Р з... > N, EE ч r че у T +
Е PFaleontographica Americana -
Fruits and Seeds of the Middle Eocene Nut Beds Flora,
Clarno Formation, Oregon
by
Steven R. Manchester
PALEONTOLOGICAL RESEARCH INSTITUTION
Officers PRESIDENE Se oe Bork EL DUMP ee JOHN C. STEINMETZ NIT PRESIDENT $99 ао RICHARD E. PETIT ЕСЕРШ & 0 een re E n ЫЕ? HENRY W. THEISEN йок ап ошку ne ee ANNIKA FARRELL DIRECTORA. ee ee eg WARREN D. ALLMON Trustees R. TUCKER ABBOTT (to 6/30/96) PETER MCLAUGHLIN (to 6/30/95) BRUCE M. BELL (to 6/30/96) SAMUEL T. PEES (to 6/30/95) CARLTON E. ВвЕтт (to 6/30/95) RICHARD E. PETIT (to 6/30/96) WILLIAM L. CREPET (to 6/30/97) GARY ROSENBERG (to 6/30/96) J. THOMAS DUTRO, JR. (to 6/30/96) CONSTANCE М. SOJA (to 6/30/97) ANTON J. EGNER (to 6/30/97) JAMES E. SORAUF (to 6/30/97) ANNIKA FARRELL (to 6/30/95) JOHN STEINMETZ (to 6/30/97) HARRY G. LEE (to 6/30/97) SUSAN B. STEPHENS (to 6/30/96) ROBERT М. LINSLEY (to 6/30/95) HENRY W. THEISEN (to 6/30/95) Amy К. McCuNE (to 6/30/97) THOMAS E. WHITELEY (to 6/30/97)
Trustees Emeritus
HARRY А. LEFFINGWELL RAYMOND VAN HOUTTE WILLIAM P. S. VENTRESS
BULLETINS OF AMERICAN PALEONTOLOGY
and
PALAEONTOGRAPHICA AMERICANA
WARREN DSATLEMON o уу een EDITOR
Reviewers for this issue
BRUCE Н. TIFFNEY ELSE MARIE FRIIS
A list of titles in both series, and available numbers and volumes may be had on request. Volumes 1—23 of Bulletins of American Paleontology have been reprinted by Kraus Reprint Corporation, Route 100, Millwood, New York 10546 USA. Volume 1 of Palaeontographica Americana has been reprinted by Johnson Reprint Corporation, 111 Fifth Ave., New York, NY 10003 USA.
Subscriptions to Bulletins of American Paleontology may be started at any time, by volume or year. Current price is US $50.00 per volume. Numbers of Palaeontographica Americana are priced individually, and are invoiced separately on request.
for additional information, write or call:
Paleontological Research Institution 1259 Trumansburg Road Ithaca, NY 14850 USA
(607) 273-6623 FAX (607) 273-6620
©) This paper meets the requirements of ANSI/NISO 239.48-1992 (Permanence of Paper).
- Paleontographica
_ Americana
Begun in 1916
| | NUMBER 58 AUGUST 24, 1994 |
Fruits and Seeds | of the Middle Eocene Nut Beds Flora,
Clarno Formation, Oregon
by
Steven R. Manchester
Paleontological Research Institution 1259 Trumansburg Road Ithaca, New York, 14850 U.S.A.
ISSN 0078-8546 ISBN 0-87710-435-2 Library of Congress Card Number: 94-65491
Printed in the United States of America Allen Press, Inc. Lawrence, KS 66044, U.S.A.
| |
CONTENTS Page A ee een ee a о е ое ee. Den Inc o г: чо ома (еко ти (ери Ни Б с ке а Te x cd E Е а 7 ОСОО е ын сокет тена ЕСЕК К ек eco До не оо бойун киш каек en Se 8 сорар аео Sein ы A аве скн ы at AEC mh хел Red Гут ue qM т ЖЕТ 9 OE реве DMN ы Вир aN ы ML и ди ы кю ы О ыры UM С ааа en 12 О О Раоа ЈАНО а А И Кеде ee a eas ei cU rm ERO eh 13 MOU BSSOfaBLESETVaON е со СЕ M а c c M M iT ы mb 13 Маи М СТВО И ee уы са E ыу bade ои 14 СЕЗА ыст уы Аш ык mos TUS du win vad а СК кы зк ке ы жо en A 15 Sy От DN МВ Sab ӨӨ en ee ose c ME M Mc oce cm А 15 ¡STO WMA TURO dais por Бо НУ ама а не канау лж М ке Е ее cud Crew chute) тул e г: Ж 19 Саа Ө ИЕШЕ ШӘ сокет и за earns, С ЕЕ mS s dae. 22 BIOECOETADLICKEONSIUSLHLIONSE ео 23 (ВО РАНО Ко Бети Бо Ши ӨТӨН sac ee 23 ООШ ОШКОШ ШИК о ТЕ ШАН С а ин чс ез eu rec ees Е а ИЕ oy c у ae у ды 26 OLSON OSY a БЕ ЛООНУ узу Иол. wy оса а и Ер у Ай 27 Ore ыссы ыы MIT ee een 28 Шека DOOR SR а ee ee АК 28 DGHSUTCIDOnisPe NONU c ue D M e era ins Cu mU MD MO bd Cue QR DRE A m AME 30 SDDLENIAUONSE RS cec боза ы cte ed ee ola ey Gee ME он се claim E me 30 Conifers Пи а Mme AX. ee ccu d e C tas KR E EMEN ENSEM оа 30 IHE EAE mE es c die M ed uec etu P а dm 30 ШОК СОЕ КН amc cc UM NECEM UC eL SE CE c MR EE c Ей 31 И промо Mp жар. P cc ur I A скы m CL NEL Mara ed cs tutior itt 31 WOMAN ВАРНЕ REESE, c p Md pM E I ове mu EM mM e rdi ПРОРОК УНЕ UE de o ctu ie exi ee en tds esa A E E ede esed rc plaid 32 Angiosperms ООА СОВЕ M Mee t eU Кы HL UU M I MC m d Muti Ja 33 A лы без син оао D e Чу К е С O cU Werde Lie s СИ ын o, 33 Е ЕО О ИИИ ИАА 34 Zr d а ан Ма dM x M P I MA ek 34 AL Та ел ОМА ИЕ M itr Cha wan I cec E NU MUN S AS EU 35 ПАЗИТИ UU ep кө cR D que rcc M eL o iM ae TN 35 MU ui uum. a ep M ON uu c Ка e mM er Mr mae ы A 35 END IE Sem ee ee ee ee e E 37 EIN ON ОВИ TL) ОЕ e ee Rc а и c. = 37 TAN ACEAE AN N tr na OLE и лане к ОВ RO LL ее е пе tc а О окови 38 ОРИ ИЕ EONA Син К ee Mee N Rad dI LU edt E EE, НИ 38 BCT ACU AORN а LE ои MM Ware па о и иза cc ко | = M 39 ISOTOTSDOUTIORDOD NONAS nih cain шешек ен RE CI к ети wah ede wean car Mees И Ne ыа E реа е ы у 39 ООС ОИ тама ре м беседа е и на е Le UR cm E M ee QI МЫ ы ы и tU 39 ШКЕАК Н единна ИИ INN fa дан erc тл O ко е CT ae 40 IA ү ы у к е NU ВИ Gu soa ы кыч ER e eet АК темя 40 SOT ACen И, Do MID уны uc сы c c E S Lu 41 [ОП КИЕК PAIL ARE Dd RM СИВ cues E эу E кта ме у КЫ о AN a АНУ бт ы г C er. . НИ 41 дум роле EE ы е qua M Би ЕН а тез reU UM и к лы ee ey I 42 Ди но OE aie акны Dy Gap ea RN tone uud c шерт x M uuu cM LE P E 43 КОЧО АИ ОЈ MERE cic орат Г л м ме ым С оо чек NM ee тосе сус E EN 44 ЕТО они ЖИК POCO ee ee direi. uade: 44 D ta ра БР Еа БЕ ма 45 а ет OR лы d E у ccr x nen 46 ЧО О ОЮУ ah E Ок A kno а HI ett 46 (Ори Ке Дуа ED О ME c ен oe ba Е en ис о PP dc d cM E M лоо Лиао NEHME NUT 47 ШЕ ШКЕ Ин ее oltre оо Co о ы c qu c Муш по P mI 47 A АЕ аа ро ену De dra MM Hu up nc ME MT 48 ПУТА БИ ДОЊА oc эз P LN pd pm I M I NM S CE E 49 КОЛО ESE О КОХ КИШКЕ А M Miu 49 ШОЧ КШ КЕДЕ ee REEL Nd C Lo лл RN i. qu. e ri d ER CU 50
КОЛЛ АР em Me dv ge M eMe x ep M st TN л ое г ee a 50
T d Ty a кк Кы cc у озук ИННИ О кн E DNE ы О л, И анас РОТИ di OIE FS OIE RE FREESE у nit н oe рол NN ух сим у ее л и у нил c MS nee RM
URN Оос OR л к и к ee ee A И е лы кор е р сы сс a RE P ЖО шс аа ко ы с т ко сс MA tee И Жк Сол со Lo ee T pue ре Ma Е I QUEEN ei Aa ЧИН Vicus Dp xeu ccce esL VEL m N ръста рапа ш у o о ныды ENANA
ee o ER CL Du ОО ERA PET lu m s I Um ec uU dene MEER AULAM
КОИ ен or nion a ro Duk О ua epi pneu Ex NE NNNM. eoo е су с ann ат oe E a ade ка КАЛИ ee ы A a E a M О ООО АО E E RM S ксы жо S c I UI d UE (oor c M TP LIN MI IE vr кес cuu S MD MM UM ME
ciusdem ecu К узо ee E eM M КИМЕ на tu uu СО EA ОА d I D M АЕ
| Оа аб o deor uote ee death, cee e Ed К шы Net UU е ни EN EEE IE A a вай 90 | FAD HGMGNTH ем ne канам Ee hatte EN Ce at ee Е КУМ оа cd. da erar 90 | WEISE ee ae c cn Macte og E СИ ы NEUE ee 91 d ШЕЙ dela нс E E А E E ee I M E d 92 | ECA CLOSDERIN т М ж E E N ee Н РА м 92 ОВА a E И M tU MR CL cu О DUE у кн A EUM күзү т iE
| LAINDEIOCISSUS о arith eater пи ин E ЛОТ КККК EN ЛУ т о itn coe на е add 93 | TAS RA NS ЕЕ СУ ee NE a UN ee er Re CN om m е 94 AA EIER пена dees er eL ee А КАФ dud d 95
| а BE. они о ооа В а у E 95 есу ере е ен ИНО 96
| PS LODO ПАВЕ Оа cR I MUR utum Tc uuu c t ay RE 96 PATON ОЛЕ А ON A aa E Кол, Жек T A MEE t t Lc I а ur у ет 97
3 AI ee en ED CET D I D UU MM tee cd more A ar 97 | AXINOSDENMAEBONENON ENE АЯ SE Ea Ку eL MEE УЕЛ S ЕГ ea MA Кук ME MUN ТА SA 97 ОПСТАО DO МАЊА e NE NL СЕН АА E RT EU Е ОКА, жо СС OP AO 98
| (БОРИСА е EEMO T Oe IX MM TOM NE M Ice Keen ctm ДРЕ ТС а CUM 98 \ (КИЛЕШӘ ЕШ nove UR rM SI РЕА Se NE ou И Р ON PULLUM пе се ар 98 | ЖОЮП е ИОН ENERO Vitex е NO ЕНУ куз ORUM SC DU. ue cM EU ICI TERES DUET Ld. ORA 99 DULOCALPUS ROME ILO Va er eee ed АИ ИО MM MUSEI MEE PC ME eI UR PP аа 99
| BEUNO CERU EORNA E E И Cc TO: OS ON МОР АТИ Mu M ЕЕС КИТЕ В AUS СР аре: AN 99 | LTD TALL CODOS E EU ЕТ ТИЛЕ Е A ER н оа ем ад 100 EFA SATILES Bele DOV ее РЕНЕ Са Н, IPM С e A ee У А a E E 100
| ФгоБийсавртринвейе О ИИ ТОИ NR АБА OM PEU M EAE Iti PM 101 ПЕХАР РО НО A m T Л НИСЕ НИ E NI UM NUMINUM e eee 101 JOCJOHGSIQCECBUTIOVIR NE eee BS НИНИ REESE LUE NEMUS E ON OAM МОЛТ ee EE 102 LighicanpuSpenenov dme qe LR ET Ee ete SS PIDEN А аа. 102 LiehigioD аа MON NS NI CHRIS PUT сае RE MET СИ EU ELEME UE MEUSE ICM 102
VEGA HON (helo ON ON DONA EE au ase SHENG Т TU e MUN E Rm M CI RN MEE I UM NEC E 103
| ОЛОР И ӨЛӨ ОИ A за ЭШИ ОНЫ NP IS ERD c ЖЕЛ О А ee КЕЧ 103 | INEDAFOSCMPN genova ee НА И 103 ОПУВИТРУШТИ ВОН ИО УЛ РАНИ N O A S 104
| PAStePHAGKTGISEN ENO Ve cU N AR TEN A ED MOX IEEE реа Пр маже DICE 104 РИО БЕЛ BEN PHONE ОН Ыы win КЛИШЕ EON Е cite. A eee cM СЕЕ. US A А MS 104
| Раббы ох н а Уа а а а А И cO ОРЕГОН none mn 105 | POUOSTOSDENINA:DENEN ONE АНЕ E ET КК А odes ep NE Feder eed ОКНО ЮКО A 105 | POL) STANDISH ONE ER трона TN TE А Em Пи AAA ANOS NEE 106 i Erun DO POMELO TET Ку Vd e E A. TA O A ж AR ER 106 ТЕГОВЕ CEN NOVERIT rM, AOS МЫ ТЬ AAA ERA ER cT TAA 106
| ШИРЕ И РОВ пот Pee ee ee A LOST Пи au ШЕЕ 107 Уа ДЕА епа торта ИТЕ LOUP AE Aceh И Н О ИТО E O PA О E O а. 107
| ОПИСИ ОЛОХ Е МАТИ RE Виан ара cU earl ААИ Ра а E Ба 107 : АОИ Чат ато Жы II К Om Cok ia все RRM т TEN 108 | SS ЕПОРАССИ ТРИ И Сево У AL О И qoe mox ios NIMES 108 DERLATIENGEIEBENE ONE ee а steilen 109 СПРИТНО ООУ N КОС атта oe MUR 109
MS DUGETOSPEHING Belle NOV EA A A O E r 110
| ISDREROSDELINGEBENNOV a УРААК УУ TIE D V И ee le ee а ПАДА НЕ I EN 110 | Saena ananena e А o uS oe een. ee era ee 111 реА ОУ A АДЫР S AE Udo рз жөө Sae E аб КЫ шыбы рыб ыан ааа cus ли uisa RN 111 ПОПИО ве ОМА ch tak teeta LC rer ЛЕКТЕ ка re i qM M onn 112
Ши тесари реши ON АЛ И ЛУ е ери те RE ЛЕ УУУ ЫКЫ EAS ANCA CUM LUE PORA AAT DET T e Сулк PT RN 112 LEONOV LE POMC ОМ АУЕ УГЕ Т ЫЕ ТУУЛ УЛТ УТ ШУ УУУ ЫТ лел пра а ТУМО NEIL ИЕ Co BR PEE 112
| IDAS TC MSE CIATION AME ENTER, eRe e ЫЛ ER DL. AS, EOS BERO IE wey N тоа A 113 LTO OSCOPNG ДӨШҮ A Е ES ue OO Ioco ee ee A rn BR 113
| TES EMITIR TE ES DI OS ORTOS TOTO CLIQ л E E ВИИИ ао RT пе MN 114 TA NADIE MAME M M D M Me aue QN i и 114 ТШ AA О УО PIA T S UE E NNI CE e AS 115
СПОРЕ ТИТ ВОНО OR ar OI Ууу QNO AL, МЕ MRAM UL Т LEER ERN RH 115
REAT TNE NONN TE tre AM Fie ОКК S EMI ct EDO EN, EN NER adi 116
| (РОДИ И ЛЕЛЕРИ ОНЕШ ШЕР А СИ ое iiu. Beate learn, 116 (Селма HAAS oF у EROR Жил Аск Ко с а Ne АЫ е ELI NUR M о ECL. 117
TAAA ЕР c лла ы сие T О РМ Ж AN Ci еа M RE UE ta кх ed 117
Carpolithus sp. Carpolithus sp. Carpolithus sp. Carpolithus sp. Carpolithus sp. Carpolithus sp. Carpolithus sp. Carpolithus sp. Carpolithus sp. Carpolithus sp. Carpolithus sp. Carpolithus sp. Carpolithus sp. Carpolithus sp. Carpolithus sp. Carpolithus sp. Carpolithus sp. Carpolithus sp. Carpolithus sp.
References Cited
Index
Text-figure . Geologic map and index map showing location of the Nut Вейѕ.............................. n nn
LIST OF ILLUSTRATIONS
ЕК E E E Не оно ШУ se eus er ee Моо О ШО ОО uae л et ци опело ae Эз UO EIC UM cM IC MIR Eq ЕЕ ОВОО ше чк к y THE АН ек AUN TR cR hog DA CE Gee OO E A ы А c Taxonomie status об Claro fruit and всей репета esi cov e b rrr retrum о иска ааа d rg t rn neue toe tes = Nut Beds genera ти не Eocene об Богоро oe aye che стани td Ag peur en usua . Present geographic distribution of extant Nut Beds genera ......................... emen Es cH IP Hd LC c E M LAU ДИ е SS qr d Ne eh ATE IP бнр Tte sud. И ОА НА te Lo I M аа ен cs eger E а Т cue uL E BANCOS OO C D T RAS eth UU A UTEM C M T
SE ID DM C dM URL Mo Vm DIL M A ОО ОО ОО A Mo EMI 5 POR ИРИ LL ОЛО Е IT MS Eod serve iue огл edu cilc Dua лы
N ee ORO ON QS ONE E I ERN an TIN UU ЕРА = UT ECCO bocce tn INN cO eL RM E T . Emmenopterys .
TWO TU cc e TER IR UO e LU WP CIS er E
Table
1 2 3. 4 а
= ООО о Acer and: Deviacer samata oer coir. БИ Rene кома е вота Oh Roi Mew ин б ee э жүзө жыды” . Ferrignocarpus . . Fimbrialata ... . Pteronepelys ... . Quintacava .... . Scaphicarpium . . Sphenosperma . . Triplascapha...
LIST OF TABLES
. Comprehensive list of fruit and seed taxa from the Clarno Nut Beds assemblage indicating number of specimens observed ..... . Present habitat and vegetational distribution of Nut Beds genera ............................................... шы
Tasa ЭНСЕП WILD European ОСО MOTAS аи а И Y ee e V Qe De Y E eee niu 5 | ка ае
. Present geographic distribution of extant genera represented in the Nut Beds Нога... ... 0.0.50 e seen 6 . Systematic list of Nut Beds genera with known modern familial affinities . . ...... eee nn
FRUITS AND SEEDS OF THE MIDDLE EOCENE NUT BEDS FLORA, CLARNO FORMATION, OREGON
\ | STEVEN К. MANCHESTER
Florida Museum of Natural History | University of Florida | Gainesville, FL 32611
ABSTRACT
Permineralized fruits and seeds from the Nut Beds flora of the Clarno Formation, north-central Oregon, are described and \ illustrated, providing the first comprehensive treatment of a North American fossil fruit and seed assemblage and facilitating | comparison with European Tertiary fruit and seed assemblages. Newly obtained radiometric dates, along with vertebrate cor- | relations, indicate а Middle Eocene age of about 44 million years for the assemblage. Based upon examination of approximately | 20,000 specimens, 145 genera and 173 species аге recognized, providing much new data for systematic, evolutionary, biogeo- graphic and paleoecologic interpretation. Seventy-five genera and 102 species have been identified to 35 living families; 70 genera | and 71 species represent form genera of unknown familial affinity and may belong both to extinct genera and to extant genera whose affinities may come to light with future work. Among the 102 species attributed to extant families, 58% belong to extant genera, 30% are extinct genera, and 22% are | stereotype genera (identical in the characters preserved to more than one modern genus). Trees апа lianas are well represented, nearly to the exclusion of herbs. The only herbaceous angiosperm identified is Ensete (Musaceae). The proportion of the identified taxa that represent lianas or climbers is high (43% of the 69 species for which growth form may be inferred) including 14 species of Menispermaceae, seven of Icacinaceae and six of Vitaceae. The disseminules range in size from about 0.5 mm to 85 mm; about 15 are winged and suitable for wind-dispersal, but most are medium to large in size and represent drupes and berries adaptive for biotic dispersal. | Among Ше extant genera determined, some are exclusively temperate in distribution today (e.g., Taxus, Emmenopterys, | Parthenocissus), while others are exclusively paratropical to tropical (e.g., Ensete, Mastixia, Iodes, Ampelocissus, and, among the vertebrate fossils, a crocodilian). Many of the genera are distributed today in both temperate and tropical areas. Despite the | оссштепсе of many taxa that are temperate today, the presence of Musaceae, Palmae and other taxa that typically are frost- intolerant today, combined with the high diversity of lianas, is taken as an indication that the climate was comparable to that | which supports paratropical rainforest vegetation today. The flora shares 24% of its genera with the Early to Middle Eocene flora of western Europe, indicating that one or more land connections were viable during or prior to the Middle Eocene. Among extant genera represented in the Nut Beds assemblage, the greatest similarity is with the extant flora of eastern Asia. The identified families are, among the conifers: Taxaceae (3 genera), Pinaceae; among the angiosperms: Actinidiaceae, Alangiaceae, Anacar- diaceae (2 genera), Annonaceae, Araliaceae, Betulaceae (2 genera), Cornaceae (6 genera), Fagaceae (2 genera), Flacourtiaceae, | Hamamelidaceae, Hydrangeaceae, Icacinaceae, Juglandaceae (4 genera), Lauraceae (3 or more genera), Leguminosae, Lythraceae, Magnoliaceae, Menispermaceae, (13 genera), Musaceae, Palmae, Platanaceae (3 genera), Rosaceae, Rubiaceae, Sabiaceae (2 | genera), Sapindaceae (2 genera), Sapotaceae, Schisandraceae, Staphyleaceae, Symplocaceae, Theaceae, Ulmaceae (4 genera), and Vitaceae (4 genera).
ACKNOWLEDGEMENTS Museum of Science and Industry, including John M. | Armentrout, Bruce Hansen, Michael С. Houck, апа | This work is dedicated to the memory of Thomas Joseph Jones, III, and members of the Oregon Agate
J. Bones (Text-fig. 1), who collected the majority of Specimens upon which this investigation is based and whose enthusiasm helped to kindle my own interest in paleobotany as a high school student. Many others assisted me in collecting at the locality as participants in paleobotanical programs sponsored by the Oregon Museum of Science and Industry during the summers of 1972 to 1989, including Alexander Atkins, Scott Blanchard, Kris Goertz, Тап Gordon, С. Bruce Hanson, Carol Hardman, Elizabeth Harding, Kathy Harvey, Joyce Lenz, Maureen Muldoon, Duane Olson, Jerome McFadden, Ellen Pasternack, John Ries, J osephine Spitzer, Michele Vowell, Eric and Mike Weinstein and Others. I thank former and present staff of the Oregon
and Mineral Society, including Gar Hurley, William and Gertrude Hall, for their support of student field programs in Clarno paleobotany.
Special thanks are due to Carrie L. Roose, for in- valuable help in sorting, measuring and cataloguing specimens, assistance with initial taxonomic analyses and in preparing the geologic map and preliminary diagrams. C. Bruce Hanson provided help in inter- preting the local geology. Wendy Zomlefer assisted with preparation of the text-figure illustrations. David Dilcher provided access to equipment and facilities as well as helpful encouragement throughout this re- search. N. Gary Lane kindly provided access to a Mi- croslice II annular diamond saw for sectioning the seeds.
8 PALAEONTOGRAPHICA AMERICANA, NUMBER 58
Text-figure 1.— Thomas J. Bones, photographed in 1977 at an excavation on the top of Face 3. Mr. Bones collected here for about 40 years, and provided the massive collections upon which much of this monograph is based.
Advice in the preparation of latex and silicone casts was given by William Simpson and Russell McCarty. Help with identifications was provided by Margaret Collinson, Robert Kaul, Peter Endress, Paul Grote, Dieter Mai, Charles N. Miller, Richard A. Scott, and Bruce Н. Tiffney. Grammatical advice on the construc- tion of new Latin binomials was provided by Benoit LePage and Christine Kampny.
Access to paleobotanical collections was provided by Scott Wing, Francis Hueber, and James P. Ferrigno (U.S. National Museum, Washington, D.C.); Howard E. Schorn (Museum of Paleontology, University of California, Berkeley), Wesley Wehr (Burke Memorial Washington State Museum, University of Washington, Seattle), Charles Beck and Robyn Burnham (Museum of Paleontology, University of Michigan), Andrew Knoll (Harvard University), Leo Hickey (Peabody Museum of Natural History, Yale University, New Haven), Chris Hill and Cedric Shute (Natural History Museum, London, England), Christiane Blanc (Mu- séum National d’Histoire Naturelle, Paris, France), Friedemann Schaarschmidt (Senckenberg Museum,
Frankfurt, Germany). Access to extant fruit and seed collections was provided by Charles Gunn (National Seed Herbarium), Peter Stevens (Harvard University Herbaria), Norris Williams and Kent Perkins (Uni- versity of Florida Herbarium). Helpful comments on the manuscript were provided by David Dilcher and Herbert Meyer. I also want to thank the reviewers, Bruce Н. Tiffney and Else Marie Friis, for invaluable comments and suggestions.
This research was supported in part by the following National Science Foundation grants: DEB 81-11-89, EAR 8707523, EAR 8904234, BSR 9007495, EAR 9322765. This publication represents work that was initiated with the facilities and support of Indiana Uni- versity and completed at the University of Florida. Publication costs were met in part by funds from the Gatorade Account of the University of Florida. This publication is number 410 in the University of Florida Contributions to Paleobiology.
