnamomeum is distinguished by its long, narrow pinnules that are acuminate to acute at the apex, presumably because the sori are always lateral. Adiantum cinnamomeum has been found only in Brazil, but may have a wider range in Amazonian South America. 3. Adiantum dawsonii Lellinger & Prado, sp. nov. Fig. 1G-—H, 3. A specie A. gracili Fée pinnulis adaxialiter glaucis, paleis paucis vel nullis, glandulis rotundis sessilibus luteolis differt. TYPE: Brazil: Goids: Southern Serra Dourada region, 20 km E of Formoso, 48°50'W, 13°45’S, on banks and margins of small stream running through hilly cerrado, 16 May 1956, E. Y. Dawson 14868 (US, 2 sheets). Plants terrestrial. Rhizomes stout, short-creeping, ca. 4-5 mm in diam., scaly, LELLINGER & PRADO: ADIANTUM GRACILE GROUP a” Gh ces BASS igh LE wdiiiler EASES SW jugs We MEE RR® avy MM le f x AAA! SESS s RASS SPLIT J 9 9 SSSidonoy Sadly lof A ws Phd iron = WYER TRE SO) OT <x US2291503). Fic. 3A. Frond. Fic. 3B. Rachis scales. Adiantum dawsonii (Dawson 14868, Fic. 3C. Pseudoindusial trichome. Fic. 3D. Pseudoindusia showing trichomes. Fic. 3E. Abaxial surface of pinnae. Fic. 3: 8 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) the scales narrowly lanceate, somewhat shiny, dark to medium brown, spar- ingly denticulate. Fronds monomorphic, 2-pinnate, 40-90 cm long, the lami- nae (15)20—30(35) cm wide; stipes approximate, black, terete or adaxially plan- ate, thinly scaly, the scales appressed throughout or sometimes distally patent, concolorous, cinnamomeous, 2—3(4) mm long, very narrowly lanceate with a long, filiform apex, weakly denticulate proximally; rachises black, terete, more densely scaly than the stipes, the scales similar to those of the stipes, except twice as wide, 3-6 mm long, and regularly denticulate; pinnae oblong-lancea- te, slightly decreasing at the base, tapering at the apex, the lateral pinnae 7— 8(10) pairs, (8)12—20 cm long, 1.25-1.75 cm wide, patent, alternate, the ter- minal pinna conform, up to 22 cm long, 2.25 cm wide, 1—1.25 times longer than the subtending pinnae and the medial pinnae, the indument of the costae similar to that of the rachises; pinnules (20)35-55 pairs, ca. 2 times longer than wide, chartaceous, free-veined without an evident midrib, the proximal pairs reduced, rounded-triangular, the medial pairs dimidiate, oblong to some- what tapering, the acroscopic base truncate, the sterile apices usually acute, the fertile apices rounded, the sterile margins irregularly serrate, the distal pinnules ca. 1/3 as long as the medial pinnules, the adaxial surface of the pinnules glabrous or with a few hairlike scales, especially near the sori, the veins prominulous to only slightly so, the idioblasts usually conspicuous, the abaxial surface of the pinnules glaucous, the scales few or none, round, sessile, pale yellow glands often scattered over the surface, the veins slightly prominu- lous, the idioblasts inconspicuous; sori oblong, 3-5 per pinnule, one com- monly apical; pseudoindusia glabrous or sparingly catenate-pilosulous, erose- ciliate at maturity; spores trilete, 25-30 um in equatorial diameter, tetrahedral- globose with prolonged angles, the surface rugulate with a thin, fragmenting outer layer, and so sometimes appearing finely cristate. DISTRIBUTION: Central Brazil (Mato Grosso and Goids) and Bolivia (Santa ruz). HABITAT: Slopes and stream banks at 500-1300 m elevation. PARATYPES.—BRAZIL: Mato Grosso: Gorge of Véu de Noiva, Chapada dos Guimaraes, Prance et al. 19082 (NY not seen, R, UC, US), Guarim Neto et al. 459 (HRCB); Serra da Chapada, Buriti, Malme Regn. Exped. I 1712 (S not seen, US); sin. loc., H. Smith s.n. (US). Goidés: Southern Serra Dourada region, 20 km E of Formoso, 48°50’W, 13°45’S, E. ¥. Dawson 14869 (US); Chapada dos Veadeiros, 1 km E of Alto Paraiso on road to Nova Roma, ca. 1300 m, W. R. Anderson 6305 (GH, HB, K, NY not seen, US); Rio Areias, Glaziou 22636 (NY, P not seen Morton photo 2589a, fragm US); sin loc., Glaziou 15726 (UC, US). Botivia: Santa Cruz: Velasco, Parque Nacional Noel Kempff M., Campamento Las Gamas, 14°48'00"S, 60°23’00"W, 900 m elev., Arroyo & Keil 200 (MO). ACKNOWLEDGMENTS The second author appreciates the financial support of the Brazilian Research Council CNPq (Proc. n. 300843/93-3) and of the Smithsonian Institution, Washington DC (Short-Term Visitor Grant). We thank also Sra. Emiko Naruto for preparing the illustration. American Fern Journal 91(1):9—12 (2001) Dryopteris Xcorrellii hyb. nov. (D. carthusiana xX goldiana), a Rare Woodfern Hybrid from Vermont WARREN H. WAGNER, JR.t Department of Botany, University of Michigan Ann Arbor, MI 48109-1048 ARTHUR V. GILMAN! P.O. Box 82, Marshfield, VT 05658 ABSTRACT.—A hybrid woodfern, Dryopteris carthusiana Xgoldiana, is described and named from a specimen (at DUKE) collected in Monkton, Vermont in 1937. This is the first reliable report of this hybrid. The report is based on morphology of the specimen; no living plants have been found for cytological study. Previous reports of this hybrid combination are excluded. Interspecific hybrids are of special interest in the study of the woodfern genus Dryopteris Adans.; more than 150 are known worldwide (Knobloch), including some 25 in North America (Montgomery, 1982). Some of these, al- though apparently sterile (not forming viable spores), are surprisingly com- mon, including D. carthusiana (Villars) H. P. Fuchs X intermedia (L.) A. Gray (= D. X triploidea Wherry) and D. cristata (L.) A. Gray X intermedia (= D. x boottii (Tuckerm.) Underw.). Many potential hybrid combinations, however, have not been reliably reported, including the one reported here, the hybrid of D. carthusiana and D. goldiana (Hook.) A. Gray (but see Montgomery [1982] for a preliminary notice of this collection). The plant shown in Figure 1 is the only known collection of this hybrid. It was collected by Donovan S. and Helen B. Correll in Monkton, Addison Co., Vermont, an area of abundant Dryopteris populations that include both par- ents. In addition to the abortive spores, it is distinguished by a number of characters that, to field workers familiar with the participating parents, are readily evident and not matched by any other hybrid combination. Specifi- cally, it differs from D. carthusiana in the less fine (bipinnate vs. bipinnate- pinnatifid) cutting of the lamina, blunter (less spinulose) toothing, blunter pin- nules, and dark (vs. pale) scales at the base of the stipe. From D. goldiana it differs in its finer cutting, triangular lower pinnae, and stalked proximal pin- nules. Sometimes, D. goldiana itself may be quite spinulose, e.g. Winslow 10102 (Vermont: Westmore: Mt. Pisgah, 19 July 1910, NEBC). However, these uncommon forms do not have triangular basal pinnae, stalked pinnules, or other characters evident in the hybrid. A closely related hybrid, D. goldiana x intermedia, is well distinguished from true D. carthusiana X goldiana (Evans and Wagner, 1964). Most significantly, that hybrid has abundantly glan- dular indusia and glands along the rachis; these are clear markers of D. inter- media parentage (Montgomery, 1982). In contrast, this hybrid is eglandular 1 Corresponding author. AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) PLANTS OF VepronT County: 4DDISov Dryopteris doldiana X spinulosa ? pu! mernerry In bog near Yonkton. GBo4 Col. D.S.Corren & H.5.Correll py. 7/29/1937 Det. Ne. (799A FIGURE 1. Correll’s hybrid woodfern, Dryopteris x correllii Wagner (D. carthusiana x goldiana), from Monkton, Addison Co., Vermont. This is the only known specimen. WAGNER & GILMAN: DRYOPTERIS X CORRELLI, HYB. NOV. 11 throughout. Also similar in overall aspect is D. clintoniana < goldiana (Dow- ell, 1908), but that hybrid has adnate pinnules, vs. the stalked pinnules of D. carthusiana X goldiana. In keeping with an American tradition of naming woodfern hybrids for prominent botanists, this singular hybrid is here named for its discoverer, the late Donovan S. Correll (1908-1983). Dryopteris <correllii Wagner, hyb. nov—TYPE: Vermont: Addison Co: Monk- ton; In bog near Monkton, 29 July 1937, D. S. & H. B. Correll 7799A (DUKE). Planta hybrida e Dryopteride carthusiana et D. goldiana exorta, inter am- bobus aspectu, a parentibus sporis abortivis differt. Frons usque ad 90 cm longinquitate (stipes 35 cm, lamina 55 cm); infimus stipitis vestitus squamis oblongi-linearibus, refulgentibus, castaneis, lamina subdeltoidea, bipinnata; pinnae proximae stipitatae; glandes laminis et indusiis desunt. Stout terrestrial fern arising from hybrid parentage (Dryopteris carthusiana x D. goldiana) and of intermediate aspect. Spores abortive. Overall frond to 90 cm long (stipe 35 cm, lamina 55 cm). Basal stipe scales large, dark shiny brown. Lamina subdeltoid, bipinnate. Proximal pinnules conspicuously stalked, triangular; basal basiscopic pinnules longer than next most basiscopic pinnules; pinnules blunt. Laminar and indusial glands absent. Only the specimen at DUKE is preserved. Duplicates may have been given to Dr. Henry Oosting for exchange, but to date we have not located any. Wagner has discussed elsewhere various collections that had been suspected of being this hybrid. Dryopteris < poyseri Wherry from Swarthmore, Delaware Co., Pennsylvania, was the first plant reported as this hybrid (Benedict, 1909), but studies showed it to be a bizarre form of D. clintoniana (Wagner and Wag- ner, 1982, Figs. 1 and 2), as originally thought by its discoverer (Clute 1908). More recently, a plant considered to be D. carthusiana X goldiana has been reported from Clarendon, Rutland Co., Vermont (Thorne and Thorne 1989). However, examination of this specimen (F. H. & E. Thorne s.n., 14 July 1981) shows it to be an odd frond of D. carthusiana. The specimen is barren of spores, sporangia, or indusia, and lacks any scales; it is more spinulose than the specimen from Monkton. Furthermore, it shows anomalous developmental characters such as forked pinnae and distally winged costae. Therefore, this report is also excluded. As far as we know at present, the Monkton specimen is unique. Why this hybrid does not occur more frequently is an enigma. ACKNOWLEDGEMENTS Wagner thanks Dr. Robert Wilbur, Curator of DUKE, for the loan of this specimen. Gilman ae Dr. Florence S. Wagner, Dr. James D. Montgomery and anonymous reviewers for com- ents on the deere te and Libby Thorne for loan of the Clarendon specimen, which will be apse at 12 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) LITERATURE CITED BENEDICT, R. C. 1909. New hybrids in Dryopteris. Bull. Torrey Bot. Club. 36:41—-49. CLUTE, 1908. Rare forms of Ferns—VI. A cut-leaved crest fern. Fern Bull.16:12 DOWELL, P 1908 New ferns described as hybrids in the genus Dryopteris. Bull. a 2 Club 35:135—-140. EVANS, : M. and W. H. WAGNER, ae 1964. Dryopteris rsa x intermedia, a natural hybrid woodfern cross of noteworthy morphology. Rhodora 66:255—266. KNOBLOCH, I. no date (ca. 1996]. Pecaontets Hybrids and nar Derivatives. Michigan State Uni- , Mi. MONTGOMERY, J. A. 2. Dryopteris in North America. Part II: the hybrids. Fiddlehead Forum 9(4):23-30. THORNE, F, L. and L. THORNE. 1989. Henry Potter’s Field pia to the Hybrid Ferns of the Northeast. Vermont Institute of Natural Science. Woodstock, Vt. WAGNER, ~. - JR. and F. S. WAGNER. 1982. The peta of Dryopteris poyseri. Michigan Bot. American Fern Journal 91(1):13—24 (2001) The Relictual Fern Genus Loxsomopsis MARCUS LEHNERT, MAREN MONNICH, THEKLA PLEINES, ALEXANDER SCHMIDT-LEBUHN, AND MICHAEL KESSLER Albrecht-von-Haller-Institut fiir Pflanzenwissenschaften, Abteilung Systematische Botanik, Untere Karspiile 2, D-37073 Gottingen, Germany ABSTRACT.—We conducted a systematic revision of the genus Loxsomopsis H. Christ (Loxomata- ceae) based on 99 herbarium sheets representing about 52 separate gatherings. Many morpholog- ical characters varied considerably, but as much of this variation was found within individual collections it was of little taxonomic relevance. Characters that varied between specimens but remained roughly constant within individual peor included the pubescence of lamina, pet- iole, and sorus, hair size and coloration, presence or absence of glaucous layers on the lamina surface, and spore surface structure. However, we found no correlation in the occurrence of two or more of these characters and were unable to morphologically define discrete species within the genus. Thus, we consider Loxsomopsis to include one variable species, L. pearcei (Baker) Maxon. The morphological variation is probably the result of the distribution of the species in small, isolated populations in ephemeral early successional habitats, leading to continuous population extinction and establishment of new population based on few propagules. Notes on ecology and distribution, full taxonomy, and specimen citations are provided. Loxsomopsis Christ is the Neotropical representative of the relictual and taxonomically isolated fern family Loxomataceae, whose other genus, Loxoma, is restricted to the North Island of New Zealand. Loxsomopsis is distinguished from Loxoma principally by its sporangia in which about two-thirds of the annulus cells are thickened (vs. only the apical cells thickened) and which open transversely (vs. longitudinally). After considerable debate as to the tax- onomic relationships of the family (Tryon and Tryon, 1982; Kramer, 1990), it is now believed to be related to the tree fern alliance, along with Cyatheaceae, Dicksoniaceae, Lophosoriaceae, and Metaxyaceae on the basis of anatomical, morphological, and molecular characters — and Given, 1979; Steven- son and Loconte, 1996; A.R. Smith, pers. co The genus Loxsomopsis was described by ‘Christ (1904) based on L. costar- icensis from Costa Rica, but the first species was actually described in 1891 by Baker as Dicksonia pearcei from Ecuador and was transferred to Loxso- mopsis by Maxon in 1933. Two further species have been described, L. leh- mannii from Ecuador by Hieronymus in 1904, and L. notabilis from Bolivia by Slosson in 1912. Because of the distinctness of the genus, there has been little debate about the generic delimitation and the placement of individual species. The delimitation of the species, however, is less clearcut and it has been suggested that only one species should be recognized (Tryon and Tryon, 1982; Tryon and Stolze, 1989). Overall, Loxsomopsis appears to be rare and/ or localized in the field and is poorly represented in herbaria. A modern treat- ment for the genus is lacking. The present study is an outcome of a course on pteridophyte taxonomy and systematics and is based on a morphological anal- 14 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) ysis of 99 herbarium sheets of Loxsomopsis representing about 52 separate gatherings from 15 herbaria. MORPHOLOGICAL CHARACTERS Stems.—Stems are terrestrial, long-creeping, dichtomously branched, dark brown, 1.5-5 mm in diameter, and with internodes 1-7 cm long. Scattered, dark, fibrous roots are produced from all sides. A physical relationship of the roots to the leaf bases was not noticed. Stems are densely pubescent with stiff bristles (defined as in Lellinger, 1985, by having pluriseriate bases and an uniseriate apex) that are dark brown, 0.8-2.4 mm long, basally pluriseriate with about 8-16 cells, and with an uniseriate apex consisting of 3—25 cells (Fig. 1A). The entire variation of bristle length and cell number can be found on individual stems. Presumably depending on the growth location, bristles can be erect or appressed, with both conditions as well as intermediates pre- sent on the same stem. The stems are solenostelic (Ogura, 1972). Root cortex anatomy corresponds to the Osmunda-type (Schneider, 1996), with a hypo- dermis of sclereids and inner cortical layers formed by relatively thin-walled cells. Petioles, rhachises and costae.—Petioles are adaxially convex to flattened- sulcate and 0.6-4 mm in diameter. Coloration varies from dark blackish or reddish brown to pale olivaceous brown, with the proximal portion generally being darker. The proximal 1-2 cm of the petioles are generally densely cov- ered in bristles resembling those of the stem. Scattered bristles are also found further along the petioles, especially in the groove. Frequently, petiole bristles have been abraded, leaving raised multicellular scars. Petiole length comprises about 45% to 65% of the entire leaf length. A single, gutter-shaped vascular bundle arrangement is present in the petioles (Kramer, 1990). Rhachises are similar to the petiole, but tend to be somewhat more strongly sulcate, the groove more pronounced in dry than in fresh material. Rhachis coloration is generally paler than that of the petiole and is markedly darker on the adaxial than on the abaxial side. The costae, when dry, are rounded to slightly sulcate with a brown base and pale yellowish medial and distal por- tions. Scattered hairs (defined as in Lellinger, 1985, as being uniseriate throughout), resembling those of the lamina, occur occasionally on the costae. Laminae.—Leaves are monomorphic, catadromous, subcoriaceous, circin- nate in the bud, and with lamina length varying from 0.2 to 1.5 m. Leaf lengths of 5 m reported by Tryon and Tryon (1982) and Tryon and Stolze (1989) were not evident in herbarium material available to us nor observed in the field (M. Kessler, pers. obs., M. Lehnert, pers. obs.). Leaf dissection varies from 1-pin- nate-pinnatifid on small leaves to bipinnate-pinnatifid on the largest leaves (Fig. 2). The pinnae are (sub)sessile, well-spaced, subopposite, asymmetric, with the proximal pinnae generally being the longest, reaching about 50% of the lamina length resulting in a narrowly to broadly deltate lamina outline. Secondary pinnae and segments are obliquely ascending, decurrent, with cre- nate to entire margins, and with the ultimate segments deltate-lanceolate. On LEHNERT ET AL.: LOXSOMOPSIS 15 Fic. 1. Morphological details of Loxsomopsis pearcei. A) Stem bristles, showing variability in size and cell number, Moran 3057; B) Variability of lamina surface hairs, Moran 3057; C) Hairs from indusium, Moran 3057; D) Young sorus with short receptacle and sporangial stalks; left side rangial stalks; left side of glabrous indusium removed, Lent 738; F) Sporangium, showing the subvertical annulus with about % of the cells indurated, Ollgaard 74387; G) Detail of pinna, show- ing free, forked-subpinnate venation, sorus location on vein ending at lamina margin, and pubes- cence on adaxial side, Goldgewicht 86369. 16 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) ZZ ~ nN Sin) LE We ‘ i wet ‘\ LGA4 er, c= — Alyy SLI, WS LPF \iieessss) J Fic. 2. Variability of lamina size, dissection, and shape. A) Distal portion of a large leaf, shown without the three proximal pinna pairs, Grayum 7212; B) Fully-developed, fertile, mid-sized leaf, Prieto 259; C) Fully developed, sterile, very small leaf, Ollgaard 90634B; D) Mid-sized leaf with notably differing segment shape, Lehmann 5715; E-G) Proximal pinnae showing differences in degree of dissection, segment orientation, and shape, Lehmann 5059, Goldgewicht 86369, Gold- gewicht 863. LEHNERT ET AL.: LOXSOMOPSIS a7 fertile leaves, acroscopic first- and second-order segments are larger than their basiscopic counterparts, whereas on sterile leaves they can occasionally be of similar size. Degree of lamina dissection, shape of pinnae and pinnules, and orientation of the pinnae vary considerably between leaves and individuals (Fig. 2). Smaller leaves tend to have longer, narrower, ascending pinnae, whereas larger leaves tend to have broader, more spreading pinnae. During lamina development, segments tend to be more rounded than on mature leaves. Veins are free, forked-subpinnate and catadromously branched (Fig. 1). Laminae of Loxsomopsis can be glabrous to densely pubescent abaxially with uniseriate, up to 2 mm long, light to brown, tortuous, septate hairs (Fig. 1). The frequency and coloration of this vestiture varies geographically, with Costa Rican specimens characterized by long, pale hairs whereas specimens from some Ecuadorean locations have dark brown hairs. Bolivian specimens tend to have a glaucous abaxial lamina surface which varies among and within collections (Tryon and Stolze, 1989). Field observations have shown that older laminae are more glaucous than younger ones (M. Lehnert, pers. obs.). Adax- ially the laminae are always glabrous. Sori.—Sori of Loxsomopsis are located marginally on veins and are directed downward from the lamina surface. Generally, a single sorus is located on the acroscopic side of each terminal pinnule or ultimate segment, although ocas- sionally a second sorus is found basiscopically. Sori are narrowly cyathiform to urceolate and have a laterally free and protruding indusium with an entire rim. Superficially, they resemble sori of the Hymenophyllaceae. Indusia are glabrous to sparsely pubescent with hairs (Fig. 1D) resembling those of the laminae (Fig. 1C). Sporangia and paraphyses are borne on a columnar receptacle whose basal portion elongates during sorus maturation, eventually measuring as much as 50% of the entire indusium length (Fig. 1D, E). As a result, young sori have fairly narrow indusia which conceal most of the sporangia, whereas in mature sori the majority of sporangia protrude beyond the more open indusium. Ripe sori are about 1 mm wide and 1.75—2.75 mm long. Paraphyses are uniseriate, unbranched, and without thickened cells, 0.5—-0.8 mm long, and consist of 5— 10 cells which are about 2—3 times longer than wide. Sporangia.—The number of sporangia per sorus varies from (10)15 to 30(50), with much of this variation found on single leaves. Sporangia are borne on short, ca. 6-rowed stalks that lengthen somewhat during sorus maturation. Sporangia are more or less pyriform (Fig. 1F), about 350 »m wide and 450 pm long, with very little variation in size among and between collections. The annulus is complete, almost vertical, bypasses the stalk, and consists of 30 to 55 cells. Annular cells are indurated, except for a group of laterobasal cells which are smaller and have thinner cell walls. These thin-walled cells make up about 1/4 of the annulus (Fig. 1F). Spores.—Spores of Loxsomopsis are trilete, tetrahedal-globose with promi- nent angles, with the laesure 1/2 to nearly equal the spore radius, and lack chlorophyll. Spores are 41-62 ym in diameter, with much of this variation found in single sori, e.g., 41-60 pm (n = 20, mean = 54 wm) measured in 18 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) RS ar Fic. 3. Spores of Loxsomopsis pearcei. A) Verrucate spore with irregular granular deposit, Gold- gewicht 86369; B) Slightly pitted spore, Moran 3051; C) Variability of surface structuring in ver- rucate spores, Ollgaard 9479. Goldgewicht 86369. In three cases (Blies 2119, Harling 24597, Lent 738) the number of spores per sporangium was determined to be 128. Spore surface is slightly pitted (Fig. 3B) to strongly verrucose (nomenclature after Lellinger and Taylor, 1997) with distinct humps or longish, coalescent ridges (Fig. 3A; see also Tryon and Lugardon, 1991:213). These structures are formed by a two- layered perispore (Tryon and Lugardon, 1991). Spores derived from a single specimen are generally similar, but there is some variation in the degree of coalescence of the tubercules (Fig. 3C). The spores are often, but not always, covered with an irregular granular deposit (Fig. 3A). Gametophytes.—Gametophytes of Loxsomopsis were described by Stokey and Atkinson (1956). They are epigeal, chlorophyllous, obcordate to somewhat elongate, centrally thickened, with thin margins, and bear long, multicellular hairs. Archegonia are borne on the lower surface of the thickened portion, especially apically, with the several- to many-celled antheridia located on the lower surface, and less often on the upper surface. Chromosome number.—Loxsomopsis has been reported as n = 46 from Cerro Vueltas, Costa Rica, by Wagner (1980). TAXONOMIC CONSIDERATIONS Numerous morphological characters exhibited considerable variation. Many of these characters varied as much within as between collections and were thus of little taxonomic relevance. Such characters include overall leaf size, stem and petiole diameters, relative petiole length, degree and shape of lamina dissection, size of the stem and petiole bristles, the number of annulus cells, the number of sporangia per sorus, and spore size. Characters that varied between specimens but remained roughly constant within individual gatherings included the pubescence of lamina, petiole, and sorus, hair size and coloration, presence or absence of glaucous layers on the lamina surface, and spore surface structure. However, we were unable to find LEHNERT ET AL.: LOXSOMOPSIS 19 N=8 N=8 N22 N=3 N=4 N=2 N=1 N=6 N=4 N=3 Me N=1 N=2 N=1 N=3 N=? Fic. 4. Geographical distribution of several variable morphological characters among 52 gather- ings of Loxsomopsis pearcei from 21 localities, indicating the number of gatherings per site (N). Characters and character states: A) Lamina pubescence: white: glabrous, gray: slight, black: dense smooth to slightly pitted, black: verrucose. Due to sterile or incomplete material, not every gath- ering has been considered for each character. The open circle shows the locality of a specimen not studied by us (Palacios and Clark 12463 (MOQ)). a correlation among these independent characters. There were some noticeable geographic trends, but exceptions were also found (Fig. 4). For example, Costa Rican specimens are characterized by rather long, pale and frequently abun- dant lamina hairs, but one specimen (Evans 2660A) was completely glabrous. Furthermore, the vestiture of some Ecuadorian and Peruvian collections re- 20 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) sembled that of the Costa Rican specimens. Dark brown hairs were restricted to Ecuadorian populations, but occurred in only about 60% of the specimens examined. Bolivian specimens were all characterized by a glaucous abaxial lamina surface. Pubescent indusia were found only in Ecuador, but not among all specimens from the region. The conspicuous differences in spore surface structure were specimen-specific and not related to the developmental stage of the spores, as has been reported, e.g., among Polypodiaceae (van Uffelen, 1997). None of these geographical trends were correlated with variation in other independent characters. Overall, the apparent independence of character variation precludes the delimitation of multiple taxa within Loxsomopsis. A comparable situation of a variable species which cannot convincingly be di- vided into discrete species is presented by Plagiogyria semicordata (Zhang and Nooteboom, 1998). While cytological or molecular analyses are lacking, we found no evidence to suggest that different ploidy levels are represented within Loxsomopsis. Spore size varied as much within as between individuals and we found no individuals with consistently malformed or abortive spores that would indi- cate hybridization between plants of different ploidy levels. We interpret the considerable morphological variation among different pop- ulations of Loxsomopsis to be a result of the localized and ephemeral nature of the growth sites of the species and its overall rarity. Since Loxsomopsis occurs mostly on disturbed sites such as landslides and roadsides, new pop- ulations are presumably regularly formed and then quickly lost as conditions are altered during vegetation succession. The result is a dynamic metapopu- lation structure in which new populations are frequently formed by one or few spores (Eriksson, 1996). Such a population structure would favor genetic drift and would explain the independent variation of morphological charac- ters. In summary, we have been unable to detect any character combinations that would permit the delimitation of different taxa within Loxsomopsis. While we cannot rule out the possibility that Loxsomopsis includes cryptospecies, we have found no evidence to support this assumption and consider the genus to include a single variable species, Loxsomopsis pearcei. GEOGRAPHIC DISTRIBUTION Loxsomopsis is known from Costa Rica, Colombia, Ecuador, Peru, and Bo- livia (Fig. 4). In Costa Rica, Loxsomopsis is confined to three separate popu- lations, corresponding to the volcanoes Poas and Barva, and to the Cordillera de Talamanca. In Colombia, specimens have been obtained from the eastern Andean chain near the Venezuelan border and from both sides of the Andes close to the Ecuadorean boundary. While not yet known from Venezuela, the species may well occur there. In Ecuador, there is one specimen of Loxsomop- sis from the northwestern Andean slope, but most collection come from the central and southern portion of the eastern Andean slope, where it has been collected along all major roads traversing the region, and presumably occurs LEHNERT ET AL.: LOXSOMOPSIS 21 as one fairly large, nearly continuous population. The species appears to be absent from the botanically rather well-known northern portion of the eastern versant. In Peru, Loxsomopsis has been collected at three scattered localities in the central portion the country. Its absence from northern Peru, including several well-collected localities in the departments Amazonas and San Martin may well represent a real distributional gap. In contrast, the lack of collections from southern Peru may be a result of the low collecting activity in the region. In Bolivia, there are three known localities for Loxsomopsis, two of which, the La Paz-Coroico and the Cochabamba-Villa Tunari roads, correspond to the most intensively studied areas in the country. We suspect that Loxsomopsis occurs in intermediate areas as well. ECOLOGY Loxsomopsis has been collected between about 1950 m and 2700 m in Costa Rica, and between 1800 m and 3600 m in the Andes. The majority of speci- mens have been gathered from roadside and forest edge habitats, but numerous specimens from Volcan Pods in Costa Rica have been found along the margin of the volcanic crater lake. Tryon and Tryon (1982) cite open habitats in ra- vines, on brushy slopes and open woodland as growth sites for Loxsomopsis. In the Andes, Loxsomopsis occurs on slopes disturbed by landslides and along roadsides, growing syntopically with aggressive colonizing ferns of the genera Sticherus (Gleicheniaceae) and Hypolepis (Dennstaedtiaceae) with which it forms dense tangles (Leén and Young, 1996, M. Kessler and M. Lehnert, pers. obs.). Leaves of Loxsomopsis are sometimes long with a pendent apical portion and lean on adjacent shrubby plants (Tryon and Tryon, 1982). Loxsomopsis has been collected on a wide variety of geological substrates, ranging from volcanic soils in Costa Rica to sandy soils and lutites in the Andes Generally, Loxsomopsis appears to be rather rare. In Bolivia, it has been recorded only four times in a total of about 300 vegetation plots (of which approximately 90 covered suitable early- to mid-successional vegetation) lo- cated in the right elevational belt within the distributional area of Loxsomopsis (M. Kessler, unpubl. data). In southeastern Ecuador, Loxsomopsis is locally common on roadsides (B. Ollgaard, pers. comm.). TAXONOMIC TREATMENT Loxsomopsis H. peek pon Herb. Boiss. II, 4:399. 1904.—Type: Loxsomop- sis costaricensis H. C A monotypic genus, as the characteristics of its single species. Loxsomopsis pearcei (Baker) Maxon, Proc. Biol. Soc. Wash. 46:105. 1933. Dick- sonia pearcei Baker, Ann. Bot. (Oxford) 5: 197. 1891. Dennstaedtia pearcei (Baker) H. Christ, Index fil. 218. 1905.—Type: Ecuador, “Eastern Andes, 8000—9000ft.” Pearce 251 (Holotype, K!; Photo, US). 22 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) Loxsomopsis costaricensis H. Christ, Bull. Herb. Boiss. II, 4:399. 1904.—Type: Costa Rica, Wercklé 279 (Holotype, P Loxsomopsis lehmannii Hieron., Bot. Jahrb. Syst. 34:435. 1904.—Type: Ecua- dor, [Morona-Santiago,] prope Chinguinda, declivibus montium Cordillera Oriental de Sigsig, 1800-2500m, Lehmann 5061 (Isotype, K!, US!). Ac- cording to Tryon and Stolze (1989) holotype destroyed in LZ. Loxsomopsis notabilis Slosson, Bull. Torr. Bot. Cl. 39:285. 1912.—Type: Boliv- ia, [La Paz, Franz Tamayo,] near Apolo, 6000 ft., Williams 1303 (Holotype, US!; Isotypes, P!, US!). Terrestrial fern with long-creeping, slender to rather stout, branched, sole- nostelic stem bearing stiff, dark brown, multicellular, basally pluriseriate bris- tles. Leaves remote, non-articulate, 0.3-1.5 m long. Petiole as long as to some- what longer than lamina, with a single, gutter-shaped vascular bundle, adax- ially convex to flattened-sulcate, with bristles at the base and sometimes fur- ther along the petiole. Lamina narrowly to broadly deltate, 1-pinnate- pinnatifid to bipinnate-pinnatifid, firm, catadromous, hypostomatic; pinnae (sub)sessile, subopposite, asymmetric, the basiscopic segments reduced, sec- ondary pinnae and segments obliquely ascending, decurrent, ultimate seg- ments deltoid-lanceolate, margin crenate to entire. Stomata paracryptic. Veins free, forked-subpinnate, catadromously branched. Sori terminal on a vein, pro- truding beyond the margin, directed downwards from plane of lamina, pro- fusely paraphysate; sorus cyathiform to urceolate, laterally free; receptacle co- lumnar. Sporangia on short, pluriseriate stalks; annulus almost vertical, con- tinuous and 3/4 indurated. Spores trilete, 41-62 j.m in diameter, slightly pitted to verrucate, mostly granulate. SPECIMENS EXAMINED Costa Rica. Alajuela: Shore of lake on Pods Volcano, 2500 m, Lent 512 (F); Near edge of lake, Pods Volcano, 2500 m, Lent 738 (F); Along shore of old crater lake of Volcan Pods, 2500 m, Mickel 3654 (NY); Pods lake, 8500 ft., in sandy soil by lake shore, Stork 2522 (US); Edge of small lake inside of an extinct sub-crater of Volcdn Pods, about 11 km NW of Heredia, 2700 m, Taylor 4496 (NY); Shore of lake, crater summit, Volcdn Pods, 2600 m, pial 4624 (UC). Heredia: Volcan Barba, 6700 ft., Evans 2660A (MO, U, Z); Headwaters of Rio San psi marik ca. 3 km NE of San Rafael de Vara Blanca, N slope of Volcén Barba, 10° 11.5’ N, 84° 07’ W, 2060 m, Grayum 7212 (MO, US); Bordes del Rio Grande, afluente del Rio Patria, en el Paso del Cais macizo del Volcan Barba, 2045 m, Jimenez 2272 (F, NY, UC, US); Southeast shoulder, near end of Route 113, 6 miles NNE of San Rafael de Heredia, Lloyd s.n. (MO, UC); San Jose, 10 km N of San sri de Heredia on Volcan Barba, 1950 m, Mickel 3001 (UC, US); Rio de las Vueltas, 2100 m, Goldgewicht 863 (US); En el Rio de las Vueltas, 2000 m, oe ae 86369 (CR); End of rt. 113, where road crosses Rio Patria about 100-200 m upstream, 2000 m, Moran 3051 (AAU). Limén: N flank of Cerro Casma, along Ujarras—San José Cabécar trail, Cordillera de Talamanca, 09° 20’ 30’’ N, 83° 13’ 30’’ W, 2250-2270 m, Grayum 10329 (CR, F, MQ); P.I. La Amistad Valle de Silencio, 09° 07’ 15’’ N, 82° 57’ 55’’ W, 2450 m, Moraga 339 (CR, UC). Prov. unknown: Wercklé 225 (P). Colombia. Narifio: Municipio de San Francisco, carretera Pasto-Mocoa, entre El Mirador y San Francisco, 1500-2200 m. Mora 4449 (COL). Norte de Santander: Municipio de Toledo, de Toledo LEHNERT ET AL.: LOXSOMOPSIS 23 a Samore, 2200 m. Lozano et al. 5472 (COL). Putumayo: road from San Francisco to Mocoa, km. 112 from Pasto, near El Mirador, 7200 ft. Plowman 4352 (COL). UADOoR. Morona-Santiago: Limén-Gualacco Road, km 16, 2100-2200 m, montane forest, Harling 600 (AAU, QCA); Muletrack Sevilla de Oro-Mendez, eastern slope (E of Cerro Negro, “» Castillo), montane forest along muletrack, 78° 34’ W, 2° 47' S, 2800-2950 m, @llgaard 9479 (AAU, F); Sigsig, 00 iente” Border: Eastern Conditions, between Ofia and the Rio Yacuambi, east es pabenten pee t., Prieto 259 (F, GH, K, MO, NY, US). Zamora-Chinchipe: Road Loja-Zamora, km 1 ° 09’ W, 4 S, 2750-2770 m, Situs Midaen 3860 (AAU, F, MO, NY); Eastern slopes of the ac ie of Loja, just E of the pass, wee scrub along old trail, 79° 10’ W, 03° 58’ S, 2800 m, @llgaard 98820 (QCA); Strasse von Loja nach Zamora, Paéramo del Nebula, Pass an der Grenze von Loja und Zamora- Chinchipe, 3600 m, Schneider 579 (Z); 10 km outside Loja along new road Loja-Zamora, just past = Nudo, 2900 m, van der Werff 9107 (AAU, MO, NY, UC); pass on road Loja-Zamora, ca. 15 km of nies ca. 3005 m, van der Werff 13456 (MO); Parque Nacional Podocarpus, Road Yangana- caren km 26, 79° 09’ W, 04° 29’ S, 2550 m, Madsen 75754 (AAU, QCA); Parque Nacional ecleiiord — Yangana-Valladolid, just S and E of the pass (Nudo de Sabanilla), muletrack ard Quebrada Honda, 78° 08’ W, 4° 27’ S, 2750-2950 m, @llgaard 91061 (GOET), Oligaard pores (AAU, QCA). Santiago-Zamora: Trail between Pailas and E] Pan, 2255-3445 m, Steyermark 54324 (F, NY, US). (Carchi: Espejo, El Goaltal, Cerro Golondrinas, vegetacion arbustiva (matorral) espeso sobre la cresta de la montafia, bosque muy humedo montano bajo, 0° 51'N, 78 08’ W, 2600-2800 m, Palacios and Clark 12463 (MO); not seen by us, but studied by A.R. Smith, pers. com.). PERU. Huanuco: pergeimny slope of the Rio Llulla Pichis watershed, on the ascent of Cerras del Sira, 9° 25’ S, 7 ' W, 2205 m, Dudley 13420 (GH); ni tal ca. 9000 ft., Macbride 4521 (F, GH, US). ey By sfiilaes Prov. Oxapampa and Pasco, 2700 m, van der Werff 8564 (MO, NY). Cuzco: Cachayocpata, entrada del Valle San Miguel La yeaa 10000 ft., Biies 2119 (US); La Convencion 73° 34’ W, 12° 37’ S, 2555 m, Dudley 10713 (GH, MO). BOLI Paz: Nor Yungas, 5 km de Chuspipata hacia Coroico, 16° 23’ S, 67° 48’ W, 2750 m, pricing 11 i (UC, GOET); Nor Yungas, Chuspipata, camino sobre loma de montafia a Coroico, Ss, 9’ W, 3000 m, Lehnert 032 (LPB, UC, GOET); Nor Yungas, entre Cotapata y Chus- aaah cee fe is 67° 50’ W, 3100 m, Smith 13139 (F, MO, UC). Cochabamba: Jo ohana Torrico, 112 km antigua carretera Cochabamba-Villa Tunari, 17° 07’ S, 65° 38’ W, 2700 m, Kessler niet (UC); José Carrasco Torrico, 130 km antigua carretera Cochabamba-Villa Tunari, 17° 07’ S, 65° W, 2000 m, Kessler 7209 (UC, GOET). ACKNOWLEDGEMENTS We thank Erika Hofmann, Ulrike Geisler and Tanja Siemsen for collaboration on specimen anal- ysis, Gabriele Weiss for the SEM photographs, A. R. Smith for comments on the manuscript, and the following herbaria and their curators for making material available to us: AAU, COL, CR, F, GH, GOET, K, MO, NY, P, QCA, U, UC, US, Z (acronyms according to Holmgren et al., 1990). LITERATURE CITED Curist, H. 1904. Loxsomopsis costaricensis nov. gen. et spec. Bull. Herb. Boissier, sér. 2, 4:393- 400. 24 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) ERIKSSON, O. 1996. Regional dynamics of — a review of evidence for remnant, source-sink metapopulations. Oikos 77:248-25 HOLMGREN, P. K., N. H. HOLMGREN and L. ) BARNETT. 1990. Index Herbariorum. Part 1: The Herbaria ht World. New York Botanical Garden, Bronx KRAMER, K. U. . Loxomataceae. Pp. beta in K. “Kubitzki, ene ed. The Families and Genera “ ae Plants. Vol. I. Pt and P. moira oo Berlin LELLINGER, D. 'B. Field Manual of the Ferns and Fern-allies of the United States and Canada. cate Institution Press, Washington LELLINGER, D. B. and W. C. TAYLOR. 1997. A classification ia spore ornamentation in the Pterido- Pp. 33-42 in R. J. Johns, editor. Holttum Memorial Volume. Royal Botanic Gardens, phyt ] ds., K. U. Kramer Kew. LEON, B. and K. R. YOUNG. 1996. Distribution of pteridophyte diversity and endemism in Peru. Pp. 77-91 in J. M. Camus, M. Gibby, and R. J. Johns, editors. Pteridology in Perspective. Royal Botanic Gardens, Kew. KHAM, K. R. and D. R. GIVEN 1979. The flavonoids of ferns in the isolated genera Loxoma cunninghamii and Loxsomopsis costaricensis. Biochemical Systematics and Ecology 7:91—93. Ocura, Y. 1972. Comparative anatomy of 5) ieee organs of the pteridophytes. Handbuch der anzenanatomie VII, 3. Gebr. Borntrager, B SCHNEIDER, H. 1996. The root anatomy of ferns: a etaabaitn study. Pp. 271-283 in J. M. Camus, M. Gibby, and R. J. Johns, editors. Pteridology in Perspective. Royal Botanic Gardens, — STEVENSON, D. W. and H. LoconTe. 1996. Ordinal and familial relationships of pteridopl era. Pp. 435-467 in J. M. Camus, M. Gibby, and R. J. Johns, editors. Pteridology in Perspective. Royal Botanic Gardens, STOKEY, A. G., and L. R. ATKINSON. 1956. The gametophytes vs orig cunninghamii R. Br. and 61. Loxsomopsis costaricensis sce Phytomorphology 6: TRrYON, A. F. and B. LUGARDON. 1991. Spores of the Peochchivin Springer-Verlag, Berlin. TRYON, R. M., and R. G. STOLZE. ee Pteridophyta of Peru. Part I. Fieldiana Botany, New Series, 20:1—145. TRYON, R. M., and A. F. TRYON. 1982. Ferns and allied plants with special reference to tropical America. sig ot gre Berlin VAN UFFELEN, G. A 7. The spore mall) in Pe Sianeipr development and evolution. Pp. 95— 117 inR. J. nibait editor. Holttum Memorial Volume. Royal Botanic Gardens, Kew. WaGnéeR, F. S. 1980. New basic chromosome numbers files genera of neotropical Sau Amer. J. Bot. 67:733—738. ZHANG, X. C. and H. P. NOOTEBOOM. 1989. A taxonomic revision of Plagiogyriaceae (Pteridophyta). Blumea 43:401-469 American Fern Journal 91(1):25—35 (2001) Growth, Leaf Characteristics, and Spore Production in Native and Invasive Tree Ferns in Hawaii LEILANI Z. DURAND AND GUILLERMO GOLDSTEIN University of Hawaii at Manoa, Department of Botany, St. John Plant Science Bldg., 3190 Maile Way, Honolulu, HI 96822 ABSTRACT.—The Australian tree fern Sphaeropteris cooperi is an invasive species in Hawaiian wet forests where it displaces Cibotium, the dominant native Hawaiian tree fern, where they co- occur. This study was undertaken in order to assess the relative growth rates and sapeoductive potential of S. cooperi and the native Cibotium species. Field measurements of rates, fertile frond production and leaf traits were made monthly over the course of one ye ar. scierh aead cooperi had a significantly higher growth rate, both in terms of height increase and frond pro tion, and maintained four times more fronds than the native Cibotium species. The mean sve height increase of the invasive tree fern was 15 cm compared to 2 to 3 cm for the native tree ferns. The leaf mass per area of S. cooperi was significantly lower than that of the native Cibotium species, and the leaf life span was significantly shorter, suggesting that the cost of construction of more fertile fronds per month than the native tree ferns. These differences in 2 history charac- teristics may help explain the rapid spread and success of S. cooperi in Hawai The invasion of non-native plants into native ecosystems has become a topic of great concern in recent years, particularly in isolated island ecosystems such as the Hawaiian Islands (Loope and Mueller-Dombois 1989, Vitousek et al. 1987). Of the more than 800 introduced plant species which have become naturalized in the Hawaiian Islands, 30 are ferns (Vitousek et al. 1987, Wagner 1995). The Australian tree fern, Sphaeropteris cooperi (Hook. ex F. Muell.) Tryon [syn. Cyathea cooperi (Hook. ex F. Muel.) Dom.], was introduced to the Hawaiian Islands as a horticultural plant and first escaped from cultivation in the 1950s (Wagner 1995). Sphaeropteris cooperi is now naturalized on the islands of Oahu, Maui, Kauai, and Hawaii (Wagner 1995) and is listed as among the worst alien plant invaders of Haleakala National Park (Loope et al. 1992). Tree ferns in the genus Cibotium are the dominant native tree ferns in Ha- waii. There are four endemic species in this genus, three of which (C. cham- issoi Kaulf., C. menziesii Hook., C. glaucum [Sm] Hook. and Arn.) are found on all of the major islands. The fourth (C. nealii Degener) is found only on the island of Kauai (Palmer 1994). These endemic ferns grow in semi-wet to wet forests from an elevation of 40 m to 1800 m, and at a mean annual temperature range between 13°C and 23°C (Becker 1976). In the Hawaiian forests, tree ferns are keystone species which have a substantial impact on the structure and function of the forest around them. For example, the native Cibotium support native Hawaiian plants as epiphytes (Medeiros et al. 1993), as well as serving as nurse logs for a number of native species which become established on the well-aerated adventitious roots of the tree ferns’ trunk (Buck 1982). The native 26 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) TABL Elevation (m), annual precipitation (mm), tree fern species present, and the number of indi- wile studied (n) at the sites used for this study. Lyon Arboretum , Wa’ahila Ridge and Mt. Ka’ala are on the island of O’ahu, Waineke Swamp is on the island of Kauai Sample size Site Elevation (m) Rainfall (mm) Species (n) Low elevation 155 4000 Sphaeropteris cooperi 6 (Lyon Arboretum) Cibotium chamissoi 8 Mid elevation 430 2000 Cibotium chamissoi 6 (Wa’ahila Ridge) Cibotium menziesii 6 High elevation 1225 2000 Cibotium chamissoi 8 (Mt. Ka’ala bog) Cibotium menziesii 8 Cibotium glaucum 8 High elevation 1250 2000 Sphaeropteris cooperi 6 (Waineke Swamp) tree ferns also sequester high levels of nitrogen and phosphorus in their leaves compared to other native Hawaiian rainforest species (Vitousek et al. 1995). This study was undertaken in order to assess the relative growth rates and reproductive potential of the native Cibotium and the invasive Sphaeropteris cooperi. We hypothesize that the invasive S. cooperi has faster growth rates and a greater spore production than the native tree ferns. In addition, we hy- pothesize that the invasive tree ferns produce fronds and spores more consis- tently throughout the year than the native ferns. This difference in growth and reproductive capacity may contribute to the ability of S. cooperi to outcompete the native tree ferns in Hawaiian rainforests. We also predict that the invasive tree fern, consistent with what has been observed in other invasive species in Hawaii (Baruch and Goldstein 1999), will have a shorter leaf life span and lower leaf mass per area (LMA) than the native tree ferns. The native Cibotium species were chosen for comparison based on their similar growth form and phylogenetic relatedness to the Australian tree fern. By studying native and invasive species of similar growth form that grow in the same habitat, predic- tions as to the relative growth and reproductive performance of the invasive and native plants can be drawn. MATERIALS AND METHODS STUDY SITES.—Four sites were selected for this study based on accessibility and species composition (Table 1). The range of elevations of the sites, from 200 m to 1250 m, encompasses much of the elevational range of the native Ha- waiian tree ferns. The tree ferns at each site were growing under similar sub- strate and environmental conditions. All sites were tree dominated with an understory of ferns and tree ferns. The tree ferns at the two high elevation sites were located in relatively well drained soils at the edge of a bog. Individual plants of the same size class (1.5 m to 2.0 m) were randomly selected, and com- parisons were made between individuals of the invasive and native tree ferns from the shaded understory. Six to eight individuals of each species at each DURAND & GOLDSTEIN: NATIVE AND INVASIVE TREE FERNS 27 site were studied (Table 1). The low elevation site was the only accessible site with sympatric native and invasive tree ferns, and the only site where the invasive tree fern was available for a long enough period of time to measure growth. The high elevation site on Oahu contained only native tree ferns. An additional site on Kauai with similar elevation and rainfall was selected for comparison. Sphaeropteris cooperi is naturalized at this site, but Cibotium chamissoi, C. menziesii, and C. glaucum were not found. Due to management efforts to control Sphaeropteris cooperi, this plant was not available at all sites. This fact, and the fact that sample sizes were relatively small, led us to be cautious when making extrapolations based on the results of this study. GROWTH MEASUREMENTS.—Stem growth rate, frond production, and leaf life span were monitored for 12 consecutive months, from November 1997 through October 1998, on mature sporophytes of the same size class at the Lyon Ar- boretum, Waahila Ridge and Mt. Kaala sites. It was not possible to measure growth and leaf life span on the high elevation population of S. cooperi as management efforts to control this plant resulted in their removal before mea- surements could be completed. A point at the soil surface on one side of each tree fern was marked and measurements of stem height from the meristem to this point were taken monthly. All newly emerging fronds were flagged, dated and monitored until senescence. The number of months between frond emer- gence and senescence was recorded as the life span of the frond. LEAF TRAITS.—The total number of living fronds was counted monthly. For the purposes of this study, a frond was defined as living when it had more than 10% green tissue remaining. The mean surface area per frond was esti- mated by multiplying the total number of pinnules per frond by the average surface area of a pinnule. Ten pinnules per frond were collected, and the sur- face area of each pinnule was measured with a Licor 3000A area meter (Li- Cor, Inc., Lincoln, Nebraska) and averaged. The surface area of the tip of each frond was measured separately, as the pinnules in this region were smaller. The average surface area per frond was multiplied by the number of mature living fronds per plant to estimate the total leaf surface area per Tissue samples from the middle of a frond were collected in the field for leaf mass per area (LMA) estimates. Samples were kept in a cooler for less than 24 hours until they could be returned to the laboratory, at which time the area was measured using a LI-3000A area meter. The leaf samples were then dried at 60°C for five days and weighed to determine LMA (the ratio of leaf dry weight to leaf surface area). The number of fronds with spores was counted each month from November 1997 through October 1998. Each frond was examined to insure that sori were intact, and no fronds with dehisced sporangia were counted as fertile. STATISTICAL ANALYSIS.