American Fern Number 4 Journal ie QUARTERLY JOURNAL OF THE AMERICAN FERN SOCIETY Editor Alan R. Smith Department of Botany, University of California, Berkeley, CA 94720 Associate Editors Gerald J. Gastony, Department of Biology, Indiana University, Bloomington, IN 47401 Christopher Haufler, Department of Botany, University of Kansas, Lawrence, KS 66045 David B. Lellinger, U. S. National Herbarium NHB-166, Smithsonian Institution, Washingt D.C. 20560 Terry R. Webster, Biological Sciences Group, University of Connecticut, Storrs, CT 06288 The American Fern Society Council for 1989 ane ang SKOG, Biology Dept., George Mason University, Fairfax, VA 22030. President AVID B. LELL LINGER, Smithsonian Institution, biCoiesi-eeg: DC oie Vice-President we CARL TAYLOR waukee, WI5 Secreta ry soa D. CAPONETTI, Dept. of Botany, ae os Tennessee, ail. TN 37916. Treasurer AVID S. BARRINGTON, Dept. of Botany, University of Vermont, Burlington, VT 05405 Re oaths Treasurer JAMES D. MONTGOMERY, Ecology III, R.D. 1, Berwick, PA 18 Back Issues Curator ALAN R. SMITH, Dept. of Botany, University of California sone ora 94720. Journal Editor DAVID B. LELLINGER, Smithsonian Institution, \idangee se > a 20 Memoir Editor JOHN T. MICKEL, New York Botanical Garden, Bronx, NY 1 *F iddlehead Forum Editor The “American Fern Journal” (ISSN 0002—8444) is an illustrated ] study of ferns. It is owned by the American Fern Society, and saree at ‘the ggcietae Herbarium, University of Vermont, Burlington, VT 05405-0086, and printed by A-R Editions, Inc., 801 Deming ay, Madison, WI 53717. Second-class postage paid at Burlington, VT, and additional entry point. Ohders for back issues should be addressed to Dr. — D. Meninoy. Ecology we KL. 3, Berwick, PA 18603. Back volumes 1910-1978 8, $5.00 to $6 ; sing 4 pages or less, $1. 25; 65-80p pages, $2.00 : 80 pag $2.5 50 each, plus shipping. Back vehi 1979 et seq. $8 gl e bach bers $2.00 each, pl I g. Ten percent aoee on orders of six volumes or more. Subscriptions $20.00 gross, $19.50 net if paid through an agency (agency fee $0.50); sent free to members of the American Fern Society (annual dues, $15.00 + $4.00 mailing surcharge beyond U.S.A., Canada, and Mexico; life membership, $300.00). Fiddiehead Forum The editor welcomes contributions from members and non-members, including miscellaneous notes, offers to exchange or ee materials, personalia, horticultural notes, and reviews of non-technical books on ferns Spore Exchange Mis. jocelyn Hordler, 16813 Lemolo Shore Drive De E., Poulsbo, WA 98370, is Director. Spores ton reque. Gifts and aaa Gifts and bequests to the Society enable it to expand its services to members and to others interested in ferns. a eet back i es aye the Journal, and cash or other gifts are always welcomed, and are tax-d = Table of Contents (A list of articles arranged alphabetically by author) ALVERSON, mashes R., Cryptogramma cascadensis, a new Parsley-fern from western North AMIOTICR 6 oo. Be ee pe ee es ee, BALLARD, eae (S66 RAVE cco i a ee ee ge le Davi S., sana eaeas EH. BAUER, and CHARLES R. WERTH, Hybridization, iculation ses joo aig Wagner, Warren Hoy ee ye ee OP, ARL, New species of Cotodania sube: Commdenia 0. ioe. ee vaophlebia, anew N genus of conemaracmmmarieed SATS aii a Mma Gham a had CUSICK, ALLISON W dextantin Maryland ............ CUTTER, ELIZABETH G. (ace tena PO oes oles ee ee ee FARRAR, DONALD don PEER ee ae ee ee ae Max: O49 bw Oo 8 eee a Re eee Bw ee oe eke os ee Re ee ee ee ee © le ee 8 kW Sb ee Oe 6 ee Re eee 8 ek ee 8 ks oe oe ee 6 ee ew 8 8 eS see Fay, ALICE, and ace BALLARD, The fl ids of Polysticl NOE 55, Ga ASTONY, GmALD LD J, cue — D. WINDHAM, Species concepts in anion The HAGENAH, shaadi (se Wagner Warren H., Jr. ,. Pe evar ee HAUFLER, CHRISTOPHE p idophytes: Introduction j ah S , Bred £ t Bs oe see HAPTINGION) 6 oo ee i ee Ci ee eh eee aes a HERD, Y. bia ~~ ane G. CUTTER, and I. bee dae teegt les effects a oengeaons and selected zolla HICKEY, R steed W. CARL TAYLOR, and NEIL T. LUEBKE, The species ab aecy in Pteridophyta with special refererice to etd oo a eens preme wer. +, (666 FICKOV) nc ee ge eee sk oe ee ec eee eee at aeoarian NAN cy A and DONALD R. FARRAR, Recover y i < obo + oe eb eae ew Oe ee ee eee a ee Be 8 ow NAUMAN, CLIFTON E. (666 Moyroud) « « .. sso 5 en cy ice ee tee new ew cere es PEPROLA. eevee (806 POCK) . oc oe oe ha a i ey ee ee eg ede eee eee @LLG , BENJAMIN, New taxa and combinations of — Lycopodiaceae ......... Paris, CATHY A., —— WAGNER, and W. WARREN H GNER, JR., Cryptic species, species delimitatio the homosporous ferns «......5....6.-.-. PECK, JAMEs H. Teoeine NEKOLA, and Donatp R. FA ; Ponce, M. Monica, and Marta A. MorseELLt, The ye jane dichotoma group of South RILEY, ins Me laos an Pa Sal ee oe es re eo Fe ee ee eee e reese se Sottis, Douctas E. (see Soltis, Pamela S.) ..... 6.6... eee ee eee eee teeter eee SOLTIS, PAMELA S., DoucLas E SOLTIS, Pau G. Wo tr, and Jitt M. Ritey, Electrophoretic in Polystichum .. ..: e6. 5 ope ee eee eee , and ETHELDA HaGENAH, A sa Pianeta hybrid, x Asplenosorus pinnatifidum x Phyllitis scolopendrium var. americana .....-.+.+-++++++sreees CO RT i a ee hs Ce nh eet cee nes ogicieaiats GEORGE, A new combination i in South American Polystichum pe eG ty Sat ROBBIN Cc. MoRAN, Primary pod p in f ferns ZIMMER aa PELDRIII: A ick d yalt tivet itical nojint GIOIITON ACROREGY cc Pee a ee Volume 79, Number 1, pages 1—32, issued 15 June 1989 Volume 79, Number 2, pages 33—93, issued 16 June 1989 Volume 79, Number 3, pages 95—126, issued 29 August 1989 Volume 79, Number 4, pages 127—158, issued 29 January, 1990 American Fern apy Jou rnal October—December 1990 QUARTERLY JOURNAL OF THE AMERICAN FERN SOCIETY Editors Carol and James Peck Department of Biology, University of Arkansas-Little Rock, Little Rock, AR 72204 Associate Editors Michael I. Cousens, Ctr. Science Education, Weber State College, Ogden, UT 84408 Gerald J. Gastony, Department of Biology, Indiana University, Bloomington, IN 47401 Christopher Haufler, Department of Botany, University of Kansas, Lawrence, KS 66045 David B. Lellinger, U.S. National Herbarium NHB-166, Smithsonian Institution, Washington, D.C. 20560 The American Fern Society Council for 1990 DAVID B. LELLINGER, Smithsonian Institution, Washington, DC 20560. President ALAN R. SMITH, University Herbarium, University of California, Berkeley, CA 94720. Vice-President W. CARL TAYLOR, Milwaukee Public Museum, Milwaukee, WI 53233. Secretary JAMES D. CAPONETTI, Dept. of Botany, University of Tennessee, Knoxville, TN 37916. Treasurer DAVID S. BARRINGTON, Dept. of Botany, University of Vermont, Burlington, VT 05405. Records Treasurer JAMES D. MONTGOMERY, Ecology III, R.D. 1, Berwick, PA 18603. Back Issues Curator CAROL AND JAMES PECK, Dept. of Biology, University of Arkansas-Little Rock, Little Rock, AR 72204. Journal Editors DAVID B. LELLINGER, Smithsonian Institution, Washington, DC 20560. Memoir Editor JOHN T. MICKEL, New York Botanical Garden, Bronx, NY 10458. Bulletin Editor The “American Fern Journal” (ISSN 0002-8444) is an illustrated quarterly devoted to the general study of ferns. It is owned by the American Fern Society, and published at the Pringle Herbarium, University of Vermont, Burlington, VT 05405-0086, and printed by A-R Editions, Inc., 801 Deming Way, Madison, WI 53717. Second-class postage paid at Burlington, VT, and additional entry point. Orders for back issues should be addressed to Dr. James D. Montgomery, Ecology III, R. D. 1, Berwick, PA 18603. Back volumes 1—68 (1910-1978) $5.00, $1.25/number; vol. 69-75 (1979— 1985) $8.00, $2.00/number; vol. 76 to present (1986—) $15.00, $3.75/number, plus shipping. Ten percent discount on orders of six volumes or more. Subscriptions $20.00 gross, $19.50 net if paid through an agency (agency fee $0.50); sent free to members of the American Fern Society (annual dues, $15.00 + $4.00 mailing surcharge beyond U.S.A., Canada, and Mexico; life membership, $300.00). Fiddiehead Forum The editor of the Bulletin of the American Fern Society welcomes contributions from members and non-members, including miscellaneous notes, offers to exchange or purchase materials, personalia, horticultural notes, and reviews of non-technical books on ferns. Spore Mrs. Jocelyn Horder, 16813 Lemolo Shore Drive N.E., Poulsbo, WA 98370, is Director. Spores exchanged and lists of available spores sent on request. G uests Gifts and bequests to the Society enable it to expand its services to members and to others interested in ferns. Botanical books, back issues of the Journal, and cash or other gifts are always welcomed, and are tax-deductible. Inquiries should be addressed to the Secretary. Table of Contents (A list of articles arranged alphabetically by author) BERNATUS, SUSAN (see Mosbl6y) 2 oe ee ee ee cea DANIEL F., Isoetes xbrittoni hyb. nov. (Isoetaceae): . bert occurring hybrid Pica x I. riparia) in the Eastern United Sta’ Couns Moet Oe re eee ee FARRAR, DONALD R. i Pete, ae ee ee ee Soe oe kh ee ee HasEBE, Mitsuyasu, Adiantum capillus-veneris chloroplast DNA clone bank: as useful heterologous probes in the systematics of the leptosporangiate ferns .......... HAUFtER, CHRISTOPHER HH. [see Ranker) 2.0 6. yes ee es Sen ————- (see sete) pee Poe ee cc eae ee HOLSINGER, KENT E., ti ti f ti t lution in h a ie a a ae ae ae oe HOSHIZAKI, BARBARA Jor, cipro EM as i ee HovENKAMP, P., The significance of rhizome morphology in the systematics of the polypodiaceows terns (anand elricta) |... ok es on ee A es ERATO, FILIPPO, Two mangiferin glycosides from Asplenium adiantum-nigrum ...... IWATSUKI, RUmiO (800 Fi). a ee ei ea KATO, ma (se@ LIM) oo es ee ei ee ees oe LELLINGER, DAVID B;; Richard Eric Holttum (1895-1990) - se ee LEON, BLANCA, New localities for Acrostichum danaeifolium in Peru................ Lin, Su-JUAN, Sporogenesis, reproductive mode, and cytotaxonomy of some species of nomeris, Lindsaea, and Tapeinidium (Lindsaeaceae) ............-.+++-+ MICKEL, JOHN T., Three new species of Elaphoglossum from Peru ............-.++--> MonTGOMERY, JAMES D., Survivorship and ie ate changes in five populations of Bo ielinin dissectum in Eastérn Pennsylvania «.... 66.062 cece cere enees MosELEY, RoseErRT K., Confirmation of Pentagramma Side me fe. PaLactos-Rios, Monica, New pteridophyte records for the State of Veracruz, Mexico... . Pack, Canvt f. (aoe Peck, fe A) ce ee ee ee ee eee ee es Peck, J. H., Influences of life history attributes on the local and distant formation of fern Pern ote ar ies wae Price, MIcHAEL G., Four new Asian Loxogramme ......-- +--+ esses erereererees et ge RRS GIS ee. ey RANKER, THOMAS A., A new combination in Bommeria (Adiantaceae) ..........-.--. ee a |g Sem Og ee SCHNELLER, J. J., Antheridogens and natural gametophyte populations ........----.-. SmarH, ALAN R. (see Windische) ........ 6-2-2 - eect erect reese eer eeeeneees Sottis, Douetas E. (see Soltis, Pamela S.) .....-.-- 00sec e eee ree teense ees So.tis, PAMELA S., Genetic — within and among populations of ferns .......... Tariom, W. Cant (see Brunton) . ... 2 2.2 cette te eee rete wees serene Wacner, Davin, Michael I. ponte (19049-1000) .. oo oc cee tee ee eee es , Polystichum kwakiutlii sp. nov., the elusive bipinnate ancestor of P. andersonii . WaGNER, FLORENCE (see Wagner, pea SS ee (see Wagner, Warren H.., Jt.) .-.. 601 -cctetcteteee ese etn eeercees WAGNER, WARREN H., JR., paaiens nothospecies i in the Appalachian Asplenium complex . , Notes on the 'fan-leaflet group of moonworts in North America with descriptions of Sens Seek TEE ow nk iw io a oo ee te en ee ee eee eee es eee eee WerTH, CHARLES R., Symposium on Population Biology of Ferns: introduction ........ ummary: The Contributions of Population Studies on POE oo ere weeks YATSKIEVYCH, GEORGE, A reconsideration of the genus Pityrogramma (Adiantaceae) in Wwesterd NON AMiGliGd 6.7. oe es ee ZHANG, XIAN-CHUN, Microgonium sublimbatum new to China.............0..0000- Volume 80, Number 1, pages 1—32, issues—June 1990 Volume 80, Number 2, pages 32—72, issued—August 1990 Volume 80, Number 3, pages 72—120, issued—December 1990 Volume 80, Number 4, pages 120—193, issued—April 1991 American Fern Volume 79 Number 1 January-March 1989 Journal QUARTERLY JOURNAL OF THE AMERICAN FERN SOCIETY A icaepsenre oamorewind Hybrid, x Asplenosorus pinnatifidus x Phyllitis colopendrium var. americana Warren H. Wagner, 2 and Ethelda Hagenah 1 Pamela S. Soltis, ae Soltis, Paul G. Wolf, and Jill M. Riley 7 New Sp f Ceradenia subg. Ceradenia L. Earl Bishop 14 A New Combination in South American Polystichum George Yatskievych 26 Shorter Notes Five Pteridophytes New to low pees H. Peck, Jeffery Nekola, and Donald R. Farrar 28 The Fl ids of Polystich tichoide Alice Fay and Robert Ballard 29 Reviews 31 Information for Authors Cover 3 The American Fern Society Council for 1989 JUDITH E. SKOG, Biology Dept., George Mason University, — is 22030. President DAVID B. LELLINGER, Smithsonian Institutio: n, Washington, DC 20560. Vice- a ecst W. CARL TAYLOR, Milwaukee Public Museum, ees Wrksss a33. Secretary JAMES D. CAPONETTI, Dept. of Botany, University of Tennessee, Knoxville, TN 37916. vrnpuneee DAVID S. BARRINGTON, Dept. of Botany, University of Vermont, Burlington, VT 05405 Records Treasurer JAMES D. MONTGOMERY, Ecology III, R.D. 1, Berwick, PA 1860 Back Issues Curator ALAN R. SMITH, Dept. of Botany, University of California, Berle, CA 94720. forge Editor DAVID B. LELLINGER, Smithsonian Institution, Washin: ngton, DC 20560. moir Editor JOHN T. MICKEL, New York Botanical Garden, Bronx, NY 10458. See Foruin Editor American Fern Journal EDITOR TOON 2 eo Dept. of Botany, esters of California, erkeley, CA 94720 ASSOCIATE EDITORS GERALDJ.GASTONY ............ Dept. of Biology, Indiana University, ear IN 47401 OTTER. 8. Dept. of Botany, University of Kansas, kaw wrence, KS 66045 DAVID B. LELLINGER ............... U.S. Nat’l Herbarium NHB-166, Smithsonian Institution, Washington, DC 20560 ticut orrs, CT The “American Fern Journal” (ISSN 0002-8444) is an il a study of ferns. It is owned by the American Fern Pacts and ade at the afta we University of Vermont, Burlington, VT 05405-0086. Second-class postage paid at Burlington, VT, point. mana f +} o12m #ha £f. : rs ohn & gs and orders fr back tues should be addrested fo to Dr. Janes emacs | Ecology III, RD. 4, Berwick, PA . Changes of address, dues, and applications for membership should be sent to the Records reasurer. rm as a . £ L 133 ie 3 L.c sj Pou skeserient ecke ae on ro ars (agency cy fee $0.50): sent free to members of the American Frm Soci al ue gig + + $4.00 mailing surcharge beyond U.S.A. Sarna eae p, $300.00 Back volumes 1910-1978 $5.00 to $6.25 cach: ete ee $1.25; 65-80 pages, $2.00 00 each; over 80 pages, $2.50 each, pe tradeing Back volumes 1979 et seq. $8.00 1. Wh OEM VULUIIITS a _ gem ERN JOURNAL, Dept. of os, University of es Fiddlehead Forum | t welcomes contributions from members and non-members, including miscellaneous aeiickems personalia, horticultural notes, and reviews of on ferns. Spore Exchange __ Mrs Jocelyn Horder, bo afmaniae scree Poulsbo, oe is Director. Spores ex- = ula, Gifts and Bequests a "Gifts and ate mt ce tt er mete nd he Fearon ferns. Botanical books, Tnquiies should bwin cash OF other gits are always “aay American Fern Journal 79(1): 1—6 (1989) A Synthetic “‘Trigeneric” Hybrid, x Asplenosorus pinnatifidus x Phyllitis scolopendrium var. americana W. H. WAGNER, JR. Department of Botany, University of Michigan, Ann Arbor, Michigan 48109 ETHELDA HAGENAH 164 Westchester Way, Birmingham, Michigan 48009 Hybrids in artificial cultures have contributed to the understanding of the systematic morphology of ferns. Since the classic work of Margaret Slosson (1902), which first proved the origin of the natural hybrid x Asplenosorus ebenoides (R.R. Scott) Wherry (=Asplenium platyneuron (L.) BSF. x Camptosorus rhizophyllus (L.) Link), the techniques of growing fern gametophytes and producing crosses have improved, and many new developments have ensued. In 1957, Wagner and Whitmire provided the first demonstration of the conversion of a sterile allodiploid fern to a fertile allotetraploid. In 1968, Lovis reconstructed a fertile hybrid species of fern (Asplenium ( x )adulterinum Milde). In this article, all taxa, fertile and sterile, of interspecific origin will be referred to as ‘‘nothospecies,” and indicated by the use of the times sign if sterile and with parentheses around the sign if fertile—i.e., (x), a convention proposed by C. Werth (pers. comm.); divergent species will be referred to as ‘‘orthospecies” and will lack the multiplication sign. Many important experiments on Aspleniaceae were accomplished in European laboratories especially, as discussed and summarized by Reichstein (1981). In almost all cases, such experimental hybridizations were carried out to test some hypothesis of the origin of a given nothospecies. By comparison, many fewer hybridizations have been undertaken simply to find out what a cross between taxon A and taxon B might look like, and what combining, for example, a creeping rhizome with an upright caudex, or hairs with scales, or discrete sori with acrostichoid sori, might yield morphologically. Yet such questions may bear upon our understanding of the determinants of structure and form; we may be able to gain insights that would otherwise be unavailable. The plants involved in this report are all members of the spleenwort family, Aspleniaceae, always popular objects for culture work and hybridization experiments because of their conveniently small size, ease of culture, rapid growth, and often very distinct forms. As a matter of fact, the bizarre hybrid that we briefly describe below was formed by accident. A terrarium containing numerous gametophytes of the lobed spleenwort, ( x )Asplenosorus pinnatifidus (Muhl.) Mickel (= mountain spleenwort, Asplenium montanum Willd. x walking fern, Camptosorus rhizophyllus (L.) Link) from south of Shoals, Martin Co., Indiana (kindly provided by Warren P. Stoutamire) was opened at the same time spores of the WOSOUR Bonrnign, JUN 9 1 19 GARDEN LIB-Apy 2 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 1 (1989) e % ; % ie i : ‘ ae é : % 5 f : : 4 % i Fic. 1. Leaf forms of the hybrid ( x ) Asplenosorus pinnatifidus x Phyllitis scolopendrium var. amer- icana. Note exaggerated basal auricl d near absence of lobation American hart’s-tongue, Phyllitis scolopendrium var. americana Fern., were being sown in a nearby culture dish by Ethelda Hagenah. She later noticed a peculiar sporophyte among the { x JA. pinnatifidus and showed it to Wagner, who diagnosed it provisionally as a hybrid. The plant was vigorous and lived from 1965 to 1973, so we were able to obtain numerous fronds and make a few observations on its chromosomes. Had we tried to predict—assuming precise “in-the-middle”’ intermediacy (Barrington, 1986)—what this trihybrid of Asplenium, Camptosorus, and Phyllitis would look like, we would have erred. for some characters. For example, by extrapolation, we would have expected it to have at least several pairs of lobes in the lower half of the blade, the scaliness of the stipe to be sparse, and the soral arrangement to be like that of Asplenium in half of the sori. WAGNER & HAGENAH: TRIGENERIC HYBRID 3 Fic. 2. Soriation of two full-sized fronds of the trihybrid, showing the close proximity of the sori and their almost uniformly double indusia. The hybrid superficially resembles a coarse, thick-textured plant of Camptosorus rhizophyllus, with more or less crispate margins and enlarged basal auricles. The very shallow undulations of the margins are probably traces of (x)A. pinnatifidus. Only one out of approximately 50 fronds had a distinct lobe above the base (Fig. 1). We expected the hybrid to be moderately lobed above the base of the blade, because the plant contains g three genomes for simple leaves (two from Phyllitis, plus one from Camptosorus with one for twice compound leaves (A. montanum), but we did not expect it to be almost wholly without lobes. (P. Mick Richardson, in litt., had detected the existence of xanthones in our hybrid, confirming the presence of the A. montanum genome.) Why the basal auricles are so large and elongate is unexplained, for neither Camptosorus or Phyllitis has them so exaggerated. The sori are abundant and closer together than they are in Phyllitis (Fig. 2). Most curious is the fact that practically all of the sori have the characteristic doubly indusiate condition, due to the facing indusia from adjacent pairs of fertile veins, found in Phyllitis but not in either the Asplenium or the Camptosorus. + AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 1 (1989) TABLE 1. Estimates of Chromosome Numbers. Meiotic figure Large pairs Large singles Small singles # Chromosomes A 15 40 73 143 B 18 34 76 146 c 14 52 70 150 Averages 15.7 42 73 146.4 es ca ae re See ee clea The frond is thick, probably the result of the influence of the coriaceous textures of the Phyllitis and Asplenium montanum parents. Vein anastomoses are absent or rare. If present at all, only one or two are found, mainly along the costa near the base of the blade above the auricles. The petiole is short, and bears numerous reddish-brown hairs and narrow scales up to 4 mm long. The base of the petiole is the only portion that is darkly pigmented. Thus the petiole characters are close to those of Phyllitis. The chromosomes of the hybrid at diakinesis were difficult to study because those of the Phyllitis genomes are decidedly larger than those of both the Asplenium and the Camptosorus, and a number of large pairs are formed. Large univalents may be confused with small pairs. None of the figures could be interpreted without difficulty, but the best three gave the estimates in Table 1. The estimated numbers are close to the expected 72 (Phyllitis) + 36 (Asplenium) + 36 (Camptosorus) = 144. The deviations of our estimated numbers are probably due to difficulties of assessing the figures rather than to aneuploidy: monosomics and trisomics and other aneuploid phenomena are apparently rare or absent among asplenioids in our experience. A certain amount of pairing between the homologous chromosomes of the tetraploid American Phyllitis is to be expected because it is probably an intraspecific polyploid derivative of the typical European diploid form. DISCUSSION WAGNER & HAGENAH: TRIGENERIC HYBRID TABLE 2. Comparison of Two Different Trigeneric Hybrid Combinations. ( x )Asplenosorus pinnatifidus (= Asplenium montanum X Camptosorus rhizophyllus, fertile form) x Phyllitis scolopendrium var. americana ( x )Asplenosorus ebenoides (=Asplenium platyneuron x Camptosorus rhizophyllus, fertile form) x Phyllitis scolopendrium var. scolopendrium Texture coriaceous chartaceous Lobation no lobes above base (with rare commonly |-several lobes above exceptions) base Areoles costal, only 1 (rarely 2), if present at _ costal, l-several near blade base Soriation Phyllitis-like, rarely otherwise. Phyllitis-like, occasionally Medi otherwise. Inframedial, 1-4 mm apa Petiole color dark-pigmented, only at base dark-pigmented, the color running Petiole scales Chromosomes numerous and dense, especially at petiole base ca. 144; 72 large ones, 72 small ones; ca. 15 pairs into lower rachis few and scattered 108; 36 large ones, 72 small ones; no pairing are obscured by inheritance from the other two parents, so that they are practically entirely free. Two unique features of Phyllitis play strong morphological roles in this hybrid: The sori are “double,” and the petiole strongly scaly and short, thus maintaining these character states in spite of the other two genomes which show only “‘single”’ sori. It is rewarding to compare the synthetic hybrid with another trigeneric combination, this one involving Camptosorus rhizophyllus, A. platyneuron (rather than A. montanum), and the diploid European variety of Phyllitis scolopendrium rather than the tetraploid American variety. This hybrid, which originated in the cultures of the late Kay Boydston at Fernwood, Michigan (Wagner, 1989), shows more lobing of the blade, as might be expected because of the lesser influence from Phyllitis, there being only one g ther than two. Probably for the same reason there are fewer scales on the petiole, which is also longer. However, the “double-sorus”’ condition is retained as a dominant feature from Phyllitis. Also, the dark leaf axis of Asplenium platyneuron is surprisingly well developed in the hybrid, considering the pale axes from the other two genomes. A summary of the major differences between the two trigeneric hybrids is given in Table 2. One of the prevalent problems in the study of fern nothospecies involves the relative expression of parental characters. Studies of various biochemical compounds, including such different entities as phenolic compounds and isoenzymes, demonstrate that the inheritance of these tend to be additive, the lectrophoreti I tographic pattern of pecies superposed upon that of the other. On the contrary, morphological characteristics, as is now well known, tend to be intermediate, i.e., somewhere between the extremes laid 6 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 1 (1989) down by the parents. Whether they show dominance of the phenotype of one parent over the other in any given trait may bear upon the interpretation on genetic controls of that morphological feature. Only rarely are all or most character states truly ‘‘medial” (Barrington’s term for being precisely “‘in the middle” between the parents, Barrington, 1986). While it is still too early to arrive at firm conclusions regarding this, it is tempting to speculate that experimental hybridization using parents with widely different morphologies may lead to possible insights concerning homologies, dominance effects, and morphogenesis. Continued work with synthesized hybrids may permit us to address questions like: Why is the ‘‘double sorus’’ of Phyllitis so strongly expressed in these hybrids? In contrast, why hasn’t the reticulate venation of Camptosorus been more strongly expressed? For those especially interested in raising and culturing ferns asa hobby, it seems to us that artificially synthesizing new crosses may serve not only as an enjoyable and stimulating pastime, but may, in the end, produce something of scientific value. Also, studies of hybrids like this one may bear on whether we recognize certain genera. Admittedly, although there is still much disagreement about this, the genera Camptosorus and Phyllitis may well be congeneric with Asplenium. Thus, whether we call the hybrids discussed here ““intergeneric’’ or “intrageneric”’ depends on our taxonomic viewpoint. ACKNOWLEDGMENTS We thank the late Dale J. Hagenah for his assistance in various ways. We also acknowledge the help of Florence S. Wagner in maki g the ch tudies, and David Bay for the photographs. Charles R. Werth g ful ti for tl ipt oo rl , Sif LITERATURE CITED BARRINGTON, D. S. 1986. The morphology and cytology of Polystichum x potteri: hybr. nov. (=P. acrostichoides x P. braunii). Rhodora 88:297—313. Lovis, J. D. 1968. Artificial reconstruction of a species of fern, Asplenium adulterinum. Nature 297:1163—1165. ; , REICHSTEIN, T. 1981. Hybrids in European Aspleniaceae (Pteridophyta). Bot. Helvetica 91:89-139. SLOssoN, M. 1902. The origin of Asplenium ebenoides. Bull. Torrey Bot. Club 29:487—495. Wacner, W. H., JR. 1954. Reticulate evolution in the Appalachian aspleniums. Evolution 8-103 118. ———. 1989. Kathryn E. Boydston (1897-1988): Michigan’s fern hvbridist of her work. Mich. Bot. 28:51—57. ; and WHITMIRE, RS. 1957. Spontaneous production of a morphologically distinct, fertile poly} y asterile diploid of Aspleni benoides. Bull. Torrey Bot. Club 84:74—89. J} exampies American Fern Journal 79(1): 7—13 (1989) Electrophoretic Evidence for Interspecific Hybridization in Polystichum PAMELA S. Sottis, Douctas E. Sottis, PAu G. WoLr, and JILL M. RILEY Department of Botany, Washington State University, Pullman, Washington 99164 Hybridization and polyploidy in Polystichum (Dryopteridaceae) have contributed to the morphological diversity and taxonomic complexity in this widespread genus. Patterns of reticulate evolution are particularly evident in the Polystichum complex from western North America (W. Wagner, 1973; D. Wagner, 1979), and several sterile interspecific hybrids have been reported (see P. Soltis et al., 1987, for review). As part of a comprehensive molecular analysis of polyploidy, its causes, and genetic consequences, we have studied two additional interspecific hybrids. In this paper we provide electrophoretic documentation of hybridization between P. andersonii and P. munitum and between P. lemmonii and P. munitum. Both hybrids were previously reported based on morphological and cytological evidence (W. Wagner, 1973). Polystichum andersonii occupies lowland coastal forests in British Columbia and southeastern Alaska, montane forests in western Washington and Oregon, and also occurs disjunctly in southeastern British Columbia, northern Idaho, and northwestern Montana. This species is tetraploid (2n= 164; W. Wagner, 1973; D. Wagner, 1979), but its ancestry is uncertain. Warren Wagner (1973) suggested that this species is of autopolyploid origin based on its distinctive leaf morphology and the presence of a vegetative bud on the rachis of the frond. However, cytological data (W. Wagner, 1973) show that P. andersonii contains a chromosome complement homologous with the common diploid P. munitum, suggesting instead an allopolyploid origin for P. andersonii. Anatomical, chromatographic, and electrophoretic data also support this hypothesis (D. Wagner, 1979; P. Soltis et al., unpubl. data). Although P. munitum is a likely candidate for one of the diploid progenitors of P. andersonii, the identity of the other parental species is currently unknown, although D. Wagner (1979) described a specimen from northern British Columbia that seems to fit the predicted morphology of the second progenitor. Polystichum lemmonii (2n= 82) occurs on open, serpentine, montane slopes in northern California, southwestern and central Oregon, and central Washington. Following D. Wagner (1979), P. lemmonii is herein considered distinct from the South American P. mohrioides. Polystichum munitum (2n=82) inhabits moist coniferous forests from California to southeastern Alaska, with disjunct populations in eastern Washington, northern Idaho, northwestern Montana, and northeastern Oregon. Populations of P. munitum are often very large, consisting of thousands of individuals. This species is highly outcrossing and experiences significant interpopulational gene flow (P. Soltis & D. Soltis, 1987). Thirty to forty plants morphologically intermediate between P. andersonii and P. munitum were observed on Deer Peak near Ketchikan, Alaska. All mature 8 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 1 (1989) plants possessed the distinctive vegetative bud of P. andersonii, yet the fronds were less dissected than typical leaves of P. andersonii, suggesting the influence of P. munitum or another once-pinnate species. Because P. munitum occurs in this region, we hypothesized that these morphologically intermediate plants were hybrids between P. andersonii and P. munitum, although neither putative parental species was found in the immediate vicinity of the hybrid population. A single plant morphologically intermediate between P. Iemmonii and P. munitum was detected in the Beverly Creek drainage of Kittitas County, Washington. Fronds were comparable in size to those of P. munitum, yet showed evidence of dissection, similar to that observed in the bipinnate P. lemmonii. Both P. lemmonii and P. munitum occur in the Beverly Creek drainage, although they occupy different habitats. Thus it seemed likely that the unusual morphology resulted from hybridization between these two diploid species. MATERIALS AND METHODS Plants.—Twenty-eight putative hybrids between P. andersonii and P. munitum were examined electrophoretically (see Table 1 for collection data). Individuals from four populations of P. andersonii were also examined. Individuals from three populations of P. munitum were analyzed alongside the putative P. andersonii X munitum hybrids, and additional data from several other populations were also used for comparisons (P. Soltis & D. Soltis, 1987; P. Soltis et al., unpubl. data). The single putative hybrid between P. lemmonii and P. munitum was compared electrophoretically to individuals of both parent species from the Beverly Creek drainage and to plants from other populations of both species (P. Soltis & D. Soltis, 1987: P. Soltis et al., unpubl. data). Leaf material was collected in the field and stored in plastic bags under refrigeration until electrophoresis was conducted. Collection data and sample sizes for all taxa are provided in Table 1. Vouchers were deposited at WS. rs on ay, } e: CANT £. D 1.42 fp Lethe? f TABLE 1. olystichum Species and Hybrids Examined Electrophoretically. P. andersonii.—Alaska: Juneau, West Glacier Trail, Mendenhall Glacier, Soltis & Soltis 1785 (22). Oregon: Linn County, Pamelia Lake, Mt. Jefferson Wilderness, Soltis & Soltis 1904 (7). British Columbia, Canada: 4 mi NE of Exchamsiks Provincial Park, Soltis & Soltis 1768 (14); 2-3 mi E of Stewart on Hwy. 37A, Soltis & Soltis 1800 (13). P. lemmonii.—Washington: Kittitas C , Beverly Creek Trail, Wenatchee National Forest, Idaho: Benewah County, 3.1 mi SW of Emida along Hwy. 95, Soltis & Soltis 1844 (5); N of St. Maries along Hwy. 5, Soltis & Soltis 1845 (7). Washington: Kittitas County, Beverly Creek Trail, Wenatchee National Forest, Soltis, Soltis, & Riley 1972 (1). -andersonii x munitum.—Alaska: Revillagigedo I Soltis 1770 (28). 1, Deer Peak Trail, Ketchikan, Soltis & TAT abs. enced, + ’ ie + Rex y; rly Creek Trail, Wenatchee o P, lemmonii Kitti National Forest, Soltis, Soltis, & Riley 1977 (1). P. S. SOLTIS ET AL.: POLYSTICHUM HYBRIDS 9 Electrophoresis Electrophoretic procedures generally followed D. Soltis et al. (1983). Leaf tissue was prepared using the tris-HCl grinding buffer of D. Soltis et al. (1983); 12% PVP was used (wt/vol). Starch gel concentration was 12.5%. Nine enzymes were examined: aspartate aminotransferase (AAT), NAD- dependent glyceraldehyde 3-phosphate dehydrogenase ([NAD]G3PDH)}, leucine aminopeptidase (LAP), malate dehydrogenase (MDH), phosphoglucoisomerase (PGI), phosphoglucomutase (PGM), 6-phosphogluconate dehydrogenase (6PGD), shikimate dehydrogenase (SkDH), and triosephosphate isomerase (TPI). AAT, LAP, PGI, and TPI were resolved on gel and electrode buffer system 8 of D. Soltis et al. (1983) as modified by P. Soltis et al. (1987). MDH, PGM, and SkDH were resolved on system 9 of D. Soltis et al. (1983); [NAD]G3PDH and 6PGD were resolved using system 1 (D. Soltis et al., 1983). Staining for all enzymes except LAP followed D. Soltis et al. (1983); staining for LAP followed D. Soltis and Rieseberg (1986). The genetic control of the enzyme banding patterns was easily inferred based on the typical subunit struct nd subcellul part talization of the en- zymes (Gottlieb, 1981, 1982). Isozymes were numbered sequentially, with the most anodally migrating isozyme designated 1. Allozymes were denoted alpha- betically, with the farthest migrating allozyme designated a. RESULTS Polystichum andersonii X munitum.—Thirteen loci were interpreted: Aat, G3pdh-1, G3pdh-2, Lap, Mdh-1, Pgi-2, Pgm-1, Pgm-2, 6pgd-2, Skdh, Tpi-1, Tpi-2, and Tpi-3. These loci designations reflect an apparent gene duplication for TPI (P. Soltis & D. Soltis, unpubl. data). Pgi-1 and 6pgd-1 were not clearly resolved in all samples and were therefore omitted from the analysis. All enzymes migrated anodally. Polystichum andersonii and P. munitum shared alleles at all loci, although distinct differences in allele frequencies were observed for Pgi-2 (Table 2). However, P. and ii exhibited fixed heterozygosity at Skdh and a five-banded pattern in the more anodal zone of activity for TPI. It is unclear whether this five-banded pattern represents fixed heterozygosity at Tpi-1 or Tpi-2 or both of these loci. Although the parental taxa are similar allozymically, electrophoretic data support the hypothesized hybrid origin of the plants from Deer Peak. The putative hybrids clearly combined the genotypes of P. andersonii and P. munitum at Pgi-2 (Fig. 1). All individuals of P. andersonii possessed Pgi-2a; this allele was in low frequency in P. munitum. The most common allele in P. munitum was Pgi-2c. The putative hybrids all displayed the heterozygous genotype Pgi-2ac, and most individuals clearly displayed unbalanced staining, with two doses of allele a and one of allele c (Fig. 1). Such dosage effects would be expected in a triploid hybrid between an allotetraploid and a diploid and have been reported for other triploid hybrids (e.g., P. Soltis et al., 1987). The hybrids also exhibited a fixed heterozygous pattern for Skdh identical to that of P. andersonii (Fig. 2) and the complex banding pattern of P. andersonii for TPI. The 10 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 1 (1989) Paes. a | pu gnig | ander TABLE 2. Allele Frequencies in P. ti . lemmonii, and P. munitum Included in this Study. Population numbers correspond to those i in Table 1 P. andersonii P. lemmonii P. coe cham Population Population Populatio Locus/Allele 1785 1904 1768 1800 1974 1844 1845 sea 1972 Aat a 1.0 1.0 1.0 1.0 0.985 1,0 1.0 1.0 1.0 b 0.0 0.0 0.0 0.0 0.015 0.0 0.0 0.0 0.0 G3pdh-1 a 1.0 a Lo a 1.0 10 1.0 1.0 a G3pdh-2 a 0.0 — 0.0 — 1.0 0.0 0.0 0.0 0.0 b LO —_ 16 — 0.0 1.0 1.0 10 1.0 Lap a 0.0 0.0 0.0 0.0 0.0 —_ — 0.240 — b 0.0 0.0 0.0 0.0 0.882 a — 0.0 — Cc 1.0 1.0 1.0 1.0 0.118 — a 0.760 — Mdh-1 a 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.036 0.0 b 1.0 1.0 1.0 1.0 1.0 1.0 1.0 0.964 1.0 Poi-2 a 1.0 1.0 1.0 7.0 0.971 0.100 O. ne Ott - 00 b 0.0 0.0 0.0 0.0 0.015 0.0 0.0 c 0.0 0.0 0.0 0.0 0.015 0.800 0 os 0.648 1.0 d 0.0 0.0 0.0 0.0 0.0 0.10 0.148 0.0 e 0.0 0.0 0.0 0.0 0.0 0. om 0.092 0.0 Pgm-1 a 1.0 1.0 Lo 1.0 1.0 1.0 1.0 0.940 1.0 b 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.060 0.0 Pgm-2 a 0.0 0.0 0.0 0.0 1.0 0.100 0.143 0.020 0.500 b 1.0 1.0 1.0 1.0 0.0 0.900 0.857 0.980 0.500 6pgd-2 a 0.0 0.0 0.0 —_— 0.129 0.100 0.071 0.0 b 1.0 1.0 1.0 —_ 0.871 0.200 0.571 0.643 1.0 c 0.0 0.0 0.0 —_ 0.700 0.357 0.357 0.0 Skdh * * * * a 1.0 1.0 1.0 1.0 1.0 Tpi-1 a 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 Tpi-2 * * * * a 0.939 0.0 0.0 0.0 0.0 b 061 1.0 1.0 1.0 1.0 Tpi-3 a 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 *denotes fixed heterozygous patterns. —indicates no data available. P. S. SOLTIS ET AL.: POLYSTICHUM HYBRIDS 11 @) @ @) ss @- - - — = _»—a << 22% 73 c aeeeee p< M MHH HHH HHH MMMHHH H MHLL Be gg a ee designat Fic. 1—3. Starch gels of Polystichum. In Figs. 1 and 3 allozymes. In all wig — below photographs designate ception: oe P, leshecweiil: Ms P, munitum, H=hybrid. . PGI in P. munitum and P. andersonii X munitum. Polystichum andersonii (not ae ani allele a. Note unbalanced staining in hybrids. Fic. 2. SkDH in P. munitum an rsonii X munitum. Hybrids possess the two-banded fixed heterozygous pattern of P. andersonii (not shown). Fic. 3. PGI in P. lemmonii, P. munitum, and their hybrid. Fic. 4. Mitotic chromosome squash of triploid P. andersonii x munitum (Soltis & Soltis 1770). 