INTRODUCTION
A diverse assemblage of well-preserved Middle Eo- cene plant remains occurs in the type area of the Clarno Formation about 3 km east of the community of Clar- no, north-central Oregon, in a locality known as the Nut Beds. The site yields abundant fossil woods (Scott and Wheeler, 1982), leaves (Manchester, 1981), and exquisitely preserved fruits and seeds (Scott, 1954; Bones, 1979). Because fossil fruit and seed floras are relatively rare in North America, the Clarno Nut Beds flora provides an important link for paleobotanical comparisons with the Tertiary of Europe, where fruit and seed assemblages are abundant and relatively well studied. The Clarno flora often has been cited in phy- togeographic discussions because taxonomic similarity with European fruit and seed floras provides evidence of broad floral continuity between Europe and North America during the Eocene (Scott, 1954; Chandler, 1964; Wolfe, 1972, 1975; Tiffney, 1985a, 1985b; Col- linson, 1988; Manchester, 1988). Only a small portion of the Clarno flora has been described, however, and the actual extent of generic and specific similarity be- tween the Eocene floras of North America and Europe has not been well documented.
The paleobotanical and phytogeographic signifi- cance of the Clarno flora was first shown by Scott (1954) in a monograph detailing the morphology and rela- tionships of ten species from the Nut Beds locality. Scott’s work demonstrated the presence of genera with Old World tropical affinities in the Eocene of western North America, and documented several genera in common with the Eocene London Clay flora of En- gland. He also called attention to the presence of both extant and extinct genera, a finding somewhat at odds with Tertiary leaf studies of the time. Subsequently,
CLARNO (EOCENE) FRUITS AND SEEDS: MANCHESTER 9
an expanded taxonomic list of Clarno fruits and seeds was published (Scott, in Chandler, 1964, p. 58) in- cluding 32 additional genera from the locality. Until now, these additional taxa have not been formally de- scribed. Bones (1979) published an atlas depicting some of the diversity present in the Nut Beds fruit and seed assemblage, including a variety ofnew but undescribed species. Several additional fruit and seed taxa from the Nut Beds have been treated in separate accounts (Man- chester, 1986, 1987a, 1987b, 1988, 1989a; Manchester and Kress, 1993), but a comprehensive treatment of the flora has not been attempted previously. The Nut Beds fossils represent one of the warmest climatic in- tervals in the Tertiary of North America, a time when thermophilic vegetation was spread well into the north- ern latitudes (Wolfe and Poore, 1982; Wolfe, 1985).
Because of the excellent details of morphology and anatomy preserved in silicified fruits and seeds from the Nut Beds flora, specimens from this assemblage provide the rare opportunity to investigate the internal structure of Eocene plant reproductive organs. Such specimens enable a well-informed assessment of tax- onomic affinities, and make it possible to test whether affinities inferred from external morphology are upheld by internal anatomy. The Nut Beds specimens thus provide insight into the structure and affinities of var- ious extinct and extant genera of fruits and seeds that are known only from molds or impressions at other fossil localities.
This monograph is an attempt to provide a thorough taxonomic treatment of all known fruit and seed spe- cies from the Nut Beds flora, including taxa of unde- termined familial affinity as well as those which can be identified confidently to modern families and gen- era. The resulting data base: 1) provides new systematic data on plants present during the Middle Eocene in western North America; 2) contributes information on the evolutionary status of various gymnosperm and angiosperm taxa in the Middle Eocene of the northern hemisphere; 3) provides insight into the environment and climate under which the Nut Beds vertebrate fauna existed; 4) enables comparison with other Eocene fossil floras of the northern hemisphere and to modern flo- ristic regions to gain some understanding of timing and routes of phytogeographic exchange.
GEOGRAPHIC AND GEOLOGIC SETTING
The Clarno Formation is a terrestrial sequence of andesitic to basaltic lavas and intrusives, ash flows, volcanic mudflows, and tuffaceous sediments exposed along the Blue Mountain Anticline in north-central Oregon. The distribution, stratigraphy, age, structure, composition and source vents of the formation are reviewed by Walker and Robinson (1990). Petrologic
and trace element studies suggest that the volcanism may have been in response to subduction zone mag- matism (Rogers and Novistsky-Evans, 1977; Noblett, 1981). The formation is well known for its paleonto- logic importance and is the focus of the Clarno Unit of John Day Fossil Beds National Monument. Leaf impressions and silicified wood occur in tuffaceous strata in many areas of the Clarno Formation (Hergert, 1961; Manchester, 1986, 1991). Fruits and seeds, which occasionally occur in the same strata as leaves, are usually preserved as impressions or compressions. However, the Nut Beds locality is unusual because it provides uncompressed, three-dimensionally pre- served, permineralized, fruits and seeds.
The Nut Beds locality (UF loc. 225) is situated in the type area of the Clarno Formation, at 44°56'36” N lat., 120?25'34" W long. (SW 4, SE Ya, sec. 27, T 7 S, R 19 E), Wheeler County, Oregon (Text-fig. 2). It is a resistant, buff-colored, cliff-forming unit about 10 me- ters thick, exposed over an area of about one hectare along the western side of Hancock Canyon (Text-fig. 3). The term “nut beds" was first applied by amateur fossil collectors in the 1940's because of the abundance of petrified walnuts and other fruits and seeds found there. Fossil plants occur throughout the vertical and horizontal extent of the exposure. The same locality is also significant for its vertebrate fauna, which consti- tutes the only known Middle Eocene terrestrial mam- malian fauna in the Pacific Northwest (Stirton, 1944; Hanson, 1973, p. 56 in Retallack, 1991).
The complex geology of the local area surrounding the Middle Eocene Nut Beds locality, and stratigraphic relations with the Late Eocene/Early Oligocene Han- cock Quarry vertebrate locality 1 km to the north (Text- fig. 3), are currently under investigation (C. B. Hanson, pers. comm.; G. Retallack, pers. comm.). Hanson (pers. comm., 1992) distinguishes five unconformity-bound- ed units in the type area of the Clarno Formation. The Nut Beds deposit occurs low in the section, within the second unit, and predates an adjacent andesitic plug! which was a source for clasts found in subsequently deposited sediments of the Clarno Formation in its type area. The Hancock Quarry fauna (Mellet, 1969; Hanson, 1973, 1991) and flora (McKee, 1970) locality is situated in Hanson’s uppermost unit, an upper Clar- no valley-fill deposit containing clasts from the an- desitic plug. Retallack (pers. comm., 1993), however, reports finding clasts of the andesitic plug! in the Nut Beds sediment and considers the Nut Beds to correlate with the upper part of a sequence of lahars and tuffs exposed | to 2 km north and east in Hancock Canyon.
The Nut Beds deposit is composed of tuffaceous silt-
! New information available; see Notes added in proof, p. 200.
PALAEONTOGRAPHICA AMERICANA, NUMBER 58
IN
SR A SS RS SA ANTÉLOPE 2 AR А oy 4
GSSIL
USN — ||| S \ A QUIET NA N A e i l D £ s А | cy po As | қ | ASHWOOD 47, 2» Хам | : 9. ? E TWICKENHA | ‘ip, {ч M UIT Q 5. wy сас | ) MUT | S N SE ern зе = | pto === | Y) хе S4 m. im) $ wr, — MITCHELL | (n 10 BETTER ai NN | V Se | я Ata euer | N тс AT i | Y [ RU , N | Plio/Pleistocene SS ||, П сее |) В |) 3 | S SE N abel ~ ` ormation AN а жале Еосепе С!агпо 9POST Е N RR Formation Ve N SEE Cretaceous marine | = f Sun ] sec ка SS vi 5 10 15 Mi у E 10 15 20 25Кт
Text-figure 2.— Geologic map and index map (inset) of the John Day and Crooked River Basins, north-central Oregon, showing location of the Nut Beds (arrow) and other paleobotanical localities of the Clarno Formation. Localities: 1. Nut Beds, 2. Hancock Quarry, 3. Dry Hollow, 4. Left Hand Canyon, 5. Horseheaven, 6. White Cliffs, 7. Indian Rocks (classic Cherry Creek site), 8. Red Gap, 9. John Day Gulch, 10. Ochocco Summit, 11. Doolittle Flat, 12. Alex Canyon, 13. Classic West Branch Creek locality, 14. Brummers Spring, 15. Sheep Rock Creek (Teater Road site). Geologic map adapted from Walker (1977).
CLARNO (EOCENE) FRUITS AND SEEDS: MANCHESTER 11
i s Se
Text-figure 3. — Oblique, northward, aerial view overlooking the Middle Eocene Nut Beds locality, with exposure faces labeled 1 to 5. The
SEN } =
dark unvegetated strata overlying the Nut Beds are a sequence of paleosols locally called Red Hill (RH) that is capped by the basal ignimbrite of the John Day Formation. Other localities in the distance include the Upper Eocene Hancock Quarry (HQ) in the upper part of the Clarno Formation and the Oligocene Dugout Gulch leaf locality (DG) in the lower part of the John Day Formation. Iron Mountain, capped by
Miocene Columbia River Basalts, forms the horizon.
stones, sandstones and conglomerates thoroughly ce- mented by silica and calcium carbonate. The strati- graphic column at the Nut Beds is about ten meters thick, consisting of a lower half with alternating beds of siltstone and sandstone mostly 0.1-0.6 m in thick- ness, and an upper half with thicker beds of alternating siltstone, sandstone and andesitic conglomerate (Text- fig. 4). Clasts of the conglomerate are subangular to rounded, generally less than 4 cm in diameter, and are matrix-supported. Permineralized wood, fruits and seeds occur throughout the vertical section, in siltstone, sandstone and conglomerate layers, with the greatest concentrations being in the uppermost strata. The ver- tebrate fossils are confined to the middle and upper strata, often in conglomerate, and are usually preserved as disarticulated bones. Impressions of leaves occur at various levels within the deposit, but are most abun- dant and well-preserved in siltstones near the base of the section (Manchester, 1981).
Abundant Equisetum stems and occasional fern rhi- zomes are found in growth position in some of the siltstone layers. In some parts of the deposit Equisetum stems may be traced vertically through 0.7 m of sed- iment (Retallack, 1981), indicating relatively rapid sediment accumulation (Scott, 1954) and shallow-wa- ter conditions. Predepositional abrasion is evident in some of the fossils (particularly fossil. woods in the upper part of the section) and suggests that the assem- blage includes transported as well as locally derived plant debris. Although well-preserved silicified wood is common, most pieces are small, predepositionally worn fragments; no standing stumps were confirmed.
The Nut Beds deposit is dissected by slumps and erosion into five outcrop faces, which, for reference, have been designated from south to north as Faces 1- 5 (Text-fig. 3). Most of the vertebrate remains were recovered from the middle and upper portions of Faces 1 and 3. The fossil fruits and seeds described by Scott
I2 PALAEONTOGRAPHICA AMERICANA, NUMBER 58
Text-figure 4.— Profile view of the stratigraphic section at Face 3, showing alternation of beds (siltstone and sandstone/conglomerate). The lowest exposed layers contain leaves as well as some of the fruits. The most productive horizon for silicified wood and fruits was found to be in the massive upper layers.
(1954) and the Hyrachyus tooth described by Stirton (1944) were obtained from two beds situated in the gully between Faces 2 and 3. Thomas Bones collected fruits and seeds throughout the deposit, but found the richest horizons in excavations at the top of Face 3 (Text-fig. 1). Beds forming the uppermost meter of section on Face 3 contain an abundance of palm pet- iolar debris, abraded wood fragments, and small rounded fossiliferous chert clasts. The chert clasts, mostly 0.5-1.2 cm in diameter, may represent frag- ments of peat that were ripped up and carried along with the other clasts, or could alternatively represent small coprolites (C. B. Hanson, pers. comm.). Leaf remains in this upper horizon are usually folded or rolled, not flat-lying, and apparently were deposited during turbulent, rather than quiet-water conditions. The sedimentary environment represented by the Nut Beds deposit has been interpreted as that of a lake delta deposit (Scott, 1954; C. B. Hanson, pers. comm.). This interpretation would account for the large amount of jumbled biotic debris, and the general thinning of the beds toward the northeast (C. B. Hanson, pers. comm.). However, typical lacustrine sediments, with fissile shales containing fish remains, as observed at other Clarno paleobotanical localities (e.g., West Branch Creek, White Cliffs; Manchester, 1990), are not present at the Nut Beds. The lower portion of the section, with
alternating siltstones and sandstones, might represent periodic flood deposition, whereas some of the thick, poorly sorted upper layers evidently represent a rapidly deposited slurry, possibly from the toe of a volcanic mudflow. Local hot springs may have been present, contributing toward rapid mineralization of biotic de- bris.
AGE
The value of the Nut Beds flora for biostratigraphic and paleobiogeographic correlations is enhanced by an accurate determination of its age. In previous paleo- botanical assessments, age interpretations for the Clar- no Nut Beds have ranged from “older than Upper Eocene" (Scott, 1954) to Early Oligocene (Wolfe, 1971, 1981). Although Wolfe (1981) attributed the Nut Beds (“late Clarno") flora to his Kummerian floral stage, of inferred Late Eocene/Early Oligocene age, his zonation is based on leaf species from a type section in the Puget Group of northwestern Washington (Wolfe, 1968). It is impossible to determine which of the Clarno fruit and seed species might correspond to the leaf species that Wolfe used to characterize his stages, and prelim- inary investigations of the Nut Beds leaf assemblage (Manchester, 1981) indicate that the flora is no more similar taxonomically to the Kummerian stage than to the older Franklinian and Ravenian stages. In any case, to eliminate potential for circularity, the age of the Nut Beds flora is best determined by means independent of paleobotanical correlation.
Radiometric dates reported for various parts of the Clarno Formation range from about 54 to 34 million years (Fiebelkorn et al., 1983), although the younger dates are questionable because the basal ignimbrite of the overlying John Day Formation has been dated at about 37 (Swanson and Robinson, 1968; Fiebelkorn etal., 1983; Vance, 1988) to 39.7 (Bestland et al., 1993) million years. The Clarno Formation represents a com- plex volcanic terrain characterized by abrupt lateral and vertical variation complicated by erosion, faulting and slumping. Therefore, it is difficult to trace indi- vidual units beyond local areas, and, as a result, the stratigraphic position of the Nut Beds in relation to other datable units of the Clarno Formation is not immediately obvious. Fortunately, the Nut Beds de- posit itself contains a vertebrate fauna that can be cor- related with North American land mammal stages, as well as volcanic clasts suitable for radiometric dating.
Stirton (1944) described a tooth of Hyrachyus from the Nut Beds, and on this basis clearly established an Eocene age for the locality. More complete cranial specimens of Hyrachyus, along with remains of five other mammalian vertebrates, including Orohippus, Patriofelis and Telmatherium, have been investigated
| | | |
| | |
CLARNO (EOCENE) FRUITS AND SEEDS: MANCHESTER 13
by Hanson (1973) and clearly represent a Bridgerian fauna (Hanson, pers. comm., 1988). This fauna occurs within the Nut Beds along with the fruits and seeds, and should not be confused (cf., Wolfe, 1981, p. 44) with the younger Clarno fauna from Hancock Quarry, 1 km northeast of the Nut Beds, that is of Chadronian or Duchesnian age (Mellet, 1969; Hanson, 1973, pers. comm., 1988). By correlation with radiometrically dated faunas of the Rocky Mountain region, the Bridg- erian vertebrate fauna of the Nut Beds indicates a Mid- dle Eocene age of about 45 m.y.
Although a potassium-argon date of 34 million years was published for the Nut Beds (Evernden et al., 1964), such a young age (Late Eocene or Early Oligocene) is very doubtful because the basal ignimbrite of the over- lying John Day Formation (situated about 30 meters Stratigraphically above the Nut Beds) yields radio- metric dates of about 37 (potassium-argon, Swanson and Robinson, 1968; fission track, Vance, 1988) to 39.7 (argon-argon, Bestland et al., 1993) million years. More recently, Vance (1988) obtained fission track dates of 43.6 and 43.7 (+ 10%) m.y. based upon zircon extracted from Nut Beds sediment. In 1987, C. B. Han- son, B. D. Turrin and I collected a sample of water- reworked tuff with intact pumice clasts from a plant and bone-bearing stratum at the middle stratigraphic level of Face 1 in the Nut Beds. Ten plagioclase crystals isolated from this sample were analyzed by the single crystal laser fusion argon 40/39 technique and yielded dates ranging from 36.38 + 1.31 to 46.8 + 3.36 my. (B. D. Turrin, pers. comm., 1988). Statistically, these dates give an arithmetic mean of 42.98 + 3.48 witha Standard error of the mean, + 1.10, and a weighted average of 43.76 + 0.29 m.y., strikingly close to the fission track results of Vance (1988). Based upon the recently obtained radiometric ages and the vertebrate Correlations, the age of the Nut Beds flora may be considered to be Middle Eocene, within the range of 43 to 45 million years old.
OTHER CLARNO FLORAS
Although this report is intended to be as compre- hensive as possible for the fruit and seed assemblage of the Nut Beds locality, it should not be regarded as an exhaustive treatment of the Clarno flora. The Clar- по Formation occurs over an area of about 4000 km2 (Walker and Robinson, 1990) and includes many fossil Plant localities (Text-fig. 2). The location and general Character of some of these localities are reviewed by Hergert (1961) and Manchester (1990). Geographic co- Ordinates for localities from which large collections have been made have been published previously (Man- Chester, 1986, p. 224; 1991, p. 717). Some of these Sites, particularly the lacustrine deposits in the West
Branch Creek and Cherry Creek drainages, include many species of leaves and compressed fruits that are not known from the Nut Beds. I have made reference to these other localities where appropriate in discussing the distribution of individual species.
Another occurrence of fossil fruits and seeds in the Clarno Formation is the Late Eocene (Duchesnean) Hancock Quarry vertebrate locality, situated 1 km northeast of the Nut Beds. This locality (UF 70) has provided abundant fruit and seed remains preserved as partially compressed carbonaceous casts and molds in claystone (McKee, 1970). Although anatomical de- tails are not preserved, external morphology enables identification through comparison with the better pre- served Nut Beds specimens. More than 500 specimens were collected from this site (collections at OMSI and UF) and is possible to distinguish about 30 taxa. Sev- eral genera are shared with the Nut Beds, including Ampelocissus, Diploclisia, Juglans, Odontocaryoidea, Palaeophytocrene, and Vitis. In addition to the genera recognized by McKee, I have observed specimens of Alangium, Eohypserpa, Iodes, Iodicarpa, Mastixioi- diocarpum and Pentoperculum in the Hancock Quarry assemblage. Thus, it is clear that some of the Nut Beds genera persisted in the same region at least until the Late Eocene. There are, however, marked floristic dif- ferences between both of these assemblages and the Lower Oligocene Bridge Creek flora of the overlying John Day Formation (Chaney, 1927; Manchester and Meyer, 1987; Manchester, 1990).
The Brummers Spring locality (UF 254), also in the Clarno Formation, is located about 90 km south of the Nut Beds locality and 20 km east of Post, Oregon (Text-fig. 2). This locality has produced a small assem- blage of silicified fruits and seeds including Juglans, Magnolia, Mastixicarpum, Quercus, and Sabal.
MODES OF PRESERVATION
Fossil fruits and seeds in the Nut Beds are preserved in various modes, ranging from compressions and im- pressions to molds and casts, to permineralizations. These different. types of preservation, as applied to fossil plant tissues, are reviewed by Scott and Collinson (1983, p. 118). Compressed fruits and seeds, preserved along with leaf impressions near the base of the section (Manchester, 1981), usually lack internal structural de- tails, but are useful in providing characters of wing morphology and venation. Three-dimensionally pre- served specimens, including casts, molds and permi- neralizations, occur throughout the deposit.
Casts and molds lacking internal anatomy are com- mon (Scott, 1954), but sometimes the internal tissues are preserved through permineralization by chalced- ony and/or calcite. A single specimen may include more
14 PALAEONTOGRAPHICA AMERICANA, NUMBER 58
than one mode of preservation; for example, a cast of the locule may occur within the carbonized or per- mineralized remains of an endocarp. Permineralized seed remains are sometimes encountered within the locule cast.
Silicification appears to have been primary, and cal- cification secondary, as in the case of walnuts that are sometimes preserved with the locule filled by chalced- ony and the nutshell infiltrated by calcite. The quality of preservation in permineralized specimens is vari- able, depending on porosity and durability of the tis- sues involved. Fleshy or pulpy outer fruit layers are rarely if ever preserved. Woody layers tend to be well- preserved, sometimes with excellent cellular detail. However, thin, hard, evidently impermeable layers typically are represented only as an airspace between the chalcedony internal mold and the siltstone outer mold of the structure, as in the case of seeds of An- nonaspermum, Fortunearites, and Bumelia.
MATERIALS AND METHODS
Approximately 20,000 specimens were examined for this study. The majority of specimens were collected between 1942 and 1983 through the extraordinary ef- forts of Thomas J. Bones. Additional specimens were collected in the 1940’s and 50’s by A. W. Hancock, R. A. Scott, and, during the summers of 1974-1989, by S. R. Manchester with the help of many high school students (acknowledged p. 7) in conjunction with pa- leobotanical programs at Hancock Field Station, a nat- ural science facility operated by the Oregon Museum of Science and Industry 1.5 km southeast of the Nut Beds.
The fossil disseminules range from less than 1 mm to 85 mm in length. Because the sediment is silica- cemented, standard washing and sieving techniques (Tiffney, 1990) are not effective in recovering speci- mens. Instead, the fossils are obtained through a labor- intensive process of prying out blocks of sediment and breaking the rock with hammers. Fruits and seeds are exposed in the matrix when intercepted by the resulting fracture planes. The percussion frequently causes the fruit or seed to pop out of the matrix. Sometimes outer layers of the fossil remain attached to the matrix mold. It is thus important to retain the associated mold for study as well as the more attractive cast. Thomas Bones collected large numbers of small seeds by crushing the matrix and sorting with the aid of a magnifying lens.
Because the type and quality of preservation varies greatly, even among specimens of the same species, it was necessary to sort the specimens under a dissecting microscope to recover the most informative samples for describing each species. Silica casts and molds re-
placing various tissues of the fruit or seed were useful for morphology and dimensions. Permineralized spec- imens were most valuable in providing the best details of internal morphology and anatomy. Permineraliza- tions were not available for all species, but in many instances a careful examination of numerous speci- mens resulted in the recovery of at least a few per- mineralized examples that could be fractured or sawn to reveal internal morphology and anatomy. Com- pression and impression specimens were also found to be useful because they enable direct comparison with specimens from other Clarno localities where permi- neralizations are not available, and because they pro- vide details of wing morphology and venation for sam- aras that usually are not evident in the permineralized specimens.
Morphology and anatomy was recorded photograph- ically with Kodak Technical Pan 2415 film processed for medium contrast. Lighting for reflected light pho- tography was provided from the upper left and lower right side of each specimen with a pair of small, diffuse tungsten lamps that could be oriented at various angles to provide optimal illumination of surface detail. Spec- imens of medium to large size (>8 mm diam.) were photographed with a Nikon FE camera with 100 mm Micronikkor lens. Macrography of smaller specimens (0.5-8 mm) was conducted with the FE camera mount- ed on a Nikon SMZ10 zoom stereo scope. Stereo-pair photographs were prepared of specimens which were subsequently fractured or serially sectioned. For ste- reo-pairs, successive pictures were taken with the spec- imen rotated 6 degrees on a teetering stand, or by using the stereo-photo option on the Nikon SMZ10 macro- scope. The translucency and differential coloration of silicified specimens initially presented problems, ob- scuring surface details in photography. These problems were overcome by coating the specimens with palla- dium using a sputter coater prior to light micrography- The palladium was subsequently removed by soaking for 30 minutes in a saturated solution of sodium cy- anide in water (Sela and Boyd, 1977). Small seeds, and minute surface details of larger specimens, were stud- ied by standard techniques of scanning electron mi- croscopy (SEM), using the Cambridge Stereoscan 250 and Hitachi S-450 SEM models.
Internal morphology and anatomy were revealed in some instances by studying fractured specimens with light and SEM microscopy and in other cases by serial sectioning. Specimens 1.5 mm to 25 mm in diameter were sectioned with a Microslice II annular diamond saw, with a very thin blade (0.05 mm thick) to mini- mize wastage (kerf loss about 0.15 mm) in the cutting process. Resulting wafers were photographed unetched in xylene, or dry after etching with hydrofluoric acid,
| |
| | |
——————————
CLARNO (EOCENE) FRUITS AND SEEDS: MANCHESTER 15
by reflected light microscopy using a Wild M400 pho- tomacroscope. For finer anatomical details the wafers were then used to make peels and/or ground thin sec- tions for transmitted light microscopy. Peels were pre- pared by the cellulose acetate peel method (Basinger and Rothwell, 1977; Basinger, 1981) after etching for two minutes in 48% HF. The peels were cleared in clove oil, mounted in Canada Balsam with slide and coverglass. Ground sections were prepared by mount- ing Ше wafers on microscope slides with Elmers Epoxy, and grinding to approximately 30 um thickness on a Buehler thin sectioning machine. Anatomical detail from thin sections was photographed with the Nikon FE camera mounted on a Nikon Labophot compound microscope using standard transmitted and/or reflect- ed light from an obliquely positioned fiber optic source. For one taxon (Hydrangea), small organically pre- served seeds were liberated by dissolving the silicified fruit in hydrofluoric acid, then washed and mounted on microscope slides for transmitted light microscopy.
External measurements were obtained with brass calipers graduated to 0.1 mm. Dimensions were cal- culated on the basis of all available specimens, or at least 25 specimens, of each species. Obviously con- torted or broken specimens were not included in the measurements. Unless otherwise indicated, the longest axis of the fruit or seed was treated as height or length. Width and thickness are the wider and narrower di- mensions, respectively, measured at right angles to the length.
Methods of identifying fossil fruits and seeds are reviewed in detail by Tiffney (1990). Discussions of €xtant and fossil seeds by Reid and Chandler (1933), Chandler (1961b, 1964, 1978), Collinson (1983), Friis (1985), Kirchheimer (1957), and Mai (1976) provided à useful guide for preliminary investigations of the Clarno taxa. Comparisons with extant seeds were car- ried out by examining representatives of most extant angiosperm families in herbaria and seed collections.