—A one-way ANOVA among all the samples was per- formed, followed by pre-planned contrasts (Moore and McCabe 1993). Con- trasts were used to assess for significant differences in LMA and leaf life span between native and invasive species at low elevation and at high elevation, 28 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) and between all natives and the invasive regardless of elevation. Monthly frond production was standardized (number of new fronds produced in a month divided by the maximum number of fronds on the plant). The stan- dardized monthly frond production was used to calculate a standard error of frond production for each species at each site. This standard error was used as a measure of seasonal variation in leaf production (Stratton et al. 2000). The same was done for spore production. Repeated measures ANOVA’s were used to analyze yearly data on new frond production, leaf surface area, height growth increase, and spore production. A Tukey’s pairwise test was used to determine significance between species. RESULTS The mean height growth of the invasive tree fern S. cooperi over the course of one year was significantly greater (P<0.001) than the mean height increase of native tree ferns (Fig. 1). After 11 months of measurements, the mean growth of the invasive tree fern was 15.4 cm, while that of the native tree ferns ranged from 1.9 cm to 3.0 cm. The most valid comparison is between S. cooperi and C. chamissoi at the same site, where the native tree fern had a mean annual growth of 3.0 cm compared to the 15.4 cm of the invasive tree fern (P<0.001). There were no substantial site-specific differences in height growth among the native tree fern species. An increase in growth during the months of April to August corresponded to an increase in rainfall during this period (Fig. 1). The number of new fronds produced per month by S. cooperi was signifi- cantly greater (P<0.001) than the number of new fronds produced by the na- tive tree ferns (Fig. 2). Sphaeropteris cooperi produced an average of 2.7 new fronds each month, while the native tree ferns at the same elevation produced an average of 0.4 new fronds per month, and all native tree ferns regardless of elevation produced an average of 0.2 to 0.4 new fronds per month. The native tree ferns also exhibited greater seasonal variation in frond production than S. cooperi (Table 2). Seasonality, calculated as the standard error of the stan- dardized monthly frond production, was 2.5 to 3.5 times greater in the native tree ferns compared to S. cooperi. A peak in new frond production was ob- served for the native tree ferns in April, and again in August/September (Fig. 2), however no corresponding peak in leaf surface area per plant was observed as leaf senescence often corresponded with leaf expansion. Sphaeropteris cooperi maintained from 12 to 18 fronds per plant over the course of one year. The native tree ferns at all elevations maintained only 3 to 6 fronds per plant over the same time period (data not shown). As a consequence, S. cooperi had a higher leaf surface area per plant than the native tree ferns (Fig. 2). The fronds of low elevation S. cooperi had a significantly lower LMA than the native tree ferns growing at the same elevation (P<0.001). Similarly, high elevation S. cooperi had a significantly lower LMA than that of native tree ferns growing at high elevation (P<0.001). The fronds of S. cooperi had a 1.3 to 3.8 times lower LMA than that of the natives at all sites (Table 2). In addi- tion, the LMA of the native tree ferns tended to increase with increasing ele- DURAND & GOLDSTEIN: NATIVE AND INVASIVE TREE FERNS 29 60 | aM Low elevation [9 High elevation 30 - 20 + ai 3 0 : | , : fh I - fl H Precipitation (cm) | b C. chamissoi (H) Mean relative height (cm) Time (months) Fic. 1. Seasonal variation in rainfall at the low sees site (Lyon ine eae and high eleva- tion site (Mt. Kaala) on the island of Oahu (a), and mean relative height growth of native and invasive tree ferns (b) over the course of one year, hoe November 1997 Pasa October 1998. Tree fern height in November was used as a reference. Symbols represent mean + SE (n = 6 to 8). Negative values — a decrease | in n height as croziers emerge from the trunk. ee oe represent the invasive tree fer i, and open ti species in the genus Cibotinn : M, and H indicate low, middle, and high elevation Mindy ating. Precipitation data from the National Weather Service vation. The leaf life span of low elevation S. cooperi was significantly shorter than the leaf life span of the low elevation native tree ferns (P<0.001). The invasive tree fern had 45 to 50 percent shorter leaf life spans than the native tree ferns at all elevations (Table 2). The number of fertile fronds per plant was significantly greater (P<0.001) in S. cooperi than in the native tree ferns (Fig. 3). The invasive tree fern had from 9 to 15 fertile fronds per month, which represented 60 to 75 percent of the total number of fronds per plant. The native tree ferns had between 1 and 4 fertile fronds per month, which represented 20 to 60 percent of the total number of fronds per plant. In addition, the native tree ferns exhibited greater seasonal variation in fertile frond production than S. cooperi (Table 2). Sea- sonality, calculated as the standard error of the standardized mean monthly 30 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) 4.5 T T T _> T T T T T T 7. T 4.0 + 4 3.5} 4 3.0 + 4 q 2.5} ? & 20+ 4 e3 o Z 15+ 4 1.0} - 0.5+ 0.0 + b < 400} ; a) S 5 300 f ; a os 5 o 200+ 4 3 3 ‘s 100} : —] 0 Time (months) Fic. 2. Number of new fronds produced each month (a) and total surface area per plant (b), from November 1997 through October 1998. Symbols are mean + SE (n = 6 to 8). Symbols as in Fig. 1. Absence of error bar indicates that the SE is smaller than the symbol. number of fertile fronds produced per month, was between 20 and 50 percent higher in the native tree ferns (Table 2). DISCUSSION Native tree ferns on Oahu (this study) and the island of Hawaii (Walker and Aplet 1994, Wick and Hashimoto 1971) had relatively low annual height in- creases compared to the invasive tree fern. The Australian tree fern grew very little in January through March, when monthly precipitation was below 20 cm. Growth increased in April through August, apparently as a result of higher precipitation. Growth in the native tree ferns at the low elevation site also increased in April. At the high elevation site we noted that growth showed DURAND & GOLDSTEIN: NATIVE AND INVASIVE TREE FERNS 31 TABLE 2. Leaf mass per area (LMA), leaf life span, standard error (SE) of standardized leaf production de and of standardized s pore production for native and invasive tree ferns at three elevations. Standard error ard error of standardized al variability in spore production. Values are means + SE. Symbo (7) next to a grouping indicates statistical significance (P = 0.01) between the invasive and native ferns Pp within that grou Leaf life span SE of leaf SE of spore LMA (g m~”) (months) production production Low elevation Invader S. cooperi Be pena 6.0 O:15T 335 1.97 Native C. chamissoi 54.7 + 2.8 11.0 + 0.30 8.49 321 Mid elevation Natives C. chamissoi 509 + 3.1 12.0 + 0.71 11.7 3.66 C. menziesii 78.9 + 4.0 10.8 + 1.1 8.32 2.45 High elevation Invader S. cooperi 34.7 + LBF n/a n/a Natives C. chamissoi 1203) 7.2 10.6 + 0.42 9.02 3.46 C. menziesii 1275 = 38 12.0 + 0.50 8.72 2.63 C. glaucum 153.2 223:0 120 :220.73 8.94 4.04 less seasonal variation and may have been influenced by higher irradiance during drier months, when cloud cover is relatively low, as well as by rainfall atterns. Frond production of the native tree ferns also exhibited seasonal var- iation, with the majority of frond production occurring from February through April. There was virtually no frond production during the months of October, November, and December. This seasonal pattern was similar to the pattern of frond production observed in Cibotium on the island of Hawaii by Wick and Hashimoto (1971). In neither study did the pattern of frond production appear to correspond with precipitation. In contrast to the native tree ferns, frond production in the Australian tree fern was consistent throughout the year. The faster growth rate and greater and more consistent frond production of the Australian tree fern S. cooperi could contribute to its ability to outcompete the native Hawaiian tree ferns. When plants of the same growth habit are compared, plants with a higher leaf surface area tend to have a higher relative growth rate (Lambers et al. 1998). Sphaeropteris cooperi had over three times more fronds per plant than the native tree ferns, and over four times more leaf surface area (Fig. 2). A greater leaf surface area per plant, along with a greater annual height increase, could allow the Australian tree fern to intercept more light than the native tree ferns, and thus potentially fix more carbon. In a previous study, we ob- served that the photosynthetic rate at light saturation for shaded plants was between 5 and 7 pmol m~ s~ for the invasive tree fern, while the photosyn- thetic rate at similar light levels was between 3.4 and 4 pmol m~ s* for the native tree ferns (Durand and Goldstein 2001). The pattern of allocation of carbon to different plant organs can also affect the annual height increase of the plant. The native tree ferns have a high starch content in their trunk, an were once harvested for this starch (Ripperton 1924, Wick and Hashimoto 32 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) 16+ 14+ 5 ae 127 D & 10} n = oS Sr C= o = 6 a 4f 2+ $4 : L O Y y. = Z a aoe 1 | ==. —_: re 5s Re eM A UME A dA LS oa Time (months) Fic. 3. The number of fertile fronds (with spores) per plant of native and invasive tree ferns, measured from November 1997 through October 1998. Symbols represent + SE (n = 6 to 8). Symbols as in Fig. 1. 1971). While no data is available on the trunk starch content of S. cooperi, in areas where S. cooperi and C. chamissoi grow together, wild pigs preferentially feed on the trunk starch of the native tree ferns. Thus, another possible expla- nation for the low annual height increase of the native tree ferns is a relatively large allocation of carbon, in the form of starch, to the fern’s trunk. The native tree ferns exhibited seasonal changes in fertility, with the greatest number of fertile fronds present in the months of September through February. The invasive tree fern did not exhibit as much seasonal variation in fertility, though there were more fertile fronds per plant in the months of July through November. The difference in seasonal spore production may allow spores of the Australian tree fern to germinate at a time when spores of the native tree ferns are not available. It has been suggested that successful invasive species either utilize resources more efficiently, or at a time when natives are inactive or unable to access the resources (Vitousek 1986). Consistent monthly spore production could give the Australian tree fern a reproductive advantage over the native tree ferns. Ferns in the genus Sphaeropteris produce 64 spores per sporangium (Gas- tony 1974), as do ferns in the genus Cibotium (Sporne 1966). Dyer (1979) es- timated that an average fertile frond of Cibotium chamissoi produces 700,000,000 spores. With between 1 and 4 fertile fronds produced per plant per year (Fig. 3), spore production per year ranges between 700 million and 2.8 billion spores. Sphaeropteris cooperi produced between 22 and 27 fertile fronds per plant per year (Fig. 3), though the number of sporangia produced per frond is not known. However, this species has a greater frond surface area than Cibotium, and given the equal number of spores per sporangia in Cibo- tium and Sphaeropteris, assuming equal spore production per frond seems a DURAND & GOLDSTEIN: NATIVE AND INVASIVE TREE FERNS 33 conservative estimate. If S. cooperi does indeed produce an equal number of spores per frond as Cibotium chamissoi, this fern has the potential to produce 15.4 billion to 18.9 billion spores per year. While both species produce a large number of spores each year, S. cooperi can potentially produce 13 to 16 billion more spores per year. Studies of the spore production of the invasive tree fern, and spore viability of both the native and invasive tree ferns, need to be con- ducted in order to have a more complete understanding of the relative repro- ductive capacity of these ferns. Variation in leaf life span has been found to be an important predictor of numerous plant responses (Chabot and Hicks 1982, Reich et al. 1992, Reich et al. 1997). The leaf life span of higher plants can range from less than one month to 25 years (Chabot and Hicks 1982). When compared across diverse ecosystems and biomes, a short leaf life span is associated with factors such as higher photosynthetic rate, higher leaf nitrogen content, lower leaf construc- tion cost, and a lower LMA (Chabot and Hicks 1982, Reich et al. 1992, Reich 1993, Reich et al. 1997). The mean leaf life span of the Australian tree fern was significantly shorter than the mean leaf life span of the native tree ferns (Table 2). As expected for a plant with a shorter leaf life span, the Australian tree fern had a significantly lower LMA than the native tree ferns (Table 2). Generally, longer-lived leaves require more secondary compounds to protect the leaves against herbivores and pathogens, whereas shorter-lived leaves with a low LMA are able to maximize photosynthetic return per unit weight of the leaves (Lambers et al. 1998, Reich 1993). The Australian tree fern produced short-lived leaves with high leaf nitrogen and high photosynthetic capacity (Durand and Goldstein 2001) at a relatively low carbon cost to the plant com- pared to the native tree ferns. In a recent study of 34 native and 30 invasive higher plants in the Hawaiian Islands, it was found that the latter had lower LMA, higher leaf nitrogen, and higher photosynthetic rates (Baruch and Gold- stein 1999). The patterns observed with the ferns in this study were consistent with the pattern found in higher plants, suggesting that this suite of traits are important determinants of the success of invasive species in the Hawaiian Islands, for both ferns and higher plants alike. The difference in leaf life span, LMA, growth rate, and spore production between the invasive and native tree ferns suggests that there are differences in life history characteristics. Tree ferns in Costa Rica that grew as pioneer species in secondary forest had higher leaf turn-over rates, faster growth rates, and more rapid production of sori than tree ferns that grew in primary forest (Bittner and Breckle 1995). Similarly, the high leaf turn-over rate, high spore production, and rapid growth rate in S. cooperi suggests that it tends to behave as a pioneer, or r selected species. The native tree ferns, on the other hand, had a much lower leaf turnover rate and slower growth rates, as well as lower reproductive output. This difference in life history characteristics may par- tially explain why S. cooperi can become established and spread quickly in Hawaiian rainforests, particularly in areas of disturbance. 34 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) ACKNOWLEDGEMENTS We would like to thank the B. Krauss Fellowship in Botany and C. Smith and the — Research Program at the University ae Baweld for Sy hee funds for yee research. We also like to thank the Oahu Natural A sistance in acai Mt. Kaala every month for a year. We also gratefully acknowledge the a Souza Center at Kokee State Park and Lyon Arboretum for their support REFERENCES Barucu, Z., and G. GOLDSTEIN. 1999. Leaf construction cost, nutrient ae and net CO, assi = paige: of native and invasive species in Hawaii. Oecologia 121(2):183—192. BECKER, R. E. 1976. The phytosociological position of tree ferns (Cibotium spp. 5) in the montane rea on the Island of Hawaii. Ph.D. Dissertation. University of Hawaii-Manoa, Hono- lulu, HI BITTNER, J., and S. W. BRECKLE. rele a growth rate and age of tree fern trunks in relation to habitats. sg Fern J. 85(2):3 Buck, M. G. 1982. Hawaiian ie ‘pian affects forest regeneration and plant succession. Research note PWS-355. Pacific Southwest Forest and Range Experiment Station, United States Department of Agriculture Forest Service, Berkeley, California. CHABOT, B. F., and D. J. Hicks. 1982. The ecology of leaf life spans. Annu. Rev. of Ecol. and System. DeBacu, P., and D. ROSEN. 1991. Biological control by natural enemies. Cambridge University Press, Cambridge U DuRAND, L. Z., and G. GoLp. LDSTEIN. 2001. Photosynthesis, photoinhibition, ae earacane use effi- ciency in native and invasive tree ferns in Hawaii. Oecologia 126:345—3 k, Gastony, G. J. 1974. Spore morphology in the Cyatheaceae. I. The perine and sporangial capacity: general considerations. Amer. J Bot. 61(6):672-680. ERS, OORTER, and M. I. vAN VUREN. 1998. Inherent Variation in Plant Growth: Phys- iological Mechanisms and ea a Backhuys Publishers, The Netherlands. dy es aad) D: R-DOMBOIS. 1989. Ch aus ual invaded islands, with special fee to Hawaii. Pp. 257-280 A A. Drake, H. A. Mooney, F. di Castri, R. H. Groves, F. t, M Rejmanek and M. Williamson, eds. Biological Invasions: A Global Perspective. John Wiley & Sons, New York. LoopE, L. L., R. J. NAGATA, and A. C. MEDEIROS. 1992. Alien plants in Haleakala National Park. Pp. 551-576 in C. P. Stone, C. W. Smith, and J. T. Tunison, eds. Alien plant invasion in native ecosystems of Hawaii: Management and Research. Cooperative National Park Resources Study Unit, ee “ Hawaii ap oo ai MEDEIROS, A. C., L. L. LOopE, S. J. ANDERSON. 1993. D tial colonizati ti roaster spp.) and alien ‘eyed pana tree ferns in a Hawaiian 1 rain forest. Selbyana —74. a A e S. and G. P. McCase. 1993. Introduction to the practice of statistics. W. H. Freeman and Company, New York. PaAtMer, D. D., 1994. The Hawaiian species of Cibotium. Amer. Fern J. 84 (3):73-85. REICH, P. B. 1993. Reconciling apparent discrepancies among studies relating life span, structure and function of leaves in contrasting plant life forms and climates: ‘the blind man and the elephant retold’. Funct. Ecol. 7:721—725 REICH, P. B., M. B. WALTERS, and D. S. ELLSWORTH. 1992. Leaf life-span in in to ms plant, and stand characteristics among diverse ecosystems. _ Monogr. 62(3):365-39 REICH, P. B., M. B. WALTERS, and D. S. ELLSWORTH nies m tropics to tundra: global conver- gence in plant functioning. Ecology 94: 13730-1373 RIPPERTON, J. D. oe ae Hawaiian tree fern as a ce source of starch. Hawaii Agri. Exp. SPORNE, K. R. 1966. as Morphology of Pteridophytes. Hutchinson & Co., Publishers, London, U.K. DURAND & GOLDSTEIN: NATIVE AND INVASIVE TREE FERNS 35 STRATTON, L., G. GOLDSTEIN, and F. C. MEINZER. 2000. Temporal and spatial partitioning of water resources among eight woody species in a Hawaiian is ‘ates Pet a 124(3):309-317. VITOUSEK, P. M. 1986. Biological invasions and ecosystem properties: can species make a differ- ? in H. A. Mooney and J. A. Drake, eds. Ecology of Biological Invasions of North America and Hawaii. Springer-Verlag, New York. VITOUSEK P. M., G. GERRISH, D, R. TURNER, L. R. WALKER, and D, MUELLER-DOMBOIS. 1995. Litterfall and nutrient cycling in four Setemess montane rainforests. J. Trop. Ecol. 11:189-203 ViTousEK, P. M., L. L. Loope, and C. P. STONE. 1987. Introduced species in Hawaii: biological effects and ee for ecological research. TREE 2:224-227 WALKER, L. R. and G. H. APLET, 1994. Growth and fertilization responses of Hawaiian tree ferns. Biotropica ig 378-383. Wacner, W. H. 1995. Evolution of “isileieaad ferns and fern allies in relation to their conservation status. one ms 49(1):31-4 Wick, H. L., G. T. HASHIMOTO. “isl Frond development and stem growth of treefern in Hawaii. U.S. Forest Service Research Note PW-237, Pacific Southwest Forest and Range Experiment Station. Berkeley, CA. American Fern Journal 91(1):36—40 (2001) SHORTER NOTE(s) Binomial for Dryopteris clintoniana x goldiana.—One strikingly handsome Dryopteris hybrid, the largest of all North American temperate Dryopteris (Thorne, F. and L. Thorne. 1989. Henry Potter’s Field Guide to the Hybrid Ferns of the Northeast. Vermont Institute of Natural Science, Woodstock, VT) and an excellent fern for hardy gardening (Mickel, J. T. 1994. Ferns for American Gardens. Macmillan, New York, NY), is without a collective epithet: D. clin- toniana X goldiana. Dryopteris Xmickelii Peck, hyb. nov—TYPE: UNITED STATES, New Jersey, Sussex Co., west side of extensive swamp south of Big Spring, 3 km south of Springdale, 17 Oct. 1969, James D. Montgomery 90895n (NY). Planta hybrida inter Dryopterem clintonianam et D. goldianam, aliis char- acteribus inter parentes media, sporis abortivis The hybrid is named to honor Dr. John ee Mickel (1934—) for his many endeavors on North American fern taxonomy, floristics, and horticulture, par- ticularly his promotion of gardening with hardy ferns. The hybrid is distin- guished readily from its parents by its abortive and irregular sized spores, intermediate frond morphology, and shared characters with both parents (see table in Thorne and Thorne, 1989, page 34). Dryopteis < mickelii has fronds up to 160 cm long, stipe to frond length ratio of 1:4, and is covered moderately with brown to dark brown scales. The blade is about 120 cm long, with a width to length ratio of about 2:5, with a very long and broad outline. The pinnae are regularly spaced, ascending relative to the rachis, with pinna shape wide, long rectangular and acuminate. Sori are borne close to the pinnule midrib. Dryopteris Xmickelii shares with D. clintoniana the following features: dark coloration at base of scales, relative length of blade, pinnae ascending from rachis, relatively narrow blade, and intermediate sorus location. D. x mickelii shares with D. goldiana the following features: wide blade, length of pinnules, falcate pinnules, relatively dark scales, and shape of pinnae. This hybrid occurs in rare and local populations in southern Ontario, south to New York, New Jersey, and Pennsylvania, and westward as outliers in Mich- igan and Ohio. This relatively narrow geographic range reflects the geographic overlap of the two parental ranges in the northeastern region of North America. The habitat of wetland woods and swamps were more common in early post- glacial times, but have declined since then, particularly in the western one- half of the range. In nature, the hybrid warrants conservation efforts wherever it still occurs. PARATYPES.—CANADA. Ontario: in woods at Ottawa, Scott s. n. (NY). UNITED STATES: Michigan: Washtenaw Co.: deep yellow birch swamp in Irwin’s woods, Wagner 9458 (US); Tuscola Co.: south of Murphy Lake, Wagner 63051 (US); tamarack swamps at Oxford, Farnwell 6117 (US). New Jersey: 37 Sussex Co.: swamp west of Springdale at Big Spring near Newton, Dowell 4929 (NY, US), Dowell 5033 (NY); Sussex Co.: roadside 1 mi south of Greendell, Edwards s. n. (NY); Bearfort Mtn., W. D. Miller 1648 (NY); West Englewood, Carhart 2b (NY) 2 sheets. New York: Green Lake, Jamesville, W. R. Dunlop s. n. (NY); Harris Swamp near Pilot Knob, Benedict s. n. (US-2202218); Kirkville, L. M. Underwood s. n. (NY); Staten Island: Arlington Station, Dowell 2801a & b (US). Ohio: Geauga Co.: Hopkins s. n. (NY). Pennsylvania: Berks Co.: swamp along spring-run 1.5 mi ne of Bernharts, Wherry s. n. (US-1849217); Delaware Co.: valley of Cruise Creek, Poyser 1286 (NY); Wakefield, Lanbor, J. J. Carter s. n. (NY). Vermont: N. Lynnfret (Limfret?), A. P & L. V. Morgan s. n. (US-154517). In the garden, its vegetative propagation should be promoted and supple- mented with modern tissue culture techniques to meet horticultural demands. It is a large and vigorous fern that forms extensive colonies through vegetative expansion while under cultivation. One specimen from New York was planted at the New York Botanical Garden, Bronx, NY in 1960 by members of the American Fern Society. Thirty-five years later, that plant had formed a clone 4 m across with over 200 apices (Mickel, 1994). JAMEs H. Peck, Department of Biology, University of Arkansas at Little Rock, 2801 S. University Ave., Little Rock, AR 72204. Cryopreservation of Shoot Tips of Selaginella uncinata.—Cryopreservation, or storage in liquid nitrogen (LN), has been successful for a wide variety of plant tissues, including shoot tips of in vitro cultures of higher plants (Bajaj, In Y. P. S. Bajaj, Cryopreservation of Plant Germplasm I, Biotechnology in Ag- riculture and Forestry 32. Springer, Berlin, 1995). At the temperature of LN, —196°C, long-term, stable storage of rare, endangered, or other valuable plant germplasm can be achieved. In an attempt to extend this technology to pteri- dophytes, LN storage of shoot tips of in vitro grown Selaginella uncinata was tested using the encapsulation dehydration procedure (Fabre and Dereuddre, Cryo-Letters 11:413—426, 1990). Shoot cultures of Selaginella uncinata (Desv. ex Poir.) Spring. were estab- lished from a plant purchased from Carolina Biological Supply Co. (voucher specimens deposited at the University of Cincinnati Herbarium, CINN, and at the CREW Herbarium). Tissues were surface disinfested in a 1:20 dilution of commercial sodium hypochlorite for five minutes, followed by two rinses in sterile, ultrapure water. The tissues were then transferred to a basal medium consisting of half-strength Murashige and Skoog salts with minimal organics (MS medium; Linsmaier and Skoog, Physiol. Plant. 18:100-127, 1965) with 1.5% sucrose and 0.22% Phytagel (Sigma Chemical Co.). Cultures were main- tained on this medium in 60 X 15 mm plastic petri plates, approximately 15 ml of medium/plate, at 26°C under Cool White fluorescent lights with a 16:8 hr, light:dark cycle. For preculture, one week prior to freezing the tissues were transferred to fresh basal medium or to basal medium plus 10 pM abscisic acid (ABA), which was filter sterilized and added to the medium after autoclaving. 38 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) AB Survival of shoot tips of S. uncinata through the steps of the encapsulation dehydration technique, 3—15 clean shoot tips per treatment. Encapsulated shoot tips were pretreated in 0.75 M sucrose, dried and exposed to LN, and samples were cultured after each step. Results are the mean of three experiments (+SE), except for long term LN exposure, for which there was one trial. Numbers of samples per trial are given in parentheses. Percent survival! 1 hr 3:5 NC Preculture medium Pretreated Dried LN exposure LN exposure Basal fe Seg a 29 = 6 Lice gt 0 127593513) C1233, 530) (8,:15, 11) (14) Basal + ABA 90 + 6 92 ct Tekst ao C12 495) (2 sh, 3) G, 9, 6) (11) ' Values with different superscripts are significantly different to the 0.05 level. For cryopreservation, the procedure of Fabre and Dereuddre (1990) was fol- lowed. Shoot tips approximately 2-3 mm in length were excised and placed in a solution of 3% alginic acid with 0.75 M sucrose in calcium-free MS me- dium. The solution containing the tips was then added dropwise to a solution of 100 mM CaCl, to form the alginate beads. These were incubated for 18 hr in liquid MS medium containing 0.75 M sucrose and were then dried for 4 hours in open dishes on filter paper under the air flow of the laminar flow hood. The dried beads containing the shoot tips were then transferred to 2 ml polypropylene cryovials and immersed directly into LN. After 1 hr, the cryo- vials were removed from the LN and allowed to warm at room temperature for 20 minutes. The encapsulated shoot tips were then transferred to basal medium for rehydration and recovery growth. Survival was measured as re- growth from a shoot tip. Results were analyzed by ANOVA (Tukey HSD Test, StatSoft, QuickSTATISTICA software). Drying and LN exposure significantly decreased survival from shoot tips which were precultured on basal medium alone (Table 1). When shoots were precultured for one week on basal medium plus 10 »M ABA, there was no decrease in survival when the tissues were dried, although survival did de- crease to approximately half when the tissues were exposed to LN. However, survival of both dried and LN exposed tissues was significantly better than that of control tissues. Shoot tips which were prepared using the encapsulation dehydration pro- cedure with and without ABA preculture were also transferred to LN for long- term storage. When these were removed after 3 1/2 yrs of storage in LN and recultured on basal medium, there was no regrowth from shoot tips which had been isolated from tissues grown on basal medium. However, 44% of the shoot tips which had been precultured on basal medium plus ABA initiated good growth after eight weeks (Figure 1). These results demonstrate that shoot tips of S. uncinata can survive LN exposure and long-term storage in LN when they are prepared using the en- capsulation dehydration procedure in combination with preculture on ABA. FiGuURE 1. Shoot tip of S. uncinata growing out of alginate bead, 8 weeks after exposure to LN. 6.7 X Encapsulation dehydration is a technique which has been successful with shoot tips of a number of species of angiosperms (Touchell and Dixon, In B. G. Bowes. Atlas of Plant Propagation and Conservation., Manson Publishers, London, 1999) as well as with gametophytes of ferns and bryophytes (Pence, The Bryologist 101:278-281, 1998 and Amer. Fern J. 90:16—23, 2000), and it is not surprising that it can also be successful with shoot tips of the fern allies as well. Survival of S. uncinata through cryopreservation was low unless the tissue was precultured on ABA. Treatment with ABA has been shown to initiate desiccation tolerance in the fern Polypodium virginianum L. (Reynolds and Bewley, J. Exper. Bot. 44:921-928. 1993a and J. Exper. Bot. 44:1771-1779, 1993b), as well as in bryophytes (Hellwege et al., Planta 198:423-432, 1996; Pence, 1998), and ABA is involved in several aspects of desiccation tolerance and water stress in seed plants (Hartung and Davies, In W. J. Davies and H. G. Jones, Abscisic Acid. Physiology and Biochemistry. BIOS Scientific Publishers, Oxford, 1991). It appears that preculture on ABA improves tolerance of S. uncinata shoot tips to the stresses of drying and LN exposure in the encap- sulation dehydration procedure, perhaps mimicking a natural response to stress in this species. Several other species of Selaginella are known to be “resurrection plants,” with a remarkable ability to tolerate extreme water loss 40 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 1 (2001) (see Proctor and Pence, In M. Black and H. Pritchard. Desiccation and Plant Survival. CAB International, Oxon, U.K., in press). Although encapsulation dehydration was successful in preserving shoot tips of S. uncinata, preliminary studies in this laboratory indicate that the survival of some other species of Selaginella through this procedure may not be as high (unpublished results). Species differ in their ability to survive LN exposure, and other techniques, such as slow freezing or vitrification, may be more suc- cessful with other species in this genus (Fukai et al., Euphytica 56:149-153, 1991; Yamada et al., Plant Science 78:81—87, 1991). There are over 700 species of pteridophytes which are of conservation con- cern worldwide, including 23 species of Selaginella (Walter and Gillett, 1997 IUCN Red List of Threatened Plants. IUCN—The World Conservation Union, Gland, Switzerland 862 Pp., 1998). Further studies should reveal how broadly applicable the encapsulation dehydration procedure is to shoot tips of pteri- dophytes and how cryopreservation may be used for preserving Selaginella germplasm for research and conservation.— VALERIE C. PENCE, Center for Re- search of Endangered Wildlife, Cincinnati Zoo and Botanical Garden, 3400 Vine Street, Cincinnati, OH 45220. Wii INFORMATION FOR aes Authors are encouraged to submit manuscripts pertinent to pteridology for pub- lication in the American Fern Journal. Manuscripts should be sent to the Editor. Acceptance of papers for publication depends on merit as judged by two or more referees. Authors are encouraged to contribute toward publishing costs; however, the payment or non-payment of page charges will affect neither the acceptability of manuscripts nor the date of publication Authors should adhere to the following guidelines; manuscripts not so prepared may be returned for revision prior to review. Submit manuscripts in triplicate (xerocopies acceptable), including review copies of illustrations and originals of illustrations. After review, submission of final versions of manuscripts on diskette (in PC- or Mac-compatible formats) is strongly encouraged. Use standard 8% by 11 inch paper of good quality, not “erasable” paper. Double space manuscripts throughout, including title, authors’ names and addresses, short, informative ab- stract, text (including heads and keys), literature cited, tables (separate from text), and figure captions (grouped as consecutive paragraphs separate from figures). Arrange parts of manuscript in order just given. Include author’s name and page number in upper right corner of every sheet. Provide margins of at least 25 mm all around on typed pages. Do not submit right-justified copy, avoid footnotes, and do not break words at ends of lines. Make table headings and figure captions self-explanatory. Use S.I. (metric) units for all measures (e.g., distance, elevation, weight) unless quoted or cited from another source (e.g., specimen citations). For nomenclatural matter (i.c., synonymy and typification), use one paragraph per basionym (see Regnum Veg. 58:39—40. 1968). Abbreviate titles of serial publi- cations according to Botanico-Periodicum-Huntianum (Lawrence et al., 1968, Hunt Botanical Library, Pittsburgh) and its supplement (1991). References cited only as part of nomenclatural matter are not included in literature cited. For shorter notes and reviews, omit the abstract and put all references parenthetically in text. Use Index Herbariorum (Regnum Veg. 120:1—693. 1990) for designations of her- baria. Illustrations should be proportioned to fit page width with caption on the same page. Provide margins of at least 25 mm on all illustrations. For continuous-tone illustrations, design originals for reproduction without reduction or by uniform amount. In composite blocks, abut edges of adjacent photographs. Avoid com- bining continuous-tone and line-copy in single illustrations or blocks. Coordinate sequence and numbering of figures (and of tables) with order of citation in text. Explain scales and symbols in figures themselves, not in captions. Include a scale and reference to latitude and longitude in each map. Proofs and reprint order forms are sent to authors by the printer. Authors should send corrected proofs to the editor and reprint orders to the printer. Authors will be assessed charges for extensive alterations made after type has been set. For other matter of form or style, consult recent issues of American Fern Jour- nal and The Chicago Manual of Style, 14th ed. (1993, Univ. Chicago Press, Chicago). Occasionally, departure from these guidelines may be justified. Authors are encouraged to consult the editor for assistance with any aspect of manuscript eparation. Papers longer than 32 printed pages may be sent to the Editor of Preridologia (Memoir Editor, see cover 2). PTERIDOLOGIA ISSUES IN PRINT The following issues of Preridologia, the memoir series of the American Fern Society, are available for purchase: 1. Wagner, David H. 1979. Systematics of Polystichum in Western North America North of Mexico. 64 pp. $10.00 postpaid. 2A. Lellinger, David B. 1989. The Ferns and Fern-allies of Costa Rica, Panama, and the Chocé (Part 1: Psilotaceae through Dicksoniaceae). 364 pp. $32.00 postpaid. Send your order with a check or money order to: American Fern Society, Inc., c/o U.S. National Herbarium MRC-166, Smithsonian Institution, Washing- ton, DC 20560. AMERICAN FERN JOURNAL ON MICROFICHE Volumes 1—61 of the American Fern Journal are available as archival quality, silver positive microfiches. Single volumes or the entire run may be purchased. The fiches are easily read with 10X or greater magnification (using a dissecting microscope and transmitted illumination or a fiche reader). Silver negative micro- fiches of vols. 1-50 are also available. The price is $4.00 per volume or $244.00 per set of 61 volumes, postpaid. Send your inquiry or order with a check or money order to: American Fern Society, Inc., c/o Dr. James D. Montgomery, Ecology II, Inc., R.D. 1, Box 1795, Berwick, PA 18603. VISIT THE AMERICAN FERN SOCIETY’S WORLD WIDE WEB HOMEPAGE: http://www.amerfernsoc.org/ AMERICAN aie FERN nant JOURNAL ae QUARTERLY JOURNAL OF THE AMERICAN FERN SOCIETY Systematics of the Northern Andean Isoétes karstenii Complex Randall L. Small and R. James Hickey 41 Shorter Note Additions and Corrections to the Pteridophyte Flora of Northeastern Argentina onica Ponce 70 The American Fern Society Council for 2001 BARBARA JOE HOSHIZAKI, 557 N. Westmoreland Ave., Los Angeles, CA 90004-2210. CHRISTOPHER H. HAUFLER, Dept. of Botany, University of Kansas, Lawrence, KS e045: ce-Presiden W. CARL TAYLOR, 800 W. Wells St., Milwaukee Public Museum, Milwaukee, WI is 14 iets JAMES D. CAPONETTI, Dept. of Botany, University of Tennessee, Knoxville, TN 37916-11 = Treasurer DAVID B. LELLINGER, 326 West St. NW., Vienna, VA 22180-4151. Membership Secnaas JAMES D. MONTGOMERY, Ecology III, R.D. 1, Box 1795, Berwick, PA 18603-9801. k Issues Curator GEORGE YATSKIEVYCH, Missouri Botanical Garden, P.O. Box 299, St. Louis, MO 63166-0299, Journal Editor DAVID B. LELLINGER, U.S. National Herbarium MRC-166, Smithsonian Seep Washington, DC 20560-0166. Memoir Editor CINDY JOHNSON-GROH, Dept. of Biology, Gustavus Adolphus College, 800 W. College ee St. Peter, MN 56082-1498. Bulletin Editor American Fern Journal EDITOR R. JAMES HICKEY al ReaD eae OH 45 45056 ph. (513) <5, 6000, e-mail: hickeyrj @muohio.edu ASSOCIATE EDITORS GERALD J. GASTONY: ... ss Bae of “pata Indiana naires Bloomington, IN 47405-6801 CHRISTOPHER H. HAUFLER .... Dept. o: ae University wie ansas, Lawrence, KS 66045-2106 ROBBIN C. MORAN w York Botanical Garden , Bronx, NY 10458-5126 JAMES H. PECK San of a acu of Arkansas—Little Rock, 1 S. University Ave., Little Rock, AR 72204 The “American Fern Journal” (ISSN 0002-8444) is an illustrated quarterly devoted to the Sg study of ferns. It is owned by the American Fern Society, and err at 326 West St. NW., Vienn VA oe 1. Periodicals 5 postage paid at Vienna, VA, and additional entry. . hs (domestic) to 12 ae (foreign) after the date of issue, and oes 8 for back issues should be addressed to Dr. James D. Montgomery, Ecology II, R.D. 1, Berwick, ag doe Eians dress, dues, and applications for membership should be sent to the Membership Secretary. eral inquiries concerning ferns should be addressed to the Secretary. por ncaa $20.00 gross, $19.50 net if paid through an agency (agency fee $0.50); sent free to members of the ne Fern Society (ann ual dues, $15.00 + $5.00 mailing surcharge beyond A.). ack volumes are sola for most years as printed issues or on microfiche. Please contact the POSTMASTER: Send address changes to AMERICAN FERN JOURNAL, 326 West St. NW., Vienna, VA 22180-4151. FIDDLEHEAD FORUM The editor of the Bulletin of the American Fern Society welcomes contributions from members and non-members, including miscellaneous notes, offers to ~~ or purchase materials, personalia, horti notes, and reviews of non-technical books on ferns Mr. Stephen McDaniel, 1716 Piermont Dr., Hacienda Hts., CA 91745-3678, is Director. Spores Gifts dh t ee < eg a pn _ nee eial al say: soy cana pag inert dame! yalronpete 1 | eee tax-deductible Inquiries should be addressed to the Secretary. American Fern Journal 91(2):41-69 (2001) F > §& Se oi Systematics of the Northern Andean &@ WV & . s ne Oe \O Isoétes karstenii Complex Se & = = RANDALL L. SMALL B SS a Department of Botany, The University of Tennessee, Knoxville, TN 3306 a = Oo R. JAMES HICKEY Department of Botany, Miami University, Oxford, OH 45056 ABSTRACT.—The Isoétes karstenii complex includes those species characterized by laevigate mega- spores, acute to free ala apices, a highly reduced labium, and distributions in the high altitude ivelata and I. precocia. Chromosome counts revealed that I. karstenii and I. precocia are diploids (2n = 22) and that I. palmeri is tetraploid (2n = 44). Estimates of chromosome number based on spore size for I. fuliginosa and I. hemivelata indicate that they are polyploid (2n = 44) The laevigate-spored taxa of Isoétes L. from the northern Andes have long presented taxonomists with considerable difficulty. These difficulties are due largely to the overall morphological similarity of the species and the wide range of morphological variation within the species. Also, few collections of these species were available for study until the field work of A.M. Cleef and collaborators and J. and S. Keeley in the 1970s and 1980s. As part of continu- ing efforts to describe and characterize the Neotropical species of Isoétes we undertook a revisionary study of this difficult group of eRe spored spe- cies, the Isoétes karstenii A. Braun complex (Hickey, 1985). The goals of this study were to: (1) determine the number of taxa that should be included within the I. karstenii complex, (2) clarify the taxonomy of the species, and (3) document intra- and interspecific morphological variation. To accomplish these goals we have collected morphological data and conducted principal components analyses of those data to infer morphologically discon- tinuous groups. In addition we have determined or estimated chromosome numbers for the included species. Finally, we present a taxonomic treatment that formalizes our understanding of the delimitations and relationships among these species. TAXONOMIC HISTORY OF THE ISOETES KARSTENII COMPLEX The Isoétes karstenii complex was first delimited by Hickey (1985) in a mor- phologically based cladistic analysis of Neotropical Isoétes. It includes those 42 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 2 (2001) TABLE 1. Summary of the taxonomic recognition of the species historically associated with the /soétes karstenii complex. A “+” indicates the taxon was accepted as valid by the author(s); an “S” that the taxon was considered a synonym of /. lechleri; a ““V” that the taxon was considered a variety of 1. lechleri. Isoétes Isoétes Isoétes Isoétes Isoétes colombi- rimbachi- Isoétes Isoétes i ana palmeri lechleri karstenii — socia cleefii Mettenius (1859) aa Braun (1862) + + + Baker (1880) + Ss S Motelay & Vendryes (1882) =A S S Underwood (1888) + 5 S. Pfeiffer (1922) + S 5 eber (1922) + + S Plamer (1929, 1932) + S S y Vv Rodriguez (1955) + S S Vv Vareschi (1968 a5 Vv Vv Gomez (198 + 5 Ss s S Fuchs-Eckert (1982) + ~ - ~ oo ~ Hickey (1985) + + + + + taxa with laevigate megaspores, truncate to free ala apices and corneous (dark- ly sclerified) leaf apices. The complex is distributed throughout the péramos of Colombia and Venezuela at elevations above 3000 m, with a single disjunct location at Mt. Chimborazo, Ecuador. Hickey (1985) included five species in this complex: Isoétes cleefii H. P. Fuchs, Isoétes palmeri H. P. Fuchs, Isoétes karstenii A. Braun, Isoétes socia A. Braun (sensu Fuchs-Eckert, 1982), and Isoétes rimbachiana H. P. Fuchs. In a compendium of South American species, Fuchs-Eckert (1982) recognized eight laevigate-spored species for Venezuela, Colombia, and Ecuador and therefore may be associated with the I. karstenii complex: I. socia, I. karstenii, Isoétes steyermarkii (nom. nud.), I. palmeri, I. cleefii, I. rimbachiana, I. colombiana (T. C. Palmer) H. P. Fuchs, and Isoétes glacialis Asplund (including Isoétes arumiana [nom. nud.] pro syn.). Fuchs- Eckert (1982), however, did not recognize these species as a distinct group. Historically these names have been treated in the literature as distinct species, or more often, as synonyms or varieties of Isoétes lechleri Mett., a central An- dean (southern Peru and Bolivia) species (Table 1). MATERIALS AND METHODS MorPHOLOGY.—Approximately 200 herbarium sheets representing about 90 discrete collection numbers of Venezuelan, Colombian, and Ecuadorian lae- vigate-spored Isoétes were examined. We gratefully acknowledge the following herbaria for use of material: B, BM, COL, F, G, GH, MO, MU, NY, U, UC, US and VEN. Sixty-three of the collections were chosen and scored for eight quan- titative morphological characters; those collections that were excluded from the analyses consisted of immature or poorly preserved material. Characters SMALL & HICKEY: ISOETES KARSTENII COMPLEX 43 used in the ordination and statistical analyses were: leaf number, leaf length (mm), leaf width (mm) at the mid-length, ala length/leaf length ratio, length (mm) of corneous portion of leaf apex, velum coverage (scored as a discrete character; see below), megaspore diameter (um), and microspore length (ym). Additional characters were scored for a subset of the collections but were aban- doned due to non-discriminatory age-specific correlation (e.g., corm dimen- sions and sporangial dimensions) or extreme variation within individuals and collections (e.g., sporangial shape and ala apex morphology). Some characters used in the analyses are also correlated with age (e.g., leaf number), or showed variability within collections (e.g., leaf length). Despite this, these characters also appeared to show group specific and discriminatory ranges of variation and were therefore retained. Terminology for morphological characters follows Hickey (1985), and for two-dimensional shapes Radford et al. (1972). Velum coverage was scored as: (1) complete to nearly complete (> 90%), or as (2) less than half covered (25-50%). For ordination analyses, velum cov- erage was therefore scored as a discrete character, complete vs. incomplete. Although microspore length and width were measured, only microspore length was chosen for the ordination and statistical analyses because most previous workers have used microspore length, rather than width, some re- porting only length values. Correlation between these two characters is high for 55 collections analyzed (r? = 0.877), and either character should adequately represent the collection. Measurements were taken only from mature plants and mature organs of those plants. Each collection, regardless of number of sheets or individuals, was considered representative of a single population unless evidence of hy- bridization (e.g., aborted spores) or multiple taxa within a collection (e.g., gross discontinuities in morphology) were present. Several individuals per collec- tion were scored for each character (with the exception of microspore dimen- sions which were taken from a single sporangium) and a mean score for each character was calculated. The number of measurements per character differed from collection to collection. Some collections consist of only a single plant on a single herbarium sheet while others consist of numerous individuals on numerous sheets. In addition, the quality of materials varied greatly: many specimens were partially destroyed by insects, many were pressed with the substrate around the corm, and many were pressed in large clumps. As a re- sult, some characters were difficult to obtain for all collections. Ten measure- ments per character per collection (25 for megaspores and microspores) were recorded whenever possible. Two types of analyses of morphological data were conducted: (1) principal component analysis (PCA) and (2) character by character scatter plots. These analyses were performed in an attempt to reveal morphological discontinuities between groups of collections and were conducted without assigning collec- tions to specific taxa, except for type specimens which were so designated. This was done to determine, without bias, whether or not distinct clusters of collections would be revealed which could be designated as putative morpho- logical species and to determine which type specimen(s) would be associated ts AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 2 (2001) with which cluster. Once morphologically cohesive groups were recognized, descriptive statistics were calculated to document putative intraspecific vari- ation. Multivariate and statistical analyses were performed using NTSYS ver- sion 1.80 (Rohlf, 1993) and Statmost v. 2.50 (DataMost). CyToLoGy.—Methods and materials for chromosome counts generally follow those of Hickey (1984). Voucher specimens for the materials from which counts were made are deposited at LOC, MU, and BM (Table 2). All specimens from which root tips were collected were grown completely submerged and potted in either fine quartz sand or a 3:1 mixture of sphagnum peat: fine quartz sand (Taylor and Luebke, 1986). Cells with mitotic figures were best obtained in root tips collected in the late morning (11:00 a.m. + one hour). Young roots about 1 cm long were harvested and immediately placed in a saturated para- dichlorobenzene solution, incubated at room temperature for 4—5 hr, fixed in Farmer’s solution (3:1 95% ethanol: glacial acetic acid) and refrigerated until squashed. Root tips were hydrolyzed in 1N hydrochloric acid at 60°C for 20- 30 min, neutralized in a 95% ethanol series, stained in Whittman’s hematox- ylin for 25 min at room temperature, rinsed in glacial acetic acid to destain, and squashed in Hoyer’s mounting media. Squash slides were examined and photographed at 600 or 1000. Color photocopies of the photographic slides were then prepared. These photocopies were placed on a light box, covered with a plain piece of white paper and the chromosomes traced onto the white paper from which counts were made. Simultaneous viewing of the photo- graphic and/or microscope slides and the photocopy aided in interpreting slides in which chromosomes overlapped. When live material was unavailable, we attempted to estimate chromosome number by plotting mean megaspore diameter vs. mean microspore length for taxa of known chromosome number and then plotting the mean value of spore sizes of the unknown taxa onto that plot. A strong correlation exists between spore size and chromosome number in Isoétes (Cox and Hickey, 1984; Britton and Goltz, 1991). For northeastern North American taxa, Kott and Britton (1983) noted that ‘ploidy level can be determined with some certainty from spore measurements, even for dried herbarium material.” Mean values for spore sizes of 27 taxa of known chromosome number were taken from Rury (1978), Kott and Britton (1983), Britton and Goltz (1991), Brunton et al. (1994), Taylor et al. (1993), Hickey (1994), Musselman and Knepper (1994), Brunton and Britton (1996a, 1996b), and Watanabe et al. (1996). While such correlations are unlikely to definitively identify the chromosome number of an unknown species, they may be useful in obtaining an estimate, or as in our case (see below), in ruling out potential ploidy levels. RESULTS MorPHOLOGY.