2n=123. Magnification is 1,000 x. putative hybrids could not be distinguished from either parent at Aat, G3pdh-1, G3pdh-2, Lap, Mdh-1, Pgm-1, Pgm-2, 6pgd-2, and Tpi-3. Chromosome counts of the putative hybrids provide further evidence for their interspecific hybrid origin because all individuals were triploid, with 2n=123 (Fig. 4). Polvetiohenas lemmonii X munitum.—Eleven loci were interpreted: Aat, G3pdh-2, Mdh-1, Pgi-2, Pgm-1, Pgm-2, 6pgd-2, Skdh, Tpi-1, Tpi-2, and Tpi-3. G3pdh-1 and Lap were not clearly resolved in the putative hybrid. Polystichum lemmonii and P. munitum differed in all populations examined for G3pdh-2 and Tpi-2 (Table 2). The putative hybrid clearly combined the 12 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 1 (1989) genotypes of P. lemmonii and P. munitum at these two loci. For 6pgd-2, P. lemmonii exhibited alleles a and b. In P. munitum allele a was in low frequency in two populations; alleles b and c were present in higher frequencies. The putative hybrid possessed the heterozygous genotype 6pgd-2bc, consistent with the hypothesis of interspecific hybridization. Data for PGI also support hybridization between P. lemmonii and P. munitum. Although P. lemmonii and P. munitum share the allele Pgi-2a, this allele was in low frequency in P. munitum and was not detected in the population of P. munitum from the Beverly Creek drainage. The hybrid exhibited Pgi-2a (typical of P. lemmonii) and Pgi-2c (typical of P. munitum), also consistent with the hypothesis of interspecific hybridization (Fig. 3). The parental species could not be differentiated consistently at any of the other loci. Such allozymic similarity between congeneric fern species is unusual (D. Soltis & P. Soltis, 1989) and may reflect relatively recent radiation in this species complex. DISCUSSION Electrophoretic data provide clear documentation of interspecific hybridization between Polystichum andersonii and P. munitum, and between P. lemmonii and P. munitum. Morphological intermediacy, additivity at one or more enzyme loci, and a triploid chromosome count for P. andersonii X munitum confirm the hybrid origin of both putative hybrids. Although only a single P. lemmonii x munitum individual was observed, hybrids of P. andersonii and P. munitum were thriving and apparently reproducing on Deer Peak. Whether this resulted from sexual or asexual reproduction requires further study. The absence of both P. andersonii and P. munitum from the Deer Peak site coupled with the vitality of the P. andersonii X munitum population suggests that this hybrid was not formed on Deer Peak but originated elsewhere and is capable of spreading via either sexual or asexual reproduction. Interspecific hybridization in western North American Polystichum is both taxonomically and geographi lly widesy 1 and apparently occurs frequently when two species occur sympatrically. This suggests that prezygotic isolating barriers are not well established in this species complex. In fact, several factors may actually promote interspecific hybridization. All species of Polystichum analyzed to date, including polyploids, are outcrossing (P. Soltis & D. Soltis, 1987; P. Soltis et al., 1988; D. Soltis & P. Soltis, 1987). Polystichum munitum is highly outcrossing: rates of intragametophytic selfing range from 0 to 3% (P. Soltis & D. Soltis, 1987). Polystichum lemmonii is also outcrossing, with gametophytic selfing estimates f six populations ranging from 0 to 18% (P. Soltis and D. Soltis, unpubl. data). Furthermore, F, the fixation index (Wright, 1965), indicates only slight deviation from Hardy-Weinberg genotypic expectations in all populations of P. lemmonij (P. Soltis and D. Soltis, unpubl. P. S. SOLTIS ET AL.: POLYSTICHUM HYBRIDS 13 coupled with poorly developed isolating mechanisms in Polystichum may in large part determine the high incidence of interspecific hybridization. The frequency of interspecific hybridization in Polystichum may also have significant implications for the evolutionary history of the numerous allopolyploid species in Polystichum. It seems likely that these allopolyploids may be polyphyletic, having arisen several times throughout their geographic ranges. For example, P. lemmonii and either P. munitum (W. Wagner, 1973) or P. imbricans (D. Wagner, 1979) are considered the diploid progenitors of the allotetraploid P. scopulinum (W. Wagner, 1973; D. Wagner, 1979). The hybrid P. lemmonii X munitum documented herein occurs sympatrically with P. lemmonii, P. munitum, and scattered individuals of P. scopulinum. It is therefore possible that P. scopulinum from the Beverly Creek drainage of central Washington arose in situ. Other populations of P. scopulinum from throughout its range may also represent independent origins and different genotypes. Evidence for multiple origins of Polystichum allopolyploids is currently being sought using isozyme and DNA markers. The significance of rampant hybridization in Polystichum extends beyond the production of typically sterile hybrids and taxonomic complexity. Frequent interspecific hybridization provides the opportunity for multiple allopolyploid events and may represent an important force in the history of allopolyploid species. ACKNOWLEDGMENTS We thank Bryan Ness for technical assistance and Dan Crawford and David bagartas = helpful comments on the manuscript. This work was supported by NSF grant BSR-8620444 an nd a grant-in-aid from the Office of Grant and Research Development, Washington State ees iby LITERATURE CITED z: \ id ] GOTTLIEB, - Me 1981. came tea sielsconsiedey 1 plant lati 8. Prog oes Smeaecore cf 1-45. 2. Conservatio SOLTIs, D. ag e H. HAULER, D. C. Darrow, Ts I. GasTONY. 1983. ewe gel electrophoresis of ferns: A Amer. Fern J. 73: and L. H. oe 1986. deg ctned sae lh in Tolmiea menziesii: Genetic insights from enzyme electrophoresis. Amer. J. Bot. 73:310—318 and P. S. SoLtis. 1987. Hirsute 1 breeding syst ft Pteridophyta: A £ LIS ie rea Amer. Naturalist 130:219-—232. _______. 1989. Polyploidy, breeding systems, and genetic differentiation in oe oahu In Isozymes in plant biology, ed. D. E. Soltis and P. S. Soltis. Portland: Dioscorides Pres Sottis, P.S. and D.E. Sottis. der, Population structure and eae of gene flow in the homosporous fern Polystichum munitum. Evolution 41:620— ————, an nd E. A. ALVERSON. 1987. Electrophoretic and morphological confirmation of fi munitum. Amer. Fern ’ J. 77: 42-49, —__——., and K.E. HOLSINGER. 1988. Estimates of =pliygietarie deems selfing and ntepopulatina gene flow in homosporous ferns. Amer. J. Bot. 75:1765—-17 WAGNER, es H. Ae of Polystichum in western North America elit ‘of Mexico. Wciacick L WacneR, W. H.., JR. set aedataie f holly ferns (Polystichum) in the western United States and adjacent Canada. Amer. Fern J. 63:99-115. American Fern Journal 79(1): 14-25 (1989) New Species of Ceradenia subg. Ceradenia - ld 46 L. EARL BISHOP # ovo ¥4 t- Herbarium, Dept. of Botany, University of California, Berkeley, California, 94720 When establishing the genus Ceradenia (Bishop, 1988), I noted that many taxa were yet to be described. Since the number of novelties will be a surprise to some workers, a few words of explanation seem appropriate. First of all, until recently no group of Neotropical Grammitidaceae has been studied thoroughly using all the nomenclaturally important material. Furthermore, it is now apparent that different species may be superficially quite similar even though they belong to distantly related groups. Asa result, the application of some nineteenth century names has been somewhat hazy, for most of the work in this century on these ferns has been done in the New World whereas the types for the older names are in Europe. And of course the descriptions published for these older names have been of little help to modern students. Therefore, specimens of undescribed species have lain in herbaria unrecognized because of the difficulty in ascertaining the precise application of published names. Lastly, these ferns appear to be inadequately collected and many are known only from their types. Parris’ careful study of the genus Grammitis in New Guinea (1983) provides an illustrative case; of 64 species accepted 21 were new. Subgenus Ceradenia has been the easier of the two in which to elucidate the included taxa. Subgenus Filicipecten includes the large C. kalbreyeri- meridensis group, which in complexity and similarity of species seems analogous to the Polypodium vulgare complex of the Northern Hemisphere. The novelties of this subgenus will be treated later. The eight new species in subgenus Ceradenia here described fall into three species alliances. Five are related to C. capillaris, a widespread species that erenow had but one closely allied described species. A single species falls with the small group of species near C. pilipes. The two remaining species, which are the first terrestrial, erect ones in the subgenus, are closely related to C. herrerae, which in turn seems allied to C. albidula. a pluricellular, unbranched, acicular trichome with thickened cell walls (the multicellular nature is not normally discernible except under higher L. E. BISHOP: CERADENIA 15 Unless otherwise noted, I have examined all specimens cited. Loans were obtained from AAU, B, BM, F, GH, K, MO, NY, P, and US, and I thank the curators for making these specimens available. Ceradenia dendrodoxa L. E. Bishop, sp. nov. (Fig. 1A)—TyPE: Peru, Amazonas, Pcia. Chachapoyas, Cerros de Calla-Calla, near Kms 403-407 of Balsa- Leimebamba road, on uppermost slopes and summit, pendent from tree branches in moist ravine, occasional, 3400-3550 m, 18 Aug 1962, Wurdack 1715 (holotype UG, isotypes F, NY, US). Ab altitudine excelsa haec filix amoena plerumque epiphytica pendulaque oritur. Rhizoma est breve ramosum caespitosum paleis castaneis lineari-oblongis vel lineari-triangularibus, basi truncatis vel subcordatis apice abrupte vel acuminate angustatis 2.0—4.5 x 0.2-0.4 mm ciliis nullis autem glandulis marginalibus caducis, cellulis medialibus 50-140 x 25-35 pm. Frondium sunt stipites nigri in senectute brunnei glandulosi esetosi teretes 0.3—0.5 mm lati 2—6 cm longi rhachides similiter teretes nigrae glandul tis carentes rectae vel flexuosae, laminae perpinnatae usque ad 75 cm longae pinnis sub angulo 25—40° a rhachide abeuntibus linearibus marginibus propter expansionem laminae circum soros repandis ad basim aut ad costam ab utroque latere contrictis aut rectis sine contrictionibus apice rotundatis vel acutis pilis glandulosis in paginis ambabus dispersis 1-4 cm x 1.0—1.5 mm costae sclerenchymate nigro dorsaliter clare evidenti, venis simplicibus aut furcatis, stomatibus 50-60 x 44—50 pm. Sori usque ad 20 paria in quaque pinna capsulis 155-170 X 135 —142 pm annulis ex 10-12 cellulis constantibus illis cellulis distalibus 28-34 ym altis sporis subglobosis vel hemisphaericis 27-32 .m in diametro longiore sub maturitate marginem excedunt. Haec species tam grandis et insignis est certe gloria cui jue arboris quam forsitan incolat. Paratypes: ECUADOR. Azuay. Rio Collay, slopes of Huagrarancha, S of El Pan, 2650-3290 m. Steyermark 53380 (F, US). Loja? Horta-Naque, 3600 m, Espinosa 1021 (NY, US). PERU. Haanuco. Tambo de Vaca, 13000 ft., Bryan 626 (F, US). This splendid, high-elevation species is known from a relatively wide range through southern Ecuador and northern Peru. The black rachis and the narrow pinnae that never show prolonged growth ally it to the C. capillaris group. From the widespread C. capillaris itself, which occupies the same range at lower elevations, it differs in its completely pinnate fronds. It appears most closely related to C. praeclara of central Peru, but that species has much wider pinnae whose margins the sori do not exceed. Both C. praeclara and C. auroseiomena bear setae on the stipe and rachis; such setae are absent in C. dendrodoxa. The insertion angle of the pinnae onto the rachis is narrower in C. dendrodoxa than any related species. On the basis of the few gatherings at hand, the Ecuadorian plants differ somewhat from the Peruvian population. These northern specimens have the pinnae smaller, of thinner texture, and more regularly constricted at the base. Moreover, the rachis is thinner and is more sharply, conspicuously flexuous. 16 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 1 (1989) Ceradenia comosa L. E. Bishop, sp. nov. (Fig. 1B)—Typk: Bolivia, Cocopunco, 10,000 ft, 24-29 Mar 1926, Tate 337 (holotype NY, isotype US). Haec filix delicatula manifeste pendula verisimiliter in arboribus epifitice viget. Rhizoma secundum exemplum unicum praesens est minus breve paulo ramosum, paleis parvis atrocastaneis anguste triangularibus ad basin plerumque pallidum truncatis vel subcordatis apice acuminatis 1.5-2.5 x 0.1—-0.3 mm marginibus per maximam partem ciliatis pilis concoloribus vel pallidioribus sed non hyalinis, cellulis medialibus 28-35 ym latis et 3-5plo longioribus quam latioribus. Frondium sunt stipites teretes brunneoli (fortasse corylinus) setis exilibus castaneis 0.5-1.5 mm per longitudinem totam etsi ad basin confertioribus etiam sub juvente pilis parvis hyalinis 1—3-furcatis praediti 0.2—0.3 mm lati 5-10 cm longi, rhachides stipitum similes sed setis carentes sub juvente glandulis dissitis instructae demum eis obscuris, laminae perpinnatae 22—40 cm longae base multum angustatae pinnis linearibus sub angulo 40—70° a rachide abeuntibus irregulatim elongatis usque ad 15 cm longis 0.7—1.5 mm latis repandis vel dentatis dentibus acutis antrorsis basi non constrictis basiscopice d tik pice truncatis vel rotundatis primum glandulas parvas in paginis ambabus exigue ferentibus demum his glandulis vix visibilibus in margine aliquot pilis hyalinis dissitis 1—3-cellulatis praeditis venis regulatim simplicibus sed aliquando furcatis dorsaliter costa paulo evidenti sed sclerenchymate suo haud exposito stomatibus 54-62 x 45-54 ym. Sori usque ad 20 paria in quaque pinna sub maturitate marginem paulo excendentes capsulis oblongis vel subglobosis 152-168 x 132-145 wm annulis ex 12—13 cellulis constantibus illis cellulis distalibus 30-36 um altis sporis subglobosis vel hemisphaericis 26—32 4m in diametro longiore in laminae expansionibus vel dentibus medialiter geruntur. Hanc speciem gracilem ob aspectum capillarem et pendulum nomino. praeclara) by its brown rachis. This feature seems to ally it more closely to the C. pilipes group. The rather irregularly elongate pinnae and the very much reduced, barely evident basal pinnae are also characteristic of these species. Among this group of species, which also includes C. fucoides and C. podocarpa, C. comosa is the only one with fully pinnate fronds. from Callana, Dpto. La Paz, Pcia. Larecaja. However, this locality is at 1400 m and there are no mountains of the given height within at least 25 km. Ceradenia praeclara L. E. Bishop, sp. nov. (Fig. 1C)—Type: Peru, Ayacucho, Cochapata, Valle de San Miguel, La Convencién, 10,300 ft, Sep 1934, Biies 2179 (holotype US). L. E. BISHOP: CERADENIA 17 i 4 ae Fic. 1. A, Ceradenia dendrodoxa, Wurdack 1715, UC. B, C. comosa, Tate 337, NY. C, C. praeclara, Biies 2179, US. D, C. auroseiomena, Little 9363, US. 18 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 1 (1989) Exemplum unicum huius filicis regionis altae novi. Rhizoma mihi praesens simplex, paleis parvis castaneis lineari-oblongis basi truncatis ad apicem angustatis 1-2 x 0.1-0.2 mm sub juvente glandulis marginalibus praeditis his demum caducis ciliis marginalibus carentibus cellulis medialibius 20-30 um latis et 2-4plo longioribus quam latioribus. Frondium pendularum sunt stipites teretes nitentes nigri sub senectute brunnei glandulis numerosis necnon setis debilibus 1-2 mm longis praediti 0.3-0.5 mm lati 1—3 cm longi, rhachides quoad indumentum aspectum coloremque stipitibus similes, laminae perpinnatae 15—30 cm longae basi paulo angustatae pinnis lineari-oblongis vel lineari-triangularibus repandulis sub angulo 50—70° a rhachide abeuntibus usque ad 25 mm longis 3.0—4.5 mm latis basi basiscopice paulo decurrentibus acroscopice conspicue constrictis margine hic ad thachidem parallo apice rontundatis vel obtusis in paginis ambabus glandulis uberius dissitis venis simplicibus perve occasionem 1-furcatis dorsaliter costae sclerenchymate nigro manifeste exposito stomatibus 48-56 x 44—52 .m ventraliter costa prominula. Sororum usque ad 15 paribus sub maturitate marginem haud vel vix attengentium capsulis subglobosis vel late obpyriformibus 145-160 x 130-145 #m annulis ex 12-14 cellulis constantibus illis cellulis distalibus 25—30 pm altis sporis hemisphaericis vel subtetraedricis 25-30 um in diametro longiore quaeque pinna medialiter vel paulum inframedialiter instruitur. Haec species insignis propriaque ut mihi videatur epithete praeclaro digna est. This attractive, distinctive species is known to me only by a single sheet from central Peru. It seems most closely related to C. dendrodoxa from northern Peru and southern Ecuador. From that species it differs in its much broader, more Ceradenia auroseiomena L. E. Bishop, sp. nov (Fig. 1D)—Type: Colombia, Caqueté, hanging down from wet limb, wet temperate forest, Caqueta side of Huila—Caqueté divide, Cordillera Oriental, 20 km SE of Garzén, 7800 ft, 2 Feb 1945, Little 9363 (holotype US). L. E. BISHOP: CERADENIA 19 Unum specimen solum huius filicis gracilis epiphyticae mihi adest. Rhizoma unicum praesens simplex paleis atrocastaneis lineari-oblongis vel lineari- triangularibus 1.0—2.5 x 0.2—0.4 mm basi pallidioribus truncatis vel cordatis marginaliter primo glanduliferis postea integris, cellulis medialibus 25—35 pm latis et 1—3plo longioribus quam latioribus. Frondium pendularum sunt stipites teretes ubi juvenes nigri sub maturitate brunnei aliquot glandulis setisque pluribus castaneis 1-2 mm longis praediti 0.2—-0.3 mm lati 4—8 cm longi, rhachides stipitibus similes sed nigrae sub maturitate permanentes solum ad senectutem brunneae praeterea setas pauciores praebentes, laminae perpinnatae 2—4 dm longae basi angustatae pinnis repandulis lineari-triangularibus sub angulo 50—70° a rhachide abeuntibus usque ad 35 mm longis 2—3 mm latis basi p apice tis in paginis ambabus glandulas dispersas gerentibus venis simplicibus (rare 1-furcatis) dorsaliter costa prominula autem sclerenchymate suo non exposito stomatibus 44—50 x 44—50 um ventraliter costa vix evidenti. Sororum usque ad 12 paria sub maturitate marginem vix attengentium capsulis subglobosis vel obpyriformibus 142—160 x 120—140 ym annulis ex 12—13 cellulis constantibus illis cellulis distalibus 26-32 ym altis sporis subglobosis vel hemisphaericis 22—28 wm in diametro longiore quaeque pinna medialiter fert. E graecor avea, ventulus, et oe.ouevn, tremefacta, hoc epitheton pro tali specie subtili pendulaque stipitibus gracilibus contraxi. This species, known from a single specimen, apparently grows at somewhat lower elevations than its nearest relatives to the south, C. dendrodoxa and C. praeclara. From these it differs in its dark brown scales and in its acuminate pinnae that show no exposure of the costal sclerenchyma dorsally and that are slightly surcurrent acroscopically at the base. Ceradenia praeclara, the more similar of these two species, has the pinna base deeply contracted acroscopically, in addition to its rounded, broader pinnae with prominently exposed costal sclerenchyma. Ceradenia dendrodoxa has no setae on the rachis or stipe, has narrower, linear pinnae with exposed costal sclerenchyma, and has sori that exceed the laminar margin at maturity. It may be that the closest relationship of C. auroseiomena is with C. phloiocharis of Central America. This latter species is more gracile, with ciliate scales and with narrower, linear, more deeply repand pinnae (1-2 mm wide) whose sori regularly attain or exceed the margins at maturity. Ceradenia phloiocharis L. E. Bishop, sp. nov. (Fig. 2A): : ye a he Fic. 2. A, Ceradenia phloiocharis, Gémez et al. 22372, UC. B,C.m irabilis, Brooke 6134, BM. CG. terrestris, Wurdack 1643, US. D,C. maxoniana, Lehmann 2400, US. instructi 0.2—0.3 mm lati 2-8 cm lon sub maturitate permanentes solum ad senect 5 cm longae basi gradatim us sub angulo 50—80° a rhachide abeuntibus mm latis 2—4plo latitudine sua inter sese angustatae pinnis repandis linearib usque ad 4 cm longis 1~2 (2.5) L. E. BISHOP: CERADENIA 21 disjunctis basi basiscopice decurrentibus acroscopice paulo surcurrentibus apice acuminatis in paginis ambabus glandulas dispersas ferentibus venis vulgo simplicibus interdum 1-furcatis dorsaliter hic illic setiferis costa venisque prominulis ac sclerenchymate costali visibili sed rare exposito stomatibus 44-52 x 38-44 um ventraliter costa venisque prominulis. Sori in quoque segmento usque ad 15 paria sub maturitat gi tteng vel excedentes capsulis late obpyriformibus 144—160 x 115—135 ym annulis ex 10—12 cellulis constantibus illis cellulis distalibus 28-34 ym altis sporis subtetraedricis vel hemisphaericis 25—30 ym in diametro longiore medialiter feruntur. E graeco ¢dAotog, cortex, et yaous, venustus, epitheton f speciem elegantem epiphyticam tali modo celebrarem. Paratypes: COSTA RICA. Limon. Cordillera Talamanca, between Rio Sini, 2-4 km W of Panama- nian border, 2300—2500 m, Davidse, Herrera, & Grayum 28979 (MO, UC), 28980 (MO). PANAMA. Bocas del Toro. Cordillera de Talamanca, 4 km NW of main peak of Cerro Fabrega, 3000-3150 m, Davidse et al. 25393 (MO). Headwaters of Rio Colubre, 2400—2550 m, Gomez et al. 22353 (MO), 22430 (MO). This elegant, slender species is apparently fairly common at appropriate elevations in a small area of the Atlantic slope near the Costa Rican—Panamanian border. The delicate fronds with linear, acuminate pinnae are quite distinctive. The two most closely related species seem to be C. auroseiomena of Colombia and C. nubigena. The former differs in its eciliate rhi scales and in its wider (2—3 mm) pinnae that are broadest at the base (linear-triangular) and only lightly repand. Ceradenia phloiocharis has linear pinnae 1-2 mm wide that are repand to the point of being subsinuately lobed. Ceradenia nubigena is a moderately variable species very nearly restricted to the summit area of Blue Mountain on Jamaica. The pinnae are generally more than 2 mm wide, they are normally rather abruptly rounded at the apex, they are usually separated on the rachis by no more than their own width, and the rachis is regularly and at least very narrowly alate between the pinnae. The pinnae of C. phloiocharis are always acuminate on mature fronds (though perhaps rounded when juvenile and sterile), they are usually separated on the rachis by two or more times their own width, and the rachis is completely terete between the pinnae. The scattered but fairly regular presence of setae on the lamina of C. phloiocharis is unique among the species of the C. capillaris group. — Two Costa Rican collections from the province of San José merit mention here. Both have black rachides that are quite exalate between the pinnae, but I believe neither to be conspecific with C. phloiocharis. Valerio 53 (US, 2 sheets) from Volcan Barva (Barba) is a collection mixed with C. fucoides, but the plants of interest here are relatively robust and have large pinnae that are sharply dentate. This frond pattern is closely matched by at least one of the many sheets of C. nubigena at hand (Sherring s.n., US). Like typical C. nubigena and C. phloiocharis, the rachis bears scattered setae. Since the rachis ala in C. nubigena may at times be scarcely detectable, I prefer at this time to refer this collection to that species, realizing that as such it represents the only record outside the Jamaican locality. The other specimen from central Costa Rica (Cerro Chiripo, Evans & Lellinger 78, US) is also a mixed collection, with two detached fronds of Grammitis 4 est ut hanc 22 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 1 (1989) jamesonioides (Fée) C. Morton and one plant of interst here. This has abruptly rounded or acute pinnae that are rather small, though well within the size range of C. nubigena. However, the rachis is devoid of setae, which is at least not typical of the Jamaica species. More critically, the base of the pinna on the basiscopic side is contracted completely to the costa before the latter’s insertion onto the rachis. This last character I have seen only in the Ecuadorian population of C. dendrodoxa. I strongly suspect this plant represents yet another undescribed, probably localized species likely related to C. capillaris, which likewise lacks setae on the rachis but which does not occur in Mesoamerica. However, because of the small distinctions involved, I prefer not to establish a new species based on this specimen until a greater range of material can be obtained. Ceradenia mirabilis L. E. Bishop, sp. nov. (Fig. 2B)—TyPe: Bolivia, Cochabamba (?), Carmen, 5-10 mi down the valley from Choro, 17° S, 66° 50’ W, hanging from a tree in warm wet forest, 8000 ft, 11 Feb 1950, W. M.A. Brooke 6134 (holotype BM, isotypes NY, US: U not seen). Filix subtilis quae in altitudinibus mediocribus epiphytice crescit. Rhizoma simplex paleis palidis linearibus 0.5—1.0 x 0.03—0.1 mm tantum 1—3 cellularum latitudine ubi juvenibus glandulis numerosis marginaliter praeditis, cellulis 40-80 x 35-50 wm. Frondium pendularum sunt stipites capillares teretes sub juventute nigri mox brunnescentes glandulis multis displicatis setis laminae bipinnatae 15—30 cm longae ad basim de medio angustatae pinnis sublineares sub angulo 30—60° a rhachide abeuntibus basi parallela decurrentibus, usque ad 25 mm longis et 8 paribus pinnulis spathulatis antrorsis acutis obtusis vel bilobatis 2 x 1 mm in pagina utraque glandulas albidis adpressis ut videtur ex 2—4 cellulis constantes gerentibus vena in quaque pinnula simplici aut furcata dorsaliter sclerenchymate costali pinnae plerumque exposito stomatibus 36-45 < 32-42 um. Sorus in quaque pinnula unicus sub maturitate latitudinem segmenti excedens capsulis globosis vel late obpyriformibus 172-200 x 160-180 m annulis ex 12-14 cellulis constantibus illis cellulis distalibus 28-32 1m altis sporis subglobosis vel sub tetraedricis 36—44 wm in diametro. tag, 11.2.50. The US sheet gives the elevation as 6000 ft, which will be considered in connection with the collecting locality. All the specimens at some place bear the collection number as 6134a, though the letter has been crossed out on all except the US sheet. At this time I assume the corrected number to be the accurate one. L. E. BISHOP: CERADENIA 23 Two of these sheets posit that a new species may be represented. Of the three other determinations suggested, Polypodium pozuzoense, P. microphyllinum, and P. pseudocapillare, only the first represents a congeneric species (P. pozuzoense = C. pilipes). The precise locality of collection is open to question. There is a small town of Choro or El Choro 2 km NE of Cocapata, Dpto. Cochabamba, 2500 m, at 16° 56’ S, 66° 42’ W. This is clearly close enough to the coordinates given to be almost certainly the Choro referred to. Carmen is a common geographical name in Bolivia; one gazeteer cites 54 such names within the country. Unfortunately none are close to the locality under consideration. Also, just which valley is meant on the label depends on the actual distance from Choro. The Rio Cocapata flows past the town of that name WNW about 10 km to join with the Rio Ayopaya (Rio Inquisivi). After this confluence the river flows north as the Rio Cotacajes. About 40 km along this last river is the Arroyo Carmen, but this is patently too distant from Choro to be the locality mentioned on the label. Also it is puzzling, if the collection site were some miles down the valley, as to why the much better known settlement of cocapata was not used as the reference point. In any case, we would expect at this distance down the valley, the elevation would be significantly lower than that of Choro (8200 ft), so that the elevation given on the US sheet (6000 ft) may be the correct one. This is the only species of the genus showing bipinnate fronds. The filiform paleae are likewise singular in that they are regularly only one or two cells wide. Otherwise, C. mirabilis fits quite well within the group of C. capillaris, which now also includes C. nubigena, C. phloiocharis, C. auroseiomena, C. dendrodoxa, and C. praeclara. Among these species the nearest relative would seem to be C. dendrodoxa. This is the only other species of the alliance that lacks setae on both the stipe and rachis. Furthermore, these two species are similar in the narrow angle of insertion of the pinnae on the rachis, the parallel, decurrent pinnae bases, and the dorsally exposed costal sclerenchyma. It should be noted, however, that C. mirabilis has distinctly larger sporangial capsules and spores than the other species of the C. capillaris group, which are otherwise fairly uniform in this regard. Ceradenia terrestris L. E. Bishop, sp. nov. (Fig. 2C)—-Type: Peru, Amazonas, Pcia. Chachapoyas, moist scrub forest on south side of Monlinopapa— Diosan pass, on moist bank, 2700-3100 m, 8 Aug 1962, Wurdack 1643 (holotype US). Speciminem singularem speciei terrestris strictae insolitae inveni. Rhizoma ramosus caespitosum, paleis castaneis lineari-triangularibus, basi cordatis vel subcondatis apice gradatim or atis: 2-5 * 0.2— 0. 4 mm in margine ciliis libus 100—170 xX 25- 35° a. Frondium arabe hs sunt stipites brunnei dense glandulosi demum glabrescentes basi teretes distaliter propter laminam decurrentem alati 0.6—0.9 lati 1-2 cm longi, rhachides alatae dense glandulosae vulgo sclerenchymata suo haud expositae, laminae pinnatifidae lineari-ellipticae basin versus paulatim angustatae apice auctu diuturno demum obtusae vel 24 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 1 (1989) rotundatae 8-15 cm longae pinnis subfalcatis lineari-oblongis sub angulo 30—50° a rhachide abeuntibus per basin decurrentem ampliatis apice rotundatis vel obtusis pilis glandularibus in utraque pagina uberrime praeditis 8-13 x 1.5—2.0 mm costa dorsaliter prominula ventraliter haud evidenti venis ordinate 1-furcatis stomatibus 48-60 x 42-52 wm. Sori usque ad 10 paria in quaque pinna vulgo ad apicem breviter sterilem carentes eis in utroque latere costae in sulculum communem saepe impressi capsulis oblongis vel obpyriformibus The other species apparently related here is C. herrerae. This is a rare species which I know only from three collections from central Peru and two from scales. It differes from both the related species here described by its pendent fronds and by the rachis when exposed dorsally being dark brown. Ceradenia maxoniana |. E. Bishop, sp. nov. (Fig. 2D)—T ype: Colombia, Tolima, auf moorigen Boden an den oberen Westgehangen des Alto de Oteras (Alto de las Oseras?), 3000-3400 m, 11 Jan 1883, Lehmann 2400 (holotype US, isotype B). acuminatis 1—2 X 0.2—0.3 mm margine eciliatis, cellulis medialibus 25—40 pm latis 1—4plo longioribus quam latioribus. Frondium erectarum sunt stipites (auctu completo non mihi viso) longitudine usque ad plus quam 40 cm pinnis lineari-oblongis vel lineari-triangularibus sub angulo 60—80° a rhachide t Pliatis api guste rotundatis vel obtusis glandulis translucidis in paginis ambabus dissitis demum saepe glabris 10-20 x 3—4 mm costa dorsaliter vix evidenti ventraliter prominula venis vulgo 1-furcatis interdum venula fertili ad venum distalem conjuncta etiam nonnunquam venula sterili denuo furcata stomatibus 46-52 x 40_48 L. E. BISHOP: CERADENIA 25 Sororum usque ad 12 paria capsulis oblongis vel obpyriformibus 180—200 x 140-160 pm annulis ex 12-13 cellulis constantibus illis cellulis distalibus 30-35 wm in diametro longiore quaeque pinna medialiter vel paulo supramedialiter praebet. Ad honorem clarissimi W.R. Maxonii qui adnotaverat hoc specimen repraesentare speciem novam hoc taxon laetabiliter dedico. This represents a second species of the subgenus with a more or less terrestrial habitat and erect fronds. From the Peruvian C. terrestris it differs in its larger size, its straight, more widely spreading pinnae, its less dense glandular indument, and its costae that are scarcely evident dorsally but distinctly prominulous ventrally. A much closer relationship seems to exist with C. herrerae. Apart from the narrower stipes (0.5—0.8) and pendent fronds, almost the only differentiation I have been able to make is in the tendency of the laminar trichomes of C. herrerae to remain white through frond maturity, while those of C. maxoniana soon become somewhat translucent and considerably less conspicuous. The original collection included examples of both C. maxoniana and C. herrerae. The US holotype consists solely of the former, a sheet at BM holds a single plant of the latter, and the specimen at B shows a single detached frond of each species. It is clear that the label notation ‘‘Laub.. . steht aufrecht’’ cannot apply to the slender, rather flexed stipes of C. herrerae, nor probably the reference to the habitat on open paramo (‘‘auf moorigen Boden’). It seems likely that we have here two very closely related species, one growing upright in bryophyte mats of the paramo and the other an epiphyte in the adjacent cloud forest. With regard to Lehmann’s locality, I take it to be a misreading of Alto de las Oseras, the peak of which is near the juncture of the borders of the departments of © Cundinamarca, Huila, and Tolima. The only other peaks in Tolima of requisite height occur on the western border, so that their western slopes would not lie within the indicated department. LITERATURE CITED BisHop, L. E. 1988. Ceradenia, a new genus of Grammitidaceae. Amer. Fern J. 78:1—5. PARRIS, : he 1983. A taxonomic revision of the genus Grammitis Swartz in New Guinea. Blumea 113-222. Note added in proof.—A recent loan for determination included another example of Ceradenia mirabilis: PERU. Ancash. Huaraz, Huascaran National Park, Quebrada Llaca, 77°27'W, 9°27'S, 4090 m, in organic matter between boulders, Smith & Buddensiek 11142 (MO). This collection is the first from Peru and represents a disjunction of ca. 2000 km from the type locality. American Fern Journal 79(1): 26—27 (1 989) A New Combination in South American Polystichum GEORGE YATSKIEVYCH Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166 Among the species encountered during taxonomic studies of the polystichoid ferns (Dryopteridaceae) was a little-known South American taxon described by Fée (1873) as Phanerophlebia aurita. Examination of the few available specimens disclosed that it lacks both the imparipinnate fronds and multiseriate sori that distinguish Phanerophlebia from Polystichum. Baker (1891) was the first to note that this taxon was anomalous in Phanerophlebia and he transferred it to Aspidium subg. Polystichum, now universally given generic status as Polystichum. It is obvious from herbarium annotations that the Brazilian Polystichum auritum (Fée) Yatskievych, stat. et comb. nov.—Phanerophlebia daurita Fée, Crypt. vasc. Brésil 2(suppl.):70 + t. 100, fig. 1. 1873.—Aspidium auritum (Fée) Baker, Ann. Bot. (Oxford) 5:313. 1891.