RESULTS AND INTERPRETATION
SYSTEMATIC DIVERSITY OF THE NUT BEDS FLORA
One hundred and forty five genera and 173 species Of fruits and seeds have been recognized among the Nut Beds collections (Table 1) and are described and illustrated in this monograph. In addition, among the less well-preserved specimens in the collections, there mày be as many as twenty more genera that are ex- Cluded from the following discussions. Four gymno- Sperm genera, representing Pinaceae and Taxaceae, are recognized, but the majority of taxa are angiosperms.
From the working total of 145 genera and 173 species, only 75 genera and 102 species have been identified as belonging to modern families. The remaining taxa, constituting approximately half of the assemblage, are of uncertain affinity with respect to extant families (Text-fig. 5A). This group probably includes both ex- tant and extinct genera, but the affinities are still un- determined at the time of this writing. Nevertheless, these species are described and illustrated in order to document, as completely as possible, the full diversity of the flora and to invite further study by other re- searchers.
The numbers presented above, based only on the fruits and seeds, understate the full diversity of the Nut Beds flora because many additional taxa are repre- sented by other plant organs. For example, Equisetum clarnoi (J. T. Brown, 1975; Retallack, 1981) is known from silicified stems and compressed strobili (Pl. 1, fig. 1); four types offerns are known from leaves and stems, and a cycad with foliage resembling Dioon is present (Manchester, 1981). Families confirmed on the basis of wood that have not been recognized among the fruits and seeds include Ginkgoaceae (Ginkgo; Scott et al., 1962), Trochodendraceae (Scott and Wheeler, 1982), and Sterculiaceae (Manchester, 1979, 1980). Although the extent of overlap is unknown, it is probable that the fossil leaves (Manchester, 1981) and woods (Scott and Wheeler, 1982) collected from the Nut Beds rep- resent some of the same species as those known from seeds or fruits treated here. However, because none of the organs have been found in direct attachment, as- sessments of conspecificity remain speculative.
It is not possible to give a precise percentage of the Nut Beds genera that are extinct vs. extant because a large proportion of the genera are still of uncertain systematic position relative to modern families and genera. Nevertheless, all genera recognized here may be divided into four groups indicating current knowl- edge of their status relative to living taxa (Text-fig. 5):
1) Extant genera. Those that possess features that are diagnostic of a living genus; these species are given the appropriate modern generic name, for example, Cornus, Magnolia, Meliosma, Quercus.
2) Extinct genera. Those that have features diag- nostic of a particular modern family, but with addi- tional characters that set them apart from living genera of that family, for example, Cruciptera, Langtonia, Pentoperculum.
3) Form genera. Those that have not been traced to a particular modern family (i.e., incertae sedis), for example, Carpolithus, Joejonesia. Form genera prob- ably include extant genera that I have failed to identify, as well as extinct genera and/or families. These genera
16 PALAEONTOGRAPHICA AMERICANA, NUMBER 58
Table 1.—Comprehensive list of fruit and seed taxa from the Clarno Nut Beds assemblage indicating number of specimens observed.
taxon quantity taxon quantity Pinaceae Palaeophytocrene pseudopersica Scott 80 Pinus sp. 1 Pyrenacantha occidentalis sp. nov. 8 Taxaceae Juglandaceae Taxus masonii sp. nov. 20 Juglandeae Torreya clarnensis sp. nov. qs Cruciptera simsonii (Brown) Manchester 12 Diploporus torreyoides gen. et sp. поу. 5 Juglans clarnensis Scott 200+ Actinidiaceae Engelhardieae Actinidia oregonensis sp. nov. 7 cf. Palaeocarya clarnensis Manchester 48+ Alangiaceae Platycaryeae Alangium eydei sp. nov. 33 Paleoplatycarya? hickeyi sp. nov. 5 Alangium rotundicarpum sp. nov. 1 Ташасоле Anacardiaceae Laurocalyx wheelerae sp. nov. 12 Pentoperculum minimus (Reid et Chandler) Laurocarpum hancockii sp. nov. 1 gen. et comb. nov. 62 Laurocarpum nutbedensis sp. nov. 1 Rhus rooseae sp. nov. 15 Laurocarpum raisinoides sp. nov. 1 Аппопасеае Lindera clarnensis sp. nov. 8 Anonaspermum cf. pulchrum Reid et Chandler 16 Leguminosae Anonaspermum bonesii sp. nov. 1 Leguminocarpon sp. 1 Anonaspermum rotundum sp. nov. 1 Lythraceae Araliaceae Decodon sp. 7 Paleopanax oregonensis gen. et 5р. поу. 3 Magnoliaceae Betulaceae Magnolia muldoonae sp. nov. 100+ Kardiasperma рагуит gen. et sp. поу. 21 Magnolia paroblonga sp. nov. 17 Coryloides hancockii gen. et sp. поу. 34 Magnolia tiffneyi sp. nov. 10 Burseraceae Е : Menispermaceae Bursericarpum oregonense Sp. поу. 40 ; Coscineae Bursericarpum sp. 6 : џ Anamirta leiocarpa sp. поу. 16 Cornaceae р ; Menispermeae Cornus clarnensis sp. nov. 5 + de д } a Langtonia bisulcata Reid et Chandler 31 Diploclisia aurif С (Hollick) comb. nov. 8 Mastixia sp. 4 Eohypserpa scottii sp. nov. 2 Mastixioidiocarpum oregonense Scott 22 Davisicarpum limacioides зразок 1 Mastixicarpum occidentale sp. nov. 2 Palaeosinomenium venablesii Chandler 70 Nyssa scottii sp. nov. 6 Tinosporeae Nyssa spatulata (Scott) comb. nov. 29 Atriaecarpum clarnense sp. nov. 4 Nyssa sp. 5 Calycocarpum crassicrustae sp. nov. 40 Fagaceae Chandlera lacunosa Scott 40 Castanopsis crepetii sp. nov. 6 Curvitinospora formanii gen. et sp. nov. 1 Quercus paleocarpa Manchester 3 Odontocaryoidea nodulosa Scott 80+ : Tinospora elongata sp. nov. 7 Flacourtiaceae a Е Е Tinospora hardmanae sp. nov. 4 Saxifragispermum tetragonalis sp. поу. 65 Tinomiscoidea occidentalis sp. nov. 1 Hamamelidaceae Thanikaimonia geniculata gen. et sp. nov. 1 Fortunearites endressii gen. et sp. поу. 6 Musaceae Hydrangeaceae Ensete oregonense Manchester et Kress 70 Hydrangea knowltonii gen. et sp. nov. 45+ Palmae Icacinaceae Sabal bracknellensis (Chandler) Mai 6 Iodeae Sabal jenkinsii (Reid et Chandler) comb. nov. 30 Todes multireticulata Reid et Chandler 14 Platanaceae Iodes chandlerae sp. nov. 13 Macginicarpa glabra Manchester 200+ Iodicarpa ampla gen. et sp. nov. ди Platanus hirticarpa sp. nov. 1 Iodicarpa lenticularis sp. nov. 16+ Tanyoplatanus cranei gen. et sp. поу. 3 Comicilabium atkinsii gen. et sp. nov. 8 Ёозассаё Phytocreneae Prunus weinsteinii sp. nov. 1 Palaeophytocrene hancockii Scott 20 Prunus olsonii sp. nov. 10
| | | | | |
Hexacarpellites hallii gen. et sp. nov. Joejonesia globosa gen. et sp. nov.
Unknown families: 70 genera, 71 species Total: 145 genera, 173 species
CLARNO (EOCENE) FRUITS AND SEEDS: MANCHESTER d Table 1.— Continued. taxon quantity taxon quantity Rubiaceae Lignicarpus crassimuri gen. et sp. nov. 4 Emmenopterys dilcheri sp. nov. 7 Ligniglobus sinuosifibrae gen. et sp. nov. 5 Sabiaceae Lunaticarpa curvistriata gen. et sp. nov. 80 Meliosma beusekomii sp. nov. 110 Microphallus perplexus gen. et sp. nov. 23 Meliosma bonesii sp. nov. 22 Nephrosemen reticulatus gen. et sp. поу. Zi Meliosma elongicarpa sp. nov. 2 Omsicarpium striatum gen. et sp. пом. 1 Meliosma cf. jenkinsii Reid et Chandler 8 Pasternackia pusilla gen. et sp. nov. 220 Meliosma leptocarpa sp. nov. 50+ Pileosperma minutum gen. et sp. nov. 10 Sabia americana sp. nov. 22 Pileosperma ovatum sp. nov. За Sapindaceae Pistachioides striata gen. et sp. nov. 2 Devincer ин aren: 28 Pollostosperma dictyum gen. et sp. nov. 260 Palaeoallophylus globosa sp. nov. 36 — аи age acti: "D я Palaeoallophylus gordonii sp. nov. 20 TR ким ас први ^ MEN Pteronepelys wehrii gen. et sp. nov. 2 Pulvinisperma minutum gen. et sp. nov. 170 Bumelia? globosa sp. nov. 40+ Pyrisemen attenuatum gen. et sp. nov. 32 Bumelia? subangularis sp. nov. B Quintacava velosida gen. et sp. nov. 1 Schisandraceae Sambucuspermites rugulosus gen. et sp. nov. 9 Schisandra oregonensis sp. nov. 27 Scabraecarpium clarnense gen. et sp. nov. 9 Staphyleaceae Scalaritheca biseriata gen. et sp. nov. 10 Tapiscia occidentalis Manchester 78 Scaphicarpium radiatum ant span. 13 EN Sphaerosperma riesii gen. et sp. nov. 2 Sphenosperma baccatum gen. et sp. nov. 1 Symplocos nooteboomii sp. nov. d Stockeycarpa globosa gen. et sp. nov. 1 Theaceae Striatisperma coronapunctatum gen. et sp. nov. 69 Cleyera grotei sp. nov. 32 Tenuisperma ellipticum gen. et sp. nov. 2 RA Tiffneycarpa scleroidea gen. et sp. nov. 1 ми Trigonostela oregonensis gen. et sp. nov. 9 Celtidoideae Tripartisemen bonesii gen. et sp. nov. 270+ Aphananthe тай sp. nov. 12 Triplascapha collinsonae gen. et sp. nov. 4 Celtis burnhamae sp. nov. 100+ Triplexivalva rugata gen. et sp. nov. 12 Celtis sp. 160+ Trisepticarpium minutum gen. et sp. nov. 1 Trema nucilecta sp. nov. 46 Truncatisemen sapotoides gen. et sp. nov. 3 Ulmoideae Ulosperma hardingae gen. et sp. nov. 4 Cedrelospermum lineatum (Lesq.) Manchester 11 Wheelera lignicrustae gen. et sp. nov. 3 Vitaceae Carpolithus bellispermus Chandler 100 Е у Carpolithus sp. 1 > Ampelocissus auriforma sp. nov. 90 Carpolithus sp. 2 2 Ampelocissus scottii sp. nov. 10 Carpolithus sp. 3 3 Ampelopsis rooseae sp. nov. 260 Carpolithus sp. 4 1 nn... angustisulcata Scott 10 Carpolithus sp. 5 Э en clarnensis sp. nov. 32+ Carpolithus sp. 6 1 itis magnisperma Chandler 9 Carpolithus sp. 7 3 Vitis tiffneyi sp. nov. 50+ Carpolithus sp. 8 1 Incertae Sedae Carpolithus sp. 9 10 Ankistrosperma spitzerae gen. et sp. nov. 48 Carpolithus sp. 10 4 Anonymocarpa ovoidea gen. et sp. nov. 1 Carpolithus sp. 11 1 Ascosphaera eocenis gen. et sp. поу. 20 Carpolithus sp. 12 1 Axinosperma agnostum gen. et sp. nov. 22 Carpolithus sp. 13 1 Bonesia spatulata gen. et sp. nov. 15 Carpolithus sp. 15 1 Comminicarpa friisae gen. et sp. nov. 5+ Carpolithus sp. 16 15 Cuneisemen truncatum gen. et sp. nov. 9 Carpolithus sp. 17 1 Dentisemen parvum gen. et sp. nov. 3 Carpolithus sp. 18 2 Durocarpus cordatus gen. et sp. nov. 1 Carpolithus sp. 19 1 Ferrignocarpus bivalvis gen. et sp. поу. 24 Carpolithus sp. 20 1 Fimbrialata wingii gen. et sp. nov. 10 Carpolithus sp. 21 1 Fragarites ramificans gen. et sp. поу. 7 Five-part flower 2 Globulicarpium levigatum gen. et sp. nov. 20 Extant families: 75 genera, 102 species
18 PALAEONTOGRAPHICA AMERICANA, NUMBER 58
include some interesting puzzles for future systematic research.
4) Stereotype genera. Those that have features di- agnostic ofa particular modern family, with a suite of characters found in more than one living genus of that family. Species assigned to stereotype genera cannot be resolved to a single extant or extinct genus; for example, Anonaspermum, Laurocarpum, Palaeosino- menium.
One must be cautious when interpreting fossil ge- neric names. Form genera should not be blanketly as- sumed to represent extinct genera, because it is just as likely that they may represent undetermined extant genera. Likewise, in the characters available for study, stereotype genera are indistinguishable from two or more extant genera, and accordingly should not be assumed to represent extinct genera.
The Nut Beds biota also includes insect remains. Two types of beetles are known from silicified adults (5. В. Manchester, pers. obs.) and one of Ше taxa pre- liminarily thought to be silicified seeds has been de- termined to represent eggs of the Phasmidae (Sellick, in press). The possibility exists that some of the form genera described in this monograph represent insect eggs or egg cases rather than plant remains. Rather than exclude taxa that could possibly be construed as insect eggs, I believe it is useful to document all of the structures because of the possible importance for bio- stratigraphy and/or future insect work. As this goes to press, I consider the following taxa worthy of inves- tigation as possible insect remains: Carpolithus bellis- permus, Scalaritheca and Pileospermum. These taxa, nevertheless, have been included in the fruit/seed counts and, if ultimately determined to be of insect origin, will slightly alter some of the percentages presented in this section.
Thirty-five extant families have been recognized among the Nut Beds fruits and seeds. Among the 75 genera of known familial affinities, 38, or 5096, have been attributed to genera that are still living today (Text-fig. 5B). These extant genera, indicated in Tables 1 and 5, are significant in showing little or no evolution in fruit morphology over a period of at least 43 million years. The number of extant genera recognized may be expected to increase with additional work as the аћп- ities of form genera come to light.
At least 26 genera (3596 of those for which familial affinities have been determined) are extinct: Atriae- carpum, Cedrelospermum, Chandlera, Comicilabium, Kardiasperma, Coryloides, Cruciptera, Curvitino- spora, Davisicarpum, Deviacer, Diploporus, Eohyp- serpa, Fortunearites, Iodicarpa, Langtonia, Mac- ginicarpa, Mastixicarpum, Mastixioidiocarpum, Odontocaryoidea, Palaeophytocrene, Palaeosinomen- ium, Pentoperculum, Saxifragispermum, Tanyopla-
tanus, Thanikaimonia, Tinomiscoidea. This figure, too, will probably change with additional work. Sometimes fossils initially considered to be extinct are subsequent- ly found to be living as neobotanical exploration and comparative work continues, as in the classic case of Metasequoia, and as more recently discovered in Di- plopanax (Eyde and Xiang, 1991) and Craigia (Kva- éek, Büzek and Manchester, 1991). Morphological dif- ferences between related fossil and extant taxa that may be viewed as warranting generic distinction by one investigator may be viewed as simply reflecting greater variability within a single genus by another. For ex- ample, the species formerly treated as Palaeonyssa by Reid and Chandler (1933) and Scott (1954) are herein regarded as Nyssa.
Relative to other paleobotanical localities of the Clarno Formation, the Nut Beds fruit and seed assem- blage is anomalously high in generic and specific di- versity. In surveying the large collections from the la- custrine shales at West Branch Creek (UF loc. 229c, 230, UCMP loc. 3904, USGS loc. 8637), I have ob- served only 48 distinct types of compressed fruits and seeds and 53 types of leaves.
The leaf flora of the Nut Beds includes about 65 genera—less than half as many as those known from fruits and seeds. Although the leaf collection acquired from the Nut Beds between 1974 and 1981 1s the result of more than 2000 person-hours, the recovery process for leaves in the heavily fractured siltstones (Man- chester, 1981) is very time-consuming and the total number of leaf specimens is less than 1000. The fruit and seed collections amassed since 1943 include more than 20,000 specimens. Thus, the discrepancies be- tween foliage and diaspore diversity may be in part due to differences in sample size.
It's unlikely that all of the Nut Beds species grew together in the same plant community; rather, it is probable that the assemblage includes elements from various parts of a large watershed. Judging from the sedimentary environment of the Nut Beds and the varying amounts of predepositional erosion evident in the fossil remains, the assemblage probably includes plants from different sources: some local, others carried from various distances upstream. Significant transport of fruits and seeds may also have been effected by birds that consumed fruits from various parts of the forest. When perched on branches close to or overhanging sites of sediment accumulation, birds have the poten- tial to introduce endocarps and seeds from more re- mote parts of the forest that would otherwise be un- derrepresented in the fossil assemblage. Such biotic transport, selective for bird-edible fruits, might explain the greater diversity of fruits and seeds than vegetative organs in the Nut Beds. Frugivorous mammals prob- ably also have played a role in fruit dispersal in Eocene
| |
CLARNO (EOCENE) FRUITS AND SEEDS: MANCHESTER 19
Unknown
48%
Modern Genera, Stereotype Genera, | Extinct Genera
Form Genera
48%
Families
52%
15%
Stereotype genera (11)
50%
Ф г. + oe 7 vr — » * 2
7 35%
7
PAPIERE Ci d Pu
2
Text-figure 5.— Taxonomic status of Nut Beds fruit and seed genera relative to extant families and genera. A, Proportion of genera identified to modern families vs. those of unknown familial affinities. B, Genera identified to modern families, with proportions of extinct, modern and Stereotype genera. C, Summary, emphasizing that only У, of genera have been determined to be extant.
forests (Collinson and Hooker, 1992). Indeed, a large Proportion of the seeds and endocarps identified ap- Dear to have been from drupes or berries that would have been well-suited for bird and/or mammal dis- Persal. The apparent absence of vitaceous leaf remains In the Clarno collections, contrasted with the abun-
dance and diversity of grape seeds, supports this con- clusion.
GROWTH HABIT AND DISPERSAL ECOLOGY
| The growth habit of a fossil species, i.e., whether it 15 a tree, shrub, Папа or herb, may be inferred from that of its living relatives. Of course, this means of Inference is not applicable to the fruit and seed taxa whose modern affinities are undetermined; therefore, fewer than half of the Nut Beds fruit and seed genera Сап be analyzed according to this approach. Never- theless, this methodology gives some insight into the
stature of Nut Beds vegetation (Table 2). Among the identified taxa, trees and lianas are particularly well represented, while herbs and aquatic plants are poorly represented.
Many of the fruit and seed genera are deduced to represent trees. Arborescent taxa include Aphananthe, Castanopsis, Juglandaceae (3 genera), Lauraceae, Mac- ginicarpa, Magnolia, Mastixia, Meliosma, Nyssa, Platanus, Quercus, Sabal, Tapiscia. Trees of lower stature and shrubs probably included Annonaceae, Cornus, cycad (known from foliage only), Fortunear- ites, Taxus, and Torreya. Arborescent taxa can also be inferred independently of fruit and seed data using fossil wood fragments from the Nut Beds. Based on curvature of growth rings, at least 40 different genera were derived from stems larger than 15 cm in diameter and were probably from trees. Some of these woods have been identified, including: Ginkgo (Scott et al.,
PALAEONTOGRAPHICA AMERICANA, NUMBER 58
Table 2.—Present habit and vegetational distribution of Nut Beds genera.
deciduous/
genus habit evergreen TRF PTRF NBLE MM MBLD MNH MC Actinidia с а + + + ar ar Alangium tsc de s + + + Es Ampelocissus (e d de Ampelopsis d dE ES Anamirta © e + + Aphananthe $ de at + ES ar Bumelia ts d + + Calycocarpum с а + Castanopsis ts e d dE 3E ar 4E 15 + Celtis t de + + + F + Cleyera s e + Cornus ts d Qu + 3r di Decodon S d ar ar Diploclisia с е 3pm + Emmenopterys t d T Ensete h e ae Es Hydrangea tsc de T + + + + + Iodes с e Е НЕ Juglans t + + + + + Lindera ts d ES + AB qe Magnolia t de + НЕ + + + dr Mastixia t e e Meliosma t de eJ 4 ar ar Н + Parthenocissus с а ar Pinus t e 7 ae Ae dE qe 4E Ar Platanus t d + + + + 7 Prunus ts de К dp EN 3E ar Pyrenacantha С ё ap dc ar Quercus + де + ap 4E зБ dis dE jr Rhus tsc de +1 s. E 4r Sabal + е ar + + Sabia с e ar Schisandra © de dE ES + Symplocos ts de + + + + + + Тарїзсїа ts d dE Taxus t @ Ar a Tinospora g © 46 + Torreya t e Ae E Trema ts e ap dE 4p Vitis с а T de a Ar
' Including submontane rainforest. 2 Including montane forest.
c = climber, s = shrub, t = tree, h = herb, d = deciduous, e = evergreen; TRF = Tropical Rain forest; PTRF = Paratropical Rain forest; NBLE = Notophyllous Broadleaved Evergreen; MM = Mixed Mesophytic; MBLD = Mixed Broadleaved and Deciduous; MNH = Mixed Northern Hardwood; MC = Mixed Coniferous. Vegetation types following Wolfe (1979).
1962), Tapirira (Manchester, 1977), Sterculiaceae (Manchester, 1979, 1980), Alangium, Betula, Faga- ceoxylon, Magnolia, Plataninium, Quercinium, Ul- minium, (Scott and Wheeler, 1982), Engelhardioxylon (Manchester, 1983), and Clarnoxylon (Manchester and Wheeler, 1993).
The diversity of climbers in the flora, at least 22 genera and 30 species, (43% ofthe 69 species for which growth form may be inferred) is particularly striking. This estimate is based upon extant genera of climbing habit (Table 2), plus extinct genera belonging to fam- ilies or subfamilies that are virtually all climbers today.
Menispermaceae, a family comprised almost exclu- sively of tropical lianas, is the most diverse family in the Nut Beds fruit and seed assemblage, with 13 genera and 14 species. In addition, seven species of Vitaceae, at least five species of Icacinaceae (those belonging 10 Iodes, Iodicarpa, Palaeophytocrene, Pyrenacantha), and the species of Actinidia, Sabia, and Schisandra prob- ably represent lianas or climbers (Table 2). By contrast with the Nut Beds figure of 43%, the highest known percentages of liana species per flora in extant vege- tation are only about 24% (Gentry, 1992).
Lianas such as those of extant Menispermaceae are
| | |
CLARNO (ЕОСЕМЕ) FRUITS AND SEEDS: MANCHESTER 21
best developed where the tree canopy is interrupted, for example along stream banks and forest edges (For- man, 1986). Preferential growth at stream side, where seeds may be dropped directly into the sedimentary environment, may have caused over-representation of lianas relative to trees, shrubs and herbs in the fossil assemblage, but I doubt that this bias fully explains the high apparent diversity oflianas. The same families of lianas are also well represented in the Eocene Lon- don Clay flora of England (Chandler, 1961b; Collinson, 1983). The diversity oflianas along with the trees men- tioned above may be taken as an indication that a multistratal, closed canopy forest had developed (Up- church and Wolfe, 1987). In addition, Crane (1987) has suggested that lianas and opportunistic shrubs may have played a more dominant role in colonizing bare ground in the Early Tertiary than today because the major radiation of extant angiosperm herbs occurred later in the Tertiary.
Very few ofthe known Nut Beds taxa represent herbs. Among Ше fruit and seed genera, Ensete is the only known herbaceous element. Stem and leaf records at- test to the presence of Equisetum clarnoi (Pl. 1, fig. 1) and a few ferns. The ferns are represented by silicified petioles of Dennstaedtiopsis and Acrostichum and by foliage of Thelypteris iddingsii (Knowlton) MacGinitie, and “Asplenium” hurleyensis Berry (Manchester, 1976). Lygodium kaulfussii Heer, which is common at many Clarno localities (Manchester and Zavada, 1987), has not been recovered from the Nut Beds. Floating aquat- ic plants such as Nymphaeaceae, Ceratophyllum, Tra- ра and Salvinia are absent. Decodon, with a single €xtant species that grows as a sprawling shrub at the edge of swamps, is the only genus of the Nut Beds flora regarded as semi-aquatic.
A small percentage of the taxa have winged dissem- inules and were probably wind-dispersed, such as the Samaras of Cedrelospermum, Deviacer, Cruciptera, Pa- laeocarya, Palaeoplatycarya, and Pteronepelys, the tufted achenes of Platanus and Tanyoplatanus, and the seeds of Emmenopterys, Fimbrialata and Hydrangea. In extant forests, wind-dispersed species tend to best Tepresented among trees and lianas of the forest can- Ору, while bird-dispersed species are concentrated in the understory (Gentry, 1983). Gentry (1983) also not- ed that the percentage of species with wind-dispersed diaspores decreases as one proceeds from dry to wet forests in the neotropics.
At least 80% of Nut Beds taxa appear to have been Well-suited for biotic dispersal. Nuts that commonly are dispersed by rodents today include Juglans, Cas- tanopsis and Quercus and it may be expected that Cor- Yloides, being similar to Corylus, was similarly adapt- ed. Although I have not observed obvious gnaw marks Оп any of the fossil nuts or seeds, it is likely that rodents
played some role in seed dispersal. Among the mam- malian fauna of the Nut Beds, Orohippus, Hyrachyus and Telmatherium were browsers that may have par- ticipated in fruit dispersal. Birds were probably the most important dispersal agents for Nut Beds plants. As inferred from extant relatives and from morphology of the preserved fruit and seed remains, the over- whelming majority of Nut Beds species represent fleshy fruits, of the kind regularly dispersed by birds. Some examples include Actinidia, Anamirta, Annonaceae, Aphananthe, Celtis, Cornus, Ensete, Mastixia, Me- liosma, Nyssa, Prunus, Rhus, Sabal, Symplocos, Tap- iscia, Taxus, and the seven species of Vitaceae.