—Included among the collections examined were type speci- mens (holo-, iso-, and/or paratypes) of Isoétes arumiana (nom. nud.), I. cleefii, I. colombiana, I. glacialis, I. karstenii, I. lechleri, I. palmeri, I. rimbachiana, I. SMALL & HICKEY: ISOETES KARSTENII COMPLEX (DOD Z£EOrI Kajaay x» Kajaay RIQUIO[OD “eorBUeUIpUND ‘esIe’] eunse’] Jo § [ood [euls 7 = UZ p1209add "J (NED) Z8PLT Kuuar RIQUIO[OD ‘vorBUIeUIpUND ‘BoRSsIyD vunse’] Jeou ood [jews T= UZ (NED ZZPZ[ Kusar BIQUIO[OD ‘oreUBUIpUND ‘eoRsIyD euNnse’y seou [ood [yeuIS rr = UZ (JOO ‘AW) 2ajpz2u0yH = /Co[-G¢T jjows BIQUIO]OD “eoreUeUIpUND ‘voRsTYyD vuNse’] Jeou ood [jeuIS th = UZ tiauypd ‘| (WE) 6ZPZ[ Kuuar RIQUIO]OD ‘RoreUBUIpUND ‘BoRsIYyD vuNse’y Jeou jood [yews TT = UZ (TOO ‘AW) 2a/vp2u0yH x» pp ows BIQUIO[OD ‘vovhog ‘ANDO_D [ap BPRASN BLIOIG ‘epespeng eT eunse] TT = UZ (TOD ‘AW) 2a7p2u0yH » CpI-[ pI ows BIqUIO[OD ‘eovkog ‘And0_D Jap vpeaodyN vIOIS ‘epequlg eT vunse’] Co = UZ (TOD ‘AW) 24p2u0yH x» gz] jjows BIQuIO[OD ‘eoeAog ‘B1090g ap ordioruN| ‘BASIg Op OUIeIeg ‘puog TZ = UZ (NW) 4ax91H ~P g J]DUS PpANZIUDA ‘epLIg| ‘Ifeqnony eunsey CZ = UZ (NIN) 4ay217 3p Z jJpws BONZIUDA “EPLIO|] ‘BISON vuNse’] TT = UZ (AW) 4ay21H ~P 9 JJDUS Bjanzous, “epLsy ‘e[Inby ap oold Jo N ury S| ‘puog co = UC (AWD) 4ax91H P ¢ OWS Blanzoua, “eplLis| ‘e[Inby ap oolg JO N WY ¢] ‘puog CC = UW (AW) 4ay21H PY p% [JOU Pjanzous, ‘pug ‘epinby op OdIg JO N Wy g ‘puog Co = UZ NUaISADY “] uoneso’] Joyono,, uones0’] uonsa[[o9 qunod “Byep UsUTISads JOYONOA puR ‘UONRdO] UONDET]JOS ‘s]JUNOD WOSOWOIYD “7 -ATAV I, 46 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 2 (2001) TABLE 3. Eigenvectors for the first three principal components (PC) from the PCA. Character PC] Pe. 2 PC 3 Leaf length 0.21957 0.77593 0.21528 Leaf width 0.63580 0.16820 —0.61498 Microspore length 0.84544 —0.13050 0.28320 Megaspore diameter 0.76647 =—O.05274 0.14659 % ala coverag 0.18744 —0.75048 —0.42986 Sclerified portion of leaf tip 0.04654 0.53506 —0.72654 Leaf number 0.54509 0.51899 0.12401 Velum coverage —0.81467 0.36054 —0.04755 socia, and IJ. steyermarkii (nom. nud.). Examination of the holotypes of I. lech- leri, I. glacialis and I. socia (all from the Andes of southern Peru or Bolivia) was performed initially to determine whether or not these taxa should be in- cluded in the subsequent analyses. Both I. glacialis and I. socia were deter- mined to be synonyms of I. Jechleri, which in turn was determined to be dis- tinct from the northern Andean taxa. For a complete discussion of the vari- ability and synonymy of the I. Jechleri complex see Hickey (1994), but a brief discussion of our rationale for excluding these taxa follows. The synonymy of Isoétes socia and I. lechleri has been proposed by all authors except Fuchs- Eckert (1982, 1992) since the former’s publication. The type specimen of I. socia (Lechler 1937b, B!) consists of a single immature plantlet and a few isolated sporophylls. These materials were apparently removed from Lechler 1937 (the type of I. lechleri) and undoubtedly represent a vegetative offset from I. lechleri, a species in which vegetative reproduction is common (Hickey, 1994). The type of I. glacialis (Asplund 4041, isotype B!) is but one of the many morphotypes of the highly variable I. lechleri and is clearly distinct, both in morphology and distribution, from Cuatrecasas 19117, the basis for I. arumiana. Sixty-three collections were used in the PCA, including all of the type spec- imens except those of Isoétes glacialis, I. lechleri, and I. socia. Eigenvectors for the first three principal components are given in Table 3; the first two principal components accounted for 36% and 23%, respectively, of the total variation. A plot of principal components one vs. two (Fig. 1) reveals six po- tentially distinct groups, which will be referred to in subsequent discussion by number. Clusters 1, 2 and 3 contain single collections: the type of I. cleefii, the “type” of I. arumiana (nom. nud.) and the type of I. rimbachiana respec- tively. Cluster 4 contains the type collections of I. karstenii, I. palmeri, I. col- ombiana, and I. steyermarkii (nom. nud.), paratypes of I. palmeri and I. cleefii, and many additional collections. Cluster 5 consists of a group of eight collec- tions without an associated type specimen. Cluster 6 contains a group of five collections without an associated type specimen. The constituents of each of these groups were examined to determine which characters separated each the others and to determine whether or not these groups constitute morphologically definable taxa. SMALL & HICKEY: ISOETES KARSTENII COMPLEX 47 |. karstenii & H 7¢ fil i ‘ cleefii {e}2 PC2 0- _@@ e |. hemivelata |, fuliginosa oreo . |. karstenii var. anomala 2 y T -1 0 1 2 3 PC 1 Fic. 1. Scatter plot of principal component one vs. principal component two based on morpho- logical data. Clusters of associated specimens are indicated and numbered. 1 = I. cleefii; 2 = I. fuliginosa; 3 = I. karstenii var. anomala; 4 = I. karstenii + I. palmeri; 5 = I. precocia; 6 = I. hemivelata. Cluster 1 (the holotype of I. cleefii) appears quite distinct from the remaining clusters in Fig. 1, yet examination of the characters suggests that the separation between clusters 1 and 4 is tenuous. Cluster 1 is separated from cluster 4 along principal component (PC) 2 by leaf length, leaf width and leaf number, yet character by character, the type of I. cleefii fell within the range of values of the individuals in cluster 4. The type collection of I. cleefii has longer, wider, and more numerous leaves than the average member of cluster 4. However, some collections in cluster 4 have values greater than or equal to those of the type of I. cleefii for some of these characters, but not for all. It appears that the combination of high values for all of these characters serve to separate the type of I. cleefii from the collections in cluster 4. Cluster 2 contains the single collection, Cuatrecasas 19117, which is the basis of I. arumiana (nom. nud.). This collection is separated from the re- maining collections along PC 1 by differences in velum coverage, megaspore diameter, and microspore length. In addition, the plant is larger than any other collection in this group, and the ligule is also unusually large (ca. 7 mm high) relative to the other species in this complex. This combination of character states makes this collection the most distinctive of those analyzed, and there- fore we suggest that it deserves species status. To avoid further nomenclatural and taxonomic confusion caused by the invalid publication and subsequent synonymy of J. arumiana (nom. nud.), we propose this taxon be recognized as Isoétes fuliginosa sp. nov. Cluster 3 contains the single collection, Rimbach 171, the type of Isoétes rimbachiana (= I. lechleri var. anomala). Differentiation of this collection from I. karstenii (from which it is otherwise nearly indistinguishable) by Palmer (1932) was based primarily on velum coverage, which also separates it from the collections of cluster 4 along PC 1 in our PCA. While velum coverage is 48 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 2 (2001) often considered to be a highly plastic character, we have found velum cov- erage to be uniformly complete in all collections of I. karstenii. Thus, the presence of an incomplete velum, along with the disjunct geographical place- ment of this collection at Mt. Chimborazo, Ecuador suggests that this collection represents a potentially unique biological entity and thus should be main- tained as a distinct taxon. We do not feel, however, that the distinction is sufficient to warrant specific status. Because of its affinities with I. karstenii it should be recognized as I. karstenii var. anomala comb. nov. Cluster 4 is a large and heterogeneous collection. Along with a large number of other collections, it contains the holotypes of Isoétes karstenii, I. colombi- ana, I. steyermarkii (nom. nud.), and I. palmeri, as well as paratypes of both I. palmeri and I. cleefii. Due to this heterogeneity and the dubious separation of I. cleefii, we performed additional analyses on this group (see below). Cluster 5 contains a group of eight collections without an associated type specimen. While placed close to cluster 4 (Fig. 1), this group of collections exhibits a unique set of character states that set it apart from the remaining specimens. Specifically, the PCA separates them from other specimens based on their diminutive size (leaf length, leaf width, and leaf number) at maturity, and the reduction of the corneous leaf tip. These specimens are also ecologi- cally differentiated as they are found solely in ephemeral pools, not permanent lakes and ponds as are the other species of this group. We propose that this taxon be recognized as Isoétes precocia sp. nov. Cluster 6 contains a group of five collections without an associated type specimen. These collections are distinguished from the remaining collections along PC 1 by velum coverage, megaspore diameter, and microspore length. This group of specimens is also unique within the complex in having trigonal rather than terete leaves. Although similar to cluster 2 (Isoétes fuliginosa) in velum coverage and spore dimensions, the two are easily distinguishable based on leaf cross-sectional shape, leaf length, ala length / leaf length ratio, leaf number, and the length of the sclerified portion of the leaf tip. We propose that this unique taxon be recognized as Isoétes hemivelata sp. nov. To examine the possibility that additional undetected groups existed within cluster 4 and to determine the status of Isoétes cleefii, two-dimensional scatter plots of all combinations of characters were produced for those collections in clusters 1 and 4. The scatter plot of leaf length vs. leaf width at the mid-point (Fig. 2) reveals two distinct groups with different character trends: one dis- playing relatively long, slender leaves, the other with relatively short, thick leaves. The group with relatively long, slender leaves contains type and par- atype collections of I. palmeri and I. cleefii (Fig. 2) as well as a number of additional collections. The other group (Fig. 2), with relatively short, thick leaves, contains the type collections of I. karstenii, I. colombiana, and I. stey- ermarkii (nom. nud.) as well as a number of additional collections. Linear regression analyses of these specimens, individually and taken together sup- port our inferences that these two groups are distinct. If all specimens in Fig. 2 are included in a regression analysis, the r? = 0.0009 (P = 0.85) indicates a lack of significant correlation. If however, the two groups (above and below SMALL & HICKEY: ISOETES KARSTENII COMPLEX 49 400 _ ° o . 300 - = ea = = Boo D Ss 2004 oO a — oO © ro | ° 100 + * 2 Leaf Width at Mid-Length (mm) Fic. 2. Two-dimensional scatter plot of leaf length (mm) vs. leaf width (mm) for those specimens included in cluster 4 of the PCA (Fig. 1). The line divides the relatively long, narrow-leaved specimens (J. palmeri) from the relatively short, stout-leaved specimens (I. karstenii). the line in Fig. 2) are analyzed separately, values of r? = 0.70 (P = 1.2 X 10°-®) and r? = 0.25 (P = 0.012), respectively, are obtained. These values indicate that leaf length and width are significantly correlated within groups, thus lend- ing support to our hypothesis that they are biologically distinct. The separation of these two very similar groups appears to have been obscured in the initial PCA by the inclusion of such a morphologically diverse selection of speci- mens. The range of variation encompassed by the initial PCA was much larger than the variation between groups within cluster 4, therefore the entire range of variation encompassed by cluster 4 was confined to a small portion of the entire PCA space. A more fine-scaled analysis was then required to detect these differences. In addition to the separation of the two groups described above, this analysis clarifies the position of the type collection of I. cleefii as an ex- treme individual along a continuum (Fig. 2). On the basis of these analyses we delimit two additional species. The group with relatively long, slender leaves includes the type collections of Isoétes palmeri and I. cleefii. These two taxa were published simultaneously (Fuchs- Eckert, 1981a,b), and thus neither has priority. We propose that this taxon be recognized as I. palmeri because the type specimen is more representative of the species overall. The group with relatively short, wide leaves includes the type collections of I. karstenii, I. colombiana, and I. steyermarkii (nom. nud.). The name J. karstenii was the first of these to be published (Braun, 1862) and therefore has priority. In order to document the variation within these species, we calculated the mean and range of each character (Table 4). Although leaf length / leaf width at the mid-point ratio was not included as a character in the PCA (because both leaf length and leaf width were included as separate characters) it is an important character in discriminating Isoétes karstenii from I. palmeri and was therefore included in descriptive statistics. Additionally, as chromosome num- AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 2 (2001) br = UZ rr = UZ 7 = TZ = UZ Joquinu suwosowlo14 8V Orel + 6E)-LE Of + on 9 + ght 8E-(L'€ + O£)-S7 (wt) Sua] asodsossTy 00L—009 7Z9O-'6b + SOS)-EIS S6r7LE + Trh)-88E ISS66Il + 88P)-88E LOS-+6OII + 6Lr)-S6E (wit) sojourerp asodsesayy %SL-OS %SL-SE 006< %06< %06< (%) A3vI9AOD WIN]PA 0 JIXE0 + T1D-8'0 FO-1'0 + TOK-T0 y7HS'0 + FD-8'0 TE+(90'0 F BI-L'0 (wut) xede jee] Jo uonsod soaws0D S10 1 0-(S0'0 ¥ PE'O)-87'0 90-00 + OE0)-970 = 870-00 + 6I'0)-I 1.0: 8€'0-(90'0 + 97'0)-F1'0 ones YBue] Jea/psua] epy vel T8179 + ¥9)-9r LSIXS'be + 08)-9r Ose{T'bs + IP7)-6ll LEIP le + SL)-9€ ones Pp jeop/pSug] sea] 6'l v71S0 + OTT 60-470 + L'0)-S'0 bI-tr'0 = O'D-S'0 v7r'0 + 91-60 (wu) YISue]-pru we YYpIM Jeo] SST p8I-L'9s + 671)-Ss bv60'6I + TS)-LE Pr6E1{T'06 + OET)-€6 Isz-(8'es + SI1)-Or (wi) ySuey Jeo] Sc 91-81 + ELI HI-@O'T + L-9 t7-9'P + PIL IZ41S'€ + E18 JOQqUINN, JB9"T (I = (8 = N) Pivjaanuay ‘J (8 = N) »1909a«d ‘7 (@Z = N) Mauypd ‘7 (pZ = N) Muassapy J pmo yi ‘pojuasaid ase (sasayjuoied Ul) JOA prepuris | URW puke ddUBI dy) Ja}ORIRYO puR UOXe} YoRe JO,] ‘pafdues suondea]]O9 Jo Joquinu oy} = N UOXe) yore JO “xa[duIOD NUAaIsupy $ajaOS]T AY] JO SIAQUIDUI JO] Sayv}s Ja}OvIeYO Burysinsunsiq: “fp ATAVL SMALL & HICKEY: ISOETES KARSTENII COMPLEX 51 —= 1000+ 12 a = 900 how 2 800 4 ao = se cay (OO 10 c F g 600 4 4 Q 500; 455 & 4 g es Bo 4007 2 2 4 a. 2 4 = 300 3 Microspore length (um) Fic. 3. Two-dimensional scatter shee - whist ia length (wm) vs. megaspore diameter (ym) for species of Isoétes with known chromosome number. tonaes fA represents a single species with its chromosome number identified = habe (e.g., = 2x = 22, “4” = 4x = 44, etc.). The two unknown species are identified by letter (““H” = I. sedate! ‘F” = I. fuliginosa). bers are important in distinguishing taxa (see below), these data are also pro- vided in Table 4 CyToLocy.—Chromosome counts were performed for Isoétes karstenii, I. pal- meri, and I. precocia (Table 2). It was found that I. karstenii and I. precocia were diploids (2n = 22) and I. palmeri was tetraploid (2n = 44). In addition to numerous counts of 2n = 44 for I. palmeri, a single individual from one population of I. palmeri was found to be a triploid (2n = 33). This individual presumably represents a hybrid between I. palmeri and a diploid species, such as I. karstenii or I. precocia, both of which are sympatric with I. palmeri in the area of this population. These cytological data support the findings of the morphological analyses in further distinguishing I. karstenii and I. precocia from I. palmeri. Live material was not available for Isoétes fuliginosa or I. hemivelata. The relatively large size of the mega- and microspores of these taxa, however, sug- gest that they are most likely polyploid. We attempted to estimate the chro- mosome numbers of these taxa by comparing their spore sizes with taxa of known chromosome number. A scatter plot of megaspore diameter vs. micro- spore length was produced with individuals identified by chromosome num- ber (Fig. 3). Individual ploidy clusters show overlap with adjacent ploidy clus- ters and thus it is not possible to confidently characterize the chromosome numbers of the unknown species. Despite this, these data clearly indicate that these taxa are polyploid, both being well out of the range of known diploids. Isoétes fuliginosa appears to be a high-level polyploid, probably at least a hexa- ploid. Isoétes hemivelata is more difficult to characterize because it falls out near taxa of three different ploidy levels, yet the data suggest that it is at least a tetraploid, possibly higher. 52 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 2 (2001) DISCUSSION Taxonomy and Analyses As a result of our morphological and cytological analyses, we recognize the following taxa within the I. karstenii species complex: I. fuliginosa; I. hemi- velata; I. karstenii var. karstenii; I. karstenii var. anomala; I. palmeri; and I. precocia. The details of this taxonomic history are given in Table 1. The conclusions presented here are based on evaluation of morphological characters using multivariate analyses to derive distinct groups of populations. In the case of these taxa, traditional approaches using intuitive taxonomic as- sessments have led to undue lumping (placing all taxa in I. lechleri) or splitting (recognition of up to eight taxa within this complex by Fuchs-Eckert 1982). Many workers have discussed the morphological plasticity in Isoétes, yet few have emphasized the fact that circumscribing that variation is imperative for understanding species delimitations (but see Kott and Britton, 1985). Evidence presented in this study supports the utility of quantitative vegetative charac- ters in delimiting species, especially when those characters are examined us- ing a multivariate approach. Cytology and Speciation Speciation in Isoétes appears to occur primarily via two pathways: (1) al- lopatric speciation of populations of the same ploidy level, and (2) speciation via allopolyploidy (Taylor and Hickey, 1992). Cytological and morphological data suggest that both modes of speciation have occurred in the I. karstenii complex. Reticulate evolution involving diploids and allopolyploids has been well documented in Isoétes (Taylor et al., 1985; Hickey et al., 1989; Taylor and Hickey, 1992), yet autopolyploidy is apparently rare, having been reported only once (Rumsey et al., 1993). Allopolyploids in Isoétes are generally char- acterized by an intermediate morphology with respect to the parental species, especially in megaspore morphology, and additive chromosome numbers and allozyme banding patterns (Taylor and Hickey, 1992). Autopolyploids would be expected to strongly resemble the parental species in all characters except chromosome number and spore size. Because the tetraploid Isoétes palmeri has completely laevigate megaspores and Isoétes allopolyploids appear to always have intermediate megaspore sur- face morphology only those diploid taxa with laevigate megaspores should be considered as possible parents. Based on the results of this study only I. kar- stenii and/or I. precocia are likely to be involved. Although conclusive evidence to address the auto- vs. allopolyploid origin of I. palmeri is unavailable, the data do suggest that the allopolyploid hypoth- esis is more plausible for several reasons. First, as mentioned, autopolyploidy is rare in Isoétes, but allopolyploidy is frequent. Second, although I. karstenii, I. precocia, and I. palmeri are all morphologically similar, I. palmeri is inter- mediate for some characters (e.g., leaf width, length of corneous portion of the SMALL & HICKEY: ISOETES KARSTENII COMPLEX 53 Fic. 4. Hypothesized relationships among the species of the I. karstenii aia Circled X’s represent a hybridization event; boxed 2X’s represents a chromosome doubling ev leaf). Finally, the presence of a triploid in one of the populations of I. palmeri suggests that hybridization occurs between I. palmeri and one (or both) of the diploids. Although this does not bear directly on the origin of I. palmeri, it suggests that hybridization does occur in this group. Isoétes karstenii and I. precocia are both diploids and therefore have their origins in divergent (rather than reticulate) speciation. These two species exist in different habitats: I. karstenii is evergreen and is found in permanent lakes and ponds while I. precocia is ephemeral and is confined to small, seasonal pools which fill during the wet season and desiccate during the dry season. Given the phenetic resemblance and geographic sympatry of these two taxa they are probably closely related. Whether one gave rise to the other or they are each descendants of a now extinct taxon is beyond the scope of this study. Species Relationships Although further taxonomic and biosystematic work needs to be done, we offer a preliminary hypothesis regarding the relationships among the species in this complex (Fig. 4). The base of these relationships rest on the two diploid species, [. karstenii and I. precocia. Hybridization between these two species followed by chromosome doubling in the F, results in the allotetraploid, I. palmeri. The origin of J. hemivelata is less clear, but one possibility is hybridization between a member of the I. karstenii complex and a species outside the com- plex, possibly I. andina or one of several undescribed Colombian or Ecuador- ian species (Hickey, unpublished data). The evidence to support this includes: I. hemivelata’s possession of character states derived from the I. karstenii com- 54 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 2 (2001) plex and from the J. andina complex, along with its undoubtedly polyploid nature. The origin of I. fuliginosa possibly occurred by hybridization between I. hemivelata and another member of the J. karstenii complex. Isoétes fuliginosa is clearly polyploid, shares the character state of reduced velum coverage with I. hemivelata, and is larger than the other members of the I. karstenii complex, but in other respects maintains the characters of the complex. TAXONOMIC TREATMENT Isoétes karstenii Complex Corm globose to horizontally elongate, two lobed (rarely three), size variable; dichotomously branching roots arising from the circumbasal fossa(e). Leaves variable in number, length, and width, flexuous or stiff; alae chartaceous, brown to black, extending upward to 50% of total leaf length, ala apices at- tenuate, acute, truncate or free; subula terete to trigonal, olive green to brown; leaf apices corneous, typically spatulate. Scale leaves, phyllopodia, stomates and peripheral fibrous bundles not seen. Sporangia (ob)ovate to elliptic, tan, concolorous to spotted, basal. Velum incomplete to complete. Ligule deltate to trullate. Labium absent or reduced to a small transverse ridge on the lower lip of the foveola. Megaspores white, 350-700 ym in diameter, laevigate on both proximal and distal faces. Microspores cocoa brown, 25—55 ym in length, echinate. Vegetative reproduction not observed. As discussed previously, these taxa have often been placed in synonymy with, or regarded as varieties of, I. lechleri. They can be distinguished from that species, however, on the basis of ala length/leaf length ratio (up to 50% in the I. karstenii complex vs. 50-75% in I. lechleri), ala apex morphology (attenuate to free in the I. karstenii complex vs. always attenuate in I. Jechleri), vegetative reproduction (common in I. Jechleri vs. absent in the I. karstenii complex), and distribution (J. karstenii complex in Ecuador, Colombia and Venezuela vs. I. lechleri in southern Peru and Bolivia). Additionally, the only member of the I. karstenii complex which strongly resembles J. Jechleri is I. karstenii. These two taxa can be further differentiated on the basis of chro- mosome number: I. karstenii has 2n = 22 and I. lechleri has 2n = 44 (Hickey, 1994). KEY TO SPECIES OF THE I. KARSTENII COMPLEX 1. Velum covering 25-75% of the sporangium 2. Alae extending 10-20% up the leaf length; leaves 150-310 mm long; velum covering 50-75% of the sporangium; microspores 45-55 wm long; megaspores 600—700 pm in CONE Si ci eS Ey atlanta RHEE pel oe ees pascal, Me Bo 1. I. fuliginosa 2. Alae extending 25-45% up the leaf length; leaves 40-200 mm long; velum covering 30-50% of the sporangium; microspores 25-45 ym long; megaspores 300-650 pm in diameter 3. Megaspores 300-450 pm in diameter; microspores 30—40 ym long; subula terete Sea a Se Ns a a re 4b I. karstenii var. anomala SMALL & HICKEY: ISOETES KARSTENII COMPLEX 55 3. Megaspores 500-650 ym in diameter; microspores 35-45 wm long; subula trigonal Be Re eee ag 8 aimee: ost ar es Cor yh Se Pec eT eee ee eS te 2. I. hemivelata 1. Velum covering >90% of the sporangium 4. Leaf apices not or only slightly sclerified (the sclerified part < 0.5 mm long); leaves = than 100 mm long, less than 1.0 mm wide at the middle; plants seasonally ephem- ON co ee he Ata Fe 4 1p. 5's ie Weds cy uy RiSMMR < es 4s Btw aoe oh eee 3. I. precocia 4. Leaf apices distinctly sclerified (the sclerified part 0.7—3.2 mm long); leaves pee ong, 0.5-2.4 mm wide at the middle; plants seasonally persistent . Leaves 25-200 mm long and 1.0-2.5 mm wide with a leaf ge width ratio ee ee Sa ica DES e ee ae cee I. karstenii var. oe . Leaves 100-400 mm reas and 0.5-1.5 mm wide with a leaf length /loaf width r ee ee i sic any wlonei kw lg «Se Mh tais wie eae are ae salvos 1) oa 1. Isoétes fuliginosa R. L. Small and Hickey, sp. nov. (Fig. 5-7; 14-15). Type: “Colombia. Departamento del Cauca: Cordillera Central, Lagunilla de las Casitas, 3700 m alt., 3 Dec 1944” Cuatrecasas 19117 (Holotype: F!; Isotypes: COL!, F!, G-2!, GH!, MO!, MU!, US!). Isoétes arumiana Fuchs-Eckert, nom. nud. Proc. Kon. Ned. Akad. Wetensch. C. 85(2):259. 1982. Based on Cuatrecasas 19117 (US!) Ob statura magna, velo incompleto et megasporis grandibus inter omnes Is- oétes species ab Andibus septentrionalibus peculiaris; Isoétes karstenii A. Braun tangit, ob megasporis laevibus et ob apicibus aliis truncatis versus lib- eris. Corm horizontally elongate, two-lobed, 9-10 (x = 9.7) mm high, 24-25 (x = 24.7) mm wide; dichotomously branching roots arising from the circumbasal fossa. Leaves 18-34, flexuous, to 310 mm long, 8-12 mm wide at base, 1.5— 2.5 mm wide at mid-length, leaf length/width ratio 134; alae chartaceous, brown to blackened, extending 11-17% of total leaf length, apices acute to truncate; subula terete in cross-section, olive green, apex obtuse, not or only slightly corneous at the tip; peripheral fibrous bundles, stomates, scale leaves and phyllopodia absent. Sporangia obovate to ovate, basal, unpigmented, 5.5— 7.8 (x = 7.2) mm long, 4.5—5.8 (x = 5.2) mm wide. Velum incomplete, covering 50—75% of the sporangial surface. Ligule deltate and auriculate, ca. 7 mm high, 5 mm wide. Labium absent or reduced to an extremely small transverse ridge on lower lip of the foveola. Megaspores immature, but ca. 600-700 pm in diameter, laevigate on both proximal and distal faces. Microspores cocoa brown, 45-55 (x = 48.2) pm in length, 30-40 (x = 35.4) ym wide, echinate. Chromosome number unknown. Isoétes fuliginosa is known only from the type collection which was taken from a small lake or pond (‘‘lagunilla’’) at 3700 m altitude in the Cordillera Central, Departamento Cauca, Colombia (Fig. 8). This collection was made in December and contains both microspores and megaspores; however, the mega- spores are immature. Further collections are necessary to ina a the phe- nology and full range of morphological variation of this spec Despite the immaturity of this collection, Isoétes fuliginosa | is oe from 56 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 2 (2001) Fics. 5-7. Isoétes fuliginosa (Cuatrecasas 19117). Fic. 5. Habit view showing reduced alae (F #1351470). Bar = 5 cm. Fic. 6. Microphyll base showing reduced alae and partial velum (US). Bar = 1 cm. Fic. 7. Variation in ala apices: truncate, acute and free respectively (F # 1351471). Bar = 5 mm. other species of Isoétes and can be recognized by its incomplete velum, large growth habit and large mega- and microspores. It is clearly related to the other members of the J. karstenii complex given its laevigate megaspores and trun- cate to free ala apices, although the incomplete velum and apparently high ploidy level indicate that an outside influence is also present. The results of the spore analyses show that I. fuliginosa is polyploid (probably hexaploid or SMALL & HICKEY: ISOETES KARSTENII COMPLEX 57 ie an DS Fan 78°W : 0 4°N e bie ts Ps t N 4b Fic. 8. Distributions of the species of the I. karstenii complex. Open triangles = I. karstenii, —— arena = I. precocia, filled squares = I. palmeri, filled circles = I. hemivelata, and the fuliginos higher) and therefore is likely an allopolyploid derivative of other nothospe- cies. 2. Isoétes hemivelata R.L. Small and Hickey, sp. nov. (Fig. 9-11, 16—17) Type: “Colombia, Departamento del Valle, Cordillera Central, vertiente oc- cidental; cabeceras de los rfos Tulua y Bugalagrande: Paramo de Las Vegas, alt 3600-3800 m, 22 March 1946” Cuatrecasas 20301 (Holotype: F!; Isotypes: M!, NY!, UC!, US!) Haec species ab Sectione Isoete karstenii praeclare distinguitur subulis tri- gonis et velis 30-50% obtegentibus. Corm globose, two-lobed, 4—12 (x = 6.5) mm high, 6-11 (x = 8.1) mm wide; dichotomously branching roots arising from the circumbasal fossa. Leaves 10— 25, stiffly erect, to 200 mm long, 4-6 mm — at base, 1.2-2.4 mm wide at mid-length, leaf length/width ratio 45-82 (x = 64); alae chartaceous, dark brown to black, extending 28-51% (x = 37%) = leaf length, apices acute to 58 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 2 (2001) Fics. 9-13. Isoétes cases: (Cuatrecasas 20301, US) and I. precocia (Oberwinkler & Ober- sides 12968, M). Fic. 9. Habit view of several plants of I. hemivelata. Bar = 2 cm. Fic. 10. Leaf base of I. hemivlt phere partial — agg Bar = 2 mm. Fic. 11. Subula apices of I. pandas showing black, corneous apice: = 2 mm. Fic. 12. Habit view of a mature fertile specimen of I. pideesctosy Bar = 1 cm. Fic. a Leaf base of I. precocia agen a well filled mega- sporangium with a complete velum and the narrow, parallel alae. Bar = 2 m SMALL & HICKEY: ISOETES KARSTENII COMPLEX 59 Fics. 14-19. Isoétes fuliginosa, I. hemivelata, and I. precocia. Fics. 14-15—-SEM of spores of indies fuliginosa (Cuatrecasas 19117, US). Fic. 14. Equatorial view of immature megaspore. Bar = 20 Fic. 15. Proximal view of microspore showing one of the two proximal surfaces. Bar = 20 um. Fics. 16-17—SEM of spores of I. hemivelata (Cuatrecasas 20301, F). Fig. 16. Proximal view of megaspore. Scale as in Fig. 14. Fig. 17. Proximal view of microspore. Scale as in Fig. 15. Fics. 18-19—SEM of spores of I. precocia 2 (Oberwinkler & Oberwinkler 12968, M). Fig. 18. Near _ajesinegeee view of megaspore. Scale as in Fig. 14. Fig. 19. Equatorial view of microspore. Scale as in Fig. 1 60 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 2 (2001) free; subula + trigonal, olive green; leaf apices blunt, often spatulate, distal 0.75—1.75 mm of apex corneous; peripheral fibrous bundles, stomates, scale leaves and phyllopodia absent. Sporangia elliptic to obovate, basal, spotted, 1.5-3.2 mm long, 1-2 mm wide. Velum incomplete, covering 30-50 % of the sporangial surface. Ligule shallowly deltate to trullate, to 1.5 mm long, to 2 mm wide. Labium absent or reduced to a transverse ridge on the lower lip of the foveola. Megaspores white, 500-650 (x = 560) pm in diameter, laevigate on both proximal and distal faces. Microspores cocoa brown, 35—40 (x = 38.5) um long, 25-30 (x = 27.6) wm wide, echinate. Chromosome number unknown. Isoétes hemivelata grows as a submerged aquatic in permanent lakes and ponds (occasionally streams) between 3300-4300 m. It is distributed through- out the central and southern Departamentos of the Colombian Andes including Meta, Risaralda, Huila, Cauca and Valle (Fig. 8). Although known from only a few collections, those included in this study were made in January, March, August and October and all contained both mega- and microspores. This suggests that this species is evergreen (as are most Andean Isoétes) and produces both mega- and microspores throughout the year. Isoétes hemivelata is one of the most distinctive elements within the I. kar- stenii complex given its trigonal subula and incomplete velum. It can be dis- tinguished from I. karstenii, I. palmeri, and I. precocia by these characters and its larger spores. It can be distinguished from I. fuliginosa by its trigonal sub- ula, smaller spores, and greater ala length / leaf length ratio. Analyses of spore size suggest that I. hemivelata is polyploid. The characters it shares with the I. karstenii complex indicate that one of the parental species comes from this complex; however, the trigonal subula and reduced velum coverage must have come from a source outside this complex. This suggests hybridization and polyploidization. One likely parent is I. andina Spruce ex Hook., a sympatric species with strongly trigonal leaves and an incomplete velum. One collection seen by us (Cuatrecasas 19058 F!, GH!, US!) apparently represents a collection of hybrids between a member of the I. karstenii com- plex and I. andina. This collection is highly variable for most characters and spore abortion is evident. While most megaspores of this collection are slightly to distinctly papillate, some are almost laevigate and all individuals of the collection have ~ trigonal leaves. Chromosome doubling in a hybrid such as this, or backcrossing to the J. karstenii complex parent could feasibly result in a taxon which strongly resembles I. hemivelata. Additional specimens examined: COLOMBIA. Departamentos Huila/Cauca: Macizo Colombiano; péramo de Las Papas, Cerros y alrededores de la laguna La Magdalena, alt. 3530 m, 16 Oct 1958, Idrobo, et al. 3307 (COL, NY). Departamento Meta: Péaramo de Sumapaz, Hoya Sitiales, Laguna La Primavera y alrededores, alt. 3510 m, 9 Jan 1973, Cleef 7555 (COL, U); Paramo de Sumapaz, Hoya El Nevado, Laguna La Guitarra y alrededores, alt. 3425 m, 23 Jan 1973, Cleef 8268 (COL, U). Departamento Risaralda: Cordillera Central; Municipie de Pereira; parque de los Nevados, alt 4250 m, Jaramillo, et al. 5700 (COL, U). Departamento del Valle: Péramo Pan de Azucar, alt 3300-3700 m, 23 Aug 1968, Espinal & Ramos 2471 (COL). SMALL & HICKEY: ISOETES KARSTENII COMPLEX 61 3. Isoétes precocia R.L. Small and Hickey, sp. nov. (Fig. 12-13, 18-19) Type: “Venezuela: Anden, Estado Mérida: Sierra de St. Domingo, Paramo de Mucubaji, Weg zur Laguna Negra, an stark vernaéBten Stellen, + 3500 m. 4. 10. 1968.” B & F Oberwinkler 12968 (Holotype: M!; Isotype: VEN!). Isoétes socia sensu Fuchs-Eckert 1982 pro parte, Proc. Kon. Ned. Akad. We- tensch. C. 85(2): 256 1982. non A. Braun. Species nova I. karstenii proxima, cujus megasporas laeves et velum com- pletum habet; differt staturo minore, apicibus foliis brevissimis corneis et ha- bitu ephemero. Corm globose to horizontally elongate, two lobed, 1.8—2.4 mm (x = 2.1) high, 2.5—4.6 mm (x = 3.6) wide; dichotomously branching roots arising from the circumbasal fossa. Leaves 6-11, flexuous, to 63 mm long, to 4 mm wide at base, 0.5 to 0.9 mm wide at mid-length, leaf length / width ratio 45-157 (x = 80); alae chartaceous, caramel-colored, extending 26-36% of total leaf length, apices acute to truncate; subula essentially terete in cross-section, pale green, apices blunt, corneous only at 0.1-0.2 mm of the tip; peripheral fibrous bun- dles, stomates, scale leaves and phyllopodia absent. Sporangia elliptic to cir- cular, basal, unpigmented, 1.6—2.7 mm (x = 2.1) long, 1.2-2.1 mm (x = 1.7) wide. Velum complete, covering 100% of the sporangial surface (rarely slightly less). Ligule ephemeral, not seen. Labium absent or reduced to an extremely small transverse ridge of tissue on lower lip of the foveola. Megaspores white, lustrous, 388—495 (x = 437) pm in diameter, laevigate on both proximal and distal faces. Microspores cocoa brown, 27-30 (x = 29) ym in length 18.9—23.3 (x = 20.5) um wide, echinate. Chromosome number: 2n = 22. Isoétes precocia grows submerged in shallow, temporary pools at 3300—4400 m. It is currently known from Estado Mérida, Venezuela and Departamentos Boyaca, Cundinamarca and Meta in Colombia (Fig. 8). Although only eleven collections of this species have been found among the specimens examined, it is probably a common plant of seasonal pools but has not been collected due to its diminutive size and ephemeral nature. We visited the type locality in March of 1992, yet were unable to find this species because the pools were dry. Collections examined in this study were made in March, April, July, Oc- tober, November and December, and all contained mega- and microspores. Fuchs-Eckert (1982) recognized that a taxon resembling the type of I. socia existed, but failed to differentiate between the vegetative plantlets of I. lechleri in Peru and Bolivia and the distinct species (J. precocia) of Colombia and Venezuela. Hickey (1985) recognized that a distinct species existed, but did not study the associated specimens and based his acceptance of the name I. socia on the work of Fuchs-Eckert (1982). Hickey (1994) has subsequently placed I. socia in synonymy with I. lechleri. Additional specimens examined: COLOMBIA. Departamento Boyaca: Sierra Nevada del Cocuy, Alto Valle Lagunillas, 1.5 km NE de la Laguna Pintada, 4390 m, 30 Nov 1972, Cleef 5704 (U); Sierra Nevada del Cocuy, Boqueron de Cusiri, alt. 4310 m, 5 Mar 1973, Cleef 8785 (U); Paéramo de la Sarna entre Sogamoso y Vado Hondo, alt. 3330 m., 8 Apr 1973, Cleef 9522 (COL, U). Departamento Cundinamarca: Péramo de 62 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 2 (2001) Sumapaz, Chisaca, 100 m SW de la Laguna Larga, alt. 3700 m, 11 Dec 1971, Cleef 153 (COL, U); Péramo de Cruz Verde, alt. 3200 m, 16 Oct 1961, Idrobo & Hotheway 4174 (COL); Péramo de az, vicinity of Laguna Chisaca, pool 50 m S of Laguna Larga, alt. 3620 m, 7 Jul 1986, Keeley & Keeley 11032 (MU); Péramo de Sumapaz, vicinity of Laguna Chisaca, pool 50 m S of Lagun Larga, alt. 3620 m, 7 Jul 1986, Keeley & Keeley 11033 (MU). Departamento Meta: Pdéramo de Sumapaz, Cerro Nevado del Sumapaz, alt. 3725 m, 21 Jan 1973, Cleef 8185A (COL). VENEZUELA. Estado Mérida: Sierra de Sto. Domingo, Péramo de Mucubaji, alrededores de la Laguna Grande, alt. 3560-3600 m 19 Nov 1959, Barclay & Juajibioy 9588 (COL, GH). 4a. Isoétes karstenii A. Braun var. karstenii. Verh. Bot. Vereins Prov. Brandenberg 3/4:332. 1862. Figs. 20-21. Type: “Venezuela. Im Gebirge von Mérida 8000’ hoch in einem See ganz unter Wasser 1857.” Karsten s.n. (Holotype B!). Isoétes lechleri var. colombiana T. C. Palmer, Amer. Fern J. 19: 18. 1929. I. colombiana (T. C. Palmer) H. P. Fuchs, Proc. Kon. Ned. Akad. Wetensch. C. 85(2):259. 1982. Type: “Colombia, Dept. Santander: Laguna de Cunta, edge of Paéramo de Santurban; alt. 3,880 m, 21 Jan 1927” Killip & Smith 17964 (Holotype PH; Isotype: NY!), mixed collection. Isoétes steyermarkii Fuchs-Eckert, nom. nud. Proc. Koninkl. Ned. Acad. We- tensch. C85:257. 1982. Based on Steyermark 55904 (US!). Corm globose to horizontally elongate, two lobed (rarely 3), 2.0-4.0 mm (x = 2.8) high, 2.8-7.8 mm (x = 5.2) wide; dichotomously branching roots arising from the circumbasal fossa(e). Leaves 8—21, flexuous to stiffly erect, to 251 mm long, 5-10 mm wide at base, 0.9 to 2.4 mm wide at mid-length, leaf length/ width ratio 35-157 (x = 77); alae chartaceous, light-brown to blackish-brown, extending 14-38% of total leaf length, apices acute to truncate; subula essen- tially terete in cross-section, olive green, apices blunt, often spatulate, distal 0.5-3.2 mm of apex corneous; peripheral fibrous bundles, stomates, scale leaves and phyllopodia absent. Sporangia widely elliptic to circular, basal, unpigmented, 2.1—5.5 (x = 3.4) mm long, 1.8-3.9 (x = 2.6) mm wide. Velum complete, covering 100% of the sporangial surface (rarely slightly less). Ligule deltate, auriculate, 0.5—-1.5 mm high, 0.75—2.0 mm wide. Labium absent or reduced to an extremely small transverse ridge on lower lip of the foveola. Megaspores white, lustrous, 400-507 (x = 467) um in diameter, laevigate on both proximal and distal faces. Microspores cocoa brown, 25-38 (x = 30) pm in length, 17.7—28.7 (x = 20.8) ym wide, echinate. Chromosome number: 2n = 22. Isoétes karstenii grows submerged in permanent lakes and ponds (occasion- ally streams) between 3300-4600 m. It is distributed throughout the eastern cordillera of the northern Andes from Estado Mérida, Venezuela, to Departa- mentos Arauca, Boyaca, Cundinamarca, Meta, Tolima and Narifio in Colombia (Fig. 8). The collections of this species were made from every month except August and October. With the exception of the single collection made in June, SMALL & HICKEY: ISOETES KARSTENII COMPLEX 63 Fics. 20-25. Isoétes karstenii var. karstenii, I. —— var. anomala, and I. palmeri. Fics. 20— O id ogee 13685, M). Bar = 20 pm. Fics. 22—-23—SEM of spores of I. karstenii var. anomala (Rimbach 71, US). Fic. 22. Distal view of ete Scale as in Fig. 20. Fic. 23. Equatorial view of micro- spore. Scale as in Fig. 23. Fics. 24-25—SEM of spores of I. palmeri (Cleef 8647, BM). Fic. 24. Distal view of megaspore. Scale as in Fig. 20. Fic. 25. Equatorial view of microspore. Scale as in Fig. 21 64 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 2 (2001) specimens from all other months contained both mega- and microspores, sug- gesting this species is evergreen and fertile year-round. This species is the most widely distributed (or at least most widely collected) member of this group. Isoétes colombiana (T. C. Palmer) H. P. Fuchs and I. steyermarkii H. P. Fuchs nom. nud. are synonyms of I. karstenii. Palmer (1929) originally separated I. lechleri var. colombiana from I. lechleri var. lechleri on the basis of the absence of an equatorial ridge on the megaspores, the small size of the microspores, and the corneous leaf apices. All of these characters agree with the current circumscription of J. karstenii. In addition, the type collection of I. lechleri var. colombiana is a mixture of fertile J. karstenii and plants of apparent hybrid origin. The hybrids are the specimens that contain ornamented and misshapen megaspores as noted in Palmer’s (1929) protologue. Isoétes steyermarkii was invalidly published by Fuchs-Eckert (1982) and no description was given. Ex- amination of the basis for this taxon (Steyermark 55904) reveals no characters that differentiate it from I. karstenii; in fact this collection is nearly identical to the type of I. karstenii. Finally, the type specimen of I. karstenii as cited by Braun (1862) is “In einem ungefahr 8000’ hoch See auf dem Péramo im Gebirge von Mérida (Ven- ezuela) ganz unter Wasser im Jahre 1853 von Prof. H. Karsten entdeckt,. . .” The specimen at Berlin which was annotated as the holotype by H. P. Fuchs bears the label “Jsoétes karstenii A. Braun. Venezuela. Im Gebirge von Mérida 8000’ hoch in einem See ganz unter Wasser. C. Dr. Karsten etc. 1857.” Al- though the locality information on this label closely matches that of Braun’s (1862) description, there is a discrepancy regarding the date of collection: 1853 in the description and 1857 on the specimen label. Stafleu and Cowan (1979) indicate that Karsten was in South America from 1848 to 1856, but returned to Berlin in 1856 where he remained until 1868. Thus, he could not have collected this specimen in 1857. A potential explanation for this discrepancy could be that 1853 was the date of collection, while 1857 was the date the material was accessioned into the Berlin herbarium. Additional specimens examined: COLOMBIA. Departamento Arauca: Cordillera Oriental, extremo sur de la Sierra Nevada del Co- cuy, Laguna el Amarillal, alt. 4065 m, 23 Mar 1977, Cleef & van der Hammen 10383 (U); Sierra Nevada del Cocuy, Laguna de la Plaza, alt. 4550 m, 20-30 Jan 1959, van der Hammen & Gonzalez 1374 (COL); Sierra Nevada del Cocuy, Laguna la Plaza, alt. 4150 m, Mar 1977, van der Hammen 4753 (COL, U). Departamento Boyaca: Sierra Nevada del Cocuy, Alto Valle Lagunillas, Laguna Cuadrada, alt. 4060 m, 26 Sept. 1972, Cleef & Florschtitz 5590 (COL, U); Sierra Nevada del Cocuy, Paramo Concavo, 3 km al N del morro Pulpito del Diable, alt. 4410 m, 27 Feb 1973, Cleef 8599 (U); Sierra Nevada del Cocuy, Alto Valle Lagunillas, medio km al WNW de la Laguna Pintada, alt. 4000 m, 4 Mar 1973, Cleef 8782 (COL, U); Municipio de El Cocuy, valle del rio Lagunillas, Sierra Nevada del Cocuy, Laguna La Pintada, alt. 3800 m, 23 May 1993, Small, Gonzalez & Ruiz 141- 143 (COL, MU); Municipio de El Cocuy, valle del rio Lagunillas, Sierra Nevada del Cocuy, Laguna La Cuadrada, alt. 3900 m, 23 May 1993, Small, Gonzalez & Ruiz 144 (COL, MU). Departamento Cundinamarca: Paéramo de Sumapaz, Chisaca, Laguna Negra, alt. 3750 m, 11 Dec 1971, Cleef 174 (COL, U); Km 31, south of Usme, Laguna Chisaca, alt. 3620 m, 3 Jan 1985, Keeley & Keeley 7908 (COL, MU); Km 31, south of Usme, Laguna Chisaca, alt. 3620 m, 3 Jan 1985, Keeley & Keeley 7909 (COL, MU); Small laguna adjacent to north side of road at km 31 between Laguna Chisaca and Laguna Larga, alt. 3650 m, 10 Dec 1985, Keeley & Keeley 10076 (MO, MU); Unnamed laguna on SMALL & HICKEY: ISOETES KARSTENII COMPLEX 65 opposite side of road from largest laguna at km 30.5, Paramo de Sumapaz, vicinity of Laguna Chisaca, alt. 3650 m, 10 Dec 1985, Keeley & Keeley vritt (MO, MU); Péramo de Sumapaz, laguna at km 30.5 opposite side of road from Laguna Chisaca, alt. 3700 m, 15 Dec 1985, Keeley & Keeley 10093 (MU); Paéaramo de Sumapaz, vicinity of Laguna Chisaca, laguna adjacent to road at km 31, alt. 3620 m, 7 Jul 1986, Keeley & Keeley 11037 (MU); Paramo, alt. 3700 m, Apr 1963, Saravia 2533 (COL). Departamento Meta: Paramo de Sumapaz, Laguna La Guitarra, alt. 3460 m, 21 Jan 1972, Cleef 832 (COL, U); sti de Sumapaz, Laguna La Primavera, alt. 3525 m, 25 Jan 1972, Cleef 1017 (COL, U); Paéramo de Sumapaz, Cerro Nevado del Sumapaz, alt. 4050 m, 30 Jan 1972, Cleef 1351 (COL, U); Péramo de Sumapaz, Cerro Nevado del Sumapaz, alt. 4090 m, 13 Jan 1973, Cleef 7766 (COL, U); Péaramo de Sumapaz, Laguna Los Sitiales, alt. 3620 m, 22 Jan 1973, Cleef 8248 (COL, U); Macizo de Sumapaz, Boqueron del Palacio, norte de Cerro del Nevado, alt. 3900 m, 5 Jul 1981, Diaz, et al. 2558 (COL); Macizo de Sumapaz, Laguna La Guitarra, alt. 3370 m, 2 Jun 1981, Diaz & Rangel 2390 (COL). Departamento Narifio: Cumbal, Laguna La Bolsa, alt. 3400 m, 21 Jan 1973, Hagemann & Leist 1968 (COL); Laguna Negra en vecindades de Pasto, alt. 3520 m, 17 Jul 1973, Leist & Méhle 2212 (COL) VENEZUELA. Estado Mérida: Laguna Grande de Mucubaji, alt. 3535 m, 23 Dec 1984, Keeley & Keeley 7861 (MU); Laguna Grande de Mucubaji, alt. 3535 m, 23 Dec 1984, Keeley & Keeley 7862 (MU); Laguna Negra, alt. 3430 m, 23 Dec 1984, Keeley & Keeley 7865 (MU); Laguna de los Patos, alt. 3670 m, 23 Dec 1984, Keeley & Keeley 7868 (MU); Laguna Negra, alt. 3400 m, 22 Nov. 1968, Oberwinkler & Oberwinkler 13685 (M); Laguna St. Barbara, alt. 3700 m, 11 Jan 1969, Oberwinkler & Oberwinkler 14272 (M); Laguna St. Barbara, alt. 3700 m, 11 Jan 1969, Oberwinkler & Oberwinkler 14274 (M); Small pond ca. 8 km N of Pico de Aguila on road to Pinango, alt. 3500 m, 15 Mar 1992, Small & Hickey 4 (MU); Small pond ca. 15 km N of Pico de Aguila on road to Pinango, alt. 3500 m, 15 Mar 1992, Small & Hickey 5 (MU); Small pond ca. 15 km N of Pico de Aguila on road to Pinango, alt. 3500 m, 15 Mar 1992, Small & Hickey 6 (MU); Laguna Negra, alt. 3500 m, 17 Mar 1992, Small & Hickey 7 (MU); Laguna Mucubaji, alt. 3500 m, 17 Mar 1992, Small & Hickey 8 (MU); Near upper limits of paramo, around small lake between Chachapo and Los Apartaderos, near El Aguila, 3930 m, 15 Apr 1944, Steyermark 55904 (F, US, VEN); Laguna Grande de Mucubaji, alt. 3500 m 19 Sep 1961, Tryon & Tryon 5839 (GH). 4b. Isoétes karstenii var. anomala (T. C. Palmer) R.L. Small and Hickey, comb. nov. Figs. 22-23. Isoétes lechleri var. anomala T. C. Palmer, Amer. Fern J. 22: 130-131. 1932. Isoétes rimbachiana H. P. Fuchs Proc. Kon. Ned. Akad. Wetensch. C. 85(2): 259. 1982. Type: “Ecuador. Mt. Chimborazo. Submerged in water of shallow pond. Rooting in mud. 4200 m.” Rimbach 171 (Holotype US!) Isoétes karstenii var. anomala differs from I. karstenii var. karstenii only in the character of velum coverage: 25-75% in var. anomala, >90% in var. kar- stenii. In Palmer’s (1932) description of this taxon (as I. lechleri var. anomala) he distinguishes it from other “I. Jechleri’” (of which he considered I. karstenii to be a synonym) based primarily on the presence of a two-layered velum which incompletely covered the sporangium. Examination of the holotype, however, revealed no indication of a two-layered velum on the sporophylls we observed. The presence of a “multilayered velum” has been reported else- where (e.g., Isoétes prototypus, Britton and Goltz, 1991). Initial investigations indicate that the apparent multilayered velum is actually due to disintegration of the mesophyll of the velum and separation of the upper and lower epider- mal layers resulting in what appears as a double velum (Hickey, unpublished data). Regardless, this taxon appears unique in the extent of velum coverage. 66 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 2 (2001) It is otherwise nearly identical to the type of I. karstenii, as noted by Palmer (1932). This taxon is known only from the type locality (Mt. Chimborazo, Ec- uador) although Rodriguez (1955) cites one specimen from Laguna Negra, Mé- rida, Venezuela as I. Jechleri var. anomala. 5. Isoétes palmeri H. P. Fuchs. Proc. Kon. Ned. Akad. Wetensch. C. 84 (2): 168-173. 1981. Figs. 24-25. Type: “Venezuela; Laguna Grande de Apartaderos, Mérida, floating on water, 21 January 1929” Pittier 13242 (Holotype US!; Isotypes G!, M!, MO!, VEN!). Isoétes cleefii H. P. Fuchs Proc. Kon. Ned. Akad. Wetensch. C. 84 (2):177-181. 