—Typr: Brazil, Est. Rio de Janeiro, source du Rio Soberbo, aux Orgues [in the Serra dos Orgaos, not the better-known Rio Soberbo in southern Est. Bahia], 3 Apr 1870, Glaziou 4431 (holotype P; isotypes P! (2 sheets), C, photos LL!, MICH!). Specimens examined: Brazil, Est. Rio de Janeiro, Terezépolis, Pedra Asst, © 1900 m, 30 Nov 1929, Brade 9516 (TEX); Serra dos Orgaos, Pedra Asst, 200 m, 31 Aug 1940, Brade 16513 (P,, US). greatest detail, using both morphological (Wagner et al., 1974; Wagner, 1979) and molecular (Yatskievych et al., 1988) approaches. Each of the four segregate genera is defined by a unique combination of advanced characters, such as anastomosing venation, multiseriate sori, and radicant frond apices. All of these characters are individually also found in species of Polystichum sensu strict , 80 the validity of Tecognizing such splinters at the generic level, other than to satisfy tradition, is suspect. On the other hand, there presently does not exist a comprehensive infrageneric classification within G. YATSKIEVYCH: POLYSTICHUM AURITUM 27 Polystichum, so reincorporation of the segregates would not serve to clarify evolutionary relationships. Combinations are available for many of the species in Polystichum, but until further research provides a better understanding of interspecific relationships within the genus, I would prefer to postpone the transfer of the remaining taxa. In Polystichum, once pinnate species can be found on most major land masses (excluding Australia and Antarctica). The largest number of such taxa is in Asia, where ca. 22 species in several complexes grow. As an example of the heterogeneity within this group, in his treatment of the genus for Japan, Ryukyu, and Taiwan (the only recent classification available for a large part of the genus), Daigobo (1972) distributed species with fronds simply pinnate or nearly so into 11 of 16 sections. The occurrence of other simply pinnate species by region may be summarized as follows: the Caribbean region (17 simply pinnate spp.); Malaysia (5 spp.); Philippines (4 spp.); mainland Africa (1 sp.); Madagascar (2 spp.); Madeira (1 sp.); South America (3 spp.); North America (3 spp.). Relationships among these various taxa remain obscure, but it is doubtful that they originated from a single ancestral group, given the diversity of morphologies present. Further studies are urgently needed. ACKNOWLEDGMENTS I am grateful to Frédéric Badré (P) for his assistance with matters of typification and location of types, and to Robbin Moran, David Barrington, and Alan Smith for helpful comments regarding the manuscript. LITERATURE CITED BAKER, J. G. 1891. A summary of the new ferns which have been di d or described since 1874. Ann. Bot. (Oxford) 5:181-222, 301-332, 455-500, 1 pl. DaIcoso, y 1972. Taxonomical studies on the fern genus aie arctce in Japan, Ryukyu, and wan. Sci. Rep. Tokyo Bunrika Daigaku, Sect. B 15:57— Fée, A. L. vi 1872. Cryptogames vasculaires du Brésil, vol. 2: saplément et ryvision, Paris: J. B. Bailliére et ils, Victor Masson et Fils, Berger-Levrault & Wacner, W. H., JR. 1979. Reticulate veins in dern f Taxon 28:87—-95. - . T. MIcKEL, and L. D. Gomez. 1974. Polystichum dubium and its bearing on the problem of generic relationships of Cyrtomium and Phanerophlebia. Amer. J. Bot. 61(5, suppl.):1974. Abstract. American Fern Journal 79(1): 28—30 (1989) Shorter Notes Five Pteridophytes New to Iowa.—During the last 16 years, a series of reports on the status of the Iowa pteridophyte flora was prepared to summarize herbarium collections, published reports, nomenclatural changes, and new field collections (Peck, Proc. Iowa Acad. Sci. 82:203—208, 1976; 83:143-160, 1976: 87:39—40, 1980; 90:28—31, 1983; 91:82—84, 1984). We report an additional five taxa to the Iowa flora based on new field collections or on re-examination of problematic specimens. With these additions, the Iowa pteridophyte flora now consists of 70 taxa, a surprisingly high total for a state originally about 85% prairie and now predominantly in intensive row-crop agriculture. Botrychium campestre W. Wagner & Farrar, Prairie Moonwort, is a North American endemic that occurs sporadically in the Great Plains of Canada and the United States. The Iowa plants were originally discovered growing in loess soils on xeric, steep hill prairies in western Iowa (Plymouth Co.) by Ted Van Bruggen in 1982. They were recognized as plants new to Iowa by Lawrence Eilers, studied by Donald Farrar (Proc. Iowa Acad. Sci., 1985), Florence Wagner, and Warren Wagner, Jr., and described as a new species (Wagner & Wagner, Amer. Fern J. 76:33-47, 1986). The Prairie Moonwort is encountered from late-April to mid-June when soil and climate conditions are moderated. It is now known from loess hill prairies in five counties in extreme western Iowa (Fremont Co., Pusateri s.n., ISC: Monona Co., Farrar 875181, ISC; Plymouth Co., Eilers s.n., ISC, Farrar 845303, ISC;: Pottawattamie Co., Farrar 835291, ISC; Woodbury Co., Farrar 835261, ISC) and froma midgrass prairie on glacial moraines in northwest lowa (Dickinson Co., Farrar 885291, ISC). The habitat at the last locality is more like sites in which the species occurs in western Minnesota. Area of Wisconsin (Peck, Contr. Milwaukee Pub. Mus. Geol. Biol. 53:1-143, 1982). It was collected in 1986 from Yellow River State Forest (Allamakee Co., Rogers 004, ISTC) in a relatively pure stand of sugar maple on a north-facing slope with a sparse understory, thick leaf litter, and deep humus. This locality is 15 km west of the nearest known population of the species at Wyalusing State Park, Grant Co., Wisconsin. Cystopteris fragilis (L.) Bernh., Fragile Fern, co-occurs in Iowa with other species and hybrids in the Fragile Fern complex that were previously reported from Iowa. These taxa are particularly abundant on algific and north-facing, moist, sandstone outcrops in northeastern Iowa (Peck, Contr. Milwaukee Pub. re-examined, based on new data provided by subsequent work on the biosystematics of the genus (Moran, Amer. Fern J. 72:41—44, 1982, Amer. Fern J. 72:93—95, 1982, Castanea 48:218—223, 1983, Castanea 48:224-229, 1983; SHORTER NOTES pe Haufler, Proc. Roy. Soc. Edinburgh 86B:81—92, 1985: Haufler et al., Canad. J. Bot. 68:1855—1863, 1985; Lellinger, 1985). Fragile Fern is now known from nine counties: Allamakee Co., Peck 7845, ISTC; Clayton Co., Peck 76619, ISC; Delaware Co., Eilers 1814, IA; Fayette Co., Peck 76620, ISTC; Hardin Co., Farrar 1102, ISC; Howard Co., Peck 7861, ISTC; Jackson Co., Peck 76626, ISTC; Lyon Co., Farrar 1248, ISC; Winneshiek Co., Peck 87243, ISTC). Cystopteris laurentiana (Weath.) Blasdell, is a North American endemic that occurs in northeastern North America, westward to the Great Lakes Region and southward into the Driftless Area (Peck, Contr. Milwaukee Pub. Mus. Geol. Biol. 53:1-143, 1982). It is a putative hybrid of C. fragilis (L.) Bernh. and C. bulbifera (L.) Bernh. that has undergone polyploidy to become a fertile hexaploid. In Iowa, C. laurentiana co-occurs with C. bulbifera, C. fragilis, C. protrusa, and C. tenuis on algific and north-facing, moist, sandstone outcrops. The small, dark, scaly, and abortive bulblets on C. laurentiana do not readily abscise, making this taxon easy to distinguish from its parents. It differs from C. tennesseensis Shaver by foliar morphology and its larger spore size. Based on re-examination of herbarium specimens and additional field work in 1987, this hybrid is now known from six counties in extreme northeastern Iowa: Allamakee Co., Peck 80624, ISTC; Clayton Co., Roosa 1814, ISTC; Dubuque Co., Peck 80617, ISTC; Howard Co., Eilers 2121, IA; Jackson Co., Peck 80607, ISTC; Winneshiek Co., Peck 87242, ISTC, Nekola sn., COE. Lycopodium inundatum L., Bog Clubmoss, was discovered 17 July 1987 near Walker in extreme southern Buchanan Co. (Nekola sn., COE), disjunct 300 km to the west from populations of this species in Illinois and Wisconsin (Peck, Contr. Milwaukee Pub. Mus. Geol. Biol. 53:1—-143, 1982). The population was found in vernal pools along a paha ridge crest of a vegetated sand dune currently being grazed. It was associated with species that are quite rare in Iowa and that were also reported with L. inundatum in abandoned sand pits in northeastern Illinois: Hypericum gentianoides, Lechea intermedia, Polygala cruciata, Polygala polygama var. obtusata, Viola lanceolata, and Xyris torta (Swink & Wilhelm, Flora of Chicago region, 1979). Lycopodium inundatum occurred only in areas with sparse cover. The microsite of the prostrate stems remains moist from seepage through summer and into autumn. By late September, the plants had released their spores (Peck 87003, ISTC).—JAMES H. Peck, Dept. Biology, University of Arkansas at Little Rock, Little Rock, Arkansas 72204, JEFFERY NEKOLA, Curriculum in Ecology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599; DONALD R. FARRAR, Dept. Botany, Iowa State University, Ames, Iowa 50011. The Flavonoids of Polystichum acrostichoides.—Hiraoka (Biochem. Syst. Ecol. 6:171-175, 1978) reported flavonoids in the leaves of five species of Polystichum: P. lepidocaulon, P. tsus-simense, r. craspedosorum, P. tripteron, and P. polyblepharum. These flavonoids are O-glycosides of the cee kaempferol (3-glucoside, 7-arabinoside, 3-rhamnoglucoside, 3-diglucoside, - 3-rhamnodiglucoside) and quercetin (3-glucoside and 3-rhamnoglucoside}, 32 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 1 (1989) of Old World ferns. For Malaysia, this can be remedied by using the amply illustrated and highly informative book recently published by Audrey Piggott, whose husband provided the photographs and notes on photography under difficult tropical circumstances. And what photographs they are! Habitat shots, habit shots, individual fronds, and pinnae all abound. A major chapter on principal vegetation types and fern habitats is like taking in a botanical travelogue. It is fine background for the main part of the book, in which nearly 80% of Malaysia’s ferns (392 species) are treated. This is by no means a standard Flora, and so there are no keys. (One can rely onR. E. Holttum’s Flora of Malaya, vol. 2. Ferns and its keys as a companion volume). Each species is listed under an up-to-date scientific name. Important synonyms are given, especially those used in Holttum’s book. Habitat, distribution, and economic information are provided, as is a short, informal description. The figure legends are informative. One really can come to know the species through the illustrations; the habit shots are especially helpful in this regard. The book concludes with a brief glossary and an index. This book is an unusual and fresh approach to the study of tropical ferns and belongs on the shelf of all who are interested in ferns: amateurs, growers, and professional botanists alike. It is available from US. book dealers, as well as from the publisher.—Davip B. LELLINGER, U.S. National Herbarium NHB-1 66, Smithsonian Institution, Washington, DC 20560. “A Nomenclatural Guide to R. H. Beddome’s Ferns of South India and Ferns of British India,” by S. Chandra and S. Kaur. 1987. x+139 pp. Today and Tomorrow’s Printers and Publishers, 24-B/5 Original Road, Karol Bagh, New Delhi 110005, India. US $15.00. Colonel Beddome’s quarto volumes, first published in the 1860’s, remain unequalled for their clear and comprehensive drawings of Indian ferns. Although they are useful for identification purposes, the volumes have fallen into disuse over the years as waves of nomenclatural changes have obscured the scientific names applied to the plates by Beddome, especially names in the genera Aspidium, Lastrea, and Nephrodium. This has now been remedied by Chandra and Kaur’s most useful book. Each of the plates in Beddome’s volumes is listed in numerical order. Both Beddome’s name and a modern equivalent are INFORMATION FOR AUTHORS Authors are encouraged to submit manuscripts pertinent to pteridology for publication in the American Fern Journal. Manuscripts should be sent to the Edi- tor. 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 ac- ceptability of manuscripts nor the date of publication. Authors should adhere to the following guidelines; manuscripts not so pre- pared may be returned for revision prior to review. Submit manuscripts in tripli- cate (xerocopies acceptable), including review copies of illustrations. Do not send originals of illustrations until they are requested. Use standard 81/2 by 11 inch paper of good quality, not “erasable’’ paper. Double s uscripts throughout, including title, authors’ names and addresses, text (including heads 98g keys), literature cited, tables (separate from text), and figure captions ires). A f man- a uscript 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. Avoid footnotes and do not break words at ends of lines. Make table head- ings and figure captions self lanat 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 typi- fication), use one paragraph per basionym (see Regnum Veg. 58:39—40. 1968). Abbreviate titles of serial publications according to Botanico-Periodicum- Huntianum (Lawrence, G. H. M. etal., 1968, Pittsburgh: Hunt Botanical Library). References cited only as part of nomenclatural matter are not included in litera- ture cited. For shorter notes and reviews, put all references parenthetically in text. Use Index herbariorum (Regnum es 106:1—452. 1981) for designations of herbaria. Til S L 13 ik e's = Jal. Loe 52 LO fit Paseo § + — Lt. page. Provide margins of at least 25 mmon all ill illustrations, design originals for reproduction without reduction or pa uniform amount. In composite blocks, abut edges of adjacent photographs. Avoid com- bining continuous-tone and line-copy in single illustrations or blocks. Coordi- nate sequence and numbering « of ‘figures (and of tables) with order of citation in text. E & ptions. Include a scale and reference to latitude and longitude i 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. Au- — will be assessed charges for extensive alterations made after type has been Foe other matters of form or style, consult recent issues of American Fern Jour- nal and The Chicago manual of style, 13th ed. (1982. Chicago: Univ. Chicago Press). Occasionally, departure from these Lessee’ may be justified. Authors teen = pt cane ai - preparation. | Papers longer than 32 salted pages may be sent to the Editor of Pteridologia arora mara see cover 2). 2 The Ferns and Fern-allies of Costa Rica, Panama, and the Choco (Part 1: Psilotaceae through Dicksoniaceae) This first of two volumes will be published in April 1989. In addition to introductory material and a key to the families, it treats about 570 species in 68 genera. Each generic treatment includes a key to the species, synonymies, habitats, distributions, and notes. Most species are illustrated with a line drawing of a frond or part of a frond so that similar species can be distinguished and identifications checked without reference to a tropical herbarium. The volume costs $32.00 postpaid. American Fern Society, Inc., U.S. National Herbarium NHB-166 Smithsonian Institution, Washington, DC 20560 American oe Fern Number 2 April-June 1989 Journal QUARTERLY JOURNAL OF THE AMERICAN FERN SOCIETY Species C ts in Pteridophytes: Introducti Christopher H. Haufler 33 George Yatskievych and Robbin C. Moran 36 nr. a:.. c. 2 c * 3y..32. Sg. 42 Ir. . cy 2 ab Uy rz, ang ict : rt . g. sk. Zz, Hvbridizati Reticulation r 2 Tu..f.. 28. Seartes ti us in Peaciddenisvtes: The Treatment Species a = c. 2 Pad ge 2 Ts. 23 | eer ssh, <. > 3 p.f. ho 2 ‘a és a ae ee ee e. y i Synthesi: Christopher H. Haufler 90 Cathy A. Paris, Florence S. Wagner, and Warren H. Wagner, Jr. 46 David S. Barrington Cleistophier Hi Haufler, i Charles R. Werth 55 Gerald J. Gastony and Michael D. Windham — 65 A. James Hickey, W. Carl Taylor, and Neil T.Luebke 78 The American Fern Society Council for 1989 JUDITH E. SKOG, Biology Dept., George Mason University, Fairfax, VA 22030. President DAVID B. LELLINGER, Smithsonian Institution , Washington, DC 20560 Vice-President W. CARL TAYLOR, Milwaukee Public Museum, Milwaukee, WI 53233. Secretary JAMES D. CAPONETTI, Dept. of Botany, University of Tennessee, Knoxville, TN we Treasurer DAVID S. BARRINGTON, Dept. of Botany, University of Vermont, Burlington, VT 0540 ce Treasurer JAMES D. MONTGOMERY, Ecology III, R.D. 1, Berwick, PA 18603. Back Issues Curator ALAN R. SMITH, Dept. of Botany, University of California, Berkeley, CA 94720. Journal Editor DAVID B. LELLINGER, Smithsonian Institution, Washington, DC 20560 Memoir Editor JOHN T. MICKEL, New York Botanical Garden, Bronx, NY 10458. - iddlehead Forum Editor American Fern Journal EDITOR Dee et Dept. of Botany, University of California, Berkeley, CA 94720 ASSOCIATE EDITORS GERALDLGASTONY .......... Dept. of Biology, Indiana University, Bloomington, IN 47401 ere AIER Dept. of Botany, University of Kansas, wrence, KS 66045 DAVIORIBILINGRN: U.S. Nat’l Herbarium NHB-166, Smithsonian Institution, : Washington, DC rai TERRY R. WEBSTER ..... Biological Sciences Group, oe of C ticut, Storrs, CT 06268 The “American Fern Journal” (ISSN 0002 8444) j terly d general 5 ve Spheres < « $1 pS << on PIE for back issues Say be addressed to Dr. James D. note pedis Il, R.D. 1, Berwick, PA 18603. Changes of address, dues, and applications for membership should be sent to the Records Tene. Sahecsitca. $20.00 gross, $19.50 net if sop ae agency (agency fee $0.50); sent free to members Fern Society (annual dues, $15. Poets + $4.00 mailing surcharge beyond U.S.A., Conade: aad hte co; life membership, $300.00). Back volumes 1910-1978 $5.00 to $6.25 each; single back numbers of 64 Pages or less, $1.25; 65-80 pages, $2.00 each; over 80 pages, $2. 50 each, plus pte Back volumes 1979 et seq. $8.00 each; single back numbers $2.00 each, plus shipping. T. di ders of six vol JULES or more. Doacm <, i ax. 2 . “ Bi AMER FERN Jou Dept. of Botany, University of Vermont. Burlington, VT 05405-0086, - The editor welcomes contributions from members and non-members, including miscellaneous nontechnal books on deroy, Purchase materials, personalia, horticultural notes otes, and reviews of Spore Exchange _ Mrs. — Horder, 16813 Lemolo Shore Drive N -E., Poulsbo, WA 98370, is Director. Spores S{uacot, __ Gifts and Bequests senyshaley e aml yser yr fammoomgomapleomee back issties af + alwatve - Tnnwiei. L Lit ay zy t + American Fern Journal 79(2): 33—35 (1989) Species Concepts in Pteridophytes: Introduction CHRISTOPHER H. HAUFLER Department of Botany, University of Kansas, Lawrence, Kansas 66045 Systematics is among the world’s oldest sciences and throughout its history there has been debate concerning concepts and definitions of species. Over 30 years ago, Mayr (1957, p. iii), in introducing a symposium on the species problem, stressed that the “species is a biological phenomenon that cannot be ignored” and ‘‘continued interest in the 2 Species problem requires no apology.” Although for practical reasons standard-of definition would seem paramount, viable scientific discipli tremain stagnant and must continually question and refine their critical parameters. Progress in systematic thought has paralleled that in related fields and has resulted in continual modifications of species concepts. Beginning with a static morphologically-based definition, subsequent species concepts acknowledged the significance of the Darwin/Wallace hypothesis of progressive evolutionary change as well as the fact that genetical features were often responsible for maintaining cohesive lineages. More recently, the advent of techniques for generating new data bases, the incorporation of information from geography and ecology in postulating mechanisms for speciation, the development of cladistic methodology for proposing phylogenetic hypotheses, and the application of morphometrics in refining descriptions have all influenced our concepts and definitions of species. Pressure from outside the field of systematics has also created continued interest in ideas about species. Just as ecologists find it important to know what taxa are inhabiting their study plots, physiologists need to be aware of the limits of variability in the individuals they study. Everyone wants to be working with properly identified species and wants others to know precisely what organisms are being studied. Thus, when systematists modify their concepts of species, the “ripple effect’’ is felt at many other levels in the biological community. Until recently, the concept of species in ferns and fern allies has not been the subject of overt debate. In part, this has been true because, as in other groups, pteridophyte systematics has been dominated by the practical definition of species as morphologically discrete units. Such continued reliance on superficial resemblance seems rather surprising because most pteridologists acknowledge that ferns and fern allies have fewer easily perceived morphological features than do seed plants. Reticence to develop and incorporate new data bases into revision of species concepts can be justified by realizing that pteridophyte species raise impediments not frequently encountered in work with other vascular plants. For example, high chromosome numbers make analyses of meiotic behavior difficult, rampant hybridization muddies already subtle species boundaries, and the high frequency of polyploidy and asexually reproducing taxa suggest that ‘“‘secondary”’ species need to be considered evolutionarily significant. Beyond these inherent MiSSOURL JUN 2 GARDEN LIBRARY 34 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) obstacles, species-level studies have been overshadowed by the perception that resolving systematic relationships at higher (family) levels of classification was of greater immediate importance. But, this situation is changing and unanswered questions about species concepts have been accumulating since we began incorporating what have been termed ‘‘biosystematic” characters into our classifications. In addition, our desire to develop explicit phylogenies has demanded that we address the species concept question with greater precision. Beginning in 1950, Manton demonstrated that studies of chromosome number and meiotic behavior could be informative in addressing problems at the species level. Her studies showed that if we were to recognize entities that were biologically real as well as taxonomically distinct, few could debate the importance of chromosome number changes in establishing dividing lines between interbreeding assemblages of populations. In delimiting “natural” groups, the new chromosomal data were of particular relevance because they usually indicated that recognized species should be subdivided. Most significantly, however, these chromosomally defined groups provided hypotheses that could be tested using other data sets. Often, from this new perspective, morphological characters that had been considered confusingly variable became species descriptors and brought taxonomic practicality to chromosomally cohesive units. In addition to chromosomal information, new data sets including micromorphological (spores, scales, trichomes, etc.) and biochemical (phenolic compounds, isozymes, DNA, etc.) characters have been used to increase the With the development and application of new characters and techniques, however, there has appeared a growing tension between systematists who use a morphological species concept (and are interested in practicality first and foremost) and those who use a biological or evolutionary species concept (and are requiring that species have some sort of real existence in nature). It is clear that as our ability to resolve independent lineages increases, so does the power as well as explaining their limitations. Based on these and other considerations, a symposium on Species Concepts C. H. HAUFLER: INTRODUCTION 35 working towards compromise summaries of their opinions. I hope that what we have assembled is a representative synthesis of current ideas about species in pteridophytes. Ideally, the conclusions presented in the following papers should enhance the role of systematics in improving our portrayal of evolutionary history. LITERATURE CITED MANTON, I. 1950. Problems of cytology and evolution in the Pteridophyta. Cambridge: Cambridge University Press Mayr, E. 1957. Preface. Pp. iii—v in The species problem, ed. E. Mayr. Amer. Assoc. Adv. Sci. Publ. 50. WAGNER, W. H., JR. 1961. Problems in the Sitar arses of ferns. Pp. 841—844 in Recent advances in botany. Montreal: University of Toronto Pre American Fern Journal 79(2): 36—45 (1989) Primary Divergence and Species Concepts in Ferns GEORGE YATSKIEVYCH and ROBBIN C. MORAN Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166 Primary divergence may be defined as differentiation among populations of homoploid organisms caused by natural selection; it often results in the formation of new species or subspecies. This definition excludes the relatively instantaneous processes of polyploidy and hybrid species formation. Monographers of pteridophyte genera name and describe species and have perforce been the individuals forced to deal with problems of primary divergence and the delimitation of species. Since scant literature exists addressing species concepts for pteridophytes, we have contrasted specific examples of usage from modern systematic treatments by various workers. We examined a sample of 50 monographs completed during the last 50 years by 36 authors (Table 1). Only 12 (24%) of these monographs contained any explicit discussion of the criteria for the species concepts and/or infraspecific categories employed, so that our analysis of these topics is based largely on inference. We perceive that three general types of species concepts have been used singly or in concert by pteridologists: the biological, morphological, and (for lack of a better term) “‘look-alike’’ concepts. The biological concept is used here in the somewhat restricted sense of genetic intersterility between species and states that if two taxa can cross to produce fertile offspring, then they are the same species. This concept has rarely been used in fern systematics because data on cross-fertilization is difficult to obtain. Hennipman and Roos (1982), however, used the ability of gametophytes to cross as evidence for reducing four taxa accepted by Hoshizaki (1972) as species to subspecies or varieties of Platycerium bifurcatum, even though the subsumed taxa were geographically and/or ecologically distinct. In contrast, the morphological concept is the recognition of discrete taxa on the basis of breaks of form, with no knowledge of the ability to hybridize. Of course, what constitutes such morphological discontinuities can be quite subjective and contentious. This concept is the tacit species concept of most fern monographers who have worked from herbarium specimens alone. The look-alike concept is similar to the morphological concept (and might best be considered a subset of it) in that both delimit taxa by morphological discontinuities. It differs, however, in that it classifies taxa hierarchically as species or varieties based upon relative morphological similarities—an author may Classify two morphologically distinct entities as varieties because they “look alike,” when by strict morphological criteria they might otherwise have been treated as separate species. An example is the Lycopodium obscurum complex studied by Hickey (1977). He clarified the taxonomy of this group by showing that a new taxon, L. isophyllum, was morphologically distinct from L. obscurum and L. dendroideum. Hickey classified isophyllum as a variety of L. obscurum rather than as a separate species, because of closer resemblance of YATSKIEVYCH & MORAN: SPECIES CONCEPTS 37 these two taxa than either to L. dendroideum. Thus the use of the varietal designation was intended to imply phylogenetic affinity of two distinct taxa (Hickey, pers. comm.). Johnson (1986) also used this concept in recognizing two distinct, nonintergrading taxa as subspecies of Marsilea vestita. SUBSPECIES AND VARIETIES Infraspecific categories must be considered in relation to species concepts when discussing primary divergence, because one person’s species is often another person’s subspecies or variety. Infraspecific categories are frequently used by monographers; of the 50 monographs surveyed (Table 1), 33 (66%) included subspecies, varieties, or both. As with species concepts, the definitions of infraspecific categories were rarely discussed and had to be inferred from the author’s work. In some instances it was impossible to determine why a monographer used an infraspecific instead of specific category. Subspecies were used in 11 (22%) of the monographs that we surveyed (Table 1) and were more consistently defined than varieties. Subspecies were generally used to describe geographic variation that either varied continuously (i.e., without sharp morphological breaks; e.g., Clausen, 1938, in Botrychium multifidum) or varied discontinuously and was based on a single character (e.g., Johnson, 1986, in Marsilea vestita). Subspecies have, however, also been used to name cytotypes (Gastony & Windham, this symposium). Varieties were used in 29 (58%) of the monographs surveyed (Table 1). There was a general lack of consistency in how varieties were defined and a few monographs contained examples of more than one kind of variety within a single genus. All of the definitions used for subspecies were also used for varieties by various monographers. Where nonintergrading geographic variation was recognized, either a single character (e.g., Stolze, 1981, in Cnemidaria uleana) or several ‘insignificant’? characters were cited, usually involving indument density, slight differences in leaf-cutting, or microscopic features (e.g., Gastony, 1973, in Nephelea woodwardioides). In some cases, monographers chose to describe the minute, nonintergrading character variation without resorting to formal taxonomic designation (e.g., Moran, 1986, in Olfersia cervina). Varieties were also used to describe nongeographic variation (e.g., Stolze, 1981, in Cnemidaria mutica). A ids, and hybrids have also been named as varieties (Barrington et al., this : symposium; Gastony & Windham, this symposium). Finally, as noted above, in some cases varieties were simply named without any discussion, and we were unable to infer from the publication on which kind of character(s) the taxon was based. The inconstancy of usage of the varietal and subspecific categories has long been a problem in systematics. Because there has been little attempt to standardize the definitions for these terms, particularly when applied to taxa involving primary divergence, it would seem that this problem will remain with us in the future. Although the International Code of Botanical Nomenclature recognizes variety as the primary level of infraspecific classification, we prefer the term subspecies for situations involving geographically defined variation, $f1Q09R A007) KA TABLE 1. Data Compiled from 50 4 } . : Ps ‘ce that discuss species concepts = 12 (24%). fonographs using I g 33 ( (66%); sub I 11 See varieties = 29 (58%); molecular data = 3 (6%); cladistics = 16 (32%). Infraspecific Discussion of categories used sp. concept Studies Author Taxon Spp subsp. var. molecular cladistic Alston et al. 1981 Selaginella 133 + 7 - ~ _ Barrington, 1978 Trichipteris 55 = + - - ~ Bishop, 1978 Cochlidium 16 - a = me — Blasdell, 1963 Cystopteris 10 + = - + Boer, 1962 Didymoglossum & oo 19 os - = ~ ~ ‘Brown, 1964 Woo 22 = + ~ _ + Clausen, 1938 het chi & Ophioglossum 50 oa + 7 ~ - Conant, 1983 Alsophila 30 + - - ~ - de la Sota, 1966 Polypodium squamatum oup 22 = ES ae - = Evans, 1969 Polypodium pectinatum- plum 26 * + ~ - ~ Gastony, 1973 Nephelea 18 a + + - - Haufler,. 1979 Bommeria 4 ee = _ + Hauke, 1963 Equisetum subg Hippochaete 7 + + - _ + Hennipman, 1977 olbitis 41 + + ~ - - Hennipman & Roos, —Platycerium 15 + + + - + 1982 Holttum, 1971 Stenochlaena 6 = ~ - ~ - Holttum, 1975 Syngramma 17 = = _ _ - Holttum, 1986 Triplophyllum 20 — - + _ - Hovenkamp, 1986 Pyrrosia 51 > + i = + Johnson, 1986 Marsilea 12 + - ~ ~ ge (6861) Z WAAWNON 62 ANN'TOA ‘TYNYNO! NYdd NVOMANV Kramer, 1957 Hennipman, 1986 Riba, 1967 Roos, 1985 Smith, 1971 Smith, 1980 Smith, 1986 Stolze, 1974 Tindale, 1965 Wilce, 1965 Windisch, 1978 Yatskievych, 1989 Lindsa Placing Niphidiu oa cleoids Ais subg. Coptophyllum Olfers Salybotey: Grammitis Pyrrosia Alsophila Drynaria Thelypteris subg. Cyclosorus Thelypteris subg. Steiropteris Cyclodium Cnemidaria Lastreopsis Pellaea Eriosorus cata oryop Slainlla rupestris group Notholae earnteedie Pityrogramma Cyathea Diellia Lycopodium sect. Complanatum Sphaeropteris Phanerophlebia C+ PS Ee 44 eee + (+t + rte f+ +i +++ S.LddDNOD SAIDAdS ‘NVYOW 8 HOAAXINS.LVA 40 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) because it has historically been restricted to this usage, and because variety has been used so many different ways that it is a priori impossible to discern what the term refers to. Such usage of the term subspecies would also agree with its use by zoologists. In any event, monographers using the varietal designation in future studies would be well advised to state their criteria for the use of this level. MODERN TRENDS AND How THEY AFFECT SPECIES CONCEPTS Several trends have become apparent in how pteridologists study variation. These include more field work (and an increasing emphasis on populational studies), cladistic and/or phenetic analysis of data, and the incorporation of biochemical and molecular sources of evidence into taxonomic studies. Field work.—One trend is the increase of field work by monographers. This is particularly true for studies involving ferns with large fronds, thick rhizomes, or other structures poorly preserved on herbarium specimens. Field studies allow collection of data not available from herbarium specimens alone, such as information on habitat, ecology, phenology, and species interactions. In addition, field work provides the opportunity to collect anatomical and cytological material, the study of which gives evidence for the delimitation of species. Field work can be especially helpful when related species grow in the same habitats, forming a “genus community” (Wagner & Wagner, 1983). Field research is often correlated with another important trend—the analysis of inter- and intrapopulational variation. Examination of such variation has proven important in assessing the degree to which morphological characters actually serve to distinguish taxa. An example comes from Moran (1987b), who found during field work on Polybotrya (a genus with strongly dimorphic sterile and fertile fronds) that three morphotypes with lobed pinnae, which traditionally had been treated as separate species, actually represented unusual individuals bearing intermediate sterile-fertile leaves and belonged to a single species. Data from interpopulational studies also have been used to document and describe geographic clines for morphological characters (Tryon, 1971, 1986; Moran, 1987a). Most of the experimental approaches to the study of plant evolution require examination of intra- and interpopulational variation before these approaches can be applied to questions at the species level (discussed below). Cladistic and phenetic analyses——Some pteridologists have attempted to quantify their findings, especially if they have generated large amounts of data. Quantification normally takes one of two forms: phylogenetic (=cladistic) analyses, or various phenetic analyses involving multivariate measures of similarity or dissimilarity (e.g., principal components, clustering). An explanation of these methods is beyond the scope of this paper, but interested readers are referred to the brief, but excellent general introduction by Simpson (1986). Although phenetic and cladistic analyses have not been widely used by YATSKIEVYCH & MORAN: SPECIES CONCEPTS 41 fern systematists (Table 1), the recent emphasis on experimental data sources will undoubtedly increase the use of quantitative methods of analysis, if only so that investigators can more efficiently deal with the often voluminous data they generate. Underlying cladistic precepts such as paraphyly (the origin of a segregate taxon from within a derived portion of an ancestral group) greatly constrain the kinds of taxonomic assignments cladists can make. Examples of numerical cladistic analyses include those of Hennipman and Roos (1982) on Platycerium, Roos (1985) on Drynarioideae, Hovenkamp (1986) on Pyrrosia, and Moran (1987a) on Polybotrya. Several other studies have employed cladistic methods in formalizing classifications or interpreting evolutionary trends without resorting to numerical cladistic analyses. Examples of nonstatistical cladistic studies include those of Hickey (1986) on Isoétes and Moran (1986) on Olfersia and Polybotrya. Cladistic analysis of morphological data is complicated by the difficulties of coding and polarizing complex characters and high levels of homoplasy. Cladistics, however, can have great utility in developing evolutionary hypotheses from more conservative types of information, such as restriction site mutational analysis of chloroplast DNA (e.g., Yatskievych et al., 1988, on polystichoid ferns). A recent example of phenetic analysis is that of Waterway (1986) on two species of Lycopodium. Using statistical methods, Waterway found that the correlation among a suite of characters supported the separation of L. lucidulum and L. porophilum, which had previously been considered conspecific by some authors. In addition to primarily morphometric studies, multivariate analyses have also been performed on several types of experimental data, notably in flavonoid and isozyme studies. The advantages of statistical analysis include reproducibility of results and effective condensation of potentially overwhelming amounts of information. Quantitative approaches to taxonomic study also have implications for how researchers view species limits, since taxonomic assignments resulting from such studies are based on various statistical levels of significance. Because titati tically compare all taxa in a study group against the same “‘vardstick,”’ relative affinities are more easily assessed than by qualitative methods. Biochemical and molecular approaches.