Although seed size ranges greatly among the Nut Beds species, the occurrence of many species with large fruits and seeds is particularly striking (Scott, 1954; Tiffney, 1984). Based on a survey of seed size in Cre- taceous and Tertiary floras, Tiffney (1984) observed that most angiosperm seeds of the Cretaceous are small; large fruits first become common in the Early Tertiary. The abundance of large fruits in Eocene floras such as the Nut Beds probably coincides both with the radi- ation of mammal and bird taxa important as dispersal agents, and with the formation of closed canopy forest (Tiffney, 1984). A well-marked correlation between large seed size and establishment in shady, stable plant associations has been statistically documented in living tropical woody plants (Salisbury, 1942; Foster and Jan- son, 1985; Foster, 1986). In a closed canopy evergreen forest, large seeds have an adaptive advantage over smaller ones because they possess more food reserves for the embryo, enabling the seed to germinate and become established under subdued light conditions. Smaller seeds typically require more intense sunlight for sufficient photosynthesis for a seedling to become established, and are commonly produced in large quantities by plants that are successful at colonizing sun-lit gaps in the forest canopy. The high proportion of relatively large, biotically dispersed fruits and seeds in the Nut Beds assemblage may thus be an indication of closed canopy forest.
CLIMATIC INTERPRETATION
Fossil plants and animals of the Clarno Nut Beds reflect climatic conditions under which the biota ex- isted. The presence of crocodilian teeth and such ther- mophilic plants as cycads, palms (Sabal) and bananas (Ensete) gives an indication that the Nut Beds biota lived under typically frost-free, subtropical or tropical conditions. The Middle Eocene coastline was situated only about 100 km to the west (Nilsen and McKee, 1979) so that the Clarno biota probably experienced the moderating climatic influence of the nearby ocean. Deposition of the Clarno Formation took place prior
22 PALAEONTOGRAPHICA AMERICANA, NUMBER 58
to the Neogene uplift ofthe Cascade and Coast Ranges that effected the present-day rain shadow.
Paleoclimatic conditions are inferred by identifying floristic and/or physiognomic similarities of a fossil assemblage to a modern vegetation type of known cli- matic parameters. Discussion of paleoclimate requires a working definition of terms, because different inves- tigators have applied widely differing concepts of what constitutes such categories as tropical, subtropical, and temperate. For consistency with other paleobotanical literature Ihave adopted the definitions of Wolfe (1979) which are based on modern mean annual temperatures (MAT) for humid to mesic regions of eastern Asia: tropical >25°C; paratropical 20-25°С; subtropical 13- 20°C; temperate 10-13°С; paratemperate 3-10?C; sub- temperate <3°C. These paramaters, along with mean cold month temperature and mean annual range of temperature, appproximately delimit distinct forest types within mesic vegetation (Wolfe, 1979).
The high proportion oflianas already discussed (43% of the species with known familial affinities) supports previous interpretations that the Nut Beds flora rep- resents tropical or paratropical forest (Scott, 1954; Wolfe 1977), possibly with a multistratal canopy (Up- church and Wolfe, 1987). Observing that in Recent vegetation, “the great majority of woody lianas are restricted to tropical forests," Gentry (1992) computed the percentage of lianas per florula in temperate eastern North America at 6%, vs. 19% in continental tropical florulas. Gentry’s work did not specifically address li- ana diversity in subtropical or paratropical situations.
Woods from the Nut Beds consistently show growth rings, indicating seasonality of temperature and/or pre- cipitation. However, true ring-porous wood structure, which typically is confined to temperate situations, has not been observed. Of the fruit and seed genera that have been identified to extant genera, at least 42% are deciduous-leaved today (Table 2). Fossil leaves from the Nut Beds are mostly of large size (67% of the di- cotyledonous leaf species exceed 10 cm in lamina length) and, at latest count (preliminary to a detailed work on the leaf flora), 33 of the dicot leaf species, or 52%, are entire-margined. Additional research on leaves and woods of the flora may provide physiognomic data useful in testing the taxonomically-based assessment provided here.
One approach to the assessment of paleoclimate is based on the growing conditions of extant taxa related to those identified in the fossil assemblage. The cli- matic preferences of present-day relatives of the fossil taxa are taken as an indication of the range of condi- tions under which the fossil flora might have existed. There are various problems inherent in using this uni- formitarian approach: 1) the climatic tolerances of a
taxon may evolve to accommodate environmental change; 2) the current distribution of a taxon is con- trolled by many factors, only one of which is regional climate (other important factors include historic bar- riers to dispersal, and effects of interspecies competi- tion and pathogens); 3) any taxa that are misidentified may introduce erroneous climatic data. Because of problems that may result from mistaken identities, I have been particularly cautious in the systematic treat- ment of genera in this monograph, placing them in form- or stereotype genera, rather than in the “nearest modern equivalent” when characters diagnostic of a modern genus could not be demonstrated. Despite the limitations of taxonomically based assessments of pa- leoclimate, such analyses provide at least a coarse guide, and may be used to evaluate conclusions reached by other methods.
Among the Nut Beds taxa that have been identified to Recent genera, most live today within the broad range of temperate to tropical vegetation (Table 2). Some of these, such as Ampelocissus, Anamirta, Di- ploclisia, Ensete, Iodes, Mastixia, Pyrenacantha, and Tinospora, are confined to tropical and paratropical vegetation today. Others are elements of temperate forest associations (Mixed Mesophytic and Mixed Broadleaved Deciduous forests, sensu Wolfe, 1979, as modified from Wang, 1961), and are not found in trop- ical vegetation, e.g., Cornus, Decodon, Emmenopterys, Fortunearites (inferred from extant Fortunearia and Sinowilsonia), Parthenocissus, Tapiscia, and Torreya. Table 2, listing extant genera of the Nut Beds flora and the vegetation types in which they occur today, shows the highest number of taxa (27 of 40 genera) occurring in the Mixed Mesophytic and Mixed Broadleaved Ev- ergreen forest communities, followed by the 24 genera with Paratropical Rain forest affinities, and 22 with Tropical Rain forest affinities. However, if one were to score the extinct genera of Menispermaceae and Icacinaceae as tropical, where these families predom- inate today, the larger proportion of Nut Beds taxa (37 of 52 genera, or 71%) would be regarded as tropical to paratropical.
Faced with a mixture of taxa, some that are exclu- sively tropical today and others that are exclusively temperate, interpretation of paleoclimate for the Nut Beds depends on the assumptions made and how the data are weighted. One might, for example, infer 4 temperate environment for the Nut Beds and suggest that many elements have subsequently evolved into more tropical situations where they are found today. Or, alternatively, the Nut Beds may represent tropical or paratropical vegetation, some of the elements of which have since evolved to accommodate more tem- perate conditions. The second scenario is favored be-
| | | |
|
CLARNO (EOCENE) FRUITS AND SEEDS: MANCHESTER 23
cause tropical to subtropical vegetation is also sug- gested by foliar physiognomic considerations (Wolfe, 1977; Manchester, 1981). It is also possible that the apparent mixture of cool and warm elements is a gen- uine reflection of mixed sources, some of the cooler elements having been transported from higher eleva- tions.
Presumably, the presence of frost intolerant taxa places more constraints on the inference of climate than the presence of frost tolerant genera. Although temperate taxa may survive in tropical situations, par- ticularly at higher elevations, tropical elements such as Ensete, a member of the banana family, would not be expected to persist in a setting with cold winters. Ensete occurs today in paratropical and tropical veg- etation of southeast Asia where the mean annual tem- perature falls within the ranges of 20-25°C (Paratrop- ical Rain forest, sensu Wolfe, 1979) and 25-28°C (Tropical Rain forest sensu Wolfe, 1979), and the mean annual range of temperature is low. The mean annual range of temperature varies from 2-19°C among dif- ferent weather stations in Paratropical Rain forest and 2-11*C in Tropical Rain forest (Wolfe, 1979). The cold month mean temperature in these vegetation types is 10°С or greater. If it is assumed that its climatic tol- егапсеѕ have not changed significantly since the Eo- cene, then the presence of Ensete suggests warm, moist, €quable climate, as is also consistent with the presence ОЁ cycads and palms.
The Nut Beds provides just one example ofthe warm Vegetation that was widespread in the Eocene of the northern hemisphere (Wolfe, 1985). Tropical to par- atropical vegetation has also been recognized in the Eocene of southern Alaska (Wolfe, 1972, 1977) and in Europe (Reid and Chandler, 1933; Chandler, 1964). A marked temperature decline is apparent by the Early Oligocene, possibly corresponding to changes in ocean Currents effected by the Late Eocene separation of South America from Antarctica resulting in development of the circum-Antarctic current (Kennett, 1977; Parrish, 1987) and the formation of deep water channels en- abling southward drainage of Arctic Ocean waters Caused by the rifting of Spitzbergen and Iceland from лш (McKenna, 1975; B. LePage, pers. comm.,
992).
BIOGEOGRAPHIC CONSIDERATIONS
The Nut Beds flora contributes a wealth of data use- ful in phytogeographic investigations. Floristic com- Parison, both with other Tertiary floras and with extant floras, provides insight into Ше spread of boreotropical forest in the Northern hemisphere. This section con- Siders the levels of taxonomic similarity of the Nut Beds flora with other fossil fruit and seed assemblages
and with living floras. Comparison with other Tertiary floras indicates particularly close similarity with the Eocene flora of western Europe. Among modern-day vegetation, the closest taxonomic resemblance is found in eastern Asia.
Comparison with other Fossil Floras
Because of the high proportion of Nut Beds genera that are extinct and/or of indeterminate familial affin- ity, it is not possible to link each of the fossil fruit/ seed species with a corresponding leaf species. There- fore, the most informative paleofloristic comparisons are those made with other fossil fruit and seed assem- blages. Detailed comparison with other Paleogene leaf floras will become possible when the leaf assemblages of the Nut Beds and other Clarno localities are docu- mented.
North American Fossil Floras
The Late Eocene or Early Oligocene LaPorte flora of northern California, best known for its leaf assem- blage (Potbury, 1935), has recently yielded a good col- lection of well-preserved, lignitized fruits and seeds, many of which are shared at the generic, and possibly specific, levels with Nut Beds taxa. A preliminary tax- onomic list of LaPorte fruit and seed flora was pre- sented by Tiffney (in Doyle et al., 1988) along with an identification key and illustrations (those taxa indi- cated by asterisks (*) are shared with the Nut Beds flora): Dracontomelon type [probably Pentopercu- lum*], hamamelidaceous seed*, Quercus*, Pterocarya/ Cyclocarya type, Magnolia?*, Мазижа (Ganitrocera type)[probably Mastixioidiocarpum*], Odontocaryoi- dea type*, Stephania type [similar to Palaeosinomen- ium*], Tinospora type*, Nyssa*, Rhamnus?, Zanthox- ylum, Sapotaceae*, Tapiscia?*, Halesia?, Symplocos*, Vitis*, Parthenocissus*. It is noteworthy that the Nut Beds flora bears greater similarity to this younger flora than to other Middle Eocene fruit and seed assemblages in North America.
The Middle Eocene Princeton chert flora of southern British Columbia, Canada includes a variety of ana- tomically preserved fruits and seeds that are revealed by sectioning and peeling the chert (Basinger and Roth- well, 1977; Cevallos-Ferriz and Stockey, 1988a, 1988b, 1989, 1990, 1991). The flora is relatively low in di- versity, but includes several taxa in common with the Nut Beds flora, including Vitaceae (Cevallos-Ferriz and Stockey, 1989), Lauraceae, sabaloid Palmae (Erwin, 1987), Decodon (Cevallos-Ferriz and Stockey, 19882), Prunus (Cevallos-Ferriz and Stockey, 1991), a mastix- ioid species similar to Mastixicarpum (Stockey, pers. comm., 1990), and the newly recognized Stockeycarpa (p. 111). The Princeton flora includes aquatic plants
24 PALAEONTOGRAPHICA AMERICANA, NUMBER 58
such as Nympheaceae (Cevallos-Ferriz and Stockey, 1989), which are unknown from the Nut Beds, and appears to lack the diversity of lianas characteristic of the Nut Beds assemblage. Metasequoia, which is abun- dant in the Princeton flora (Basinger, 1981) is absent from the Clarno Formation, though it becomes abun- dant in the Lower Oligocene ofthe overlying John Day Formation (Manchester and Meyer, 1987). The full diversity of the Princeton flora remains to be docu- mented so I hesitate to provide statistical comparison between the Princeton and Nut Beds floras. However, the presence of several shared genera indicates that the geographic ranges of some of the Nut Beds taxa ex- tended into the northern part of the continent.
Although it is known mostly from leaf impressions, the Middle Eocene (Lower Ravenian) assemblages of the Kushtaka and Kulthieth Formations in the Gulf of Alaska appear to be floristically very similar to the Nut Beds assemblage (Wolfe, 1972, 1977). Shared or similar genera include Alangium, Anamirta, Diplocli- sia, Limacia (cf. Davisicarpum), Mastixia, Meliosma, Palaeophytocrene, ?Pyrenacantha, Sabalites, and Vi- tis. These assemblages are similar to the Nut Beds flora in the high proportion of lianas. Geophysical and pa- leobotanical data suggest that the Gulf of Alaska floras are part of a tectonic block that was located farther south in the Eocene (Cowan, 1982; Saleeby, 1983; Ax- elrod, et al., 1991).
In the southeastern United States, the Middle Eo- cene Claiborne flora of Tennessee and Kentucky is best known for fossil leaves (Berry, 1930; Dilcher, 1971), but has also yielded about 90 distinct taxa of fruits and seeds (Grote, 1989). Given the similarity in age, one might expect a high percentage of taxa to be shared between the Clarno and Claiborne floras. Although some of the same families are present, e.g., Annona- ceae, Hamamelidaceae, Juglandaceae, Leguminosae, Magnoliaceae, Sapotaceae, and Theaceae, generic sim- ilarity is minimal, including only Magnolia (Grote, 1989), Nyssa (Dilcher and McQuade, 1967), Palaeo- carya (Dilcher, Potter and Crepet, 1976; Manchester, 1987), and Sabal (represented by leaves, Daghlian, 1978). Unlike the Nut Beds flora, the Claiborne in- cludes a large number oftaxa with present-day relatives in tropical America (Roth and Dilcher, 1979; Taylor, 1990). The Clarno flora lacks the high diversity of legumes found in the Eocene of southeastern North America (Herendeen and Dilcher, 1990; Herendeen, 1992). The apparent absence in the Claiborne flora of families that are diverse in the Clarno flora, such as Icacinaceae, Menispermaceae and Vitaceae, is con- spicuous. These differences suggests major floristic dis- continuity between northwestern and southeastern North America in the Middle Eocene. The floristic differences between the Clarno and Claiborne assem-
blages might be due in part to climatic and edaphic factors, but may also reflect barriers to dispersal, in- cluding the prior mid continental sea of the Cretaceous and rise of the Rocky Mountains. High paleoelevations that have been inferred for the southern Rocky Moun- tains by the Late Eocene (Meyer, 1992) may have re- inforced floristic differentiation of western and eastern Eocene floras of North America.
The Brandon Lignite flora of Vermont is relevant in this comparison as the only well-documented Tertiary megafossil flora of northeastern North America. Al- though not independently dated, the flora is thought to be Oligocene (Tiffney, 1981) or perhaps Early Mio- cene (Tiffney, 1985b). The Brandon flora appears to lack tropical elements such as Menispermaceae and Icacinaceae. Genera shared with the Nut Beds include Alangium (Eyde, Bartlett and Barghoorn, 1969), Mag- nolia (Tiffney, 1977), Vitis and Parthenocissus (Tiffney and Barghoorn, 1976), Nyssa (Eyde and Barghoorn, 1963), and Symplocos (from pollen, Traverse, 1955; from endocarps, Tiffney, pers. comm., 1992). The composition and geographic placement of the Brandon flora provide support for the concept ofa Tertiary land connection across the North Atlantic (Tiffney, 1985b).
European Fossil Floras
Most of the families and many of the genera present in the Nut Beds are also known from the Eocene of Europe (Text-fig. 6). Comparable floras include the Early Eocene London Clay of southern England and the Middle Eocene Geiseltal and Messel floras of Ger- many (Table 3). The largest and most intensively in- vestigated of these, with about 200 described genera of fruits and seeds, is the London Clay (Reid and Chan- dler, 1933; Chandler, 1961b, 1964, 1978; Collinson, 1983). The London Clay assemblage includes pyritized fruits and seeds collected from several localities of the London Clay Formation along the southern coast of England. The London Clay fossils were deposited in marine sediment, and include disseminules of man- grove vegetation and shark teeth that are absent in the Nut Beds. However, most of the London Clay plants are terrestrial, evidently deposited near the mouth of a large river, and provide a good record of regional vegetation.
Similarity between the Clarno and London Clay flo- ras was recognized previously on the basis of prelim- inary investigation of the Nut Beds assemblage and has contributed to the concept of an Eocene boreo- tropical flora (Scott, 1954; Chandler, 1964; Wolfe, 1975; Tiffney, 1985b; Mai, 1989). It is now possible to quan- tify the level of similarity based on the entire Nut Beds fruit and seed collection. Of the 145 fruit and seed genera currently recognized from the Nut Beds assem- blage, 30 genera (2096) are shared with the London
| | | | | |
| | |
CLARNO (EOCENE) FRUITS AND SEEDS: MANCHESTER 25
Clay flora (Table 3; Text-fig. 6). In addition, about 15 Species (10%) are the same or are nearly indistinguish- able from those of the London Clay. These data cor- roborate other evidence for the existence of a land- connection between Europe and North America during the Early Tertiary (McKenna, 1975; Tiffney, 1985a, b). Dated as Early Eocene (Collinson and Hooker, 1987) the London Clay biota predates the Middle Eocene Nut Beds biota by about five million years. The timing, routes and rate of dispersal of individual taxa across land connection(s) between western Europe and west- ern North America are not well understood, but the younger age ofthe Nut Beds deposit indicates that the interchange of these taxa between Europe and North America must have occurred during or prior to the Middle Eocene.
Fruit and seed floras of central Europe are not as diverse, or have not been as thoroughly documented, as the London Clay flora, but include additional taxa shared with the Clarno flora. The Middle Eocene Gei- seltal flora from extensive lignite deposits in eastern Germany includes well-preserved leaves (Rüffle, 1976), pollen (Krützsch, 1976) and fruits and seeds. Mai (1976) recognized 25 genera and 28 species of fruits and seeds from the Geiseltal, nine of which are shared with the Nut Beds: Castanopis, Iodes, Magnolia, Meliosma, Nyssa, Pinus, Sabal, Tapiscia, and Alangium. The Middle Eocene Messel flora in western Germany (Schaarschmidt, 1988) is a lacustrine deposit that in- cludes well-preserved compressions of leaves (Wilde, 1989), and fruits and seeds (Collinson, 1982, 1988). About fifty genera of fruits and seeds are present in the Collections from Messel, about one quarter of which are shared with the Nut Beds. The shared genera in- Clude Cruciptera, Iodes, Magnolia, Mastixia, Melios- та, Palaeocarya, Palaeophytocrene, Palaeosinomen- ium, Pentoperculum, Prunus, Tapiscia, Tinospora, Vitis?, and Tripartisemen (Collinson, 1982, 1988 and Pers. obs.)?. Of these, Palaeocarya, Tripartisemen (“?Lythraceae” in Collinson, 1988) and Cruciptera (Manchester, 1991) are not recorded from the London Clay.
Combining data from the London Clay, Messel and Geiseltal floras (Table 3), the percentage of Nut Beds genera shared with the Early to Middle Eocene of Eu- Tope is 24. Some of the Nut Beds taxa have not been Tecorded from the Early and Middle Eocene of Europe but are known from later Tertiary deposits. Examples include Actinidia (Kirchheimer, 1957; Friis, 1985), Cedrelospermum (Manchester, 1987b, 1989a)?, Celtis (Manchester, 1989c), Decodon (Friis, 1985), Hydran- Sea (Mai, 1985), Rhus (Friis, 1969), and Sabia (Geis- nn
* Cedrelospermum is now also recognized at Messel. See Notes added in proof, p. 200.
Total
40 4
London Clay 30 4
20 -
Messel
Number of Genera
Geiseltal
Text-figure 6.— Nut Beds fruit and seed genera shared with Eocene floras of Europe. Genera listed in Table 3.
sert and Gregor, 1981). Most of these are no longer native to Europe, having succumbed to Plio-Pleisto- cene climatic cooling.
Paleogeography
The geographic routes of exchange between the Nut Beds flora and the London Clay, Messel and Geiseltal floras of western and central Europe are ofconsiderable phytogeographic interest (Tiffney, 1985a, b). Geo- physical data suggest that land bridges were open both across the North Pacific and across the North Atlantic during the Eocene (Parrish, 1987). If the connection was across Beringia, then one would expect to see marked floristic similarities with the Eocene of Asia. Although floristic ties with the Recent flora of eastern Asia are strong (discussed next section), the antiquity of this connection needs to be documented. In this regard, it would be desirable to make detailed com- parisons with floras of similar age in Asia. However, I am not aware of comparable Asian Eocene fruit and seed floras. Careful comparisons among leaf and pollen records are needed for a better understanding of the similarities and differences between North American and Asian Eocene floras, a project beyond the scope of the present study.
Even if taxa were successful in spreading between North America and Asia via Beringia, the Turgai sea- way that separated Asia and Europe through at least the Early Eocene may have blocked exchange with Eu- rope (Tiffney, 1985b). Exchange across the North At- lantic via connections through Greenland and Iceland is viewed as more likely (McKenna, 1975; Tiffney, 1985b), but Early Tertiary floras so far studied from Greenland and Spitzbergen give no clue to the presence of such thermophilic taxa as Menispermaceae and Ica- cinaceae.
Comparison with Recent Floras
The only genera of the Nut Beds flora that are still native to Oregon today are Pinus, Taxus, Celtis, Cor-
26 PALAEONTOGRAPHICA AMERICANA, NUMBER 58
Table 3.—Clarno fruit and seed taxa shared with European Eocene floras.*
Alangiaceae
Alangium Le, Су Anacardiaceae
Pentoperculum LC Annonaceae
Anonaspermum Le Burseraceae
Bursericarpum EG Cornaceae
Cornus LC
Langtonia Ee
Mastixia LC, M
Nyssa EGG Fagaceae
Castanopsis G Flacourtiaceae
Saxifragispermum LC Icacinaceae
Тодез LC, G, M
Palaeophytocrene LC, M Juglandaceae
Cruciptera M
Palaeocarya ECE M Lauraceae
Laurocalyx EC
Laurocarpum EC Magnoliaceae
Magnolia LC, G, M
Menispermaceae
Atriaecarpum ще:
Davisicarpum BG
Diploclisia LC
Eohypserpa ще
Palaeosinomenium LC, M
Tinospora LC, M
Tinomiscoidea LC Palmae
Sabal е 76, Rosaceae
Prunus M Sabiaceae
Meliosma LC, б, М Sapindaceae
Palaeoallophylus ще Staphyleaceae
Tapiscia LC, G,M Symplocaceae
Symplocos LC Theaceae
Cleyera TG Vitaceae
Vitis LC, M
Ampelopsis LC
Parthenocissus LC Incertae Sedis
Carpolithus bellispermus LC
Tripartisemen M
LC = London Clay; G = Geiseltal; M = Messel. * See also Notes added in proof, p. 200.
nus, and Quercus. The majority of extant genera of the Nut Beds flora are found today in other geographic regions. Table 4 and Text-fig. 7 show the distribution of Recent genera known from the Nut Beds in selected modern phytogeographic provinces (southeast Asia, Malesia, western North America, Central America, South America, Europe and Africa). Strongest ties are with Southeast Asia and Malesia.
Within the Americas, the closest floristic ties are with eastern North America and with central America. Twenty-three genera, or 56% of the extant genera known from the Nut Beds, occur in eastern North America. Although most of these occur in other continents as well, Calycocarpum and Decodon are endemic to east- ern North America. Twenty-two genera, or 54%, live today in Central America. However, none of these are presently endemic to the region. Sabal, distributed in the southeastern U.S. and tropical America, is one of the few genera of the Clarno assemblage that is limited to the New World in its present-day distribution. The genus was formerly more widespread, however, with Eocene records in Europe (e.g., Mai, 1976) and Asia (Huzioka and Takahasi, 1970).
Twenty-four percent of the Nut Beds genera occur
today in Europe, although none are endemic there. This percentage, although significantly less than those computed for the present-day floras of southeastern Asia, Malesia, southeastern North America and central America (Table 4), matches that presented above in comparing the Nut Beds flora with Early to Middle Eocene floras of Europe (24%). However the similarity of these values for Eocene and Recent flora of Europe probably is an artifact of the quantitative analysis. All 152 genera, both extinct and extant, are included in the base number for comparing fossil assemblages, whereas only the 41 extant genera form the basis of comparison with the Recent flora. It is important to note that although 36 of the Nut Beds genera are shared with the Eocene flora of Europe, only ten are found in Europe today. Many of the extant genera shared be- tween the Clarno and the London Clay floras are now absent from Europe and are disjunct between Asia and eastern North America and/or central America, for example, Nyssa, Hydrangea, Symplocos; and others are endemic to southeastern Asia.