981. Type: “Colombia, Cundinamarca: Péramo de Sumapaz, péramo y bosque alto-andino cerca de Lagunitas al S. de San Juan. Laguna Gober- nador: orilla W.” Cleef 8308 (Holotype COL!; Isotype COL!). Corm globose to horizontally elongate, two lobed, 1.3-6.0 mm (x = 2.9) high, 3.0-8.5 mm (x = 5.0) wide; dichotomously branching roots arising from the circumbasal fossa. Leaves 5-25, flexuous, to 350 mm long, 5-8 mm wide at base, 0.5 to 1.5 mm wide at mid-length, leaf length/width ratio 118-350 (x = 241); alae chartaceous, dark brown to blackened distally, pale proximally where leaf was imbedded in substrate, extending 10-27% of total leaf length, apices acute to truncate; subula essentially terete in cross-section, pale green, apices blunt, often spatulate, distal 1.0-2.5 mm of apex sclerified and black- ened; peripheral fibrous bundles, stomates, scale leaves and phyllopodia ab- sent. Sporangia elliptic to obovate, basal, unpigmented, 2.5-5.9 (x = 3.8) mm long, 1.9-4.0 (x = 2.8) mm wide. Velum complete, covering 100% of the spo- rangial surface (rarely slightly less). Ligule deltate, auriculate, to 1 mm high, to 1 mm wide. Labium absent or reduced to an extremely small transverse ridge on lower lip of the foveola. Megaspores white, lustrous, 380-550 (x = 460) jm in diameter, laevigate on both proximal and distal faces. Microspores cocoa brown, 26—36 (x = 30.4) um in length, 17.9-26.4 (x = 21.3) wm wide, echinate. Chromosome number: 2n = 44. Isoétes palmeri grows submerged in permanent pools, ponds, and lakes be- tween 3340-4435 m. It is distributed in the eastern cordillera of the northern Andes from Estado Mérida, Venezuela, through the Departamentos Arauca, Boyaca, Cundinamarca and Meta of Colombia (Fig. 8). The presence of both microspores and megaspores in all months for which collections are available (nine out of twelve) suggests that this species is ev- ergreen and fertile year-round. Isoétes palmeri'’s affiliation with the I. karstenii complex is apparent by its laevigate megaspores, acute to truncate alae apices, and corneous leaf tips. For the most part, I. palmeri and I. karstenii are easily differentiated morphologi- cally; I. palmeri has long, slender leaves which appear flexuous, and JI. kar- stenii has shorter, wider leaves which usually appear rigid and stiff. Occa- sional collections appear intermediate between these two extremes. These col- SMALL & HICKEY: ISOETES KARSTENII COMPLEX 67 lections often represent juvenile individuals of either species, but may also include hybrids. Isoétes cleefii is a synonym of I. palmeri. Although the holotype and one of the paratypes of I. cleefii have leaves that are slightly longer and wider than typical I. palmeri, these individuals are merely extreme variants within a wide range of intraspecific variation. Both I. palmeri and I. cleefii were published by Fuchs-Eckert (1981a,b) at the same time, thus neither name has priority. We select the name I. palmeri to represent this species because the type col- lection is more representative of the species. Additional speciemens examined: COLOMBIA. D t A Si Nevada del C , Cabeceras de la Quebrada E] Playon alt. 4245 m, 11 Mar 1973, ape’ 9088 (COL, U). Departamento Boyaca: Paéramos al NW de Belen, ca. 600 m NE de la Laguna el Alcohol, alt. 3850 m, 29 Feb 1972, Cleef 2059 (COL, U); Péramo de la Rusia, NW-N de Duitama, alt. 3570 m, 13, Dec 1972, Cleef 7144 (COL, U); Sierra ands del Cocuy, Péramo Concavo, alt. 4435 m, 28 Feb 1973, saocd 8647 (COL, U); Paramo de la Rusia, NW- N de Duitama, Laguna Negra, alt. 3745 m, 14 Dec 1972, Cleef 7221 (COL, U); Sierra act del Cocuy, Péramo Concavo, 2.5 km al N de morro Pali del Diablo, alt. 4435 m, 28 Feb 1973, Cleef 8648 (COL, U); Péramo entre Pefia Arnical y Alto de Mogotes; lagunita 0.5 km al SE de la Laguna Grande, alt. 3340 m, 1 Apr 1973, Cleef 9273 (COL, U); Paéramo entre Pena de Arnical y Alto de Mogotes, cerca de la Laguna Grande, alt. 3350 m, 10 Apr 1973, Cleef 9551 (COL, U); Péramos al Belen, vereda San Jose de la Montana, alt. 3725 m, 3 May 1973, Cleef 9699 (U); Péramo de Pisva, carretera Socha-La Punta km 77, Laguna Colorada, alt. 3425 m, 22 May 1973, Cleef 9898 (COL, U); Largest laguna, 17 km NW of Belen on road to San Jose de peace aes 3700 m, 1 Jan 1985, Keeley & Keeley 7887 (COL, MU). Departamento Cundinamarca: Paramo entre ean y San Cayetano, Laguna Verde, 3600 m, 30 Nov 1971, Cleef 80 (COL,U); Péramo de PUBS hoya la Laguna Larga, alt. 3750 m, 14 Dec 1 1971, Cleef & Jamarillo 262 (COL, U); Péramo de Palacio, Lagunas Buitrago, alt. 3600 m, 16 Dec 1971, Cleef 296 (COL, U); Paéramo de Sumapaz, Andabobos, alt. 3800 m, 9 Feb 1972, Cleef 1553 (COL, U); Péramo de Palacio, Lagunas de Buitrago, alt. 3580 m, 25 May 1972, Cleef 4112 (COL, U); Pa oO az, lagunitas al S de San Juan, La Gobernador, alt. 3815 m, 26 Jan 1973, Cleef 8308 (COL); Paramo de Chingaza, cerca de la “Palla” 10-20 Jan 1965, Huertas & Camargo 6025 (COL); Cordillera Oriental, Macizo de Sumapaz, Laguna de Chisaca, alt. 3680 m, 14 Apr 1958, Idrobo 2741 (COL); Laguna Seca, N of Zipaquire, alt. 3700 m, 30 Dec 1984, Keeley & Keeley 7876 (MU); Laguna Seca, N of Zipaquire, alt. 3700 m, 30 Dec 1984, Keeley & Keeley 7878 (COL, MU); Péramo de Sumapaz, vicinity of Laguna Chisaca, Laguna Larga, alt. 3650 m, 10 Dec 1985, Keeley & Keeley 10072 (MO, MU); Paéramo de Sumapaz, vicinity of Laguna Chisaca, Laguna Larga, alt. 3650 m, 10 Dec 1985, Keeley & Keeley 10073 (MO, MU); Pdéramo de Sumapaz, vicinity of Laguna Chisaca, small pools S of Laguna Larga, alt. 3650 m, 10 Dec 1985, Keeley & Keeley 10077 (MO, MU); Péramo de Sumapaz, vicinity of Laguna Chisaca, small pools ca. 100 m above SE end of Laguna Chisaca, alt. 3700 m, 13 Dec 1985, Keeley & Keeley 10088 (MO, MU); Paéramo de Sumapaz, Laguna Chisaca, alt. 3700 m, 13 Dec 1985, Keeley & Keeley 10090 (MU); Péramo de Sumapaz, vicinity of Laguna Chisaca, small pool ca. 200 m above SE end of laguna on NE side of road at km 30.5, alt. 3700 m, 13 Dec 1985, Keeley & Keeley 10091 (MU); Pdramo de Sumapaz, vicinity of Laguna Chisaca, laguna ca 100 m NE of agunita, alt. 3700 m, 13 Dec 1985, Keeley & Keeley 10092 (MO, MU); Paéramo de Sumapaz, agian of Laguna — aca, Laguna Larga, alt. 3620 m, 7 Jul 1986, Keeley & Keeley 11029 (MU); P. o de Sum vicinity of Laguna Chisaca, shallow pool on ridge SE of first lake, alt. 3620 m, 7 7 jal 1986, ee & Keeley 11039 (MU); Carretera Cogua-San Cayetano, Péramo de Laguna Seca, alt. 3660 m, 3 Aug 1970, Piedrahita 276 (COL). Departamento Meta: Péramo de Sumapaz, Laguna la Primavera, alt. 3510 m, 9 Jan 1973, Cleef 7556 (COL, U); Paramo de Sumapaz, Lagunas el Sorbedero y El Nevado, alt. 3635 m, 16 Jan 1973, Cleef re (COL, U); Péramo de Sumapaz, Laguna la Guitarra, alt. 3425 m, 23 Jan 1973, Cleef 8267 (COL 68 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 2 (2001) ACKNOWLEDGMENTS We thank the staff of the Herbario Nacional de Colombia, especially Favio Gonzalez and Maria pi Murillo, for assistance with field and herbarium work in Colombia; A. Clive Jermy for This work was supported by two sre se from the State of Ohio Academic Challenge Program and W.S. Turrell Herbarium Fund Grant # LITERATURE CITED BAKER, J. G. 1880. A synopsis of the species of Isoetes. J. Bot. Br. Foreign 18:65—70, 105-110. BRAUN, A. 1862. aie uber einige auslandische Arten der Gattung Isoétes. Verh. Bot. Ver. Prov. 333. BriTTON, D. M., and J. P. GOLTz. 1991. Isoetes prototypus, a new diploid species from eastern Canada. Can. J. Bot. 69:277—281. BRUNTON, D. F., and D. M. BrirTon. 1993. Isoetes prototypus (Isoetaceae) in the United States. BRUNTON, D. F., and D. M. BRITTON. 1996a. Taxonomy and distribution of Isoetes valida. Amer. Fern. J. 86:16—25. BRUNTON, D. F., and D. M. BRITTON. 1996b. The status, ns gepanet wie identification of Georgia Quillwort viatiny georgiana; Isoetaceae). Amer. Fern. J. 86:105—113. BRUNTON, D. F., D. M. BRITTON, and W. C. TayLor. 1994. ae i roar sp. nov. (Isoetaceae):a new pares from the southeastern United States. Castanea 2-21. Cox, P. A., and R. J. HICKEy. 1984. Convergent megaspore use a Isoetes. Amer. Naturalist 124:437—441., FUCHS-ECKERT, H. P. 1981a. Isoétes Palmeri H. P. Fuchs, eine neue Isoétes-Art des Péramo. Proc. Kon. Ned. Akad. Wetensch. C. 84 (2):165-174. FUCHS-ECKERT, H. P. 1981b. Isoétes psig H. P. Fuchs, eine weitere neue Isoétes-Art aus dem kolumbianischen Péramo. Proc. Kon. Ned. Akad. Wetensch. C. 84:175-182 FucHs-ECKERT, H. P. 1982. Zur dead annhes von Vorkommen und Verbreitng der siidamer- ee Isoétes-Arten Proc. Kon. Ned. Akad. Wetensch. C. 85:2 FUCHS-ECKER . 1992. Supplementum ad indicem Isoétalium. ee 12: 99-159. Gomez, L. D. 1980. Vegetative reproduction in a Central American Isoétes (Isoétaceae). Its mor- ho. ny ee reno and taxonomical implications. Brenesia 18:1-14 Hickey, R. J. 1 Chromosome numbers in neotropical Isoétes. Amer. HICKEY, R. J. si aon studies of neotropical Isoétes. Ph.D. aoniain, 8s sae of Connecticut, HICKEY, R. J. 1994. 28. solace, in Esebehicraas dey of Peru, part VI, (R. M. Tryon and R. G. Stoltze, eds.) aie g oy Bot. N.S. 3 Hickey, R. J., W. C. TAYLOR, ioe “s ma LUEBKE. 1989. The species concept in Pteridophyta with csgeoes sei to Isoétes. Am. Fern. J. 79:78-89. Korr, L. 5., D. M. BRITTON. 1983. Spore morphology and taxonomy of Isoétes in northeastern No 6 Maudie. Can. J. Bot. 61:3140-31 Kort, L. S., and D. M. BRITTON. 1985. Role of morphological pera ata of leaves and the — region in the taxonomy of Isoétes in northeastern North America. Amer. Fern. J. ‘44-55 ainan G. HL 1859. Filices Lechlerianae, fasc. 2 MOTELAY, L., and. A. VENDRYES. 1882. Seaiuahie fo Jandtonn: Act. Soc. Linn. Bordeaux 36:309-— 405. MUSSELMAN, L. J., and D. A. KNEPPER. 1994. Quillworts of per aci Amer. Fern. J. 84:48-68. PALMER, T. C. 1929. lncecions Lechleri Mett. Amer. Fern. J. 1 PALMER, T. C, 1932. More about Isoetes Lechleri Mett. ioe dss J. 22:129-132. PFEIFFER, N. E. 1922. Monograph of the Isoetaceae. Ann. Missouri Bot. Gard. 9:79-232. SMALL & HICKEY: ISOETES KARSTENII COMPLEX 69 RADFORD, A. E., W. C.DICKISON, and C. R. BELL. 1972. Vascular Plant Systematics. Univ. North Carolina, Chapel Hill. RODRIGUEZ, G. 1955. Revision del genero Isoetes en Venezuela. Bol. Mus. Cienc. Nat. I ROHLF, F. J. 1993. NTSYS-pc: numerical taxonomy and multivariate analysis system, ante 1. 80. Exeter seit Applied Biostatistics, es Setauket, NY. pied F. J., P. THOMPSON, and E. SHEFFIELD. 1993. Tiploid Isoetes echinospora (Isoetaceae; Pter- dophyta) in northern England. Fern cat 14:2 Rury, P. M. 1978. A new and unique, sateomibeay op s-grass (Isoétes) from Georgia. Amer. Fern STAFLEU, F. A., and R. S. Cowan. 1979. Taxonomic Literature, 2nd ed. Vol. II: H-Le. Regnum Vegetable Vol. 98. Bohn, Scheltema & Holkema, Utrecht dr. W. Junk b.v., Publishers, The ague. i W. C., and R. J. Hickey. 1992. Habitat, evolution and speciation in Jsoetes. Ann. Missouri. t. Gard. 79:613-622. es W. C., and N. T. LUEBKE. 1986. Germinating spores and growing sporelings of aquatic Isoétes. Amer. Fern. J. 76:21-—24. TAYLOR, W. C., N. T. LUEBKE, D. M. BRITTON, R. J. Hickey, and D. F. BRUNTON. 1993. seater) Pp. 64-75 in: Flora of North America, Vol. 2, Pteridophytes and Gymnosperms. Oxford U versity Press, New York TAYLOR, W. C., N. T. LUEBKE, sia M. B. SMITH. 1985. Speciation and arenes in North Amer- ican eee age Roy. Soc. Edinburgh Sect. B (Biol. Sci.) 86B: _ UNDERWOOD, L. M. 1 The distribution of Isoetes. Bot. Gaz. April: 8 VARESCHI, V. 1968. Sie Rn in: T. Lasser (ed.), Helechos, Flora de eo ae Vol 1(1 WaTANABE, M., M. TAKAMIYA, T. MATSUSAKA, and K. ONO. 1996. Biosystematic studies on the genus Isoetes (Isoetaceae) in Japan. III. Variability within qualitative and quantitative mor- phology of spores. J. Plant Res. 109:2 WenrR, U. 1922. Anatomie und Systematik a Gattung Isoétes L. Nova Hedwigia 63:219-262. American Fern Journal 91(2):70—72 (2001) SHORTER NOTE Additions and Corrections to the Pteridophyte Flora of Northeastern Argen- tina.—Two new records as well as material to confirm two non-vouchered records were collected by colleagues at the Instituto Darwinion and Parques Nacionales during botanic expeditions to Misiones in northeastern Argentina. Misiones is a strip of land between the Uruguay and Parana rivers, with lon- gitudinal mountains reaching approximately 800 m in elevation. This region is under exploration in order to study and document the floristic composition of the southernmost subtropical forests in South America. These records are elements of the Araucaria angustifolia forests and rocky “campos”, two pri- mary communities that should be maintained and preserved within national and state reserves. Megalastrum crenulans (Fée) A. R. Smith & R. C. Moran, Amer. Fern J. 77: 127. 1988. Aspidium crenulans Fée, Cr. vasc. Brés. 1: 139, t. 47, f. 1, 1869. Dryopteris crenulans (Fée) C. Chr. DESCRIPTIONS AND ICONOGRAPHIES.—Fée, op. cit.; Sehnem, Aspididceas, 186, f. 45, in R. Reitz (ed.) Flora Ilustrada Catarinense Parte I (ASPI) Itajaf, SC, Brasil, 1979. DISTRIBUTION AND EcoLoGy.—This fern grows in southern Brasil, Paraguay and has been cited also as occuring in the Venezuelan Guyana (Smith, Pterido- phytes pp. 1-334, in A. Steyermark, P. E. Berry & B. K. Holst (Gen. Eds.) Flora of the Venezuelan Guayana 2, Missouri Botanical Garden, 1995). Here it is recorded for the first time for Argentina, elevating the number of species of Megalastrum in this country to five (Ponce, Monogr. Syst. Bot. Missouri Bot. Gard. 60: 1-79. 1996). It inhabits subtropical Araucaria forests. This species is characterized by the occurrence of glandular trichomes mixed with simple trichomes, both of which are short and dense on the abaxial face, bulliform scales on the costule, and large, brown and glandular indusia. The latter character is not common in Megalastrum. The forms crenulans and glandulosa (Rosent.) C. Chr. were described by Christensen (Kongel. Danske Vidensk. Selsk. Skr. Naturvidensk. Afd. 8(6): 3-132. 1920); material from Mi- siones is referable to the latter. A revision of the genus is necessary to clarify species boundaries. SPECIMEN STUDIED.—ARGENTINA. Misiones, Dpto. San Pedro, Parque Provincial Cruce Caballero, 690 m, 26° 31’ S, 53° 59’ W, 15-IV-1996, F. O. Zuloaga, O. Morrone & M. Mulgura 5551 (SI). Elaphoglossum pachydermum (Fée) T. Moore, Index Fil. 12. 1857. Acrostichum pachydermum Fée, 2eme. Mém. Foug. 47. 1845. SHORTER NOTE 71 DESCRIPTIONS AND ICONOGRAPHY.—Sehnem, Aspididceas, 24, Fig. 5.2, in R. Reitz (ed.) Flora Ilustrada Catarinense Parte I (ASPI) Itajai, SC, Brasil, 1979. DISTRIBUTION AND ECOLOGY.—This species grows in Brasil, Paraguay and is here reported for Argentina. It is rupiculous, occurring on rocky promontories and steep slopes in montane areas. This species has erect to short creeping rhizomes and fimbriate scales from the base to the middle of the petiole. The sterile fronds are longer than the fertile fronds, the lamina is linear-oblong, attenuate-decurrently based, coria- ceous in texture, and with minute, reddish to brown, stellate trichomes on the abaxial surface. Elaphoglossum pachydermum is in the E. latifolium (Sw.) J. Sm. complex, and was thought to be confined to the Antilles (Mickel, pers. comm.). Previ- ously it was recorded for Argentina by Hassler (Trab. Inst. Bot. Farmacol. 45: 1-102. 1928) as E. latifolium, but the voucher specimem (Parodi 84) could not be located for verification. With this report, E. pachydermum is documented for the first time for Ar- gentina. The related species, E. macrophyllum (Mett. ex Kuhn) H. Christ and E. macahense (Fée) Rosenst., are known from Brasil. SPECIMEN STUDIED.—ARGENTINA. Misiones. Dpto. San Ignacio, Parque Provincial Teyti Cuaré Pefién del Teyti Cuaré, rupicola en zonas protegidas, 28/09/1998, Biganzoli, F. & D. Giraldo-Cafias 394 (SI). Cheiroglossa palmata (L.) C. Presl, Suppl. Tent. Pterid. 57. “1845” 1846. Ophioglossum palmatum L. Sp. Pl. 1063. 1753. DESCRIPTION AND ICONOGRAPHY.—Tryon & Stolze (Fieldiana, Bot. n.s. 20: 9, f. 2. 9) DISTRIBUTION AND EcoLocy.—This fern grows in southern Florida (USA), the Antilles, southern Mexico to Brazil, and northeastern Argentina. It is also dis- junct in Madagascar and southeastern Asia. It is dae cane cis suberect to pendant on trunks in subtropical primary and marginal fores Wagner’s treatment of the Ophioglossaceae See pp. 44-48, in G. Davise et al. (Eds.) Flora Mesoamericana. Vol. 1, Univ. Nac. Aut6noma de México, 1995) is accepted here. Wagner considers Cheiroglossa a genus justi- fiably separated from Ophioglossum. This species was cited by Molfino (Phy- sis, Buenos Aires, 8: 259-260. 1925) and Lichtenstein (Darwiniana 6: 380-441. 1944) as being in Misiones, but without documenting herbarium specimens. SPECIMEN STUDIED.—ARGENTINA. Misiones. Dpto. Gral. Manuel Belgrano: Deseado, Reserva Pri- vada Caé-pora, epifita sobre “Sotacaballo’’, invierno/1996, Chaves & Chebez s.n. (SI). Dicranopteris flexuosa (Schrad.) Underw., Bull. Torrey sige orig 34: 254. 1907. Mertensia flexuosa Schrad., Gott. gel. Anz. 1824: 863. DESCRIPTIONS AND ICONOGRAPHS.—Fée, Crypt. vasc. bras. 1: 199, t. 72, f. 1 (sub Mertensia scalpturata Fée); 200, t. 73, f. 2 (sub M. spissa Fée) 1869. Sehnem, 72 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 2 (2001) Gleiqueniaceas, in R. Reitz (ed.) Flora IJustrada Catarinense Parte I (GLEN): 25, f. 9. 1970 DISTRIBUTION AND ECoLoGy.—Southern Alabama (USA), the Antilles, southern Mexico, Central America, Colombia and Guyanas to Bolivia, southern Brazil, Paraguay, and Argentina. Terrestrial and rupiculous at shady and moist sites in rocky “campos” and marginal forests. Dicranopteris flexuosa is the correct name for the neotropical counterpart of the paleotropical D. linearis (Burm.) Underw. Roth (Faculty of Natural Scienc- es, University of Gotreborg, Sweeden, 1986), in her thesis on the neotropical Gleicheniaceae, suggested that D. flexuosa may represent a subspecies of D. linearis, but that combination has never been made. SPECIMEN STUDIED.—ARGENTINA. Misiones, Dpto. San Ignacio, Parque Provincial Teyui Cuaré. Pefién del Teyt Cuaré, rupicola, ocasional en zona protegida y htimeda, localmente abundante, 28/09/1998, Biganzoli, F. & Giraldo-Carias 397 (SI). Thanks to colleagues and institutions of the Inventario Florfstico de la Prov- incia de Misiones project (PMT-PICT 01511) directed by F. Zuloaga and sup- ported by the Consejo Nacional de Investigaciones Cientfficas y Técnicas (CONICET) Argentina.—MOonica PONCE, Instituto de Botanica Darwinion, La- bardén 200, B1642HYD San Isidro, Argentina. NT INFORMATION FOR AUTHORS: Authors are encouraged to submit manuscripts pertinent to pteridology for pub- lication in the American Fern Journal. Manuscripts should be sent to the Editor. Acceptance of papers for publication depends on merit as judged by two or more referees. Authors are encouraged to contribute toward publishing costs; however, the payment or non-payment of page charges will affect neither the acceptability of manuscripts nor the date of publication. Authors should adhere to the following guidelines; manuscripts not so prepared may be returned for revision prior to review. Submit manuscripts in triplicate (xerocopies acceptable), including review copies of illustrations and originals of illustrations. After review, submission of final versions of manuscripts on diskette (in PC- or Mac-compatible formats) is strongly encouraged. Use standard 8% by 11 inch paper of good quality, not “erasable” paper. Double space manuscripts throughout, including title, authors’ names and addresses, short, informative ab- stract, text (including heads and keys), literature cited, tables (separate from text), and figure captions (grouped as consecutive paragraphs separate from figures). Arrange parts of manuscript in order just given. Include author’s name and page number in upper right corner of every sheet. Provide margins of at least 25 mm all around on typed pages. Do not submit right-justified copy, avoid footnotes, and do not break words at ends of lines. Make table headings and figure captions self-explanatory. Use S.I. (metric) units for all measures (e.g., distance, elevation, weight) unless quoted or cited from another source (e.g., specimen citations). For nomenclatural matter (i.e., synonymy and typification), use one paragraph per basionym (see Regnum Veg. 58:39—40. 1968). Abbreviate titles of serial publi- cations according to Botanico-Periodicum-Huntianum (Lawrence et al., 1968, Hunt Botanical Library, Pittsburgh) and its supplement (1991). References cited only as part of nomenclatural matter are not included in literature cited. For shorter notes and reviews, omit the abstract and put all references parenthetically in text. Use Index Herbariorum (Regnum Veg. 120:1—693. 1990) for designations of her- Illustrations should be proportioned to fit page width with caption on the same page. Provide margins of at least 25 mm on all illustrations. For continuous-tone illustrations, design originals for reproduction without reduction or by uniform amount. In composite blocks, abut edges of adjacent photographs. Avoid com- bining continuous-tone and line-copy in single illustrations or blocks. Coordinate sequence and numbering of figures (and of tables) with order of citation in text. Explain scales and symbols in figures themselves, not in captions. Include a scale and reference to latitude and longitude in each ma Proofs and reprint order forms are sent to authors by the printer. Authors should send corrected proofs to the editor and reprint orders to the printer. Authors will be assessed charges for extensive alterations made after type has been set. For other matter of form or style, consult recent issues of American Fern Jour- nal and The Chicago Manual of Style, 14th ed. (1993, Univ. Chicago Press, Chicago). Occasionally, departure from these guidelines may be justified. Authors are encouraged to consult the editor for assistance with any aspect of manuscript preparation. Papers longer than 32 printed pages may be sent to the Editor of Pteridologia (Memoir Editor, see cover 2). PTERIDOLOGIA ISSUES IN PRINT The following issues of Pteridologia, the memoir series of the American Fern Society, are available for purchase: 1. Wagner, David H. 1979. Systematics of Polystichum in Western North America North of Mexico. 64 pp. $10.00 postpaid. 2A. Lellinger, David B. 1989. The Ferns and Fern-allies of Costa Rica, Panama, and the Choco (Part 1: Psilotaceae through Dicksoniaceae). 364 pp. $32.00 postpaid. Send your order with a check or money order to: American Fern Society, Inc., c/o U.S. National Herbarium MRC-166, Smithsonian Institution, Washing- ton, DC 20560. AMERICAN FERN JOURNAL ON MICROFICHE Volumes 1—61 of the American Fern Journal are available as archival quality, silver positive microfiches. Single volumes or the entire run may be purchased. The fiches are easily read with 10 or greater magnification (using a dissecting microscope and transmitted illumination or a fiche reader). Silver negative micro- fiches of vols. 1-50 are also available. The price is $4.00 per volume or $244.00 per set of 61 volumes, postpaid. Send your inquiry or order with a check or money order to: American Fern Society, Inc., c/o Dr. Jamés D. Montgomery, Ecology III, Inc., R.D. 1, Box 1795, Berwick, PA 18603. VISIT THE AMERICAN FERN SOCIETY’S WORLD WIDE WEB HOMEPAGE: http://www.amerfernsoc.org/ QR) A399 AMERICAN FERN nant J O U R N A ' July-September 2001 QUARTERLY JOURNAL OF THE AMERICAN FERN SOCIETY SPECIAL ISSUE The Evolution and Diversification of the Lycopods Proceedings of a Symposium Held August 1999 at the XVI International Botanical Congress Missouri Botanical Garden St. Louis, Missouri Jointly Sponsored by: VI International Botanical Congress Green Plant fgpicuate Research eRe resorts Group U.S. Department of Agricultur National Science Romndatinn. Department of Energy American Fern Society Symposium Organizers: Niklas Wikstr6m The Evolution and Diversification of the Lycopods W. Carl Taylor 73 Early Lycophyte Evolution Patricia G. Gensel and Christopher M. Berry 74 Isoetalean Lycopsid Evolution: from the Devonian to the Present Kathleen B. Pigg 99 The Triassic Lycopsids Pleuromeia and Annalepis: Relationships, Evolution, an Léa Grauvogel-Stamm and Bernard Lugardon 115 Diversification and Relationships of Extant Homosporous Lycopeds _—Niklas Wikstrém 150 The Utility of Nuclear ITS, a LEAFY H Intr d Chi last atpB-rbcL Spacer Region Data in Phylogenetic fe and Species Delimitation i in Isoétes Sara B. Hoot and W. Car! Taylor 166 The American Fern Society Council for 2001 BARBARA JOE HOSHIZAKI, 557 N. Westmoreland Ave., Los Angeles, CA 90004-2210. President CHRISTOPHER H. HAUFLER, Dept. of Botany, University of Kansas, Lawrence, KS 66045-2016. oo W. CARL TAYLOR, 800 W. Wells St., Milwaukee Public Museum, Milwaukee, WI 53233-147 < etary JAMES D. CAPONETTI, Dept. of Botany, University of Tennessee, Knoxville, TN 37996-1110. Treasurer VID B. LELLINGER, 326 West St. NW., Vienna, VA 22180-415 Membership Secretary a D. MONTGOMERY, Ecology III, R.D. 1, Box 1795, oe PA 18603-' ase Back Issues Curator R. JAMES HICKEY, Botany Dept., Miami University, Oxford, OH, 45056 Journal Editor DAVID B. LELLINGER, U.S. National Herbarium MRC-166, Smithsonian Institution, Basen er DC 20560-0166. Memoir Editor CINDY JOHNSON-GROH, Dept. of psi Gustavus Adolphus College, 800 W. College Ave., St. Peter, MN 56082-1498. Bulletin Editor American Fern Journal R. JAMES HICKEY y Departm ami University, Tae OH 45056 ph. (513) cae: paehera hickeyrj @muohio.edu ASSOCIATE EDITORS GERALD J. GASTONY 2.0.2. ares of ao Indiana University, Bl a IN cohesion CHRISTOPHER H. HAUFLER Dept. of — University of Kansas, Lawrence, KS 6604 ROBBIN C. MORAN New York Botanical cnc eee NY — se JAMES H. PECK Dept. of Biology, University of Arkansas—Little Rock, 2801 S. University ie. Little Rock, AR 72204 The “American Fern Journal” (ISSN ies sib is an illustrated quarterly devoted to the ee study of ferns. It is owned by the American Fern Society, and sas at ae West St. NW., VA 22180-4151. Periodicals postage paid at ee VA, and Claims for missing issues, made 6 months (domestic) to 12 sti (forei en) after the date = issue, and orders for back issues should be addressed to Dr. James D. Montgomery, Ecology III, Berwick, PA 18603-9801. Paes of address, dues, and applications for membership should be sent to the Membership eR Rae RAAT sh i h o the Secretary. Subscriptions $20.00 gross, “$19. 50 net if paid sane an 7 ag pega fee $0.50); sent free to of the American Fern Society (annual dues, $15.00 + $5.00 mailing surcharge beyond “life membership, $300.00 + $140.00 mailing surcharge a U.S.A Back volumes are available for most years as printed issues or on microfiche. Please contact the Back Issues Curator for prices and availability pac eee Send address changes to AMERICAN FERN JOURNAL, 326 West St. NW., Vienna, VA 22180-4151. cS in a: FIDDLEHEAD FORUM The editor of the Ries of the American Fern Society welcomes contributions from members and non-members, includin scellaneous notes, offers to —— or purchase materials, personalia, horticultural notes, and reviews of non-technical books on fi SPORE EXCHANGE Stephen McDaniel, 1716 Piermont Dr., Hacienda Hts., CA sa De is Fad i Pit and lists of available spores sent on request. _http://www.amerfern: soc.org/sporex GIFTS AND BEQUESTS ifts and bequests to the Society enable it to expand its services to members and to others interested in ferns. Back issues of the Joumal and h ther gift 1 y 1 4 a tax -deductible. Inquiries should be addressed to the Secretary. American Fern Journal 91(3):73 (2001) The Evolution and Diversification of the Lypopods The lycopods, a diverse and ancient group of pteridophytes, form a major lineage of vascular plants. This lineage includes three orders, Lycopodiales, Selaginellales, and Isoetales. Each order contains a single extant family, but the group has a long and diverse fossil record beginning in the Devonian or possibly late Silurian. This fossil record is replete with many curious forms that have long captured the interest and imagination of those interested in plant evolution. Although lycopods have attracted the attention of paleobota- nists and morphologists for many years, it is only within the last few years that systematists have initiated studies to critically evaluate taxa and their modes of speciation. While there is agreement that the lycopods form a clade distinct from all other vascular plants, there is some argument in the interpre- tation of the diversity and evolution within the lycopod lineage. This is an exciting time for systematic biologists. We are beginning to uncover, discover, and explore the modes and mechanisms of divergent and reticulate evolution in the lycopods, revealing their incredibly long history and amazing diversity. The following five papers are facsimiles of oral presentations given on 4 August 1999, in a symposium entitled “The Evolution and Diversification of the Lycopods.” This symposium, a part of the XVI International Botanical Con- gress held in St. Louis, Missouri, brought leading researchers together to pre- sent their current findings relevant to the phylogeny of the lycopods. Niklas Wikstrém, Léa Grauvogel-Stamm, and Carl Taylor organized the symposium. The authors express their gratitude to Brent D. Mishler, the Green Plant Phylogeny Research Coordination Group, the U.S. Department of Agriculture, the National Science Foundation, the Department of Energy, and the organizers of the XVI International Congress for their assistance and support. The authors are also grateful to the American Fern Society, Inc. for sponsoring the publi- cation of their papers.—W. Carl Taylor, Department of Botany, Milwaukee Pub- lic Museum, Milwaukee, WI 53233 MISSOURI BOTANICAL nFC 3 1 2001 GARDEN LIBRARY American Fern Journal 91(3):74—98 (2001) Early Lycophyte Evolution PATRICIA G. GENSEL Department of Biology, University of North Carolina, Chapel Hill, NC 27599 CHRISTOPHER M. BERRY Department of Earth Sciences, Cardiff University, Cardiff, CF10 3YE, Wales, United Kingdom ABSTRACT.—Lycophytes, comprising the groups historically known as the lycopsids and zostero- phylls, have the longest history of any group of vascular land plants. The early evolution of the WHAT Is A LYCOPHYTE AND WHEN DID LYCOPHYTES APPEAR? The lycophytes ( = club mosses and related plants) are regarded as a distinct lineage of vascular plants with a lon evolutionary history. Considerable di- versity is evident for much of their existence, although only four (or 7+ if one accepts either Ollgard’s (1987) or Wagner and Beitel’s (1992) splitting of Ly- copodium into several genera) herbaceous genera are recognized today (Lyco- podium (and segregates), Phylloglossum, Selaginella, and Isoetes). Club mosses have long been defined as plants with microphyllous leaves, axillary or adax- ially borne reniform sporangia with dehiscence along their distal margin, and mostly exarch xylem maturation; such plants are now included in the class Lycopsida (“lycopsids”) by Kenrick and Crane (1997). Extinct taxa attributed to Zosterophyllopsida (and possibly including barinophytes) are considered closely related to lycopsids, either as ancestral or sister taxa (Gensel, 1992; Hueber, 1992; Bateman, 1992; Kenrick and Crane, 1997), being similar in spo- rangium shape and xylem attributes but differing from lycopsids in lacking leaves and leaf-borne sporangia. Additionally they have vascularized sporan- gial stalks (Fig. 1). REDEFINITION OF “LYCOPHYTES” AND “LYCOPSIDS”.—Several workers emphasize the monophyly of vascular plants (Mishler et al., 1994; Stewart and Rothwell, 1993; Kenrick and Crane, 1997). Many types of evidence, molecular and mor- phological, demonstrate that lycophytes sensu lato form a distinct lineage of vascular plants from early in their evolutionary history (Raubeson and Jansen, 1992; Kranz and Huss, 1996; Kenrick and Crane, 1997). In Kenrick and Crane’s recent, extensive morphological phylogenetic analysis of living and extinct plants, monophyly is supported, and zosterophylls and lycopsids are inter- preted as two classes in a larger group, subdivision Lycophytina (Fig. 1). This GENSEL & BERRY: EARLY LYCOPHYTE EVOLUTION 75 lon t t Crenaticaulis t . ascidian Sawdoniaceae Deheubarthia t Sawdonia t +——®- Thrinkophyton t pennies Barinophyton citrulliforme + Barinophyton obscurum ¢ | Barinophytaceae ’ Protobarinophyton t Lycophytina [7 Euphyllophytina rere ia et he pe . *- -——@ Gosslingia t meds a te -[ Lycopsida ; Sere pee dl Zosterophyllopsida L___ Tarelia t |——3 Cooksonia cambrensis t —_—_——@- Zosterophylium divaricatum t¢ |_—a- Renalia t -—————- Zosterophyilum fertile t Math ae }———# Uskiella t }+————-#- Zosterophyltum lanoveranum t i -———# Yunia t +} Rebuchia t + Sartimana aaeereeeeey eee eapobaadunee! Zosterophyllum deciduum t 3 Cooksonia caledonica ¢ 4 Huperzia Rhyniopsida Asteroxyion t cineianban’ Lycopsida Lycophytina > suaedimasneatile ae Noma t a it Banks Acmeatogay t -——®- Zosterophylium myretonianum t +} Gumuia t \_____—» Huiat Hicklingia t Rhynia t Psilophyton ¢ A B @ Zosterophyllophytina sensu Banks Fic. 1. Strict consensus trees showing relationships among several polysporangiophyte taxa and among lycophytes according to analyses conducted by Kenrick and Crane (1997). A) Adapted from Kenrick and Crane, fig. 4.31. B) From Kenrick and Crane, fig. 5.25, with permission. Zosterophyl- lopsida and Lycopsida are highlighted. See text for discussion. Not all taxa discussed in the text are included, partly because of differences in preservation and therefore absence of characters for some taxa larger group is defined (p.237—238) by the following synapomorphies: “more or less reniform sporangia, marked sporangium dorsiventrality, isovalvate de- hiscence along distal sporangium margin, conspicuous cellular thickening along sporangial dehiscence line, sporangia on short, laterally inserted stalks, exarch xylem differentiation.” Within the Lycophytina, the strict consensus tree of their analysis (p. 172) shows a polytomy (Fig. 1B) from which emerge the Zosterophyllopsida, a clade largely consisting of lycopsids, and a large number of individual taxa. Zosterophyllopsida sensu Kenrick and Crane (Fig. 1B) are defined by only two synapomorphies, namely circinate growth and a unique, two-rowed arrangement of sporangia. Among the single unresolved taxa excluded from Zosterophyllopsida sensu stricto are several plant genera (Rebuchia, Discalis) originally placed among the zosterophylls, several species 76 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) of Zosterophyllum including the type species (Z. myretonianum Penhallow), and some recently discovered Chinese forms. Thus Kenrick and Crane regard plants traditionally considered zosterophylls as paraphyletic. The incomplete data among the taxa analyzed for all characters employed, and lack of reso- lution at lower levels in the cladogram, indicates further investigation of char- acters and character homologies is needed. Thus these conclusions should be regarded with caution until additional data are available. Kenrick and Crane (1997) define the second clade, the Lycopsida (Fig. 1B), by the synapomorphies of microphylls, stellate xylem strand, vascularized leaves, sporophylls (close association of sporangium and leaf), and loss of spo- rangial vasculature. Lycopsida includes a clade composed of three possibly transitional genera, Asteroxylon, Drepanophycus, and Baragwanathia, and a second larger clade of plants traditionally considered as lycopsids. The first clade is composed of taxa that do not possess all of the characters definitive of lycopsids—Asteroxylon and Drepanophycus have cauline sporangia with a vascularized stalk, and Asteroxylon lacks full vascularization of lateral ap- pendages. Neither has a sporangium associated with a leaf, as occurs in mem- bers of the lycopsid clade and most probably in Baragwanathia (Hueber, 1992). TIME OF APPEARANCE.—The occurrence of the genus Baragwanathia in the pre- sumed late Silurian (Ludlow; Garratt et al., 1984; Rickards, 2000) of Victoria, Australia suggests a possible pre-Devonian diversification of lycophytes. Zos- terophylls also are recorded from these strata, but detailed descriptions are lacking (Tims and Chambers, 1974). Most other occurrences of late Silurian plant remains are much smaller in size and simpler in organization than either Baragwanathia or Zosterophyllum (Fig. 2). An interesting exception is the as- semblage of zosterophylls and related plants from well-dated late Ludlow stra- ta of Bathurst Island (Kotyk et al., in prep.). These new forms support an early diversification of the lycophyte lineage. POSSIBLE PRECURSORS.—Most polysporangiate plants (i.e. plants with branched sporophytes) with round to reniform sporangia referred to either as rhynioph- yte?, cooksonioid, or rhyniophytoid (Taylor, 1988; Hueber, 1992) represent possible close relatives or precursors to the lycophytes but currently there are no obvious candidates. Outline drawings (Fig. 2) of several Late Silurian and Early Devonian rhyniophytoids show that these plants mostly are very small and exhibit variable sporangial shapes and modes of dehiscence. Their small size, simple construction, lack of anatomical preservation, and fragmentary preservation make more precise inference impossible, but those with globose- reniform sporangial construction may represent likely candidates for lycophy- te precursors. A few larger forms with similar sporangial morphology have been recorded in pre-Devonian or earliest Devonian deposits, such as the spec- imen from the Ludlow of Bathurst Island (Basinger et al., 1996; Kotyk, 1998) which resembles a pseudomonopodial Cooksonia or very simple Zosterophyl- lum (fig. 2 I). Renalia, and some Cooksonia species, inferred to have arisen from within the plexus of latest Silurian-earliest Devonian rhyniophytoids (but their rela- GENSEL & BERRY: EARLY LYCOPHYTE EVOLUTION 77 Silurian plants (A - 1) B Cc @ = nt = A x3 D 2, E G | x250 F @ x32 GOs | \\ x3 \O x24 q Lochkovian plants (J - O) oo || *4 { . x64 M } * contains cryptospores ] a co dyads & obligate tetrad © ® | © trilete miospores N x23 (@) x25 Fic. 2. Selected late Silurian and earliest Devonian meso- and megafossils. A—I) Pridoli plants. A) from Cai et al. (1993); B—D) from Fanning et al. (1991); E—G) from Edwards (1996). H) from Kotyk (1998). I) from Basinger et al. (1996). J—O). Lowermost Devonian plants from Edwards (1996). tionship to these simple plants thus far is not extensively tested by cladistic analyses), also represent likely precursors. When Renalia and Cooksonia were included in the phylogenetic analyses of Kenrick and Crane, they formed part of the large basal polytomy of the polysporangiophytes and were interpreted as possible stem-lineage zosterophylls. Further investigation of these early plants is needed. 78 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) The present contribution notes major advances and addresses several areas of concern in regard to early lycophyte evolution and diversification, including current views of zosterophylls, early lycopsids, and evolution of major lyco- phyte structures. CURRENT DATA ON ZOSTEROPHYLLS As a probable sister group(s) or stem group(s) to lycopsids, zosterophylls sensu lato (Fig. 3A—3C) may provide evidence of early evolutionary patterns within these lineages. However, presently recognized taxa, while demonstrat- ing greater diversity of form and some new combinations of characters, do not alter conceptions of relationships between zosterophylls and lycopsids nor aid in clarifying ordinal or familial delimitation within zosterophylls. The two main zosterophyllopsid clades sensu Kenrick and Crane (1997) can be typified by two genera, Sawdonia ornata (Fig. 3A) and Oricilla bilinearis (Fig. 3C). Differences include sporangium orientation and presence/absence of emer- gences (multicellular appendages). The Sawdonia clade (Kenrick and Crane’s Sawdoniaceae) exhibits more upright orientation of sporangia and emergences of varying morphology and arrangement, while the Oricilla clade (Kenrick and Crane’s Gosslingiaceae) exhibits naked axes and sporangia stalks attached per- pendicular to the axis. The Zosterophyllum depicted in Fig. 3B exhibits heli- cally arranged sporangia and is most similar to those species of Zosterophyl- Jum that form part of the basal polytomy in Kenrick and Crane’s analyses. Beyond these two well- supported clades, we now have considerable new in- formation about some of the more basal lineages, discussed next. Plants which appear very similar to Zosterophyllum (Fig. 3B), except that sporangia are more scattered and borne on longer stalks than in many previ- ously described Zosterophyllum species, are found in Pragian deposits of Chi- na and Bathurst Island, Canadian Arctic Archipelago (Kotyk, 1998). These do not quite fit existing generic or species delimitations, but as they become more completely known, they might aid in a better understanding of the genus Zos- terophyllum or closely allied taxa. The separation of several Zosterophyllum species into separate clades in Kenrick and Crane’s analysis is based mostly on symmetry of sporangial arrangement and presence/absence of circinate ver- nation. However, these characters are not easy to evaluate and their coding may not reflect that data about them are incomplete; most particularly, several taxa lacked anatomy. It is difficult to document the presence or (known) ab- sence of circinate vernation (and true terminations). The character “symmetry of sporangial arrangement” may be difficult to ascertain because determining if sporangia are helically arranged or are subopposite in two rows is difficult in compression remains. A two-rowed pattern could be produced by a widely spaced, very simple helix coupled with twisting of either sporangial stalks or the axis (see Gerrienne, 1996b for an analysis of this feature), Current analyses indicate these new Zosterophyllum species and other Zosterophyllum-like plants may represent stem zosterophylls for which only a few distinguishing characters are present. GENSEL & BERRY: EARLY LYCOPHYTE EVOLUTION 79 AN 1mm < = 2mm > | er Ii, Fic. 3. Selected putative basal lycophytes or zosterophyllopsids. A, C) Representatives of Zos- terophyllopsida sensu Kenrick and Crane. A) Sawdonia ornata (Dawson) Hueber, based on spec- Zosterophyllum with helically arranged sporangia, a stem group taxon according to Kenrick and Crane (1997). D) Gosferia curvata Gerrienne from the Early Devonian of Belgium. Note thickening near base of sporangium. Redrawn from Gerrienne (1991). E) Ensivalia deblondii Gerrienne, a probable zosterophyllopsid from the Early Devonian of Belgium. F) Faironella valentula Gerrienne, affinity uncertain, from the Early Devonian of Belgium. Sporangium is borne inside a recurved, thickened spiny structure. E, F) redrawn from Gerrienne (1996). 80 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) Another possible basal taxon, Gosferia curvata Gerrienne (1991, 1999), from the Early Devonian of Belgium, is unique in having long, basipetally recurving stalks, each terminated by a reniform sporangium (Fig. 3D). The closest simi- larity of this plant is to some unidentified zosterophyll-like plants in China and Bathurst Island and to some Silurian-Pragian rhyniophyte-like plants. The occurrence of a swelling near the sporangium base where it attaches to the stalk also is of interest; similar swellings are known in some Drepanophycus species (Li and Edwards, 1995) and in Kaulangiophyton (Gensel, pers. obs.). What it represents in any of these plants is currently unknown (but see below). Ensivalia deblondii Gerrienne (1996a) from the Early Devonian of Belgium has spiny, anisovalvate sporangia terminating stout stalks. The abaxial sporan- gial valve is enlarged and spiny, inside of which the more delicate, smooth adaxial valve rests (Fig. 3E). Odonax borealis Gerrienne (1996b), also from Belgium, and some unnamed zosterophylls from the Early Devonian of New Brunswick also exhibit modification of the sporangium/stalk in the form of enlargement of either stalk or distal sporangium valve and presence of emer- gences, features which perhaps help protect the sporangium. Further evaluation of the variability of emergence types (see Gensel, 1992 and Kenrick and Crane, 1997 for examples) among zosterophylls is needed as well as, perhaps, clarification of the homologies of different types of emer- gences. For example, in Nothia aphylla (Lyon, 1964) El Saadawy and Lacey (1979), the so-called emergences are produced by raised areas of axial tissue (cortex and epidermis), resulting from expansion of existing cells, with api- cally located stomata (Kerp et al., 2001). Are these structures truly emergences or do they represent short, discontinuous ridges of the stem? The unusual prismatic emergences of Serrulacaulis (Berry and Edwards, 1994) and those now known to occur in Bathurstia (Kotyk and Basinger, 2000) also add to diversity of those structures but do not provide any further tests of homology between emergences and microphylls. Some Early Devonian plants in which sporangia are arranged in an obvious strobilus may relate to the lycophyte clade, either among zosterophylls or ly- copsids. Demersatheca, based on only a few leafless fertile specimens from the Pragian of China, exhibits reniform sporangia deeply sunken into the axis (Fig. 4D) and forming tight strobili (Li and Edwards, 1996). Other plants with stro- bili include Distichophytum and Bathurstia, both attributed to the zostero- phylls, and some un-named plants of uncertain affinity from Bathurst Island. This morphology blurs the distinction between zosterophyll and barinophyte to some extent. Barinophytes are defined, according to Kenrick and Crane, by a very com- pact, unbranched strobilus, indehiscent sporangia, intrasporangial heterospo- ry, and sporangia borne on a horizontally extended axis which wraps around the stem (their “clasping” sporangium orientation). However, their relation- ships are unclear as the Barinophytaceae form part of a polytomy that includes some major groups of zosterophylls in Kenrick and Crane’s analysis. More data are needed about sporangium shape, attachment, and dehiscence in both the early strobilate plants and some plants usually attributed to barinophytes, such GENSEL & BERRY: EARLY LYCOPHYTE EVOLUTION 81 Fic. 4. Plants from the Pragian of China that may be related to lycophytes. A) Adoketophyton subverticillatum Li and Edwards with a leaf-like appendage subtending each sporangium. Re- drawn from Li and Edwards (1992). B, C) Eophyllophyton bellum Hao, in which lobed leaf-like structures terminate lateral axes. Globose to reniform sporangia attach on the upper surface of the lobes. Redrawn from Hao and Beck (1993). D. Demersatheca contigua Li and Edwards, with spo- rangia sunken in the axes. Redrawn from Li and Edwards (1996). as Protobarinophyton. Faironella valentula Gerrienne (1996a) is also of interest in this discussion as a possible link between zosterophylls and barinophytes. Its sporangia are attached to the inside of a curved appendage (Fig. 3F), and while affinities presently are uncertain, it thus bears some resemblance to Bar- inophytaceae. Adoketophyton, a genus described from the Pragian of China by Li and Ed- wards (1992), demonstrates fertile regions with leaf-like structures that bear sporangia in their axils (Fig. 4A). Crane and Kenrick (1997) suggest the leaf- like structures might represent sterilized sporangia that became modified as a “leaf” as a result of changes in developmental processes such as multiplication and co-option. Another plant from the Pragian of China, Eophyllophyton bel- lum (Hao) Hao and Beck (1991) has the habit of a typical Early Devonian plant, but bears lobed, laminate leaf-like structures (Fig. 4B, 4C). Whereas this plant is at the base of the euphyllophyte clade of Kenrick and Crane, largely because 82 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) Fic. 5. Line drawing of Drepanophycus qujingensis Li and Edwards. Note cauline sporangia not associated with leaves. Redrawn from Li and Edwards (1995). of its centrarch protostele, its globose-reniform sporangia are located adaxially on leaves and its tracheids possess G-type wall thickenings like those present in many zosterophylls and stem-lycophytes. In our opinion, this plant pos- sesses an odd suite of characters for basal euphyllophytes and it is perhaps more appropriately considered in the context of zosterophylls or lycopsids. Both Adoketophyton and Eophyllophyton thus provide idiosyncratic evidence of the development of laminate appendages, but little information on their homologies. PRE-LYCOPSIDS This term has been applied to plants with many lycopsid characters, but lacking the consistent association of sporangium with a leaf or lacking fully vascularized leaves. Genera included are Drepanophycus, Kaulangiophyton, Asteroxylon, and in some instances, Baragwanathia (for another interpretation, see Hueber, 1992). They form a sister clade to other lycopsids in Kenrick and Crane’s analysis and are treated there as a stem-based group of Lycopsida, the Drepanophycales. However, some aspects of their structure remain controver- sial, and re-evaluation of these features may influence their position relative to lycopsids or the interpretation of their evolutionary significance. Based on current knowledge of Drepanophycus, the following appears cer- tain: two species of Drepanophycus, D. qujingensis Li and Edwards and the type species D. spinaeformis (Goeppert) Li, Hotton and Hueber, exhibit both rhizomatous and aerial axes covered with vascularized microphyllous leaves and cauline sporangia (Fig. 5; Li and Edwards, 1995; Li et al., 2000). Among GENSEL & BERRY: EARLY LYCOPHYTE EVOLUTION 83 vegetative remains assigned to Drepanophycus, leaf morphology varies con- siderably, from long and falcate to short and thorn-like. It has been suggested that the axes with the former represent aerial regions, those with the latter rhizomatous axes. Sporangia borne on comparatively short stalks (usually shorter than leaves) are interspersed among the microphyllous leaves, with no apparent association with them. Kaulangiophyon akantha exhibits a similar growth habit to that of Drepan- ophycus, but with all axes covered in stout, thorn-like emergences (Gensel et al., 1969). Studies are underway to test assertions that Kaulangiophyton is very similar to or congeneric with Drepanophycus (Kasper, 1972; Schweitzer and Giesen, 1980), or that Kaulangiophyton has sporangia associated with leaves (Hueber, 1992). Careful examination of original Maine collections and new material ha New Brunswick shows no indication of vascularization in the emergences. Sporangia were originally described as cauline, terminating a elas long nak and extending beyond emergences (Gensel et al., 1969). Hueber (1992) noted features of the sporangial stalk that led to his hypothesis that a leaf extends beneath the sporangium but is obscured by the sporangium and thus is not visible on impression remains. We note that the stalks of some sporangia attach to a thickened pad or ring of coalified material that itself is borne on an axial protrusion. It resembles the pad at the base of sporangia in Stockmansella and Huvenia. A similar thickening may occur in Drepanophy- cus qujingensis. However, in contrast to Hueber’s ideas, no extra appendage, emergence, or leaf lamina has been demonstrated in attachment to this thick- ening in Kaulangiophyton. Additional information on these and other char- acters of Kaulangiophyton and assessment of relationships will be presented in a later paper. ROOTING STRUCTURES IN EARLY LYCOPHYTES Information about rooting structures, other than rhizoids, in basal lycophy- tes is increasing, being recorded for the zosterophylls Zosterophyllum, Ba- thurstia, Crenaticaulis (Gensel et al., 2001), and Hsua (Li,1992), for Drepano- phycus (Rayner, 1984; Li and Edwards, 1995; Gensel et al., 2001) and for the rhyniacean Stockmansella (Taeniocrada) langii Fairon-Demaret (1985) by Schweitzer (1980). Two patterns are recognized by Gensel et al. (2001) and may correlate with habit. The first pattern occurs in tufted forms, such as some Zosterophyllum species, which produce downward trending axes from a cen- tral region that appear root-like. Upward trending axes departing from the same point are interpreted as shoots. In the other pattern, slender smooth axes depart from the rhizomatous regions, and in some cases divide to varying de- grees. These frequently lie across, not on, the bedding plane. Also, they are sometimes produced during the formation of K-branches as in Bathurstia and Drepanophycus (Gensel et al., 2001). In these plants, an axis departs, then orks. One of the resultant axes is oriented towards the rhizome apex and appears identical to shoots and the other, lacking either emergences or leaves and directed away from the apex, appears root-like (Figs. 6A, 6B). Based on 84 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) 1cm Fic. 6. A) Line drawing of the zosterophyll Bathurstia denticulata (Hueber) Kotyk and Basinger, showing numerous K-branches in which one portion is a rooting structure and the other a shoot- like structure, frequently in crozier form. B) Line drawing of Drepanophycus from the Pragian of Bathurst Island, showing several K-branches in which one part is a shoot and the other a smooth, root-like organ. r = rooting organs. Redrawn from Kotyk (1998), with permission and Gensel et al. (2001). current evidence, these rooting structures appear to be similar and perhaps homologous to shoots. Gensel et al. (2001) postulate that among lycopsids, and perhaps most non-seed plants, roots arose via a dichotomy of the shoot system, with one apex transformed into a root apex. This condition occurred over time progressively earlier in ontogeny, until the root arose in the embryo. The po- sition of the root primordium lateral to a shoot apex in many pteridophyte embryos may be significant in this regard. Furthermore, roots and shoots thus are homologous structures. None of the rooting structures among zosterophylls or lycopods have any kind of emergences or leaves on them except perhaps some fine hairs on those in Crenaticaulis. These differ from any ornament on aerial axes. Thus, part of the transformation to a root apex may have been to lose any potential to develop leaves or emergences. GENSEL & BERRY: EARLY LYCOPHYTE EVOLUTION 85 PROTOLEPIDODENDRALES (PROTOLEPIDODENDRACEAE, HASKINSIACEAE, AND ARCHAEOSIGILLARIACEAE): EARLY LYCOPSID DIVERSIFICATION Following the Late Silurian-Early Devonian establishment of zosterophy]- lopsids and lycopsids, diversification of lycopsids is evidenced by abundant and widespread remains of lycopsids with distinctive anatomical and mor- phological features. Many of these form a group currently recognized as the Protolepidodendrales. These plants were prominent in Middle to early Late Devonian times and show more advanced and variable leaf morphology and vascular anatomy than pre-lycopsids. Sporangia where known are epiphyllous. Two families and a satellite taxon (Thomas and Brack-Hanes, 1984) are rec- ognized within protolepidodendralean lycopsids. The Protolepidodendraceae are well known from several parts of the world in Middle Devonian times. They consisted of lycopsids with helically ar- ranged, forked leaves, and globose to elongate sporangia located on the adaxial surface of the leaf. Their characteristic leaves, which are divided into three (Colpodexylon, Minarodendron) or five or more (Leclercqia) segments, are the most striking features of the plants. In many of these taxa, sporophylls resem- ble vegetative leaves. Sporangia in Leclercqia attach by a single discrete layer of cells to the leaf just proximal to its division, and the free part of the spo- rangium is directed towards the axis. A single line of dehiscence extends the length of the sporangium. Mode of sporangial attachment in other protolepi- dodendrids is not as clear, but they apparently lack stalks. Permineralized remains of some of these plants show a solid xylem column with several nar- row ridges of protoxylem around the periphery (Leclercqia, Minarodendron), resulting in a shallowly ribbed protostele, or a more deeply lobed xylem strand (Colpodexylon). Leclercqia is notable for being a ligulate, homosporous taxon (Grierson and Bonamo, 1979). Berry (1996) summarized the considerable new data and major features of these plants and other representatives of the Pro- tolepidodendrales (Table 1). Pre-Middle Devonian Protolepidodendraceae have been found. Kasper and Forbes (1979) recorded Leclercqia sp. from the late Early to Early Middle De- vonian of Maine and Kasper (1977) noted the presence of a new species of Leclercqia in the Emsian of New Brunswick. In both of these occurrences, axes bear five-segmented leaves, but all segments are oriented in the same direction (Fig. 7B) rather than being three-dimensionally arrayed as in L. complexa Banks, Bonamo and Grierson. At the time of recording them, the age deter- minations of sediments were uncertain. Since then, dispersed spore data in- dicate the New Brunswick sediments are undoubtedly Emsian. Recent collec- tions from these Emsian New Brunswick outcrops by Gensel document a sec- ond occurrence of the genus, these plants being very similar in leaf morphol- ogy and sporangial attachment to L. complexa (Fig. 7A), differing mainly in their smaller size. Remains of a plant with leaves possessing a total of nine segments also have been found in the Emsian of New Brunswick. In these forms, four segments are oriented in an upwardly directed, flattened lamina to each side of a long, AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) “paqo] a[91s010.1g ‘UMOUYUN SdINjION.NS 9.104 “sSuruoyory) euapidaqns Aq papunodns saseq Jeo] [RUOSeXOY ,,poyeYUl,, ‘,YI9e1, poyuatO AqyeIstp ‘yesoyey, ‘uouTWOI’d Jo Jed eB YIM ‘“TeUTWR] SOAR] L661 SpareMpy 19 ALlog puaua (/€6] ploury uosydvogy 1061 volsply (l4addoy) nian sisoapys4sy (UOXB} 991998S) TVAOVRLV TIDISOAVHOUV ‘ayv[NoNuap [asoj01g ‘umOUyUN saiods !ploaogo 0} asogoys eIsuRsods ‘QyBISeY JO aVNISES ‘ayejooouR] opeyq ‘ayejoned saava7y IL61 YIAoyuag uopydynyrsjay 9661 SpremMpy Jo ALieg pusuls ¢gG] SyUeY Ja UOSIOUID ,vIsuLysDET AVAOVISNIASVH ‘gyefnonuap 10 (uojAxapodjoy) paqoy 1ayita 9f91S0101g ‘UMOUY doyM SNoJodsowoY Sploao aesuoja vIsuRIOdS ‘[BUOISUDLIp-daiy) JO ayeuR]d ‘syusWISs MOLIEU 6 O} dn OVUT PapIAIp sdArey ZEH| PurlAaM Ja jasne.ry Uvospuapopidajojosg prol Syurg uazAxapodjoz 0661 V1 “ospuaposvUpy ZL6| UOSIALID 19 OWKUOg ‘syueg ,.o1ba4aj9a'7 JANRIOSOA 0} [RONUAPI ssa] JO a10Ur s{|Aydosods ‘sjoymopnesd Jo sadtjay SuIjquiasas suJayed Ul payasul sacra] ‘spleyoen panid pasopsog YIM WafAx ‘yorrxa ‘g1aisojoid suiays :(qUaisIsied soARe]) ssvds Jeo] OU ‘sanssH Asepuodeas ou “IIqey SuIdaes9 ‘snoyUOzIYys Ajquqoid ‘saxe poyouesg {A[SNOWOJOYSIP YIM sIUe[d @ (MO]aq YSLAISe Se UMOYS-URTUOAAG A[Iey o1e[) URIS UT SddUaLINIZO BOS Ing URTUOASG ae] AjIee puL o[PpIy UL JUepUNge A[[eIDedsy @ SHTVAGNAGOCIda TOLOUd ‘spisdoo4] uvajespuapopidajojoid jo stsdoukg ‘| aIaV GENSEL & BERRY: EARLY LYCOPHYTE EVOLUTION 87 Fic. 7. Illustrations of canes from the Emsian of New Brunswick. A) Leclercqia collected by Gensel. This is most similar to L. complexa, 4X. B) Leclercqia sp. described in abstract by Kasper (1974). Oval bodies represent sporangia (arrows). 