—Most of the biochemical and molecular research on pteridophytes has focused on speciation involving hybridization and/or polyploidy, and its use for studies on primary divergence and species delimitations remains to be adequately addressed. Of the various techniques currently in use, three sources of data show the greatest promise for elucidation of primary series limits: secondary compounds, isozymes, and nucleic acids. Secondary compounds, particularly flavonoids, have been a staple of experimental systematics during the last 25 years. A recent use of flavonoid analysis in a problem involving primary divergence is the work of Seigler and Wollenweber (1983). They showed that, in the absence of clear morphological characters, the diploid Notholaena standleyi was statistically separable into 42 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) three groups of populations on the basis of geography, substrate preferences, and types of flavonoid exudates. Although the flavonoid evidence suggests that these taxa probably do not interbreed, they have not been formally named. Thus, in this case, the data from experimental sources have not (yet) affected species recognition. Several comparative analyses of enzyme variation have proven the usefulness of measuring both intra- and interpopulational variation (see most other papers, this symposium). Although much of the recent work involving pteridophytes has focused on polyploid or hybrid speciation, a recent study by Werth et al. (1985) on the Appalachian Asplenium complex had bearing on the question of primary divergence. They showed that A. rhizophyllum, one of the three diploid species in the complex (often segregated as Camptosorus) was more closely related genotypically to A. platyneuron than was A. montanum, a ‘“‘typical”’ Asplenium. Other recent isozyme studies also include data on primary divergence: Haufler (1985) upheld the distinctions between three morphologically defined species of Bommeria, but combined two other taxa under B. subpaleacea, and Yatskievych (1989) upheld the distinctions between two morphologically similar species of Phanerophlebia. Few studies have been published on ferns involving comparative analysis of nucleic acids; however, many of the molecular techniques now available provide highly conserved characters for analysis of primary divergence. Stein et al. (1979) used evidence from DNA denaturation/renaturation studies to challenge an existing hypothesis, based primarily on paleobotanical evidence, concerning species relationships in North American Osmunda. More recently, Stein et al. (1986) presented qualitative comparisons of the chloroplast genomes of these species in support of the earlier molecular data. These studies did not, however, affect the status of the species previously recognized in Osmunda. Yatskievych et al. (1988) presented a cladistic study of restriction site mutations in the chloroplast genomes of groups of morphologically similar species of Polystichum and Phanerophlebia. They showed that the presence of vein-anastomoses, which was previously thought to be of major importance for classification within Phanerophlebia, arose twice, and that free-veined P. nobilis and net-veined P. remotispora were conspecific. They also found that Polystichum munitum and P. imbricans, both from the Pacific Coast of North America, are quite dissimilar based on their chloroplast DNAs even though they are morphologically (Wagner, 1979) and isozymically (Soltis & Soltis, 1987) very similar. These data are in agreement with those of Soltis and Soltis (1987) on the chloroplast genomes of various populations of this species pair. In most molecular studies the underlying genetic basis of the taxonomic characters is readily verifiable. Molecular researchers generally examine genotypic, rather than phenotypic, variation, data that can provide strong, indirect evidence of crossability between taxa. The species accepted thus come closer to biological rather than morphological species, even though no direct evidence of crossability has been examined. In this way, molecular studies allow pteridologists to test whether species previously defined by the morphological concept are also valid species by the biological concept. YATSKIEVYCH & MORAN: SPECIES CONCEPTS 43 SUMMARY Although monographers of fern genera rarely discuss criteria for their specific and infraspecific categories, three types of species concepts have been used by pteridologists: the biological, morphological, and look-alike concepts. Infraspecific categories were used by 66% of the monographs surveyed, and in some instances it was impossible to determine why a monographer used an infraspecific instead of the specific category. The term ‘‘variety’’ has lost meaning since it is defined in so many ways that it is impossible to know a priori what kind of variation it was intended to denote. Monographers are urged to define this term explicitly when using it in future publications. In ferns, newer data sources have not been widely used to study primary divergence and have had varying effects on — recognition. The increase of field studies has sh d species by allowing the collection of data unavailable from herbarium specimens alone. Field work also allows the study of inter- and intrapopulational variation, the lack of which has often resulted in the same species being described more than once. The increase of cladistic studies has had limited effect on species recognition in ferns, because species limits have largely been predefined before these studies were done. Quantitative methods have rarely been used in fern taxonomy and no new species have been named based on such methods. The increase of biochemical and molecular approaches has in most cases supported morphological species; however, two examples were found where species were combined on the basis of such evidence. Molecular and biochemical data sources have been used primarily to address relative degrees of interspecific relationships (i.e., in refining classifications), rather than to recognize species. This may partially be explained by the fact that practical taxonomists have avoided naming “‘cryptic” species, 1.e., those entities not readily separable by macromorphological characters (Paris et al., this symposium), because experimental evidence is generally not useful in distinguishing taxa in the field or herbarium. Molecular studies, which can detect lack of gene flow, may prove useful in allowing pteridologists to test whether morphological species are also species in the biological sense. LITERATURE CITED Aston, A. H.G., A.C. JeRMy, and J. M. RANKIN. 1981. otis genus Selaginella in tropical South America. Bull. Brit. Mus. (Nat. Hist), Bot. 9:233— Bisuop, L. E. 1978. Revision of th ). Amer. Fern J. 68:76—94. BLASDELL, R. F. 1963. A ‘iecaiberaphic study of the fern genus Cystopteris. } Mem. Torrey Bot. Club 21(4):1-102. Boer, J. ie eh 1962. The new world species of Tri t. Didymogl d Microg a Bot. Neerl. 11:277-330. CLAUSEN, - T.1938.A graph of the Ophiogl m. Torrey Bot. Club 19(2):1-177. Conanr, D. S. 1983. A revision of the genus Alsophila ephsiso in the Americas. J. Arnold rbor. 64:333—382. BARRINGTON, D. S. 1978. A revision of the genus Trichipteris. Contr. Gray Herb. 208:3—93. DE LA Sota, E. R. 1966 Revisién de las especies Argentinas del grupo Polypodium squamatum L. Polypodiaceae (s. str.). Revista Mus. La Plata, Secc. Bot., n.s., 10: 69-186 + 8 plates. 44 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1 989) Evans, A. M. 1969. Interspecific See in the Polypodium pee peels complex. nn. Missouri Bot. Gard. 55:193-— GasTony, G. J. 1973. A revision of =e fern genus Nephelea. Contr. Gray Herb. 203:81-148. HAUFLER, C. H. 1979. A biosystematic revision of Bommeria. J. Arnold Arbor. 60:445—476. . 1985. oe variability and modes of evolution in Bommeria (Pteridaceae). Syst. Bot. 10:92-10 HAUKE, R. . fe ‘A taxonomic monograph of the genus testes subgen. Hippochaete. Beih. a Hedwigia 8:1-123 + 12 tables, 3 graphs, and 22 sede ae E. 1977. A monograph of the fern genus Bolbitis heheiatuiiioe Leiden Univ. Press, gids Bot. Ser. 2:1—331. aie M. C. Roos. 1982. A monograph of the fern genus pevestam EPebenodiseee) Verh. n. Ned. Akad. Wetensch., Afd. Natuurk., Tweede Reeks 8 HICKEY, R. 1 1977. The Lycopodium obscurum complex in North America. en Fern J. 67:45—48. - 1986. The early evolutionary and morphological diversity of bites with descriptions of wo neotropical species. Syst. Bot. 11: 309—321 HOLtTTuM, R. E. 1971. The genus Stenochlaena J. Smith with a description of a new species. Amer. en 61:119-123. 5. A comparative account of the fern-genera Syngramma J. Sm. and Taenitis Willd., ne discussion of their relationships to each other and to other genera. Kew Bull. . =343 +2 pla tes llied to Tectaria V. Triplophyllum, a new genus of Africa ean Kew Bull. 41: 237-260. Hosuzant i 1972. Morphology and phylogeny of Platycerium species. Biotropica 4:93— VEN , P. 1986. Monograph of the fern genus Pyrrosia (Polypodiaceae). Leiden on Ser. 0. JOHNSON, D. M. 1986. arian of the new world species of Marsilea (Marsileaceae). Syst. Bot. nogr. 11 KRAMER, K. U. 1957. ae poene a8 genus Lindsaea in the new world with notes on allied genera Acta Bot. aaa 6:9 LELLINGER, D.B. 1 ae renee species of Plagiogyria sect. Carinatae. Amer. Fern J. 61:11 en arthe oh —o of the fern genus snl home Amer. Fern J. 62:101— n: Botany of the Guayana steiiecucaie Xi. a wien York Bot. Gard. 38:2— 9. LLoyp, R. M. 1971. Systematics of the onocleoid ferns. Univ. Calif. Publ. Bot. 61:1—93. MICKEL, J. T. 1962. Am Spots oay study of the fern genus Anemia, subgenus Coptophyllum. Iowa State Coll. J. Sci. 36:349-482. Moran, R. se 1986. The shtatie fern genus Olfersia. Amer. Fern J. 76:161-178. "_———. 1987a. Monograph of the neotropical fern genus Polybotrya (Dryopteridaceae). Bull. ee Nat. Hist. Surv. 34:1-138. . 1987b. Sterile—fertile leaf Ronis and the evolution of soral types in Polybotrya (Dryopteridaceae) Syst. Bot. 12: 617— PARRIS, S 5.1 G itis Swartz (G itid ew ayy Blumea 29:13-— Ravensnure, bt J. and se _HENNIPMAN. A966. The Pyrrosia species Suan ana to Fil icales) Anales Inst. Biol. Univ. Nac. México, ser. 38 Bot. 1:61—1 Roos, M.C. 1985. Phylogenetic systematics of the Drynarioideae (Polypodiaceae). Ph.D. dissertation, Univ. of Utrecht. SEIGLER, D. S.andE # Witte 1983. Chemical variation in Notho] tandleyi. Amer. J. Bot Wee : Simpson, M. G. 1986. Phenetic and phylogenetic classification. Aa 313-327 in: Fundamentals of plant Beare A. E. Radford. New York: Harper and Row YATSKIEVYCH & MORAN: SPECIES CONCEPTS 45 SMITH, A. R. 1971. ee of the neotropical! species of Thelypteris section Cyclosorus. Univ. ae Publ. Bot. —143. gee: of Thelypteris subgenus rs including Glaphyropteris Peridophsta) Univ. Calif. Publ. Bot. 76:1- putea 6. Revision of Eager tropical fern lodi Amer. Fern J. 7 SOLTIS, P. : ne D. E. SoLtis. 1987. Genetic dlerentiton veniam = closely uae specins of Polystichum sciriss acl Aeteael Amer. J. B 55. (Abstr. STEIN, D. B., W. F. THompson, and H. S. gt sp Seageiag on DNA sequences in the Osmundaceae. J. Molec. Evol. 13:215—232. STEIN, DB, J. D. phuemee and Lad ae ciate 1986. _ Structural evolution and flip-flop Curr. Gen, 10:835-—841. STOLZE, . G. 1974. A iaxouoniic’ revision 7 the genus Cnemidaria (Cyatheaceae). Fieldiana, Bot. s. 6. 37:1-98 TINDALE, M D. 1965. A monograph of the genus Lastreopsis Ching. Contr. New South Wales Natl. erb. 3:249—339 + 23 plates. TRYON, A. F. 1957. A revision of the fern genus Pellaea section Pellaea. Ann. Missouri Bot. Gard. 44:1 3. . 1970. A monograph of the fern genus Eriosorus. Contr. Gray Herb. 200:54-174. TRYON, R. M. 1941. Revision of the genus Pteridium. Contr. Gray Herb. 134:1-68. ———. 1942. A revision of the genus Doryopteris. Contr. Gray Herb. 143:3—77 . 1955. Selaginella rupestris and its allies. Ann. Missouri Bot. Gard. 42:1-99 + 6 plates. . 1956. A revision of the American species of Notholaena. Contr. Gray Herb. 179:1—106. . 1960. A review of the genus Dennstaedtia in America. Contr. Gray Herb. 187:23—52. 1962. Taxonomic fern song Il. Pityrogramma (including Trismeria) and Anogramma. Contr moh Herb. 189: 52— pe A revision of the genus Cyathea. Contr. Gray Herb. 206: 19-100. . 1986. The biogeography of species, with Bot. Rev. 52:117-156. Wacner, D.H. 1979. Systematics of Polystichum in western North America, north of Mexico. Pteridologia 1:1— WAGNER, W. H., JR. 1952. The fern genus Diellia. Its structure, affinities, and taxonomy. Univ. Calif. Publ. Bot. 26:1—211. and F. S. WAGNER. 1983 Botrychium (Ophioglossaceae). Taxon 32: 51— pee M. J. 1986. A reevaluation of ge oo and its relationship to L. idulum eon a Syst. Bot. 11:263-27 WERTH, é ~ S. 1. GUTTMAN, and W. H. ESHmAUGe. mo Electrophoretic hn of reticulate 10:184-— 7.99.20 1nn LUV, i“ ai per eee Pune £ ae | y of new worida evolution inthe pan aint A x. Sys WILcE, J. H. Section Complanata of the genus Lycopodium. aa Nova | 19:1-233+ 40 pla pocorn. C 1977. Synopsis of the genus Sphaeropteris (Cyatheaceae) with a revision of the neti oni soon ed species. Bot. J st. 98:176—198. Varaxiev¥ch G. 9. Systematics of the fern genus Cyrtomium in the new world. Ph.D. ues iy University, Bloomington > B. STEIN, and G.J. GasToNny. 1988. Chicroplest DNA evolution and systematics of Phanerophlebia iDeyopterdacese) and related fern genera. Proc. Natl. Acad. Sci. U.S.A 85:2589 American Fern Journal 79(2): 46—54 (1989) Cryptic Species, Species Delimitation, and Taxonomic Practice in the Homosporous Ferns CaTuy A. PARIS Department of Botany, University of Vermont, Burlington, Vermont 05405 FLORENCE S. WAGNER and WARREN H. WAGNER JR. Department of Botany, University of Michigan, Ann Arbor, Michigan 48109 Biologists generally agree that species are to be delimited on the basis of genetic discontinuities. The two species concepts that depend on such discontinuities to delimit species are the biological and the evolutionary species concepts. A biological species is a group of interfertile populations that is reproductively isolated from other such groups and that occupies a specific niche in nature (following Mayr, 1982). An evolutionary species is a single lineage of ancestor-d lant populations that maintains its own identity from other such lineages, that fits into its own ecological niche, and that has a unique evolutionary history (Simpson, 1961; Grant, 1981; Wiley, 1981). It thus differs from the biological species concept in that it is equally applicable to both sexually and asexually reproducing organisms. Under both the biological and the evolutionary species concepts, genetic discontinuities between sister species are thought to arise stochastically following speciation. It is assumed that as time passes, the two diverge progressively in a suite of morphological, physiological, and ecological attributes. Although most botanists espouse an evolutionary species concept in their theoretical writings, in their classifications they often recognize only species that have distinctive structural characters by which the taxa can be identified. In so doing, they are employing the morphological species concept of their predecessors. Thus a conflict arises between theory and practice. As long as evolutionary species are structurally well differentiated, there is usually good agreement between species defined using either biological or morphological criteria. After all, a morphological species is ‘‘an inference as to the most probable limits of the biological species’’ (Dobzhansky, 1951), and the gaps by which such species are recognized are presumed to have arisen along with the reproductive isolation between them. But in the case of cryptic species, morphologically similar or identical natural populations that are reproductively isolated (Mayr, 1970), a species defined using morphological criteria comprises two or more genetically isolated evolutionary species. The practical consequences of unrecognized cryptic species range from a simple underestimate of diversity in a study group to effects of substantial economic importance. For instance, the parasitic Hymenopteran species Aphytis maculicornis, imported into California to control olive scale, has been found by Hafez and Doutt (1954) to comprise at least three cryptic species. Of these, only one has proven effective as a biological control agent (DeBach, 1969). Another example, from human epidemiology, is that of the European mosquito PARIS, WAGNER & WAGNER: CRYPTIC SPECIES 47 Anopheles maculipennis. This species was for a long time considered the vector of malaria among human populations in Europe. More recent studies have shown that six cryptic species occur within A. maculipennis, only two of which (A. labranchiae and A. sacharowi) carry malaria among humans (Mayr, 1970). In systematic research on the ferns, new data from morphometrics, cytogenetics, and electrophoresis are greatly improving our ability to resolve problems in taxonomically difficult groups. In the process of resolving them we are discovering cryptic species in many lineages where they had previously been unsuspected. Satisfactory taxonomic treatment of these cryptic species depends on a resolution of the conflict between the evolutionary species concept we embrace in theory and the morphological species pt we common ly employ in practice. CHARACTERIZATION AND RECOGNITION OF CRYPTIC SPECIES Cryptic species (also called sibling species by some authors, e.g., Mayr, 1963) were defined by Stebbins (1950) as “... population systems which were believed to belong to the same species until genetic evidence showed the existence of isolating mechanisms separating them.” Grant (1981) defined them as ‘‘... good biological species which are virtually indistinguishable morphologically.” Wiley writes: “Cryptic species . . . are species that cannot be diagnosed by morphology, but that act as independent evolutionary lineages in nature” (Wiley, 1981). Although each author’s definition reflects his particular approach to the species question, several elements are common to all three. Cryptic species have the following characteristics: 1. They are poorly differentiated morphologically. 2. They represent distinct evolutionary lineages because they are reproductively isolated. 3. They have historically been misinterpreted as members of a single species. These characteristics provide a set of criteria by which putative cases can be evaluated. A classic example from the genus Drosophila illustrates application of these three criteria to cryptic species of fruit flies (Dobzhansky and Epling, 1944; Dobzhansky, 1951). Populations of D. pseudoobscura and D. persimilis were originally treated as one species until researchers observed hybrid sterility barriers between them in laboratory cultures. Once reproductive isolation of the two species was recognized, subtle differences between them were found in attributes of the wings and male genitalia. A series of subsequent studies identified additional differences in physiology, behavior, and chromosome morphology. In the following discussion we use the criteria of subtle morphological differentiation, reproductive isolation, and historical confusion to evaluate the evidence for cryptic species in several problematic genera of ferns. Reproductive isolation is used as the primary criterion for determining whether or not two 48 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) morphologically similar populations represent cryptic species. In the absence of such evidence, the alternative explanation, that the populations are merely infraspecific variants, is favored. Discussion of cryptic species in plants often centers on polyploids and their diploid progenitors (e.g., Grant, 1981: see also Barrington et al., 1989). This paper, however, addresses the conceptually different problem of cryptic species at the same ploidy level. We also explore the factors that have obscured species boundaries in the ferns and demonstrate how new systematic methods are increasing our ability to define those boundaries. Finally we consider the practical problems posed by cryptic species to fern systematics and taxonomy. RECENTLY RESOLVED CRYPTIC COMPLEXES IN THE FERNS A. The Adiantum pedatum complex Cryptic species have recently been documented within the Adiantum pedatum (maidenhair fern) complex in eastern North America. They are the typical maidenhair of rich deciduous woods and the serpentine maidenhair, a small serpentine endemic with fastigiate axes. Both are diploid (Paris, 1986; Paris & Windham, 1988). The serpentine diploid has traditionally been interpreted as an infraspecific variant of the typical maidenhair (Fernald, 1905: Cody, 1983). Other authors have not recognized the taxon, presumably because many of the characters used to diagnose it are susceptible to environmental modification (Fernald, 1905; Wylie, 1949; Paris, unpubl. data) and because for most key characters it overlaps with the typical maidenhair (see Lellinger, 1985). Isozyme data, however, have demonstrated that the diploids are well differentiated genetically (Paris & Windham, 1988); indeed, at several loci, the two taxa share no alleles in common. The average Nei’s genetic identity for populations of the serpentine and non-serpentine maidenhairs is only 0.495, a value comparable to those available for congeneric fern species (Haufler, 1987) and quite low as compared with angiosperm congeners (I = 0.67, Crawford, 1983). Isozyme electrophoresis has also permitted the detection of still another maidenhair—a tetraploid population at Belvidere Mountain in Vermont. Additivity in the tetraploid of isozyme markers for the diploid taxa showed that the tetraploid was derived from a hybrid between the serpentine and the typical maidenhairs. Gametophyte progeny tests indicated that heterozygosity is fixed in the allotetraploid. Non-pairing of the two genomes provided further evidence that the serpentine and the typical maidenhair ferns represent distinct species (Paris & Windham, 1988). With a new understanding of these entities and their relationships, morphological characters were reevaluated in a discriminant analysis. Three morphologically distinct groups corresponding to the taxa emerged, with the tetraploids occupying an intermediate position between their diploid progenitors (Paris & Windham, 1988). PARIS, WAGNER & WAGNER: CRYPTIC SPECIES 49 It is now evident that the serpentine and the typical maidenhair ferns differ in a number of morphological characters, but that differences between them were obscured by the unrecognized tetraploid and by phenotypic plasticity. Thus they represent a good example of cryptic species: as was the case with Dobzhansky’s drosophilas, once the isolation of the taxa was perceived, morphological characters were found to differentiate them. B. Botrychium subg. Botrychium Additional examples of cryptic species are available from recent systematic work in Botrychium subg. Botrychium (moonworts) in western North America (e.g., Wagner & Wagner, 1983a, 1983b). Botrychium isa taxonomically difficult genus in a number of respects: first, in subgenus Botrychium many of the plants are small and so are overlooked by collectors. In consequence, they tend to be poorly represented in herbaria. Furthermore, morphological characters are difficult in subgenus Botrychium: not only are they quite subtle, requiring careful definition and comparison, but also they are readily influenced both by the age of the sporophyte and by the habitat in which it grew. The susceptibility of Botrychium species to environmental modification is best demonstrated by variation within B. simplex, the typical form of which is moderately dissected and occurs in dry upland fields. In deep forests and at bog edges, a delicate, relatively undissected form occurs (“‘var. tenebrosum”’ of Clausen), whereas a robust and ternately dissected form (Milde’s ‘‘var. compositum”) is found in low, moist meadows. In some Botrychium taxa, morphological variation i lly controlled. A well-known example is provided by B. dissectum forma dissectum and B. dissectum forma obliquum (subg. Sceptridium), morphologically distinct entities that maintain their distinctive characteristics when growing side by side. These forms were originally thought to be species, then varieties. When it gnized that the two are fully interfertile and that a range of intermediates exists, they were relegated to forms (Fernald, 1921). Given such complex and apparently inconsistent patterns of variation in Botrychium, how can sound taxonomic judgments be made? Common garden experiments, useful in addressing such questions in other taxa, are problematic in Botrychium because complex and sensitive mycorrhizal relationships make them difficult to grow. Also, botrychiums are very slow-growing, producing but one leaf per year. Nevertheless, there are two approaches that permit the genetic and environmental components of variation to be distinguished in Botrychium. These are the Genus Communities Method and the Method of Mutual Associations (Wagner & Wagner, 1983a). The Genus Communities Method is based on the tendency of congeneric species to grow together in the same habitats. These genus communities provide a natural common garden experiment: if problematic taxa maintain their differences consistently and persistently when growing together, it is evidence that their morphological differences are genetically fixed. The method was used successfully to differentiate Botrychium hesperium and B. echo (Wagner & 50 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) Wagner, 1983a, 1983b). Both species grow, often together, in the southern Rockies at elevations between 2,500 and 3,500 m, in rocky soil on grassy slopes, along roadsides, and at lake edges. The taxa are so similar that, although the original collectors noticed their differences, they did not initially recognize them as species. However, extensive fieldwork with large population samples demonstrated consistent differences in a number of gross- and micro- morphological characters and in phenology (Wagner & Wagner, 1983b). These differences were maintained where the species grew together, confirming that their distinctive characteristics are heritable. Further evidence for the reproductive isolation of B. hesperium and B. echo is provided by the sterility of the rare interspecific hybrids that are found at sites where the two species co- occur. The basis of reproductive isolation between Botrychium species has only recently been elucidated. Studies of genetic variation in Botrychium virginianum (Soltis & Soltis, 1986), B. dissectum (McCauley et al., 1985), and B. simplex (W. Hauk, pers. comm.), which together represent all three North American subgenera, suggest that these species are selfers (intragametophytic self-fertilization, sensu Klekowski, 1969). Given that all Botrychium species have bisexual subterranean gametophytes, perhaps selfing is common in the genus, and represents the major isolating mechanism between homoploid species in genus communities. The Method of Mutual Associations, closely related to the Genus Communities Method, is useful if the two taxa of interest do not grow together in the same place. In such cases, a third taxon is brought into the picture and is used as an assay for variation in the other two. The method is based on the principle that if taxon A grows together with taxon C in one place, and if taxon B grows with C in another place, and C is morphologically uniform from habitat to habitat, then the differences between A and B are probably genetically fixed (Wagner & Wagner, 1983a). The Method of Mutual Associations was used effectively to differentiate B. mormo, the little goblin fern, from the dwarf form of B. simplex, which it closely resembles. Because B. mormo and B. simplex do not occur together, B. minganense, which occurs with each, was used as the assay species and permitted the recognition of a number of genetically based differences between the two (Wagner & Wagner, 1983a). In Botrychium subg. Botrychium, unlike the other examples given so far, the identification of subtle morphological differences between similar species preceded the demonstration of their reproductive isolation. Nevertheless, these moonworts meet the three criteria of cryptic species and may usefully be numbered among them. C. The Pityrogramma triangularis complex Cryptic species within the Pityrogramma triangularis (goldback ferns) complex of western North America are especially problematic because so far they appear to be morphologically indistinguishable, Alt and Grant (1960) made the first attempt to resolve biosystematic relations} ips within the Pityrogramma PARIS, WAGNER & WAGNER: CRYPTIC SPECIES 51 triangularis complex (1960). Within P. triangularis var. triangularis, they found two quite similar yet phologically separable diploids, which they called type Aand type B. A previously unknown tetraploid taxon discovered in their sample was interpreted, on the basis of leaf morphology, as the derivative of a hybrid between types A and B. Chromosome pairing behavior in the tetraploid and backcross triploids suggested an alloploid origin for the tetraploid and indicated that types A and B were reproductively isolated (Alt & Grant, 1960). Subsequent chemosystematic studies of P. triangularis flavonoids (Star et al., 1975a, 1975b; Smith, 1980) have shown that the situation in P. triangularis var. triangularis is even more complex than Alt and Grant recognized. Var. triangularis comprises an array of six distinct flavonoid chemotypes, two among the diploids and four among the tetraploids: 2X chemotypes: ceroptin kaempferol 4'-methy] ether 4X chemotypes: ceroptin kaempferol 4’-methy]l ether and 7,4’-dimethy] ether kaempfero galangin 7-methy] ether The partial to full sterility of inter-chemotype hybrids found in the study provides evidence for at least partial genetic isolation of the chemotypes at each ploidy level (Smith, 1980). The extent to which morphological and phytochemical characters are correlated in the P. triangularis complex is so far unknown, as is the relationship between types A and B of Alt and Grant and the chemotypes of Smith. In consequence, it is unclear whether there are good field characters by which reproductively isolated taxa within var. triangularis can be recognized. The challenge to taxonomists posed by the cryptic entities in Pityrogramma triangularis was recognized both by Alt & Grant (1960) and by Smith (1980). The latter summarized the situation in these words: “One dilemma met here is a common tormentor of vascular plant systematists: namely, how can one rationally treat taxonomically those members of a group which, even though they are reproductively isolated, nevertheless are distinguishable only by complex methodology generally beyond the reach of most people interested in identification, while at the same time imperfectly isolated entities receive formal taxonomic status because they possess superficially distinguishing marks.” We consider Smith’s dilemma in the last section of this paper. ARE CRYPTIC SPECIES COMMON IN THE FERNS? At present it is not clear whether cryptic species are more common among ferns than other groups of organisms. The results of recent studies suggest that they are: in addition to the examples discussed above, cryptic species have also 52 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) been detected in Cystopteris (Haufler, pers. comm.), Gymnocarpium (Pryer & Windham, 1988), Polypodium (Haufler & Windham, 1988), and Woodsia (Windham, 1987). This flush of recent discoveries presents the impression that cryptic species are especially prevalent in ferns, but that impression may be an artifact of the recent incorporation of biosystematic data into systematic studies in the ferns. According to Mayr (1970), cryptic species are probably common in many groups of organisms and will be detected with increasing frequency as sensitive methods such as isozyme electrophoresis are applied to systematic problems. Perhaps, on the other hand, cryptic species do occur more commonly in the ferns than other kinds of organisms. This may be because in the ferns there is no selection for visual recognition cues during speciation. Many angiosperm species are pollinated by animals that rely on visual cues for the recognition of species. Speciation may involve a shift from one pollinator to another and concomitant evolution of a new set of visual cues. In the ferns, however, animals have no role in the movement of gametophytes. Selection is not therefore expected to elicit the evolution of novel recognition characters in new species. Mayr (1963) summarized the relationship between the prevalence of cryptic species and the means of mate recognition ina group: “Sibling species are apparently particularly common in those kinds of species in which chemical senses (olfactory and so on) are more highly developed than the sense of vision. Although indistinguishable to the eye of man, these sibling species a idently dissimilar to each other, as is shown by cross-mating experiments. Sibling species are apparently rarest in organisms such as birds that are most dependent on vision in the role of epigamic characters.”’ especially difficult to detect in ferns because in many lineages, such as Adiantum and Botrychium, reticulate evolution and phenotypic plasticity have blurred species boundaries. Also, because ferns have a relatively simple plant body, few structural characters are available for taxonomic analysis (Haufler, 1985, 1987). Thus the pteridologist working with prepared specimens may have less power to resolve problematic complexes than do specialists in other groups such as angiosperms. CRYPTIC SPECIES AND TAXONOMIC PRACTICE IN THE FERNS The examples above demonstrate that cryptic species are good biological and evolutionary species. They are not very satisfactory morphological species, however, because they are virtually indistinguishable using structural characters. How one treats cryptic species taxonomically, then, depends on one’s species concept. Whereas most investigators agree that a system of classification should reflect as nearly as possible the phylogenetic relationships of the taxa being classified, there has always been argument about the extent to which this objective is PARIS, WAGNER & WAGNER: CRYPTIC SPECIES 53 possible, let alone practical. Although we recognize that they will not be applicable to every situation in the ferns, we present the foll the taxonomic treatment of cryptic species. We suggest that cryptic species represent independent evolutionary lineages and so deserve species names. At the same time, we acknowledge the need for a multipurpose classification, one useful for the herbarium curator, the conservation biologist, and the park naturalist, as well as the specialist. We recommend that reproductively isolated taxa be given species names if morphological characters have been found to differentiate them, even if the character differences are very subtle. Those investigators who have a specific interest in the group and can differentiate the cryptic entities will thereby have the means to communicate about them. For purposes of routine identification, however, specimens can be keyed out to species group, with an indication that two or more cryptic species may be represented in the sample (Ross, 1974; Grant, 1981; Wiley, 1981). In the case of species that are so far indistinguishable without recourse to special methodology, as in the chemotypes of Pityrogramma triangularis, species epithets may be superfluous. Although these taxa are evolutionary species, the practical problems of identifying specimens in such cases makes the names useless to all but the chemosystematist. Even though the taxa are unnamed, manuals and floras should note that biochemically differentiated species exist within the complex. It is probable that with continued study, characters will be found to separate the cryptic entities; we recommend that the species then be named. ACKNOWLEDGMENTS We are grateful to David S. Barrington and Lynda Goldsmith for their helpful reviews of this manuscript. CAP thanks Drs. R. M. Tryon, Jr. and E. Mayr for thought-provoking discussions of h Iped to cryptic species. A grant-in-Aid of Research from Sigma Xi, T’ e support the work on Adiantum. The 'Botrychium research was carried out in large part ‘under NSF Grant DEB 8005536, “Evoluti p Botrychium.”’ LITERATURE CITED ALT, K.S. and V. Grant. 