The strongest geographic affinities are with south- eastern Asia (Indochinese Region, sensu Takhtajan, 1986) and Malesia (Malesian Region, sensu Takhtajan,
| | | |
CLARNO (EOCENE) FRUITS AND SEEDS: MANCHESTER 27
Table 4.— Present geographic distribution of extant genera rep- resented in the Nut Beds flora.
genus seAs Mal wNA eNA cAm sAm Eur Afr
Actinidia x x Alangium x x x Ampelocissus х 7 x ? Ampelopsis x ? x Anamirta x x Aphananthe x x х х Bumelia х x х Calycocarpum x Castanopsis х x у Celtis x x x x x x x x Cleyera x x Cornus x x x x x x x x Decodon x Diploclisia x x Emmenopterys x Ensete x x x Hydrangea x x x x Zodes x X х Juglans E x x x x xe Lindera х Y x Magnolia x X x x Mastixia х х Мейозта x x x x Nyssa x x x x Pathenocissus x х x х Pinus x х х x x в х х Platanus x x х * х Prunus х x х x x х Pyrenacantha x x x Quercus x x x x x x x Rhus x x x x x x x x Sabal x х х Sabia x Ж Schisandra x x x Symplocos x х x x x x Tapiscia x Taxus x x х x x x Tinospora x x x Tor reya x i x x Trema x X x x x Vitis x x х х х x
Totals 97 25 14 25 22 12 10 13
SeAs = Southeast Asia; Mal = Malesia; eNA = eastern North Amer- lca; sAm = South America; Eur = Europe; Afr = Africa.
1986). Among Recent floras, the highest number of genera shared with the Nut Beds assemblage is found In eastern Asia. Thirty-seven genera, or 9096, are native to southeastern Asia. The flora of Malesia shares 25 genera with the Nut Beds. Genera that are exclusively Asian or Malesian today include Actinidia, Mastixia, Anamirta, Sabia, Tapiscia. Although Meliosma occurs In central and South America as well as in Asia, four Of the Nut Beds species conform most closely to sub- genus Meliosma, section Meliosma, which is exclu- Sively oriental in its Recent distribution (Van Beuse- kom, 1971). The close floristic similarity with Southeastern Asia and Malesia probably reflects the
Number of Genera
Region
Text-figure 7. — Present geographic distribution of Extant Nut Beds genera. Abbreviations and data from Table 4.
status of this area as a refugium for once-widespread thermophilic genera that could not withstand the ef- fects of climatic cooling and glaciation at the end of the Tertiary in other parts of the northern hemisphere.
INTRODUCTION TO SYSTEMATIC PALEONTOLOGY
New species and genera have been named according to procedures required by the International Code of Botanical Nomenclature (Greuter, 1988). In order for a new species to be validly published, the new name must be accompanied by a description or diagnosis (art. 32.1). Fossil plants are excluded from the require- ment that the diagnosis be written in Latin (art. 36.1). A new genus and species may be created, and the names of a genus and a species may be simultaneously vali- dated, by provision ofa single description or diagnosis (art. 42.1). I have therefore given a combined generic and specific description for the many new, currently monotypic, genera presented here. Type and cited specimens are housed in the indicated museum col- lections. Specimens are cited as holotypes, paratypes, lectotypes and syntypes as defined in the Botanical Code of Nomenclature (Greuter, 1988). In addition, specimens that are not a part ofthe original description of species, but which are helpful in understanding the species may be designated hypotypes, as defined pre- viously in zoological (Schenk and McMasters, 1956) and paleobotanical (e.g., MacGinitie, 1953, p. 79) lit- erature.
In naming species of fruits and seeds in this work, Ihave tended toward conservatism, and have probably erred more on the part of lumping than splitting. By contrast, I consider that the classic monographs on the London Clay flora (e.g., Reid and Chandler, 1933; Chandler, 1978) reflect a greater tendancy for taxo- nomic splitting. An important objective has been to document the major taxonomic groups represented,
28 PALAEONTOGRAPHICA AMERICANA, NUMBER 58
thereby highlighting taxa available to the research com- munity for more detailed individual systematic inves- tigations. Many of the extant families and genera rec- ognized have never been analyzed by cladistic methodology, and may be in need of revision. In at- tempting to classify the entire fossil flora, however, it has been necessary to accept the taxonomic treatments currently available, rather than attempting to revise intrafamilial classifications.
Species are assigned to a modern genus only when sufficient diagnostic characters are present to rule out affınities with other extant genera and/or families. Oth- er species are assigned to extinct, form, or stereotype genera according to the criteria set out on pp. 17-18.
In paleobotanical literature, the generic name Car- polithus Schlotheim traditionally has been used as a catchall for fossil fruits and seeds of unknown system- atic affinities (Andrews, 1970). Different "species" of Carpolithus are often so different from one another that they probably represent different biological genera and families. I have given unique generic names to most of the form genera described from the Nut Beds but have placed an additional 21 taxa as unnamed species of Carpolithus. Thus, in comparing the generic diver- sity of different fossil localities, it is important to be mindful ofthe hidden diversity lurking under the name Carpolithus. It should also be noted that different spe- cies assigned to stereotype genera such as Laurocarpum and Anonaspermum may in reality represent different genera, but information from other plant organs would be required to determine the “real” generic diversity represented by such taxa.
Although leaves and woods co-occur with the fruits and seeds at the Nut Beds locality, no actual attach- ments have been found, and the high floral diversity combined with the large allochthonous component of the flora makes it difficult or impossible to determine which of the woods, leaves and reproductive structures might be conspecific. Therefore, I have refrained from assigning the same epithet to different isolated organs. It is conceptually intriguing to reconstruct fossil “whole- plant species” from their isolated organs; such recon- structions cannot be proven correct, however, unless there are specimens preserving direct attachment be- tween the organs. In some instances evidence of co- occurrence at numerous localities and shared modern familial affinity of isolated organs provides good evi- dence to infer that particular species of leaves and re- productive structures were perhaps produced by the same biological species, as in the case of the Clarno plane tree (Manchester, 1986). In the discussion ac- companying each family of fruit or seed recognized here, I indicate known records based on other organs
from the Clarno Formation, recognizing that some of them could in reality represent the same biological species.
ORGANIZATION
The descriptive systematic section is organized into two parts: The first part includes genera that can be confidently assigned to extant plant families. These are arranged alphabetically by family, treating the gym- nosperms first, followed by the angiosperms. The sec- ond part treats genera of uncertain familial affinity, ordered alphabetically, followed by a numerical se- quence of species informally attributed to Carpolithus. Families have been organized alphabetically, rather than phylogenetically, because there is more agreement among botanists as to the sequence of the alphabet than to the best phylogenetic arrangement. Phyloge- netic schemes are subject to continual change and re- finement and, while clearly important in expressing current views of relationships, are cumbersome and unstable as indexing systems. Table 5 presents a phy- logenetically arranged list of the Nut Beds fruit and seed genera following, with some modification, the ar- rangement of Cronquist (1981). Unless otherwise in- dicated, the numbers for species diversity and the geo- graphic ranges of extant genera were obtained from Mabberley, 1989.
TERMINOLOGY
Geometric symmetry, in particular reflectional point symmetry (Lockwood and MacMillan, 1978), has proven very useful in describing the three-dimensional shapes encountered among fruits and seeds. An object is said to be bilaterally symmetrical if there is a single plane of mirror symmetry. Following the same logic, I have used the term quadrilaterally symmetrical when there are two planes of mirror symmetry. These two planes usually intersect at right angles along the long axis of the fruit.
In describing the fossil seeds, I have generally used the term seed coat, rather than testa or tegmen, because the terms testa and tegmen have been used in a re- stricted sense that implies knowledge of ontogeny (Cor- ner, 1976; Schmid, 1986). Seed coat is a neutral de- scriptive term for the wall of the seed, whereas the terms tegmen and testa (sensu Corner, 1976) require an interpretation that is better left to discussion. Like- wise, for fruits, I have applied the term endocarp in а loose sense to refer to the hard, inner layer of the fruit, regardless of its ontogenetic derivation. It is this hard layer of the pericarp (fruit wall) that is most often preserved in the fossil specimens, although occasion- ally the meso- and exocarp are preserved. Cell layers
| | | |
| | |
CLARNO (EOCENE) FRUITS AND SEEDS: MANCHESTER 29
Table 5.—Systematic list of Nut Beds genera with known modern familial affinities. Angiosperms listed in phylogenetic sequence similar
to that of Cronquist (1981).
Gymnospermae Order Coniferales Family Taxaceae Genus Taxus Torreya Diploporus Family Pinaceae Genus Pinus Angiosperms Class Dicotyledonae Subclass Magnolididae Order Magnoliales Family Magnoliaceae Genus Magnolia, 3 spp. Family Annonaceae Genus Anonaspermum, 3 spp. Order Laurales Family Lauraceae Genus Lindera Laurocalyx Laurocarpum, 3 spp. Order Illiciales Family Schisandraceae Genus Schisandra Order Ranunculales Family Menispermaceae Genus Anamirta Atriaecarpum Calycocarpum Chandlera Curvitinospora Davisicarpum Diploclisia Eohypserpa Odontocaryoidea Palaeosinomenium Tinospora, 2 spp. Tinomiscoidea Thanikaimonia Family Sabiaceae Genus Meliosma, 5 spp. Sabia Subclass Hamamelidae Order Hamamelidales Family Platanaceae Genus Macginicarpa Platanus Tanyoplatanus Family Hamamelidaceae Genus Fortunearites Order Urticales Family Ulmaceae Genus Aphananthe Celtis, 2 spp. Trema Cedrelospermum Order Juglandales Family Juglandaceae Genus Juglans Cruciptera cf. Palaeocarya Paleoplatycarya? Order Fagales Family Fagaceae Genus Quercus Castanopsis Family Betulaceae Genus Coryloides Kardiasperma Subclass Dilleniidae Order Theales Family Actinidiaceae Genus Actinidia
Family Theaceae Genus Cleyera Order Violales Family Flacourtiaceae Genus Saxifragispermum Order Ebenales Family Sapotaceae Genus Bumelia?, 2 spp. Family Symplocaceae Genus Symplocos Subclass Rosidae Order Rosales Family Hydrangeaceae Genus Hydrangea Family Rosaceae Genus Prunus, 2 spp. Order Fabales Family Leguminosae Genus Leguminocarpon Order Myrtales Family Lythraceae Genus Decodon Order Cornales Family Alangiaceae Genus Alangium Family Cornaceae Genus Cornus Langtonia Mastixia Mastixioidiocarpum Mastixicarpum Nyssa, 3 spp. Order Celastrales Family Icacinaceae Genus Paleophytocrene, 2 spp. Iodes, 2 spp. Pyrenacantha Comicilabium lodicarpa, 2 spp. Order Rhamnales Family Vitaceae Genus Ampelocissus, 2 spp. Ampelopsis Parthenocissus, 2 spp. Vitis, 2 spp. Order Sapindales Family Staphyleaceae Genus Tapiscia Family Sapindaceae Genus Palaeoallophylus, 2 spp. Deviacer Family Burseraceae Genus Bursericarpum, 2 spp. Family Anacardiaceae Genus Pentoperculum Rhus Order Apiales Family Araliaceae Genus Paleopanax Subclass Asteridae Order Rubiales Family Rubiaceae Genus Emmenopterys Monocotyledons Subclass Arecidae Order Arecales Family Palmae Genus Sabal, 2 spp. Subclass Zingiberidae Order Zingiberales Family Musaceae Genus Ensete
30 PALAEONTOGRAPHICA AMERICANA, NUMBER 58
may be described as uni-, bi-, or multiseriate, indi- cating the thickness of the layer in numbers of cells. The orientation of cells relative to the seed or fruit surface morphology may also be important. Cells are oriented anticlinally, if their long axes are perpendic- ular to the trend of the adjacent external or internal surface. Cells are oriented periclinally if their long axes are parallel to the trend of the adjacent external or internal surface. Periclinally oriented cells may at the same time be horizontally or vertically oriented with respect to the principal axis of the fruit or seed.
MEASUREMENTS
If a species is known from three or fewer specimens, then the individual measurements of each are listed in the description. If more than three are measured, then the range, mean (avg.) and standard deviation (SD) are presented followed by the number of specimens mea- sured (n). Broken specimens vary in completeness and sometimes provide only one or two of the three spatial dimensions (length, width and/or thickness), hence, the number of measurement indicated for length may dif- fer from those given for width and/or thickness.
ABBREVIATIONS
Specimen numbers are prefixed by the appropriate institutional abbreviation, indicating where the spec- imen is deposited. The largest collections are at the US National Museum (USNM), Washington D.C. and the Florida Museum of Natural History, Gainesville (UF). Additional smaller collections were studied at the Uni- versity of California Museum of Paleontology, Berke- ley, California (UCMP), the University of Michigan Museum of Paleontology, Ann Arbor, Michigan (UM), Harvard University, Cambridge, Massachusetts (HU), the Burke Memorial Washington State Museum, Uni- versity of Washington, Seattle (UWBM), and the Or- egon Museum of Science and Industry, Portland (OMSI). Reference is also made to fossils from other localities deposited at the Natural History Museum, London (BM) and the Peabody Museum of Yale Uni- versity, New Haven (YPM) and to extant seeds from the National Seed Herbarium of the Beltsville Agri- cultural Research Center, Beltsville, Maryland (BARC), the herbaria of the Smithsonian Institution, Washing- ton, DC (US), Missouri Botanical Garden (MO), Rijk- sherbarium, Leiden (L), University of Florida (FLAS) and the Arnold (A) and Gray (GH) herbaria of Harvard University, Cambridge, Massachusetts.
The authors’ names for families, genera and species are presented at the first occurrence in the text, and may be abbreviated. Standard abbreviations for the
names of botanists are presented in Mabberley (1989, pp. 661—706).
In the plate captions, the abbreviations preceding the magnification refer to techniques used in photog- raphy. “РС”, or palladium-coated, is indicated for specimens that were coated to enhance reflected light microscopy (see materials and methods). “TL” refers to transmitted light microscopy, and ““RL” to reflected light micrography, while “SEM” denotes scanning electron microscopy. When not specified, specimens were photographed by reflected light, generally with the light source from the upper left and lower right.
Family PINACEAE Lindley Genus PINUS L.
Pinus sp. Plate 1, figures 6-9
Description.— Pollen cone elongate, with central axis 1.2 mm thick surrounded by numerous helically ar- ranged sporophylls, about 30 sporophylls per cycle; cone width 5.2 mm, length more than 12 mm (speci- men incomplete); lower part of cone with a bulge formed by large curved scales 4.5-5.5 mm long; sporophylis 1.5-1.8 mm long, thickened and upcurved distally; pollen sacs paired, abaxial on each sporophyll, 0.7-0.8 mm long.
Specimen.— UF 9334.
Discussion.— Although no seeds of Pinus have been identified from the Nut Beds, the genus is represented by wood fragments, dispersed pollen, and by a pollen cone. The pollen cone, described above, is known from a single specimen preserved in siltstone from the upper part of Face 1 of the Nut Beds (Pl. 1, fig. 6). Sectioned longitudinally (Pl. 1, figs. 7-9), the specimen shows the arrangement of microsporophylls typical of Pinaceae. The pollen itself is poorly preserved, possibly imma- ture. Similar pollen cones (Pl. 1, fig. 3), preserved as compressions, occur in association with foliage of a 5-needle pine and seeds of the articulate type (Pl. 1, figs. 4, 5) at the West Branch Creek locality of the Clarno Formation. Although the single specimen from the Nut Beds is broken so that the full length cannot be determined, comparison with the West Branch Creek specimens indicates a length of about 4 cm. Abraded pinaceous seed cones also occur in the Nut Beds (е.Р., Pl. 1, fig. 2), but because of poor preservation, they have not been identified to genus.
Family TAXACEAE Gray
The Taxaceae, or yew family, has five extant genera and about 20 species of trees and shrubs (Price, 1989). Although the family is well documented by foliage of
| |
CLARNO (EOCENE) FRUITS AND SEEDS: MANCHESTER 31
various genera in the Tertiary of Europe and North America, reports based on seeds are rare. Three genera of taxaceous seeds occur in the Nut Beds: Taxus, Tor- reya, and a new extinct genus, Diploporus. A single taxaceous leaf, with cuticle corresponding to that of Torreya, was recovered from the White Cliffs locality of the Clarno Formation. The leaf is noteworthy for its large size: 70 mm long, 7 mm wide.
Genus TAXUS L.
Taxus masonii sp. nov. Plate 2, figures 1-11
Etymology.—This species is named for Herbert L. Mason, in recognition of his pioneering investigations of fossil conifers from western North America.
Description.—Seed bilaterally symmetrical, lentic- ular in cross-section, nearly elliptical in face view, base truncate, apex keeled; length 4.3-6.2, avg. 5.3 mm (SD=0.53, n=10), width across plane of symmetry 4.5— 5.6, avg. 4.9 mm (SD=0.34, n=10), thickness 3.3-4.0, avg. 3.7 mm (SD=0.28, n=10), length/width ratio 1.0- 1.1; basal truncation perpendicular to central axis of Seed, elliptical with a pair of small circular vascular Scars; apical keel slightly arched as viewed from above, with a central micropylar scar; seed coat about 250- 300 um thick, composed of isodiametric cells 25-60 um in diameter; two vascular strands, one along each Side in the dorsiventral plane, arising from the basal truncation and passing into the seed cavity in the lower % of the seed.
Specimens.— Holotype: USNM 355474. Paratypes: UCMP 10622, UF 6643, 6644, 9474—9478, 9801,9802, USNM 354997, 354998, 355000, 355406.
. Discussion.— Taxus is represented by permineral- 12е4 seeds from the Nut Beds that are readily recog- nized by their lensoid cross section, keeled apex and truncate base (Pl. 2, figs. 1-8). The pointed end rep- Iesents the micropyle, and the basal truncation cor- Tesponds to attachment of the aril with the seed. The Pair of vascular scars on the basal truncation (РИ 2 figs, 7, 8) are diagnostic for the genus. Although extant
ахиѕ seeds may vary from two- to three-fold sym- Metry, these fossil seeds are nearly all bilaterally sym- Metrical. The morphology and vascular supply of the Seed, as seen in serial transverse sections (Pl. 2, figs. 9-11), is very similar to that observed in extant species (Dupler, 1920).
Taxus grows today in central Malesia and Mexico 45 well as in north temperate areas. The genus has also been confirmed from the Oligocene and Miocene of eastern Europe through careful study of cuticularly pre- Served leaves (Куасек, 1982).
Genus TORREYA Arn.
Torreya clarnensis sp. nov. Plate 2, figures 12-16; Text-figure 8A
Etymology.— The epithet clarnensis refers to the Clarno Formation, source of the specimens.
Description.—Seed bilaterally symmetrical, obovate in face view, fusiform in lateral view, ellipticalin trans- verse section; length 19.5-20.8 mm, width 10.7-11.3 mm, thickness across the dorsiventral plane 9.5-9.6 mm, length/width ratio 1.7-1.9; base acute-rounded, apex rounded 1n face view, acute in lateral view, with the dorsal and ventral faces meeting to form a keel in the apical % of the seed; micropylar scar at the mid- point of the apical crest; two circular vascular scars situated in the dorsiventral plane, one on either side of the seed; the vascular scars positioned at the level of greatest seed width, about Y of the length from the apex.
Specimens.— Holotype: UF 6510. Paratypes: USNM 353966, 422380.
Discussion.— Torreya clarnensis is represented only by a few specimens but they present enough characters to establish firmly the affinities with extant Torreya. The apical keel, the pair of pores (vascular scars) pierc- ing the seed coat near the apex in the dorsiventral plane (Pl. 2, fig. 13), and the narrowed, acute-rounded base are characters diagnostic of the extant genus. The pair of vascular pores are closer to the apex (Text-fig. 8A) than they are in Diploporus (Text-fig. 8D), but not as close to the apex as in the extant species of Torreya (Text-figs. 8B, C).
Torreya is a genus of trees with five extant species distributed in North America and eastern Asia. The North American species are restricted to a few sites in California and northwestern Florida. The Asian spe- cies are limited to an east-west belt extending from Honshu over the Yang-tse valley to eastern Burma. The genus has also been recognized on the basis of upon fossil foliage from the Late Eocene Florissant and Oligocene Bridge Creek floras (MacGinitie, 1953), and from Oligocene and Miocene localities Europe (Kva- сек, 1982). E
Genus Diploporus gen. nov.
Etymology.—Diplo (Gr = twofold) + poros (Gr = hole, passage) referring to the pair of vascular scars in the seed coat.
Generic diagnosis.—Seed subovoid, base rounded, apex wedge-shaped; apical keel slightly arched as viewed from above, with a central micropylar scar; seed coat
32 PALAEONTOGRAPHICA AMERICANA, NUMBER 58
Text-figure 8.—Seeds of Taxaceae in face view and lateral view, showing apical keel and variable position of the pair of vascular pores in the seed coat. A, Torreya clarnensis sp. nov., with pores positioned about of seed length below apex. B, T. grandis Fort. from Kuling, China, showing pointed base, apical pores. C, T. taxifolia Arn. from northwestern Florida showing apical pores. D, Diploporus torreyoides sp. nov., showing apical keel, pair of pores in the lower half, rounded base. Scale bar = | cm.
pierced by a pair of circular vascular scars, one on each side in the dorsiventral plane at, or slightly below, the equator.
Type species.— Diploporus torreyoides sp. nov.
. Diploporus torreyoides sp. nov. Plate 3, figures 1-16; Text-figure 8D
Etymology.— The epithet refers to similarity with seeds of extant Torreya.
Description.—Seed bilaterally symmetrical, subo- void, ovate to circular in face view, pyriform in lateral view, wide-elliptical in transverse section; length 6.7- 8.0, avg. 7.9 mm (SD=0.5, n=25), width 5.4-8.0, avg. 6.9 mm (SD=0.6, n=25), thickness 3.6-7.0, avg. 5.8 mm (SD=0.78, n=25), length/width ratio 1.0-1.3; base
rounded; apex rounded in face view, angled in lateral view, with the dorsal and ventral faces meeting to form a sharp crest in the apical У of the seed with a central micropylar scar; two circular vascular pores situated in the dorsiventral plane, one on either side ofthe seed, positioned at the level of greatest seed width, about Y to % of the length from the apex; aril preserved in some specimens, completely covering seed, 1.0-1.5 mm thick, outer surface of aril with numerous short longitudinal grooves; seed coat smooth; embryo straight, 1.0 mm in diameter, 6.4 mm long, aligned with the central axis of the seed.
Specimens.— Holotype: UF 8542. Paratypes: UE 8501-8541, USNM 355072 (2 specimens), 355471, 355479, 355481 (9 specimens), 355489 (17 speci- mens), 424744, 424745, 435098, OMSI Pb217,
| | |
CLARNO (EOCENE) FRUITS AND SEEDS: MANCHESTER 33
Pb1348, internal casts; UF 6531, 9587, 9803, 9804, USNM 355569, 446078, specimens with intact seed coat and aril.
Discussion.—Represented by more than 75 speci- mens, this species is among the more common fossils from the Nut Beds, and is particularly abundant near the top of Face 3. Most specimens are smooth-sur- faced, representing internal molds of the seed coat (Pl. 3, figs. 9-12); a few possess an additional outer layer about one mm thick with a wrinkly external surface (Pl. 3, figs. 1-8). This outer layer is poorly preserved anatomically but apparently represents the aril. Salient features of these fossils are the rounded base, keeled apex and the pair of conspicuous pores, one on either side in the dorsiventral plane (Pl. 3, figs. 6, 8, 11—13). Longitudinal and transverse sections reveal a slender cylindrical embryo (Pl. 3, figs. 14-16) that is medial within the seed.
Diploporus is interpreted as an extinct genus of Tax- aceae based upon several important similarities with Taxus and Torreya, including the pair of vascular bun- dles, the central linear embryo and the apical keel that is slightly arched as viewed from above, with a central micropylar scar. The seed base is rounded, however — Dot truncate as in Taxus, and not pointed as in Torreya. The tissue surrounding the embryo within the seed Cavity is cellularly preserved in many of the specimens, and has a speckled appearance (Pl. 3, figs. 14-16) like that of Torreya. Diploporus seeds are larger than those of Taxus and smaller than those of Torreya. The pair of vascular pores in the dorsiventral plane corresponds to those seen in Torreya, but in the fossil these scars are positioned at or slightly below the equator, rather than near the apex (Text-fig. 8).
Seeds of this morphology have not, to my knowl- edge, been described from the Tertiary before, and it is of some interest to confirm the presence of extinct Taxaceae surviving into the Tertiary. As this manu- Script was going to press, I noticed an apparent close Similarity of this genus to the Cretaceous genus Vesquia Bertrand from the Wealden Formation of Belgium (Al- Vin, 1960). Although I have not seen specimens of Vesquia, the genus was described in good detail and illustrated with line drawings (on the basis of two well Preserved liginfied seeds) by Alvin (1960). Vesquia seeds are similar in size and shape to those Diploporus, with а corresponding pair of pores (“two formamina") in the lower half. However, the specimens of Vesquia are More globose and appear to lack an apical crest. In addition, the basal surface of the Vesquia seeds have 4 pair of grooves leading to the foramina, which are Not seen in Diploporus. Vesquia is also considered to belong to the Taxaceae and, among extant genera, com-
pares most closely to Torreya (Alvin, 1960). It is pos- sible that additional work may reveal that the Clarno taxon is a hold-over from the Mesozoic. Indeed, both of the living genera, Torreya and Taxus, are believed to extend back to the Jurassic (Florin, 1958; Harris 1976).
Family ACTINIDIACEAE Hutch.
This family includes four modern genera: Clema- toclethra, Sladenia, Saurauria and Actinidia. The lat- ter, which provides the kiwi fruit of commerce, is the best known. Saurauria has been reported on the basis of leaves from the middle Eocene of Alaska (Wolfe, 1977) and Wyoming (MacGinitie, 1974). Seeds as- signed to the fossil genus Pasternackia (described, p. 104) are similar to those of Saurauria, although affin- ities remain uncertain. The following constitutes the first recognition of Actinidia from the North American fossil record.
Genus ACTINIDIA Lindl.
Actinidia oregonensis sp. nov. Plate 4, figures 1—4
Etymology.—The epithet refers to the state of Ore- gon.
Description.—Seed bilaterally symmetrical, anatro- pous, elongate-elliptic in face view, laterally flattened, keeled in the plane of symmetry, lensoid in cross sec- tion; length 3.2-4.4, avg. 3.8 mm (SD=0.46, n=5), width 2.2-3.0, avg. 2.5 mm (SD=0.29, n=7), thickness 1.0-1.1, avg. 1.1 mm (SD=0.05, n=7); base with ob- tusely pointed micropyle and laterally adjacent hilum, apex rounded; seed coat ca. 0.1 mm thick, surface re- ticulate, made up of thick-walled, polygonal (five- to seven-sided), isodiametric cells 190 to 240 um de- creasing in size toward the raphe and antiraphal edges of the seed; 15—20 cells wide in face view.