3.6X. C) New leclercqioid plant wi 8- segmented leaf. Note long distal segment, upward trending lateral ones in side view on upper left, an imprint of a sporangium associated with leaf between arrowheads to lower right, other sporangia at arrows, 2.4X. D. New leclercqioid plant with 9-segmented leaf, lateral leaf segments in surface view, 4X. 88 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) central, horizontal to downward- extended one (Figs. 7C, 7D). Elongate spo- rangia attach, perhaps at one end, to the upper surface of the undivided por- tion of the leaf (Fig. 7C, arrows). Another genus with segmented leaves, Cer- vicornus wenshanensis, was reported from the Pragian of Yunnan, China (Li and Hueber, 2000). This genus has leaves that are divided into two groups of four (eight segments total) segments, arrayed three-dimensionally. Although lacking important diagnostic evidence such as sporangia and anatomy, Cervi- cornis may be the oldest record of Protolepidodendraceae. Thus, more diver- sification of this group prior to the Middle Devonian is evident. In terms of younger examples, a lycopsid of similar size as all examples of Protolepidod- endraceae, and superficially identical in appearance, has been reported from Lower Carboniferous sediments of Argentina. Frenguellia Arrondo et al. (1991) has leaves with a long median segment, and two pairs of shorter opposite lateral tips, two of which point backward towards the stem. Sporangia are borne on unmodified leaves. However, we are concerned about the lack of conclusive and direct evidence for the present dating of these fossils; they may be substantially older. A second protolepidodendralean type is represented by the Haskinsiaceae (Figs. 8A-8C). This family differs from the Protolepidodendraceae in having more laminate, hastate or sagittate leaves and perhaps in sporangium shape. Leaf morphology varies less within this group than in Protolepidodendraceae. Epiphyllous, globose or flattened ellipsoidal sporangia have been demonstrat- ed in two species of Haskinsia from Venezuela, although mode of dehiscence and spores are unknown (Berry and Edwards 1996). Haskinsia from New York State has a primary xylem column with peripheral ridges of protoxylem sim- ilar to those of Leclercqia and Minarodendron (Grierson and Banks 1983). Archaeosigillariaceae, including the genera Archaeosigillaria and Gilboaph- yton, are sometimes regarded as members of the Protolepidodendrales but fer- tile examples are lacking. Archaeosigillaria has recently been restricted to the type specimen (Berry and Edwards 1997). Archaeosigillariaceae sensu Berry and Edwards (1997) includes Middle-Late Devonian plants which have a hex- agonal pattern of swollen leaf bases on the stem surface, and permanent lan- ceolate (more or less hastate) leaves (Figs. 8E, 8F). A permineralized specimen of Gilboaphyton goldringiae Arnold from Gilboa (Grierson and Banks 1963, PI. 37, figs 4, 7) demonstrates a protostele with approximately eight peripheral lobes in transverse section. Specimens of Gilboaphyton griersonii from Vene- zuela and New York State have similar leaf bases, and Berry and Edwards (1997) were able to infer the presence of elongate-oval, possibly parenchy- matous, regions beneath each leaf base (Fig. 8D). These are suggestive of pos- sible greater differentiation of stem tissue and may represent a prototypical form for one or more types of parenchyma/aerenchyma (e.g. parichnos) asso- ciated with leaf traces found in Late Devonian and Carboniferous lycopsids. Among the Protolepidodendrales, where xylem anatomy is known, second- ary wall pitting patterns include simple and bordered pits as well as annular and scalariform pitting. Minarodendron tracheids also exhibit pitlet sheets which resemble to some extent pitting found in younger lycopsids, termed GENSEL & BERRY: EARLY LYCOPHYTE EVOLUTION 89 3mm { ion Dp \2| Fic. 8. Haskinsiaceae and Archaeosigillariaceae. A) Haskinsia- line drawing of sterile specimen of H. hastata from Venezuela. B) Fertile leaf and sporangium of H. sagittata Edwards and Bene- detto emend. Berry and Edwards fromVenezuela. C) Fertile leaf and sporangium of H. hastata. D, F) Gilboaphyton griersonii from the Devonian of Venezuela. E) Gilboaphyton goldringiae from New York State. A—C) Redrawn from Berry and Edwards (1996), D—F) redrawn from Berry and Ed- wards (1997). Williamson’s striations or fimbrils. Fertile examples, where known, were ap- parently homosporous. No basal regions or rooting structures have been de- scribed. A ligule has been documented only in Leclercqia. Few protolepidodendralean taxa were included in cladistic studies largely because many of them are not completely enough known. Leclercqia and Min- arodendron, the most completely-known examples, consistently form a clade that is sister, or paraphyletic, to Selaginellaceae or as part of a polytomy with 90 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) Selaginellaceae and the rhizomorphic lycopsids (Kenrick and Crane, 1997; Bateman et al., 1992). Further, earlier putative lycopsids such as Drepanophy- cus, Asteroxylon, and Baragwanathia usually are sister to all of these clades. Most troubling in Protolepidodendrales is our current lack of knowledge about their rooting structures. This feature has the potential to reveal crucial infor- mation about relationships of the Protolepidodendrales to other groups, in- cluding the arboreous forms. The lack of well-preserved rooting structures is worrisome in the context of the large amount of above-ground material known and the excellence of some of the descriptions. One possible implication is that these plants may not have had specialized rooting structures. Alternative- ly, rooting structures may have existed, but were not preserved in the same horizons where the stems occur. INNOVATIONS IN GROWTH HABIT, ROOTING, AND REPRODUCTIVE STRUCTURES IN DEVONIAN LYCOPSIDS Until recently, the Protolepidodendrales, or certain less well-known re- mains, were the most likely group from which the Late Devonian and Carbon- iferous arborescent, ligulate, heterosporous lycopsids may have arisen. A cru- cial difference between Protolepidodendrales and aborescent forms however, is that the latter have a bipolar shoot system and erect growth habit. New discoveries from China have provided important evidence about changes in growth habit and rooting structures among Devonian lycopsids. SMALL TREE-LIKE LYCOPSIDS OF THE MIDDLE DEVONIAN.—The earliest examples of true bipolar growth, resulting in upright growth of a trunk and downward growth of a dichotomous rooting system, are to be found in the late Middle Devonian (Givetian) and earliest Late Devonian (Frasnian) of China. The two plants in question are Longostachys (Cai and Chen, 1996) and Chamaedendron (Schweitzer and Li, 1996). Both are small in height (0.5-1.5m), having a slen- der trunk that branches isodichotomously at the top to form a crown of narrow branches and terminal cones, and at the base to form a downward-pointing, conical- shaped, rooting system which divides into four major segments (Fig. 9). Roots may dichotomize further but lack any sign of attched rootlets or stigmarian-type rootlet scars. Longostachys is the larger of the two plants and has secondary growth in both the trunk and the rooting system. Sterile leaves are long and narrow, with spiny margins. Fertile leaves are spoon-shaped (the bowl containing the spo- rangium nearest the axis), and also have spiny margins (Fig. 9). Both of these plants are interesting because they have major features characteristic of rhi- zomorphic lycopsids (habit, dichotomous rooting system, bipolar growth) but lack rootlets. LATE DEVONIAN ROOTING STRUCTURES.—The famous ‘Naples tree’ (Lepidosigil- laria whitei) from the Late Devonian of New York State has previously been interpreted to have a swollen base covered in stigmarian rootlets (White, 1907; Pigg, 1992, 2001). This specimen is poorly preserved, and suffering pyrite de- GENSEL & BERRY: EARLY LYCOPHYTE EVOLUTION 91 15cm Fic. 9. Reconstruction of Longostachys, showing the four-parted — structure and leafy stems, and (inset) a single sporophyll. After Cai and Chen 1996, with permis cay, and the presence of rootlets and rootlet scars can no longer be verified. The poor quality of White’s photographs showing the basal region of the plant compared to his other photographs of the upper trunk would suggest it was no better preserved when he worked on it. Thus we cannot accept this spec- imen as showing the earliest example of a lycopsid rhizomorph. Earlier Middle Devonian stumps named Eospermatopteris (Goldring 1924) which several peo- 92 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) ple considered to be possible lycopsid bases (as noted by Pigg, 1992), are far more likely to be cladoxylalean in affinity (Boyer and Matten 1996; Berry and Fairon-Demaret in progress). Clevelandodendron, from the Late Devonian (Fa- mennian) Cleveland Shale, has an isoetalean habit but again stigmarian root- lets were not demonstrated (Chitaley and Pigg, 1996). Branched rooting struc- tures are reported, but not described in any detail, for the Upper Devonian Cyclostigma kiltorkense (Johnson, 1913). Leptophloeum from the Upper De- vonian (Famennian) of South Africa has a bulbous base from which many thick dichotomous roots emerge all around (Gess and Hiller 1995). These roots are up to 10 mm wide in comparison with the 45 mm bulbous base width and seem too large to be rootlets. This suggests an absence of rootlets in this plant. Lastly, cormose lycopsid plant bases, some with attached rootlets, have been reported from the Upper Devonian (7late Famennian age) Red Hill, Pennsyl- vania deposits (Cressler, 1998, 1999; Scheckler et al., 1999). The previous earliest western occurrence of an undisputed lycopsid rhizo- morph (Pigg, 2001) is the report by Jennings et al. (1983) of Protostigmaria eggertiana from the Mississippian of Virginia, USA. Chinese occurrences in- clude Stigmaria rugulosa Gothan from the Upper Devonian Wutung Formation near Nanjing (Gu & Zhi 1974). Berry and Wang Yi collected dichotomous stig- marian roots with characteristic rootlet scars from Nanshan Quarry, in Jiangsu Province, also from the Wutung Formation (unpublished) where other fossils included Eviostachya and Hamatophyton (Li, Cai and Wang 1995). The facts, as outlined above, demonstrate that the emergence of the lycopsid rhizomorph remains a mystery. According to Rothwell and Erwin (1985) ‘there now can be little doubt that the rooting structures of all rhizomorphic lyco- phytes are shoot systems modified for rooting’ and implied rootlets therefore were homologous to leaves (p. 95). Bateman et al. (1992) however recognize that the stigmarian rhizomorph ‘is a shoot-like developmental system’, but prefer to regard it as a unique organ reflecting limited developmental options within the arborescent lycopsid bauplan’. The new evidence from China sug- gests that the earliest lycopsids with bipolar growth, although possessing a dichotomous, downwards-pointing rooting system, did not possess stigmarian rootlets, supporting Bateman et al.’s hypothesis. Such rootlets are not found until the uppermost Devonian in China and either uppermost Devonian or Lower Carboniferous in the U.S.A. By the time tree-shaped lycopsids appeared in the Middle Devonian, leaves were well developed as structures, and it seems inconceivable that leaves might have been involved in the development of a downward-directed rhi- zomorph as implied by Rothwell and Erwin’s hypothesis. If the K branching recognized above in zosterophylls and pre-lycopsids (Gensel et al., 2001) does in fact represent the earliest stages of bipolar growth in the Lycophytina (sensu Kenrick and Crane 1997) we can propose homologies between structures in zosterophylls and arborescent lycopsids (Fig. 10). In some zosterophylls and in Drepanophycus, K-branching occurs, the posterior (or downwards)-directed branch being naked and the apically (or upwards)-directed branch bearing ena- tions or leaves. The two branches are homologous structures, deriving from a GENSEL & BERRY: EARLY LYCOPHYTE EVOLUTION 93 Zosterophyll or Lycopsid Longostachys Lepidodendrid Early Devonian Middle Devonian Latest Devonian / Carboniferous Fic. Hypothesis of homology of rooting structures between creeping early lycophytes (zos- ail or lycopsid), Longostachys, and a Carboniferous arboreous lepidodendrid. Shoot and rooting structures are homologous, being derived by a branching event. Early rooting structures lack lateral pita rootlets arose later, possibly in early arboreous forms. See text for further explanatio regular dichotomous branch as found in all zosterophylls, but the posterior- directed branch having developed the characters of roots and loss of enations/ leaves. In the earliest pseudobipolar lycopsids (Middle Devonian of China) the K branching occurs precociously at the initial growth of the sporophyte, caus- ing upward growth of the bifurcate, leafy aerial axes and downward growth of the bifurcate, naked rooting system. Final development of the arborescent lycopsid body plan occurs when rootlets evolve de novo on the rooting system to increase efficiency. INNOVATIONS IN FERTILE STRUCTURES AMONG MIDDLE AND LATE DEVONIAN Ly- copsips.—Middle Devonian Protolepidodendrales bear spores in single adaxial sporangia where known. Characteristically the fertile leaves are unmodified compared with the sterile ones. Often this means that the sporangia are borne more or less exposed on narrow leaf pedicels before the leaf forks (e.g. Leclerc- gia, Colpodexylon). In Haskinsiaceae, however, the laminate nature of the leaf and its upcurved pedicel (Figs. 8B, C) means that the sporangia are slightly more protected by the leaf (Berry and Edwards 1996). In those Protolepidod- endrales where spores are known, the plants are homosporous. Sporophylls are often found concentrated into areas of the axes, but show no closer inser- tion nor increased overlap to areas of sterile leaves. In Chamaedendron, leaves are slightly broader near their base, but in Lon- gostachys sterile leaves are narrow and fertile leaves are distinctly enlarged near their base in the area of attachment of the sporangia (Fig. 9), the shape being suggestive of an upturned spoon. The leaf margins of these Chinese lycopsids are spiny. The enlarged bases of the sporophylls are curved around 94 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) sporangia which contain large megaspores, demonstrating that the plant is probably heterosporous (microsporangia are unknown). Leaves are not well enough preserved to know for certain if ligules were present or not. In these small, bipolar, tree-shaped lycopsids the sporophylls are crowded closer to- gether forming distinctive fertile regions, and individual leaves overlap to some extent. Such fertile regions are therefore intermediate between the lax protolepidendrid fertile areas and the compact cones of many more advanced lycopsids. Lycopsid strobili from Kazakhstan, correlated to the upper part of the Giv- etian on the basis of plant fossils, are reported to contain a random mixture of mega- and micro-sporangia (Senkevich et al., 1993). These demonstrate an early record of compaction of sporophylls to form a tight, protective strobilus in lycopsids. A new probably arborescent lycopsid from the uppermost Devonian of Wuxi, China (Berry, Wang, and Cai, in prep), with branches slightly wider than those of Longostachys, demonstrates the earliest known occurrence of a lycopsid that has distinct micro- and megasporangiate cone-like structures. Megasporangiate ‘strobili’ contain leaves similar to those of Longostachys (but up to 90 mm long) whereas the microsporangiate strobili contain closely packed sporo- phylls that are much narrower. The megasporangiate strobili of this lycopsid are not terminal, but occur at dichotomies on the distal axes. Thus some mod- ification and compaction of sporophylls is well underway prior to the end of the Devonian. CONCLUSIONS Important new information has been obtained in recent years about the early history of lycophytes, particularly concerning anatomy and morphology of shoots, the occurrence of rooting structures, and greater diversification and differentiation of fertile regions. Recent broadly based phylogenetic analyses of vascular plants present interesting hypotheses of relationships, the most significant result being recognition of zosterophyllopsids, lycopsids, and of several basal lineages not yet well known. These analyses increase our under- standing of lycophytes, zosterophyllopsids and lycopsids as natural groups, but do not at this stage help significantly to clarify relationships within and between these groups. Moreover, some characters and possibly character states employed in these phylogenetic studies require modification or re-interpreta- tion in order to more fully resolve relationships. For example, the phylogenetic positions of several plants more traditionally viewed as zosterophylls remain unresolved as do the relationships of presumed basal lycophytes from China which show extremely interesting combinations of characters. Interpretation and understanding of important morphological characters and character-state transformations which appear to be critical in delineation of major clades are presently particularly unsatisfactory, and remain a major obstacle to resolution of early lycophyte phylogeny. Important examples include helical versus lin- ear insertion of sporangia, presence/absence of circinate vernation, and inter- GENSEL & BERRY: EARLY LYCOPHYTE EVOLUTION 95 pretation of sporangial shape, orientation, and mode of dehiscence. Re-eval- uation is also particularly needed in linking early lycopsids (?protolepidod- endrids) to the bipolar, increasingly arborescent Late Devonian/Carboniferous forms. Although we can view a simple transformation sequence of, for exam- ple, the Middle Devonian protolepidodendralean fertile leaf via the Middle / Upper Devonian Longostachys sporophyll to the compacted Carboniferous cone, we do not see a corresponding pattern in the evolution of the stigmarian rooting system using the conventional homologies of rootlets and leaves. With dramatic advances in our knowledge of fossils as likely in the next decade as has occurred during the past three decades, and perhaps better understanding of the developmental basis for evolutionary change, we anticipate greater un- derstanding of the early radiation of lycophytes in the years to come. LITERATURE CITED ARNOLD, C.A. 1937. Observations on fossil plants from the Devonian of eastern North America. Ill. Gilboaphyton goldringiae, gen. et _ nov. from the Hamilton of eastern New York. Con- trib. Mus. Palaeontol. Univ. Mich. 5:7 ARRONDO, O., S. N. CESARI and P. R. as 1991. Frenguellia a new As of lycopods from the Early Carboniferous of Argentina. Rev. Palaeobot. Palynol. 70:187-197. BANKS, H. P. 1944. A new Devonian lycopod genus from ee Ses York. Amer. J. Bot. 31: 649-659 BANKS, H. P., P. M. BoONAMO, and J. D. GRIERSON. 1972. Leclercgqia complexa gen. et sp. nov., an lycopod from the late Middle Devonian of eastern New York. Rev. Palaeobot. Palynol. ad 19—40. BASINGER, J. F., M. E. Koryk, and P. G. GENSEL. 1996. Early land plants from the Late Silurian- Early Devonian of Bathurst Island, irene Arctic Archipelago. In: Current Research 1996B; Geological piers of Canada, pp. BATEMAN, R. M. 1992. Morphogenetic eae ppampenes and phylogeny of Oxroadia gracilis png emend., and O. conferta sp. nov.: orn ically preserved lycopods from Oxroad Bay, SE Scotland. Palaeontographica B228:2 BATEMAN, R. M., W. A. DIMICHELE and D. A. WILLARD. 1902 experimental cladistic analysis of anatomically preserved arborescent lycopsids from th on cap as phylogenetics. Ann. Missouri. Bot. Gard. 79:500-559. Berry, C. M. 1996. Diversity and distribution of Devonian Protolepidodendrales (Lycopsida). The Palaeobotanist 45: ile” 16. vo Berry, C. M. and D. Epwarps. 1994. New data on the morphology and anatomy of the Devonian a hana Hueber and Banks from Venezuela. Rev. Palaeobot. Palynol. 81: 41-150. eae e M. and D. Epwarps. 1996. The herbaceous lycophyte genus Haskinsia Grierson and Banks from the Devonian of western Venezuela, with observations on the leaf morphology and fertile specimens. ee J. Linn. Soc. 122:103-122 Berry, C. M. and D. Epwarps. 1997. A new species of the lycopsid Gilboaphyton Arnold from th erouites of coeeanet and New York State with a revision of the closely related genus Archacosiillara Kidston. Rev. Palaeobot. Palynol. 96:47—70. nd BOYER, J. S. a ATTEN. 1996. Anatomy of ee eriana from the upper Middle Devonian ( = srncan of New York. Abstracts—I.0.P.C. V — 1996, p. 13. Santa Barbara: Department of Geological Sciences. CAI, ai and L. CHEN. 1996. On a Chinese Givetian lycopod, L llus ZI and Feng, seis se morphology, anatomy and vontaclrniiiad: oo Abt. " porate 238 Cal, C. Y., Y. Dou and D. oe. 1993. New observations on a Pridoli plant assemblage from 96 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) north Xinjiang, northwest China, with comments on its evolutionary and palaeogeographic significance. Geol. Mag. 130:155-170. CHITALY, S. and K. Picc. 1996. Clevelandodendron ohioensis, gen. et sp. nov., a meaigptet upright lycopsid from the Late Devonian Cleveland Shale of Ohio. Amer. J. Bot. 83 CRANE, P. R. 1990. The phylogenetic context of microsporogenesis. In Hitinay. if asi Knox, R. B., eds. Microspores: stig and ontogeny. Academic Press, London CRANE, P. R. and P. KENRICK. 1997. Diverted development of reproductive organs: a source of ho loge alee in land plants. Pl. Syst. Evol. 206:161-174. an ae W. L . A diverse Late rina flora preserved in a fluvial setting, north-central Pen sii LH a S.A. Amer. J. B CRESSLER, W. L. 1999. Site-analysis ea floristics of the Late Devonian Red Hill locality, Pennsyl- vania, an Archaeopteris-dominated plant community and og tetrapod site. Unpublished Ph.D. Dissertation, Univ. of Pennsylvania, Philadelphia PA, 5 EpDWarbps, D. 1996. New gon into early land ecosystems: a ive jl a Lilliputian world. Rev. Palaeobot. Palynol. 90:159-174 EL Saapawy, W., and W. es Lacey. 1979. Observations on Nothia aphylla Lyon ex Heeg. Rev. Palaeobot. Palynol. 27:119-147. FAIRON-DEMARET, M. 1985. Les ser fossils de l’Emsien du Sart Tilman, Belgique. I. Stockman- sia langii (Stockmans) comb. nov. Rev. Palaeobot. Palynol. 44:243—260. FANNING, U., J. B. Ri RICHARDSON and D. ante 1991. A review of in situ spores in Silurian land plants. pp. 25-47 in S. Blackmore and S.H. Barnes, eds, Pollen and spores, patterns of di- versification. The Seatsenutics Association Special Volume No. 44. Oxford: Clarendon Press. GarraTT, M. J. 1984. The appearance of Baragwanathia (Lycophytina) in the Silurian. Bot. J. Linn. Soc. 89:355—358. GENSEL, P. G. 1982. Oricilla, a new genus referable to the a oe a the late Early evonian “ hye ser n New Brunswick. Rev. Palaeobot. Palynol. 37:3 GENSEL, P. G., 1992. Phylogenetic relationships of the zosterophylls and ean evidence from net el paleoecology, and cladistic methods of inference. Ann. Missouri Bot. Gard. 79: 450-473. GENSEL, P. G., A. E. KAsPER, and H. N. ANDREWS. 1969. Kaulangiophyton, a new genus of plants from the ates of Maine. Torrey Bot. Club Bull. 96:265—276. GENSEL, P. G., E. Koryk and J. F. BASINGER. 2001. Morphology of above- and below-ground structures “a Early Devonian (Pragian-Emsian) plants. In Gensel, P. G. and D. Edwards, eds. Plants Invade the Land: Evolutionary and Environmental Perspectives. Columbia University rat GERRIENNE, P. 1991. Les plantes emsiennes de Marchin (bord nord du Synclinorium de Dinant, ee Il. Forgesia curvata gen. et sp. nov. Compte Rend. Acad. Sci. Paris, Sér. II 313: Prete 2 pm Contribution a |’étude paléobotanique du Dévonien inférieur de Belgique. Les genres nouveaux Ensivalia et Faironella, Mém. Acad. Royale Sci. Belgique Publ., Cl. Sci. Coll. 4°, 3 Ser., T. 1:1-94 GERRIENNE, P. 1996b. Lower Devonian plant remains from Marchin (northern margin of Dinant en Belgium). IV. Odonax borealis, gen. et sp. nov. Rev. Palaeobot. Palynol. 93:89-— GERRIENNE, P. 1999. spies a new name for Forgesia (fossil plants). Taxon 48:61. Gess, R and N. HILLer. . A preliminary catalogue of fossil algal, plant, arthropod, and fish remains fron a “jin mee black shale near Grahamstown, South Africa. Ann. Cape Prov. Mus. (Nat. Hist.) 19:225-304. GOLDRING, W. 1924. The Upper Devonian forest of seed ferns in Eastern New York. New York State us. Bull. 251:5-92. GRIERSON, he D. and H. P. BANks. 1963. Lycopods of the Devonian of New York State. Paleontogr. Amer. 4: 217-295 RIERSON, I a and H. P. BANKS. 1983. A new genus of lycopods from the Devonian of New York State. Bot. J. Linn. Soc. 86:81-101 GENSEL & BERRY: EARLY LYCOPHYTE EVOLUTION 97 GRIERSON, J. D. and P. M. ener ree Leclercqia complexa: earliest ligulate lycopod (Middle Devonian). Amer. J. Bot. 66: Gu and Zui. 1974. Paleozonic wince Pl Ching Fossil Plants of China, 1. Sci. Press, Beijing, 262 pp. lin Chinese] Hao, S. and C. B. Beck. 1993. Further observations on Eophyllophyton bellum from the Lower Devonian — of Yunnan, China. Palaeontographica, Abt. B. Palaophytol. 230:27—41. HUEBER, F, M. 1992. Thoughts on the early lycopsids and zosterophylls. Ann. Missouri Bot. Gard. Feb ee JENNINGS, J.R., E. E. KARRFALT and G. W. oMpigtta 1983. Structure and affinities of Protostig eggertiana. Amer J. JOHNSON, T. 1913. On Sdnndon is kiltorkense Haughton sp. Sci. Proc. Roy. Dub- lin Soc. 13:500-528. Kasper, A. E, 1974. The reproductive organs of Kaulangiophyton and Drepanophycus. Amer. J. ot. 61:16. KASPER A. E. 1977. Anew vhs sot of the Devonian lycopod genus Leclercqia from New Brunswick, anada. Amer. J. Bot. 154:39. aperenen A. E. and W. H. nie 1979. The Devonian lycopod OE ON oe the Trout Valley Formation of Maine. Geol. Soc. of Maine, Maine Geology Bull. 1:49- KENRICK, P. and P. R. CRANE. 1997. The origin and early sedi ay of land plants: a cladistic study. Smithsonian Series in Comparative Evolutionary Biology. Washington: Smithsonian Institution Press. 441 p. Kerp, H., H. Hass, and V. MosBRUGGER. 2001. New data on Nothia aphylla Lyon 1964 ex El- ete et Lacey 1979, a poorly known plant from Lower Devonian Rhynie Chert. In: Gen- sel, P. G. and D. Edwards, eds. Plants po oon the Land: Evolutionary and Environmental escehiienh Columbia University Press. P. 52-82. KipsTOoN, R. 1901. ae lycopods and sphenophylls. Trans. Nat. Hist. Soc. Glasgow Trans. Ser. 3, 6:25—14 Kortyk, M. E. 1998. ae Silurian and Early Devonian Fossil Plants us apne Island, Arctic Canada. M.Sc. thesis, University of Saskatchewan, Saskatoon, Can. Kotyk, M. E. and J. F. BASINGER. 2000. The Early Devonian (Pragian) cme Bathurstia denticulata Heuber. Can. J. Bot. 78:193—207. KRaAnz, H. D. and V. A. R. Huss. 1996. Molecular evolution of pteridophytes and their Payee 4 to seed plants: bau from gr 18S rRNA gene sequences. Plant Sys. Evol. 2 KRAUSEL, R. and H. WEYLAND, 1932. Pflanzenreste aus dem Devon. IV. pee an ar Seackesibetgiarik 14:391—403. Li, C. S. 1990. Minarodendron cathaysiense (gen. et comb.nov.), a lycopod we wa late Middle Devonian of Yunnan, China. Palaeontographica, Abt. B, Palaophytol 220:9 Li, C. S. 1992. Hsiia robusta, an Early Devonian plant from Yunnan Province, Gin a its bearing on some structures of early land plants. Rev. Palaeobot. Palynol. 71:121-147 Li, C. S. and D. Epwarps. 1992. A new genus of Early Devonian land plants with novel strobilar construction from the Lower Devonian Posongchong Formation, Yunnan Province, China. Palaeontology 35: 257-272. Li, C. S. and D. Epwarps. 1995. A re-investigation of Halle’s Drepanophycus spinaeformis Goepp. from the Lower Devonian of Yunnan Province, southern China. Bot. J. Linn. Soc. 118:163- 192. Li, C. S. and D. Epwarps. 1996. Demersatheca Li et Edwards, gen. nov., a new genus of early land plants from the Lower Devonian, Yunnan Province, China. Rev. easing agi oe 93: 77-88. Li, C. S. and F. M. Hueser. 2000. Cervicornus wenshanensis, gen. a Pragian (Early re Nasiaie plant with forked leaves from Yunnan, China. Rev. cae it Palynol. 109:113- 19, Li, C. ;. F. M. Hueser and C, L. Hotron. 2000. A neotype for Drepanophycus spinaeformis Gép- pert 1852. Can. J. Bot. 78:889-902. Li, X., C. Cat and W. Yi. 1995. Hamatophyton verticillatum (Gu & Zhi) emend. A primitive plant of sphenopsida from the Upper Devonian-Lower Carboniferous in China. Palaeontographica, Abt. B, Paliophytol 235:1-22. 98 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) Lyon, A. G. 1964. — fertile region of Asteroxylon mackiei K. and L. vionyesd 203:1082—1083. MISHLER, B. D., L. A. LEwis, M. A. BUCHHEIM, K. A. RENZAGLIA, D. J. GARBARY, C. F. DELWICHE, F. W. T. . KANTz and R. L. CHAPMAN. 1994. per eri ae of the “green algae” and “bryophytes. ” Ann. Missouri. Bot. Gard. 81:4 QMLLGAARD, B. 1987. A revised classification of the hie el s. lat. Opera Botanica 92:153- Picc, K. B. 1992. Evolution of isoetalean lycopsids. Ann. Missouri Bot. Gard. 79:589-612. Ficc RB. wa Isoetalean lycopsid evolution: from the Devonian to the present. Amer. Fern J. 91: 99-11 RAUBESON, L. m4 and R. K. JANSEN. 1992. se ea gran evidence on the ancient evolutionary split in one land plants. Science 255:1 RAYNER, R. J. 1 New finds of ea, spinaformis Goppert from the Lower Devonian of Scie? Tena: Roy. Soc. Edinb. E 363. RickarDs, R. B. 2000. The age of the earliest club mosses: the Silurian Baragwanathia flora in Victoria, Australia. Geol. Mag. 137:207—209. ROTHWELL, G.W, and Erwin, D. M. 1985. The rhizomorph apex of Paurodendron; implications for homologies among the rooting organs of lycopsida. Amer. J. Bot. 72:86—98. KLER, S. E., W. L. CRESSLER, T. CONNERY, S. KLAVINS and D. L. PostNikoFr. 1999. Devonian shrub and tree epee eaperiig 2 XVI International Botanical Congress, paseatiie 13. SCHWEITZER, H. J. and P. Gre 0. Uber Taeniophyton inopinatum, Protolycopodites devoni- cus, und Cladoxylon scoparia aus dem Mitteldevon von Wuppertal. Palaeontographica, Abt. B, Palaophytol 17 SCHWEITZER, H. J. and C.S. . on Chamaedendron nov. gen., eine ae a epg lycophyte aus dem Frasnium siidchinas. Palaeontographica Abt. B, Palaophytol. 2 SENKEVITSCH, M. A. 1971. New Devonian Lycopodiales. Mat. Geol. Miner. ees .. “Nau- ka” Kazakhskoi SSR, 4:88-92. SENKEVITSCH, M. A., A. L. JURINA and A. D. ARKHANGELSKAYA. 1993. On fructifications, morphology and anatomy of Givetian lepidophytes in Kazakhstan (USSR). Palaeontographica Abt. B, Pa- laophytol 230:43-58 STEWART, W. N. and G. Ro OTHWELL. eae Paleobotany and the evolution of plants. Cambridge Uni rp cs Cambridge. 52 TAYLOR, T. N 8. The origin of lend plants: some answers, more questions. Taxon 37:805-833. THOMAS, B. A. ghee S. D. BRACK-HANES. 1984. A new approach to family groupings in the lyco- phytes. Taxon 33:247—2 sees Tims, J. D. and T. C. CHAMBERS. 1984. Rhyniophytina and Trimerophytina from the early land flora of Victoria, Australia. Palaeontology 27:265-279. WacneRr, W. H. and J. BEITEL. 1992. Generic classification of modern North American Lycopodi- aceae. Ann. Missouri “a Gard. 79:676-686. Wuire, D. 1907. A remarkable fossil tree trunk from the middle Devonic of New York. Bull. New York State Museum, (Geol. Pap.) 107:328-340. American Fern Journal 91(3):99—114 (2001) Isoetalean Lycopsid Evolution: from the Devonian to the Present KATHLEEN B. PIGG Department of Plant Biology, Box 871601, Arizona State University, Tempe, AZ 85287-1601 .s fs. p eS ABSTRACT.—The evolution of the isoetalean id t vascular plants, from Late, (or possibly Middle), Devonian to the. current day genus Isoetes. The fhe ate fossil members of this group are the arborescent lepidodendrids that dominated the Late Carboniferous coal swamps. Simpler unbranched isoetaleans with elongate stems also predated, coexisted sae and postdated the coal swamp trees, extending well into the Mesozoic. Whereas certain syna morphies such as stigmarian rootlets, bipolar growth and secondary tissues unite the clade, tains features characterize smaller subgroups of differing age, growth form and possibly, evolutionary lineage. Although some of these features are well known for plants of given time periods, partic- ularly the Carboniferous, trends in character evolution have never been adequately docume nted olete microspores, sunken sporangia and elaborate ligules with glossopodia occur within elongate- stemmed Triassic forms. The dominant plant habit of modern Isoetes, a reduced cormose form that lacks appreciable stem elongation, originated at least by the Jurassic and typifies late Mesozoic and Cenozoic isoetaleans. The lycopsids have always held a particular fascination for paleobotanists and neontologists alike because of their diverse morphologies, long strati- graphic record and their phylogenetic position as the outgroup to the rest of the vascular plants (Gifford and Foster, 1989; Kenrick and Crane, 1997). Whereas most lycopsids are similar to other pteridophytes in their growth habit and life history, isoetalean plants are characterized by distinctive mor- phological innovations, including secondarily derived bipolar growth, second- ary cortical (and possibly vascular) tissues and arborescence. This group is generally thought to make its first appearance in the Late Devonian, or perhaps earlier, and reach its greatest diversity and ecological importance in the Late Carboniferous. Isoetaleans continue into the Early Mesozoic, where they are represented worldwide by numerous species of Pleuromeia Corda, Annalepis Fliche and several other forms. In the Jurassic, plants quite similar to modern Isoetes L. first appear. They continue to diversify throughout the Cretaceous and Tertiary and up to the present day. The best known members of the Isoetales are the arborescent lepidoden- drids, a group that dominated the Late Carboniferous ( = Middle-Upper Penn- sylvanian) coal swamp of Euramerica. These heterosporous trees had multi- branched crowns of varying forms, long microphyllous leaves attached to char- acteristic diamond- or hexagonal-shaped leaf cushions, thick secondary cor- tical tissues, well developed stigmarian rooting systems and often monosporangiate cones with reduced numbers of megaspores and specialized 100 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) modifications for dispersal (Phillips, 1979; Bateman et al., 1992; Phillips and DiMichele, 1992). Most of the arborescent forms became extinct by the Late Carboniferous of Euramerica concomitant with the loss of the most extensive coal swamp habitats (Bateman, et al., 1992; Stewart and Rothwell, 1993). How- ever, growing alongside the Middle Pennsylvanian tree lycopsids and surviv- ing after them into the Late Carboniferous (Upper Pennsylvanian) were smaller (1-2 m tall), unbranched or rarely branched isoetaleans such as Chaloneria Pigg & Rothwell and Sporangiostrobus Bode (Pigg and Rothwell, 1979, 1983a, 1983b, 1985; Wagner, 1989). It is now known that isoetaleans with this elon- gate, unbranched growth habit originated by the Late Devonian (e.g., Cleve- landodendron Chitaley & Pigg, Chitaley and Pigg, 1996) and were widespread throughout the Triassic where they belonged to at least two broad groups, Pleuromeia-like and Annalepis-like forms and their relatives (Pigg, 1992; Grau- vogel-Stamm and Lugardon, 2001). Today the order Isoetales is represented by a single genus, Isoetes, which remains one of the most unusual living pteri- dophytes and retains many morphological, anatomical and developmental fea- tures that reflect its origins (e.g., Stewart, 1947; Gifford and Foster, 1989; Bate- man, et al., 1992; Pigg, 1992). Recent studies have provided new information about the origin, evolution and diversity of the Isoetales. In this contribution, four phases of evolution in isoetalean lycopsids are addressed: (1) presumed Devonian origins of the Is- oetales; (2) their maximum diversity in the Late Carboniferous, (3) Mesozoic plants with the growth habit of modern Isoetes, (4) new occurrences of Cre- taceous and Tertiary forms. Lastly, problems of relating fossil taxa to extant Isoetes are addressed and suggestions are outlined for a better understanding of the isoetalean lineage through time. ORIGIN OF ISOETALEAN LYCOPSIDS Historically, the Isoetales, the rhizomorphic heterosporous lycopsids, have been assigned to a variety of orders including the Lepidodendrales, Pleuro- meiales, Isoetales and some forms occasionally to Protolepidodendrales (Chal- oner, 1967; Pigg and Rothwell,1983b; DiMichele and Bateman, 1996; Kenrick and Crane, 1997), and even, probably mistakenly, the Selaginellales (see Roth- well and Erwin, 1985). In this paper I follow DiMichele and Bateman (1996), who recognize the entire group within the order Isoetales, based on a suite of synapomorphies including bipolar growth, secondary tissue production, and stigmarian rootlet formation. Whereas much is known about both vegetative and reproductive structure of fossil isoetaleans, particular attention has been given to the basal parts, or rhizomorphs of the plants, especially the size and shape of their lateral lobes or horns. A great deal of diversity has been documented for these rhizomorphic features which vary from elongate, dichotomizing rooting systems of Stigmaria Brongniart to smaller bilaterally symmetrical lobed or rounded forms (e.g., Pleuromeia), to radial unlobed forms (e.g., Paurodendron Fry). All of these thizomorphic forms can be homologized (Rothwell and Erwin, 1985; Grauvo- PIGG: ISOETALEAN LYCOPSIDS 101 gel-Stamm and Lugardon, 2001). It is now clear that the shape of the rhizo- morph, or even its symmetry, is not nearly as significant as its production of stigmarian appendages or “rootlets” with their characteristic anatomy and the likelihood of their secondary derivation from an aboveground shoot system. This fascination with the basal portions of the plants has been particularly centered around the stigmarian rooting system and its ubiquity among Late Carboniferous coal swamp lepidodendrids. There has been interest for some time in finding evidence for the origin of Stigmaria, with the underlying as- sumption that this would lead toward understanding the origin of the isoeta- lean clade as a whole. Devonian and Early Carboniferous plants that obtained tree stature and/or those which produced rhizomorphic plant bases have been sought as the ancestors of lepidodendrids. The best known candidates have been the enigmatic ‘Naples tree” Lepidosigillaria whitei Kraéusel & Weyland from New York State and Cyclostigma kiltorkense Haughton, from Late De- vonian strata in several areas of Ireland and Bear Island in the Arctic, and hina. Lepidosigillaria whitei Krausel & Weyland, a large stem with a plant base that appears rounded and lacking a stigmarian rooting system, occurs in the Late Devonian of New York State (Grierson and Banks, 1963). The original specimen was over 5 m tall and demonstrated that well developed, arborescent lycopsids had clearly evolved by the Late Devonian (White, 1907; Krausel and Weyland, 1949; Grierson and Banks, 1963; Chaloner, 1967). Because of lateral compression during fossilization, the exact nature of the plant base of Lepi- dosigillaria is unknown. However, there is some increased thickening of basal cortical tissues, suggesting secondary tissues were produced. Several distinc- tive decortication layers are recognized on the stem surface, and small (less than 3 cm long), very narrowly attached leaves occur at higher levels. Al- though Lepidosigillaria has been described as having rootlets attached to the plant base (White, 1907), preservation is poor and their presence cannot be verified (Gensel and Berry, 2001). A second Late Devonian lycopsid, Cyclostigma kiltorkense Haughton, has also been difficult to interpret. This plant was originally described from the Upper Devonian of Ireland, and later found in Bear Island, in the Arctic, and China (Johnson, 1913; Chaloner, 1967, 1968; Schweitzer, 1969; Cai and Wu, 1994). Cyclostigma is described as a large, heterosporous plant with charac- teristic rounded leaf scars (Chaloner, 1967) and is reported to have a bilobed plant base (Johnson, 1913). Specimens showing plant bases may be known (Chaloner, pers. comm. 1999) but they have never been clearly illustrated and have yet to be studied in detail. It also remains unclear whether all plant remains assigned to this genus represent the same plants. Clearly, reinvesti- gation of this problematic genus is in order. Despite the currently limited information about Lepidosigillaria and Cyclos- tigma, their occurrences have established the presence of large isoetalean plants at least as early as the Late Devonian. In recent years additional Late Devonian and older, Middle Devonian lycopsids with a presumed relationship to the Isoetales have been reported. The Middle Devonian Longostachys latis- 102 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) @ ia Fic. 1. Line diagram of the rootstock of Longostachys latisporophyllus. Redrawn from Cai, Chon- gyang, and Lizhu Chen, 1996. porophyllus Zhu, Hu & Feng emend. Cai & Chen was described from Givetian strata of southern China (Cai and Chen, 1996; Gensel and Berry, 2001). Lon- gostachys is reconstructed as a small, branching heterosporous lycopsid tree that was up to 1.5 m tall with a stem 3.5 cm wide and a plant base that had a three-dimensional funnel-like structure with thick, rootlike appendages (Fig. 1). Because of its bipolar growth and centralized rooting system, this Middle Devonian plant may have some significance to isoetalean origins. The rooting structure of Longostachys bears some resemblance to stigmarian appendages, however it apparently lacks the dichotomizing stigmarian axes and character- istic stigmarian rootlet scars. Given the plant’s Middle Devonian age it may be that these rooting organs are modified branching axes that cannot be clearly defined as stem or root, and may be more comparable to the H & K branching characteristic of zosterophylls (Gensel and Berry, 2001). However, well-defined leaves and strobili are produced by Longostachys, indicating that true leaves and stems had evolved by then. Whatever their exact homologies, the rooting structures of Longostachys suggest a potential intermediate step in the evolu- tion of bipolar growth and the origin of stigmarian rooting systems in isoeta- leans. Additional Late Devonian isoetaleans have been discovered in Pennsylvania and Ohio. A flora from Red Hill, Pennsylvania contains cormose lycopsid plant bases (Cressler, 1998, 1999; Scheckler, et al., 1999). Five specimens with at- tached rootlets have been recovered from oxbow lake facies of this fluvial deposit. One specimen has a stem width of 10 cm and a flared base of 12 cm, and with two lobes visible and is estimated to have four lobes. A second spec- imen is approximately 2 cm wide with a flared base. Roots are borne in or- thostichies, with largest roots adjacent to one another. Associated, usually un- PIGG: ISOETALEAN LYCOPSIDS 103 branched, stems have various decortication patterns including Knorria Stern- berg, Helenia Zalessky and Cyclostigma, as well as Lepidodendropsis Lutz-like pseudowhorls. Reproductive structures are not yet known (Walter L. Cressler, University of Pennsylvania, pers. comm. 1999). The Elkins, West Virginia lo- cality that provided the earliest seed plant, Elkinsia, is of comparable age to Red Hill, and also contains lycopsid stems but plant bases are not yet known from Elkins (Gar W. Rothwell, Ohio University, pers. comm.1999). Another recently recognized Late Devonian plant with an elongate isoeta- lean plant habit is Clevelandodendron ohioensis Chitaley & Pigg from the Cleveland Shale (Chitaley and Pigg, 1996). The genus is based on a single specimen which represents the entire plant. Clevelandodendron is an un- branched, slender plant with a compact terminal bisporangiate cone at the apical end and a partially preserved plant base proximally. The specimen is 125 cm long from base to apex and 2 cm wide for most of its length. Although details of the plant base are obscure, it appears to bear several thick append- ages that taper distally and an elongate stigmarian root system is lacking. The terminal bisporangiate cone contains trilete megaspores that are laevigate (smooth) and trilete microspores that conform to either Calamospora Schopf, Wilson & Bentall or Punctatisporites (Ibrahim) Potonié & Krempf. Ligules are not known. Clevelandodendron is of particular significance because it dem- onstrates that elongate, slender isoetalean lycopsids with an unbranched plant habit, more common in later isoetalean forms, were present as early as the Late Devonian. Plant remains with similarities to Clevelandodendron are also known in younger, Lower Carboniferous strata at