1960. Cytotaxonomic observations on the goldback fern. Brittonia 12:153—170. BARRINGTON, D. S., C. H. HAUFLER, and C. R. — 1989. Hybridization, reticulation, and species — in the: ferns. Amat, Fern J. 79:55— Copy, W.J.198 I calderi bsp Woche 85: 93-96. CRAwrForD, D.J. 1983. Phylogenetic and systematic inferences from electrophoretic studies. Pp. 57-287 in Isozymes in plant genetics and breeding, Part A, eds. S.J. Tanksley and T. J. Orton. Amsterdam: Elsevier. DEBacu, P. 1969. Uniparental, sibling, and semi-species in relation to taxonomy and biological control. Israel J. Entomol. 4:11-28 DoszHANSKy, T. 1951. Genetics and the origin of species, 3rd ed. New York: Columbia University Press. and C. Eptinc. 1944. Contributions to the genetics, taxonomy, and ecology of Drosophila pseudoobscura and its relatives. Carnegie Inst. Wash. Publ. 554:1-183. tl t North America tN 54 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) FERNALD, . ae 1905. Analpine sigecwaienser Rhodore 7:190—192. —_——__, 1. The Gray Nova Scotia, 1920. Rhodora 23:130-152. GRANT, V. Tee oe eo and ed. New York: Columbia University Pres Harez, M. and R. L. Doutr. 1954. Biological evidence of atgud ae in ADAGE maculicornis (Masi) iluabsien. Aphelinidae). Canad. Entomol. 86:90—96. HAUFLER, C. H. 1985. Pteridophyte jovpincenng biology: the aeduailias approach. Proc. Royal Soc. Edinburgh 86B:315 . 1987. Electophoress is fan our concepts of evolution in homosporous ee Amer. J. Bot. 74:953—966. iy Windham. fae Exploring the origin of Polypodium vulgare, again. Amer. J. Bot. Kuskowsx E 2 JR. 1969.R luctive biology of the Pteridophyta, II. Theoretical considerations. nn. Soc., Bot. 62:3 —359. LELLINGER, he B. 1985. pee fa of the ferns & fern allies of the United States & Canada. pbs viian re C; geese ogee ee Press. Mayr, E. 19 Cambridge, MA: Belknap P f Harvard University eat 1970. sa ce eal species, and evolution. Cambridge, MA: Belknap Press of Harvard University Pres . The sow of biological thought. Cambridge, MA: Belknap Press of Harvard University res McCauLe LEY, D.E., D. B. ine wil M. REILLY. 1985. Inbreedi 1 tl te of self-fertilization in a grape fern, Botrychiu um d issec tl Apier Le Bot. Ta: 1978- 1982. Paris, C. A.1986. Ab ti plex i t North America. M.S. Thesis, Rainer of Vermont, Burlington Vi. and M.D. WinpHaM. 1988. A biosystematic investigation of the Adiantum pedatum complex in eastern Nau America. Syst. Bot. 13:240—255. PRYER, K.M. and M.D. WinpHaM. 1988. A Apecemagia of Gymnocarpium dryopteris (L.) Newman in North America. ae J. Bot. 75:142. Ross, H. H. 1974. ea syst eading oe Addison-Wesley SIMPSON, G. G. 1 a] fanimal taxonomy. New York: Columbia Hap msapntd a SMITH, D.M.1 a tua y the Pityrog g - Bull. Torrey Bot. Club 107:134-145. Sottis, D. E. and P. S. Sottts. 1986. Electropt id fi I ling in the fern Botrychium virginianum (Ophioglossaceae). Amer. J. Bot. 73:588—592. mide et ee oe H. Roster, T. J. Masry, and D. M. SM MITH. 1975a. Fl id tin pi ts f i shee Seceese 14: ner 2278. , D. SIEG T. Masry, and D. M.S. MITH. 1975b id d tetraploid of two exudate chemotypes of Pityrogramma triangularis. Biochem. Syst. Ecol. 7109-1 STEBBINS, G. L. 1950. Variation ates chins in plants. New York: Columbia University Press. Wacner, W. H., JR. and F. S. W 83a tematic tool in the stud new world Botrychium | (Ophioglossaceae). Taxon 32: 51-63. and Botrychium h ; d sed w vith it. Amer Tt. Fern J. 73:53—62. WItey, E. na 1981. Phylogenai the tevin and practice of phylogenetic systematics. New York: John Wiley & Son WINDHawy, M. D.1 1987, cme and electrophoretic studies of the genus Woodsia in North rica. Amer. J. Bo WYtrE, R. B. ae Variations in simian — among Adiantum pedatum plants growing in a rock cavern. Amer. J. Bot. 3 287. y we American Fern Journal 79(2): 55—64 (1989) Hybridization, Reticulation, and Species Concepts in the Ferns Davip S. BARRINGTON Department of Botany, University of Vermont, Burlington, Vermont 05405-0086 CHRISTOPHER H. HAUFLER Department of Botany, University of Kansas, Lawrence, k 66045 CHARLES R. WERTH Department of Biological Sciences, Texas Tech University, Lubbock, Texas 79409 Hybrids and hybrid species are common among ferns, and they account for many of the problems in species definition in the group. Most systematic inquiry into the evolutionary process in ferns has addressed hybrid species, because meaningful explanations of their origins are feasible (Manton, 1950). As aresult, complexes of hybrids, hybrid species, and their progenitor species have been popular subjects for experimental work. Here, we address the definition and changing perception of these hybrid species in the light of improvements in the data available to systematists. Once we have established basic definitions, we demonstrate the utility of recent advances in defining hybrid species of ferns. With this orientation, we investigate the status of hybrid species in the context of reigning species concepts. Renewed reproductive interaction between populations or species following a period of isolation characterizes all hybrids; hence hybrids are often spoken of as the products of secondary contact. Hybrids are unique in that they arise when isolating mechanisms fail; thus they are evolutionarily a consequence of the disruption of the divergence process that leads to ordinary (primary) species. Consequently, the hybrid is at once a novelty and a rehash: it is a novel combination of genetic and morphological features already present in its progenitors. These features need not be intermediate: see Grant (1975) on transgressive segregation and Barrington, 1986a. Fern hybrids are predominantly sterile (Knobloch, 1976), though there is a small, disparate set of variously fertile hybrids (in Pteris, Walker, 1958; in Dryopteris, Whittier & Wagner, 1961; in the Cyatheaceae, Conant & Cooper-Driver, 1980). The origin and evolutionary significance of sterile hybrids have been the subject of most studies relevant to a discussion of species concepts in pteridophytes (Lovis, 1977). Many fern species are thought to be derived from hybrids. Traditionally, these taxa are argued to be 1) species because they breed true and they are autonomous and 2) hybrid because comparison with allied taxa yields evidence of intermediacy. Both allopolyploid and allohomoploid species have been reported in the ferns. The allopolyploid species is an old concept with much experimental support (for a historical synopsis see Manton, 1950). New allopolyploid species ordinarily arise from sterile hybrids via doubling of chromosome sets and consequent restoration of fertility. The typical angiosperm 56 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) route to polyploid species, via function of unreduced gametes (deWet, 1980) is by contrast rare in the ferns (but see Gastony, 1986). Allopolyploids have been documented with extensive work in numerous fern genera, especially in the Dryopteridaceae and Aspleniaceae (reviewed by Lovis, 1977). In contrast, the allohomoploid species is an idea new to fern evolutionary biol gy (Conant & Cooper-Driver, 1980) in need of further testing. New alloh ploid species ari via isolation of fertile F, and recombinant hybrids by intragametophytic selfing and/or geographic isolation. Conant and Cooper-Driver (1980) have presented evidence for allohomoploid speciation in species of the Cyatheacaeous genus Alsophila (including Nephelea). The uniform haploid chromosome number of n=69 (Love et al., 1977) among all scale-bearing Cyatheaceae is consistent with the idea that allohomoploid hybrid speciation has predominated in the family. o emphasize the significantly different phylogenetic origin of hybrid species, Wagner (1969, 1983) has suggested that hybrid species (nothospecies) should be recognized as qualitatively different from divergent species (orthospecies). He argues that nothospecies are qualitatively different because 1) their origin is a consequence of a return to sympatry and the breakdown of isolating mechanisms, not the action of natural selection and drift on allopatric populations, 2) they originate during one or more discrete episodes each involving only two individual parents, and 3) they are not novelties but combinations of previously existing entities. Wagner’s criteria for distinguishing hybrid species are that they have a pattern of character states intermediate between their progenitors, that they be distinct enough from their parents to be treated at the same taxonomic level as their progenitors, and that their reticulate origin be documented. Certainly the origin of hybrid species is distinctive, but distinguishing hybrid species from divergent (primary) species depends on our capacity to recognize the features of hybrid species that document the process by which they originated from their progenitors. Our purpose is to explore the status of taxa now regarded as hybrid species: how have these taxa been treated taxonomically, and do they meet criteria for species definition? oe THE RESOLUTION OF Hysrips, Hysrip SPECIES, AND RETICULATE COMPLEXES Successive technical and theoretical advances have led to increased resolution of complexes of hybrids, hybrid species, and their progenitor species. The earliest realization that plants hybridized was based on a morphological criterion—intermediacy (see Wagner, 1983, and Barrington, 1986a, for discussion of intermediacy). For instance, Berkeley (1866) proposed that Asplenium ebenoides was A. platyneuron x A. rhizophyllum, citing its intermediacy as evidence. Eaton (1879) detailed the characters and invited a test by reconstitution in culture, which Slosson (1902) accomplished in a careful set of experiments. Morphological comparison of reconstituted hybrids with their putative wild counterparts has since been a popular test, especially with European workers (e.g., Walker, 1961; Lovis, 1968). However, morphology alone has not been adequate to allow discrimination of hybrids and hybrid species in complexes with little structural divergence between progenitor species (Paris et BARRINGTON, HAUFLER & WERTH: HYBRIDIZATION 57 al., 1989). Further, purely morphological criteria do not provide sufficient basis for robust hypotheses about the origin and evolution of hybrid species (Thorpe, 1984). Morphological criteria related to spore abortion have been used to document fern hybrids (e.g., Tryon, 1948; Wagner & Chen, 1965; Wagner et al., 1986). Features such as collapsed, unopened sporangia, irregular spore shape and size, and failure of indusium eversion have proven to be powerful evidence, in combination with morphological intermediacy, of hybridity. Evidence from chromosome number and pairing behavior have long provided criteria for recognizing hybrids and hybrid species and distinguishing them from their progenitor species (Manton, 1950). The combination of morphology and chromosomal studies has yielded fully resolved evolutionary hypotheses for several complexes, for example the European polystichums (Manton, 1950) and the Appalachian aspleniums (Wagner, 1954). In these complexes, the diploid progenitor species are all extant and distinctive, and their derived tetraploids are readily discernible using qualitative morphological comparisons. Summation of biochemical markers provides a powerful basis for discriminating populations of hybrids and hybrid species from phylogenetically patristic intermediates, which can resemble hybrid taxa (Endler, 1977). Two general kinds of markers have been explored, phenolic compounds and isozymes. Interpretation of phenolic data is relatively easy, compared to morphology, because hybrids and hybrid species sum marker compounds characteristic of their progenitor species. For example, Smith and Levin (1963) used chromatography of flavonoids to confirm Wagner’s hypothesis for reticulate evolution in the Appalachian Asplenium complex. Chromatographic analysis of phloroglucinols has yielded similar confirmations in the genus Dryopteris (reviewed in Euw, 1980). However, work on phenolics has proven inadequate to solve problems in many reticulate complexes. Chromatographic patterns are often not clear, the status of chromatographic spots as homologous character states is uncertain, and there is not enough variability in these secondary compounds to provide sufficient species-specific markers. In the past seven years, allelic variants of isozymes (allozymes) have been used as genetic markers for the study of relationships among hybrid species of ferns. Data on the electrophoretic mobility of these allozymes have been used to 1) confirm or choose between established hypotheses of reticulate relationships, 2) resolve difficulties in poorly understood complexes, and 3) reveal previously unsuspected complexes. ; Several examples of hypotheses confirmed or chosen are now available. Comparison of isozymes among species of the Appalachian Asplenium complex (Werth et al., 1985a) has confirmed the relationships proposed by Wagner (1954). More recently isozyme evidence has been used to decide the ancestry of Dryopteris celsa, an allopolyploid hypothesized as originating from hybridization of D. ludoviciana with either D. goldiana (Walker, 1962; Wagner, 1971) or D. marginalis (Hickok & Klekowski 1975). Werth (1989) presented isozyme data that strongly support D. goldiana as the second progenitor of D. celsa: the data showed D. celsa to sum m ker alleles for D. ludoviciana and D. 58 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) goldiana—electrophoretic markers for D. marginalis were missing. Gastony (1986, Gastony & Windham, 1989) was able to distinguish between two alternative hypotheses for the allopolyploid origin of Asplenium plenum by using isozyme electrophoresis. Recently, Barrington and Conant (unpubl. data) have used isozyme electrophoresis to confirm the hybrid status of the proposed Alsophila amintae x A. portoricensis, an important first step in testing Conant and Cooper-Driver’s (1980) hypothesis for allohomoploid speciation in tree ferns. Isozyme studies have also been effective in resolving problems in reticulate complexes, especially in those in which some progenitor species pairs are cryptic species (Paris et al., 1989). Haufler has advanced the resolution of the problematic complex in North American Cystopteris using isozyme electrophoresis. He demonstrated that the poorly understood Cystopteris tenuis (also known as Cystopteris fragilis var. mackayi) is an allotetraploid species derived from C. protrusa of the central Appalachians and a diploid, which although unknown as yet, is also implicated in the origin of the wide-ranging tetraploid C. fragilis (Haufler, 1985). The clear qualitative nature of the marker allozymes also allowed the strong inference that C. tenuis and C. fragilis share a diploid ancestor, a distinction that had not been tenable using only morphological and cytological data. Finally, isozyme studies have revealed several hybrids and hybrid species that were previously unresolved at any taxonomic level. A previously unsuspected reticulate complex was recently discovered in Central America. Montane polystichums from Costa Rica include an allotetraploid species, Polystichum talamancanum, derived from an endemic diploid, P. concinnum, and an unknown second progenitor that shares a genome with the Andean tetraploid P. orbiculatum (syn. P. polyphyllum). The backcross hybrid is now known (Barrington, 1985a) as well as three other hybrids among central American montane species of Polystichum (Barrington 1985b and unpubl. data), suggesting that reticulation in the montane tropics progresses much the same as in temperate areas. Isozyme work has also revealed previously unsuspected reticulate evolution in the genus Adiantum (Paris & Windham, 1988; Paris et al., 1989 documented morphometric distinctions between typical Adiantum pedatum, A. pedatum subsp. calderi, and their derived allotetraploid using both Principal BARRINGTON, HAUFLER & WERTH: HYBRIDIZATION 59 Components Analysis and Discriminant Function Analysis. In another case, Barrington (1986a) used Principal Components Analysis to provide a clear definition of the differences between triploid Polystichum x potteri and its tetraploid progenitor P. braunii, with which it has commonly been confused. Critical morphometric work, combined with cytological and isozyme analysis, constitute an excellent array of techniques for the resolution of hybrids and hybrid species in polyploid complexes. As we have improved our perception of hybrids and hybrid species we have come to recognize more species, based on morphological, cytological, and biochemical criteria. In any given epoch, newly resolved taxa are accepted reluctantly, whereas those resolved in the last generation are supported vehemently. To most workers today, Dryopteris carthusiana and D. campyloptera seem distinct from D. intermedia, but specialists argue about the recognition of more recently resolved taxa, such as Cystopteris laurentiana and C. tennesseensis. Arguments about the correct taxonomic treatment of Cystopteris tenuis and the newly discovered allotetraploid ally of Adiantum pedatum have hardly begun. The question of relative divergence between progenitors of polyploid species in different evolutionary alliances, which is dependent on a well-documented sample of complexes, including those just being discerned, has yet to be seriously considered. HYBRID SPECIES AS SPECIES Do species taxa whose heritage includes hybridization meet criteria as species? We will address the biological reality and taxonomic utility of hybrid species in the context of each of three species concepts; morphological, biological, and evolutionary. Morphological species are the species of the field biologist and the herbarium taxonomist interested in discerning the diversity of life and developing conservation strategies for them. Hybrids and their derivatives have been difficult to treat as morphological species for five reasons. First, hybrids and their derivative species are confused with their progenitor species because they tend to be intermediate between them, and they often occur with them in similar habitats. Second, hybrid species may originate more than once, combining different structural variants of their progenitor species in each case. Third, at least allopolyploids tend to vary phenotypically toward their progenitor species in habitats where they occur together, the Vavilov effect (see D. Wagner in Barrington, 1985c). For instance, the allotetraploid Polystichum talamancanum varies toward its forest-dwelling progenitor in shade, but toward the alpine tetraploid P. orbiculatum (with which it shares a genome) when it grows in the sun (Barrington, 1985a and unpubl. data). Fourth, hybrid species commonly hybridize with their progenitors, yielding backcrossed plants that further obscure the morphological boundaries between species. Fifth, anomalous interactions between species yield cytological versions of the same hybrid: triploid and tetraploid versions of Polystichum x potteri are difficult to distinguish without morph tri lysis (Barrington, 1986b). In spite of these 60 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) confusing factors, morphological criteria, corroborated by chromosomal and enzyme data, allow the taxonomist to distinguish hybrid species. Mayr (1940, 1963) developed the biological species concept based on reproductive isolation and cohesiveness as a consistent category to be used in modern studies of populations and speciation. More recently, botanists have argued that at least among plants the biological species cannot be a real evolutionary entity because plant species evidence neither cohesiveness nor consistent isolation from other evolutionary units (Mishler & Donoghue, 1982). They suggest that we abandon the requirement that species evidence genetic cohesion and autonomy and demand only that species be monophyletic assemblages of populations. Hybrid species are certainly autonomous entities, since many of them reproduce independently of their progenitors. In terms of their reproductive isolation, hybrid species are good biological species, although the origin of the mechanisms isolating them differs from that of divergent species. Allopolyploid species are reproductively isolated from their progenitors because the backcrosses are sterile and therefore do not constitute a means of gene flow between the two progenitors. (However, recurrent polyploidizations do allow gene flow from diploid to polyploid.) Allohomoploid species may be geographically or reproductively isolated from their progenitors (Conant & Cooper-Driver, 1980). Thus, the sympatry involved in the origin of hybrid species is closely followed by the re-establishment of an effective mechanism isolating the hybrid novelty from its allies. Deciding whether or not hybrid species are cohesive will require more extensive data on fern population biology than are now available. Probably most hybrid species comprise populations derived from different hybridization events (Haufler & Soltis, 1986). For example, the allotetraploids Asplenium bradleyiand A. pinnatifidum show fixed heterozygosity for different allele combinations (Werth & Windham, in prep.). The apparent parallel origin of allotetraploids such as Dryopteris cam pyloptera (D. austriaca) in Europe and North America (Gibby, 1977) represents an extreme case of multiple origins that presents a challenge to applying the biological species concept to hybrid species. On the one hand, populations on two continents have a negligible chance of interacting reproductively, and the allotetraploid will continue to be generated from the diploids in each place as long as the two can hybridize. On the other hand, a careful analysis of the genetic and structural features of the allotetraploids and their diploid progenitor populations on both continents sometimes reveals that they are genetically and morphologically indistinguishable: Walker (1961) argued on the basis of synthesized hybrids that D. intermedia of eastern North America and D. maderensis of Madeira are conspecific. Comparative analysis of more allopolyploid populations from different continents is needed to resolve this issue. Note that hybrid species howing multiple origins would fail the modified species concept of Mishler and Donoghue (1982), since they are demonstrably polyphyletic. Simpson’s evolutionary species concept (Simpson, 1961) has been developed for plant biologists by Grant (1981) and applied to the needs of systematists using cladistic methods by Wiley (1981). Hybrid species meet the criteria for BARRINGTON, HAUFLER & WERTH: HYBRIDIZATION 61 evolutionary species: they have spatio-temporal identity and their own evolutionary tendencies and historical fate. However, they are unusual in that they originate without divergence. Unlike divergent species, hybrid species are derived from two lineages, not one, and they can displace their progenitors over time (Stebbins, 1971). This displacement may be of major importance in the ferns: allopolyploid ferns may be predominantly selfing and thus better at migration than their predominantly outcrossing progenitors (Haufler, 1989). Regardless of their origin and fate, hybrid species are spatio-temporal lineages, and thus are evolutionary species. Are there two distinguishable kinds of species, hybrid species (nothospecies) and divergent species (orthospecies), as Wagner contends (1969, 1983)? In the case of the allopolyploids, the answer to this question depends on an understanding of chromosomal evolution in the ferns. High chromosome numbers in the ferns have been interpreted in two ways: one school argues that they are the result of repeated polyploid speciation—progenitor extinction events (Haufler, 1987; Werth & Windham, in prep.), the other that they are primitively high (Haufler & Soltis, 1986; Soltis & Soltis, 1987), perhaps because selection for a greater number of linkage groups would reduce problems of homozygosity (Buckley & Lloyd, 1985). Development of the argument for repeated polyploidization depends on distinguishing between neo- and paleopolyploids (Wagner & Wagner, 1980). Neopolyploids are documented recent allopolyploids with chromosome numbers that are multiples of the lowest number now known for their evolutionary group; paleopolyploids are hypothesized ancient allopolyploids that have the lowest known chromosome numbers for their evolutionary group (that is, the base number)—they include the progenitors of the neopolyploids. Ferns with the lowest chromosome numbers for their genus have so far consistently shown diploid gene expression (Haufler & Soltis, 1986; Haufler, 1987; Wolf et al., 1987). If there have been repeated cycles of polyploidy, then there must have been consistent, exhaustive diploidization (gene silencing) of the paleopolyploids and concomitant extinction of the progenitor diploids. If gene silencing has indeed had a major role in fern evolution, it probably constitutes an isolating mechanism facilitating divergent evolution among allopolyploid species, since two populations, each silenced for reciprocal alleles of the same gene, would yield hybrid sporophytes with reduced gametophyte viability (Werth & Windham, in prep.). Consequently, hybrid speciation (secondary speciation) would be succeeded by speciation at the polyploid level (tertiary speciation—Haufler, 1989). Some evidence of silencing has begun to accumulate (Werth et al., 1985a,b; Elisens & Crawford, 1988; Werth & Windham, in prep.), and further inquiry may demonstrate that gene silencing has been important in polyploid evolution. At the moment neither interpretation of high basic chromosome numbers in the ferns is unequivocally supported: both should be further tested. If there have been repeated cycles of polyploidy in the evolutionary history of ferns, many species dubbed orthospecies (because they have the lowest known chromosome number for the genus) may be diploidized nothospecies. The problem is that all of the criteria for recognizing nothospecies are comparative: 62 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) they require analysis of both the hybrid species and its progenitors. Paleopolyploids cannot be identified as hybrid versus divergent species, since they have the base number for their evolutionary group, and their diploid progenitors must be extinct. The nothospecies is useful if neopolyploids are a new kind of species distinct from their progenitors. However, it is not useful if paleopolyploidy has had a role in the evolutionary history of fern species, since detectable nothospecies (neopolyploids) would simply be the latest products of an ongoing, uniform process. CONCLUSIONS Perception of the taxa arising from hybridization and allopolyploidy has changed dramatically in the last 120 years, and it is likely to continue to change as we develop more tools for discerning interactive, derivative species from their progenitors. From our synthesis, we conclude that hybridization yields species, be they allohomoploid or allopolyploid, since these entities meet morphological and biological criteria. Though the question, “Are hybrid species good species in the evolutionary sense?” requires more thought, the answer appears to be “Yes.” Hybrid species are lineages in space and time, in spite of their atypical origin and extinction. Whether there are two qualitatively different kinds of species, nothospecies and orthospecies, depends on whether there have been repeated cycles of polyploidy, since the nothospecies is a qualitatively different and useful concept only if there have not been repeated cycles. If tertiary speciation via reciprocal gene silencing is a significant mode of generating diversity, then hybrid species may be more similar to divergent species in evolutionary be treated as species if they meet the criteria for species and they can be ACKNOWLEDGMENTS We thank Cathy A. Paris for critical reading of the manuscript. DSB acknowledges the support of the University of Vermont Institutional Grants Program through Grants BSCI85-1 and BSCI87-3. CHH acknowledges the support of the National Science Foundation through its Grant 84-15842. oo acknowledges the support of the National Science Foundation through its Grant Number -11684. LITERATURE CITED BARRINGTON, D. S. 1985a. The morphology and origin of a new Polystict hybrid from Costa Rica. Syst. Bot. 10:199—204. . 1985b. Hybridisation in Costa Rican Polystichum. Proc. Roy. Soc. Edinburgh 86B:335-—340. - 1985c. The pr t evolutionary and t ic status of the fern genus Polystichum: The 1984 Botanical Society of America Pteridophyte Section Symposium. Amer. Fern J. 75:22—28. - 1986a. The morphology and cytology of Polystichum x potteri hybr. nov. (=P. acrostichoides x P. braunii). Rhodora 88:297— ; - 1986b. Variability in intermediate character states in fern hybrids. Amer. J. Bot. 73:733. BARRINGTON, HAUFLER & WERTH: HYBRIDIZATION 63 BERKELEY, M. J. 1866. On i hybrid fern from Philadelphia. J. Roy. Hort. Soc. 1:137—140, t.II rer I o 7 vf —3). — D. iy — M. Lioyp. 1985. Anew! tic system hypothesis. Amer. J. Bot oi9— eae . Ss. wa G. Cooper-DRIVER. 1980. Autogamous allohomoploidy in Alsophila and Nephelea (Cyatheaceae): A new hypothesis for speciation in homoploid homosporous ferns. Amer. J. Bot. 67:1269—1288. DEWET, i M. J. 1980. Origins of ek sea pp. 3—15 in Polyploidy: biological relevance. ed. W. H. wis. New York: Plenum eae c 1879. The ferns of Dostie? paren vol. I. Salem, M. husetts: S.E at ‘assino ELISENS, W. J. and D. J. CRAWFORD. 1988. Genetic variati d differentiation in the genus Mabrya (Sc phaleece amend): srabetntic and reer ae inferences. ‘Amer. J. Bot. ee. 7 ENDLER, J. A.1977.G } iati speciation and clines. Princeton, NJ: Pri ton Uni it Press Euw, J. von, M. LounasMaa, T. REICHSTEIN, and C.-J. Monee 1980. Chemotaxonomy in Dryopteris and related fern genera. Stud. Geobot. 1:275-31 — G. J. 1986. pasiaaansing evidence for the pen of fern species by unreduced spores. er. J. Bot. 73:1563—1569. M.D. Aaa ag ite The treatment and definition of ag: I pecies. Amer ey 79:65—77 Gipsy, M. 1977. The origin wot Dryopteris campyloptera. Canad. J. Bot. 55:1419-1428. GRANT, V. tig Genetics of Flowering Plants. New York: Columbia ir got Press. Plant speciation. New York: Columbia University HAUFLER, “s on 1985. Sipe a! evolutionary biology: the sea lee aE approach. Proc. Roy. Cc Edinburgh 86B:315—323 . 1987. Electrophoresis is modifying our concepts of evolution in homosporous pppoe ieee Amer. J. Bot. 74:953—96 ward a uae of Sica modes and mechanisms in homosporous peridonhses Biochem. Syst. Ecol. In press and D. OLTIS. 1986. Genetic evidence ‘indicates that sconces ae with high chamosome ae may be diploid. Proc. Natl. Acad. U.S.A. 83:4389— Hickok, L.G. and E.J. KLEKowskI, JR. 1975. Chromosome behavior in cyeas ferns: a reinterpretation of Appalachian Dryopteris. Amer. J. Bot. 62:560—569. KNOBLOCH, I. W. 1976. cnr geet hybrids. Publ. Mus. Michigan State Univ., Biol. Ser. 5:273-—352. Korr, L: ond D. BRITTON. -A comparative st ay of sainteging cen nig of the three f Polyp pie o. Canad. J. Bot. 60:1360—1 L6vE, A., D. LOVE, and R. E. G. PiIcHI- -SERMOLLI. 1977. itexoneica atlas of ie pore’ aduz, Federal Republic of Germany: J, Cram Lovis, J. D. 1968 gil. Fern hybridizing at the Uni ity of Leeds. Brit. Fern Gaz. 10: 13- on ——-. . Evolutionar | ferns. Adv. Bot. Res. 4:229—415. MANTON, : 1950. Probie of cytology and evolution in the Pteridophyta. Cambridge, England: mbridge University Press. Mayr, E. 1940 Speciation phenomena it in sap Amer Prien 74: 249- 278. SU mT en ver we srs - D. and M.J. DonoGHUE. 1982. Species concepts: a case for pluralism. Syst. Zool. 7491-503 Paris, C. ve and M. D. WINDHAM. 1988. A pea eee investigation of the Adiantum pedatum complex in eastern North America. Syst. Bot. 13:240—255. —, bil in Wa GNER, JR., a and F. S. WAGHES 180. dig 2 SiS species definition, and th Fern J. 79:46—54. Sh C. G. re Principles of animal taxonomy. se ene neon University Press. SLosson, M. 1902. The origin of Asplenium ebenoides. Bull. Torrey Bot. Club 29:487—495. 64 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) SMITH, D.M. and D.A. LevIN. 1963. nie pipes saga - reticulate evolution in the se Oo hoot Sie Asplenium I 952- Soxtis, D. E. and P. S. Sottis. 1987. Polyploid t i Pp Pteridophyta: a re-evaluation. Amer. Naturalist 130: 219-232. STEBBINS, G. L. ph Chromosomal —_— in higher plants. Reading, MA: Addison -Wesley. THORPE, R. S. 1984. Primary and se ondary transition zones in speciation and population diferentation ie co pit of range expansion. Evolution 38:233—243. TRYON, z fe JR. 1948. Some woodsias from the north shore of Lake Superior. Amer. Fern J. 139-170 WAGNER, W. H. , JR. 54 Reti Lot ] +} Ls ; Bunaliati 2.102 135: —_——.. ig The wy and taxonomic treatment of Sao Bioscience 19:785—789, 795. olution of Dryopteris in relation to the A ppalachians. Pp. 147-191 in The disbutonl history of the biota of the southern Appalachians. Part II. Flora, ed. P. C. Holt. Research Division Monograph 2. Blacksburg, VA: Virginia Polytechnic Institute and nse Univers ose v +} . Sg ean MN cee Js} oars 8 sf . pees in Advances in cladistics, vol: 2, ed. N. I. Platnick and V. A. Funk. New York: Columbia “sp Sages Press and K. L. CHEN. 1965. Abortion : at tleg and sporangia as a tool in the detection of Dryopters ide. Amer. Fern J. 5 d F.S. WaGNER. 1980. Palyaiekay in pteridophytes. Aa 199-214 in Polyploidy: eiaisaeal relevance, ed. W. H. Lewis. __ York: Plenum Press , F. S. WAGNER, and W. C. TayYLor. 1986. Detecti bortive sy rbari is of wees vig Amer. Fern J. 76:129-140. WALKER, S. 1961. Cytogenetic studies in the Dryopteris spinulosa complex. II. Amer. J. Bot. pelea 1962. Further studies in the sage sil ed the origin of D. clintoniana, D. celsa, and slated taxa. Amer. J. Bot. 49:497—503 WALKER, T. G. 1958. bbl teas is species of Pteris. Evolution 12:82—92. WERTH, C.R. 1989. The use of isozyme data for inferring ancestry of polyploid species of pteridophytes. Biochem. Syst. and Ecol. In pres: oe sie a WINDHAM. Evidence of gene silencing in polyploid remcsdes implications l lev el. I : in Gy ecmocespearsresiaagcots 1985a. R ‘ing gins of il i iyplbtd pecies in Asplenium. — 228: i ——., —————. 1985b. Electrophoretic — of reticulate evolution in the Apel a dele ana Syst. Bot. 10: 184-1 WHITTIER, and W. H. WaGNneR, JR. 1961. The \ i 1 ination in Dryopteris tf o WILEY, Be O. 1981. Phylogenetic systematics. poten gg J. Wiley & Sons. Wotr, P. G.,C. H. HAUFLER, and E. SHEFFIELD. 1 f tic diploid the bracken fern (Pteridium coo alae st 236: 947-949. American Fern Journal 79(2): 65—77 (1989) Species Concepts in Pteridophytes: The Treatment and Definition of Agamosporous Species GERALD J. GASTONY Department of Biology, Indiana University, Bloomington, Indiana 47405 MICHAEL D. WINDHAM Department of Botany, University of Kansas, Lawrence, Kansas 66045 Pteridophyte species have generally been defined on the basis of relatively major morphological differences between sets of populations. At least implicitly, these morphological discontinuities are taken to indicate lack of gene flow and are thought to reflect the genetic discontinuities that make one species distinct from another. When functionally diploid, outcrossing fern species are circumscribed in this way, morphologically recognizable taxonomic species may also be good biological species. Unfortunately, in the case of agamosporous ferns (formerly designated ‘“‘apogamous”’ or “‘apomictic’’), the criteria used to define a biological species do not apply. Unlike the members of a biological species, agamosporous individuals cannot interbreed. However, they can cross with related sexually reproducing taxa to generate reproductively competent offspring, which biological species are not supposed to do. Thus the reproductive behavior of agamosporous ferns precludes application of a strict biological species concept, and the treatment and definition of agamosporous species is somewhat problematical. This issue merits consideration because agamosporous taxa constitute about 10% of all fern species for which the type of reproduction is known (Walker, 1984 p. 125). In this paper, we review the salient features of the typical life cycle of agamosporous pteridophytes. We then discuss the origins of several agamosporous fern taxa and indicate how we would treat them taxonomically. We conclude with a species concept accommodating both sexual and agamosporous taxa. MATERIALS AND METHODS Electrophoretic samples of Notholaena grayi Davenp. and the Pellaea atropurpurea (L.) Link complex were obtained from living plants maintained in greenhouses at the University of Kansas. Enzymes were extracted by crushing a small section (ca. 50 mm?) of immature leaf tissue in ten drops of the phosphate grinding buffer-PVP solution of Soltis et al. (1983). The grindate was absorbed into paper wicks which were inserted into 12.5% starch gels for electrophoresis. Phosphoglucomutase (PGM) was resolved on gel/electrode buffer system 6 of Soltis et al. (1983). Leucine aminopeptidase (LAP), hexokinase (HK), and triosphosphate isomerase (TPI) were resolved on the modification of buffer system 8 discussed by Haufler (1985). Malate dehydrogenase (MDH) was resolved using a modification of gel/electrode buffer system 11 (Soltis et al., 1983) in which the concentration of histidine-HCl was doubled. Shikimate 66 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) dehydrogenase (SkDH) and isocitrate dehydrogenase (IDH) were resolved using the 0.04M citrate buffer of Clayton & Tretiak (1972) titrated to a pH of 7.5 with N-3(3-aminopropyl])-morpholine. TPI, HK, SkDH, and IDH were assayed using the agarose staining schedules of Soltis et al. (1983). PGM, LAP, and MDH were assayed using recipes provided by Werth (1985). Stained gels were photographed using a red filter and Kodak Technical Pan 2415 high contrast film. RESULTS AND DISCUSSION Background.