Specimens.— Holotype: UF 6292. Paratypes: UF 6293, 6294, 6498, 8580, 9218, USNM 355370.
Discussion. — This species is represented by chalced- ony seed casts. The distinctive reticulate surface of these fossil seeds (Pl. 4, figs. 1-4) matches that in extant Actinidia seeds, where the outer layer of the seed coat is formed by polygonal cells with strongly thickened lateral walls. This surface ornamentation, combined with seed shape, and position of the hilum and micro- pyle, confirms the identification of these fossils as Ac- tinidia.
Actinidia is a genus of about 30 species of climbers distributed in Indomalesia and eastern Asia. Although this is the first report of fossil Actinidia from North
34 PALAEONTOGRAPHICA AMERICANA, NUMBER 58
B
Text-figure 9.—Transverse sections of Alangium endocarps from the Nut Beds. A, A. eydei sp. nov., showing relatively thin endocarp wall, and one locule somewhat larger than the other. B, A. rotun- dicarpum sp. nov., showing thick walls, and locules equal in size.
America, seeds of the genus occur from the upper Oli- gocene to Pliocene of Europe (Kirchheimer, 1957; Do- rofeev, 1963; Mai and Walther, 1978; Friis, 1985).
Actinidia oregonensis seeds are large in comparison to the four extant species that I have seen (A. arguta Franch. et Sav., A. callosa Lindl., A. chinensis Planch., A. strigosa Hook. f. et Thoms.). Whereas the fossils average about 4 mm in length, the extant species ex- amined, and most of the European fossil specimens, are about 2 mm long. Actinidia seeds from the Pliocene of the Netherlands, however, range from 1.8 to 4mm long (Kirchheimer, 1957). Kirchheimer (1957) placed all of the European specimens in a single fossil species, A. foveolata Reid and Reid, but Friis (1985) stated that the type material of A. foveolata is larger than the Dan- ish and Polish material and has a greater number of cell rows across the width of the seed.
Family ALANGIACEAE DC.
The Alangiaceae are a tropical family with one genus and about 20 living species in Africa and from China to Australia. Alangium grows as a tree, less often as shrubs or lianas. The genus has an excellent fossil rec- ord in Europe and North America (Eyde, Bartlett and Barghoorn, 1969; Mai, 1970). Although allied to the Cornaceae, and placed in the Cornales by Cronquist (1981), Alangium differs by the presence of laticifers and alkaloids (Eyde, 1988). The genus is divided into four sections that are distinguished on the basis of style, stigma, stamen, floral vascular system and endocarp characters (Eyde, 1968). The paleogeographic history of Alangium is reviewed by Kriitzsch (1989).
The Nut Beds flora includes two species of Alangium based upon fruits. Silicified wood of this genus is also known from the Nut Beds (Scott and Wheeler, 1982).
Based upon the presence of scalariform perforation plates, small intervascular pits and long vessel ele- ments, Scott and Wheeler (1982) determined that Alangium oregonensis wood conforms to Metcalfe and Chalk's (1950) “group A” (A. javanicum type), and differs from “group В” (A. chinense type). This obser- vation is of interest because the fruit morphology of the species considered below corresponds closely to that of extant A. chinense Rehder.
Genus ALANGIUM Lam.
Alangium eydei sp. nov. Plate 4, figures 5-14, Text-figure 9A
Etymology.—This species is named for the late Richard Eyde, in recognition of his contributions in understanding the morphology of living and fossil Alangium and other Cornales.
Description.—Endocarp bilocular, ovate in face view, laterally compressed in the plane of septum; base rounded, apex obtuse-pointed; length 4.9-10.3, avg. 7.7 mm (SD=1.7, n=13), width 5.1-7.6, avg. 6.3 mm (SD=0.75, n=11), thickness 3.2-6.4, avg. 4.3 mm (SD=0.75, n=15); one locule slightly smaller than the other; each locule with a single seed; endocarp surface relatively smooth with faint broad ridges; position of the septum marked on the exterior of the endocarp by a peripheral furrow between the two carpels; septum ca. 0.4 mm thick, containing six to eight canals that radiate in the median plane and terminate before reaching the edge of the endocarp, forming an asterisk pattern; endocarp wall ca. 0.5 mm thick, made up of isodiametric polygonal sclereids 17-30 um; locule lin- ing uniseriate, 17.5 um thick, composed of cuboidal sclereids.
Specimens.— Holotype: USNM 424646, endocarp. Paratypes: HU 60003, 60004, UCMP 10710, UF 6311, 6508, 6645, 6646, USNM 354012-354015, 354016 (2 specimens), 354018, 354019, 354045, 422393, 424645, 446073, endocarps; UF 9479, USNM 354020, septal canal casts; USNM 354017 (20 specimens), isolated locule casts.
Discussion.— Alangium eydei is represented in the Nut Beds by occasional endocarp casts (Pl. 4, figs. 5- 8) and permineralizations (PI. 4, figs. 11—12). The latter may be sectioned to reveal internal morphology and anatomy (Pl. 4, fig. 13). Frequently all that remains of the endocarp is a carbonaceous powder that disinte- grates in the field, leaving the two chalcedony locule casts side by side, resembling a sandwich (Pl. 4, figs- 9, 10). Sometimes star-shaped chalcedony casts of the canals within the septum are recovered (Pl. 4, fig. 14) but the most common remains of Alangium in the Nut
| | | |
CLARNO (EOCENE) FRUITS AND SEEDS: MANCHESTER 35
Beds collections are isolated silica locule casts, pre- served as thin, elliptical wafers with a faint stellate impression on one side corresponding to the septal canals.
This species, like most of those known from the Tertiary (Eyde, Bartlett and Barghoorn, 1969; Mai 1970), corresponds to section Marlea of extant Alan- gium in having a bilocular, thick-walled endocarp. Eyde (1968) noted that the extant species of this section can be arranged in a sequence to show the gradual loss of one ofthe carpels. In Alangium kurzii Craib the locules are more or less equal in size, whereas in A. platani- Јоћит (8. et Z.) Harms, A. alpinum W. Smith et Cave and A. rotundifolium (Hassk.) Bloemb. one of the car- ре! is much smaller than the other, often with a tightly closed abortive locule, and in A. griffithii, only one Carpel is evident. The fossil closely resembles A. chi- nense Rehder in having one locule smaller than the Other (Text-fig. 9A) , but with both locules fertile. Eyde (1968, fig. 3h) illustrated a transverse section of an endocarp of extant A. chinense that is almost identical to that seen A. eydei (Pl. 4, fig. 13). The smaller of the two carpels generally is at least 80% of the size of the larger carpel both in the fossil and in A. chinense.
Alangium eydei bears a close resemblance to A. jen- kensii Chandler (1961b) from the London Clay, al- though the largest of the Nut Beds specimens is slightly Smaller than the single London Clay specimen. Al- though A. jenkinsii clearly belongs to Alangium, Chan- dler (1961b) noted that “owing to the poor condition Of the single specimen it is difficult to define specifi- cally." A. vermontanum Eyde et Barghoorn from the Oligocene Brandon Lignite of Vermont (Eyde, Bartlett and Barghoorn, 1969) differs from А. eydei in having a much more reduced smaller locule, and in the retic- ulate sculpture of Ше ventral side of the larger locule. The Clarno fossil differs from 4. dubium (Unger) Mai from the Miocene of central Europe (Mai, 1970, p. 474) in the lack of transverse ridges on the endocarp Surface.
Alangium rotundicarpum sp. nov. Plate 4, figures 15, 16, Text-figure 9B
Etymology.— Когипаи (L = circular, round), + kar- Роз (Gr = fruit), referring to the rounded outline ге- Sulting from thick wall of the endocarp.
Description.—Endocarp subglobose, bilocular; length 8.5 mm, width 7.2 mm, thickness 7.7 mm; locules approximately equal in size, each with a single seed; Position of the septum marked at the endocarp surface bya peripheral furrow between the two carpels; septum 0.5 mm thick, containing six to eight canals that radiate In the median plane and terminate before reaching the
edge of the endocarp, forming an asterisk pattern; en- docarp wall 2.75 mm thick dorsally and ventrally, thin- ner laterally, made up of isodiametric to anticlinally elongate sclereids, 40-150 um in diameter; locule lin- ing not well preserved, 30 um thick.
Specimen.— Holotype: USNM 435005.
Discussion.— This species, known from a single per- mineralized specimen, differs from A. eydei in its glo- bose endocarp (Pl. 4, fig. 15) with thick lateral walls and in the equivalent size of the two locules (Text-fig. 9B). Although the exterior surface of the endocarp is not preserved, the intact portion is about five times thicker than the complete wall of A. eydei. Transverse sectioning revealed that the locules of A. rotundicar- pum are approximately equal in size (Pl. 4, fig. 16), in contrast with the unequal development of locules in A. eydei (Pl. 4, fig. 13). The endocarp includes anti- clinally elongate sclereids, whereas those of A. eydei are more or less isodiametric. In size of the endocarp and unusually thick endocarp wall, this species resem- bles A. krutzschii Mai from the middle Eocene Gei- seltal flora of Germany (Mai, 1970); the locules of that species, however, are much smaller.
Family ANACARDIACEAE Lindley
The Anacardiaceae, or sumac family, includes about 70 extant genera of tropical to temperate distribution including trees, shrubs and lianas, and is divided into five tribes: Anacardieae (eight genera), Spondiadeae (17 genera) Rhoeae (40 genera), Semecarpeae (five genera), Dobineeae (one genus) The Spondiadeae, characterized by drupes with hard, often multilocular, stones having oval germination apertures, is well rep- resented in the fossil record, with records of both extant and extinct genera in the early Tertiary of Europe (Reid and Chandler, 1933; Collinson, 1983) and North America.
The Anacardiaceae are represented in the Nut Beds flora by at least two genera of fruits: Rhus, of the tribe Rhoeae, and the extinct genus Pentoperculum of the tribe Spondiadeae. In addition, the fruit described herein as Pistachioides (p. 105) resembles that of extant Pistachia, and might also belong to the Anacardiaceae. Anacardiaceous woods from the Nut Beds include Tapirira (tribe Spondiadeae; Manchester, 1977), and at least one additional genus remaining to be described (S. R. Manchester, unpublished).
Genus PENTOPERCULUM gen. nov.
Etymology.— Penta (Gr = five) + operculum (L = lid), referring to the five germination valves ofthe fruit.
Type species.—Pentoperculum minimus (Reid et Chandler), comb. nov.
36 PALAEONTOGRAPHICA AMERICANA, NUMBER 58
Pentoperculum minimus (Reid et Chandler) comb. nov. Plate 5 figures 1-17
Dracontomelon minimus Reid et Chandler 1933, p. 302, pl. 13, figs. 20-24.
Description.—Endocarp subglobose to more or less oblate, five-(rarely four- or six-) carpellate, roundly pentangular (rarely quadrangular or hexagonal) in transverse section, length 4.4-6.5, avg. 5.7 mm (SD=0.49, n=23), equatorial diameter 4.5-8.5, avg. 6.8 mm (SD=0.89, n=24), length/equatorial diameter ratio 0.74-1.11, avg. 0.90, widest at the equator; base round to flat with a slight depression ca. 1 mm in diameter at the extreme base, apex rounded; locules five (rarely four or six), single-seeded, radially ar- ranged, vertically elongate, 3.5-5.0 mm long, 1.5-2.1 mm wide, 2.1—3.3 mm in dorsiventral dimension; each locule with a convex, dorsal-apical, germination valve, resulting in five (rarely four or six) elliptical bulges at the surface of the endocarp extending from the equator almost to the apex; germination valves 440 um thick, two-partite with a median longitudinal slit of shutter- like splitting; lacunae alternating with the locules at the equator of the fruit corresponding in position to the exterior angles of the fruit; these lacunae about as wide as the locules, but shorter both dorsiventrally and axially; central axis of the fruit with a cylinder of vas- cular tissue extending from base spreading and dissi- pating apically; endocarp wall and septa composed of isodiametric sclereids 28-45 um in diameter with in- terspersed tracts of fibers; locule lining 25-30 um thick, composed of horizontally aligned fibers, these fibers 7 um in diameter and at least 10 um long.
Specimens.—UCMP 10633, UF 5708-5711, 6608- 6610, 10634, UM 31604, USNM 312762 (Bones, 1979, pl. 3, fig. 6), 353967 (10 specimens), 353969, 422390, 424851, 435106-435114, endocarps; UCMP 10635- 10637, UF 6664-6671, 9317, 9736 (100 specimens), USNM 435115 (7 specimens), isolated locule casts. Holotype: BM v22546 from the London Clay.
Discussion. —Specimens of Pentoperculum are pre- served in the Nut Beds as white chalcedony casts of the endocarp and as permineralized endocarps with intact seeds. Most specimens are pentalocular (e.g., Pl. 5, figs. 1, 2, 8), although two specimens are four-loculed (e.g., Pl. 5, fig. 15) and three are six-loculed (USNM 435107-435109). In some cases the endocarp has dis- integrated leaving a cavity in the matrix with protrud- ing chalcedony locule casts. Isolated locule casts are common (Pl. 5, figs. 5, 6). Because the fossil does not correspond precisely to any living genus, the new ge- neric name Pentoperculum is established to accom- modate this species. Pentoperculum minimus was a
widespread species that occurred in the Eocene of Hun- gary (Kovacs, 1957), as well as England (Reid and Chandler, 1933) and western North America.
This species was first recognized on the basis of pyr- itized endocarps and locule casts from the London Clay flora (Reid and Chandler, 1933), an example of which is included in Pl. 5, fig. 14. The Nut Beds specimens are indistinguishable both in external and internal morphology from the type material. Reid and Chan- dler (1933) demonstrated similarities to the extant ge- nus Dracontomelon, and their work confirmed affini- ties with the tribe Spondiadeae. However, new observations based upon sections of the fossil and ex- tant endocarps indicate significant differences between Pentoperculum and Dracontomelon. Contrary to the interpretation of Reid and Chandler (1933), serial sec- tions (e.g., Pl. 5, figs. 9, 10) indicated that the lacunae are not linked to peripheral apertures in the fossil, a condition considered diagnostic of the extant genus. The fossil fruits are more symmetrical than those of Dracontomelon, with more or less equal development of locules. In addition, the fossil endocarps are much smaller than those of the living species of Draconto- melon. In contrast to Dracontomelon, in which each of the locules opens by the “uncorking” of an undi- vided apical plug (Hill, 1933), the locules of Pento- perculum open by means of a bipartite operculum (Pl. 5, fig. 8).
Various modes of germination occur in the fruits of the Spondieae and are described in detail by Hill (1933, 1937). The apertures of Pentoperculum appear to have had shutter-like valves similar to those observed in Haematostaphis (Hill, 1933) and Pseudospondias (Hill, 1937). In these genera, “the lid separates into two equal halves which are pushed apart and thrown aside by the radicle as it emerges” (Hill, 1933, pp. 881—882).
Genus RHUS L.
Rhus rooseae sp. nov. Plate 6, figures 1—8
Etymology.— This species is named after Carrie L. Roose, recognizing her helpful assistance and support of this project.
Description.—Fruit bilaterally symmetrical, subel- liptic, unilocular, laterally flattened in the plane of symmetry, transversely elliptic in cross section, base and apex obtuse-rounded; width greater than height; height 3.3-4.3, avg. 3.8 mm (SD=0.29, n=8), width 4.2-5.1, avg. 4.7 mm (SD=0.4, n=5), thickness 1.5- 2.0, avg. 1.8 mm (SD=0.25, n=9); mesocarp with 25- 30 prominent longitudinal vascular bundles; endocarp surface smooth, shiny, with a circular funicular de- pression in the plane of symmetry, and a slight bulge
CLARNO (ЕОСЕМЕ) FRUITS AND SEEDS: MANCHESTER 37
on the opposite side; endocarp wall 400 um thick, com- posed of three layers: inner layer 250 um thick, uni- Seriate, composed of anticlinally oriented columnar cells 250 um high and 20-30 um wide; the next layer much thinner, also uniseriate, 33 um thick, composed of rectangular cells typically 45 um wide; the overlying layer uniseriate, composed of columnar cells similar to the first, but only 80 um thick, capped by a uniseriate to triseriate layer of rectangular cells.
Specimens.— Holotype: USNM 355036. Paratypes: UF 9350-9352, 9417-9422.
Discussion.—Sumac fruits, Rhus rooseae, typically are preserved in the Nut Beds as translucent, silicified endocarps. The prominent columnar cell construction is exhibited both in fractured specimens by SEM (PI. 6, figs. 5, 6) and with transmitted light in thin sections (Pl. 6, fig. 8). Only one specimen, designated the ho- lotype, has the mesocarp intact showing the prominent longitudinal vascular bundles (Pl. 6, figs. 1-3).
Fruit and seed morphology of Rhus are reviewed by von Teichman and Robbertse (1986). R. rooseae shares many characters with extant Rhus including lateral compression in the plane of bisymmetry, width greater than height, longitudinal vascular bundles forming ribs in the mesocarp, and endocarp comprised of two or more layers of anticlinally elongate columnar cells. Al- though Rhus has been reported commonly on the basis Of leaves from the North American Tertiary, this is the first North American report based upon fruits. The genus has also been documented on the basis of fruits Similar to extant Rhus toxicodendron L. from the Mid- dle Miocene of Denmark (Friis, 1979). The Danish fossil differs from R. rooseae in possessing only about 12, rather than 25-30, longitudinal ribs.
Family ANNONACEAE Juss.
The Annonaceae are a family of trees, shrubs and lianas with about 128 living genera and 2050 species. The family is mostly tropical, with the exception of few genera that extend into temperate regions. Three Species of Annonaceae can be identified from the Nut Beds on the basis of internal molds of the seed coat, Showing the characteristic ruminate endosperm.
Genus ANONASPERMUM Ball emend. Reid et Chandler
Reid and Chandler (1933) emended the fossil genus Anonaspermum Ball to accommodate anatropous seeds with ruminate endosperm of the various types found Ш the family Annonaceae, and with an encircling band Of fibers formed by the raphe and chalaza, which di- Vides the seed into more or less symmetrical halves, With the hilum terminal. Most of the fossils that have been attributed to this genus are internal molds of the
seed coat, or endosperm casts, with the seed coat miss- ing. Distinctive surface patterns result from the con- figuration of the ruminate endosperm.
The Clarno specimens of Anonaspermum have been segregated into three species using criteria similar to those used in distinguishing London Clay species. Fourteen species of Anonaspermum were described from the London Clay (Reid and Chandler, 1933; Chandler, 1961b; summarized, Collinson, 1983), based upon differences in seed shape, pattern of ridges in endosperm, and nature of the encircling fiber band. As noted by Collinson (1983, pp. 73-74), “Many different fossil forms may be recognized but it is very difficult to relate these to living genera. Similar endosperm pat- terns may occur in different living genera and different endosperm patterns may occur within the same living genus”. It is possible that the three Nut Beds species recognized here were actually produced by different genera, but because of the limited systematic resolution provided by seed characters in this family, they are all be placed in the stereotypic genus Anonaspermum.
Anonaspermum cf. pulchrum Reid et Chandler Plate 6, figures 9-15
Description. —Seed bilaterally symmetrical, short- to long-ellipsoid, all faces convex rounded, sometimes slightly flattened laterally in the plane of symmetry; length 5.5-10.3, avg. 8.0 mm (SD=1.47, n=10), width 3.8—6.2, avg. 5.2 mm (SD=0.85, n=12), thickness 3.4— 4.9, avg. 4.3 mm (SD=0.42, n=12); without a marked longitudinal median depression; internal mold of seed coat with rumination ridges many, narrow, transverse, interrupted in course; these ridges mainly continuous across the center, without forming pits, nodules or punctae; sometimes with short grooves along the ridge crests; ruminations tetrapartite in transverse section; with a strap-like ridge 1-1.4 mm wide in the plane of symmetry formed by fibers of the raphe and chalaza; hilum terminal at the broader end of the raphe ridge; seed coat thin (80-100 um), smooth.
Specimens.— HU 60061, UCMP 10676, 10677, UF 5233, 5234, 6574, USNM 353971 (7 specimens), 353972, 353975, 353976, 424794-424797, 434983.
Discussion. — This species is represented in the Nut Beds mostly by internal molds of the seed coat (i.e., endosperm casts). In one specimen (UCMP 10677) part of the thin seed coat is adhering. In seed shape, size and pattern of endosperm ruminations, this spe- cies corresponds closely to Anonaspermum pulchrum Reid et Chandler from the London Clay flora. Reid and Chandler (1933) described the raphe and chalaza as lying in a rather inconspicuous marginal groove, but in most of the Clarno specimens these form a ridge (Pl. 6, figs. 10, 13). Perhaps this difference is due to
38 PALAEONTOGRAPHICA AMERICANA, NUMBER 58
abrasion of the London Clay specimens resulting in loss of the fibrous band. In some of the Nut Beds specimens, the fibrous band is partially or entirely bro- ken away (e.g., Pl. 6, fig. 15) revealing a marginal groove like that of the London Clay species. Reid and Chan- dler (1933) called attention to similarity with the living Indomalaysian species Dasymaschalon clusiflorum Merrill, but refrained from suggesting a close relation- ship because of the likelihood that similar seeds could be found elsewhere in the family with a more exhaus- tive search of modern genera.
Anonaspermum bonesii sp. nov. Plate 6, figures 16-18
Etymology.— This species is named for the late Tho- mas J. Bones, who collected most of the Anonasper- mum specimens known from the Nut Beds.
Description.—Seed roughly ellipsoidal, pointed api- cally, with a slight flattening near the hilum, rounded basally, bilaterally symmetrical, laterally compressed in the plane of symmetry such that the thickness is less than Y of the width, with some shallow depressions on the flattened face; length 11.7 mm, width 7.4 mm, thickness 2.5 mm; internal mold of seed coat with ruminations transverse in the midsection of seed, but radially oriented near the base and apex, thin, closely spaced, irregular, disappearing toward the center of the face, giving rise to fine punctae; inner layer of seed coat with a criss-cross pattern of straight to somewhat curved fibers.
Specimen.— Holotype: USNM 353975.
Discussion.—This species, known from a single si- licified seed, differs from A. cf. pulchrum in its pro- nounced lateral compression (Pl. 6, fig. 18), such that the thickness is less than % of the width, and by the orientation of ruminations. Although the ruminations are transversely oriented across most of the seed in A. cf. pulchrum (Р1. 6, figs. 9, 10, 14), they are radially oriented in the apical and basal 3 of the A. bonesii seed (Pl. 6, figs. 15, 17).
Anonaspermum rotundum sp. nov. Ра е Stress 2
Etymology.— Rotundus (L = circular, round) refer- ring to the seed shape.
Description.—Seed wide-ellipsoidal, rounded api- cally and basally, laterally compressed in plane of the raphe; length 13.5 mm, width 12.0 mm, thickness 6.0 mm; internal mold of seed coat with ruminations ra- dially oriented from the center, thin, short, irregular, shorter toward the center; raphe band extending nearly 350° around periphery of the seed.
Specimen.— Holotype: USNM 435099.
Discussion.—A. rotundatum is known from a single
specimen which is preserved as an internal mold of the seed coat. The specimen differs from those of the other two Clarno species by its larger size and nearly circular outline.
Family ARALIACEAE Juss.
The Araliaceae, as traditionally circumscribed, are a family of about 57 genera and 800 species including trees, shrubs and lianas mostly of tropical distribution. The more cosmopolitan, mostly herbaceous Apiaceae (Umbeliferae) with 418 genera and 3100 species, are very similar and should be included within Araliaceae to have a monophyletic family (W. Judd, pers. comm., 1991; and see Mabberley, 1989).
Genus PALEOPANAX gen. nov.
Etymology.— Paleo (Gr = old) + panax (L, Gr = ginseng) referring to the ginseng family, Araliaceae. Type species.— Paleopanax oregonensis sp. nov.
Paleopanax oregonensis sp. nov. Plate 7, figures 3-4
Description.— Fruit wide-elliptical in face view, bi- carpellate, bilaterally symmetrical, dorsiventrally compressed, with an epigynous perianth bulge; base flat or somewhat cordate, apex rounded; length 4.5- 5.6 mm, width 5.4—6.6 mm, estimated thickness 1—2 mm; carpels D-shaped in face view, with two to three arched longitudinal grooves, adjoined ventrally at the central fruit axis by their straight margins, outer mar- gins convex, smooth; pedicel > 2.0 mm long, 70.3 mm thick; two styles arising parallel to each other from the apex, 2.0-3.0 mm long, recurving distally.
Specimens.— Holotype: OMSI Pb810. Paratypes: OMSI Pb1779, Pb1014.
Discussion.— This species, represented only by com- pression fossils, was recovered from the leaf horizon at the base of Face 3 in the Nut Beds. The preservation of the styles indicates that they were tough and per- sistent, or that the fruits were deposited without rig- orous transport.
Fruits of Paleopanax conform to those of the Apiales (Araliaceae + Apiaceae) in the possession of epigynous perianth, styles as many as the carpels and schizocarp morphology. Although samaroid schizocarps also oc- cur in Sapindaceae and Aceraceae, the epigynous tepals rule out these families. In addition the apical bulge in the perianth region (Pl. 7, fig. 3) may be interpreted as a nectary disk, as is characteristic of Apiales. Paired thin mericarps are characteristic of Apicaceae and also occur in some Araliaceae. Specialized carpophores of the type found in Apiaceae do not appear to be de- veloped in this fossil.
Similar fruits occur in the extant genus Pseudopan-
CLARNO (EOCENE) FRUITS AND SEEDS: MANCHESTER 39
ax. Fruits of P. davidii (Franch.) W.R. Philipson from western Hupeh, China (Pl. 7, fig. 5) are particularly similar to the fossil. However, Pseudopanax has a sin- gle style that divides distally into two arms, whereas the fossil has two distinct styles.