several localities in eastern North America, France and Ireland. For example, anatomically preserved stems of similar dimensions are present in Lower Carboniferous localities in France (Brigitte Meyer-Berthaud, Université Montpellier, pers. comm. 1996) and ad- ditional permineralized stems showing cortical and stelar features transitional between the traditional Protolepidodendrales and Lepidodendrales occur at a variety of localities in North America, (e.g., Andrews, Read and Mamay, 1971; Cichan and Beck, 1987; Roy and Matten, 1989), Ireland (Matten, 1989; Roy and Matten, 1989), Scotland (Beck, 1958), and France (Meyer-Berthaud, 1984). Several bisporangiate cones with spore types similar to those of Clevelandod- endron have been described from Upper Devonian and Lower Carboniferous strata of France and eastern North America (Arnold, 1933, 1935, 1939; Ma- thews, 1940; Meyer-Berthaud, 1984). These bisporangiate cones with laevigate (smooth) megaspores are unlike the stratigraphically younger bisporangiate cones of Flemingites Carruthers. The genus Flemingites bears Lycospora Schopf, Wilson & Bentall microspores with an equatorial flange, and Lageni- cula (Bennie & Kidston) Potonié & Krempf or Lageniosporites Potonié & Krempf megaspores with an apical spore extension called a gula (Brack-Hanes and Thomas, 1983). Thus, not all Paleozoic bisporangiate lycopsid cones can be assigned to Flemingites (Chitaley and Pigg, 1996). Species of Flemingites are thought to have been borne by small lepidodendrid trees such as Paraly- copodites brevifolius (Williamson) DiMichele (e.g., Brack, 1970; DiMichele, 104 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) 1980). Thus there may be at least two lineages of Paleozoic lycopsids that produced bisporangiate cones: those related to Clevelandodendron and those bearing Flemingites cones. One interesting Lower Carboniferous plant of isoetalean affinities is the Pro- tostigmaria/Lepidodendropsis plant. Protostigmaria Jennings plant bases are reported to have had up to 13 lobes and a maximum width of 32 cm. This multilobed rooting structure was initially described by Jennings (1975) and later studied by Jennings, Karrfalt and Rothwell (1983). These authors sug- gested that Protostigmaria grew like Isoetes, with new lobes being produced by forking at the ends of furrows. The presumed above-ground parts of Pro- tostigmaria, assignable to Lepidodendropsis Lutz are found both at the same and at several additional nearby sites in the Mississippian Price Formation of Virginia (Patricia G. Gensel, personal communication, 1999). Above-ground parts of Lepidodendropsis suggest the plant was a tree that branched dichot- omously and may have had determinate growth. It was apparently eligulate but heterosporous and had linear leaves. Middle Devonian fossils assigned to Lepidodendropsis may represent a mix of several different plants that may not be equivalent to this Lower Carboniferous plant. Study and correlation of Pro- tostigmaria plant bases and vegetative stems at several decortication layers of Lepidodendropis and reproductive remains from the same localities will pro- vide the opportunity to reconstruct this interesting Lower Carboniferous ly- copsid (Gensel and Pigg, in progress). LATE CARBONIFEROUS DIVERSITY During the Late Carboniferous, diversity and ecological importance of isoe- talean plants reached their maximum. Arborescent lepidodendrids are exten- sively well known as both permineralizations in coal ball floras, and as coal- ified compressions. Whole plant reconstructions, reproductive biology and pa- leoecological aspects of these plants have been reviewed elsewhere (e.g., Phil- lips, 1979; Bateman, et al., 1992). The extensive information available about Carboniferous lepidodendrids has allowed for the whole plant reconstruction of several species including the bisporangiate Paralycopodites, and larger monosporangiate trees with special- ized dispersal units that have been modified presumably for water dispersal (Phillips, 1979). These include Diaphorodendron DiMichele and Synchysiden- dron DiMichele & Bateman, genera with two different crown branching forms that both produce Achlamydocarpon varius (Baxter) Taylor & Brack-Hanes, a megasporangium with a specialized, thickened wall surrounding a single func- tional megaspore (DiMichele and Bateman, 1992). Another tree lycopsid, Lep- idophloios Sternberg, produces Lepidocarpon Scott, a cone that breaks up into individual megasporophyll units that contain the megasporangium with a sin- gle functional megaspore surrounded by an elongate and elaborated sporophyll (Phillips and DiMichele, 1992). Another tree form, Sigillaria Brongniart, bore monosporangiate cones on lateral cauline peduncles. Permineralized megas- porangiate cones of Sigillaria, Mazocarpon (Scott) Benson produced 4-8 large PIGG: ISOETALEAN LYCOPSIDS 105 megaspores that are found each associated with a portion of sporangial wall and a pad of sterile tissue (Schopf, 1941; Pigg, 1983; Phillips and DiMichele, 1992). Parallels of these specialized forms with the seed habit have been com- monly made (e.g., Phillips, 1979; Thomas, 1981). We now know that in addition to the large arborescent lepidodendrids, which themselves were quite diverse, isoetaleans had a variety of other plant habits. These included the small scrambling “‘pseudoherbs” Oxroadia Alvin and Paurodendron Fry, as well as the unbranched monopodial plants Chalo- neria Pigg & Rothwell and Sporangiostrobus Bode. Oxroadia, a plant of Lower Carboniferous age, is known from a number of localities in England and Scot- land (Bateman, 1992), while Paurodendron has been described from Middle and Late Pennsylvanian coal ball floras of eastern and midcontinent North America (Rothwell and Erwin, 1985). Comparative developmental studies of apical and lateral growth of rhizomorphs in Paurodendron, Stigmaria, Na- thorstiana Richter and modern Isoetes have been instrumental in understand- ing homologies across the isoetalean clade (Karrfalt, 1984; Rothwell, 1984; Rothwell and Erwin, 1985; Rothwell and Pryor, 1991). The genus Chaloneria is reconstructed as a whole plant based on perminer- alized stems, fertile areas with in situ spores, decortication surfaces, and mega- gametophytes (Pigg and Rothwell, 1979, 1983a, 1983b, 1985). In contrast to the well known lepidodendrids, Chaloneria was more like its Devonian rela- tive, Clevelandodendron, in having an unbranched, elongate stem. Instead of compact cones Chaloneria exhibits a less well differentiated fertile area that occurs either apically with alternating mega- and microsporangiate zones (C. cormosa Pigg & Rothwell) or alternating vegetative and fertile zones (C. per- iodica Pigg & Rothwell; DiMichele, Mahaffy and Phillips, 1979; Pigg and Roth- well, 1983a, 1983b). These and presumably related forms share the megaspore type Valvisisporites Ibrahim (Gastaldo, 1981) and microspores of Endosporites Wilson & Coe (e.g., Brack and Taylor, 1972; Hanes, 1975). Compressed lycopsid reproductive structures bearing the same spore types as Chaloneria are given the name Polysporia Newberry (Chaloner, 1958; Dra- bek, 1976a, 1976b; DiMichele, et al. 1979; Pigg and Rothwell, 1983b). Previous reports of Polysporia were from incompletely preserved specimens, and it was difficult to tell whether these structures were compact cones or fertile areas such as those of Chaloneria. A recent study of well-preserved, complete cones of a new species of Polysporia from France, P. doubingeri Grauvogel-Stamm & Langiaux shows that Polysporia was at least sometimes a compact cone rather than an undifferentiated fertile area (Grauvogel-Stamm and Langiaux, 1995). This study brings to question whether all plants that produced these charac- teristic spore types, well known from Middle Pennsylvanian coals (Pigg, 1992), had the same type of growth habit (Pigg and Rothwell, 1983b, Grauvogel- Stamm and Langiaux, 1995). Sporangiostrobus Bode is known from coal ball floras of the Middle Penn- sylanian of Kansas and compressed floras from the Stephanian ( = Upper Pennsylvanian) of Spain (Pigg and Rothwell, 1983b; Wagner, 1989 and refer- ences cited therein). This plant parallels the Chaloneria/Polysporia plants in 106 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) stem anatomy, size, unbranched to rarely branched habit and presence of fer- tile areas rather than compact cones. These plants grew in nearly monotypic stands, a situation also common in later pleuromeian forms (e.g., Retallack, 1975; Fuchs et al., 1991; Pigg, personal observation, 1991). In contrast to Chal- oneria/Polysporia plants, Sporangiostrobus bore the highly variable Densos- porites (Loose) Schopf, Wilson & Bentall microspores, characterized by a prom- inent equatorial ridge, the cingulum, and Zonalesporites (Bartlett) Leisman megaspores. Whereas Pigg and Rothwell (1983b) suggested that Sporangiostro- bus may be assignable to the same family as Chaloneria, the Chaloneriaceae, similarities of the two groups may be more the result of convergence in growth habit. Several other Carboniferous forms which, for various reasons, have been allied with the isoetaleans are reviewed by Pigg (1992). ORIGIN OF MODERN ISOETES PLANT HABIT AND MORPHOLOGICAL FEATURES As the Paleozoic ended and most of the large, specialized lepidodendrid plants died out, a variety of several smaller, elongate, unbranched forms re- mained. These are represented in the Mesozoic by the genera Pleuromeia Cor- da, Annalepis Fliche and several other diverse forms (Pigg, 1992; Skog and Hill, 1992; Retallack, 1997; Grauvogel-Stamm and Lugardon, 2001). Around a dozen species of Pleuromeia have been named from numerous localities throughout the world including Germany, Russia, China, Japan and Australia (Pigg, 1992). Some pleuromeian genera, such as Cylomeia (White, 1981) and Lycomeia (Dobruskina, 1985) have been proposed on basis of the geographic occurrence, morphology and mono- vs bisporangiate nature of plants. How- ever, whereas numerous species of Pleuromeia have been described and some are well known (e.g., Grauvogel-Stamm, 1993), the overall diversity and inter- relationships of these taxa remain obscure. In addition to pleuromeians there are a suite of Triassic plants assigned to Annalepis and related taxa which have been described by Grauvogel-Stamm and colleagues (Grauvogel-Stamm and Duringer, 1983; Grauvogel-Stamm and Lugardon, 2001). This latter group appears distinct from pleuromeians on the basis of a number of features including sporophyll morphology and spore char- acters. In some features Annalepis may resemble some later isoetaleans of Cre- taceous-Tertiary ages (Grauvogel-Stamm and Lugardon, 2001). A third source of Triassic lycopsids is the discovery of several permineral- ized forms. These include Takhtajanodoxa Snigirevskaya from Russia (Snigir- evskaya, 1980a, 1980b), forms attributed to Pleuromeia, also from Russia (Sni- girevskaya and Srebrodelskaya, 1986) and the more recently reported permi- neralized lycopsids from the Lower Triassic of Australia (Cantrill and Webb, 1998). The affinities of permineralized Triassic forms and their relationships to the more widely known compressed remains have not been fully resolved. In the early 1980’s Snigirevskaya described the permineralized Triassic plant Takhtajanodoxa from Russia (Snigirevskaya 1980a, 1980b). This plant has stem, leaf and megaspore anatomy similar to Chaloneria and other Pennsy]- vanian lycopsids. What is surprising, however, is the presence of elaborately PIGG: ISOETALEAN LYCOPSIDS 107 shaped ligules with curving basal extensions of the glossopodium (Snigirev- skaya, 1980a, 1980b; Pigg, 1992). This type of ligular structure is lacking in Paleozoic permineralized lycopsids but present in extant Isoetes (Gifford and Foster, 1989). Permineralized stems described as Pleuromeia were also char- acterized from Russian Lower Triassic strata (Snigirevskaya and Srebrodel- skaya, 1986). These specimens show small, protostelic stems lacking second- ary tissues with mesarch leaf traces, and fimbriate metaxylem tracheids with scalariform and reticulate wall thickening patterns. More recently, perminer- alized forms have been described from the Lower Triassic Bowen Basin of eastern Australia by Cantrill and Webb (1998). These authors describe a suite of disarticulated lycopsid plant organs that include one bisporangiate and one monosporangiate cone type, two types of stems, one with prominent aeren- chyma in the cortex, and one type of rooting structure that bears stigmarian- like rootlets. From the presently known information, at least two types of ly- copsids thus appear to be present in these remains. As with late Carboniferous coal ball lycopsids, for the first time the exciting potential exists for “whole- plant” reconstructions of Lower Triassic isoetaleans, based on the co-occur- rence, anatomical similarity and organic interconnections between these sep- arate plant organs. It is during the Mesozoic that we see the origin of a number of morphological features present in modern-day Isoetes. These features include: 1) a change in microspores from trilete to monolete suture (and probably also change in tetrad arrangement); 2) sporangia that are sunken into the adaxial surface of the spo- rophyll; 3) an elaboration of the basal portion of the ligule into a glossopo- dium; 4) the origin of a velum and/or a labium, covering the proximal, distal, or entire upper surface of a sporangium. All of these features can be seen among various Triassic-aged plants. Micro- sporophylls of Annalepis are characterized by monolete microspores of the Aratrisporites (Leschik) Playford & Dettman type (Grauvogel-Stamm and Lu- gardon, 2001). Sporangia are notably sunken into the sporophyll surface in several species of Pleuromeia such as P. rossica Neuburg and P. longicaulis (Burges) Retallack (Neuburg, 1960, 1961; Retallack, 1975). Ligules with an ex- tended basal glossopodium much like that present in many extant species of Isoetes occur in the permineralized Triassic Russian plant Takhtajanodoxa (Snigirevskaya 1980a, 1980b; Pigg, 1992). Whereas it appears that no single currently known Triassic isoetalean taxon possesses all of these features, all are found within various representatives of Triassic lycopsids. As the Triassic and later Cenozoic lycopsids are better doc- umented, it will be possible to better understand character evolution in the various lineages of Triassic lycopsids (Grauvogel-Stamm and Lugardon, 2001). This suggests the Triassic was a time of important radiation and change in several key morphological characters prior to the appearance of the extant, reduced Isoetes forms. The origin of the modern Isoetes plant habit with a nonelongated stem was previously thought to occur in the Triassic (Pigg, 1992). However, evidence for this type of plant habit is scant, as the best documented Triassic Isoetes-like 108 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) ~, ¥ a ee ss a \ j Fics. 2-5. Tertiary and extant isoetalean lycopsids. Fig. 2. cf. Isoetites from Dennison Gap, Wy- oming. Florida Museum of Natural History. UF 18120-12604. Fig. i ni s PIGG: ISOETALEAN LYCOPSIDS 109 Although it is clear that all of these forms are isoetalean plants, none show undisputed evidence of a non-elongating stem such as is usually characteristic of modern Isoetes. They should certainly be included within the Isoetales, based on their particular morphological features, yet I am reluctant to accept them as belonging to the modern genus. The earliest strong evidence of the modern plant habit is that seen in the Jurassic Isoetites rolandii Ash & Pigg from western North America (Ash and Pigg, 1991). Concomitant with this change in plant habit is the loss of vege- tative leaves: modern Isoetes and its younger fossil relatives typically produce only fertile leaves, thereby minimizing the vegetative phase of its life history. CRETACEOUS AND TERTIARY—SOME NEW OCCURRENCES Over a dozen Isoetes-like plants have been described in Cretaceous and Ter- tiary strata (see Pigg, 1992, and Skog and Hill, 1992 for review). Several newly recognized forms have been recently described. One unusual example is Mon- ilitheca Krassilov & Makulbekov described from the Upper Cretaceous of Mon- golia by Krassilov and Makulbekov (1996). The genus consists of fragmentary linear megasporophylls up to 18 mm long and 1 mm wide that bear within the sporangium, a single row of megaspore tetrads. These tetrads are borne along almost the entire length of the megasporophyll, which has a short sterile tip. Megaspores are trilete, and reticulat Isoetalean plants described from the ‘Lower Cretaceous of Tunisia were named Isoetites daharensis Barale (1999). These fossils are represented by corms 10 cm long X 6 cm wide with attached sporophylls and roots. Sporo- phylls are ligulate and bear sporangia containing trilete megaspores 480 pm in diameter. Isoetalean plants have also been recognized in the lower Creta- ceous of Brazil, in the Santana Formation (Dilcher et al., 2000). In this flora, Isoetes-like plants are represented by clusters of sporophylls that occur along with ferns, conifers, angiosperms, and abundant gnetalean remains. In the Tertiary, plants from the Paleocene Joffre Bridge locality in central Alberta, Canada originally identified by Hoffman as lilean monocots are clearly isoetalean (Hoffman, 1995, Plate 14, figs. 101, 102). The specimen is a plant base approximately 5.5 cm long and around 3 cm high with a crown of elon- gate leaves at the top of the specimen. At the base dense, helically arranged, rounded stigmarian rooting scars are present, a feature diagnostic for the is- oetalean clade (Fig. 4). Around twenty stigmarian rootlets, each measuring 3 mm wide and over 7 cm long, are found in attachment to the plant base. Additional isoetalean specimens are known from the Lower Eocene Denni- son Gap locality of the Wasatch Formation near Sweetwater, Wyoming, where they occur with platycaryoid foliage, fructifications and Azolla (Steven R. Manchester, Florida Museum of Natural History, written communication, 1992; Fig. 2, 3). These specimens are probably assignable to either I. horridus or I. serratus of Brown, depending on details of the leaf margin that are not readily visible, and show the characteristic air channels that are often found in Isoetites leaves (Fig. 3; Brown 1939, 1958, 1962). 110 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) CONCLUSIONS AND SUGGESTIONS FOR FURTHER RESEARCH In recent years, valuable new types of information have appeared that prom- ise to help clarify long-held questions about the evolution of the isoetalean clade. Notable are newly discovered 1) Middle Devonian forms from China that may have some bearing on the origins of this order, 2) permineralized Triassic fossils that provide a missing “middle” to our knowledge of anatom- ical features between Paleozoic forms and extant Isoetes, and 3) Cretaceous and Tertiary forms that document additional diversity of the group. Phyloge- netic analyses of anatomically preserved Carboniferous forms have been suc- cessful in delimiting interrelationships among the most complex isoetaleans which have been documented as “whole plants” and are understood ecologi- cally (e.g., Bateman et al., 1992; Phillips and DiMichele, 1992). However, to date it has not been possible to extend these types of studies beyond that stratigraphic range because of the limitations of preservation and difficulty of homologies. Whereas researchers studying Devonian, Pennsylvanian or Triassic aged fos- sils or extant Isoetes have made partial comparisons, homologies of characters through geological time have been more difficult. Part of the problem has been the use of terminology developed for one group of lycopsids for a different group. In some cases, terms and concepts have had to be “stretched” to ac- commodate real differences between isoetalean lycopsids of different ages. What is needed is an analysis of overall diversity within given characters, such as sporophyll morphology, without terminology imposed a priori. Lastly, with all of the work done on fossil and extant isoetaleans to date, the question of how to define the modern genus Isoetes in relation to its fossil relatives has never been resolved. In the present contribution I have referred to the extant Isoetes plant habit as having a reduced, non-elongating stem. However, there is considerable morphological variation in extant species. In addition to “typical” plants there are forms with branching axes such as I. andicola (formerly Stylites Amstutz), and the unusual rhizomatous mat-form- ing I. tegetiformans Rury (Fig. 5). Some species of Isoetes have even been pg to have small rhizomorphic structures (W. Carl Taylor, pers. comm. 1999). However, despite this diversity, it is well known among neontologists that morphology has been of little value in understanding speciation in Isoetes because of the high degree of homoplasy and convergence to habitat (e.g., Taylor and Hickey, 1992). Despite numerous discussions, it is still unclear what is basal in the group based on morphology because of the strong imprint of ecological influence. Molecular studies that clarify extant species relation- ships may help us resolve this problem and offer clues to the relationships of the extant and relatively recent fossil isoetaleans. Further developmental stud- ies of Isoetes may also aid in a better understanding of homologies. Neverthe- less, the fossil record still promises to resolve much more about diversity and evolution within this interesting group throughout its long evolutionary his- tory. PIGG: ISOETALEAN LYCOPSIDS a1 ACCKNOWLEDGEMENTS I thank W. Carl Taylor, N. Wickstrum, and Lea Grauvogel-Stamm for inviting my ee to this symposium and the American Fern Society for travel funds to the symposium; Char M. Christy, Walter L. Cressler, Melanie L. DeVore, Else Marie Friis, Patricia G. Gensel, judith - Gordon, David M. Jarzen, Steven R. Manchester and Ruth A. Stockey for access to material and field localities and assistance; Melanie L. DeVore, Patricia G. Gensel, Lea Grauvogel-Stamm, R. — Hickey, Gar W. Rothwell and Wesley C. Wehr for their comments on the manuscript; Ste- 9980388 REFERENCES ANDREWS, H. N., C. B. READ, and S. H. MAMAy. 1971. A Devonian lycopod stem with well-pre- served cas tissues. Palaeontology 14:1-9. aaa ns A. 1933. A lyc aaa strobilus from the Pocono Sandstone of Pennsylvania. r. J. bot. 20:114-11 pay . A. 1935. Notes on some American species of Lepidostrobus. Amer. J. Bot. 22: 23-25. ARNOLD, C. A. eis Observations of fossil plants from the Devonian of eastern North America. ant remains from the Catskill Delta deposits of northern Pennsylvania and southern ae York. Contrib. Mus. Paleontol., Univ. Mich. 5:271-314 ASH, S. R., and K. B. PicG. 1991. A new Jurassic Isoetites [feockales) from the Wallowa Terrane in Hells Canyon, Oregon and Idaho. Amer. J. Bot. 78:1636—1642. BARALE, G. 1999, Sur la présence d’une nouvelle espéce d’Isoetites dans la flore du Crétacé infér- ieur de la région Bote (Sud tunisien): implications paléoclimatique et phylogéné- tiques. Can. fe = 196. BATEMAN, RM gv roan’ reconstruction, pperiiberait and phylogeny of Oxroadia gracilis oa panes and O. conferta. sp. nov.: Anatomically preserved egal ly- copsids from the Dinantian of Oxroad Bay, SE Scotland. eee 228B:29— BATEMAN, R. M., W. A. DIMICHELLE, and D. A. WILLARD. 1992. Experimental cladistic oer of anatomically preserved arborescent lycopsids from the Carboniferous of Euramerica: an essay on paleobotanical phylogenetics. Ann. Missouri Bot. Gard. 79:500—559. BEcK, C. B. 1958. “Levicaulis arranensis”’, gen. et sp. nov., a anaes axis from the Lower Car- Bock, W. 196 tudy on fossil Isoetes. J. Paleontol. 36:53— Brack, S. D. 1970. On a new structurally preserved acicxtecont lycopsid fructification from the Lower Pennsylvanian of North America. Amer. ot. 57:317-330. BRACK, S. D., and T. N. eae 1972. The alteeicoctare and organization of Endosporites. Mi- mae on 18:101—10 BRACK-HANES, S. D., and B. A. te 1983. A re-examination of Lepidostrobus Brongniart. Bot. J. Linn. Soc. 86:125-133. BROWN, r hy 1939. Some American fossil plants belonging to the Isoetales. J. Wash. Acad. Sci. —269. BROWN, - om 1958. New occurrences of the fossil quillworts called Isoetites. J. Wash. Acad. Sci. 48:358—-361. BROWN, R. W. 1962. Paleocene flora of the Rocky Mountains and Great Plains. Prof. Pap. U. S. Geol. Surv. 375:1-119. Cal, C, and X. Wu, 1994. First discovery of Cyclostigma- ‘gine stem ent — pit from the Upper Devonian of Chaeohu i Anhui. Acta Palaeontol. Sin. 33:81— Cai, C., sees L. CHEN. 1996. On a Chinese Givetian Icopod, Longostachys ar ee ng and Feng, emend.: = morphology, anatomy an CANTRILL, D. J., and J. A. Wess. 1998. Permineralized pleuromeid lycopsid remains from the Early Triassic Arcadia Formation, Queensland, Australia. Rev. Palaeobot. Palynol. 103:189-211. 112 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) CHALONER, W. G. 1958. Polysporia mirabilis Newberry, a fossil lycopod cone. J. Paleontol. 32:199- 209. CHALONER, W. G. 1967. ery Pp. 437-845, in: E. Boureau, ed. Traité de Paléobotanique. I. Masson et Cie, Pari CHALONER, W. G. 1968. The cone ir ae ir ap kiltorkense Haughton, from the Upper Devonian of Ireland. J. Linn. Soc., Bot. CHITALEY, S, and K. B. Picc. 1996. Sah ade as ohioensis, gen. et sp. nov., y abeuey iain lycopsid from the Late Devonian Cleveland Shale of Ohio. Amer. J. Bot CICHAN, M. A., and C. B. BECK. 1987. A woody lycopsid stem from the New pea aa ‘ixaat Mi ssp 4 Kentucky. Amer. J. Bot. 74:1750-1757 CRESSLER, W. 1 ead Late ee flora ceeerved in a fluvial setting, north-central sue. U.S.A. Amer. J. Bot. 8 CRESSLER, W. L. 1999. Site- saahee and ris of the Late Devonian Red Hill locality, Pennsyl- vania, an Archaeopteris-dominated plant community and early ye te site. Unpublished Ph.D. dissertation, University of an. Philadelphia, PA, 1 DILCHER, D. L., A. F. ARIM-DE-LACERDA, A. M. F. BARRETO, and M. E fe BERNARDES-DE-OLIV- EIRA. ome bagigrs fossils from the Santana Formation, Chapade do Araripe, Brazil. Amer. J. Bot. DIMICHELE, “ a “1980. Paralycopodites Morey & Morey, from the Carboniferous of Euramerica. A reassessment of the generic affinities and evolution of “Lepidodendron” brevifolium Wil- liamson. Amer. J. Bot. 67:1466—1 ston DIMICHELE, W. A., and R. M. BATEMAN. 1992. Diaphorodendraceae, fam. nov. (Lycopsida: Carbon- iferous): systematics and Lonel relationships of Diaphorodendron and Synchysiden- ron, gen. nov. Amer. J. Bot. 79:605-617 DIMICHELE, W. A., and R. M. BATEMAN. 1996. The bara eye lycopsids: a case-study in paleo- bietaniesl classification. Syst. Bot. 21:535-5 DIMICHELE, W. A., J. F. MAHAFFY, and T. L. ee 1979. Lycopods of Pennsylvanian age coals: Polysporia. Can, J. Bot. 57:1740-1753. DOoBRUSKINA, I. A. 1985. Some problems of the systematics of the Triassic lepidophytes. Palaeon- tological Journal 19:74—88 DRABEK, K. 1976a. Polpsporic mirabilis ronal 1873 from the Nyfany locality (Westphalian D). Cas. Nar. Mus., Odd. Prir. 145:93-9 DRABEK, K. eating Polysporia robust de n. from the Carboniferous of Central Bohemia. Cas. Nar. Mus., Odd. Prir. 145:204— Fucus, G., L. GRAuVvOG uot, and D. MADER. 1991. Une remarquable flore a Pleuromeia et Anomopteris in situ du Buntsandstein moyen (Trias inférieur) de l’Eifel (R. is Allemagne). pre ae pe dap caisson et paléogéographie. rao sam aoc 222B:89-120. GASTALDO, R. - An ultrastructural and taxonomic study of Valvisporites nik (Zerndt) Bhardwaj, a yop ‘eimai from the mide Pennsylvanian of southern I]linois. Micro- paleontology 27 GENSEL, P. G. and C. hr sca 2001. Early lycophyte evolution. Amer. Fern J. 91:99— Girrorb, E. M., and A. S. FosTER. 1989. Morphology and evolution of vascular plants. aa edition. W. H. Freeman & Co., New York. emer hens cs ace L. 1993. Pleuromeia sternbergii (Miinster) Corda from the Lower Triassic of any t observations and comparative morphology of its rooting organ. Rev. Pa- nee Palynol. 77:185—212. GRAUVOGEL-STAMM, L., and P. DuRINGER. 1983. Annalepis zeilleri FLICHE 1910 emend., un organe reproducteur de lycophyte de la Lettenkohle de |’Est de la France. Morphologie, spores in et paléoécologie. Geologische Rundshau 72: be GRAUVOGEL-STAMM, L, and J. LANGIAUX. 1995. Polysporia doubingeri n. sp., un nouvel organe reproducteur de lycophyte du Stérhanion (Carbon eg she ane de Blanzy-Montceau (Mas- sif Central, France). Sci. Géol. Bull. Strasbourg 48:6 rine TAMM, L. and B. LuGaRDON. 2001. The waite cops Pleuromeia and Annalepis: relationships, evolution, and origin. Amer. Fern J. 91:115-14' PIGG: ISOETALEAN LYCOPSIDS 113 GRIERSON, J. D., and H. P. BANKs. mes esac of the Devonian of New York state. Palaeonto- graphica Araceae 4 (31):220- Hanes, S. D. 5. Structurally staan lepidodendracean cones from the Pennsylvanian of Nort Asin Unpublished PhD hic’ Ohio University, Athens, Ohio HorrMaN, G. L. 1995. Paleobotany and paleoecology of the Joffre Bridge readout locality (Paleo- cene), Red Deer, Alberta. Unpublished Master’s thesis, University of Alberta, Edmonton, Can- JENNINGS, Re 1975: sion ae a new plant organ from the Lower Mississippian of Virginia. pe pons 18 eh oe Z RE: E aed LT, and G. bs huihipsiniaes 1983. Structure and affinities of Protostig- a eggertiana. Amer. J. Bot. 70:9 ice. T 1913. On ms NyccaoraNg pamela kiltorkense Haughton sp. Sci. Proc. Roy. Dub- lin Soc, 13:500-52 KARRFALT, E. E. 1984. is observations on Nathorstiana (Isoetaceae). Amer. J. Bot. 71:1023— 30 1030. KENRICK, P, and P. R. CRANE. eign lip origin and early diversification of land plants. Smithsonian Institution Piosa, Washing peorcacnnld V. A., and N. M. Sane 1996. Isoetalean Sitio sganea ag a megaspores m the Upper Cretaceous of Mongolia. Rev. Palaeobot. Palynol. 94: siidtoas, R., and H. WEYLAND. 1949. Gilboaphyton und die a cern met Seawiadick 0:129-15 MATHEWS, G. B. 1940. New lepidostrobi from central United States. Bot. Gaz. 102:26—49. MarTEN, L. C. 1989. A petrified lycopod from the Uppermost Devonian of Hook Head, County Wexford, Ireland. Bot. Gaz. 150:323-336. MEYER-BERTHAUD, B. 1984. Les axes de lycophytes a structure anatomique conservée du carboni- fere basal (Tournaisien) de la Montagne Noire: Trabicaulis gen. nov. et Landyrodendron gen. nov. Palaeontographica 190B:1-36. NEUBERG, M. F. 1960. Pleuromeia Corda from the Lower Triassic deposits of the Russian Platform. Trudy Inst. Geol. Nauk. Ser. Stratigr. 43:63-94 [In R an]. NEUBERG, M. F. 1961. New data on the morphology of Pree Corda from the Lower Triassic of the Russian Platform. Dok. Akad. Nauk. SSSR 136:445—448. [In Russian]; 200-203 [English summary]. Puituips, T. L. 1979. Reproduction of heterosporous arborescent lycopods in the Mississippian— Pennsylvanian of Euramerica. Rev. Palaeobot. Palynol. 27:239-289 PHILLIPS, T. L, and W. A. DIMICHELE. 1992. Comparative ecology and life-history biology of “en ey lycopsids in Late Carboniferous swamps of Euramerica. Ann. Missouri Bot. Gard. 0-588. PicG, 4 B. 1983. es oe and reproductive biology of the sigillarian cone Mazocarpon. t. Gaz. 144: 600-61 PiGG, K . 1992. Evolution 2 isoetalean lycopsids. Ann. Missouri Bot. Gard. 79:589-612. Picc, K. B., and G. W. ROTHWELL. 1979. Stem-root transition of an Upper Pennsylvanian woody lycopsid. Amer. J. Bot. 66:914—924. Pic, K. B., and G. W. ROTHWELL. 1983a. Chaloneria, gen. nov.: heterosporous lycophytes from the Pennsylvanian of North America. Bot. Gaz. 14 —147. Pica, K. B., and G. W. ROTHWELL. 1983b. Mesapaietaphyto development in the Chaloneriaceae fam. nov., permineralized ae Iseotales (Lycopsida). Bot. Gaz. 144:295-302. Picc, K. B., and G. W. ROTHWELL. 1985. Cortical development in Chaloneria cormosa (Isoetales), and the biological derivation of compressed lycophyte decortication taxa. Palaeontology 28: 545-553. RETALLACK, G. 1975. The life and times of a Triassic lycopod. Alcheringa 1:3-29 RETALLACK, G. J. 1997. Earliest Triassic origin of Isoetes and quillwort evolutionary radiation. J. Paleontol. 71: icatiser ROTHWELL, G. W. . The — of Stigmaria (Lycopsida), rooting organ of Lepidodendrales. Aine J. Bot. is pirohe 114 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) ROTHWELL, G. W., and D. M. ERwIN. 1985. The rhizomorph apex of Same implications for homologies among the rooting organs of Lycopsida. Amer. J. B 6-98. ROTHWELL, G. W., and J. S. PRvyor. 1991. Developmental dynamics cs jee nbneaete lycophytes— apical and lateral growth in Stigmaria ficoides. Amer. J. Bot. 78:1 Roy, . A., and L. C. MaTTEN. 1989. Lycopods from the New Albany oe Palaeontographica 12B:1—45. Pa aig S. E., W. L. CRESSLER, T. CONNERY, S. KLAVINS, and D. L. POSTNIKOF. 1999. Devonian shrub and tree dominated landscapes. XVI International Botanical Congress, Abstracts: 13. SCHOPrF, J. M. 1941. Contributions to Pennsylvanian paleobotany: Mazocarpon oedipternum sp. nov. and sigillarian pies oh Illinois Geological Survey, Report of Investigations 75:1-40 ae H. J. 1969. Die Oberdevon-Flora der Bareninsel. 2. Lycopodiinae. Pelvouiaabsod ees SkoG, J. E., and C. R. HILL. 1992. The Mesozoic herbaceous lycopsids. Ann. Missouri Bot. Gard. 79:648-675 merninas N. S. 1980a. Takhtajanodoxa Snig., a new link in evolution of lycopsids. Pp. 4 5 «oy oy me Systematika i evolyutsiya vysshikh rastenii, Nauka, Leningrad e Hus sian]. sais Ug on N. S. 1980b. A new fossil genus of Isoetopsida in the early Triassic of East Siberia. Zarn. SSSR 65:95—96 [In Russian]. ena N. S., and I. N. SREBRODELSKAYA. 1986. Petrified stems rf pe (Lycopo- diophyta) with a preserved anatomical structure. Bot. Z*urn. 71:4 STEWART, W. N. 1947. Comparative study of stigmarian ahah ah Pe see roots. Amer. J. STEWART, W.N., and G. W. ROTHWELL. 1993. Paleobotany and the evolution of plants. 2nd edition, Cambridge University Press, Cambridge. TayYLor, W. C., and R. J. Hickey. 1992. Habitat, evolution, and speciation in Isoétes. Ann. Missouri. Bot. Gard. 79:613-622. THOMAS, B. A. 1981. Structural adaptations shown by the Lepidocarpaceae. Rev. Palaeobot. Pa- lynol. 32:377-388. WaGneR, R. H. 1989. A late Stephanian forest swamp with ee ee cogs ae by volcanic ash fall in the Puertollano Basin, central Spain. Int. J. Coal Geol. 1 WANG, Z. 1991. Advances on the Permo-Triassic lycopods in North ae : ae rae from the mid-Triassic in northern Shaanxi Province. Palaeontographica 222B:1 White, D. 1907. A remarkable fossil tree trunk from the middle Devonic of Now York. Bull. New Wnire, M. E. 1981. Cylomeia undulata (Burges) gen. et comb. nov., a lycopod of the Early Triassic strata of ai South Wales. Rec. Austral. Mus. 33:723-734. American Fern Journal 91(3):115-149 (2001) The Triassic Lycopsids Pleuromeia and Annalepis: Relationships, Evolution, and Origin LEA GRAUVOGEL-STAMM EOST-Géologie, Université Louis Pasteur, 1 rue Blessig, 67084 Strasbourg, and ISEM, UMR 5554 CNRS, Montpellier, France BERNARD LUGARDON Université Paul Sabatier, Laboratoire fae Biologie Sa 39 Allées Jules Guesde, 31000 oulouse, Franc ABSTRACT.—Two kinds of isoetalean lycopsids widely prevailed in the Triassic, the Pleuromeia- type and the Annalepis-type, the latter including a plexus of closely related genera. Comparative terconnected and closely related to Isoetes. Morever they suggest that Annalepis is p n- tral to Isoetes, via Isoetites. Besides several of the morphological and ultrastructural features of the Triassic lycopsids and Isoetes also appear to be present in some of the most anci ycop- sids, suggesting that the lineage including the modern Isoetales has a very remote origin. The lycopsids have the longest evolutionary history of any vascular land plants, covering about 400 million years and spanning every geological period from the Siluro-Devonian to the present. Although their structural diversity has decreased considerably since their peak in the Carboniferous, they were still widespread in the Triassic of both hemispheres. However the Mesozoic lycopsids were not arborescent like many of the Carboniferous ones but rather consisted of slender, herbaceous or pseudoherbaceous plants having an un- branched habit. Two kinds of lycopsids widely prevailed in the Triassic, the Pleuromeia-type and the Annalepis-type, the latter including a plexus of close- ly related genera, which are all regarded as belonging to the isoetalean lycop- sids (Pigg, 1992). Pleuromeia Corda was long regarded as one of the intermediates between the Carboniferous arborescent lepidodendrids and living Isoetes and thus thought to be part of a reduction series (Solms-Laubach, 1899; Potonié, 1904; Magdefrau, 1931). This concept was questionned by Jennings (1975) who no- ticed that the lycopsids with a cormose rhizomorph already existed in the pper Devonian, long before the lepidodendrids with a dichotomously branched Stigmaria rooting system. However DiMichele and Bateman (1996) also suggested, on cladistic foundations, that the isoetaleans evolved from the lepidodendraleans and that the bilateral condition emerged from the radial one in the Devonian. Nevertheless Bateman (1994, 1996) recognized that the precise relationships between both groups are ambiguous The comparative study of Pleuromeia and Annalepis presented here, using new macromorphological and ultrastructural data, shows that these Triassic isoetalean lycopsids appear closely related to Isoetes and that Annalepis seems to be closer than Pleuromeia to the living genus. Moreover this study consti- 116 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) tutes a new basis for comparison with other lycopsids and for evaluating and possibly clarifying the relationship between the isoetaleans and the lepidod- endrids. THE TRIASSIC LYCOPSID PLEUROMEIA: COMPARATIVE STUDY Although the lycopsid Pleuromeia Corda was long regarded as being phy- logenetically highly significant, it was never estimated at its true value because many characteristics of this widespread Triassic genus were not known with enough accuracy. The progress in knowlege of its rhizomorphs, growth habit, reproductive organs and spores now allows to propose a comprehensive com- parative analysis of this outstanding genus. THE RHIZOMORPH OF PLEUROMEIA—As in the other rhizomorphic lycopsids (Bateman 1996), the rooting organ of Pleuromeia has proven to be, phyloge- netically, highly informative since it shows many similarities with that of Is- oetes, suggesting that these lycopsids are closely related. The reinvestigation and morphological analysis of several fairly well-preserved rhizomorphs of P. sternbergii (Miinster) Corda in light of new data on Isoetes permit us to dem- onstrate the structural and developmental correspondence between both. Thus most of the features of the rhizomorph of P. sternbergii can be interpreted in terms of those of the living genus (Grauvogel-Stamm, 1993). Like the rhizomorph of Isoetes, that of Pleuromeia is lobed and shows a bilaterally symmetrical furrow system, when seen from below (Figs.1a-f). The four-lobed rhizomorph of P. sternbergii has a central furrow bifurcating distally into two peripheral furrows which run along the midline of the lobes and extend upward into their extremities (Fig. 1c). Its roots have the same structure as those of Isoetes and Stigmaria. As in Isoetes but unlike Stigmaria, they are arranged on both sides of the furrows according to two intersecting row sys- tems, one roughly parallel to the furrows (series) and the other diverging from them at a relatively high angle (orthostichies) (Fig. 1f). In the two-lobed rhi- zomorph of P. epicharis Wang and Wang (1990) from the Lower Triassic of China, the roots are also arranged in series and orthostichies on both sides of the straight furrow, as in the two-lobed rhizomorph of Isoetes (Figs.1a,b,e). This bilateral arrangement suggests that the roots were produced as in the living genus. That is, they are produced by a basal linear root-producing mer- istem underlying the lobed stele and running parallel to the furrow, and the roots emerged at the furrows and were progressively displaced by continued meristematic activity. The lobing of the rhizomorph developed in relation to this meristem and this process of root production. However the difference in cortical development in Pleuromeia and Isoetes resulted in a difference in lobe and furrow arrangement (Figs.1a-d). Indeed in Isoetes where the cortex is strongly developed, the peripheral furrows of a four-lobed rhizomorph seem to run between the lobes (Fig. 1d). In fact, the correct developmental corre- spondence requires comparing the lobes of Pleuromeia with those of the root- bearing stele in Isoetes, without its cortical lobes. Observations on the diameter GRAUVOGEL-STAMM & LUGARDON: TRIASSIC LYCOPSIDS 117 --O- - -- Oo rath ‘ = es ’ - * On. . . . a Org ’ Fic. 1. Comparative morphology of the rhizomorphs of Pleuromeia and Isoetes. a—Side-view (above) and underside view (below) of a two-lobed rhizomorph of Pleuromeia, such as those of P. epicharis (Wang and Wang, 1990) or P. sanxiaensis (Meng, 1995) from the Triassic of China. b— Cross section of a two-lobed rhizomorph of Isoetes at the level of the basal meristem and furrow (inferred from Karrfalt and Eggert, 1977). This rhizomorph has a straight furrow like the two-lobed rhizomorph of Pleuromeia in Fig. 1a, but its two lobes (stippled) are more developed due to the great cortex production. c—Side-view (above) and underside view (below) of a four-lobed rhizo- morph of Pleuromeia, like that of P. sternbergii (Grauvogel-Stamm, 1993) or P. rossica (Neuburg, 1960; Dobruskina, 1982). The bilaterally-symmetrical furrow system consists of a central furrow lobes. d—Cross section of a four-lobed rhizomorph of Isoetes at the level of the meristem and furrow system (inferred from Karrfalt and Eggert, 1977). The farrow-system is at to that of Pleuromeia in Fig. 1c, but the high cortex production resulted in the development of cortical lobes (stippled) between those of the root-bearing stele. Thus the correct developmental correspondence requires comparing the lobes of Pleuromeia with those of the root-bearing stele in Isoetes (cortical rotrusions excluded). e—Root arrangement on both sides of the straight furrow in a two-lobed high angle (orthostichies) (modified from Karrfalt and Eggert, 1978). f—Root arrangement in series and orthostichies on both sides of the bifurcated furrow system in the four-lobed rhizomorphs of Pleuromeia or Isoetes (modified from Karrfalt and Eggert, 1978). of the root scars of P. sternbergii in relation to their place on the rhizomorph demonstrate that root production proceeded in an acropetal direction as in Isoetes. The significant increase of the root scar diameter from youngest to oldest (also observed in P. epicharis, unpublished observations, L. G.-S.), their numerical increase with the aging and thickening of the rhizomorph, and the maintenance of their arrangement give evidence for long retention of the roots, 118 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) possibly their persistence throughout the life of the plant as in Selaginella. This is unlike Isoetes in which the roots are regularly lost and replaced. The sequence of root initiation, the orientation of the root traces inside the lobes and thus the direction of their emergence and the marked lengthening of the peripheral furrows which occurred during ontogeny show how the rhizo- morph of Pleuromeia developed laterally to the detriment of its downward growth. This new restatement of the structure of the rhizomorph of P. stern- bergii and the comparative study with the living Isoetes accurately emphasizes the numerous points of structural and developmental correspondence between these plants. Fossil lycopsids with a cormose rhizomorph that is furrowed, lobed and bilaterally symmetrical are rather rare in the literature. Besides P. sternbergii, such a rhizomorph has been described in P. rossica Neuburg from the Lower Triassic of the Russian platform (Neuburg, 1960; Dobruskina, 1982) and P. ep- icharis from the Lower Triassic of China (Wand and Wang, 1990). The rhizo- morphs of P. sternbergii and P. rossica are four-lobed and quite similar (Fig. 1c). In contrast, the rhizomorph of P. epicharis which is two-lobed, shows a straight, unbranched furrow (Fig. 1a). Also P. sanxiaensis Meng (1995) from the early Middle Triassic of South China has a two-lobed rhizomorph. Besides Pleuromeia, three other lycopsids with lobed, furrowed, and bilaterally sym- metrical rhizomorphs have been described: Protostigmaria eggertiana Jennings from the Lower Carboniferous of Virginia, USA (Jennings, 1975; Jennings et al., 1983), Nathostianella Glaessner and Rao (1955) from the Lower Cretaceous of South Australia, and Nathorstiana Magdefrau from the Lower Cretaceous of Germany (Karrfalt, 1984). Unlike P. sternbergii, they have no prominent lobes. In the lobed rooting organ of Cylomeia White (1981) from the early Triassic of Australia, no furrows have ever been mentioned and the other data are too imprecise to interpret this feature. In Chaloneria cormosa Pigg and Rothwell from the Pennsylvanian of North America, which is said to have a rounded, slightly lobed, and bilaterally symmetrical cormose rooting organ (Pigg and Rothwell, 1979; 1983; Pigg, 1992), the furrow system has not been mentioned and the root arrangement has not been described. The lobe and furrow system of Chaloneria was not fully documented because of the incomplete preserva- tion of cortical tissue on plant bases of available specimens (Pigg, personnal communication, 1999). Among the cormose lycopsids, there are others in which the rhizomorph is radially or nearly radially symmetrical, such as Pau- rodendron fraiponti Fry from the Upper Pennsylvanian of Ohio, USA (Roth- well and Erwin, 1985) and Oxroadia gracilis Alvin from the Lower Carbonif- erous of Scotland (Stewart and Rothwell, 1993). Moreover, several other cor- mose lycopsids clearly lack an extensive branched rooting system but are not well enough preserved to permit a precise study of their rhizomorph. These are Lepidosigillaria whitei Krausel and Weyland, Cyclostigma kiltorkense Haughton, possibly Eospermatopteris erianus (Dawson) Goldring from the De- vonian (see Pigg, 1992, 2001), Bodeodendron Wagner/Sporangiostrobus feist- mantelii (Feitmantel) Némejc from the Late Stephanian of Spain (Wagner, 1989) and Clevelandodendron ohioensis Chitaley and Pigg (1996) from the GRAUVOGEL-STAMM & LUGARDON: TRIASSIC LYCOPSIDS 119 Late Devonian of Ohio, USA. In all of the cormose lycopsids cited above, the structure of the rhizomorph, its development, and its root arrangement are not known with as much precision as in species of Pleuromeia. The bilaterally symmetrical cormose rhizomorphs of Pleuromeia and Isoetes seem to differ greatly from the radially symmetrical stigmarian rooting organs of the lepidodendrids which are dichotomously branched and in which the roots are helically arranged. However, Rothwell and Erwin (1985) demonstrat- ed that the radially and bilaterally symmetrical rhizomorphs are homologous and that they are merely growth variations rather than indicators of two major lines within the rhizomorphic lycopsids as suggested earlier by Jennings (1975) and Jennings et al. (1983). These homologies were the basis for includ- ing both the Isoetales and Lepidodendrales of the traditional classifications in the rhizomorphic or isoetalean clade (Rothwell and Erwin, 1985; Stewart and Rothwell, 1993; DiMichele and Bateman, 1996). GROWTH HABIT AND REPRODUCTIVE ORGANS OF PLEUROMEIA.—The similarity in rhizomorphic structure which suggests a close relationship between Pleuro- meia and Isoetes strongly contrasts with the differences in growth habit and reproductive structure of these plants. In Pleuromeia (Fig. 2c-g) the sterile leaves differ greatly from the fertile ones which form a well defined cone at the apex of the stem whereas in Isoetes (Fig. 2h, 7d), the fertile and sterile leaves are alike and do not form a cone (Jermy, 1990). However such differ- ences seem to be frequent in related lycopsids. The genera Polysporia New- berry and Chaloneria which belong to the Chaloneriaceae and produce mor- phologically similar microspores and megaspores, are characterized by well- defined terminal cones in Polysporia (Grauvogel-Stamm and Langiaux, 1995) and by fertile zones in Chaloneria in which the sporophylls resemble the ster- ile leaves (Pigg and Rothwell, 1983). Likewise, in the extant genera Phyllog- lossum and Huperzia of the Lycopodiaceae which produce comparable spores, there is a compact strobilus in the first genus while the second genus has either fertile zones or compact strobili (Ollgaard, 1990). All the Pleuromeia species, which have been discussed by Dobruskina (1985) and Pigg (1992), consist of unbranched plants having basal cormose rhizomorphs and terminal, well defined cones. The sizes of the plants are rather variable but many species are relatively small. Even among specimens of P. sternbergii from the type-locality (Bernburg, Germany) for which Mag- defrau (1931) indicated a height of 2—2.5 m, there are fertile specimens which only reach 23 cm (Fig. 2d) (Grauvogel-Stamm, 1999). Those from the Eifel in Germany (Fig. 2c) are 1-1.5 m tall (Mader, 1990; Fuchs et al., 1991). Pleuro- meia rossica from the Lower Triassic of Russia was no more than 1 m high (Neuburg, 1960) and P. jiaochengensis Wang and Wang (1982) from the Lower Triassic of China usually reached 20-30 cm high, sometimes 50 cm (Fig. 2e). Recently, several new species have been described from the early Middle Tri- assic of southern China (Meng, 1995,1996) which consist of plants mostly 50 cm high. However, P. marginulata Meng (Fig. 2f) includes specimens more AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) Sy = r hig Sf eS5t ET ete ed aS itty! seecitainel a = ove © ® Fic. 2. Comparative growth habit of the Paleozoic and Triassic Pleuromeia-like lycopsids and the extant Isoetes. (scale-bar = 10 cm, except in Fig. 1g,h = 1 cm). a—Clevelandodendron ohioensis; Devonian of Ohio, USA (redrawn from Chitaley and Pigg, 1996). b—Chaloneria cormosa; Penn- ,1983). c—Pleuromeia sternbergii; Lower Triassic (Olenekian) of the Eifel, Germany (from Fuchs et al., 1991). d—P. sternbergii from the Lower Triassic of Bernburg, the type-locality, (reconstructed after the illustrations of Bischof, 1853, Figs.1,2). e—P. jiaochengensis; Lower Triassic (Induan) of Shanxi, China (redrawn from structed after the illustrations of Meng, 1995: Pl.1 fig. and Meng, 1996: Pl.1 fi iaensis; early Middle Triassic (Anisian) of South China (reconstructed after the illustrations of Meng, 1995, 1996, Pl.1 Fig. 9). h—Isoetes brochoni, showing leaves 3-7 cm long (redrawn from Motelay, 1893). than 1 m high. In contrast, P. sanxiensis Meng is only 5 cm tall, with a 2,5 cm long cone (Fig. 2g). The trunk of Pleuromeia is usually devoid of leaves but is covered with helically arranged leaf scars. However some of the specimens of P. sternbergii from Eifel had trunks covered with leaves (Fig. 2c). These leaves have a strong midvein and, according to Magdefrau (1931), they were succulent and contain transverse ridges extending perpendicularly on both sides of the midvein. As suggested by Grauvogel-Stamm (1999, Fig. 4), these transverse ridges might correspond to air chambers, as in Isoetes and Isoetites phyllophilla Skog, Dilcher and Potter (1992). However this feature requires further investigation. Cone size is variable in Pleuromeia. According to Magdefrau (1931), the cone of P. sternbergii reached 20 cm long. However, other cones of that species GRAUVOGEL-STAMM & LUGARDON: TRIASSIC LYCOPSIDS 121 are only 5 cm long (Bischof, 1853, Figs.1,2; Grauvogel-Stamm, 1999). Likewise according to Magdefrau (1931), the cones were monosporangiate, containing either microspores or megaspores. Nevertheless, the small cone figured by Bis- chof (1853) is clearly bisporangiate (Grauvogel-Stamm and Lugardon, in prep- aration). Its microsporophylls and megasporophylls are inter-mixed as in P. rossica (Neuburg, 1960). The cones of P. sternbergii appear to be mono- or bisporangiate depending perhaps on the size and the age of the plants. The two cones on which Magdefrau (1931) relied were rather large whereas that of Bischof (1853) is small. In Isoetes too, the size of the plants has an influence on the production of microsporophylls and/or megasporophylls (Williams, 1943). The sporophylls of the different Pleuromeia species show a remarkable sim- ilarity in shape and structure since they consist of a circular to oval lamina and a large, globose sporangium which covers nearly all of the adaxial surface (Figs.3a-j). Distally, the margin has a slight notch or a tiny pointed tip at the apex. The ligule is attached immediately distal to the sporangia and is often obvious in the apical border of the sporophylls (Figs.3a,c,d,h-j). Trabeculae similar to those of Isoetes have also been observed in the sporangia of several species. According to Snigirevskaya (1989), the sporophylls of Pleuromeia are similar to Isoetes leaves having lost their distal sterile portion. Indeed, as sug- gested by this author, they had a tip which was longer than usually depicted but which dried and decayed in maturing cones. However, examination by one of us (L.G.