—To understand why agamosporous taxa cannot interbreed and yet can cross with related sexually reproducing taxa to produce new, true- breeding agamosporous lineages, one must be conversant with details of the agamosporous life cycle. These details have been well documented (Manton, 1950) and recently reviewed (Lovis, 1977; Walker, 1979, 1984). The life cycle of obligately agamosporous ferns involves alternation of two quite separate phenomena—the avoidance of meiotic reduction in sporogenesis (producing diplospores and gametophytes at the same ploidy level as the sporophyte that begets them) and the spontaneous development of a new sporophyte from the gametophyte without fertilization (apogamy or apomixis). We follow Léve & Love (1975) in adopting the term ‘“‘agamospory” for this overall reproductive process involving both diplospory and apogamy. At least two different major variations in the sporogenetic part of the agamosporous life cycle are known (Lovis, 1977; Walker, 1979). Because the implications for speciation and the recognition of species in agamosporous ferns are the same no matter which sporogenetic system operates, we will concentrate on the more commonly encountered Dépp-Manton scheme since the examples from our work follow this pattern. Sporogenesis in this system (Fig. 1) begins with the archesporial cell undergoing three successive mitotic cell divisions so that the sporangium contains eight cells of sporophytic ploidy. Thereafter two alternative courses are followed. In one type of sporangium (Fig. 1, upper sequence) the eight cells undergo a fourth mitotic division, yielding sixteen spore mother cells with the sporophytic chromosome complement. During meiosis, the chromosomes in these cells fail to pair regularly, instead forming univalents, bivalents, and multivalents. As a result, the chromosomes are not distributed evenly, and the resultant spores are ch lly unbal 1 and abortive. In the other type of sporangium (Fig. 1, lower sequence), the fourth mitotic division of the eight cells starts normally. The chromosomes gather on the equator and divide, but there is no nuclear or cell division. Instead, a restitution nucleus is formed having double the original chromosome number. Thus at the end of the fourth division (Fig. 1, heavy arrow) there are eight spore mother cells, each with twice the sporophytic ploidy. Because the fourth division was endomitotic, each chromosome now has an identical sister chromosome with which to pair. Pairing is therefore regular as the eight spore mother cells undergo meiosis, and sporogenesis yields 32 viable spores with the same chromosome number as the sporophyte. GASTONY & WINDHAM: AGAMOSPOROUS SPECIES 67 MITOTIC DIVISIONS MEIOSIS 3n n n C)—C)— © Abortive Spores 3n 3n 3n 3n oA. 16 32 64 Cs) foo \ on an C)—C)— © Functional spores 8 16 32 Fic. 1. Diagrammatic representation of sporogenesis in the Dopp-Manton agamosporous system. Circles represent cellular stages. Arrows represent the divisions of mitosis and meiosis. 3n denotes the sporophytic chromosome complement at whatever ploidy. Numbers below circles indicate the number of cells at that stage in the process. Heavy arrow represents the irregular mitotic division (endomitosis) leading to a restitution nucleus. See text for further explanation. A few agamosporous taxa such as Notholaena grayi and N. aliena show variations in the Dépp-Manton scheme that result in the formation of 16 spores per sporangium (Windham, unpubl. data), and some sexually reproducing ferns commonly produce 32-spored sporangia (Vida et al., 1970; Hickok & Klekowski, 1974; Smith, 1974). Therefore, spore counts may be used to generate hypotheses concerning the life cycle of a given taxon, but the existence of agamospory must be verified chromosomally or by growing gametophytes and observing sporophyte production in the absence of fertilization. When restitution nuclei of momentarily doubled ploidy undergo meiosis in Dopp-Manton sporogenesis (Fig. 1), the pairing of sister chromosomes precludes genetically significant recombination and segregation. Barring mutation, the genotype of each spore is identical to that of the parental sporophyte. The unreduced spores are disseminated and develop into gametophytes that look normal except that (1) functional archegonia are suppressed and (2) a new sporophyte develops spontaneously from gametophytic tissue without fertilization. The absence of syngamy means that no genetic variation is introduced through the union of genetically dissimilar gametes. These factors make it reasonable to assume that once an agamosporous lineage is established, it is essentially clonal, genetically invariant except for non-deleterious mutations. Although lack of functional archegonia and syngamy precludes the introduction of variant genetic material into an agamosporous lineage, Dopp- Manton agamosporous taxa can produce antheridia with functional (non- reduced) sperm and can thereby act as male parents in crosses with archegoniate gametophytes of sexually reproducing taxa. Hybrids resulting from such crosses inherit the complete agamosporous mechanism and are therefore able to 68 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) reproduce faithfully their new genotype (Walker, 1966). Such hybrid lineages always feature an increase in ploidy over that of either parent, because the unreduced ploidy of the paternal gametophyte is added to that of the reduced maternal gametophyte. The reproductively competent, agamosporous offspring of such hybridizations bridge the morphological discontinuities between otherwise discrete evolutionary lines and thereby complicate taxonomic identification and species circumscription. Application of species concepts to agamosporous taxa and their taxonomic treatment depend in part on how agamosporous taxa arise and how they are therefore related to their sexually reproducing progenitors. In the past, writers have generally regarded agamosporous fern taxa as being of hybrid origin (Lovis, 1977, p. 389; Walker, 1984, p. 127). This conclusion is based partly on meiotic chromosome behavior in the 16-spore-mother-celled sporangia of agamosporous taxa (Manton, 1950) and partly on the reasoning that “there is a very high proportion of triploid cytotypes which can only have arisen by hybridization” (Walker, 1979, p. 117). The assertion that triploid cytotypes can only have arisen by hybridization overlooks the possibility that triploid or other polyploid cytotypes might be autopolyploids that have arisen through intraspecific fertilization in which one or both gametophytes are diploid because they are derived from unreduced spores. To determine whether polyploid cytotypes do arise in nature via gametophytes from unreduced spores, Gastony (1986) tested Morzenti’s (1967) complex hypothesis that tetraploid Asplenium plenum results from a cross between an unreduced triploid gametophyte of A. curtissii and a haploid gametophyte of A. abscissum. The mechanism of unreduced spores was tested in an allopolyploid system rather than an autopolyploid one because variant species-specific electrophoretic markers could be identified and traced in an allopolyploid phylogeny whereas such markers would necessarily not be available in an autopolyploid phylogeny. Gastony’s electrophoretic data matched expectations under Morzenti’s hypothesis and rejected the competing hypothesis, confirming that unreduced spores in nature do produce gametophytes that generate polyploid sporophytic cytotypes. Unreduced gametophytes have also been implicated in the origin of scattered triploid sporophytes of Cystopteris protrusa (Haufler et al., 1985), Woodsia mexicana and Bommeria hispida (Windham & Haufler, 1985). Thus, it is clear that unreduced spores do yield sexually functional unreduced gametophytes in nature that are capable of crossing with haploid or other diploid gametophytes to produce triploid or tetraploid sporophytes. In the absence of other evidence of interspecific hybridization, the triploid and tetraploid cytotypes of agamosporous taxa need not be explained by interspecific hybridization. Instead, they may be autopolyploids (intraspecific polyploids) derived through the mechanism of naturally occurring unreduced spores. If at least some agamosporous fern taxa are autopolyploids, why do the endomitotic spore mother cells of their 32-spored sporangia show normal bivalent formation rather than multivalents? In an agamosporous autotriploid, for example, these cells would contain six homologous chromosome sets and GASTONY & WINDHAM: AGAMOSPOROUS SPECIES 69 multivalents should be observed. This situation could be explained if chromosome associations during meiosis were dependent on both structural homology and the action of certain regulatory genes. Genetic control of chromosome pairing has been reported for several species of Asplenium (Braithwaite, 1964; Bouharmont, 1972a, 1972b), and preliminary evidence suggests that it may be widespread among agamosporous taxa exhibiting Dépp- Manton sporogenesis. Rigby (1973) identified trivalent chromosome associations in sporangia of Pellaea atropurpurea that failed to undergo a premeiotic endomitosis (Fig. 1, upper sequence), suggesting that the three sets of chromosomes in this agamosporous triploid are at least partially homologous. However, endomitiotic sporangia (Fig. 1, lower sequence) from the same plant yielded only bivalents, providing no indication that all six chromosome sets were capable of associating to form multivalents. The same situation has been observed in several other agamosporous triploids (Windham, unpubl. data), including Asplenium monanthes, Cheilanthes bonariensis, Cheiloplecton rigidum, Notholaena aschenborniana, and Argyrochosma limitanea. Although the genetic mechanism is poorly understood, it appears that multivalent formation in these ferns is suppressed whenever an even number of homologous genomes is present in the cell. If such control of pairing is common among agamosporous taxa (as preliminary evidence suggests), the absence of multivalents during meiosis cannot be used as evidence against autopolyploid origins of agamosporous ferns. Case studies.—Mode of origin, degree of genetic continuity with sexual progenitors, reproductive interactions, and morphological distinctions must all be taken into consideration when characterizing agamosporous taxa and determining their taxonomic treatment. The following examples from our work on the relationships of agamosporous taxa in Pellaea, Notholaena, and Cheilanthes illustrate the value of modern biosystematic data when attempting to generate meaningful species concepts for this problematical group of pteridophytes. Pellaea andromedifolia, endemic to California and Baja California Norte, includes widespread sexually reproducing diploid populations and ympatrically interspersed, ag p ly reproducing triploid and tetraploid populations (Tryon, 1957, 1968; Gastony & Gottlieb, 1985). Sporophytes from sexual and agamosporous populations of all ploidy levels look alike. The only way to distinguish them morphologically is by counting spores per sporangium, with 64 in the sexuals and 32 in agamosporous individuals. Thus there is no morphological evidence of interspecific hybridization in the origins of the polyploids. Details of electrophoretically detected genetic variation in natural populations of P. andromedifolia of both reproductive types were provided by Gastony & Gottlieb (1985). Comparative electrophoresis showed that agamosporous triploid sporophytes and their respective gametophyte progenies are genetically identical. This confirmed the lack of recombination and segregation expected from chromosome pairing behavior in Dépp-Manton sporogenesis. Furthermore and most importantly, ll alleles coding allozy in 70 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) agamosporous populations were entirely a subset of those in the sexual populations. Thus there is no electrophoretic evidence that hybridization with another species was involved in their origin. As in the Asplenium plenum complex, reproductive interactions involving gametophytes from unreduced spores of P. andromedifolia could account for both their triploidy and their possession of only P. andromedifolia electrophoretic bands. A similar situation was observed during investigations of Notholaena grayi (Windham, unpubl. data), a species found in the southwestern U.S. and adjacent Mexico. Tryon (1956) indicated that N. grayi produced 32 spores per sporangium, but a survey of herbarium specimens identified some populations having only 16 spores per sporangium. Subsequent cytogenetic work revealed that 32-spored plants were sexual diploids whereas those with 16 spores per sporangium were agamosporous triploids. As in Pellaea andromedifolia, the two cytotypes are morphologically very similar, suggesting that interspecific hybridization was not a factor in the origin of the agamosporous taxon. Preliminary electrophoretic data for 19 enzyme loci support this conclusion. Figures 2—7 illustrate zymogram patterns of Notholaena grayi for the enzymes PGM, LAP, TPI, SkDH, IDH, and MDH. In each of these figures, the two lanes at the far left and the two on the far right represent sexual diploids, and the intervening lanes show all of the variation so far encountered among agamosporous triploids. For the loci coding PGM-2, LAP, TPI, SkDH, IDH, MDH-1, and MDH-2, all alleles found in the agamosporous individuals were also detected gth | diploids. The only orphan allele (found in the triploid but not in the diploid) occurred at the locus coding the cathodal MDH isozyme (MDH-3) in two plants from a single population (Fig. 7). This allele may be found upon further sampling of the diploid, or it may represent a recent mutation in the triploid that has not yet spread beyond the population of origin. In either case, morphological and electrophoretic data suggest that the agamosporous triploid form of Notholaena grayi arose through autopolyploidy. Further insight into the relationships between agamosporous taxa and their sexual progenitors is provided by the Pellaea glabella complex. At the time of the last taxonomic revision of this group (Tryon, 1957; Tryon & Britton, 1958), it was said to consist of three varieties: (1) sexual diploid var. occidentalis in South Dakota, Wyoming, Montana, and Alberta, (2) agamosporous tetraploid var. simplex in Alberta, British Columbia, Washington, Utah, Colorado, Arizona, and New Mexico, and (3) ag p tetraploid var. glabella widely distributed in eastern North America. Tryon (1957) suggested either allopolyploid or autopolyploid origins for the agamosporous varieties. The allopolyploid hypothesis proposed that sperm of agamosporous triploid P. atropurpurea fertilized a haploid egg of centrally distributed sexual P. glabella var. occidentalis, yielding agamosporous tetraploid var. glabella to the east and agamosporous tetraploid var. simplex to the west. Tryon’s autopolyploid hypothesis derived both agamosporous varieties directly from var. occidentalis, the only sexual diploid known to her. Several years later, Wagner et al. (1965) discovered a sexual diploid race of Pellaea glabella var. glabella in Missouri that was indistinguishable from GASTONY & WINDHAM: AGAMOSPOROUS SPECIES 7 PGM-2 LAP 1 TPt 2 SkDH IDH 1 2 MDH 3 Fics. 2-7. Zymograms of Notholaena grayi. Numbers in the margin identify different isozymes in enzymes coded by multiple loci. In each figure, the two lanes on the far left and the two on the far } + 1 dinlnids. Int ; ] t + ipl id right ae - o f oO f 7Z AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) agamosporous tetraploid var. glabella except in its chromosome number and number of spores per sporangium. They suggested that this new sexual race made an allopolyploid origin for agamosporous var. glabella unlikely, proposing instead that the agamosporous tetraploid was an autotetraploid derivative of the sexual diploid of the same variety. Comparative enzyme electrophoretic data for P. atropurpurea and the sexual and agamosporous taxa in the P. glabella complex were presented by Gastony (1988). All sampled populations of agamosporous var. simplex were invariant for the analysed enzyme patterns. Variation in allozyme patterns was found among the populations of agamosporous var. glabella, and the chromosome numbers of all such populations confirmed that they represent truly tetraploid agamosporous variety glabella and not variant backcrosses to sexual plants. These electrophoretically variant tetraploid populations may represent primary apomicts that have accumulated mutations since their origin or populations that have arisen through independent origins from sexual individuals with different genetic constitutions. Electrophoretic data further showed that P. atropurpurea was not involved in the origin of either agamosporous tetraploid. Based on population samples of the four glabella taxa from throughout their ranges, Gastony (1988) found that, like agamosporous varieties simplex and glabella, the two sexual taxa have several distinctive allozyme patterns for the enzymes considered. He concluded that (1) the enzyme banding patterns of agamosporous tetraploid var. simplex are contributed by sexual var. occidentalis and that (2) the somewhat variable banding patterns of agamosporous tetraploid var. glabella are a subset of those in sexual var. glabella. As in the case of P. andromedifolia and Notholaena grayi, there are no distinctive enzyme bands that have been contributed by another species and thus no electrophoretic evidence to support a hypothesis that these agamosporous tetraploids arose through interspecific hybridization. In fact the electrophoretic data are consistent with the hypothesis that they are autotetraploids derived from the respective sexual diploids presumably through the mechanism of unreduced spores. Although Pellaea atropurpurea was not involved in the origin of either agamosporous tetraploid variety of P. glabella, ongoing studies indicate that much of the morphological variation observed in P. atropurpurea results from hybridization between these triploid plants and sympatric sexual taxa. Electrophoretic and ch | data reveal that the so-called ‘simple form” of P. atropurpurea (discussed and illustrated by Tryon, 1972) arose through hybridization between that species and the rare Mexican diploid, P. notabilis. The resultant agamosporous tetraploid is fully fertile and has spread throughout much of Mexico (Windham, unpubl. data). It is morphologically intermediate between the putative parents and shows additivity at 10 enzyme loci, including SKDH (Fig. 8) and HK (Fig. 9). Pellaea atropurpurec has also hybridized with P. truncata in the southwestern U.S., producing an agamosporous tetraploid with more dissected leaves and sparsely pubescent rachises. At one locality in Oklahoma, P. atropurpurea has even hybridized with the sexual tetraploid P. wrightiana to form several pentaploid plants of intermediate morphology. GASTONY & WINDHAM: AGAMOSPOROUS SPECIES 73 nen nF nonh wae Fics. 8—9. Zymogr | ting tl igin of the ‘simple form” of Pellaea atropurpurea. A=P. atropurpurea; N =P. notabilis; H.= hybrid. 8. SkDH. 9. HK. L ( EES. Similar evidence of hybridization between P. atropurpurea and P. glabella var. occidentalis has been observed in South Dakota, Wyoming, and Alberta (Gastony, 1988). Pellaea atropurpurea is by no means unique in its tendency to hybridize with sympatric congeners. Hybridization between agamosporous species and sexual relatives has also been reported in Pteris (Walker, 1962), Phegopteris (Mulligan et al., 1972), and Asplenium (Morzenti, 1966). An additional example of hybridization revolves around the agamosporous tetraploid “form” of Cheilanthes wootonii provisionally called C. yavapensis by Reeves (1979). Cheilanthes wootonii is an agamosporous triploid (Windham, 1983), and one would assume that it was involved in the origin of the tetraploid “form” of that species. However, electrophoretic data (Reeves & Windham, in prep.) suggest that C. “yavapensis” is a hybrid between C. lindheimeri (an agamosporous triploid quite distinct from C. wootonii ) and C. covillei (a sexual diploid). This illustrates well the taxonomic confusion that can result when electrophoretic data are lacking for reproductively competent, agamosporous- sexual hybrids. Application to species concepts.—Several conclusions relevant to the treatment and definition of agamosporous species emerge from the foregoing case studies. (1) Once an agamosporous taxon is initiated, its genotype is perpetuated and not disrupted by meiotic segregation, recombination, or syngamy. (2) Individuals or populations that are electrophoretically detectable genetic variants of the rest of an agamosporous taxon at its base ploidy may reflect mutations that have accumulated in these populations since a single origin event or they may indicate multiple origins of the taxon from genetically variant sexual progenitors. (3) At least some agamosporous taxa at the primary level (such as those in P. andromedifolia, Notholaena grayi and the P. glabella complex) appear to be of autopolyploid origin. Taxonomic treatment of agamosporous lineages must recognize these genetic realities if taxonomy is to reflect phylogeny. How we would treat agamosporous lineages taxonomically is exemplified by the following discussion of the taxa used in our case studies. Some autopolyploid primary agamosporous lineages are genetically indistinguishable from their sexual progenitors at the enzyme electrophoretic level (except for banding pattern fixation caused by non-segregation and gene dosage effects associated with polyploidy) and are morphologically indistinguishable (except for number of spores per sporangium). Genetic 74 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) diversity in these agamosporous lineages is a subset of the pool of genetic diversity of the progenitor sexual taxon. In this case, the agamosporous taxon is not genetically discontinuous from the sexual taxon in a qualitative sense but merely quantitatively so by virtue of its extra gene set or sets. The situation in Pellaea andromedifolia is an example. Here genetic and morphological continuity argue that sexual and agamosporous lineages represent a single species with different reproductive behaviors that may be recognized in the sporophytes only by counting spores, counting chromosomes, or observing electrophoretic banding patterns. The case of the sexual and agamosporous lineages of P. glabella var. glabella is analogous to that in P. andromedifolia. Here also electrophoretically detectable genotypes of the agamosporous tetraploid comprise a subset of the allelic variation in the sexual diploid, and except for counting spores or chromosomes the two lineages cannot be distinguished reliably (Gastony, 1988). As in P. andromedifolia, the distinction between these taxa seems best recognized at the level of varieties. Comparable taxonomic treatment may be appropriate for the autotriploid agamosporous lineage of Notholaena grayi. Other autopolyploid agamosporous lineages may show similar qualitative genetic continuity with their sexual progenitors (being fixed for a subset of the sexuals’ alleles, although with extra doses of them) but show slight divergence from the sexuals in other respects. In the case of P. glabella var. simplex versus var. occidentalis, there is sufficient morphological discontinuity (even if partially attributable to a gigas effect in the tetraploid) that the sexual and agamosporous lineages have long been recognized as distinct varieties or even species (Brunton, 1979; Lellinger, 1985). In this case, qualitative genetic continuity again argues against distinction at the species level, but a slight morphological discontinuity (with an underlying, although undetermined, genetic basis) usually permits distinguishing the two taxa without counting spores or chromosomes. In addition, these lineages are almost completely allopatric. The fact that var. simplex survives fairly well under greenhouse cultural conditions in Indiana whereas var. occidentalis suffers great mortality under identical cultural conditions suggests that these taxa also differ physiologically. This degree of divergence of the sexual and agamosporous taxa is greater than that in P. andromedifolia and has accordingly been recognized at slightly higher rank, viz. at the level of subspecies (Gastony, 1988). In the case of agamosporous triploid P. atropurpurea, distinctive enzyme electrophoretic patterns indicate genetic discontinuity with other Pellaea taxa, and this is paralleled by its morphological discontinuity. We do not yet have enough data to determine whether it is of autopolyploid or allopolyploid origin. Although its agamosporous lifestyle prevents interbreeding among plants of its own lineage, electrophoresis does reveal genetic variation throughout its range, even at the triploid level. It therefore features the genetic characteristics of a sexual “biological’’ species (coherent genetic variation that is discontinuous with the coherent genetic variation of other species) and correlated from other species. It can hybridize with sexual taxa and thereby produce reproductively functional hybrids because of momentary GASTONY & WINDHAM: AGAMOSPOROUS SPECIES 75 chromosome doubling just before meiosis in Dopp-Manton sporogenesis. In this it hardly differs from “good,” sexually reproducing, “biological” plant species that are able to hybridize and whose hybrid offspring become reproductively functional if their chromosome number doubles. Agamosporous taxa such as P. atropurpurea are appropriately recognized at the rank of species. gamosporous lineages can and do act as male parents in crosses with sexually reproducing relatives. The offspring of such crosses are agamosporous at a ploidy level higher than that of their agamosporous parent and are capable of reproducing and perpetuating their new lineage through agamospory. Tetraploid Cheilanthes “yavapensis”’ with a substantial geographic range from southern New Mexico to northern Arizona provides an example of such an allopolyploid origin. Its geographic range and reproductive capability indicate a degree of evolutionary performance typical of successful species, and we consider it appropriate to recognize taxa such as this at the rank of species. The respective hybrids of Pellaea atropurpurea with P. notabilis and with P. truncata also belong to this category. If the derivative of an agamosporous-sexual cross consists only of one or a few individuals resulting from isolated, independent hybridization events, it may be treated taxonomically simply as an occasional hybrid plant, a potential lineage that has not yet demonstrated evolutionary performance warranting acceptance as a species. Individuals of this kind can be accommodated under the hybrid formula name uniting the epithets of their parents with a “x.” This, for example, may be the preferred treatment for the agamosporous derivative Pellaea atropurpurea X wrightiana. We regard a species as a coherent evolutionary lineage whose allele frequencies change through time as its genome varies in response to selection and other perturbations. This is a genetic species concept in which the species is genetically equivalent to that in the widely accepted biological species concept. In the biological species concept, two hanisms maintain the genetic integrity of the species: (1) panmixis or potential panmixis maintains the genetic coherence of the species, and (2) the inability to produce fertile offspring in reproductive interactions with other species under natural conditions maintains genetic discontinuity with other species. For sexually reproducing taxa, the occasional production of hybrids does no violence to this concept when the hybrids are sterile. When chromosome doubling renders these hybrids fertile and they perpetuate their lineage, often with range expansion as in some members of the Appalachian Asplenium complex (Wagner, 1954), we have no difficulty in accepting them as species of hybrid origin, while still regarding the parental taxa as species. In our genetic species concept, agamosporous species are genetically equivalent to sexual species in the biological species concept, but the mechanisms by which they maintain their genetic integrity are different. The absence of both syngamy and genetically significant recombination replace panmixis in maintaining the genetic coherence of the species, and the inability to produce fertile hybrid offspring at the same ploidy level in paternal interactions with sexual species maintains genetic discontinuity with other 76 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) species. Thus the genetic species concept is equally applicable to both sexual and agamosporous taxa and its genetic consequences are comparable to those of the biological species concept. This genetic approach to species definition permits us to formulate treatments of agamosporous taxa of various derivations. These treatments can be both genetically meaningful and taxonomically practical. Species of Pellaea, Notholaena, and Cheilanthes provide examples in which agamosporous taxa are appropriately recognized as species, subspecies, and varieties when data from morphology, cytogenetics, and enzyme electrophoresis are given combined consideration. ACKNOWLEDGMENTS —— reported here v was supported by NSF Grants BSR-6206056 and BSR-8516666 to G.J.G. J. Scelonge and Valerie R. Savage. Research by M.D.W. was supported by =a eign of Kansas General Research aha een ans toC. H. Haufler. M. D. W. also thanks Geo 4 te a a general. LITERATURE CITED BOUHARMONT, J. 1972a. Meiosi ly produced diploid plants of Asplenium manner Gee kt 3: 253- 258. 972b. Meiosis in geoeat a produced diploid plants of Asplenium septentrionale. Aes Fern Gaz. 10:237 BRAITHEWAITE, A. F. 1964. pata ae of apogamy in ferns. New Phytol. 63:293—305 BRUNTON, DI F. 1979. Taxonomy, distribution, and ecology of the cliff-brake Seat (Pellaea: Polypodiaceae) in Alberta. Canad. Field-Naturalist 9:288—295. CLAYTON, a3 W. and 1), . hataag 1972. lee rebeoriens ey for pH control in starch gel Res. B GastTony, G. J. 1986. Midlhoes evidence for the origin aye rate species by unreduced spores. er. J. Bot. 73:1563-1569. - 1988. The Pellaea glabella complex: electrophoretic evidence for the derivations of the er. Fern J. 78:44—67. and L.D. GOTTLIEB. 1985. Genetic variation in the homosporous fern Pellaea andromedifolia. Amer. J. Bot. 72:257—267. HauFLer, C. H. 1985. Enzyme variability and modes of evolution i in Bommeria (Pteridaceae). Syst. Bot. 10: pei »M. DW WINDBAM, D. M. BRITTON, and S. J. RoBINSON. — — and its evolutionary anad. J. Bot. 63:1855-186 Hickok, L. G. and E. j. KLEKowskI. 1974. Inchoate speciation in gars opteris: an analysis of the synthesized hybrid C. richardii x C. pteridoides. Evolution 28: 439— LELLINGER, D. B. 1985. A field manual of the ferns & fern allies of the piers States & Canada. ashington: Smithsonian Institution Press. Love, A. and D. Love. 1975. Plant Chromosomes. Vaduz: J. Cramer Lovis, J. D. 1977. Evolutionary patterns and processes in fence. Pp. 229-415 in Advances in oe research, Vol. 4, eds. R. D. Preston and H. W. Woolhouse. New York: Academic MANTON, 1 ten Problems of cytology and evolution in the Pteridophyta. Cambridge: University Press. GASTONY & WINDHAM: AGAMOSPOROUS SPECIES 77 MorzeENnrtI, V. M. 1966. Morphological and cytological dat theastern United States species of the Asplenium het ili I ] Amer. Fern J. 56:167-177. . 1967. Asplenium plenum: a fern which suggest | method of f Amer. J. Bot. 54:1061—1068. MULLIGAN, G. A., L. CiInq-Mars, and W. J. Copy. 1972. Natural int li between sexual and a species of the beech fern genus Phegopteris Fée. Canad. J. Bot. 50:1295-13 REEVES, T. 1979. A eee of the fern genus Cheilanthes ae Physapteris (Adiantaceae). h.D. dissertation, Arizona State University, Tempe, RicBy, S.J. 1973. Chromosome pairing in obligately a ese Pellaea atropurpurea and Pellaea glabella var. at a Rhodora 75: 122-131, SMITH, A. R. 1974 g. Californi d Arizona. Madrono 22:376-378. Soitis, D. E., GH: sale base D. C. Darrow, =~ - GAONY: 1983. Starch gel electrophoresis of ferns: A compila g Amer. Fern 73 9 TRYON, A. F. 1957. A revision e the fern genus Pellaea section Pellaea. Ann. Missouri Bot. Gard. 44:125-19 . 1968. smn of sexual and apogamous races in the fern genus Pellaea. Rhodora 0:1-24 1972. Spores, chromosomes and relations to the fern Pellaea atropurpurea. Rhodora 74:220— and D. NM BRITTON. 1958. Cytotaxonomic studies on the fern genus Pellaea. Evolution 12:137-145 TRYON, R. M. 1956. oe of the American species of Notholaena. Contr. Gray Herb. 179:1—106. Vina, G., C.N. Pace, T.G. WALKER, and T. REICHSTEIN. 1970. Cytologie der Farn-Gattung Cistedithen in oe und auf den Canarischen Inseln. Bauhinia 4:223—253 Wacner, W.H., JR. 1954. Reticulate evolution in the Appalachian Asplasiisas. Evolution 03-118. , D. R. FARRAR, and K. ns CHEN. 1965. Anew lf f Pellaea glabella from Missouri Amer. Fern J. 55:171— WALKER, - . nee Bape and evolution in the f nus Pteris L. Evolution 16:27—43. ——. 6. Apomixis and eae reproduction in ferns. Pp. 152-161 in Reproductive b lar plants, ed. J. G. Hawkes. Bot. Soc. Brit. Isles Conf. Rep. 9. 1979. “The Evcueiea ‘of hese. Pp. 87-132 in The experimental biology of ferns, ed. A. F. Dyer. New York: Academic Press _ 1984. Chromosomes and evolution in pteridophytes. Pp. 103-141 in Chromosomes in evolution of eukaryotic groups, Vol. 2, eds. A. K. Sharma and A. Sharma. Boca Raton, Florida: CRC Press WERTH, C. . 1985. implenienting an isozyme laboratory at a field station. Virginia J. Science 36:53-—76. WINDHAM, a D. 1983. vas ferns of Elden Mountain, Arizona. Amer. Fern J. 73:8: and . 1985. Autopolyploid evolution among oe Se Amer. J. Bot. 72:919. American Fern Journal 79(2): 78—89 (1989) The Species Concept in Pteridophyta with Special Reference to Isoétes R. JAMES HICKEY Botany Department, Miami University, Oxford, Ohio 45056 W. Cari TAYLOR and NEIL T. LUEBKE Botany Section, Milwaukee Public Museum, Milwaukee, Wisconsin 53233 The definition, description, and identification of Isoétes species has always been difficult. This is primarily the result of an overall morphological simplicity and extreme phenotypic plasticity that combine to reduce the number of stable and therefore diagnostic characters. Such aspects of the biology of Isoétes have, historically, resulted in rather diverse specific treatments and circumscriptions. Recent studies in the genus have tended to emphasize previously overlooked characters and to apply both modern experimental techniques and the principles of population genetics in order to resolve many of these inconsistencies. Investigations utilizing cytological data, germination and hybridization studies, and protein electrophoresis have shown that over the range of a species its boundaries vary due to clinal variation, non-patterned geographical variation, interspecific hybridization, and the production of fertile allopolyploids. Such approaches have greatly increased our understanding of the dynamics of this genus. At the outset, it should be stated that it is not the objective of this paper to discuss the tl tical implications, or the limitations and benefits of the various species concepts that are currently in existence. Such discussions are numerous and can be found throughout the pages of many of the major systematic journals such as Systematic Botany, Systematic Zoology and Biology and Philosophy. Rather, this paper is designed to illustrate both the current and the historical application of the various species concepts in pteridophytes with particular reference to Isoétes. It is hoped that the reader will come away with an appreciation of the progressive nature involved in the elucidation of species and for the (practical) necessity of applying multiple species concepts during any given systematic inquiry (see also Zander, 1985). To illustrate such processes the following examples will be presented: the I. storkii aggregate of Costa Rica; the I. melanopoda complex of the southeastern United States; and the Isoétes species of northeastern North America. Asa prelude to these examples, a brief review of the species concept and a discussion of the role of herbarium material will be presented. Species concepts.