Family BETULACEAE Gray
The Betulaceae are a family of six extant genera, with an excellent fossil record in the northern hemi- sphere (Crane, 1989). The family is represented in the Nut Beds by the two fruit types described below, Cor- yloides and Kardiasperma, and by wood attributed to Betula (Scott and Wheeler, 1982). The family is rep- resented at other Clarno Formation localities, includ- ing West Branch Creek and White Cliffs, by leaves and infructescences of Alnus and by fruits that may be sim- ilar to Palaeocarpinus. Also a fruit resembling a small Shuttle-cock, Calycites ardtunensis Crane, is known from West Branch Creek, White Cliffs and Gosner Road localities of the Clarno Formation, and may represent an extinct genus of the Betulaceae (Crane, 1989).
Genus CORYLOIDES gen. nov.
Etymology.—The generic name refers to the simi- larity of these fossils to nuts of Corylus. Type species.—Coryloides hancockii sp. nov.
Coryloides hancockii sp. nov. Plate 7, figures 6-12, 16-18
Etymology.—The species is named in honor of the late Alonzo W. Hancock, amateur fossil collector and founder of the field station adjacent to the Nut Beds locality.
Description.—Nut nearly perfectly spherical, round- €d apically and basally; unilocular, single seeded; out- ermost surface as seen in casts and external molds Strongly ribbed meridionally; ribs spaced 2—4 mm apart at the equator; fruit wall 0.8-1.2 mm thick with a large longitudinal vascular bundle corresponding to each of the external ribs. Nut cast spherical, length 24.8-37.5, avg. 28.8 mm (SD=2.65, n=30), equatorial diameter 23.0-32.6, avg. 27.5 mm (SD=2.48, n=30), with a Prominent basal scar about 20-25 mm in diameter and а circular stylar protrusion 1.5-2.5 mm in diameter at the apex; surface of nut cast above the basal scar with faint longitudinal ridges corresponding in position to the ribs on the fruit surface; perimeter of the basal scar forming a circumscissle band ca. 2—3 mm wide with a Smooth, circular upper margin at which point the lon- 8ltudinal surface ribs begin, and a more jagged lower Margin with the broken ends of vascular bundles cor- responding to the surface ribs, where they may have been torn away from the involucre; nut surface slightly
recessed over the basal scar. Seed more or less globose, ca. 20-24 mm in diameter, longitudinally wrinkled, with a large circular hilum, about 9.5 mm in diameter at one end.
Specimens.—Holotype: UF 8497. Paratypes: UF 8492-8496, 8499, 9898, USNM 40544 (Scott, 1954), 354408, 354425, 354620, 354630, 354632, 354634- 354636, 354641-354643, 354645, 354647, 354650, 354654, 354658, 354677, 354757, 354811, 354903, 354922, 354984, 354986, 354992, UWBM 35268, 35269, nut casts; UF 8498, nut with seed partially exposed; UM 29942, seed.
Discussion. — Coryloides is distinguished by a nearly spherical, meridionally ridged nut with a large circular circumsissle mark on one end (Pl. 7, figs. 7, 8, 9, 10) and a small stylar scar at the other (Pl. 7, figs. 6, 8). The chalcedony-replaced nuts are attractive specimens with the shape and density of a large game marble. The fruits occur mainly at the top of Faces 3 and 4 in the Nut Beds, and are not known from other localities of the Clarno Formation. Although these specimens were informally identified as palm fruits (Gregory, 1969; Bones, 1979), the large circular band at the basal end of the nut is inconsistent with the morphology of extant Palmae. The cellular anatomy of the nutshell usually is not preserved but in one of the more com- pletely preserved, although somewhat crushed, speci- mens, transverse sectioning reveals the thickness and sculpture of the outer wall, and shows the presence of a large vascular bundle corresponding in position to each of the surface ribs (Pl. 7, figs. 17, 18).
Scott (1954) figured a fruit cast (p. 86, pl. 16, fig. 24) and an isolated seed cast (p. 87, pl. 16, figs. 29, 30) of Coryloides hancockii among his incertae sedae. A si- licified fruit that has been fractured, revealing an intact seed cast (Pl. 7, fig. 12) provides the basis for linking isolated seed casts, such as that figured by Scott, as belonging to the same species as the fruit casts. The sedimentary mold surrounding the silicified nut cast shows that the external surface was more strongly ribbed (Pl. 7, figs. 11, 16) than the nut cast alone would sug- gest. The large basal scar indicates probable detach- ment from an involucre as is seen in the nuts of various Fagaceae and Betulaceae.
The globose form, unilocular, single-seeded struc- ture and prominent basal scar of these nuts at first suggested affinities with cupulate fruits of Quercus and Lithocarpus in the Fagaceae. Fruits with such promi- nent longitudinal ridges are not known in the Fagaceae, and the peculiar, more or less crenate pattern of vas- cular bundles seen on the basal scar is unlike anything observed among extant fagaceous fruits (Kaul, pers. comm., 1989). In addition, the nutshells of Quercus, Lithocarpus and other Fagaceae do not contain large,
40 PALAEONTOGRAPHICA AMERICANA, NUMBER 58
widely spaced longitudinal vascular bundles as seen in this fossil.
The affinities of this fruit appear to be close to Cor- ylus in the Betulaceae. The nature ofthe scar left behind on the nut of Corylus, where Ше involucre, or “hull”, detaches (Pl. 7, figs. 13, 14), is very similar to that seen in Coryloides. This feature, together with the presence of a single locule and seed and the relatively thin nut- shell containing unusually large longitudinal vascular bundles (Pl. 7, fig. 15), is consistent with the fossil nuts. However, extant Corylus fruits are smaller and not as perfectly globose as those of Coryloides. The basal scar of Coryloides is nearly circular, whereas that of Corylus usually somewhat distorted and elliptical, indicating a difference in the symmetry of involucral attachment. These features, and the prominently ribbed surface of the nuts indicate that the fossils do not belong in the extant genus. Whether Coryloides is a sister genus to Corylus, or a case of parallel evolution in another line of the Betulaceae is open to speculation.
Genus KARDIASPERMA gen. nov.
Etymology.— Kardia (Gr = of the heart) + sperma (Gr — seed). Type species.— Kardiasperma parvum sp. nov.
Kardiasperma parvum sp. nov. Plate 8, figures 1—5
Etymology.— Parvum (L — little) referring to the small size of the locule casts.
Description.— Locule casts small, dorsiventrally compressed, obcordate, quadrilaterally symmetrical, base rounded or faintly cordate with hilum central, apex blunt-pointed in face view (ca. 90°); length 1.3— 2.4, avg. 2.1 mm (SD=0.23,n=19), width 1.0—2.0, avg. 1.7 mm (SD=0.25, n=19), thickness 0.5-1.2, avg. 1.0 mm (SD=0.14, n=19); surface smooth with a keel in the major plane ofsymmetry and with a median groove running dorsally and ventrally from base to apex in the minor plane of symmetry.
Specimens.— Holotype: USNM 435077. Paratypes: UCMP 10710-10713, UF 6602-6604, USNM 355706- 355710, 424705-424710, 424750, 424768, 435076, 435078—435081.
Discussion.— These specimens are morphologically identical to the locule casts of Palaeocarpinus from the Paleocene of Almont, North Dakota (Crane et al., 1990, p. 24). There is a marked difference in size, however. Locule casts from the Almont locality are 45.5 mm wide, i.e., two to five times larger than the species described above. Although the locule casts of Kar- diasperma parvum have not been found intact within complete fruits, they correspond in size to fruits of an undescribed species of small, spiny-bracted fruit sim-
ilar to Palaeocarpinus known from compressions (Pl. 8, fig. 20) at the West Branch Creek and White Cliffs localities of the Clarno Formation. Palaeocarpinus is an extinct genus of the Coryleae known from the Pa- leocene of England (Crane, 1981) and North America (Crane et al., 1990; Sun and Stockey, 1992).
Family BURSERACEAE Kunth.
The Burseraceae, or torchwood family, comprises 16 genera and about 500 species of tropical trees and shrubs with present centers of diversity in America and northeastern Africa. The family is divided into three tribes (Lam, 1931, 1932; Daly, 1989), based prin- cipally upon characters of the fruits. In the tribe Bur- sereae (Bursera, Commiphora, Aucoumea, Boswellia) pyrenes are concrescent but not connate, and have a dry, dehiscent pericarp. In the Protieae (Protium, Ga- ruga, Tetragastris, Hemicrepidospermum, Crepidos- permum) pyrenes are separate and the outer pericarp may be either fleshy or dry and dehiscent. In the Can- arieae (Canarium, Scutinanthe, Dacryodes, Santiria, Trattinickia, Haplobus) pyrenes are connate, forming a plurilocular endocarp with a fleshy, rarely more or less dry, indehiscent pericarp. General fruit and seed morphological characters for the family are reviewed by Reid and Chandler (1933, pp. 266-7).
Genus BURSERICARPUM Reid et Chandler
Bursericarpum oregonense sp. nov. Plate 8, figures 6-17
Etymology.— The epithet refers to Oregon, the state in which the Clarno fossils were collected.
Description.— Endocarp single-seeded, bilaterally symmetrical, tear-dropshaped, pointed apically, rounded basally, subtriangular in cross section, with a convex, rounded dorsal face and two more or less flat ventral faces meeting in a median longitudinal angle of 70—90°, this angle sharper above the placenta; length 4.7-8.7, avg. 6.0 mm (SD=0.97, n=42), width 2.8—5.7, avg. 4.3 mm (SD=0.82, n—42), thickness 2.2-4.9, avg. 3.7 mm (SD=0.85, n=42); transverse placental slit concave upwards, located centrally on the ventral angle of the endocarp; endocarp wall 0.3 mm thick, com- posed of more or less isodiametric sclereids 15-35 um in diameter; exocarp not anatomically preserved, but showing a reticulum of coarse vascular bundles de- veloped on dorsal and ventral surfaces; the bundles of this reticulum becoming closer together and forming a complex at the placental area. Locule casts similar in shape to the endocarps, with more pronounced ven- tral concavity in the position of placental slit, a sharper (less rounded) ventral angle, a more pointed apex, and a curved ridge near the base of the dorsal side corre- sponding to the edge of the germination valve.
CLARNO (ЕОСЕМЕ) FRUITS AND SEEDS: MANCHESTER 41
Specimens.—Holotype: USNM 424885. Paratypes: HU 60010, UCMP 10655-10660, UF 6647-6648, 6657-6663, USNM 326717 (Bones, 1979, pl. 3, fig. 7, lower left), 326718 (Bones, 1979, pl. 3, fig. 7, lower right), 355491, 424877-424884, 424886-424888, 434951-434960, 434964.
Discussion.— Bursericarpum oregonense is герге- sented by numerous chalcedony locule casts (Pl. 8, figs. 16, 17), occasional permineralized endocarps (Pl. 8, figs. 6-10, 12, 15), and a few complete fruit casts. The rounded angles of the endocarps suggest that they were borne freely (not fused) within the fruit. In addition, complete fruit casts with the reticulate-veined exocarp intact (Pl. 8, figs. 13, 14) show only a single endocarp рег fruit. The absence of fruit casts bearing two or more endocarps suggests that fruits of this species were typ- ically single-seeded, a feature of the tribe Bursereae.
Among extant genera, Bursericarpum oregonense most closely resembles Protium, Tetragastris and Bur- Sera. Endocarps of Protium and Tetragastris, of the tribe Protieae, generally are larger than the fossils, and are often borne two or more per fruit. Bursera endo- Carps are borne one-per-fruit and are similar in size to the fossil. This species is particularly similar in en- docarp morphology to the London Clay fossil Burser- icarpum aldwickense Chandler (1961b). A separate Species name is necessary, however, because exocarp vasculature, and the number of pyrenes per fruit, are unknown in B. aldwickense and because features of the Seed, readily apparent in the London Clay material, are not known for the Nut Beds specimens.
Bursericarpum sp. Plate 8, figures 18, 19
Description.—Endocarp single-seeded, bilaterally Symmetrical, tear-drop-shaped, pointed apically, Tounded basally, subtriangular in cross section, with a Convex, rounded dorsal face and two more or less flat ventral faces meeting in a median longitudinal angle of 70—90°, length 2.0-3.2 mm, avg. 2.8 mm (SD=0.5, 1—5), width 1.3-2.3 mm, avg. 1.9 mm (SD=0.4, n=5), thickness 1.0-2.0, avg. 1.6 mm (SD=0.4, n=5); pla- Cental slit transverse, located centrally on the ventral angle of the endocarp, curved apically through the en- docarp.
Specimens.—UF 6654-6656, USNM 434961, 434963.
Discussion. — This species is based upon specimens that are morphologically very similar to Bursericarpum Огевопепзе, but which are much smaller. Most are pre- Served as chalcedony casts of the endocarp; locule casts have not been observed. It is possible that these spec- Imens may be immature or abortive endocarps of B. Oregonense.
Family CORNACEAE Dumort.
The Cornaceae, or dogwood family includes six ex- tant genera: Cornus, Mastixia, Nyssa, Camptotheca, Davidia and Diplopanax (Eyde, 1988; Eyde and Xiang, 1990). Although Nyssa and Davidia have often been segregated as Nyssaceae, Eyde (1988) reviewed the cri- teria used to separate Cornaceae and Nyssaceae and found them inadequate to support the continued rec- ognition of Nyssaceae. Most species of the Cornaceae are arborescent and are distributed from temperate to tropical regions. The Nut Beds assemblage includes both extant and extinct genera of the family: Cornus, Mastixia, Mastixioidiocarpum, Mastixicarpum, Lang- tonia and Nyssa. Other relatives of Cornaceae present among the Nut Beds fruits include Alangiaceae (p. 34) and Hydrangeaceae (p. 50).
Genus CORNUSL.
Cornus clarnensis sp. nov. Plate 9, figures 1-15
Etymology.— The epithet is named after Clarno, Or- egon, the small settlement west of the Nut Beds.
Description. — Endocarp nearly globose to ellipsoid- al, bilocular (to trilocular), length 4.2, 4.4, 5.0 mm, width 3.8, 4.4, 4.3 mm; length/width ratio 1.0-1.3; base rounded, apex obtusely pointed; surface with ca. 10 fine longitudinal grooves; endocarp wall 0.5-0.6 mm thick, without lacunae, composed of more or less isodiametric polygonal sclereids, 38-70 um in diam- eter, decreasing in size toward the outer surface; sep- tum 0.2-0.25 mm thick without central vascular strand but with two peripheral vascular strands forming prominent longitudinal grooves along the endocarp surface in plane ofthe septum; locule lining uniseriate, composed of thin-walled cuboidal cells 45 um in di- ameter along the distal surface ofthe locule and 25 um on the proximal surface (nearest the septum); each loc- ule with one seed.
Specimens.— Holotype: USNM 422378. Paratypes: UF 6310, 9763, USNM 355397, 424702.
Discussion.— Cornus clarnensis is represented only by five specimens: a chalcedony endocarp cast and four permineralized endocarps. Among the four specimens with internal structure preserved, three are bilocular (e.g., Pl. 9, figs. 7, 8) and the other is trilocular, but otherwise identical in anatomy (Pl. 9, fig. 12). The lack of a central vascular bundle in the septum supports placement in Cornaceae. The species corresponds to extant Cornus in the smooth ellipsoidal outline of the endocarp, the number of locules and the isodiametric sclereids making up the endocarp. Based upon analysis of many characters of extant species, Eyde (1988) di- vided Cornus into two groups that he named infor-
42 PALAEONTOGRAPHICA AMERICANA, NUMBER 58
B
D С
Text-figure 10.— Mastixioid fruits from the Nut Beds. A, Mastixia sp. В, Mastixioidiocarpum oregonense Scott. С, Mastixicarpum occidentale sp. nov. D, Langtonia bisulcata Reid and Chandler. Epicarp stippled where preserved. Scale bar = | cm.
mally according to the color of their fruits, the Blue line (ca. 50 spp.), and Red line (ca. 15 spp.). In shape and distribution of surface grooves, C. clarnensis re- sembles extant C. florida L., a member of the “big- bracted” subgroup of the “showy bract” group of the red line dogwoods (sensu Eyde, 1988). The fossil is distinguished from endocarps of the Cornelian cherry subgroup (or “dunstania” type) by the absence of la- cunae in the endocarp wall.
A thorough review ofthe fossil record of Cornus was presented by Eyde (1988). Cornus clarnensis appears to be the oldest known fruit record of the big-bracted dogwoods. However, multilocular Cornus endocarps of the dunstania type occur in the late Paleocene of North Dakota (Crane et al., 1990) and in the early Eocene London Clay flora (Eyde, 1988). Although Reid and Chandler (1933) attributed the London Clay spe- cies to an extinct genus, Dunstania, Eyde (1988) argued that the *dunstanias" belong within Cornus and, be- cause of lacunae in the endocarp wall, correspond to the Cornelian cherry subgroup.
Genus MASTIXIA Blume
Mastixia is an extant tropical genus of 13 species distributed in Indomalesia. Several fossil species of Mastixia (reviewed by Kirchheimer, 1957; Mai, 1970) and closely related extinct genera (Kirchheimer, 1936, 1957) are common in the middle to late Tertiary of central Europe, prompting application of the term “Mastixoid flora" to these warm-climate fossil assem- blages (Kirchheimer, 1940). Knobloch and Mai (1986) reported several species of this complex from the Late Cretaceous (Maastrichtian) of Germany and presented a key distinguishing the European fossil mastixioid genera. The mastixioid group is represented in the Nut Beds flora by Mastixia, Mastixioidiocarpum, Mastix- icarpum, and Langtonia.
Mastixia sp. Plate 10, figures 1-3; Text-figure 10А
Description.—Endocarp ovoid, unilocular, single- seeded, surface smooth; length 8.0-11.5 mm (esti- mated from incomplete specimens), width 5.7—7.0, avg. 6.3 mm (SD=5.7, n=5), dorsiventral thickness 6.0- 6.7, avg. 6.4 mm (SD=0.33, n=4); dorsal surface with a sulcus formed by a deep longitudinal infold of the endocarp wall extending about 2.8 mm into the locule; infold thick (1.8-2.4 mm); locule U-shaped in cross section; endocarp wall 0.6-0.8 mm thick, composed of two layers: the inner layer 0.3-0.4 mm thick with fibers 14-20 um in diameter and more than 400 um long, mostly oriented transversely; outer layer approx- imately the same thickness, 0.3-0.4 mm, composed of more or less isodiametric to elongate sclereids, 20-35 um wide and up to 200 um long; germination valve elongate; dehiscence plane running nearly the full length of the endocarp, extending from the dorsal limbs of the locule through the thickness of the endocarp wall; two longitudinal lacunar tubes up to 0.5 mm in di- ameter situated parallel and adjacent to the dehiscence plane, one on each side of the locule; epicarp not pre- served. Locule cast ellipsoidal, concave dorsally, trough- shaped, base rounded, apex pointed; length 7.3 mm, width 4.8 mm, thickness 4.1 mm.
Specimens.—UF 6303, 6502, 9369, 9470, permi- neralized endocarps; USNM 354035, locule cast.
Discussion.— This species, represented by a few per- mineralized endocarps and a locule cast, conforms 10 extant Mastixia on the basis of its relatively smooth endocarp surface, the dorsal infold and germination valve and the anatomical structure of the епдосагр consisting of an inner layer of transversely aligned fi- bers surrounded by an outer layer of sclereids (Pl. 10, fig. 3). Kirchheimer (1936, pl. 5, fig. 10) and Eyde
CLARNO (ЕОСЕМЕ) FRUITS AND SEEDS: MANCHESTER 43
(1963) illustrated endocarps of comparable living spe- cies including M. trichotoma Blume. Eyde (1963) ob- served that: “the endocarp splits exactly where two Ovular traces pass through the endocarp from base to Placenta. These vascular strands form hollow space where the ovular trace deteriorates”. The same struc- tures are observed in the fossil (Pl. 10, fig. 2; Text-fig. 10A) and are described above as lacunar tubes. The dorsal infold is much thicker in this fossil than in most extant species. However, the two species of subgenus Manglesia (M. euonymoides Prain and M. octandra Matthew), which I have not seen, are described as hav- ing thick infolds (“septum swollen to one third the diameter of the fruit", Matthew, 1976).
This species differs from the other Clarno mastixioid Species by its smaller size, thinner endocarp wall and Smooth surface. The lack of preserved epicarp (me- Socarp and exocarp) may indicate that the outer layers were fleshy and hence lost prior to fossilization. The Smooth surface of the endocarp and small size differ- entiates this species from the nodular surfaced endo- carp of Mastixia cantiensis and from the gently ribbed M. parva from the London Clay flora (Reid and Chan- dler, 1933).
Genus MASTIXIOIDIOCARPUM Scott
Mastixioidiocarpum oregonense Scott Plate 10, figures 4-11, Text-figure 10B
Mastixioidiocarpum oregonense Scott 1954, р. 84, pl. 16, figs. 16— 18.
Emended Description.—Fruit ovoid, rounded api- cally and basally, unilocular, single-seeded, length 13- 23.8, avg. 20.2 mm (SD=3.91, n=6), width 9.2-15, avg. 11.7 mm (SD=2.01, n=8), dorsiventral thickness 8.2-15.0, avg. 11.6 mm (SD=2.45, n=7); endocarp longitudinally ribbed with nine to 14 prominent irreg- ular sharp ridges up to 1.5 mm in relief that are irreg- ularly divided into excrescences or protuberances; en- docarp surface with a dorsal sulcus (longitudinal infold) €xtending 3.8-5.8 mm into the locule (hence, locule U-shaped in cross section); endocarp wall 2.5-4 mm thick measured to the crest of the highest ribs; ger- Mination valve elongate; dehiscence plane running Nearly the full length of the endocarp, extending from the dorsal limbs of the locule through the thickness of the endocarp wall; two longitudinal lacunar tubes up to 0.5 mm in diameter running through the endocarp adjacent to the dehiscence plane, one on each side of the locule; endocarp consisting of an inner layer 0.3 to 9.4 mm thick made up of horizontal, periclinally ori- ented fibers, 17-23 um in diameter, grading into the Outer layer 2.2-3.7 mm thick composed of fibers, 28—
0 um in diameter, arranged in tracts of varying ori-
entation; locule lining formed by transversely aligned cells; epicarp not preserved. Locule cast narrowly sad- Че shaped, the ventral face convex, the dorsal face concave, U-shaped in transverse section, the arms proximate, lateral margins rounded, but thinning to- ward the ends and acute at base and apex, traces of a median raphe strand visible on the ventral surface.
Specimens.— Holotype: UM 29938. Hypotypes: HU 60013, OMSI Pb196, UF 9206-9208, 9210, 9269, USNM 354022-354028 (these specimens together in the same piece of matrix), 354029, 354031, 354033, 354034, 354036, 354037, 354598, endocarps; UCMP 10661, 10662, USNM 354036, 354038, 354039 (3 specimens), 354041 (3 specimens) 354042, 354049, locule casts.
Discussion. — Mastixioidiocarpum oregonense is dis- tinguished from the other mastixioids of the Nut Beds by its prominently sculptured endocarp surface (Pl. 10, figs. 4, 8; Text-fig. 10B). It is preserved as perminer- alized endocarps, locule casts and endocarp molds. Commonly, the endocarp itself has disintegrated, leav- ing behind its mold in the matrix, with a chalcedony locule cast inside. In one such example, the locule cast was removed (Pl. 10, figs. 9—11) and a replica of the original endocarp, complete with the irregular surface ribs, was prepared by making a latex cast of the sur- rounding sedimentary mold (Pl. 10, fig. 8). Molds with the same distinctive type of sculpture are known from other localities of the Clarno Formation, including West Branch Creek, Gosner Road, and from the Eocene Sep- ulcher Formation of Yellowstone National Park (Princeton 22488, specimen at USNM).
These endocarps are morphologically similar to those of extant Mastixia, but differ by the prominent, coarse sculpture and thick fruit wall. Scott (1954) established the genus Mastixioidiocarpum for this Nut Beds spe- cies, because not enough details were preserved to en- able a secure assignment to any of the extinct genera known from the European Tertiary (Kirchheimer, 1936; Reid and Chandler, 1933). Since Scott's work was pub- lished, however, many additional specimens have been recovered, some with excellent anatomical detail, en- abling a more informed comparison with other fossil and modern material. The resulting new information, included in the above description, supports retention of the generic name Mastixioidiocarpum, because the Clarno fossil 1s readily distinguished in its morphology from the European fossil mastixioid genera reviewed by Holy (1975) and Knobloch and Mai (1986). In sculpture and thickness of the endocarp, Mastixioidi- ocarpum resembles Eomastixia Chandler (= Ganitro- cera Kirchheimer) as emended by Holy (1975). How- ever, Eomastixia generally is two-loculed (up to four- loculed) and, when unilocular, there is usually evidence
44 PALAEONTOGRAPHICA AMERICANA, NUMBER 58
ofan abortive ovule (Holy, 1975). Mastixioidiocarpum specimens are exclusively unilocular, without any in- dication of an abortive ovule.
Genus MASTIXICARPUM Chandler
Mastixicarpum occidentale sp. nov. Plate 10, figures 12-15; Text-figure 10C
Etymology.— Occidens (L = of the west), referring to the western occurrence of this species.
Description.—Fruit ovoid, unilocular, length 21.0- 22.0 mm, width 14.5-16.0 mm, dorsiventral thickness 11.8-15.5 mm; endocarp with faint longitudinal ridg- es; dorsal infold present but not well exposed surfi- cially; infold extending 6 mm into the locule; endocarp wall 3.6-4.0 mm thick, composed ofinterwoven tracts of fibers, the fibers 38-80 um in diameter; vascular bundles located in a narrow zone transitional between endocarp and epicarp; epicarp surface smooth, exocarp 0.8-1.0 mm thick, composed of anticlinally elongate sclereids, 28—55 um in diameter, 80-300 um long. Loc- ule boat-shaped, the dorsal face concave, and the ven- tral convex, c-shaped in cross section; dehiscence plane running nearly the full length of the endocarp, extend- ing from the dorsal limbs of the locule through the thickness of the endocarp wall.
Specimens.— Holotype: UF 8740. Paratype: UCMP 10705.