-S.) of sporophylls of P. sternbergii and P. rossica in side view shows that they have a tip extending from the lamina without direction change and that this tip is a little longer than the border of the sporophylls in face view (Grauvogel-Stamm, 1999), but never as long as suggested by Snigirevskaja (1989). Lycopsids with the same growth habit as Pleuromeia already existed in the Paleozoic: Clevelandodendron ohioensis Chitaley and Pigg (1996) from the Late Devonian of Ohio, USA and Chaloneria cormosa Pigg and Rothwell (1983) from the Pennsylvanian of North America. Clevelandodendron ohioen- sis consists of 1.25 m tall unbranched plants covered with helically arranged leaf scars, and possessing a well defined bisporangiate cone containing trilete microspores and megaspores (Fig. 2a). Its partially preserved rhizomorph is said to bear “thick appendages”. Chaloneria cormosa is also unbranched, and is 2 m tall with a bilaterally symmetrical cormose rhizomorph (Fig. 2b). Its reproductive organs consist of a terminal fertile zone up to 21 cm long, with alternating megasporangial and microsporangial regions resembling the sterile nes. Meyen (1987) has suggested that Pleuromeia evolved from Chaloneria, and Stewart and Rothwell (1993) state: ‘““Chaloneria seems to bridge the gap be- tween the Upper Devonian treelike lycopsids with their cormose rhizomorphs and the cormose Triassic Pleuromeia’’. The discovery of Clevelandodendron ohioensis may indicate that a lineage comprising Pleuromeia-like plants al- ready existed in the Late Devonian and was contemporaneous with the begin- ning of the radiation of arborescent lepidodendrids. This suggests that the di- 122 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) WE ¢ WOO OC) ® Fic. 3. Comparative shape of the sporophylls in Pleuromeia. Note the globose sporangium oc- cupying most of the upper surface; note also the ligular pit (Ip) distal to the sporangium, often visible in the narrow apical border of the sporophyll. (scale-bar, between h and i = 1cm). a—P. sternbergii; Lower Triassic of Germany (redrawn from Solms-Laubach, 1899; Potonié, 1904). b— Paltinis; Lower Triassic of North China (redrawn from Wang and Wang, 1989). c—P. sternbergii; Lower Triassic of North China (redrawn from Wang and Wang, 1989). d—P. rossica; Lower Triassic of Russia (redrawn from Neuburg, 1960; Dobruskina, 1982). e—P. epicharis; Lower Triassic of North China (redrawn from Wang and Wang, 1989, 1990). f—P. patriformis; Lower Triassic of North China (redrawn from Wang and Wang, 1989). g—P. jiaochengensis; Lower Triassic of North China (redrawn from Wang and Wang, 1982). h—P. sanxiaensis; early Middle Triassic of South China (redrawn from Meng, 1995, Fig. 3b,c). i—P. marginulata; early Middle Triassic of South China (redrawn from Meng, 1995, Fig. 3a). )j—P. hunanensis; early Middle Triassic of South China (re- drawn from Meng, 1995, Fig. 3d). versification within the Isoetales was well established prior to the Carbonif- erous (Chitaley and Pigg, 1996). It may also question the statement of Di- Michele and Bateman (1996) which suggests that the isoetalean line evolved from the lepidodendrids. THE SPORES OF PLEUROMEIA.—Like the structure of the rhizomorphs, the ultra- structure of the spores of Pleuromeia appears to be phylogenetically infor- mative, suggesting a close affinity with Isoetes. The microspores of both genera, especially, show many ultrastructural sim- ilarities (Lugardon et al., 1997, 1999). This is all the more noteworthy since Isoetes microspores have a rather large number of distinctive characteristics which clearly differentiate them from all the other extant spore types, as stressed below. In order to underscore the similarities in ultrastructure and GRAUVOGEL-STAMM & LUGARDON: TRIASSIC LYCOPSIDS 123 correspondences of the homologous walls of Pleuromeia and Isoetes micro- spores, a description of both is given hereafter (Fig. 4). Isoetes microspores are monolete spores easily recognizable, through light microscope, by an apparently bipartite sporoderm consisting of an ovoid, un- adorned inner part wrapped in a wide, thicker, usually slightly ornamented outer part (Fig. 4a). The inner part is generally regarded as the exospore in morphological studies, while the outer part is variously interpreted and termed (perine, sexine. . .). TEM studies (Lugardon, 1973, 1980) have shown that these microspores possess three distinct aceto-resistant walls (Fig. 4b). The inner- most one, which really represents the exospore, is a thin, almost plain-sur- faced, bilayered wall that is characterized mainly by its aperture and by special areas of the wall called “laminated zones’ ( = “zones pluristrates” in the initial description of Lugardon, 1973). The aperture is distinguished by a sim- ple reduction of the exospore thickness, without any other special modifica- tion. The laminated zones are small, appreciably thickened areas of the exo- spore arranged on both sides of the aperture, in which the exospore inner layer is tangentially cleft into several irregularly segmented laminae. The middle wall, that is called “para-exospore”’, consists of interconnected elements com- posing a spongy layer which is detached from the exospore all around the spores except in some points of the lateral regions of the proximal face; more- over this wall forms a high projection, usually divided apically into two lips, above the aperture. The outermost wall, which is termed perispore, is a rather thick, complex wall closely applied to the para-exospore surface. Ontogenet- ical fine studies (Lugardon, 1990; Tryon and Lugardon, 1991) showed that the para-exospore elements develop at the same time as the exospore outer layer, through simultaneous accumulation of the same sporopollenin on the exo- spore inner layer and on the tenuous framework of the para-exospore that is formed very early in the course of sporogenesis. Moreover, these studies showed that the perispore develops after the para-exospore and exospore com- pletion, with materials of different chemical nature, like the typical perispore of homoporous fern spores (it is generally admitted that a genuine perispore is formed only in pteridophytes having a plasmodial tapetum; nevertheless, TEM studies proved that, besides the microspores of recent Isoetaceae, those of a number of Selaginellaceae, as well as the spores of rather numerous Ly- copodiaceae, have a perispore comparable to that of the fern spores, although the lycopsids possess a secretory tapetum; Lugardon, 1990). Thus the wide sporoderm outer part of Isoetes microspores comprises two quite distinct com- ponents, i.e an uncommon structure, the para-exospore, that is ontogenetically related to the exospore, and a veritable perispore. The microspores of Pleuromeia are trilete and rounded or rounded-trian- gular in equatorial outline (Fig. 4c). The morphological studies achieved with light and scanning electron microscopes (Neuburg, 1960; Yaroshenko, 1975) showed that these microspores have an outer envelope largely separated from a much thinner inner part often called inner body. The outer envelope appears spongy and faintly ornamented. The inner body is roughly smooth-surfaced and reveals under the light microscope, especially in specimens having lost 124 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) . 4, Comparative ultrastructure of the apertural area in the microspores of Isoetes and Pleu- romeia (scale-bar = 10m in a, c and 2 pm in b, d). a—Isoetes—Diagrammatic representation of the microspore proximal face showing the outer envelope (OE) and the inner body (IB). The dotted spots on both sides of the monolete aperture mark the location of the laminated zones observed with TEM, but indiscernible in intact spores observed with light microscope. b—Isoetes. Ultra- structural features of the apertural area of a microspore sectionned along x-y, as shown in Fig. 4a. Note that there are three distinct resistant walls: the exospore (E) with the aperture (a) consisting of a thinner, non-folded wall portion, bordered with laminated zones (z); the para-exospore (PE) broken above the apical para-exospore lips. c—Pleuromeia—Diagrammatic representation of the microspore proximal face showing the outer envelope (OE) and the inner body (IB) with three interradial papillae (pap) arranged between the arms of the trilete aperture and visible with light microscope only in particular conditions of preservation. d—Pleuromeia—Ultrastruct features microspore apertural area sectionned along x-y, as shown in Fig. 4c. Two distinct walls are preserved: the exospore (E) with the aperture (a) and laminated (z) similar to those in Isoetes; e para (PE) which is thicker but structurally analogous with that of Isoetes microspores. in? th tuo GRAUVOGEL-STAMM & LUGARDON: TRIASSIC LYCOPSIDS 125 the outer envelope, three rounded papillae arranged between the rays of the trilete mark near the proximal pole (Fig. 4c). Ultrastructural investigations in Pleuromeia rossica (Lugardon et al., 1997, 1999) and P. sternbergii (Grauvogel- Stamm and Lugardon, in prep.) proved that the inner body is a two-layered exospore, uniformly solid and thick except in the proximal region where it shows two notable characteristics (Fig. 4d). In the aperture area indeed, the exospore simply appears thinner without other noticeable differentiation, as in Isoetes. Secondly, in each of the three areas lying between the aperture rays, the exospore shows a well-defined laminated zone which is quite similar to those of Isoetes microspores, and which obviously corresponds to one of the three interradial papillae observed with light microscope. The outer envelope wholly consists of interconnected elements which show the same electron per- meability and reactions to chemical reagents as the outer layer of the exospore, and are surely made up of the same sporopollenin. It is attached to the exo- spore in the peripheral regions of the proximal face. In the other spore regions, this outer envelope is usually separate from the exospore, and forms a well- marked protrusion divided apically into two lips above each of the three ap- erture rays. Thus, the ultrastructural features of the exospore are very similar in Pleu- romeia and Isoetes microspores, in spite of the fact that those of Pleuromeia are trilete whereas those of Isoetes are monolete. Likewise, the main features of the outer envelope of Pleuromeia microspores remarkably resemble those of the para-exospore in Isoetes, so that both walls obviously appear homolo- gous, and can be equally called para-exospore. Besides the trilete or monolete condition, the only noteworthy differences lie in the lesser development of the para-exospore and the presence of an extra wall, the perispore, in Isoetes. The combination of the exospore and para-exospore features which is quite identical in Isoetes and Pleuromeia microspores is also that of the dispersed spore genera Endosporites and Aratrisporites which are respectively produced by Chaloneria/Polysporia and Annalepis (Lugardon et al. 2000b, Grauvogel- Stamm and Lugardon, in prep.). The three interradial papillae have been clear- ly shown in Endosporites by Brack and Taylor (1972). This combination of features differentiates these microspores from all the other spore types of ex- tant, and probably fossil, pteridophytes. The aperture consisting of a simply thinner, unfolded exospore area appears unique, and contrasts with the well- marked prominent fold showing various outlines and more or less intricate structures that characterizes the aperture of most pteridophyte spores, includ- ing the microspores of Selaginellaceae (Lugardon, 1972, 1980, 1986; Tryon and Lugardon, 1991). Moreover, although the laminated zones are present in the microspores of many Selaginellaceae, they are totally missing in a rather im- portant number of species of this large family (Tryon and Lugardon, 1991: Figs. 231.21, 25, 26,28). Likewise, the para-exospore appears to exist only in very few species of Selaginellaceae in most of which the exospore is devoid of laminated zones, and the para-exospore is almost solid and does not form a sharp protrusion above the aperture (Tryon and Lugardon, 1991: Figs. 231. 26, 28). Among the recent and fossil lycopsids in which the spore ultrastructure 126 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) is known to date, the microspores which most resemble those of Pleuromeia and Isoetes are those of the present Selaginella selaginoides (Lugardon, 1972) and those from the Pennsylvanian lycophyte cone Selaginellites crassicinctus (Taylor and Taylor, 1990). In both, indeed, the microspores have an exospore with laminated zones and a fairly spongy para-exospore forming prominent lips above the aperture rays. However, these spores are unambiguously distin- guishable by the specially narrow, strongly marked apertural fold of the exo- spore (known only in Selaginellaceae microspores among recent pteridophyte spores, as it can be seen in Tryon and Lugardon, 1991). Thus, not only the microspores of Isoetes and Pleuromeia, as well as those of Annalepis (Aratris- porites) and Chaloneria/Polysporia (Endosporites), show very close ultrastruc- tural similarities, but also they differ markedly from the spores of all the other pteridophyte groups, including the especially variously structured micro- spores of Selaginellaceae. It is regrettable that there are no adequate data on the fine features of the microspores of other fossil rhizomorphic lycopsids, particularly the lepidod- endrids, which could be compared to those of Pleuromeia. The different TEM studies on the lepidodendrid microspores Lycospora, achieved by Thomas (1988), Taylor (1990), Hemsley and Scott (1991), Scott and Hemsley (1993), suggest that these microspores also have a thin exospore partly separated from a thicker outer envelope presumably equivalent to a para-exospore. However the Lycospora studied by these authors, like those presently investigated (un- published observations, B.L.), are rather poorly preserved and have not clearly shown the ultrastructural characteristics of their apertural area which, very likely, would provide significant information on the relationships of the lepi- dodendrids with the other lycopsid lineages. The megaspores of Pleuromeia are also trilete with a slight ornamentation and a spongy aspect rather comparable to that of the microspores (Zhelezkova, 1985; Marcinkiewicz and Zhelezkova, 1992). Ultrastructurally (Lugardon et al., 2000a; Grauvogel-Stamm and Lugardon, in prep.), their exospore consits of a very thin, solid inner layer and a thick spongy outer layer which constitutes most of the wall and is usually divided by a tangential gap in its innermost area, except near the aperture. These features, as well as those of the aperture, are comparable to those of the Isoetes megaspores (Lugardon, 1986; Taylor, 1993 ; Tryon and Lugardon, 1991). However they are also analogous to those of all the other, recent or fossil, lycopsid megaspores, so that they do not pro- vide any special phylogenetic information. Nevertheless, the ultrastructural study of especially well-preserved megaspores of P. rossica (Lugardon et al., 2000a) demonstrates that the spongy outer layer consists of elements which usually show a rod-like shape, are mostly arranged parallel to the wall surface and are rather widely spaced in moderately crushed spores. These features of the exospore outer layer, which are quite obvious and unvarying throughout the wall in those P. rossica megaspores, precisely represent the few character- istics regarded as distinctive of the isoetalean megaspores (Kovach, 1994). Thus the megaspores indicate also, although in a less demonstrative way than GRAUVOGEL-STAMM & LUGARDON: TRIASSIC LYCOPSIDS 127 the microspores, the relation joining the Triassic Pleuromeia to the living Is- oetes. Furthermore, the exospore inner layer of Pleuromeia megaspores includes, along both sides of the three apertural arms, a number of laminated zones which are structurally quite similar to those of the microspores. Such lami- nated zones, that have not been observed in Isoetes megaspores, undoubtedly correspond to the “papillae, pustules, cushions, nipple-like projections. . .” described in several morphological studies of fossil lycopsid megaspores. In particular, rather abundant papillae, i.e. 15-20 on either side of each apertural ray, can be observed with LM in the megaspores of P. sternbergii (Grauvogel- Stamm and Lugardon, in prep.). However, for the moment these structures do not provide any reliable information on the affinities of Pleuromeia since their distribution among the lycopsids is not clearly characterized. THE TRIASSIC LYCOPSID ANNALEPIS AND RELATED TAXA: COMPARATIVE STUDY. Besides Pleuromeia, there is another noteworthy Triassic lycopsid, Anna- lepis Fliche, which has long been misunderstood. The type-species, A. zeilleri Fliche from the Middle and Upper Triassic of eastern France, is represented by isolated “scales” described by Fliche (1910) who noticed their resemblance to seed scales of the conifer Araucaria but finally placed them as an incertae sedis. Grauvogel-Stamm and Duringer (1983) showed that the scales of A. zeil- leri are sporophylls belonging to the lycopsids and containing microspores assignable to Aratrisporites. Such sporophylls are often termed “scales” in descriptions, likely on the basis of their shape and maybe on the analogy of the wording used in the conifer seed scales. Undescribed new material from the Ladinian (late Middle Triassic) of Germany now allows to give some further details, particularly about the rest of the plant (Grauvogel-Stamm and Lugar- don, in preparation). GROWTH HABIT AND REPRODUCTIVE ORGANS OF ANNALEPIS ZEILLERI FLICHE.—The sporophylls of A. zeilleri are preserved as compressions having 2.5—5 cm long and being slightly trapezoidal in shape (Fig. 5a, 6a). They consist of a long and wide proximal portion flaring abruptly at the base which is somewhat thickened, widening distally and extending by a thick and short, roughly tri- angular distal portion ending with a small pointed tip. The presence of a trans- verse wide groove followed by a bulge, distal to the sporangium, between the proximal and distal portions in the sporophylls in compression and adaxial view, indicates that their junction was more or less at right-angle. Indeed, the groove corresponds to the zone of juncture of the two portions and represents exactly the place where the distal portion was straightened up (Fig. 5d-e). This right-angled juncture remained partly preserved and appears as a groove be- cause of the thickness of this zone while the thinner distal portion has been folded back by the sediment pressure and appears in the same plane as the proximal portion in the fossilized sporophylls. Sometimes the groove+ bulge appears as a transversal fold (Fig. 5f). Thus, when the sporophylls were packed 128 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) Fic. 5. The sporophyll of Annalepis zeilleri Fliche in adaxial view (a-c): (a) in compression, entire; (b) in compression but lacking the lateral wings; (c) in life condition (modified from Grau- the course of fossilization: (d) sporophyll in life condition, with both portions roughly perpen- dicular to each other (the arrow indicates the direction of flattening during fossilization); (e) flat- tened sporophyll with a bulge distal to the transverse groove; (f) flattened sporophyll showing a small fold above the groove. (pp = proximal portion; dp = distal portion, upturned in life position as in c and d; g = transverse groove; b = transverse bulge; f = transverse fold; j = thickened junction zone; sp = sporangium; w = wing; lp = ligular pit; pl = rod-like placenta, located on the abaxial side but often visible on the adaxial side of the sporophylls in compression, especially when the sporangium is empty. GRAUVOGEL-STAMM & LUGARDON: TRIASSIC LYCOPSIDS 129 Ch) SS : Mt Ot bia Oe N op Gx: @ A © Fic. 6. Comparative morphology of the sporophylls of the Annalepis-type and related genera. (Scale-bar = 1 cm). a—Annalepis zeilleri Fliche (from Grauvogel-Stamm and Duringer, 1983). The arrow shows a scale having lost its lateral wings. b—A. Jatiloba Meng (redrawn from Meng, 1998). c—A. angusta Meng (redrawn from Meng, 1995). d—A. sangzhiensis Meng (redrawn from Meng, 1995). e—A. brevicystis Meng (redrawn from Meng, 1998). f—Annalepis sp. (unpublished). Note the range of variation in shape, size and apex length. The arrows indicate two scales having lost their lateral wings. g—Isoetes ermayinensis Wang (redrawn from Wang, 1991). h—Tomiostrobus radiatus Neuburg (redrawn from Dobruskina, 1985). i—Tomiostrobus belozerovii, T. fusiformis, T. bulbosus, T. convexus, T. gorskyi and T. migayi, from left to right (redrawn from Sadovnikov, 1982). j—Lepacyclotes ellipticus Emmons (redrawn from Fontaine, 1900). k—Cylostrobus sydneyensis Helby and Martin (redrawn from Helby and Martin, 1965). |—Skilliostrobus australis Ash (redrawn from Ash 1979). m—Isoetites daharensis Barale (redrawn from Barale, 1999). n—Isoetites horridus Brown from the Tertiary of North America (redrawn from Hickey, 1977). o—Isoetites choffatii (Saporta) (redrawn from Teixeira, 1948) . 130 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) together in the cones, their proximal portion was perpendicular to the cone axis whereas their distal part was upturned and more or less parallel to this axis (Fig. 5c). The adaxial surface of the proximal portion was occupied, in the middle and on the whole length, by a long, oval and well-delimited spo- rangium. The borders of the proximal portion which project on both sides of the sporangium appear to be membranous, as also noticed by Fliche (1910). The boundary between the median part bearing the sporangium and its mem- branous borders is not always very clear but it is particularly obvious when the imprints are covered with coaly organic matter which is thinner on the borders and therefore less dark than on the other parts of the scales. These membranous borders, called wings or alae according to the authors, likely tended to fall or to be damaged. In some specimens they are partly detached and in others they are completely missing (Fig. 5b), so that it appears clearly that the proximal part is widened distally and that the wings are restricted to the area below this distal widening (see Grauvogel-Stamm and Duringer, 1983; compare figs. 2 and 3 of Pl. 4). The long oval sporangium seems divided by a longitudinal midline which likely corresponds, as in Isoetes (Hall, 1971), to the trace of a rod-like placenta which was located on the abaxial side of the proximal portion of the sporophyll and to which the sporangium wall was attached. The presence of a ligule, which was not recognized when Grauvogel- Stamm and Duringer (1983) described these scales, is clear but this ligule is usually decayed and represented only by its attachment point, which is lo- cated immediately distal to the sporangium. Undescribed new material from the Ladinian (late Middle Triassic) of Ger- many shows that the distal portion of these sporophylls was fleshy. The coal- ified organic matter covering their imprint is thick and the cell imprints are usually quite visible on its brilliant surface. These sporophylls were arranged in cones reaching at least 8-10 cm in diameter. The occurrence, together with these scales and cones, of several rounded furrowed rhizomorphs, probably 4- lobed, each still attached to a stem (10 cm long preserved) which is covered with oval leaf-scars (the leaves are unknown) is informative since it shows that this plant had the same growth habit as Pleuromeia and consisted of an un- branched monopodial axis with a bilaterally symmetrical rooting base and a large, terminal and compact cone (Fig. 7a). The impressive and numerous roots, 3-7 mm wide and more than 30 cm long, which are still attached to the rhizomorphs, may indicate that this plant was robust and strongly anchored. As shown by Grauvogel-Stamm and Duringer (1983), the sporophylls of the cones were monosporangiate, containing either monolete microspores (Ara- trisporites Leschik) or trilete megaspores (Tenellisporites marcinkiewiczae Reinhardt and Fricke). No specimens offer evidence as to whether the cones were bisporangiate or monosporangiate. The ultrastructural features of the monolete microspores are comparable to those of Pleuromeia and Isoetes (Grauvogel-Stamm and Lugardon, in preparation). THE OTHER ANNALEPIS SPECIES.—Five further species of Annalepis have been described from the Anisian (early Middle Triassic) of South China: A. furo- GRAUVOGEL-STAMM & LUGARDON: TRIASSIC LYCOPSIDS 131 \\ wittiga/// \\" iy / Fic. Comparative growth habit of Annalepié, ienetiiee and Isoetes. (Scale-bar = 5 cm). a— peel cies zeilleri Fliche. Note the apical d the basal rounded rhizomorph bearing numerous long roots, more than 30 cm long, arranged in two sets on both sides of the furrow. The leaves are unknown but their scars cover helically the stem. Ladinian of Germany (unpublished). b—lIsoetites daharensis Barale. Note the basal rounded rhizomorph and the stocky stem covered ly were packed in a cone as suggested by the specimens figured by Barale, which are at the origin of this modified reconstruction. Lower Cretaceous of Tunisia. (modified from Barale, 1999). c— et al., 1992). d—Isoetes boryana, having 20-30 lacunate leaves 10—20 cm long (redrawn from Motelay and Vendryés, 1882 ngqiaoensis Meng, A. latiloba Meng (Fig. 6b), A. angusta Meng (Fig. 6c), A. sangzhiensis Meng (Fig. 6d) and A. brevicystis Meng (Fig. 6e). The last three species are shown to have a long, narrow leaf tip emerging abruply from a rounded apex (Meng, 1995, 1996, 1998, 2000). The sporophylls representing A. brevicystis are by far bigger than the others and greatly differ in shape. Meng (1998) moreover described several further sporophylls with a long leaf tip con- taining transverse partitions which he assigned to A. zeilleri but which prob- ably correspond to another species. Indeed none of the specimens correspond- ing to the original material of A. zeilleri described from the Ladinian (late Middle Triassic) of France show such features (Grauvogel-Stamm and Durin- ger, 1983). In all the sporophylls described by Meng, the sporangium is tongue- shaped and is longitudinally divided by a midline. Unfortunately none of the photographic illustrations of the Chinese specimens clearly shows the struc- ture of these sporophylls, such as it is shown in the rough reconstructions figured by Meng (1995, 1998, 2000). Furthermore, an assemblage of numerous undescribed sporophylls from the Ladinian (late Middle Triassic) of Germany which is here provisionally as- +32 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) signed to Annalepis sp. (Fig. 6f), represents a further new species which is still under study (Grauvogel-Stamm and Kelber, in preparation). Indeed, these sporophylls which originally have been attributed to A. zeilleri by Kelber and Hansch (1995) contain features which distinguish them from this species. They show a great variability in shape, size and apex length, which probably de- pends on their place within the cone. Transverse partitions have never been observed in any of the long apex of these sporophylls. However, they seem to have lateral membranous borders as in A. zeilleri (Fig. 8a,b) and the tongue- shaped sporangium shows a longitudinal midline as in the other Annalepis- like sporophylls. In spite of a roughly similar organization, the sporophylls of Pleuromeia and Annalepis differ in several respects. Their general outline and the shape of their sporangium are quite different. The wings bordering the sporophyll prox- imal part and the longitudinal midline running along the sporangium in An- nalepis are missing in Pleuromeia. In contrast to the sporophylls of Annalepis that are clearly differentiated into a proximal and a distal portions which were very likely more or less perpendicular to each other in lifetime, those of Pleu- romeia do not show such a morphological partition and are approximately on the same plane from the base to the apex. Moreover, in Pleuromeia the micro- spores and megaspores are trilete whereas in Annalepis only the megaspores are trilete while the microspores are monolete (Grauvogel-Stamm and Durin- ger, 1983), as those of Isoetes. THE ALLIED GENERA: FROM ANNALEPIS TO ISOETITES.—While it is easy to recognize a sporophyll as belonging to the genus Pleuromeia, there are difficulties in identifying those of the Annalepis-type. Most of the latter have been long mis- understood, being first regarded as araucarian seed scales and therefore often assigned to the genera Araucarites Pres] or Pseudoaraucarites Vladimirovitch (Dobruskina, 1985). Now, eventhough they are regarded as lycopsids, these sporophylls are assigned to different genera as Tomiostrobus Neuburg, Skil- liostrobus Ash, Cylostrobus Helby and Martin and Lepacyclotes Emmons, and classified, together with Annalepis, as satellite-taxa in the Pleuromeiaceae (Thomas and Brack-Hanes, 1984). However their taxonomic position is greatly in flux. Dobruskina (1985) stressed the similarity between Annalepis and To- miostrobus whereas Retallack (1997) synonymized Annalepis with Lepacyclo- tes. Moreover Sadovnikov (1982) and Retallack (1997) synonymized Skillios- trobus with Tomiostrobus. Nevertheless, since the genus Annalepis, particu- larly the type-species A. zeilleri, is the best known, it is here regarded as the type of this kind of sporophyll. Lepacyclotes was described earlier by Emmons (1856) but the precise structure and the in situ spores have never been studied. Other taxa for which the structure and the shape of the sporophylls are un- known include Lycostrobus scotti Nathorst and Austrostrobus ornatum Mor- belli and Petriella. As for Lycostrobus chinleana Daugherty from the Triassic of southwestern USA, it has proven to be an equisetalean reproductive organ (Grauvogel-Stamm and Ash, 1999) and not a lycopsid, as previously believed, or an arborescent quillwort, as recently asserted by Retallack (1997). GRAUVOGEL-STAMM & LUGARDON: TRIASSIC LYCOPSIDS 133 f a s @)— \ o “-\-. Oe @ wie Fic. 8. Comparative morphology and terminology of the — of Annalepis, Lepidoden- drids and Isoetes in adaxial views. (scale-bar = 5mm, except in d,e = m). a—Annalepis zeilleri Fliche (pp, proximal portion; dp, distal portion; lp, lieular pit; sp, ee w, wing). Note the clearly delimited wings (stippled) of the proximal portion (modified from Grauvogel-Stamm and Duringer, 1983). b, c—Annalepis sp. (undescribed): two sporophylls within the range of variation. The wings (stippled) are clearly delimited in b, while they are not in c. d—Isoetes coromandelina Lin. F ils: a scale-leaf with an aborted sporangium (asp) (redrawn from Srivastava and Wagai, 1996, pasa from Bochenski, 1936). Sab Sanger ps jenneyi (White) Abbott. Sporophyll showing a proximal pedicel below the sporangium and an apically directed lamina (la). Due to its strong distal widening, the pedicel has been ea T-shaped by Abbott (modified from Abbott, 1963). i—Lepidocarpopsis lanceolatus (Lindley and Hutton) Abbott. The pedicel clearly bears well-de- stippled), sporangium (sp), ligula (li), and the terminal subula (s) 3,4 cm long (modified from Kubitzki and Borchert, 1964; terminology after Hickey 1985) Since the precise structure and shape of the sporophylls are essential for identifying and comparing these lycopsids, only the genera for which these features are known with considerable accuracy and which resemble the An- nalepis-type sporophylls are listed in Table 1 and shown in Fig. 6, Some Is- oetites are also listed but their inclusion raises some questions on the exact interpretation of this fossil genus, as commented below. 134 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) TABLE 1. The crap -like sporophylls. Data taken from the following sources: 'Fliche (1910); *Meng (1995); 3Meng (1998); Meng (2000); ‘Dobruskina (1985); “Sadovnikov (1982); 7Wang (1991); ‘Emmons (1856); °Ash (1979); tletby and Martin (1965); ''Teixeira (1948); '?Barale (1999); Brown (1958), Hickey (1977) Taxon Age Occurrence Figure si nrtcande (type-species) Middle Triassic NE France; Germany Fig. 6a A. ine ensis* Middle Triassic Southern China A. lati Middle Triassic Southern China Fig. 6b A. angusta* Middle Triassic Southern China Fig. 6c A. sangzhiensis* Middle Triassic Southern China Fig. 6d A. brevicystis? Middle Triassic Southern China Fig. 6e i vat ohana Middle Triassic Germany Fig. 6f oetes erma Middle Triassic North China Fig. 6g Tmicarrabee le (type-species) Lower Triassic South Siberia, Russia Fig. 6h . belozerovii® ower Triassic East Siberia, Russia Fig. 61 T. fusiformi Lower Triassic East Siberia, Russia Fig. 61 bulbosus® Lower Triassic East Siberia, Russia Fig. 61 T. gorskyii® Lower Triassic East Siberia, Russia Fig. 61 T. pees Lower Triassic East Siberia, Russia Fig. 61 T. convexus® Lower Triassic East Siberia, Russia Fig. 61 Lepeveoe Sera (type-species), Upper Triassic North Carolina, USA Fig. 6j yn. L. elliptic ae ior fer neyensis"® (type-species) Lower Triassic Australia Fig. 6k Skilliostrobus emeieg (type-species) ower Triassic Australia Fig. 61 Isoetites daharens Lower Cretaceous Southern Tunisia Fig. 6m Isoetites cevichas™ Tertiary North America Fig. 6n Isoetites choffati'' Lower Cretaceous Portugal Fig. 60 The sporophylls characterizing these taxa share a remarkable similarity with Annalepis. They have a comparable shape and structure (Fig. 6a-o). Most of them (Annalepis, Tomiostrobus, Isoetes ermayinensis, Skilliostrobus, Cylostro- bus) contain microspores assignable to Aratrisporites (Balme, 1995). In all, the sporangium is elongate-oval and shows a longitudinal midline which has been variously interpreted. In Annalepis zeilleri, Grauvogel-Stamm and Duringer (1983) suggested that it might represent the trace of the vascular bundle (Fig. 6a). In Isoetites daharensis (Fig. 6m), this line, which is obvious too and quite resembles that of Annalepis, has been interpreted as being the internal bound- ary of the velum (Barale, 1999). In Isoetes ermayinensis (Fig. 6g) it has been regarded as a strong keel or a vascular bundle (Wang, 1991). In Tomiostrobus (Figs.6h,i), Sadovnikov (1982) interpreted it as a keel whereas Dobruskina (1985) regarded it as the vascular bundle. Fontaine (1900) also regarded it as a keel in Lepacyclotes ellipticus (Fig. 6j). The fact that this line appears in variably the same in the different scales indicates that it is a eicatiette feature of this kind of sporophyll and that it has the same significance in all. A comparison with Isoetes shows that it does not represent the velum which is a thin membrane covering partially or entirely the sporangium and which never has a longitudinal slit-like free edge in the middle. In contrast, the lon- gitudinal midline more likely indicates that, as in Isoetes (Hall, 1971), a rod- GRAUVOGEL-STAMM & LUGARDON: TRIASSIC LYCOPSIDS 135 like placenta including the vascular bundle existed on the abaxial side of these sporophylls, to which the sporangium wall was attached, and that it became apparent on the adaxial side of the compressions particularly when the spo- rangium was empty. According to J. Hickey (pers. com), in Isoetes this placenta is represented by a pad of vascularized tissue which often intrudes into the sporangium and runs lengthwise. Hall (1971) moreover noticed that in dried specimens of I. tenuifolia and I. pitotii, the rod-like shaped placenta shrinks and leaves an especially deep and profound groove. In some specimens of Tomiostrobus (Figs. 6h,i) and Skilliostrobus (Fig. 61) which one of us (L. G.-S.) observed, and according to the illustrations of Isoetes ermayinensis (Fig. 6g) and Cylostrobus sydneyensis (Fig. 6k), most of the scales seem to have a proximal portion with lateral wings, as in A. zeilleri (Fig. 6a). These lateral wings are clearly shown in the transverse section of the sporo- phyll of Tomiostrobus figured by Sadovnikov (1982, Fig. 2, at right), though not noticed by him. They are comparable to those shown in transverse sections of Isoetes sporophylls (Pitot, 1959; Hall, 1971). In some specimens the wings seem to be missing or partly destroyed, as in Skilliostrobus (see Ash, 1979, Figs. 7 A—D) or I. ermayinensis (see Wang, 1991, Pl. 1 figs. 9-11, 13, 14). How- ever the question of the wings would require further investigations since their presence has not been noticed in most of the sporophylls listed in Table 1. The long, oval and contiguous depressions, often visible on both sides of the sporangium in most of the scales, such as Annalepis sp., A. zeilleri, A. latiloba and Lepacyclotes circularis, correspond to the imprint of the sporangia of the underlying and overlying sporophylls when they were packed together in the cone. The wide transverse groove visible at the junction of the proximal and distal, thick portions of the scales of L. circularis, as seen in adaxial view and as figured by Bock (1969, Figs. 93-94), is quite similar to that observed in A. zeilleri (Grauvogel-Stamm and Duringer, 1983, Pl. 4 figs. 2,3); this strongly suggests that the proximal and distal portions were originally more or less perpendicular to each other in Lepacyclotes too. The sporophylls of Tomios- trobus which one of us (L.G.-S.) observed, also show that the junction of the proximal and distal portions was more or less right-angled. Likewise, in An- nalepis sp., the distal portion also seems to have been upturned. Also Meng (1998) noticed the presence of a transverse ridge in A. latiloba which suggests that the sporophyll distal portion was more or less perpendicular to the prox- imal one. However various interpretations and/or misinterpretations have occurred because these sporophylls often have a more or less blunt, seemingly incom- plete apex and/or because the scale tip has been considered as being the ligule, making the interpretation of this area confusing. Moreover, some of these spo- rophylls have been regarded as having lost their leafy distal portion. The spo- rophylls of Isoetes ermayinensis Wang (Fig. 6g) are so similar to those of A. zeilleri (Fig. 6a) that the inclusion of that taxon in Isoetes is surprising. In fact, Wang (1991) interpreted the numerous, associated elongate leaf imprints as their detached apical parts. These imprints contain transverse partitions that 136 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) he supposed to be collapsed air chambers, similar to those of Isoetes. However no specimens show them in organic connection. A similar leaf imprint found associated with sporophylls of A. brevicystis in the Anisian of South China has also been regarded as their tip (Meng, 1996, Pl. 2 Fig. 14). Sterile parti- tioned leaves have never been found associated with A. zeilleri from the Tri- assic of eastern France (Grauvogel-Stamm and Duringer, 1983). Retallack (1997) noticed that Wang’s sporophylls of I. ermayinensis are “unlikely to have borne the wider undulose leaf fragments on their tips as in Isoetes’” and re- assigned I. ermayinensis to the genus Lepacyclotes. Likewise, the reassignment of Lepacyclotes circularis and L. ellipticus to the genus Isoetites, as proposed by Brown (1958), does not seem to be justified and has been questioned by Chaloner (1967) and Bock (1969) since these scales have a short and well delimited triangular apex and are devoid of any traces of a long sterile leafy portion comparable to that of Isoetes. Skog and Hill (1992) also questioned this reassignment and noticed the resemblance between Lepacyclotes circu- laris and Annalepis. Among the fossils attributed to the genus Isoetites, there are also isolated scales with a small pointed apical tip resembling the scales of Annalepis, which have been supposed to be the fertile base of Isoetes leaves. It is the case, especially, of Isoetites horridus (Fig. 6n) from the Tertiary of North America (Brown 1958, Figs. 1,3; 1962, Pl. 9 Fig. 1) and I. choffati Saporta (Fig. 60) from the Cretaceous of Portugal (Teixeira, 1948, Pl. XXVI figs. 7-8). However the small apical tip is not preserved in some specimens, as those of I. horridus from the early Tertiary of Dakota figured by Hickey (1977, Pl. 1 figs. 6,7) and described as being dumb-bell shaped by Collinson (1991). Moreover, according to the illustration of Collinson (1991, Fig. 7.3h), which shows a circlet of such supposed fertile leaf bases, one can see that Hickey (1977) rep- resented these sporophylls upside down and that the narrower extremity cor- responds to the base while the broader one corresponds to the distal extremity to which the long leafy part was supposed to be attached. It is noteworthy that this circlet closely resembles those consisting of Annalepis sporophylls. Since the leaf tip is often missing in the Isoetites remains, Collinson (1991) suggested that this tip was deciduous in the fossil genus. In Isoetes, the long leafy part is usually not deciduous, except in some terrestrial species where the distal portion of the leaves erodes away leaving only a leaf base, typically with a sporangium (J. Hickey, pers. com.). However, the fossils attributed to Isoetites call for several remarks. An over- view of these fossils shows that the precise shape and structure of the fertile parts are usually poorly described and illustrated and that their relationship with the sterile leaves is not clear. For most taxa, the fertile and sterile parts are not shown in organic connection and several rhizomorphs and stems seem to contain only sterile leaves, e.g., Isoetites phyllophila (Skog et al., 1992), I. horridus (Brown, 1962; Hickey, 1977), I. indicus (Bose and Banerji, 1984), I. serratus (Brown, 1939), I. rolandii (Ash and Pigg, 1991) and Cylomeia (White, 1981) which is now reassigned to Isoetes beestonii (Retallack, 1997). It is sur- prising that none of them shows any trace of a sporangium at the base of the leaves. Likewise, the late Cretaceous and Tertiary specimens assigned to Is- GRAUVOGEL-STAMM & LUGARDON: TRIASSIC LYCOPSIDS 137 oetites by Pigg (2001) do not show clearly if there were sporangia at the base of the leaves. Long sterile leaves, without partitions, still attached to an axis have also been found associated with the sporophylls of Annalepis sp. in the Triassic of Germany and have been assigned to Isoetites sp. (Kelber and Hansch, 1995). In Isoetes, a great number of leaves are usually fertile (Jermy, 1990) and when once the plant becomes fertile, all subsequently produced leaves will either be fertile or at least initiate a sporangium (J. Hickey, pers. com.). During the ontogeny of the Isoetes leaves, the sporangium develops very early (Fig. 8e), is already discernible in the leaf primordia (Bhambie, 1963; Kubitzki and Borchert, 1964), and even the vegetative leaves have an aborted sporangium at the base (Pitot, 1959; Rauh and Falk, 1959; Kubitzki and Borch- ert, 1964). Thus the real nature of many Isoetites leaves remains quite ambig- uous. Moreover, some of the described Isoetites species do not represent ly- copsids. According to Skog and Hill (1992), Isoetites gramineus (Ward) Bock more likely is an osmundaceous fern and J. bulbosus Drinnan and Chambers more probably corresponds to “pentoxylalean fertile short shoots”’. Likewise, the interpretations about the scales of Isoetites having a small tip and resembling the Annalepis sporophylls are still questionable. Do they rep- resent fertile leaf bases which have lost their leafy extremity, and is the small tip visible in some of them actually a ligule? In Isoetes, the ligule is a delicate ephemeral parenchymatous flap which is mostly deciduous and tends to dis- appear at maturity (Pitot, 1959; Bhambie, 1963; Goswami, 1976; Kott and Brit- ton, 1985). Therefore, it seems unlikely that the ligule would persist if the long leafy extremity of the sporophyll was destroyed, as for example in I. horridus where a specimen representing a circlet of sporophylls still attached to an axis has been interpreted by Brown (1962, Pl. 9 Fig. 1,3) as being a “corm with sporangia tipped by ligules’”’. The Isoetites scales with a small apical tip seem more likely to represent entire sporophylls rather than the basal part of fertile leaves, and their small apical tip more likely represents their actual extremity and not the ligule. Moreover, the fact that these Isoetites sporophylls with short pointed or blunt apices are often found associated with long, apparently ster- ile, leaves and are still closely packed (see Brown, 1958 Fig. 3; Hickey, 1977; Collinson, 1991, Fig. 7h; Barale, 1999, Fig. 5)—as in the circlets of sporophylls of I. horridus, A. zeilleri, Annalepis sp. and Lepacyclotes ellipticus - suggests that they represented distinct organs, occupied a place distinct from that with sterile leaves, and became detached or abcised from an axis. Isoetites dahar- ensis (Fig. 7b) is particularly interesting in this respect since it has a rounded rhizomorph with roots and a short stem bearing distal scale-like sporophylls and elongate sterile leaves (Barale, 1999). Such a plant does not conform to Isoetes (Fig. 7d) in which the sterile and fertile leaves are morphologically similar (Jermy, 1990) but more closely resembles Annalepis or Pleuromeia in which the fertile and sterile leaves differ and are located in distinct places on the plant. Isoetites daharensis (Fig. 6m) is all the more interesting because its sporophylls are of the Annalepis-type, as are those of I. horridus (Fig. 6n). However, contrary to the latter, in J. daharensis the leaves are not lacunate. In most of these fossil lycopsids, the length of the corm plus stem is unknown. 138 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) In I. daharensis (Fig. 7b), the rounded corm + stem are 11 cm tall (Barale, 1999) as in Annalepis zeilleri (Fig. 7a). In Isoetites phyllophila (Fig. 7c), the rhizomorph is elongate, conical, and 6 cm high (Skog et al., 1992). However, as noticed by Skog and Hill (1992), there are no broadly ranging studies on the shape of the stems at various stages in development of extant species of Isoetes stripped of their appendages and de-corticated, so it is difficult to com- pare fossil and living plants. In Stylites for example, the stem reaches 20 cm long (Rauh and Falk, 1959). Likewise, there are no broadly ranging studies on the distribution of the fertile and sterile leaves in extant Isoetes. Clearly, most of the Isoetites specimens need further investigation and better illustrations, all the more since this genus is regarded as representing primitive Isoetes and is included in the subgenus Euphyllum, together with three extant species of Isoetes (Hickey, 1986, 1990; Taylor and Hickey, 1992). Nevertheless, this comparative analysis of the sporophylls of Annalepis and related taxa (Tomiostrobus, Lepacyclotes, Skilliostrobus, some Isoetites) dem- onstrates that all these sporophylls have a similar basal unit which is narrow at the base, widens distally and extends by a short, roughly triangular distal portion ending with a more or less long pointed apical tip. In all there is a long oval adaxial sporangium and a ligule immediately distal to it. In most of them moreover, the proximal portion appears to bear lateral wings which are restricted to the area below the widened portion of the leaf. The sporophyll of Isoetites horridus figured by Hickey (1977, Pl. 1 figs. 6,7) is especially in- teresting in this respect because it shows particularly well the distal widening (Fig. 6n), suggesting that initially it also had lateral wings restricted to the proximal part. RESEMBLANCES WITH ISOETES AND LEPIDODENDRID SPOROPHYLLS.