—For the purposes of this paper, the species concepts of consequence are: the typological species concept (TSC), the morphological species concept (MSC) and the biological species concept (BSC). The TSC is associated with all poorly known groups. Typically it is associated with rare or at least rarely collected species and is almost always employed when only a single collection is known for the species; i.e., simply, the species is circumscribed upon the only materials available—the type. The MSC, the traditional concept of HICKEY, TAYLOR & LUEBKE: ISOETES 79 most herbarium workers, is broader in scope and relies upon morphological hiatuses between individuals and/or populations in defining species. This concept i the variation known from as many populations as possible, and generally ‘includes field observations. Under the framework of the MSC the herbarium taxonomist frequently employs extrapolations from better known taxa to justify the inclusion of somewhat disparate elements into a single species (e.g., Tryon, 1955, p. 3). The BSC is dependent not on morphological boundaries but on reproductive ones. The systematist employing the BSC defines and describes interbreeding populations as the basic taxonomic unit. The role of herbarium material—A major factor involved in the transition from a TSC to an MSC is the incorporation of data on variability. Such data are generally obtained by examination of herbarium specimens. For most European and North American species, such material is both sufficient and readily available. This is not true in the tropics, however, where many species are -_ a a ae = > ' - = < a | 2 af “< ES ig = o = = wi = = i) Ss ~ hee ” a w” z z = s o oO a a = = o a. aw wi nn = — ro) = = © ad S od oO bs ce < pom 4 “ge 5 a > he > z - - ° ” re = & “ a 24 ° os ¥ : = . 3 —] z 0- 18 1860 1870 1880 1890 1900 1910 ‘1920 ' 1930 1940 1950 1960 1970 19 -59 -69 79 -89 -99 ‘09 -I9 -29 -39 -49 59 -69 -79 -8 Fic. 1. Status of Neotropical Isoétes. Bar graph indicates cumulative numbers of herbarium —— available for study, by decade, as of 1985; darker separation at 1960 indicates median: of colle ice dak solid line aape a “best- fit” "approximation of trend. Solid stars represent the number o ickey p star in the 1980-decad ‘ts the alt ti timate of Fuchs-Eckert (1982). L970U } 80 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) represented by 3 or 4 specimens and a large number are still represented by only a single known collection. In the Neotropics this situation is gradually changing. Since 1960, the total number of Neotropical Isoétes collections has grown from 226 to over 560, a 150 + % increase in the last 37 years relative to the previous 120 (Fig. 1). An analysis of these collections (Fig. 1) suggests that we are approaching (in the Andes at least) an end to the collecting of new (morphological) species. These herbarium collections represent the major resource for the development of the morphological species concept in Isoétes. Before 1960, for example, the average number of collections for Neotropical species was four; today that average is 12. Because the actual number of collections per taxon is usually higher than this (many Brazilian species are still only known from the type collection), estimates on interspecific variability can be made for most of them. As a direct result, we are now able to make reasonable hypotheses relative to species delimitations for many Neotropical species. The formation of alternative hypotl garding the delimit of these species, such as can be seen when comparing the works of Fuchs-Eckert (1982) and Hickey (1985) (Fig. 1), represent the base line models for future biosystematic research. In the examples presented below, the accumulation of morphological data and its incorporation into a MSC were crucial steps in the elucidation of our current hypothesis about an appropriate working species concept. THE ISOETES STORKII AGGREGATE In 1931, T. C. Palmer described a plant from Volcan Pods in central Costa Rica which ultimately became known as I. storkii Palmer. The type collection provided sufficient data to describe adequately the features of the plant and to differentiate it from other currently known species. Since then, several additional populations (at Cerro de la Muerte, Cerro Chirrip6, and Cerro Fabrega in Panama) have been found. Material from Cerro Chirrip6 was segregated as I. tryoniana Gémez (1970) and the Cerro de la Muerte material was tentatively suggested as a new species (“‘I. mickelii”’) by Lellinger (based on herbarium identifications). Morphological analyses (Hickey, 1985; Gémez, 1970) of the three Costa Rican populations indicate that there are observable differences in megaspore size, megaspore ornamentation, velum coverage, leaf number, leaf length, and habit. The segregation of I. tryoniana and proposed segregation of “‘I. mickelii” were the direct result of the identification of such diff All three of these taxa are diploids and presumably arose through geographic isolation and subsequent divergence. A somewhat different hypothesis was proposed (Cox & Hickey, 1984; Hickey, 1985; Hickey, ms. submitted) as a result of an alternative interpretation of the morphological data (summarized in Fig. 2). Morphological variation in megaspore size (axis b), megaspore ornamentation (axis c), leaf number (axis d), and leaf length (axis e) were compared with differences in elevation (axis a). The inverse relationship between axis a and axes b—e suggests that the three populations represent a single species exhibiting clinal variation. Morphological variation among the populations was interpreted to be primarily HICKEY, TAYLOR & LUEBKE: ISOETES 81 83°W vP Fic. 2. Character polygonals for three Costa Rican populations of Isoétes storkii. Axis a = elevation (2500-3500 m); axis b = mean megaspore size (500—700 pm); axis c = megaspore ornamentation (slight, moderate, pronounced); axis d = mean leaf number (20— 100): axis e = mean leaf length (0-100 mm). VP = Volcan Pods; CM = Cerro de la Muerte; CC = Cerro Chirripo. the result of plastic responses to differences in elevation and the concomitant variations in temperature and insolation. These competing hypotheses were recently tested (Hickey, unpubl. data) using starch gel electrophoresis on four presumptive populations from Volcan Pods and Cerro de la Muerte. Forty-three individual sporophytes were examined for variation at eighteen loci representing eight enzyme systems. The four populations were fixed for similar allozymes at 14 of these loci. Allelic variation in the four variable loci was low and overall genetic similarity was high (Fig. 3). Unlike other Isoétes species studied (I. mexicana, I. cubana, I. pallida, and I. flaccida; Hickey unpubl.) virtually no intrapopulational variation was detected in I. storkii. The electrophoretic studies support the single species hypothesis of Hickey (1985) and suggest two models to account for the genetic characteristics of the populations studied. The first model hypothesizes that the Cerro de la Muerte and Volcén Pods populations were derived from a very few long distance dispersed spores; the paucity of genetic variation then is the result of an initial lack of genetic variability. A corollary to this model is the virtual lack of 82 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) GENETIC SIMILARITY 0.8 0.9 1.0 | | j vP DISTANCE (km) FROM CERRO-1 | | | o Ww Fic. 3. Estimates of genetic similarity plotted against geographic distance for four populations of Isoétes storkii. VP = Volcan Pods; C1, C2 & C3 represent three subpopulations from Cerro de la Muerte. interpopulational gene flow. This model is essentially one of island biogeography where the highest peaks of the Talamanca range represent discrete, ecological islands. Alternatively, the lack of genetic variability within and between populations may be stochastic and result from severe genetic bottlenecks. Vulcanism (Volcan Pods) and fires (Cerro de la Muerte) are likely probable causes for population fluctuations in these areas. Such bottlenecks could deplete genetic variation even in the face of limited introduction of additional propagules. These models, either alone or in concert, would account for the low genetic variation seen in the Cerro de la Muerte and Volcan Pods populations. Both scenarios suggest that the populations studied are peripheral, derivative populations and both predict the existence of a genetically more diverse source population at either Cerro Chirrip6o or Cerro Fabrega. The morphological variation seen in the Costa Rican populations of I. storkii is may reflect an initial stage in allopatric speciation (Tryon, 1986). THE ISOETES MELANOPODA COMPLEX In the southeastern United States there are a number of diploid Isoétes which are endemic to granitic outcrops. These species are known, on the basis of enzymatic (Hickey et al., 1989), morphological, and phenological (Table 1; Fig. 4) synapomorphies, to be closely related to the more widespread I. butleri HICKEY, TAYLOR & LUEBKE: ISOETES 83 Engelm. and I. melanopoda Gay & Durieu. All of these species are tropical in origin (Hickey, 1985; Hickey et al., 1989) and, while doubtfully holophyletic, are of monophyletic origin. With few exceptions, these taxa were initially characterized on the basis of one or at best a few collections and the transition from a TSC to a MSC was a gradual one. The morphological analyses of Boom (1979), Matthews and Murdy (1969), Reed (1965) and Rury (1978) have outlined the infraspecific variation in these plants. Not surprisingly, these studies have resulted in somewhat divergent opinions regarding the status of several taxa. This complex is now undergoing electrophoretic and biosystematic analyses and, while the data are not complete, several trends are apparent. Using triose phosphate isomerase (TPI) as an example, it will be noted that phylogenetically ‘ x a0, 3; 4,5, 6 ISOETES MELANOPODA COMPLEX Fic. 4. Geographic distribution and phylogenetic reconstruction of the Isoétes melanopoda complex. Character numbers refer to those listed in Table 1. B = I. butleri; Mp = I. melanopoda; P = I. piedmontana; Ms = I. melanospora; T = I. tegetiformans. 84 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) } ] TABLE 1. Polarized Mory gical Cl ter States Used for Phylogenetic Analysis of the I. melanopoda Alliance. Outgroups for polarization are those established by Hickey (1986) and the I. panamensis alliance. Polarity Character Pleisomorphic Apomorphic 1. Phenology fall spr 2. Velum coverage absent partial— complete 3. Fibrous bundles present absent 4. Leaf X-section trigonal +/-terete 5. Megaspore color white brown-black 6. Scales/Phyllopodia present absent 7. Microspore surface echinate papillate 8. Substrate pH acidic asic 9.G absent 10. Leaf arrangement spiral distichous 11. Corm shape globose elongate basal taxa such as I. butleri, I. melanopoda, and I. piedmontana (Pfeiffer) C. Reed are genetically more variable and possess a greater number of allelic variants over several loci (Table 2, Fig. 5). These species also show a greater number of accumulated electrophoretic autay phies. As one moves geographically and phylogenetically away from the basal taxa, the number of allelic variants and autapomorphies diminish. Distal (= derived) taxa are genetically depauperate and their allelic constitution represents a subset of that found in the more basal species. These data suggest either recent or ongoing speciation via allopatry. Isolation among the taxa, in part geographical and in part ecological (as in Selaginella spp.; Tryon, 1971), appears to be incomplete. While the lack of marker alleles in the distal taxa precludes a definitive testing of reproductive isolation, morphological analyses suggest that reproductive isolating barriers have not yet evolved. In fact, Matthews and Murdy (1969) suggested that the lack THE ISOETES OF NORTHEASTERN NORTH AMERICA Perhaps the most critical data dealing with the nature of the species in Isoétes came from work on the taxa of northeastern North America. For well over a century, these plants have attracted the attention of botanists working in this region. Initially, the taxonomy of these Isoétes was influenced by a TSC, resulting in over thirty published species, varieties, and forms. However, the last MSC account of this region accepts only eight species and no infraspecific taxa (Kott & Britton, 1983). More recently, the application of experimental techniques HICKEY, TAYLOR & LUEBKE: ISOETES 85 TABLE 2. Distribution of TPI Alleles within the Isoétes melanopoda Complex. Species Character Number Locus Allele B Mp P Ms 4 1; I A + oe I B + 3: I > + + + + 4, I D + 5. I E + - 6. II A + + + ie II B + - - 7 8. II G ei 9. II D + 10. II E + RE Ae Tf, II F + gas II G . i 13. II H + 2 = cmme } es 13 . 8 ll 2 a _15 ’ 4 , 12 10 ry 6 5 3 Fic. 5. Overlay of TPI allele distribution onto cladogram in Fig. 4. One of several eauanty parsimonious reconstructions; this assessment favoring allele loss over convergence. Data as in Table 2. 86 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) |. lacustris 2n=110 |. heterospora 1. X hickeyi (1. X harveyi) 2n=66 2n=77 |. tuckermanii 2n=44 1. echin. X tuck. |. X foveolata 2n=33 |. echinospora _ |. X eatonii 1. engelmannii 2 anaes 2n=22 Chromosome doubling |. riparia. 2n=44 Fic. 6. Hypothetical pedigree for some Isoétes in northeastern North America. has made it possible to confirm the presence of interspecific hybrids, to identify a polyploid series, and to recognize the reticulate nature of the relationships of these taxa. The controversy regarding the occurrence of interspecific hybrids in Isoétes began in 1896 when Dodge remarked that species of Isoétes “intergrade at times, Isoetes tuckermanii, for instance, appearing to hybridize with Isoetes echinospora.” Eaton (1900) took exception to Dodge’s claim stating that hybridity is extremely rare, if not altogether absent in the genus.” Jeffrey and Hicks (1925) and Jeffrey (1937) reported on a quillwort from Nova Scotia that produced megaspores that varied tremendously in size and shape, microspores that were mostly empty, and meiotic cells that showed lagging chromosomes HICKEY, TAYLOR & LUEBKE: ISOETES 87 during microsporogenesis. They attributed these phenomena to interspecific hybridization. Kott and Britton (1983) maintained that there was insufficient evidence to state that hybridization occurs naturally among the Isoétes of northeastern North America. They asserted that either ethological or sterility barriers exist among most taxa. The first experimental data pertinent to these conflicting views came from Boom (1980), who reported successful, artificial crosses between four species of Isoétes. The ease with which hybrids formed in culture suggested to Boom that hybridization followed by polyploidization was a feasible mode of evolution in the genus. Thus, Boom’s work provided a rational, testable hypothesis regarding allopolyploid speciation in Isoétes. 3 Evidence from spore morphology, chromosome number, enzyme electrophoresis, and in vitro germination and hybridization experiments was provided by Taylor et al. (1985) to support the hypothesis that sterile, interspecific hybrids occur in nature and that some of these diploids could have been involved in the evolution of fertile allopolyploid species. Specifically they proposed that I. echinospora Durieu, a circumboreal diploid and I. engelmannii A. Braun, an eastern North American diploid, hybridized to form I. x eatonii Dodge (pro sp.), a sterile, diploid plant occurring mainly within the overlapping range of the putative parents. In addition, they suggested that I. riparia A. Braun, a fertile tetraploid distributed along the eastern seaboard, is an allotetraploid derivative of I. x eatonii. Their data also indicate that I. x foveolata A. Eaton (pro sp.), a sterile triploid, represents a cross between I. engelmannii and I. tuckermanii A. Braun, a fertile tetraploid restricted to New England and eastern Canada. More recent work by Taylor and Luebke (1986, 1987, and unpubl. data) and D. Britton (pers. comm.) reveals that I. x harveyi A. Eaton (pro sp.) and I. x heterospora A. Eaton (pro sp.), both commonly encountered in northern New England and the Maritime Provinces, are heptaploid hybrids resulting from crosses between I. tuckermanii A. Braun and I. lacustris L., a circumboreal decaploid. In this same region, they also have identified a less common hybrid between I. echinospora and I. tuckermanii. Luebke and Taylor (1985, 1988) reported the discovery of interspecific hybrids between I. echinospora and I. lacustris. All of these putative hybrids produce aborted spores and possess additive electrophoretic profiles. The thorough documentation of interspecific hybrids in addition to the presence of polyploid species indicate that allopolyploid speciation has been an important factor in the evolution of Isoétes in northeastern North America (Fig. 6). CONCLUSIONS The formulation of a working species concept in any genus is an historical, stepwise process. In many animal groups and most definitely in Isoétes, the progression has been from a typological to a morphological to a biological species concept. The transitions between these steps are the result of differing objectives and methodologies, each of which contributes unique insights toward an understanding of the various species involved. 88 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) The TSC-MSC transition is highly dependent upon significant numbers of collections and an analysis of variation. As discussed above, this transition is still far from complete for tropical Isoétes but the collections of D. Breedlove, A. Cleef, J. Cuatrecasas, S. and J. Keeley, B. Ollgaard, and J]. Steyermark, for example, have made such a transition possible for at least some parts of the Neotropics. While the primary focus and importance of this transition is in the establishment of workable taxa, the accumulated data often form the hypotheses and needed background for the MSC-BSC transition. This is shown particularly well by the work in the I. storkii analyses and in the ongoing work with the I. melanopoda complex. The establishment of the biological species concept for Isoétes is a recent development. In part, this has been because its establishment is inseparable from an understanding of the nature and modes of speciation (Tryon, 1986). In Isoétes, we are just beginning to investigate and understand these mechanisms. However, the data accumulated to date are compelling. The works of Boom (1980) and Taylor et al. (1985) have clearly shown that pre-zygotic isolating mechanisms do not exist between aquatic species in eastern North America and that natural interspecific hybrids occur in areas of sympatry. The works of Hickey, Luebke, and Taylor indicate that allopatric divergence may well be the initial mode of speciation in the genus. Spatially isolated, divergent taxa produce sterile F, hybrids upon subsequent contact suggesting that post-zygotic isolating mechanisms form rapidly. A 56% incidence of polyploidy in the genus (Hickey, 1984, and unpubl. data) indicates that the results of Taylor and Luebke’s work on the species of northeastern North America may be typical of the genus as a whole. That is, polyploidization is a frequent associate of hybridization and that allopolyploidy may be a significant, secondary speciation mechanism in Isoétes. In conclusion, the data currently available suggest that species of Isoétes appear to have evolved in two ways. First, species have evolved gradually as the result of spatial isolation and genetic divergence; second, species have formed abruptly by interspecific hybridization followed by chromosome doubling. Available data also suggest that the genus is currently undergoing rapid and continuing speciation (both divergent and abrupt), a situation that often confuses taxonomic boundaries under the restraints of a morphological (or any other non-dynamic) species concept. All of the species of Isoétes that have been intensively investigated readily conform to a dynamic interpretation of the biological species concept. With additional study, our species concept, interpreted as a working hypothesis, will undoubtedly change. ACKNOWLEDGMENTS This research was supported in part by an NSF grant (BSR 86-0672) to R. James Hickey. HICKEY, TAYLOR & LUEBKE: ISOETES 89 LITERATURE CITED Boom, B. M. 1979. Systematic studies of the genus Isoetes in the southeastern United States. M.S. Thesis. The University of Tennessee, Knoxville. . 1980. Intersectional hybrids in Isoétes. Amer. Fern J. 70:1 Cox PA. ee R. : cna 1984. Convergent megaspore Gas anid Isoetes. Amer. Naturalist 124:4 DoncgE, R. ey i pads and Fern Allies of New ange ea aa ¥ Bs Clute & Co. EaTon, A. A. 1900. The 68g Isoetes in New Engla FucHS-ECKERT, H. P. . Zur heutigen Kenntnis von sis i ommen aaa Verbreitung der suda mriknischen cate, Arten. Proc. Ned. Akad. Wetensch. C85:205—260. GOMEZ, L D419 eee a la Pteridologia costarricense. I. Nuevas especies. Revista Biol. Trop. Hi : 7 Hickey, R. J. Tae 5. Reisionary studies of Neotropical Isoetes. Ph.D. Dissertation. The University of Connecticut, Stor . 1984. Chitinase numbers in Bonga ® Isoétes. Amer. Fern J. 74:9—13. . Ms. submitted. Isoétaceae. Flor Se I. GUTTMAN, and W. H. HAs wens 1989. Evidence for post-translational modification of riose phosphate isomerase foie in Isoétes. eek J. Bot. 76:215—221 JEFFREY, " C. 1937. ane tology of a I s. Bot Mag (To ky }51: a aan pets C.H onete TI duction divisi er. scikiealist 59:410—426. Kort, L. . eee D. M. BritToNn. 1983. Spore ais and taxonomy of Isoetes in northeastern North America. Canad. J. Bot. 61:3140—316 LUEBKE, N. T. and W. C. Taytor. 1985. an of sae American Isoétes hybrids. Amer. J. Bot. ve: f 24 MatTTHEWS, J. F. and W. H. Murpy. 1969. A study of Isoétes common to the granite outcrops of the southeastern Piedmont, United at Bot. Gaz. ii catbeae 130:53-61. PALMER, T. C. 1931. Tropical American Isoetes. Amer. Fern J. 21: REED, C. F. 1965. Isoetes in southeastern United eee oy 12: 369 400. Rury, P. M. 1978. i and unique, mat-f gr gia. Amer. Fern 68:99— TAYLOR, W. C. and in a: LUEBKE. 1986. cere ER: hybridization and the taxonomy of North merican Isoétes. es J. Bot. 73:7 and ———_—_. 1987. Spore ponharteee and evolution in Isoétaceae. XIV International Botanical Congress Mica p. 292. Berlin (West), Germany. and .Isoétes Xt t I echinospora and I. macrospora. Tia Fern J. 78 , ———., and M. B. SMITH. on Speciation and hybridisation in North American quillworts. Bic Roy. Soc. Edinburgh 86B:2 TRYON, R. — , Selaginelig rupestris and its allies. ante Missouri ney ae 42: i “99, 92.20 10 1986 ee biogeography of species, with special reference to ferns. Bot. Rev. (Lancaster) 52:11 ZANDER, : a es Ss essay on speci ts developed during revisi y studies. Bryologist 15-220 American Fern Journal 79(2): 90—93 (1989) Species Concepts in Pteridophytes: Summary and Synthesis CHRISTOPHER H. HAUFLER Department of Botany, University of Kansas, Lawrence, Kansas 66045 Systematic pteridologists may be at a crossroads in their perception and treatment of species. Each of the papers in this symposium series presents a historical perspective, the state of the art, and a view to the future in considering what impact current research will have on the evolution of our theories, the direction of our research, and, ultimately, the development of our discipline. The first two papers (on primary divergence [Yatskievych and Moran, 1989] and reticulate evolution [Barrington et al., 1989]) cover the more familiar considerations of pteridophyte species. Both of these papers review how we have approached species in the past and explore how ongoing studies are modifying both the theoretical and the practical perception of species. Both contributions showed that to date species recognition has not been affected significantly by modern approaches. This may change, however, because fieldwork and examination of living specimens are now becoming standard components of revisionary studies. Thus far, molecular data have been used primarily to test hypotheses about the origin and interrelationships of species and have proven remarkably useful in evaluating polyploid complexes. The third contribution on cryptic species (Paris et al., 1989) explores the possibility that in some cases, traditional approaches may not yield an accurate picture of what constitutes natural evolutionary units. This paper raises the additional specter that these are not isolated cases and cryptic species may be much more common than currently recognized. The fourth contribution on agamosporous taxa (Gastony & Windham, 1989) presents convincing evidence that asexual species should be accorded greater status and consideration as dynamic evolutionary elements. Although typically regarded as entities without a future, agamosporous taxa can be genetically variable because of multiple origins and can add substantially to the taxonomic confusion of many groups by hybridizing with sexual congeners to produce new agamosporous species. The final paper on Isoétes species (Hickey et al., 1989) illustrates well that through the application of biosystematic apy hes to the “fern allies,” all of the predicted influences of cryptic species, ecological variants, and hybridization are conspiring to complicate modern species concepts in the microphyllous vascular cryptogams. THE CHANGING VIEW OF SPECIES In the papers constituting this symposium, several recurrent themes relating to the development of modern perceptions of species are evident. 1) Fieldwork has played an important role in providing a firmer foundation for assessing the limits of intraspecific variability. The collection of live plants has HAUFLER: SUMMARY 91 been especially important in considering the primary divergence of species and the environmental component of morphological variability. Obviously, in modern revisions, fieldwork must continue to figure prominently. 2) Evidence from modern experimental approaches is having a great influence on changing our perception of the forces driving speciation, developing intraspecific variability, and maintaining isolation of species. Since 1950, the value of chromosomal information has been recognized, but more recently teaming chromosomal and isozymic data has proven to be a powerful approach for building new systematic hypotheses. Questions that cannot be answered by one technique are often resolvable by the other. It seems clear that revisions will need to incorporate evidence from at least one if not both of these techniques in developing a persuasive set of taxonomic guidelines. In addition, it is clear that chloroplast DNA analysis will be used with increasing frequency as an important data base for developing hypotheses concerning interspecific and intergeneric relationships. 3) Because it is evident that many of the morphological features that have figured prominently in defining species in the past are subject to parallel and convergent evolution and are heavily influenced by ecological conditions, we must continue looking for new characters and developing morphometric analyses of our accumulated data. The likely emergence of more and more cryptic species demands that we pay closer attention to less prominent characters in proposing species boundaries. 4) We must acknowledge the value of employing explicitly phylogenetic analyses in developing hypotheses of relationship. There are many programs and protocols (e.g., PAUP and PHYLIP) that are available for phylogenetic analysis of biochemical and morphological data sets. Revisions based exclusively on intuitive perceptions of evolutionary links between taxa can no longer be considered sufficient. CONCEPTS AND DEFINITIONS Three principal concepts emerge from the amalgamated contents of the papers in this symposium. Although the emphases given to these views of species are varied and the actual names applied may differ, I think that the fundamental ideas may be encapsulated by the following summary. The morphological species concept refers to groups whose boundaries are diagnosed by discontinuities in critical, qualitatively or quantitatively definable features of the available specimens. The biological species concept is applied to groups that do not necessarily differ morphologically but do have barriers to interbreeding. In nature, biological species may be difficult to recognize because they can be isolated solely by genetic differences, ecological tolerances, geological barriers, geographical distance, or by a combination of these features. The evolutionary species concept has been proposed to place a historical parameter on biological species and require definition of ancestral/descendent relationships. Even agamosporous species can be called good evolutionary species because they represent monophyletic clones of organisms. It seems, therefore, that the 92 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 2 (1989) different ideas about species depend on the amount and type of data available; as more data are available, more inclusive species concepts can be employed. The crux of the matter is that our delimitation of a species should always represent a testable hypothesis. Ideally, there should be one concept that is the most robust, and that is applicable to both sexually and asexually reproducing species. We should attempt to delimit species that are consistent with this concept. Of those discussed above, the most widely applicable is the evolutionary species concept. To properly propose an evolutionary species, however, we should know 1) its morphological characteristics and how these differ from taxa that resemble it most clearly, 2) its breeding behavior and/or how it remains isolated from other species, 3) its ecological and geographical range, and 4) how it is related genealogically to its congeners. Clearly, this represents more than we know about most pteridophyte species and it is unreasonable to demand that all of this information be in hand before new species are proposed. If we consider the evolutionary species concept as a goal of modern systematists, what should we call the steps or preliminary hypotheses along the way? Perhaps practical definitions can be employed as mileposts marking the route towards the evolutionary concept goal. These definitions can represent the best hypotheses that can be proposed based on the available data and the application of these definitional mileposts can recognize the progressive acquisition of new data. When morphological analyses of herbarium specimens are combined with geographical and ecological data (and at least this set of data should be considered basic in recognizing new taxa), the morphological definition is employed. Including data from natural or artificial crosses, chromosomes, isozymes, DNA, etc. leads to a more “biological” definition. By adding historical perspectives on the origins of taxa through paleobotanical and geological data as well as explicit arguments on character state evolution and ancestors, an evolutionary species can be proposed. I do not mean to imply that this is an endpoint—even the most seemingly robust evolutionary scenarios are still hypotheses open to further testing. CONCLUSIONS AND PREDICTIONS Several conclusions about pteridophyte species can be made from the information in this symposium. First, in pteridophytes, because there is not a good fossil record and extinction has erased much of the historical evidence for polarizing characters, it will continue to be difficult to propose evolutionary species at the diploid level. Second, Paris et al. (1989) have demonstrated that expanded data bases can modify our perception of species limits and can be used to develop strong biological species definitions even in the face of ambiguous morphological criteria. Certainly the demonstration that cryptic species may be prevalent in pteridophytes means that we must remain open minded regarding the sorts of applicable data. Ultrastructural and biochemical traits cannot be ignored simply because they are difficult to generate and observe directly. At the same time, we must evaluate critically what constitute significant characters in defining species. Third, in polyploid taxa (whether sexual or agamosporous), HAUFLER: SUMMARY 93 chromosomal and isozymic data often allow us to make precise statements about the origin of lineages and so in many cases we may be able to apply a rigorous evolutionary species concept. Finally, it may be predicted that at all ploidy levels, the expanded application of chloroplast DNA data will be valuable in proposing and testing phylogenetic hypotheses. ACKNOWLEDGMENTS m grateful to all participants for their contributions to. is syroposiutt. I thank in particular David Barrington, Jim Hickey, and Cathy Paris for ummary. LITERATURE CITED BARRINGTON, D. S., C. H. HAUFLER, and C. R. soe 1989. Hybridization, reticulation, and species concepts in the ferns. Amer. Fern J. 79:55—64 GasTony, G. J. and M. D. WINDHaAM. 1989. ‘Species concepts in pteridophytes: The treatment and definition of ag ies. Amer. Fern J. 79:65—77. Hickey, R.J., W. C. TAYLOR, ‘and N.T . LUEBKE. 1989. The species concept in Pteridophyta with special reference to Isoetes. Amer. Fern J. 79:78—89. PARIS.C. A..F: ie — and W. ae WAGNER, i 1989. Cryptic sagt oegae delimitation, and Fern J. 7 onomic p YATSKIEVYCH, . and R. C. Moran. 1989. ‘Primary divergence and species ak in ferns. Amer. Fern J. 79:36—45. INFORMATION FOR AUTHORS Authors are encouraged to submit manuscripts pertinent to pteridology for publication in the American Fern Journal. Manuscripts should be sent to the Edi- tor. 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 ac- ceptability of manuscripts nor the date of publication. Authors should adhere to the following guidelines; manuscripts not so pre- pared may be returned for revision prior to review. Submit manuscripts in tripli- cate (xerocopies acceptable), including review copies of illustrations. Do not send originals of illustrations until they are requested. Use standard 81/2 by 11 inch paper of good quality, not ‘erasable’ paper. Double space manuscri throughout, including title, authors’ names and addresses, text (including heads and Keys). literature cited, tables (separate from text), and figure captions (grouped a ). Arrange parts of man- uscript 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. Avoid footnotes and do not break words at ends of lines. Make table head- ings 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 typi- fication), use one paragraph per basionym (see Regnum Veg. 58:39—40. 1968). Abbreviate titles of serial publications according to Botanico-Periodicum- Huntianum (Lawrence, G. H. M. et al., 1968, Pittsburgh: Hunt Botanical Library). References cited only as part of nomenclatural matter are not included in litera- ture cited. For shorter notes and reviews, put all references parenthetically in text. Use Index herbariorum (Regnum Veg. 106:1—452. 1981) for designations of herbaria. cay 4 eR TIk cae Bn Am ee ps age +h, d to fit pa page width p page. Provide margins of at least 25 mm onall illustrati F ti tone illustrations, design ction without reduction or by uniform amount. In composite blocks, abut aes of adjacent photographs. Avoid com- bining continuous-tone and line-copy in single illustrations or blocks. Coordi- nate sequence and | numbering of ‘figures (and of — with order of citation in text Explain scale oti pti Includea scale and reference to latitude and longitude i in each map Proofs and reprint order forms are sent to authors a the printer. Authors should send corrected ta to the editor and reprint orders to the printer. Au- thors will be assessed made after type has been set. For other matters of form or style, consult recent issues of American Fern Jour- nal and The Chicago manual of style, 13th ed. (1982. Chicago: Univ. Chicago Press). racer departure from these may be justified. Authors with any 5 coon longer than 32 printed pages may be sent to the Editor of Pteridologia (Memoir Editor, see cover 2). American on Fern Number 3 July-September 1989 Journal QUARTERLY JOURNAL OF THE AMERICAN FERN SOCIETY Cryptogramma cascadensis, a New Parsley-Fern from Western North America Edward R. Alverson 95 Zygophlebia, a New Genus of Grammitidaceae L.EarlBishop 103 flare A - lash "Warren H. Wagner Jr., Joseph M. Beitel, and RobbinC.Moran 119 r pa ts p Pe 2. ae oe Pad AWN, c * ia; es . 2 LL. nr * s. oe thn Ctoata nt Veracruz, Mexico Ramén Riba 122 Shorter Note Cy ‘ pt ~ ‘nf . st. Bs ryl aoe : Allison W. Cusick 125 The American Fern Society Council for 1989 JUDITH E. SKOG, Biology Dept., George Mason University, sig ae 22030. President DAVID B. LELLINGER, Smithsonian Institution, Washington, DC sme Vice-President W. CARL TAYLOR, Milwaukee Public Museum, Milwaukee, WI aa Secretary JAMES D. CAPONETTI, Dept. of Botany, University of Tennessee, sandr TN 37916. Treasurer DAVID S. BARRINGTON, Dept. of Botany, University of Vermont, Burlington, VT 05405. Records Treasurer JAMES D. MONTGOMERY, Ecology III, R.D. 1, Berwick, PA 1860 Back Issues Curator R. SMITH, Dept. of Botany, University of California, ste eh 94720. Journal Editor DAVID B. LELLINGER, Smithsonian Institution, Washington om 205 Memoir Editor JOHN T. MICKEL, New York Botanical Garden, Bronx, NY 10458. ae iddlehead Forum Editor American Fern sci EDITOR Oe Oe Dept. of Botany, University of California, Berkeley, CA 94720 ASSOCIATE EDITORS GERALDJ.GASTONY .........2.. Dept. of Biology, Indiana University, Bloomington, IN 47401 wks EUR ga 1S CNSR ee Dept. of Botany, presente of Kansas Lawrence, KS 66045 DAVIUE. LELLINGER ......2..:... U.S. Nat’] Herbarium NHB-166, Smithsonian | Institution, ashington, DC 20560 TERRY R. WEBSTER ..... Biological Sciences Group, cies of Connecticut, Storrs, CT 06268 The “American Fern Journal” (ISSN 0002-8444) i to the gener study of ferns. It is owned by the American Fern Society, and seelrnaiye at the Matas Herberium: University of Vermont, Soe VT 05405-0086. Second-class postage paid at Burlington, VT and d additional entry poin' and orders for back i issues should be addressed to Dr. so D. ‘Bina. alle Ill, R.D. 1, Berwick, PA 18603. e — of address, dues, and applications for membership should be sent to the Records reasurer. +s TIL rte «L rag Seri $20.00 gross, $19. 