Discussion.—This species is known from two frag- mentary but well preserved permineralized specimens from the Nut Beds, and by a well preserved specimen from the Brummers’s Spring Clarno locality (UF 9886; loc. 254). Mastixicarpum occidentale differs from the other Clarno mastixoids by the presence of a resistent epicarp (Pl. 10, figs. 13-15, Text-fig. 10C). The epicarp in the other species was probably fleshy and easily degraded as in extant Mastixia. The endocarp of this species is smooth in comparison to the strongly ribbed endocarps of Mastixioidiocarpum discussed above.
Among fossil mastixioids recognized in the Tertiary of Europe, four genera have persistent epicarps (Knob- loch and Mai, 1986): Mastixicarpum Chandler, Mas- tixiopsis Kirchheimer, Retinomastixia Kirchheimer, and Tectocarya Kirchheimer. Of these, the Nut Beds species most closely resembles Mastixicarpum and Tectocarya. Tectocarya has a pair oflacunae separated by a median septum in the upper % of the dorsal infold, and has a thin epicarp that may become free from the endocarp, whereas Mastixicarpum lacks the longitu- dinal septum in the infold and has a thick epicarp that is tightly attached to the endocarp. Based upon sections of Tectocarya rhenana Kirchheimer that I have ex- amined from Diiren, Germany, there are no prominent secretory ducts, and the epicarp is composed of very
small isodiametric cells. The Clarno species corre- sponds most closely in morphology and anatomy to Mastixicarpum, based upon sections of M. limnophil- um that I have examined from the Miocene of Wiesa, Germany. The two species correspond in thickness of the epicarp, size and orientation of epicarp cells, and in the presence of secretory canals at the junction of the endocarp and epicarp.
Recent recognition of the extant Chinese genus Di- plopanax as a living mastixioid (Eyde and Xiang, 1990) is of considerable interest in interpreting paleobotan- ical remains. Diplopanax, like Mastixia, has an ovoid unilocular fibrous endocarp with an elongate dorsal germination valve. Although there is a dorsal infold resulting in a locule that is u-shaped in cross section, the infold is not apparent at the surface of the endocarp, which has a circular cross section. Mastixia fruits have a soft epicarp, whereas that of Diplopanax is woody or leathery. Also, the endocarp of Diplopanax is rid- dled with longitudinal secretory canals that are not observed in the endocarp of Mastixia. Eyde and Xiang (1990) suggested that Mastixicarpum and Diplopanax are equivalent. However, in both Mastixicarpum lim- nophilum and M. occidentalis the secretory canals are mostly peripheral to the endocarp (e.g., Pl. 10, fig. 15), largely confined to the epicarp, whereas in Diplopanax stachyanthus Hand.-Mazz. the secretory canals are uniformly distributed, giving a speckled appearance to the endocarp in cross section.
Genus LANGTONIA Reid et Chandler
Langtonia bisulcata Reid et Chandler Plate 11, figures 1-10; Text-figure 10D
Langtonia bisulcata Reid et Chandler 1933, p. 453, pl. 25, fig. 18- ER
Description.— Fruit oblong, more or less circular in cross section, rounded basally and apically, length 16.0- 35.0, avg. 24.9 mm (SD=5.27, n=7), width 10.0-26.5, avg. 14.8 mm (SD=4.56, п=7), dorsiventral thickness 9.9-22.0, avg. 14.1, mm (SD=4.12, n=7); bicarpellate, with two single-seeded elongate locules, sometimes with one of the locules poorly formed, abortive; locules w-shaped in cross section, due to a pair of longitudinal infolds of the endocarp wall; longitudinal planes of dehiscence in the endocarp wall defining a pair of ger- mination valves, one corresponding to each locule; each germination valve extending the full length of the loc- ule, similar in width to the locule that it serves; septum without a central vascular column; endocarp made up of tracts of fibers oriented in various directions; epicarp about 1 mm thick, composed of thin-walled, periclin- ally elongate cells 20-30 um high and 40-120 um wide;
CLARNO (EOCENE) FRUITS AND SEEDS: MANCHESTER 45
Secretory ducts present both in endocarp and epicarp; seeds conforming in shape to the locule. Locule casts elongate, dorsiventrally flattened 12.5-20.6 mm long, 5.5-12.7 mm wide, 2.2-6.0 mm thick, pointed apically and basally, sometimes with fine transverse striations; dorsal side with a median longitudinal ridge flanked on either side by a prominent groove, ventral side convex-rounded with a thread-like median longitudi- nal raphe groove or ridge.
Specimens.—HU 60016, UF 8729-8735, 8740, 8742, USNM 312746 (Bones 1979, pl. 4, fig. 5 top), 354032, 354046, 354052, 354053, 354351, 354423, 354937, 354941, 422524, 424810, 424875, endocarps; UF 8736-8739, USNM 312747 (Bones 1979, pl. 4, fig. 5, left), 354043, 354047, 354050, locule casts. Holo- type: BM v22984 from the London Clay.
Discussion.—Langtonia is represented in the Nut Beds by permineralized endocarps (Pl. 11, figs. 1-3) and occasional isolated locule casts (Pl. 11, figs. 5, 6). Fruits have been recovered from the basal leaf layer of Face 3, as well as higher in the Nut Beds section. The preservation of fruit morphology and anatomy is
i et as seen in transverse sections (Pl. 11, figs. 4, -10).
This species was first described and recognized as an extinct genus from the London Clay flora by Reid and Chandler (1933). They attributed Langtonia to Ше Cornaceae and placed it near Mastixia because of its long, infolded germination valves. It differs from Mas- tixia in having two infolds of the locule, rather than One, resulting in a w-shaped, rather than u-shaped, Cross section. Kirchheimer (1936) expressed doubt as to the mastixioid affinities of this genus. Anatomy re- vealed in peels of the well preserved Clarno material (Pl. 11, figs. 9, 10), however, shows variously oriented tracts of fibers forming the endocarp in conformity With the anatomy of Mastixia, Diplopanax and Nyssa. In addition, the presence of secretory canals in the €ndocarp and epicarp is a feature shared with Diplo- Panax and Mastixicarpum. Therefore, Langtonia Should be reinstated as a member of the Cornaceae.
Most of the Nut Beds endocarps are larger than those described by Reid and Chandler (1933) from the Lon- don Clay (length 9-20, avg. 15.5 mm; width 4.5-12, avg. 9.6 mm). The size ranges overlap, however, and Ше specimens are morphologically indistinguishable, and I have therefore placed the Nut Beds specimens Ш the same species.
Genus NYSSA L.
Nyssa spatulata (Scott) comb. nov. Plate 11, figures 11—19
Palaeonyssa spatulata Scott 1954, p. 83, pl. 16, fig. 8-12.
Description.—Endocarp oblong, ovoid or ellipsoid- al, trilocular, roughly circular to rounded-triangular in cross section, base and apex rounded, with nine round- ed longitudinal ridges extending from base to apex, length 20.1-30.0, avg. 24.0 mm (SD=3.8; n=15), width 12.0-21.5, avg. 14.6 mm (SD=2.8, n=16), endocarps trilocular, with large, rounded germination valves vis- ible as slits in the apical half ofthe endocarp; wall and septa composed of swirling tracts of fibers, the fibers 15-30 um in diameter; central axis not vascularized, epicarp not preserved. Locules single-seeded, elongate, spatulate, c-shaped in cross section, concave outward, broadening and rounded apically, narrowed and point- ed basally.
Specimens.— Holotype: UM 29936 (Scott, 1954, pl. 16, fig. 8, 9). Paratype: USNM 40543 (Scott, 1954, pl. 16, fig. 10-12). Hypotypes: OMSI Pb197A, Pb914, UF 6834-6842, USNM 354541, 354582-354584, 354586, 354591, 354594, 354595, 354604, 424874.
Discussion.—This species was described by Scott (1954) on the basis of four specimens. Although it is not abundant in the assemblage, it is now known from approximately 25 specimens, resulting in the broader range of dimensions indicated in the above description. All of the specimens show three locules (rather than three to four as in the London Clay species).
Reid and Chandler (1933) correctly deduced the close affinity of this taxon to extant Nyssa, but because of some apparent differences, particularly the presence of three to four locules, placed the London Clay fossils in a new genus, Palaeonyssa. Scott (1954) recognized close similarity between the Clarno and London Clay fruits, and placed the Nut Beds specimens in a new species of Palaeonyssa. These fossils differ from most modern species of Nyssa in being consistently trilocular (Pl. 11, figs. 16-18), rather than unilocular, and by their larger size. Thin sections reveal that the endocarp is made up of swirling fibers (Pl. 11, fig. 19), as is also true of extant species of Nyssa.
Nyssa has about six extant species (three in North America, one in China, one in Indomalesia and one in southern Central America) that are readily distin- guished from one another on the basis of fruit mor- phology (Eyde, 1963). The new discovery of an extant species of Nyssa from Costa Rica and Panama, N. talamancana Hammel and Zamora (1990), is helpful in the interpretation of fossil remains. Fruits of N. talamancana measure 4—5 cm in length and 2-2.5 cm in diameter, and are larger than those of any other living or fossil species of the genus. In contrast to the other extant species of Nyssa, which typically have only one locule, and in rare instances two locules (Eyde, 1963), stones of N. talamancana are bilocular to tri- locular (Hammel and Zamora, 1990). Taking into con-
sideration the wider range of endocarp morphology imparted to Nyssa by the tropical American species, the taxon Palaeonyssa now appears to fall neatly within the extant genus.
Nyssa scottii sp. nov. Plate 12, figures 1-2
Etymology.—This species is named for Richard A. Scott in recognition of his careful work on the Clarno fruit and seed flora.
Description.—Endocarp spindle shaped, fusiform in lateral view, roughly circular in cross section, triloc- ular, base and apex bluntly pointed, surface with coarse meridional ribs; length 12.6-16.5, avg. 13.8 mm (SD=2.3, n=6), width 6.1—9.5, avg. 7.5 mm (SD=1.18, n=6), endocarp with three inverted u-shaped germi- nation valves over the apical Y of the endocarp; en- docarp wall 1.2 mm thick composed of fibers 12-30 um in diameter, arranged in swirling groups, central axis not vascularized.
Specimens.— Holotype: USNM 354588. Paratypes: OMSI Pb207, UF 6843, USNM 354582, 354600, 435150.
Discussion. — Nyssa scottii is represented by several translucent silica casts and one permineralized speci- men (UF 6843). The shape, size, surface sculpture and type of germination valve are diagnostic of Nyssa. The specimens are tricarpellate as in N. spatulata, rather than unicarpellate as is the usual condition in extant species. One of the specimens (Pl. 12, figs. 1, 2) has one of the germination valves broken away, showing a locule that is D-shaped in cross section with the convex surface inward. The short length of the ger- mination valves is a character that distinguishes Nyssa from the mastixioid genera (Chandler, 1926, pp. 36— 37). This species is readily distinguished from М. spa- tulata by its smaller size range (length 12-17, vs. 20— 30 mm).
Nyssa sp. Plate 12, figures 3—5
Description.— Fruit unilocular, endocarp more or less elliptical in face view, dorsiventrally compressed, rounded basally, dorsal side convex, ventral side con- cave; length 24.6, 30.0, 48.0 mm (est. from incomplete specimen), width 12.9, 23.5, 25.0 mm, dorsiventral thickness 5.1, 12.6, 10 mm, surface relatively smooth, with thin, shallow longitudinal grooves; endocarp wall 3.2 mm thick, composed of tracts of swirling fibers, the fibers 15-30 um in diameter. Locule cast dorsi- ventrally compressed.
Specimens. — UF 9757, UM 66135, USNM 446088.
Discussion.— Three unilocular specimens of Nyssa have been recovered from the Nut Beds. The longi-
PALAEONTOGRAPHICA AMERICANA, NUMBER 58
tudinal grooves on the endocarp (Pl. 12, fig. 3) and thick wall composed of swirling tracts of fibers (Pl. 12, fig. 5), are consistent with assignment to Nyssa. There is no indication of abortive locules and, in the absence of any bilocular specimens in the Nut Beds collections, it seems likely that this represents a species distinct from the trilocular species, N. spatulata and N. scottii. Although it is incomplete, the larger of the specimens (Pl. 12, fig. 3) appears to have been longer than any of the trilocular specimens.
The relatively smooth surface, size, and marked dor- siventral compression of these endocarps suggests af- finities with N. brandoniana (Lesquereux) Eyde and Barghoorn, from the Brandon lignite of Vermont (Eyde and Barghoorn, 1963) and extant N. javanica (Bl.) Wangerin of southeast Asia and Malesia (Eyde, 1963).
Family FAGACEAE Dum.
The Fagaceae (oak family) are represented in the Nut Beds by at least two genera of fruits: Quercus, and Castanopsis. Two kinds of fagaceous foliage have been recovered from the Nut Beds (Manchester, 1981). In addition, two species of fagaceous wood have been recognized (Scott and Wheeler, 1992): Quercinium crystallifera, with anatomy corresponding both to ex- tant Lithocarpus and to evergreen species of Quercus, and Fagaceoxylon ostryopsoides. Fagaceoxylon Scott et Wheeler is an extinct genus with characters diag- nostic of fagaceous wood including aggregate rays, ор- posite to alternate intervascular pitting, mostly simple perforation plates, vasicentric tracheids and variable ray-vessel parenchyma pits. As noted by Scott and Wheeler (1982), however, the pore arrangement of Fa- gaceoxylon, i.e., diffuse porous with radial multiples and clusters arranged in “flame-like” tracts, is un- known in any living genus.
The Fagaceae have an excellent fossil record in the Tertiary of the northern hemisphere (Daghlian and Crepet, 1983; Crepet and Nixon, 1989 a, b; Kvacek and Walther, 1989). The Nut Beds flora provides the oldest known fruit records of Quercus and Castanopsis.
Genus CASTANOPSIS (D. Don) Spach.
Castanopsis crepetii sp. nov. Plate 12, figures 6-11
Etymology.— This species is named after William Crepet, recognizing his contributions to the North American fossil record of Fagaceae.
Description.— Fruit a globose nut completely envel- oped by cupule, length (15)-39 mm, equatorial di- ameter (15)-36 mm; base and apex rounded, surface coarse; nutshell 1.8-2.0 mm thick, composed of iso- diametric angular sclereids 12-50 um in diameter ОГ
CLARNO (EOCENE) FRUITS AND SEEDS: MANCHESTER 47
ganized concentrically into islands of densely packed Cells among smaller cells, with occasional vascular bundles passing through the nutshell; cupule adnate to the nut, 1.8-2.0 mm thick, composed of similar, less densely packed cells. Locule cast rounded-conical, truncate basally, pointed apically, with meridional grooves defining broad-rounded ridges.
Specimens.—Holotype: UF 9267, large perminer- alized fruit. Paratypes: UF 9268, USNM 354811, 446079, fruits; UF 9597, 9598, locule casts.
Discussion.—This species is represented by permi- eralized fruits and occasional locule casts. The ho- lotype isa large, fragmentary permineralized specimen showing the globose outer surface of the cupule (Pl. 13, fig, 6), with portions of the cupule and nutshell removable to show the apically tapered, basally trun- Cate, locule cast with longitudinal grooves (Pl. 12, figs. 7, 8). Thin sections of the wall clearly show the junction Of cupule with the nut (Pl. 12, figs. 9, 10).
In the Fagaceae, some species of both Castanopsis and Lithocarpus have globose nuts with complete cu- Pular coverage and adnation between the nut and cu- Pule (Camus, 1952-54; Kaul, 1988, 1989). These gen- ста apparently overlap with one another, as summarized by Kaul (1989, р. 68): “Generic and infrageneric tax- onomy in the Lithocarpus—Castanopsis complex has a long history of instability (Forman, 1966). Some recent authors have accepted the two genera in their broad Senses, acknowledging some species with intermediate Character states."
Castanopsis piriformis Hick. et A. Camus from An- Nam, southern Vietnam (Pl. 12, figs. 12-15) isan extant Species that resembles C. crepetii in size, globose shape Of the fruit, wall structure and locule morphology. Transverse sections of the endocarps of these two spe- Cles reveal remarkable similarity, both at the macro- Scopic and microscopic levels. In both the fossil and the modern species, adnation of cupule and nut is clear- ly evident (Pl. 12, figs. 9, 14) and the nutshell wall is Made up of more-or-less angular sclereids that are Srouped concentrically to form conspicuous “islands” 85 viewed in transverse section (Pl. 12, figs. 11, 15).
he anatomical correspondence confirms the identity i this fossil to the Fagaceae, and supports placement Within Castanopsis (providing that the generic assign- ment of C. piriformis is correct). Castanopsis also has a 800d fossil fruit record in Europe based upon species with small, angular nuts (Mai, 1989).
Genus QUERCUS L.
Quercus paleocarpa Manchester Plate 12, figures 16-18
Quercus paleocarpa Manchester, 1976, p. 82, pl. 5, fig. 29.
Description.— Acorn with prolate to ovate nut cov- ered А to У; from base to apex by a thick woody cupule; complete fruit length (nut with cupule) 22 mm, width 15.5-21.8 mm; nut length (without cupule) 14.0-18.0, width 11.4-15.6 mm; nut base rounded, apex obtuse; relatively smooth, finely longitudinally striate; cupule bowl-shaped, height 9.0-10.4 mm, width 18-21 mm, with seven concentric exterior transverse ribs; the ribs decreasing in thickness from the base to the apical rim.
Specimens :— Holotype: ОМІ Pb176, nut with abraded cupule. Hypotypes: USNM 414508, isolated cupule; USNM 312758 (Bones, 1979, pl. 6, fig. 9), complete, abraded nut.
Discussion.—This species, representing the oldest known acorn, is represented only by a few specimens. The most complete specimen, with both the nut and cupule intact (Pl. 12, figs. 16, 17) was abraded prior to deposition, so that the arrangement of involucral scales on the outer surface of the cupule cannot be deter- mined. A well-preserved detached cupule (Pl. 12, fig. 18) shows the outer surface in an unworn condition, revealing involucral scales arranged in concentric rings. I assume that each of these specimens represents the same species, although a larger population would be desirable to rule out the possibility that two different taxa are present.
Although the cupules of extant species of Lithocar- pus and Quercus are often indistinguishable (Kaul, 1988), the coarse, rather than papery, lamellae of the cupule and the limited extent to which the cupule cov- ers the nut suggest that the species belongs to Quercus rather than Lithocarpus. Cupules with the involucre arranged in concentric rings as in this fossil occur in the strictly Asian subgenus Cyclobalanopsis of Quer- сиз, as well as in Lithocarpus.
Nixon (1989) noted that the cyclic arrangement of cupule scales/spines, which also occurs in Chrysolepis, Castanopsis, Castanea and Trigonobalanus, is appar- ently pleisiomorphic within the Fagaceae, with the im- bricate-spiral arrangement found in the derived forms of Quercus. If this interpretation is correct, then it is not surprising that Quercus paleocarpa, the oldest known Quercus fruit, should possess a cupule with cy- clically arranged scales.
Family FLACOURTIACEAE DC.
The Flacourtiaceae are a family of about 90 genera and about 875 species of shrubs, less often trees, that are chiefly tropical in distribution, but with a few that are temperate. The species described below, belonging to an extinct genus, is the only member of this family recognized from the Nut Beds. The extant genus On- coba, recognized in the London Clay flora (Reid and
48 PALAEONTOGRAPHICA AMERICANA, NUMBER 58
Зтт
Text-figure 11.—Saxifragispermum tetragonalis sp. nov., trans- verse section showing four-valved capsule with parietally attached seeds.
Chandler, 1933) does not appear to be present in the Nut Beds collections.
Genus SAXIFRAGISPERMUM Reid et Chandler
Saxifragispermum tetragonalis sp. nov. Plate 13, figures 1-16; Text-figure 11
Etymology.— Tetra (Gr = four) + gonia (Gr = angle, corner), referring to the four-angled shape of the fruit.
Description.—Fruit an elongate-obovoid, tetracar- pellate, tetravalved loculicidal capsule, nearly circular to rounded-quadrangular in cross section; length 6.4— 11.8, avg. 9.2 mm (SD=1.42, n=20), width 3.0-5.6, avg. 4.2 mm (SD=0.63 n=19); base more or less round- ed, sometimes with a small projection at the pedicel scar, apex pointed; placentation parietal, with four elongate placental ribs, one positioned medially on each of the four valves of the fruit and extending the full length of fruit; seeds numerous, six to eight radiating from any given level on the placenta, arranged verti- cally in rows about 12 high; interior of fruit solid with sinuously arranged fibers 5-18 um in diameter having a general longitudinal orientation, among which the seeds are embedded; fruit wall comprised of two layers: inner layer forming a thin lining (80-120 um thick) of longitudinally oriented fibers 12-30 um in diameter (this is the layer that forms the placental ribs), outer layer (150-200 um thick) formed of similar fibers ori- ented horizontally and periclinally; this outer layer showing lines of loculicidal dehiscence; pedicel scar with five apertures, one central, and four lateral, rep- resenting extensions of the placentae. Seeds ovoid, as- cending, rounded basally, tapered apically, 0.5-0.7 mm long, 0.35-0.45 mm wide; seed coat composed of an outer layer of mostly quadrangular cells about 15-20 um wide producing small spine-like simple hairs 18- 25 um long and an inner layer of transversely elongate hexagonal cells about 35-40 um wide and 12-14 um high.
Specimens.—Holotype: UF 5236 (Bones, 1979, pl.
4, fig. 4). Paratypes: UCMP 10614-10617, UF 6312, 6471, 6473-6478, 6575, 6649-6653, 6972, 9897, USNM 353999, 354000 (21 specimens), 354001 (7 specimens), 354002 (2 specimens), 354006-354008, 354009 (7 specimens), 354010 (2 specimens), 354011, 355469.
Discussion. — This species is represented by numer- ous permineralized specimens and occasional casts. Part of the outer wall is often stripped away when the specimens are removed from the matrix, sometimes showing the rows of seeds (Pl. 13, figs. 3-5). Serial transverse sections reveal as many as eight seeds sur- rounding each placenta in a single slice (Pl. 13, figs. 7— 9). The placentae are preserved in white silica and appear to have been hollow or parenchymatous (Pl. 13, fig. 9). Scanning electron microscopy of fractured specimens illustrates details ofthe seeds and surround- ing fibers (Pl. 13, figs. 11-14).
These valvate capsules resemble the fruits that Reid and Chandler (1933) named Saxifragispermum spi- nosissimum (emended diagnosis Chandler, 1961b) from the London Clay. Both have median longitudinal pa- rietal placentae with about a dozen rows of seeds on each placenta; in both taxa the entire cavity of the locule is virtually filled with fibers (Pl. 13, figs. 9, 11- 13). Seeds of the London Clay fossil were originally described as spinose and small projections from the seed coat were figured by Reid and Chandler (1933, pl. 8, fig. 35). However, Chandler (1961b) redescribed the seeds more cautiously: “Testa apparently spines- cent (unless this is an effect of decay)”. Thin sections of the Nut beds specimens confirm the presence of small spines or spine-like trichomes on the seed coat (PI. 13, figs. 15, 16).
Placentation and seed morphology indicate that Saxifragispermum is related to the Flacourtiaceae. Al- though the generic name suggest affinities to Saxifra- gaceae (Reid and Chandler, 1933), Chandler (1961a) reported that “Examination of more material and re- consideration of the whole evidence necessitates а cor- rection of the supposed relationship to Saxifragaceae. The single locule with its broad parietal placentae bear- ing several rows of seeds is unlike any known Saxifra- gaceae, a family which normally shows two or more locules. It is highly suggestive of Flacourtiaceae. Моге- over the succession of coats in the seed and the char“ acter of the chalaza are so closely comparable with the same features in Oncoba that the probability of rela- tionship with Flacourtiaceae is confirmed.”
Although clearly congeneric with the London Clay fruits, the Nut Beds material can confidently be ге“ garded as a distinct species. The London Clay fruits are much wider in diameter, and have larger seeds. 5.
CLARNO (ЕОСЕМЕ) FRUITS AND SEEDS: MANCHESTER 49
Spinosissimum fruits range from 11.5 to 15.75 mm, whereas those of S. tetragonalis range only from 6.4— 11.8 mm. The London Clay fossils range from four to five carpels, whereas the many Nut Beds specimens are all tetracarpellate. Carpolithus subfusiformis (Bow- erbank) Reid et Chandler (1933), which is known from tri- and tetracarpellate fruits, also shows similarity to and may, in fact, be conspecific with S. spinosissimum.
Family HAMAMELIDACEAE R. Br.
The Hamamelidaceae (witch hazel family), with 29 extant genera and about 90 species, is divided into four Subfamilies: the Hamamelidoideae, Rhodoleioideae, Exbucklandioideae, and Liquidambaroideae (Endress, 19893). The Liquidambaroideae are known to have been present in the Eocene based upon Liquidambar leaves in the Green River Formation, Utah (Mac- Ginitie, 1969), and fruits identified as Steinhauera Presl from the Geiseltal of Germany (Mai, 1976), but have Not been identified from the Clarno Formation. In- fructescences from the Clarno Nut Beds formerly iden- üfied as Altingia (Scott in Chandler, 1961a; Bones, 1979) have been shown to represent Macginitiea (Pla- tanaceae; Manchester, 1986). The Hamamelidoideae are represented in the Nut Beds flora by the single fruit and seed species described below, and by inflorescences that may or may not represent the same taxon (Pl. 14, figs. 14—16). The oldest fossils with close affinities to the Hamamelidaceae are flowers with intact stamens ànd pollen from the Late Santonian or Early Campan- làn of Scania, Sweden (Endress and Friis, 1990).
Genus FORTUNEARITES gen. nov.
4 Etymology.— Fortunearites is named for its similar- ПУ to extant Fortunearia. Type species.— Fortunearites endressii sp. nov.
Fortunearites endressii sp. nov. Plate 14, figures 1-13
Etymology.— The species 15 named for Peter К. En- Tess, recognizing his important contributions in un-
derstanding the systematics and phylogeny of Hama- Melidaceae.
Description— Infructescence a spike, with a stout axis UP to 53 mm long, 2.4 mm in diameter, bearing nu- Merous closely packed sessile fruits; infructescence about 15 fruits in length with up to eight fruits en- countered at a given