—Besides the ul- trastructural similarities of the spores of Annalepis and Isoetes (Lugardon et al. 2000b; Lugardon and Grauvogel-Stamm, in prep.), there are still some other morphological resemblances between both genera which deserve to be noticed. Indeed, some Annalepis sporophylls (Fig. 8c) closely resemble the devel- oping leaf primordia of Isoetes (Fig. 8e), which have a short pointed apex devoid of air chambers and an almost fully developed ligule, proximal to this apex (Kubitzki and Borchert, 1964). Such a resemblance shows how an An- nalepis sporophyll may have developed into an Isoetes sporophyll. It would be instructive to know the shape and growth sequence of the wings in young Isoetes leaves but this development has not been studied (J. Hickey, personal information, September 1999). Likewise it is noteworthy that the Annalepis sporophylls much resemble the scale leaves which occur in many Isoetes species and differ greatly from the typical Isoetes long leafy sporophylls. These scale leaves which are mostly unknown to Isoetes non-specialists are present at the growing apex and at the base of the corm, but are also said to occur below the spirally arranged leafy sporophylls, sometimes intermingled with them, and to form two or three whorls in old corms (Duthie, 1929; Stolze and Hickey, 1983; Hickey, 1986; Srivastava and Wagai, 1996). Thus for example, within the range of variation GRAUVOGEL-STAMM & LUGARDON: TRIASSIC LYCOPSIDS 139 of Annalepis sp. there are sporophylls (Fig. 8c) which are very similar to the scale leaves of Isoetes montezumae A.A. Eaton (Stolze and Hickey, 1983, Fig. 10c) or I. coromandelina Lin. fils (Srivastava and Wagai, 1996, Pl. 4 Fig. 1) (Fig. 8d). Likewise, the sporangium bearing scales with a small pointed apical tip (Fig. 60) which occur in Isoetites choffati (Teixeira, 1948) among the long leafy Isoetes-like sporophylls, and which resemble the sporophylls of Anna- lepis, might also represent such scale leaves. According to Hickey (pers. com.), the scale leaves of Isoetes are environmentally induced and occur in distinct zones (alternating with typical fertile leaves) as a result of seasonality. Also according to Hickey (1986), they have a protective function for the corm and represent leaf primordia in which the development has become arrested. How- ever, their resemblance with Annalepis sporophylls suggests that they also might have a phylogenetical significance and might represent vestigial rem- nants of ancestral sporophylls, indicating that the genus Annalepis is an an- cestor of Isoetes. The sporophylls which have been described recently from the Triassic of South China and which have been assigned to several new Annalepis species, e.g. A. angusta (Fig. 6c), A. sangzhiensis (Fig. 6d), A.brevicystis (Fig. 6e) and A. zeilleri (sensu Meng 1998) are also particularly interesting in the compari- son between Annalepis and Isoetes. Indeed, these sporophylls which are char- acterized by a long, narrow (spur-like) outgrowth, arising distal to the ligular pit, running across the short, wide, triangular to rounded distal portion and extending far beyond it, also closely resemble the scale leaves which occur in Isoetes and which are characterized by a spine-like apex. In I. mahadevensis for instance, the scale apices are said to appear as spur-like outgrowths and to emerge abruptly from an almost rounded region (Srivastava and Wagai, 1996). In Annalepis zeilleri (sensu Meng 1998) moreover, this apical outgrowth is described as containing transverse partitions. Meng (1998) interpreted it as the leaf tip of the sporophylls. According to Hickey (pers. com.), it is comparable to the ‘“‘subula” of the Isoetes leaves, i.e. their distal awl-like shaped outgrowth. Meng (1998, 2000) also noticed the similarity of the Annalepis sporophylls with Isoetes leaves, and suggested that A. brevicystis, especially, might be an- cestral to the living genus. The new Chinese Annalepis species are moreover particularly interesting because they allow to understand some unexplained features of Tomiostrobus radiatus (Fig. 6h). Indeed, the two parallel lines shown to arise distal to the ligule and to extend apparently out from the fragmentary apex of the sporo- phylls in the rough drawings of Dobruskina (1985, Fig. 1b-e, g, i, n), more likely represent partially preserved apical spine-like outgrowths comparable to those described by Meng (1995, 1998, 2000) in several Chinese Annalepis species and compared by Hickey (pers. com.) to the subula of the Isoetes leaves. In a sporophyll of T. radiatus figured by Yaroshenko (1988, Fig. 11), the long apex appears moreover to contain transverse ridges in the middle, though not noticed by the author, which might represent collapsed air cham- bers. However, none of the fragmentary sporophylls attributed by Sadovnikov (1982) to several distinct Tomiostrobus species (Fig. 61) seems to contain such 140 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) partitions. Likewise, in none of the specimens of Tomiostrobus from the Lower Triassic of East and North Siberia (Russia) which one of us (L. G.-S.) could observe, the apex shows such features. Comparable unexplained spur-like out- growths seem to exist in Annalepis sp. and A. zeilleri but they are badly pre- served and have not yet been described (Grauvogel-Stamm and Lugardon, in preparation; Grauvogel-Stamm and Kelber, in preparation). However, in spite of these resemblances there are some differences in the shape of the fertile area and lateral wings between the Annalepis sporophylls and the typical Isoetes long fertile leaves. In Isoetes the proximal portion is wide at the base, narrows distally and extends upwards in a long leafy extrem- ity (Fig. 8j), whereas in Annalepis and allied genera the proximal portion is narrow at the base, widens distally and extends by a short more or less tri- angular portion ending in a small pointed apical tip (Fig. 8a, b). In Annalepis moreover the lateral wings are restricted to the area below the distal widening whereas in Isoetes the winged borders, now commonly termed alae, extend upwards over a variable length (Kott and Britton, 1985). It is worthy of note that in most of the extant Isoetes species, these alae are also restricted to the proximal portion of the leaf (Hickey, 1986; Taylor and Hickey, 1992) but, in contrast to Annalepis, they are not limited by a distal widening of that prox- imal portion. In fact the shape of the proximal part and lateral wings of Annalepis spo- rophylls more resembles that of the basal unit of the sporophylls of the Car- boniferous lepidodendrids Lepidostrobus Brongniart, Lepidostrobopsis Abbott and Lepidocarpopsis Abbott, as already noticed by Grauvogel-Stamm and Dur- inger (1983, Figs. 3a, b). These lepidodendrid sporophylls that have usually a long, wide, upturned leafy distal portion termed lamina in most descriptions, also have a basal unit consisting of a proximal portion bearing an adaxial sporangium and sometimes lateral wings, which is termed pedicel (Fig. 8f-i). It is noteworthy that the proximal portion of these sporophylls widens distally as in Annalepis or Isoetites horridus (Hickey, 1977) and that the wings, when present, are restricted to the area below the distal widening of the proximal portion. The sporophyll of Lepidophyllum jenneyi White (1899, Pl. 59 Fig. 2) which shows very clearly the distally widened proximal portion was regarded as having lost its lateral wings while it was interpreted by Abbott (1963) as having a T-shaped pedicel and was transferred by this author to the genus Lepidostrobus (Fig. 8h). In Lepidocarpopsis lanceolatus (Lindley and Hutton) Abbott on the contrary, the pedicel is clearly alate, exceeding the width of the sporangium (Fig. 8i). The sporophylls of Cyclostigma kiltorkense Haughton (Fig. 8f) from the Upper Devonian of Ireland (Chaloner, 1968) and Lepidostro- bus bohdarowiczii Bochenski (1936) from the Carboniferous of Poland (Fig. 8g) also have such a T-shaped pedicel and a long distal lamina, but their ped- icel seems to be devoid of lateral wings. However, according to Taylor and Hickey (1992), the alation which characterizes the sporophylls of the arbores- cent lepidodendrids and those of the isoetaleans is a plesiomorphy and there- fore is not phylogenetically informative. GRAUVOGEL-STAMM & LUGARDON: TRIASSIC LYCOPSIDS 141 THE TERMINOLOGY IN THE LYCOPSID REPRODUCTIVE ORGANS Comparisons of the Annalepis sporophylls with those of the arborescent lycopsids and those of Jsoetes underscore a problem of terminology. Indeed, as there is no standardization for the different groups of lycopsids, various terminologies occurred which make rather complex and confusing the com- parative studies. As regards to the lepidodendrids, the proximal and distal portions of the sporophylls are respectively termed ‘‘pedicel” and “lamina” by most of the authors (Abbott, 1963; Chaloner, 1967; Emberger, 1968; Bateman et al. 1992; Stewart and Rothwell, 1993). Emberger (1968, p.179) noticed that the proximal portion is equivalent to a petiole bearing a sporangium. In con- trast, Taylor and Taylor (1993, p.270) defined the lepidodendrid sporophylls (Lepidocarpon Scott) as consisting of two lateral laminae and a distal exten- sion. The similarities between the Annalepis sporophylls and those of the lep- idodendrids led Grauvogel-Stamm and Duringer (1983) to use for Annalepis the terminology applied to the latter by most authors. However the resem- blances between Annalepis and Isoetes emphasized for the first time in this study, required to change the terminology employed previously, as it has been done above. In contrast to the lepidodendrids, in Isoetes the proximal portion of the sporophyll is termed “vagina” while its distal portion is called “subula”. The current use in the terminology of Isoetes of these Latin words has been intro- duced by Hickey (1985) whereas, previously, they were employed only in the Latin diagnosis (Braun, 1868; Duthie, 1929; Wanntorp, 1970). However the use of these Latin words complements those which were already in use for Isoetes, such as “ala”, “ligula”, “‘velum” and “labium’”. As for the lateral wings (or alae) which are present in the proximal portion of the Isoetes leaves, they are regarded as the restricted remnants of the lamina (Hickey, pers. com.). In fact the meaning of these Latin words describes the features of the different por- tions of the leafy sporophylls. Thus, the portion containing the alae is called “vagina” which means “sheath”, because the wings are slightly enveloping, as shown for example in the drawings of Wanntorp (1970). On the other hand, as the distal outgrowth of the Isoetes sporophylls is awl-like shaped, it has been called “subula” which means “awl”. The terminology such as applied to Isoetes seems to be applicable to some of the Chinese species of Annalepis (Meng 1995, 1998, 2000), as noticed by Hickey (pers. com. 1999) who recognized a subula in some of them, but it does not seem applicable to all the Annalepis-like sporophylls, particularly those in which there is no subula. However, the leaf primordium of Isoetes also seems to be devoid of a subula. Thus, only the shape and growth sequence of the wings and the subula in young Isoetes leaves would allow to establish the exact homologies with Annalepis but this development has not yet been stud- ied (Hickey, pers. com., September 1999). CONCLUSIONS The comparative analysis of the Triassic lycopsids Pleuromeia and Anna- lepis using new macromorphological and ultrastructural data provides signif- 142 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) icant support for their close relationship with each other and with the living genus Isoetes and more ancient lycopsids, some of which extend back to the Devonian. A close affinity between Pleuromeia and Annalepis appears unquestionable. Preliminary observations of some rather well preserved specimens of Anna- lepis indicate that they have furrowed, bilaterally symmetrical rhizomorphs as in Pleuromeia. The sporophylls, with their adaxial sporangium and distal lig- ule, are comparable in both genera, except that the sporophylls of Annalepis are wedge-shaped and have well-marked membranous wings which are seem- ingly missing in Pleuromeia. Moreover, although they are monolete, the mi- crospores of Annalepis clearly show the same ultrastructural features as the trilete microspores of Pleuromeia, i.e. a thin exospore with a thinner, non- folded apertural area combined with laminated zones, and a thicker spongy outer wall largely dissociated from the exospore and divided into two pro- truding lips above the aperture. Most of the features shared by Pleuromeia and Annalepis also appear ve similar to those of the living Isoetes. The comparative study of the rhizo- morphs of Pleuromeia and Isoetes demonstrates that these organs have many points of structural and developmental correspondence. The ultrastructural features of the exospore and the spongy outer wall in the Pleuromeia and Annalepis microspores are remarkably similar and homologous with the walls of the Isoetes microspores (the latter having in addition a third, superficial wall, the perispore). This similarity between fossil and recent microspores is all the more striking because the fine structure of the Isoetes microspores dif- fers markedly from that of all the other extant pteridophyte spores, including those of Selaginella. Indeed, despite some ultrastructural resemblances reflect- ing the relationships between Isoetes and Selaginella, the microspores of the latter show a number of unambiguous differences which strongly suggest that Isoetes is much more closely related to Pleuromeia and Annalepis than it is to Selaginella. The resemblances in morphology of the sporophylls also de- serve to be noted. Snigirevskaya (1989) stressed the similarity between the sporophylls of Pleuromeia and the basal fertile region of Isoetes leaves. How- ever, the clearly winged sporophylls of Annalepis show a stronger resemblance to Isoetes. In fact, the most important difference lies in the growth habit of the Triassic genera and that of Isoetes. Like some lycopsids from the Carboniferous (Chaloneria cormosa) and the Devonian (Clevelandodendron ohioensis), the Triassic lycopsids Pleuromeia and Annalepis have an elongated stem covered with entirely sterile leaves and a terminal, well-defined cone with short spo- rophylls, whereas the living genus consists of a short corm with long lacunate leaves, most of which bear a basal sporangium. However these differences do not seem to be very significant since they are rather common in the lycopsids where they are regarded as simple variations among related genera. The fact that in Isoetes many or all the leaves are fertile can be explained either by the loss of sterile leaves (Pigg, 2001) or, more likely, by the evolutionary reduction and transfer of the reproductive function to the vegetative leaves, as in the Gymnosperms and the Angiosperms (Nozeran, 1955). GRAUVOGEL-STAMM & LUGARDON: TRIASSIC LYCOPSIDS 143 All the similarities shown by both Triassic genera and Isoetes surely indicate close connections between them. Nevertheless, at first glance, they are not clearly indicative of a sister-group or an ancestor-descendant relationship and do not reveal the exact role played by Pleuromeia and Annalepis in the evo- lutionary history of Isoetes. However, there are some further noteworthy re- semblances between Annalepis and Isoetes which strongly suggest that An- nalepis is closer to the extant genus than Pleuromeia and that it is ancestral to Isoetes. Recently Meng (1998, 2000) also made this assumption. Besides their winged sporophylls/leaves noted above, there is, indeed, a striking sim- ilarity between the Annalepis sporophylls and the leaf primordia and scale leaves of the extant genus. Especially the scale-leaves which occur in many Isoetes species, and appreciably resemble Annalepis sporophylls, might be vestigial structures and might represent remnants of ancestral sporophylls, which might indicate that originally Isoetes had distinct sterile and fertile leaves as in Pleuromeia and Annalepis. The microspores of Annalepis are monolete as in Isoetes, whereas they are trilete in Pleuromeia. Moreover, the similarities between Annalepis and some Cretaceous and Tertiary fossils attri- buted to the genus Isoetites give further support to an ancestor-descendant relationship between Annalepis and Isoetes. The newly described species Is- oetites daharensis from the Lower Cretaceous of Tunisia (Barale 1999), which is the best known among these plants, shows quite clearly that its sporophylls are remarkably similar to those of Annalepis and differ morphologically from the sterile leaves. Furthermore, its rhizomorph is cormose as in the Triassic genus, although not bilaterally symmetrical. As discussed above, Isoetites hor- ridus from the Tertiary of North America (Brown 1939, 1958; Hickey 1977; Collinson 1991) was probably organized similarly, bearing its Annalepis-type sporophylls and sterile lacunate leaves in distinct areas, unlike Isoetes. These similarities strongly suggest that Isoetites and Annalepis are closely related and that Isoetites may have evolved from Annalepis. Thus the Triassic genus Annalepis seems to have played a more direct role in the evolutionary history of Isoetes than Pleuromeia. As regards to Isoetites daharensis and I. horridus, they may represent a group of lycopsids intermediate between Annalepis and Isoetes, and may indicate which structural changes occurred during this evo- lution. Thus the evolutionary series ‘Annalepis—Isoetites—Isoetes’ appears very likely. Unfortunately, most of the Isoetites species are poorly known. They are usually identified on the sole basis of their lacunate leaves whereas the most important features, such as the precise shape and structure of the fertile parts and the microspore ultrastructure, remain unknown. The fine structure of Minerisporites mirabilis, the megaspore of I. horridus (Collinson, 1991), shows rather clearly the features of the isoetalean megaspores, but it does not provide any more precise informations on the affinities of Isoetites. Further and more thorough investigations of all the Isoetites material would be essen- tial for definitely establishing the evolutionary series from the Triassic lycop- sids to Isoetes. Besides, the fact that Annalepis and Pleuromeia are contem- poraneous suggests that an ancestor-descendant relationship cannot exist be- tween them. Therefore it is inferred that Pleuromeia either became extinct, or 144 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) originated within the nonalate species of Isoetes. However the latter hypoth- esis seems unlikely in view of the apparent homogeneity of the living genus which strongly suggests a monophyletic origin for it. Some of the typical features of Pleuromeia and Annalepis are also present in more ancient lycopsids, providing some evidence of their ancestral affini- ties. The bilateral symmetry of the rhizomorph is known from the Devonian (Pigg, 2001). The sporophylls of the Carboniferous lepidodendrids, especially those with lateral wings, and those of more ancient lycopsids such as Cyclos- tigma kiltorkense from the Devonian of Ireland (Chaloner, 1968) are compa- rable in some respects to Annalepis. The special ultrastructural features of the microspores shared by the Triassic Pleuromeia and Annalepis, as well as by the extant Isoetes, also occur in Polysporia (Lugardon et al., 2000), Chaloneria, and other Carboniferous lycopsids. Moreover, it is of note that the trilete mi- crospores of Bisporangiostrobus harrisii Chitaley and McGregor (1988) from the Upper Devonian of Pennsylvania contain proximal papillae as do the mi- crospores of Chaloneria/Polysporia, Annalepis, Pleuromeia and Isoetes. This is all the more interesting since Cyclostigma kiltorkense, which has cones sim- ilar to those of B. harrisii (Stewart and Rothwell, 1993), also has a bilobed, bilaterally symmetrical rhizomorph as in Pleuromeia, Annalepis and Isoetes. Moreover several Frasnian-Famennian (Late Devonian) dispersed spores have ultrastructural characteristics comparable to those of the microspores of the Triassic genera and the extant genus (unpublished observations, B.L.). Thus it appears that the main features characterizing the rhizomorph and the micro- spores of living Isoetes appeared very long ago and remained outstandingly unchanged from the Devonian to the present. It is likely that these features characterize a particular group among the rhizomorphic lycopsids and that this group probably has a very remote origin. However, this group cannot be defined at the present time, notably because there are no accurate data on the microspore ultrastructure of most of the fossil lycopsids, mainly the arbores- cent lepidodendraleans. Methodical fine studies of in situ microspores of these lycopsids would surely supply significant information on the origin, relation- ships and delimitation of the different groups of isoetalean lycopsids. ACKNOWLEDGEMENTS. We thank W. Carl Taylor for inviting one of us (L.G.-S) to coorganize the symposium on “Evo- lution and diversification of the Lycopods” in the XVI International Botanical Congress held in St. Louis, Missouri, 1-7 August 1999, and Jim Hickey, Kathleen Pigg and an anonymous reviewer REFERENCES ABBOTT, M.L. 1963. Lycopod fructifications from the Upper Freeport (no 7) coal in southeastern Ohio. Palaeontographica, Abt.B, 112: 93-118. GRAUVOGEL-STAMM & LUGARDON: TRIASSIC LYCOPSIDS 145 ASH, : R. 1979. peoevirayng gen. nov., a new lycopsid cone from the Early Triassic of Australia. cheringa ASH, [ R., and : - an 1991. A new Jurassic agen nerean from the Wallowa terrane in Hells Canyon, Oregon and Idaho. Amer. J. Bot. 78(12): 1636-1642. BALME, B.E. 1995. Fossil in situ spores and orate grains: an oie catalogue. Rev. Palaeobot. Palynol. 87: 81-323. BARALE, G. 1999. Sur la présence d’une nouvelle espéce d’Isoetites dans la flore du Crétacé infér- ieur de la ar Ge = Tataouine spe Tunisien): implications paléoclimatiques et phylogéné- tiques. Can. J. B : 189- BATEMAN, R M. 1994. ile aidan change in the growth architecture of fossil rhi- zomorphic lycopsids: scenarios constructed on cladistic foundations. Biol. Rev. Cambridge Philos. ye 527-597. BATEMAN, R.M. . An overview - lycophyte phylogeny. iii 405-415 in Pteridology in Per- spective, os Ia Camus et al., Royal Botanic Gardens, BATEMAN, R.M., ina and a ‘4 WILLARD. 1992. oatenenind cladistics of anatomically preserved hie rescent lycopsids from the oe of Euramerica: an essay on paleo- botanical phylogenetics. Ann. Missouri Bot. Gard. 79: 500-559. BHAMBIE, S. 1963. Studies in Pteridophytes. III On the structure and eng sa! of = leaf and sporophyll of Isoetes coromandelina. Proc. Indian acad. Sci., Sect. B, 29: 169-19) BISCHOF, 1853. Abbildungen der Sigillaria Sternbergii. Zeitschr. Ges. eeniiriady = 1; 257; Bd.2: PI.VIII, Halle. BOCHENSKI, T. 1936. Uber a (bliiten) 0 Lepidophyten aus dem produktiven Karbon Polens. Ann. Soc Bock, W. 1969. The American pens flora = lb -aiecennibiae, Geologic Center Research Series 3-4, North Wales, Pennsylvania Bose, M.N. and J. BANERJI. 1984. The fossil pra ‘of Kachchh. I—Mesozoic megafossils. The Pa- sacrteingts 33: 1-189. BRACK, S. D. and T.N. TAYLOR. 1972. The ultrastructure and organization of Endosporites. Micro- agape 18: 101-109. BRAUN, A. 1868. Uber die ee Arten der Gattung Isoétes. Monatsber. KGnigl. Preuss. Akad. eae Berlin, 523-545. Brown, R.W. 1939. Some rinicoe fossil plants belonging to the Isoetales. J. Wash. Acad. Sci. 261-269. BROWN, R. W. 1958. New occurrences of the fossil quillworts called Isoetites. J. Wash. Acad. Sci. 48: 358-361. BROWN, R.W. 1962. oe flora of the Rocky Mountains and Great Plains. Geol. Surv. Prof. Paper 375: 1— CHALONER, W.G. 1967, -Lycophya, Pp.435-781 in Traité de Paléobotanique, ed. E.Boureau II Mas- son et Cie., Par ane W.G. ie The cone of Regan renng iai Haughton, from the Upper Devonian Ireland. J. Linn. Soc., Bot., paises, S. and D.C. McGREGOR. ae ‘Bisprangistobs harrisii gen. et sp. nov., an eligulate lycopsid cone with Duosporites megaspores and Geminospora microspores from the Upper Devonian of Pennsylvania, U.S.A. Pa te sseinuutiags B 210: 127-149. CHITALEY, S., and K.B. PicG. 1996. Clevelandodendron ohioensis, gen. et sp. nov., a slender upright lycopsid from the Late Devonian Cleveland Shale of Ohio. Amer. J. Bot. 83(6): 781-789. COLLINSON, M. 1991. Diversification of modern heterosporous pteridophytes. Pp. 119-150, in: Pol- n and spores, eds. S. Blackmore and S.H. Barnes. Systematic Association Special vol. 44, Clarendon Press, Oxford. DIMICHELE, W.A. and R.M. BATEMAN. 1996. The rhizomorphic lycopsids: a case-study in paleo- tanical classification. Syst. Bot. 21(4): 535-552 DOBRUSKINA, I.A. 1982. Triassic floras of Eurasia. Akad. Sci. USSR Trans., 365:1—193 (in Russian). DosRUSKINA, I.A. 1985. Some problems of the aise of the Triassic lepidophytes. Paleont. ly shee 74-88 (transl. from Paleont. Zhurn., 3: 90-104, in Russian). 146 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) DuTHiE, A.V. 1929. The species of Isoetes found in the union of South Africa. Trans. Roy. Soc. South Africa 17: 523-545. slag L. 1968. Les plantes fossiles dans leurs rapports avec les végétaux vivants. Masson & aris. diana, E. 1856. Geological Report of the Midland counties of North Carolina. G. Putnam, New York, 352 p- FLICHE, P. 1910. Flore fossile du Trias en Lorraine et Franche-Comté (avec des considérations sigs _ M.R. Zeiller). Berger-Levrault, Paris-Nancy, 297 p.; Bull. Soc. Sci. Nanc cy Ill, XI, a0 eee "1900. Notes on fossil plants collected by Dr. Ebenezer Emmons from the older Mesozoic rocks of North Carolina. U.S. Geol. S urvey, Ann. abit 20: (2): 277-315. Fucus, G., L. GRAUVOGEL-STAMM and D. MApER. 1991. Une remarquable flore a Pleuromeia et nan rnin in situ du Buntsandstein moyen (Trias inféiews) de l’Eifel (R. hore Morphologie, ogee et paléogéographie. Palaeontographica Abt. B, 222(4—-6): 89-1 GLAESNER, M.F. and V.R. RAO. 1955. Lower Cretaceous plant remains from the vicinity of Sn cus Goswami, H.K. 1976. A revision of ligule and labium in Isoetes. Acta Soc. Bot. Palawine. XLVI, (1-2): 69-76. GRAUVOGEL-STAMM, L. 1993. Pleuromeia sternbergii (Miinster) Corda from the Lower Triassic =f any—further observations and comparative morphology of its rooting organ. Rev. laeobot. Palynol. 77:185-212 GRAUVOGEL-STAMM, L. 1999. Bivenimeie sternbergii (Miinster) Corda, ein characteristische Pflanze des deutschen Buntsandsteins. Pp. 271-281, in Trias—Eine ganz andere Welt. Europa im GRAUVOGEL-STAMM. L. and S. AsH. 1999. “Lycostrobus” chinleana, an equisetalean cone from the Upper Triassic of the southwestern United States and its phylogenetic implications. Amer. J. Bot. 86(10): 1391-1405. GRAUVOGEL-STAMM, L, and P. DurRINGER. 1983. Annalepis zeilleri Fliche 1910 emend., un organe reproducteur a Lycophyte de la Lettenkohle de |’Est de : France. Morphologie, spores in situ et paléoécologie. Geologische Rundschau 72(1): 23-5 GRAUVOGEL-STAMM, L. and J. LANGIAUX. 1995. Polysporia pies n.sp., un nouvel organe producteur ae lycophyte du pci , fére supérieur) de Blanzy-Montceau (Massif Central, France). Sci. Géol., Bull. 4 HALL, J.B. 1971. a on an ei a. Bot. J. Linn. Soc. 64: 117-139. HELBy, R. and A.R.H.MarTIN. 1965. Cylostrobus gen. nov., cones ‘e oe plants from the Bot Whetorn North D ota. The Geol. Soc. America, Mem. 150: 1-183. HICKEY, R.J. 1985. series studies of Neotropical Isoetes. Ph.D. dissertation, The University of Connecticut. Storr HICKEY, R.J. 1986. The oak evolutionary and ae ree ntay oe of Isoétes, with descriptions of two new neotropical species. Syst. Bot. 11(2): 3 HICKEY, R.J. 1990. Studies of neotropical Isoétes L. I. aie a new subgenus. Ann. Missouri Bot. Gard. 77: 239-245. JENNINGS, JR. 1975: eee a new plant organ from the Lower Mississippian of Virginia. adage. 18(1): 1 JENNINGS, J.R. R. KARRPALT = G.W. ROTHWELL. 1983. Structure and affinities of Protostigmaria eggertiana. heat er. J. Bot. 70(7): 963-974 JERMY, A.C. 1990. Isoetaceae. Pp. 26-31, in The families and genera of vascular plants. I Pteri- dophytes and gymnosperms. Eds. K.U. Kramer and P.S. Green, Springer—Verlag, Berlin KARRFALT, E.E. 1984. Further observations on Nathorstiana (Isoetaceae). Amer. J. Bot. 71(8): 1023- 1030. GRAUVOGEL-STAMM & LUGARDON: TRIASSIC LYCOPSIDS 147 KARRFALT, E.E. and D.A. EGGERT. 1977. The comparative morphology and ee = Isoetes L. I. Lobe and furrow development in I. tuckermanii A.Br. Bot. Gaz. 138(2): 236 neem. E.E. and D.A. EGGERT. 1978. The comparative morphology and ribet af Ionia L. Ill. The sequence of root initiation in Lape and four-lobed plants of J. tuckermanii A. Br. and J. nuttallii A. Br. Bot. Gaz oy 2 KELBER, K.-P. and W. HANscu. 1995. si spliaisoare Die Entratselung einer tiber 200 Millionen ie Kort, L.S. and D.M. BRITTON. 1985. Role of morphological characteristics of leaves and the spo- jig region in the taxonomy of Isoetes in northeastern North America. Amer. Fern J. 75(2): 44— KOVACH, e ‘. 1994. A review of Mesozoic oka ultrastructure. Pp.23—37, in Ultrastructure of fossil spores and pollen, eds. M.H. Kurmann and J.A. Doyle—R.B.G., “sear K. and R. BORCHERT. 1964. Morphologische Studien an Isoetes cain a 7% Braun und Bemerkungen tiber das Verhaltnis der Gattung Stylites E. Amstutz zur Gattung Isoetes L.. Ber. Deutsch. Bot. Ges. 77: 227-233. LUGARDON, B. 1972. Sur la structure fine et la nomenclature des parois microsporales de Selagi- nella denticulata (L) Linck et faa (L) Link. Comptes Rend. Acad. Sci. Paris, Sér. D, Sciences Lippi 274: 1 menclature a anes fine des parois acéto-résistantes des microspores d’'Is patie, Compt Rend. Acad. Sci. Paris, Sér. D, Sciences Naturelles 276: 3017-3020. LUGARDON, B. 1980. Isospore and microspore walls of living Pteridophytes; identification possi- bilities ne different pre ig instruments. Proceedings IV Internat. Palynol. Conf., B.S.I.P. Lucknow, India, 1: 199-206. LUGARDON, B. 1986. Données a tig tales sur la fonction de l’exospore chez les oo Pp: wee 264, in sind and spores: Form and function. Eds. S. Blackmore and I.K. Ferguson (Linn Symp. Ser. N°12), Academic Press, London LUGARDON, . err Pinvidoplivte sporogenesis: a survey of spate wall ontogeny and fine structure in a polyphyletic plant group. Pp. 95-120, in Microspores: Evolution and Ontogeny, eds.S. Blackmore and R.B. Knox. Academic ieee, London. LUGARDON, B., L. GRAUVOGEL-STAMM AND I. DOBRUSKINA. 1997. Sur |’ultrastructure des micro- spores et mégaspores de Pleuromeia rossica (Lycophyta) du Trias inférieur de Russie. XV Symposium APLF, Lyon 1997, Abstract: 41. LUGARDON, B., L. GRAUVOGEL-STAMM AND ‘I DoOBRUSKINA. 1999. The microspores of Pleuromeia rossica Neuburg (Lycopsida; Triassic): comparative ultrastructure and phylogenetic impli- cations. Comptes Rend. Acad. Sci. Paris, Sér. Ila, Sciences de la Terre et des Planétes 329: 435-442. LUGARDON, B., L. GRAUVOGEL-STAMM AND I. DOBRUSKINA. 2000a. Comparative ultrastructure of the megaspores of the Triassic lycopsid Pleuromeia rossica rhe 4 Comptes Rend. Acad. Sci. Paris, Sér. Ila, Sciences de la Terre et des Planétes 330: 50 LUGARDON, B., L. GRAUVOGEL-STAMM, R. COQUEL AND C. coe ak 2000b. Spores a ultrastructure isoétalienne ches des Lycophytes triasiques et plus anciennes (Carbonifére, Dévonien). Colloque OFP, Lyon 2000, Abstract: 11 and OFP Informations Maper, D. 1990. Palaeoecology of the flora in aera and Keuper in the Titinde ot Middle rope. 2 i 1-1582; Gustav Fischer Verlag, Stuttgar eben ta K. 1931. Zur Morphologie und pylogentschon Bedeutung der fossilen Pflanzgat- tung Saal Beih. Bot. Centralbl. 48: MARCINKIEWICZ, T. and E.V. ZHELEZKOVA. 1992. a compatitoit between dispersed megaspores of m =I Q > a6! fs g m8 MENG, F. 1995. Flora of the Badong Formation. Pp. 6-27, in Nonmarine biota ane sedimentary ge ¢ the Badong Formation in the Yangzi and its neighbouring area, <i F. Men Xu, Z. Zhang, J. Lin and H.Yao. Press China Univ. Geosciences, Wuhan, 76p (in Chinese ; English abstract). MENG, F. 1996. Middle Triassic lycopsid flora and its paleoecology in South China. The Palaeo- Re. 45: 334-343. 148 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) _ F. 1998. Studies on Annalepis from range Triassic along the Yangtse River and its bearing n the origin of Isoétes. Acta Bot. Sinica 40: 768-774. ene F. 2000. Advances in the study of Middle ‘Triassic plants of Yangtse Valley of China. Acta Palaeontologica Sinica 39 (sup.): 154-166 MEYEN, S.V. 1987. Fundamentals of Palaeobotany. Chapman & Hall, London. MoTELAY, L. 1893. Isoetes brochoni Motelay spec. nov. Actes Soc. Linn. Bordeaux, 45: 45 NEUBURG, — F. 1960. Pleuromeia Corda from the Lower reaper ra of the Russian ened Akad. Nauk. SSSR, Geol. Inst. Trudy, 43: 65-92 (in Rus — R. 1955. A propos des aSagaeap de transfert ies anon reproductrices. Rec. Trav. nst. bot. iy 7: 75-8 Q@LLGAARD, B. 19 a Pp.31-38, in: jon families and genera of vascular plants. I ne AEH and Gymnosperms, eds. Kramer K.U. and P.S. Green, ean Berlin. PicG, K.B. 1992. Evolution of Isoetalean ocak Ann. Missouri Bot. Gard. 7 Picc, K.B. are Isoetalean lycopsid evolution: from the Devonian to the eae yp ts tis J 91: 99-11 Picc, K.B. a G.W. ROTHWELL. 1979. co giaaee transition of an Upper Pennsylvanian woody lycopsid. Amer. J. Bot. 66(8): 914-92 Picc, K.B. and G.W. ROTHWELL. G.W. si ‘Chalonecks rane nov.: heterosporous lycophytes from e Pennsylvanian of North America. Bot. Gaz. 144:132—147 Pitot, A. nage Contribution a I’ gee oi Isoetes aditcacines Isoatea melanotheca Alston. Bull. Inst. ncisapeee 9 Fy 21: ae pny! tT 1904 i iler Pfl iscl 1 mesozoischen Formationen. oii Preuss. Geol. Landesanst. saa Liat II, 38:1-15, Ber- RAUH, W. and H. FALK. 1959. Stylites E. Amstutz, eine neue Isoétacee aus den Hochanden Perus. 1. Teil: Morphologie, Anatomie und Entwicklungsgeschichte der bapa a Abh. Sitzungsber. Seidalberpex Akad. Wiss., Math.-Naturwiss. K]., Abt.B 1959: 3-160. RETALLACK, G.J. 1997. Py Triassic origin of Isoetes and quillwort sca radiation. J. Paleont. 71(3): 500— opi 880 G.W. and D. *; ae 1985. The rhizomorph apex of aap saan ae for mologies among the rooting organs of Lycopsida. Amer. J. Bot. 7 secant G.N. 1982. The morphology, systematics and SgiNiereo of in baht Tomiostrobus. Paleont. J. 16(1): 100—109 (transl. from Paleont. Zhurn. 1: 104-112, on nee an). Scott, A.C. and A.R. HEMSLEY. 1993. The spores of the TON lycopsi t from ergs, Fife, Scotland. Special Papers in Palaeontology, 49: 31-4 Sxoc, J.E. and C.R. HILL. 1992. The Mesozoic herbaceous lycopsids. Ann. aera Bot. Gard. 79: 648-675. So, J, D.L. DILCHER, and F.W. PoTTEr. 1992. A new ang of nape ti the mid-Cretaceous Dakota Group of Kansas and Nebraska. Amer. Fern J. 82(4): 1 spe OR N. S. 1989. Further considerations on the od of elie Corda. In Paleo- ristic and stnctigreilite questions. Akad. Nauk. SSSR: 74-88, see cine et ayeeons SOLMS- fama H. 1899. Ueber das genus eodiedanl Bot. Zeitung, 2. Abt. 5 7-243. SRIVASTAVA, G. K. and S.O. WaGal. 1996. Scales and phyllopodia of some se a of Isoetes L. Geophytology, 25: 137-14 STEWART, WN. and G.W. ROTHWELL. . 1993. Paleobotany and the evolution of plants. 2nd ed., 521p. C oOo b ; ware R.G. and R,J. nck. 1983. pss Fp. 62-67 in R.G. Stolze (ed.): The flora of Gua- mala. Part Ill. Fieldiana Bot., ae T.N. 1990. iiunbniecndice in fossil plants, Pp. 121-145, in Microspores—Evolution and Ontogeny, eds. S. Blackmore and R.B. Knox. Academic Press, London. TAYLOR T.N. and _ TAYLO OR. 1993. The eies pete evolution of fossil plants. Prentice Hall, En- glewoods SC New Jersey 07 TAYLOR, W.A. 19 lg wall sna aiotine in Isoetes. Amer. J. Bot. 80: 165-171. TAYLOR W.A. <a. TAYLOR. 1990. Persistent ultrastructural features in microspores of hetero- sporous nace Grana 29: 219-228. GRAUVOGEL-STAMM & LUGARDON: TRIASSIC LYCOPSIDS 149 Tay or, W.C. and R.J. HICkEy. 1992. Habitat, evolution and speciation in Isoetes. Ann. Missouri Bot. Gard. 79: 613-622. TEIXERA, C. 1948. Flora Mesozoica Portuguesa. Servigos Geologicos de Portugal, Lisboa THOMAS, B.A. 1988. The fine structure of the Carboniferous ca aja microspore Lyeospora per- forata Bharadwaj and pea to Pollen and Spores 1-8 THoMaS, B.A. and S.D. BRACK-HAN 984. A new approach ms ae, groupings in the Lycophy- tes. Taxon 33(2): 247-255. TRYON, A.F. and B. LUGARDON,. 1991. Spores of the Pteridophyta: Surface, wall structure and di- versity based on electron microscope studies. Springer-Verlag, New Wacner, R.H. 1989. A late Stephanian forest swamp with Sporangiostrobus fossilized by volcanic ash fall in the Puertollano Basin, central Spain. International Journal of Coal Geology, 12: 523-552. WANG, Z. 1991. Advances on the Permo-Triassic lycopods in North China. I. An oie from the Mid-Triassic in Northern Shaanxi Province. Palaeontographica 222B (1—3):1- Wanc, Z. and L. WANG. 1982. A new species of the lycopsid cbpninggts from a nae Triassic of Shanxi, China and its ecology. Palaeontology, 25(1): 215-2 Wang, Z. and L. WANG. 1989. oo early Triassic fossil plants in ne Shigianfeng Group in North China. pape Geo 4(1): 23-40. Wane, Z. and L 1990. “we early Triassic seo plants from upper part of the Shigianfeng Group in ea mer Shanxi Geol., 5(2)195— Wanntor?P, H.-E. 1970. The genus Isoetes in South we Africa. Svensk Bot. Tidskr. 64(2): 141- 157 speier oe 1899. Fossil flora of the Lower Coal Measures of Missouri. US Geol. Surv. Monographs, 1-307, Washington ware, iM E. 1981. Gylomeie undulata (Burges) gen. et comb. nov., a lycopod of the early Triassic a of New South Wales. Rec. Austral. Mus. 33(16): 723-734. wei S. 1943. On Isoetes australis S. Williams, a new cig from Western Australia. Part I. General Morphology. Proc. Roy. Soc. Edinburgh, B, LX YAROSHENKO, O.P. 1975. Morphology of Pleuromeia rossica oats Densoisporie nejburgii spores. Paleont. J. 9(3): 373-379 (transl. from Paleont. Zhurn. 3: 101-1 YAROSHENKO, O.P. 1988. The microspores of Tomiostrobus radiatus Neuburg. Palynology in SSSR, Acad. Novosibirsk Pp. 77-79 (in Russian ZHELEZKOVA, E.V. 1985. The structure of mega a wall of Pleuromeia rossica eimai gag m the Lower Triassic of Yaroslavl District. Bot. Zhurn. (Moscow & Leningrad) 70: pan (in Russian). American Fern Journal 91(3):150—165 (2001) Diversification and Relationships of Extant Homosporous Lycopods NIKLAS WIKSTROM Department of Palaeontology, The Natural History Museum, Cromwell Road, London SW7 5BD, United Kingdom ABSTRACT.—A series of phylogenetic analyses using nucleotide sequence data have resolved many aspects of the relationships in a group of land plants that until recently had received comparatively little attention, the homosporous lycopsids (Lycopodiaceae). The family includes no more than 400-500 living species but the group has evolved as an isolated lineage since the Early Devonian (390 Mya). Despite this ancient history, patterns ging through the phylogeneti lyses imply that most diversification in this group is comparatively recent. The Lower Jurassic stem section Lycoxylon indicum indicates a minimum age for the split between Lycopodium and Lycopodiella at 208 Mya, and reticulate fossil spores from the Early Jurassic indicate that early cladogenesis in Lycopodium is of equivalent age. In the diverse predominantly epiphytic Huperzia group, biogeo- graphic data indicates that 85-90 % of all living species result from cladogenic events postdating the final rifting of S. America and Africa in Mid to Late Cretaceous. The timing of these events coincides with the radiation of Angiosperms, and the diversification of epiphytic Huperzia was likely mediated by the development of broad leaved Angiosperm rain forests. Results indicate a single origin of epiphytism in Huperzia, but there have been at least two reversals to the terrestrial habit in the neotropics. The diversification of a large secondarily terrestrial clade, including about 80 montane high altitude species, was likely triggered by the Andean orogeny in the Mid Miocene, no more than 15 Mya. Lycopsids are an ancient group with an extensive fossil record, and small herbaceous species are among the most easily recognizable and conspicuous elements of the early land flora. Because of their lengthy fossil record, the oup was soon recognized as the product of an early cladogenic event in land plant evolution (Banks, 1968, 1975; Bower, 1908). Recent phylogenetic anal- yses support this view and indicate that lycopsids are sister to all remaining vascular plants (Doyle, 1998; Duff and Nickrent, 1999; Kenrick and Crane, 1997; Kranz and Huss, 1996; Raubeson and Jansen, 1992). The fossil record provides strong evidence that the three extant lineages (Lycopodiaceae, Sela- ginellaceae and Isoetaceae) also originated in the Palaeozoic. The occurrence of Protolepidodendrales in the Early Devonian (Kenrick and Crane, 1997) sug- ests a minimum age of 390 Myr for the split between Lycopodiaceae and the ligulate clade including Selaginellaceae and Isoetaceae. Furthermore, the oc- currence of Isoetales in the Late Devonian (Kenrick and Crane, 1997) suggests a minimum age of 377 Myr for the split between extant Isoetaceae and Sela- ginellaceae. Despite the antiquity of these events, little is known about the diversification of the crown groups. Although the fossil record includes taxa showing similarities to extant Lycopodiaceae (Skog and Hill, 1992; Thomas, 1992), Selaginellaceae (Thomas, 1992, 1997) and Isoetaceae (Pigg, 1992; Re- tallack, 1997), their specific relationships to living species remain unclear. One of the major problems has been the lack of well developed phylogenetic WIKSTROM: EXTANT HOMOSPOROUS LYCOPODS 151 hypotheses. Without such hypotheses, there is no way of telling whether the fossils are nested within the crown groups, or are part of the stem lineages. Isoetes bestonii Retallack for example has been identified as far back as the Triassic (Retallack, 1997). If shown to be part of the crown group of living Isoetes, this would certainly contribute to our understanding of the patterns of diversification of living species. Separated from the ligulate lycopsids since the Early Devonian, the homo- sporous lycopods (Lycopodiaceae) have evolved as an independent lineage for at least 390 Myr. Morphologically, living species show striking similarities to some of the oldest vascular plants (Heuber, 1992; Kenrick and Crane, 1997), but from a palaeobotanical perspective Lycopodiaceae have been particularly problematic (Skog and Hill, 1992; Thomas, 1992). The macrofossil record is relatively meager and palaeobotanical work is hampered by recognition prob- lems. Absence of clear morphological synapomorphies for the family makes assignments of fossil taxa difficult, and this is often based on absence of the characteristics of more distinctive groups such as Selaginellaceae. Another problem involves the superficial similarities between Lycopodiaceae and organ systems of other plant groups. These problems notwithstanding, Skog and Hill (1992) hypothesized that the major living groups within Lycopodiaceae were established during the Late Jurassic-Cretaceous. In other words, despite its apparent relict status, modern species diversity in Lycopodiaceae evolved comparatively recently and in parallel with that in more diverse, ecologically prominent groups such as flowering plants. Over the last couple of years, I and my coauthors have been developing a phylogenetic hypothesis for living species of Lycopodiaceae (Wikstrém and Kenrick, 1997, 2000a, 2000b; Wikstrém et al., 1999). Using molecular sequence data from the plastid rbcL gene, and from the trnL-trnF intron and spacer regions, we have conducted a number of phylogenetic analyses at different levels of the hierarchy in the family. These analyses provide empirical support for the idea that most cladogenesis in this ancient family is comparatively recent. Here I review some of the progress that has been made in our under- standing of Lycopodiaceae relationships and the implications for interpreting the evolution of the family. RELATIONSHIPS LYCOPODIACEAE MONOPHYLY AND PHYLLOGLOSSUM.—Recent cladistic analyses have shown that monophyly of Lycopodiaceae is at best weakly supported by morphological characters (Crane, 1990; Kenrick and Crane, 1997). Other stud- ies have failed to identify unequivocal family synapomorphies (Wagner and Beitel, 1992), or indicate that Lycopodiaceae may be paraphyletic to a hetero- sporous lycopsid clade comprising Selaginellaceae, Isoetaceae and extinct rel- atives (Bateman, 1992). Molecular characters, on the other hand, furnish very strong support for monophyly of the family (Wikstrém and Kenrick, 1997, 2000a). Their rbcL sequence data unequivocally support Lycopodiaceae mono- phyly. Furthermore, molecular data resolve a basal split in Lycopodiaceae sep- 152 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) Marchantia polymorpha Andraea rupestris Sphagnum palustre Psilotum nudum 4 Equisetum arvense Lt @ Angiopteris evecta 2 Ephedra tweediana 3 sai . 64 Zamia inermis 13 99 Pinus edulis 20 a Psedotsuga menzi lsoetes melanopoda lsoetaceae ‘sbiicptuiibaitas Selaginellaceae Selaginella selaginoides Phylloglossum drummondii Huperzia selago Huperzia wilsonii (Epiphyte) Huperzia hippuridea Huperzia campiana (Epiphyte) Huperzia cumingii Lycopodiaceae 18 Huperzia linifolia (Epiphyte) Lycopodiella inundata 7] Lycopodiella alopecuroides Lycopodium clavatum “Strobilate clade” Lycopodium obscurum Lycopodium digitatum By Fic. 1. A single most parsimonious tree indicating nop: of Lycopodiaceae. The tree is reproduced from the analyses of Wikstrém & Kenrick (1997) of rbcL sequence data. The tree is printed as a phylogram with branch lengths proportional to the amount of change along each branch and branch support is indicated by decay indices (above) and bootstrap values (below) for each branch. arating a “‘strobilate clade”, picemaenae: Lycopodium L. and Lycopodiella Holub, om a Huperzia Bernh.—Phylloglossum Kunze clade. The single most parsi- monious tree from Wikstrém and Kenrick (1997) is reproduced in figure 1 and shows branch support as measured by decay indices (Bremer, 1988; Donoghue et al., 1992) and bootstrap values (Felsenstein, 1985). The support for mono- phyly of Lycopodiaceae is not entirely unexpected. Perhaps more surprising is the close relationship between Phylloglossum and Huperzia (Fig. 1). Phyl- loglossum is a monotypic genus occurring only in Australia and New Zealand, WIKSTROM: EXTANT HOMOSPOROUS LYCOPODS 153 and its highly divergent morphology has made its relationships to other spe- cies in Lycopodiaceae difficult to understand. Previous taxonomic treatments have either implicitly or explicitly classified it incertae sedis with respect to other groups in the family (Holub, 1985; Ollgaard, 1987; Wagner and Beitel, 1992). A close relationship between Phylloglossum and Huperzia is however consistent with a number of morphological features such as spore morphology (Breckon and Falk, 1974; Tryon and Lugardon, 1991), sporangial epidermis morphology (Ollgaard, 1975), phytochemistry (Markham et al., 1983) and chro- mosome number (Blackwood, 1953). Other characters, and in particular the morphology of the perenniating tuber in Phylloglossum, led workers such as Bower (1935), Bruce (1976a) and Hackney (1950) to consider a close relation- ship between Phylloglossum and Lycopodiella. The rbcL sequence data, how- ever, lend unequivocal support for the Huperzia-Phylloglossum relationship. The evolution of Phylloglossum from either a Huperzia or a Lycopodium- Lycopodiella like ancestor implies a remarkable architectural transformation. This include the development of a new organ (perenniating tuber), a substan- tial reduction of the microphyll-bearing stem system, and the development of a unique stem anatomy. It is noteworthy though that the aspects of morphology likely to be the least affected by this overall architectural change, such as spore and sporangial epidermis morphology and phytochemistry, are also the most consistent with the rbcL data in supporting a Huperzia relationship. THE “STROBILATE CLADE”’.—The ‘“‘strobilate clade” includes the two genera Ly- copodium and Lycopodiella (sensu Oligaard, 1987), and although comprising no more than about 40 species each, 11 genera (Holub 1964, 1975, 1983, 1985, 1991) or 13 sections (Ollgaard 1987) have been recognized. My usage of generic and subgeneric names follows that of Ollgaard (1987). One factor contributing to the inflation of subgeneric groups is the segregation of several divergent species as monotypic groups. However, this segregation also reflects a genuine morphological pattern where subgeneric groups are very distinct, each includ- ing very similar species, but where there has been considerably more difficult to find similarities between such groups. Works looking at various morpho- logical features such as stem anatomy (Bierhorst, 1971; Jones, 1905; Ogura, 1972), spore morphology (Tryon and Lugardon, 1991; Wilce, 1972), distribu- tion of mucilage canals in trophophylls and sporophylls (Bruce, 1976b), branching pattern (Ollgaard, 1979), chromosome numbers (Wagner, 1992), spo- rangium epidermis morphology (Ollgaard, 1975), and gametophyte morphol- ogy (Bruce, 1976c) have tended to reinforce this pattern, and Ollgaard (1990) hypothesized that the subgeneric groups represent ancient evolutionary line- ages. dee the success of recent taxonomic work in recognizing subgeneric groups and our increased knowledge about the morphological variation in these plants, there have been few explicit attempts to investigate the phylo- genetic relationships of Lycopodium and Lycopodiella, Wagner and Beitel (1992) analyzed the relationships of N. American species and Wikstrém and Kenrick (1997) included representatives of both genera in their analyses. Both 154 AMERICAN FERN JOURNAL: VOLUME 91 NUMBER 3 (2001) ; H. selago oo Phylloglossum SP gee H. hippuridea 100 H. campiana L. magellanicum ; fi Section Magellanica L. fastigiatum 96 L. volubile Section Pseudodiphasium L. alpinum : Te L. wightianum Section Complanata 1 L. digitatum L. annotinum ‘ Section Annotina 82 L. clavatum : . 1 1 ; Section Lycopodium 4 100 L. vestitum L. obscurum Section Obscura 3 : eas WO} ay L. jussiaei 8 93 100 L. scariosum £. deuterodensum ¢ Section Pseudolycopodium L. casuarinoides ‘ Section Lycopodiastrum Section Diphasium 58 bak 3 rh ae l Section Lycopodiella La. cernua ‘00 Tw La. glaucescens Section Campylostachys 3 * La. pendulina a4 La. lateralis ; Section Lateristachys La. caroliniana ‘ Section Caroliniana Fic. 2. Strict consensus of two most parsimonious trees indicating relationships within the “stro- bilate clade”. The tree is reproduced from the analyses of Wikstrém and Kenrick (2000b) of com- bined rbcL gene and trnL intron sequence data. Branch support is indicated by decay indices (above) and bootstrap values (below) for each branch. Si bg ic groups (secti Ollgaard 1987) are shown to the right. studies however, only included a limited number of species and the majority of the subgeneric groups recognized were not included. To address the rela- tionships of Lycopodium and Lycopodiella, and specifically the relationships among subgeneric groups, we recently undertook an analysis using combined rbcL gene and trnL intron sequence data from a more extensive sample of taxa (Wikstro6m and Kenrick, 2000b). The strict consensus tree from this analysis is reproduced in figure 2 showing branch support (decay indices and bootstrap values), and subgeneric groups (sections sensu Ollgaard, 1987) to the right. As seen in figure 2, the analyses support monophyly of both Lycopodium and Lycopodiella as well as monophyly of the sections. Branch support is however relatively low for a number of groups and some of the questions concerning WIKSTROM: EXTANT HOMOSPOROUS LYCOPODS 155 relationships among subgeneric groups remain unresolved. There are however some interesting patterns that not only provide an opportunity to calibrate the tree against the fossil record (see below) but that are supported by morpho- logical data. Lycopodium casuarinoides Spring. (section Lycopodiastrum) is grouped as sister to all remaining Lycopodium. This basal split in Lycopodium is sup- ported by at least two morphological features: the lack of spore muri (Tryon and Lugardon, 1991; Wilce, 1972) and the presence of thick, lignified and sinuate sporangium Cell walls in L. casuarinoides (Ollgaard,