50 net if paid airaige an agency cra fee $0.50); sent free to — ve — Society (annual dues, $15.00 + $4.00 mailing surcharge beyond __Back volumes 1910-1978 $5. 00 to $6.25 ‘each; single back numbers of 64 pages or less, $1.25; Op ig pag , $2.50 each, mae HEAP! Back volumes 1979 et seq. $8. 00 7 1 ts L eo An Fs PP VOLIUTNeS Dances corcn. C. 3 ep oe eet nate Scien JOURNAL, Dept. of Botany, University of tor 1 tributions fro M -members, including miscellaneous = notes, offers to = or ‘Purchase materials, eens hactiouttieal notes, and reviews of Spore Exchange ee sin Horder, 16813 Lemolo Shore Drive N.E., Poulsbo, WA 98370, is Director. Spores Ag tok. it ote ta na men ten ir ferns. Botanical | —— a +r American Fern Journal 79(3): 95—102 (1989) Cryptogramma cascadensis, a New Parsley-Fern from Western North America EDWARD R. ALVERSON Department of Botany and Plant Pathology, Oregon State Uni ity, Corvallis, Oregon 97331-2902 The genus Cryptogramma R. Br. is widely distributed in the temperate and boreal regions of the northern hemisphere and South America. All but one of the 8-10 taxa recognized belong to section Cryptogramma Prantl, which comprises plants commonly knownas “‘parsley-ferns.”’ Species belonging to thi tion, as exemplified by the Eurasian species C. crispa (L.) R. Br., are small, tufted ferns possessing dimorphic fronds, with erect fertile leaves bearing contracted, linear segments, and shorter, spreading, finely dissected sterile fronds. North American parsley-ferns currently known include the widespread C. acrostichoides R. Br., and C. sitchensis (Rupr.) T. Moore [C. acrostichoides var. sitchensis (Rupr.) C. Chr.], a taxon of Alaska and adjacent northwest Canada. The delicate calciphile C. stelleri (S. Gmelin) Prantl, the sole member of section Homopteris (Rupr.) C. Chr., is widespread in northern North America. Biosystematic studies of Cryptogramma have been conducted with the aim of producing a modern taxonomic treatment of the genus as it occurs in North America. This work has led to the conclusion that in addition to the taxa currently known from North America, an additional undescribed species is worthy of recognition. The purpose of this paper is to validly publish the name of this species in preparation for the treatment of Cryptogramma in Flora of North America, and to provide a discussion of the diagnostic features of the new taxon. Additional biosystematic data and an analysis of evolutionary patterns, currently in preparation, will be forthcoming. Cryptogramma cascadensis Alverson, sp. nov. (Fig. 1)—TyPE: U.S.A., Washington, King Co., 5 km NW of Snoqualmie Pass, above Source Lake along the old trail to Snow lake, on open southeast facing talus of an avalanche track below Chair Peak, T23N R11E S30, 1110 m, 6 Oct 1984, Alverson 876 (holotype OSC, isotypes MO, NY, ORE, UC, WS, WTU). Ab C. acrostichoides R. Br. foliis textura tenuiore, deciduis, non in anno sequenti marcentibus, et ab C. crispa (L.) R. Br. foliis sterilibus raro quadripinnatis differt. Small, clumped, finely dissected ferns with deciduous fronds. Rhizomes decumbent to erect, strongly multicipital, stout, 4-8 mm wide including old attached frond bases; densely clothed with broadly lanceolate to linear scales up to6 x 2mm:scales generally bicolorous. Fronds strongly tufted, the fertile erect, 5—25 cm long, the sterile spreading, 3-20 cm long; deciduous and usually not isting into the following year, soft, glabrous. Stipes ca. 1 mm wide when dry, ar collapsing and strongly furrowed, green to st dark brown only at very base; stipe scales like those of the rhi m , becoming 7 sparse above. Fronds deltate to ovate-lanceolate, 1/2 to equaling petiole, 2—3 MISSOURI BOTANICAL SEP 6 1989 GARDEN LIBRARY 96 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 3 (1989) Urexon State University PLANTS OF WAS Fic. 1. Photograph of the holotype of C. cascadensis. E. R. ALVERSON: CRYPTOGRAMMA 97 times pinnate, thin and translucent when dried. Segments of sterile fronds typically cuneate-based, oblong to flabellate, widest at a point 1/2 to 2/s of the way above the base; the apical 1/2 to 1/3 regularly dentate, and often more deeply incised every 2nd—4th tooth; segments of fertile fronds ascending to erect, strongly differentiated from sterile fronds, linear, 3-12 x 1-2 mm; veins free, branching once or twice, veins ending in elongate, clavate to linear, surficial hydathodes. Sori round to oblong, coalescing at maturity, fertile segments revolute, protecting sporangia, at maturity often becoming plane with drying and exposing the receptacular surface; spores tetrahedral, yellow, avg. 49.6 »m in diameter; receptacular paraphyses stalked, capitate, unicellular, often abundant, particularly along the margins of the fertile segments. 2n = 30 II. pea Pe yee Rep tati peci CANADA. British Columbi Range, 20 Aug 1885, Macoun s.n (CAN, NY); Mt. Cheops, 7500 ft, 13 Aug 1904, Heacock 459 (COLO, GH, NY, RM, US); The Lions, 4600 ft, 27 Jul 1961, Peterson s.n. (UBC); Mt. Revelstoke N.P., between Millar Lake and Jade Lake, 7000 ft: 15 Aug 1969, Soper Shct I k& Szczawinski 12,491 (CAN, CAS, V); Mt. Lindeman, W of Chilliwack Lake, 14 Sep 1984, Ceska & Ogilvie 18,487 (V). U.S.A. California. Eldorado Co., Susie Lake, 7650 ft, 13 Aug 1909, McGregor 108 (CAS, NY, US); Fresno Co., Vidette Meadows, 25 Jul 1916, Campbell s.n. (CAS); Modoc Co., below Eagle Lake, Warner Mts., 7600 ft, 13 Aug 1918, Jepson 7961 (JEPS, MICH); Mono Co., Slate Cr. Basin, E of Mt. Conness, 11,000 ft, 9 Sep 1934, Cl 993 (CAS, US); Nevada Co., 0.5 mi W of Basin Pk., 8200 ft, 6 Sep 1971, True 6997 (CAS); Plumas Co., Mt. Harkness, Lassen Volcanic N.P., 7400 ft, 23 Jul 1957, Gillett 887 (CAS, JEPS, MICH); Shasta Co., Mt. Lassen, timberline, 9 Aug 1931, Copeland 1424 (CAS, POM, UC); Siskiyou Co., a Horse Camp, Mt. Shasta, 8250 ft, 18 Aug 1938, Cooke 11,502 (CAS, GH, UC); Tehama Co., Brokeoff Mt. Trail, Lassen Volcanic N.P., 7000 ft, 13 Aug 1957, Gillett 1056 (CAS, JEPS, MICH); Tuolumne Co., W of Fairview Dome, Yosemite N.P., 22 Aug 1922, Hall 170 (UC). Idaho. Adams Co., Black Lake, Seven Devils Mts., 20 Jul 1931, Johnston s.n. (CAS); Blaine Co., divide between Alpine Creek and Twin Lakes, Sawtooth Primitive Area, 10,000 ft, 30 Jul 1944, Hitchcock & Muhlick 10,500 (NY, UC, UTC, WS, WTU); Bonner Co., Priest River Experimental Forest, 22 Jul 1943, Daubenmire 43,261 (NY, WS, WTU); Elmore Co., 1 mi S of Lower Spangle Lake, Sawtooth Primitive Area, 19 Jul 1944, Hitchcock & Muhlick 10,146 (CAN, CAS, GH, NY, RM, UC, US, UTC, WS, WTU); Idaho Co., Cool Water Mt., 11 Jul 1936, Gail s.n. (ID); Kootenai Co., without locality, 1891, J.B.L. 11, (UC); Shoshone Co., Freezeout Summit, 27 Jul 1958, Baker 15,415 (ID, NY, WTU); Valley Co., Brundage Mt., 7600-7800 ft, 5 Jul 1937, Pennell & Constance 20,749 (US). Montana. Missoula Co., Squaw Peak, 2375 m, 7 Jul 1964, Harvey & Pemble 7075 (MONTU); Ravalli Co., Mt. Jerusalem, 9000 ft, 11 Aug 1968, Lackschewitz & Fageraas 693 (MONTU). Oregon. Baker Co., W. base of Red Mtn., Wallowa Mts., 7500 ft, 5 Sep 1957, Head 1637 (NY, OSC); Clackamas Co., Breitenbush Lake area, 24 Aug 1962, Rodin 6926 (ARIZ); Deschutes Co., Hidden Lake, Paulina Mts., 8 Jul 1928, Detling 28 (ORE); Douglas C , Old Bailey Mt., 7 Jul 1924, Applegate 4125 (CAS, WILLU); Hood River Co., near Eden Park, Mt. Hood, 6000 ft, 5 Jul 1926, English 174 (WS); Jackson Co., Mt. Pitt, 7000 ft, 27 Jul 1887, Colville & Applegate 233 (US); Jefferson Co., Three-Fingered Jack, 6500-7000 ft, 5 Sep 1976, Crosby 1073 (OSC); Klamath Co., Wizard Island, Crater Lake N.P., 21 Aug 1949, Baker 6363 (ID, NY, OSC, RSA, UC, WS, WTU); Lane Co., West Lava Camp, McKenzie Pass, 5200 ft, 22 Aug 1937, Ireland 1025 (ORE); Linn Co., Mt. Washington, 6000—7000 ft, 31 Aug 1976, Crosby 991 (OSC); Wallowa Co., Chimney Lake, Wallowa Mts., 21 Jul 1950, Kruckeberg 2367 (CAN, COLO, ID, NY, RSA, WS, WTU). Washington. Chelan Co., slopes of Mt. Stuart, 1520 m, 28 Aug 1893, Sandberg & Leiberg 1821 (NY); Ferry Co., Twin Lakes, 3500 ft, 5 Sep 1927, St. John 8874 (WS); King Co., Snow Lake trail, 3600 ft, 22 Sep 1986, Alverson 1036 (OSC); Kittitas Co., Stafford Creek drainage, 1.5 mi SE of Earl Peak, 5800 ft, 14 Aug 1981, Alverson 534 (ORE); Lewis Co., Reflection Lake, Mt. Ranier, 5000 ft, 23 Aug 1901, Flett 1923 (NY, WTU); Pierce Co., Glacier Basin, Mt. Ranier, 6700 ft, Aug 1925, Grant s.n. (CAS, WTU); Skamania Co., Mt. St. Helens, 4500 ft, 4 Aug 1925, St. John et al. 7453 (WS); Snohomish Co., Lake Serene, 2600 ft, 10 Jul 1983, Alverson 584 (ORE); Yakima Co., Wodan’s Vale, Mt. Adams, 4 Oct 1902, Suksdorf 2793 (WS). 98 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 3 (1989) Morphological comparisons.—This new species has not previously been distinguished from C. acrostichoides. The two species are most clearly separated on the basis of habit, texture, and micromorphology, although they also differ more subtly in typical outline and segment shape of sterile fronds. Evidence from common-garden trials demonstrates that the traits by which C. cascadensis and C. acrostichoides differ are intrinsic to each species, and not a product of differing environmental conditions. Numerous transplants of both taxa have been cultivated in a greenhouse for three years, and all of the diagnostic differences have been maintained. The best diagnostic feature distinguishing the two taxa is that the sterile fronds of C. cascadensis are fully deciduous in the autumn, while those of C. acrostichoides are evergreen or nearly so. In natural habitats, the withered fronds of C. cascadensis quickly decay and detach from the rhizome, with little accumulation of dead foliage around the base of the plant. The fronds of C. acrostichoides are strongly marcescent, often resulting in the accumulation of a substantial quantity of dead foliage and organic matter at the base of the plant. Though diagnostic for C. acrostichoides, these old fronds are often removed by collectors in an effort to ‘‘tidy up’’ specimens, so this character may not always be evident on herbarium specimens. Fronds of C. cascadensis are relatively thin and soft, and upon drying become more or less translucent. The sterile fronds of C. acrostichoides are thicker, with a coriaceous texture when mature. The characteristic color of C. cascadensis is grass green, while mature fronds of C. acrostichoides area darker verdigris green. Furthermore, the abaxial surfaces of the sterile fronds of C. acrostichoides are lighter in color than the adaxial surfaces, in contrast to the uniformly colored surfaces of C. cascadensis. The stipes of C. acrostichoides, particularly those of the fertile leaves, are rigid and straw-like, even upon drying. Those of C. cascadensisare less firm, and upon drying, collapse inward, so that when dry the diameter of the stipes is about 1 mm, compared to the 2 mm di ter typical inC. acrostichoides. Hydathodes occurring at vein endings on the adaxial surface of the fronds of C. cascadensis are elongate, clavate to linear, and are flush with the lamina surface. The hydathodes of C. acrostichoides are sunken below the lamina surface, due to the thicker, more coriaceous lamina. They are also shorter and wider, typically ovate to shortly clavate in outline, only oecasionally approaching linear in outline. Scattered along both surfaces of the fronds of C. acrostichoides are very small, appressed, unbranched cylindrical trichomes. These trichomes are most numerous in the sulca of the rachis and along the costae of the pinnae and pinnules. Such trichomes are essentially lacking in C. cascadensis, and their absence is ful ter that distingui this species. Frond outlines and shapes of ‘sterile segments are extremely variable in t fronds of a single plant, a factor that has caused considerable confusion i in the past (Fernald, 1935). Both species are character- ized by a distinctive ‘‘profile,” but within-species var iability is sufficiently great that these characters are not always diagnostic. Sterile laminae of C. cascadensis E. R. ALVERSON: CRYPTOGRAMMA 99 are typically deltate in outline, while those of C. acrostichoides are often more ovate-lanceolate, but plants with more deltate laminae do occur. Shape of the ultimate segments varies in a similar manner; those of C. cascadensis are typically flabellate in outline, while those of C. acrostichoides tend to be oblong to ovate-lanceolate, but with considerable variation in eac In favorable habitats, such as moist subalpine talus slopes, C. cascadensis is very strongly multicipital, growing into large patches bearing many hundreds of fronds. Cryptogramma acrostichoides is typically much less strongly multicipital, usually forming smaller, discreet clumps. Evolutionary considerations.—The significance of these morphological distinctions is bolstered by biochemical data obtained from allozyme analysis. Nei’s Genetic Identity statistics obtained from an electrophoretic study of 13 populations C. acrostichoides and C. cascadensis showed a high degree of genetic differentiation between the two species, with a mean genetic identity of 0.36 for interspecific population comparisons (Alverson & Windham, unpubl. data). This value is comparable to genetic identities of congeneric fern species obtained in previous electrophoretic studies (Haufler, 1987). Cryptogramma cascadensis shares several morphological features with the European C. crispa, a tetraploid species with 2n = 60 pairs of chromosomes (Manton, 1950), such as the thin-textured, deciduous sterile fronds and surficial hydathodes. However, C. cascadensis, like C. HErosLChoiies, is Ariki with s from C. 4 2n = 30 pairs of chromosomes (Fig. 2). Cryptog crispa also differ Fic. 2. Meiotic ck iC. lensis (Al 1036, OSC). 100 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 3 (1989) cascadensis in several morphological characters, including sterile leaves typically larger and more finely dissected, with many more ultimate segments; firm, straw-like petioles; and concolorous rhizome scales. While the true evolutionary relationship between C. cascadensis and C. crispa has not been established, their morphological and cytological distinctness suggests that they are not conspecific. Other species of Cryptogramma sect. Cryptogramma from different geographic regions, such as C. fumariifolia (Phil.) Christ of South America, C. raddeana Fomin of northeast Asia, and C. brunoniana Hook. & Grev. of the Himalayas, all appear by virtue of their evergreen sterile fronds to be more closely allied to C. acrostichoides than to C. cascadensis or C. crispa. Geographic distribution—The epithet cascadensis refers to the Cascade mountain range, where this species is a typical inhabitant of rocky subalpine habitats. Cryptogramma cascadensis occurs along the entire length of the Cascades, from southern British Columbia, through Washington and Oregon, to the volcanic peaks of Mt. Shasta and Mt. Lassen (Fig. 3). The distribution of C. cascadensis extends southward in California along the high Sierra as far as Fresno County. A second center of distribution for this species is in the northern Rocky Mountains of Idaho, western Montana, southeast British Columbia, northeast Washington, and northeast Oregon. Here C. cascadensis is a representative of the coastal or Cascadian floristic element that is disjunct in the high rainfall regions of the northern Rockies (Daubenmire, 1975). Cryptogramma cascadensis typically grows on granitic and volcanic rocks; it occurs, for example, on nearly every volcano in the Cascade Range. Elevations range from as low as 1000 m in the Washington Cascades and as high as 3400 m in the Sierra Nevada. A complete listing of herbarium specimens annotated as C. cascadensis, including locality data, is available from the author upon request. ryptogramma cascadensis is not known with certainty from the Olympic Mountains of Washington, although suitable habitats must exist. Two herbarium sheets with C. cascadensis reputedly collected in the Olympic Mts. have been examined (without specific locality, Piper 1905, WTU; Elwha Basin, Leach & Leach s.n., ORE), but both are mixed sheets also containing plants of C. acrostichoides. In the absence of further corroborating evidence, these records are presumed to be in error. Ecological considerations—Morphology of C. cascadensis, particularly the deciduous habit and thin frond texture, suggest that this is a species most suited to mesophytic habitats in regions with deep winter snow accumulations. In contrast, C. acrostichoides is a relatively xerophytic species, with thick evergreen fronds that withstand significant moisture stress, and in addition can photosynthesize in autumn, winter, and early spring (if not covered by snow). This assessment is supported by field observations. At the type locality in Washington near Snoqualmie Pass, where both species are present, colonies of C. cascadensis are generally found in habitats that are released from the snowpack later in the season, either because of concave microtopography, or because of heavy snow accumulation due to winter avalanches. In the subalpine of the Sierra Nevada, at the outlet of Heather Lake, Eldorado Co., California, C. cascadensis grows in its typical habitat, a cool, north-facing talus slope. At the 101 -E. R. ALVERSON: CRYPTOGRAMMA 400 MILES more herbarium is. Each dot represents one or dens Fic. 3. Geographic distribution of C. casca collections. 102 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 3 (1989) same locality, C. acrostichoides is abundant on the opposing south-facing slope, where in late August of 1987 the plants were completely withered by drought. Like many cheilanthoid ferns, C. acrostichoides regularly survives periods of drought with dehydrated, curled, and brittle fronds that readily rehydrate and resume normal functioning when moisture becomes available. Cryptogramma cascadensis apparently does not possess this drought tolerance ,and must avoid moisture stress by occupying mesic microsites, such as Heather Lake. When cultivated plants of C. cascadensis were left unwatered, their fronds at first wilted, then withered and died, and did not come back to life when watered again. In the texture and color of its foliage, C. cascadensis is remarkably similar to Athyrium alpestre ssp. americanum, a common associate in moist subalpine habitats, a convergence attributable to selection for similar ecological conditions. wel oad correlation of a suite of subtle but distinct morphological featu ith ts and a coherent geographical es "pnvidos ample evidence that an additional distinct member of the Cryptogramma crispa complex, Cryptogramma cascadensis, is present in western North America ACKNOWLEDGMENTS I thank K. Chambers, C. Haufler, T. Kaye, A. Smith, D. Wagner, W. Wagner, and M. Windham for assistance received during the course of this study. Field work was supported by the Oregon State University Herbarium and a Sigma Xi Grant-in-Aid-of-Research. I am grateful to the curators of the following herbaria for the loan of specimens: ARIZ, ASU, CAN, CAS, DAO, GH, ID, al MICH, MONTU, NY, ORE, OSC, POM, RM, RSA, SD, SOC, UC, US, UTC, WILLU, WS, and WTU LITERATURE CITED DAUBENMIRE, R. 1975. ao plant geography of eastern Washington and northern Idaho. J. iog —. M. L. 1935. “Critical plants of the upper orn Lakes foes . Ontario and Michigan. dora 37:2 HAUFLER, C.H. 1987. Electrophoresis is oe our span vi Soden in homosporous pteridophytes. Amer. J. Bot. 74:9 66. MANTON, 1950. Problems of Cytology Hot Evolution in the Pteridophyta. Cambridge: University American fern journal 79(3): 103-118 (1989) g | 000 gq E20 Zygophlebia, a New Genus of Grammitidaceae L. EARL BISHOP Herbarium, Department of Botany, University of California, Berkeley, California 94720 In my studies for the generic realignment of the Neotropical, anhydathodous Grammitidaceae, it has become clear that there is a small group of paraphysis-bearing species that is quite distinct from Ceradenia (Bishop, 1988). This assemblage includes seven American species and one or more from Africa. Of the American species two are fairly common and widespread, three are rare and localized, and two are new. Surprisingly, there has never been a suggestion of the interrelationship of any of the described species, but the continuity of these taxa now seems obvious. The decision to erect a new genus, instead of including these species in Ceradenia as subgenus, has required careful consideration. In addition to the correlated morphological and anatomical characters, an important factor has been that the ease of identification and mutual hierarchic integrity of the related genera would be undermined by the inclusion of these species in a single genus. Zygophlebia is named for the strong tendency for the fronds to show areolate venation. Regular anastamoses are present in four species, in the others they are irregular to a varying extent. Anastamosing veins are uncommon in the Grammitidaceae, and at least in the New World, Zygophlebia is the only genus with pinnate or pectinate-pinnatifid species that have regularly anastamosing veins (Fig. 3C). The most densely areolate species is Z. cornuta in which the fertile veinlets are usually prolonged beyond the sorus to form costal areolae and the sterile veinlets are once or twice forked and form an irregular reticulum or at least a series of distal areolae. The clearly related Z. sectifrons shows generally regular costal areolae with free or irregularly connivent sterile veins. Zygophlebia mathewsii and Z. werffii exhibit a fairly regular intramarginal vein formed by the fusion of the sterile veinlets. The fertile veinlets are only irregularly prolonged and fused with the next distal sterile veinlet. The largest plants of Z. longipilosa exhibit a similar venation pattern, but the smaller plants show only irregular fusion of the sterile veinlets, as is the case of the smallest plants of Z. mathewsii. The species with veins most regularly free are Z. dudleyi and Z. eminens. However, in these the sterile veinlets are often forked and irregular, and marginal fusion is not uncommon. Irregular anastamoses are seen in various species of Ceradenia subg. Filicipecten. The most common type involves the connivence of a shortly prolonged fertile veinlet with the sterile fork of the same vein. Costal areolation formed by the fertile veinlet’s fusion with the next distal sterile veinlet or the marginal connivence of sterile veinlets is quite rare. Except at the very base of the pinnae of certain species, even the sterile veinlets are rarely forked distally. It is of note that Z. werffii and large examples of Z. longipilosa, both usually with intramarginal veins, are smaller than larger examples of Ceradenia, so that the argument that areolation in these ferns is a direct response to larger size is vitiated. It does seem clear that in Zygophlebia itself relatively free vein patterns 104 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 3 (1989) are generally correlated with smaller frond size. The pattern of irregular anastomoses in Zygophlebia species with mostly free veins to me strongly suggests their derivation from a larger, areolate ancestor. On the other hand, the lack of a similar venation pattern in even the largest species of Ceradenia seems to point to an ancestral stock with free veins. Few contemporary pteridologists would recognize genera based solely on venation. The clearest difference between Zygophlebia and Ceradenia is found in the paraphyses. Morphologically, these are similar in both genera. They consist of a uniseriate stalk, the terminal cell of which bears two glands distally. Some species of Zygophlebia have only these two glands; others bear a third distolaterally from the subterminal stalk cell. Ceradenia species have paraphyses with 1—4 subterminal cells bearing glands. The lowest of these glands may be supported by a stalk cell. Paraphyses of each genus protect the developing sporangia and may b idered functionally mature at the time the capsules are exserted between them. At this time the paraphyses of Zygophlebia are brown and thickly viscid, so that in dried specimens at least, the entire sorus adheres into a single sticky mass. Microscopically, the brown glands never show an external accumulation at any stage. In Ceradenia the paraphyses are waxy white, yellowish, tan, or red (unless colorized by specimen preparation) and show no tendency to adhere to each other or to the sporangia. The appearance of the glands microscopically is striking. Each is overlain by colorless, rough-textured excretion. This substance is insoluble in water, insoluble or weakly soluble in ethanol, and quickly dissolved by xylene. The positive affinity for osmic acid stain suggests a composition of long-chained fatty acids. This accumulation is still quite evident even when the color is changed by specimen preparation. Another striking, if recondite, difference between these genera is found in the rhizome. In both the stele is siphonostelic, but in Zygophlebia there are ventrolateral perforations not associated with the leaf traces (Fig. 1). [have seen such gaps in Z. mathewsii, Z. villosissima, Z. sectifrons, Z. cornuta, Z. dudleyi, and Z. werffii. Material for sectioning has not been available for Z. longipilosa and Z. eminens, but their close relationship to other species makes it unlikely that either would be aberrant in this respect. No species of Ceradenia, of either subgenus, is known to exhibit such accessory stelar perforations. Two additional characters common to all Zygophlebia species merit note. Simple or branched hairs with terminal, ultimately sticky-viscid glands occur in all species on the lamina and usually on the stipe as well. In Z. mathewsii and its relatives these are usually branched and spreading from the surface of attachment. In the Z. sectifrons group the hairs, especially on the laminar surface, are generally simple and appressed. As the frond matures they become adherent to the surface. In the most extreme case, that of Z. sectifrons, the hairs on mature fronds appear as dark, scarcely differentiated ‘squamulae’ whose morphology is scarcely discernible. In Z. werffii the marginal and ventral laminar hairs are mostly branched and spreading (Fig. 3C), whereas those of the dorsal surface are simple and appressed. The rhizomes of all species support the stipes on distinct, bulbous L. E. BISHOP: ZYGOPHLEBIA 105 vl leaf trace race ~ : leaf gap bud trace re ) _ oe root trace a ee j ventral gap root trace ventral gap Fic. 1. Zygophlebia sectifrons. Diagram of stelar cross section. phyllopodia. Most larger members of the family have phyllopodia of some sort, but they have been little discussed. In Zygophlebia they are well developed. The cortical tissue of which they are composed is usually distinctly darker (dark brown or blackish) than the stipe. A bud is conspicuously visible at the stipe base of younger fronds as a whitish, parenchymatous mass on the exteriolateral side of the phyllopodium. This structure is scarcely noticeable in older fronds unless developed as a branch. The abscission layer is clearly functional. The fallen stipe leaves a shallowly concave scar normally surrounded by an annular thickening. The eight species of Zygophlebia fall into three distinct groups. Zygophlebia mathewsii, Z. longipilosa, and Z. villosissima all have laminar setae, which the other species lack, and all three show fundamentally distal soriation. The sori never extend to the rachis and they usually reach the tips of the pinnae (occasional specimens of Z. mathewsii may have sterile segment tips). The Z. sectifrons group also includes Z. cornuta, Z. eminens, and Z. dudleyi. These four epiphytic species letely lack laminar setae and their soriation is essentially basal. The sori often reach the rachis and the segments show a sterile apical portion. Among these species Z. eminens and Z. dudleyi appear closely related by virtue of their smaller, paler scales and the presence of scattered, long setae along the dorsal rachis. Zygophlebia sectifrons and Z. cornuta have long, relatively narrow, medium brown scales, areolate venation, and lack setae on the rachis (except in occasional specimens of Z. sectifrons). Zygophlebia werffii stands apart from the rest. It is terrestrial and nearly pinnate, with ciliate scales, blackish stipes, erect, unflexed fronds, and numerous, short setae on the rachis. Like the species of the Z. sectifrons group, the pinnae lack setae and the soriation is basal, so that the relationship of Z. werffii may be closer here than to the Z. mathewsii group. Although these groups are quite distinct, because of the few species involved, it seems unnecessary to recognize them formally as infrageneric taxa. 106 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 3 (1989) I believe Zygophlebia to represent a primitive genus among the Grammitidaceae. To support such a statement requires a discussion of phyletic trends in the family. The fund tal question in this regard is whether the most ancestral type was a fairly large plant with a more complex venation and stelar organization, or whether the entire family arose from a stock of already much reduced plants. I believe the development of the Grammitidaceae to be an example of ferns that have in general become morphologically less complex over evolutionary time. Demonstrating a sequential series of character states is usually straight- forward; interpreting the phyletic direction of such a series requires care. A reasonable deduction from the tenets of natural selection is that if a complex structure is lost, it is unlikely to reevolve in the same form. A major anatomical discontinuity in the Grammitidaceae is the presence or absence of an internal endodermis. One might expect the simpler state of lacking this structure to be associated with smaller rhizomes. However, many small species show the more complex organization (e.g., Cochlidium, Lomaphlebia, many southeast Asian species of Grammitis sensu Parris, 1983), whereas Adenophorus, which has _ some rather large species (A. tamariscinus, A. tripinnatifidus, A. periens), lacks an internal endodermis, as do some large Neotropical species such as G. semihirsuta. These observations suggest that once lost, the internal endodermis is not subsequently redeveloped even in a phyletic line becoming larger in rhizome size. Accessory stelar perforations are found at least in Zygophlebia, Ctenopteris, and Lomaphlebia, all of which show an internal endodermis. Although the argument in this case is admittedly tentative, considering the previous discussion of these perforations in connection with Zygophlebia and Ceradenia, it seems likely that these are more easily lost in phyletic lines showing size reductions than they are apt to be developed de novo with increased size. The ancestral frond pattern in the Grammitidaceae was probably pectinate. The genus Enterosora provides an illuminating situation in this regard. Although closely related to Zygophlebi d Ceradenia, most of the species have more or less simple fronds. The largest species (e.g., E. trifurcata (L). comb. ined.) do have deeply lobed, pinnatifid fronds, but the lobes here are very broad and mostly rounded, not at all like the pectinate laminae found throughout Zygophlebia and Ceradenia subg. Filicipecten. This suggests that pinnatifid laminae in Enterosora have derived from simple fronds and that pectinate fronds to not directly result from such a derivation. A dorsiventral rhizome probably represents the more primitive condition in the family relative to the radially symmetrical rhizome. Except for debris-collecting epiphytes such as the Asplenium nidus group, it is reasonable that a creeping, dorsiventrally organized rhizome is functionally adaptive for all but very reduced epiphytes. Most Grammitidaceae with radially symmetrical rhizomes, such as Grammitis sensu stricto and Cochlidium, are quite small. Ceradenia subg. Ceradenia, also with sucha rhizome, includes some rather large species, but, as I will discuss elsewhere, I believe the ancestral stock of that subgenus was probably a small species, such as C. jungermannioides or C. pruinosa. L. E. BISHOP: ZYGOPHLEBIA 107 Among those species with pinnate or pinnatifid fronds, forked veins are almost certainly primitive. In groups in which the fertile veinlet has been suppressed (such as the G. cultrata and G. moniliformis groups), larger species do not develop forked veins, even though careful examination of the veins will often clearly suggest their derivation from a more elaborate arrangement. I have posited that in Zygophlebia, areolate veins represent the more primitive state, and I think it likely that the ancestral Grammitidaceae also had such a venation pattern. Whether the presence of hydathodes is ancestral or derived can best be supposed by correlation. The putatively primitive character states just discussed are found in the three large, anhydathodous American genera (Ceradenia, Zygophlebia, Enterosora). On the other hand, most of the derived states are more common in the Grammitidaceae with hydathodes. In summary, the primitive Grammitidaceae were probably relatively large ferns with pectinate, anhydathodous fronds, areolate or at least forked veins, and a dorsiventral rhizome with an internal endodermis and accessory stelar perforations. In other words, they were very much like some species of Zygophlebia, and I take this genus to be the most primitive extant example of the family. Unless otherwise noted, I have examined all specimens cited. Loans were obtained from AAU, B, BM, C, F, GH, K, MO, NY, P, S, and US, and I thank the curators for making these specimens available. Full specimen citations for Z. mathewsii and Z. sectifrons are available on request. 266 |254 7 Zygophlebia L.E. Bishop, gen. nov—Type: Polypodium sectifrons Kunze ex Mett. . n34arim t circulum specierum majorum quart gnatio inter se ad nunc non percipiebatur. Rhizoma est validum dorsiventrale phyllopodios bene effectos ferens interne siphonostelam amphiphloicam cum endoderme interna et perforationibus accessoriis ventrolateralibus, externe paleas concolores nitentes vel subnitentes in margine glandulis unicellularibus et per unam speciem ciliis praeditas ostendit. Lamina profunde pinnatifida vel perpinnata (una specie) cum aut sine setis, pilos glandulares simplices vel ramosos gerens, venis 1—2(3)-furcatis nunc regulatim conniventibus nunc plusminusve liberis, hydathodis caret. Sori submediales vel mediales ad venulam acroscopicam siti, setis circumsoralibus in 2 speciebus invenitis, sporangia capsulis 195-320 x 155—265 pm harum annuli ex 12—18 cellulis constantes, sporas hemisphaericas vel subtetraedricas aliquando binucleatas 28-60 »m in diametro longiore includentibus, necnon paraphyses quae gerrent distaliter 2—3 glandulas brunneas viscidas autem his strata externa exsudati solidi ceracei microscoplo visibili carentibus, comprehendent. KEY TO THE NEOTROPICAL SPECIES 1. Lamina setose; sori usually distal on the segments, not extending to the rachis. 2. Rhizome scales dark, lustrous b tae scattered the dorsal lamina ve 1. Z. mathewsii (Costa Rica, Panama, Andes) .....----+++-rest ttt’ 108 AMERICAN FERN JOURNAL: VOLUME 79 NUMBER 3 (1989) 2. Rhizome scales (at least the younger ones) tan or light brown; setae on the dorsal lamina localized on the costa and around the sorus (Brazil) ....... bray es Ce lees SSS eS. 2. Z. longipilosa 1. Lamina without setae (unless near rachis); sori basal, often reaching the rachis, but not extending to the segment tips. 4. Rhizome scales medium brown, up to more than 10 mm: rachis usually without scattered setae dorsally. 5. Pinnae linear-elongate, set at 20—60° angle to the rachis, separated by 1-5 times their width (Caribbean, Costa Rica, Panama, Venezuela, WM a ee 3. Z. sectifrons 5. Pinnae linear-triangular, set at 70-90° angle to the rachis, separated by 0.5—1.5 times their width (Costa Rica, Panama) ....4.Z. cornuta 4. Rhizome scales tannish brown, usually less than 8 mm long; rachis with scattered setae dorsally. . 6. Stipe with few, scattered setae; pinnae approximate, separated by less than theie width (ecuador). . -.. oo 5. Z. eminens 6. Stipe densely setose in proximal half; pinnae distant, separated by at Om te Wint (Pany 6. Z. dudleyi “1. Zygophlebia mathewsii (Kunze ex Mett.) L.E. Bishop, comb. nov. Polypodium mathewsii Kunze ex Mett., Abh. Senckenberg. Naturf. Ges. 2:74. 1856.~Grammitis mathewsii (Kunze ex Mett.) C. Morton, Amer. Fern J. 60:66. 1970.“Tyrr: “Peru (Mathew).” There are three Mathews’ sheets at B. One isa frond ofa larger plant with a questioned collection number (3281) which is not duplicated at other herbaria. The two other sheets bear the number 1811 and represent a smaller expression of the species. Of these I select the sheet from Mettenius’ herbarium a ectotype. Other isolectotypes are at BM, K, P. Goniophlebium villeminianum Fée, Mém. foug. 7:63, t.27, £.3. 1857.