ee ee Dmep ees - a Lan : wie Wee te ef Facies ada Get, b danste is Hare WS 9 oie os Te ce fee PRP Le Spee thet abe s . Ses thant: Ee ee cab eons de re Cee ste i weer ce wag WEE Ehret tale deta te. pe thteskies ree Tee oer % pe tuleis he itt ey tea teense teteeads wed oy eid Nyylsuespyes ed PLP ge gE TAN ee Ake Wb teat ady abs thee ss sedbet ane Sead gawd recat ere Seren ees vB fhe eae athe tg Ma Pe ge TN SS se See iy ew lite “= 7 a . . - = a wo _ 7 3 ‘ wo 7 = 7 0 ‘ 7 a7 ‘Ss ty a : a 2 au q. eal i. a Dod te ay ei 1 oe . 7 mt MADRONO A WEST AMERICAN JOURNAL OF BOTANY VOLUME XLV 1998 BOARD OF EDITORS Class of: 1998—FREDERICK ZECHMAN, California State University, Fresno, CA JON E. KEELEY, U.S. Geological Service, Biological Resources Division, Three Rivers, CA 1999—TimoTHy K. Lowrey, University of New Mexico, Albuquerque, NM J. MARK PorTER, Rancho Santa Ana Botanic Garden, Claremont, CA 2000—PAMELA S. SOLTIS, Washington State University, Pullman, WA JOHN CALLAWAY, San Diego State University, San Diego, CA 2001—-ROBERT PATTERSON, San Francisco State University, San Francisco, CA PAULA M. SCHIFFMAN, California State University, Northridge, CA 2002—-NORMAN ELLSTRAND, University of California, Riverside, CA CARLA M. D’ ANTONIO, University of California, Berkeley, CA Editor—KRISTINA A. SCHIERENBECK California State University, Chico Department of Biology Chico, CA 95427-0515 kschierenbeck @csuchico.edu Published quarterly by the California Botanical Society, Inc. Life Sciences Building, University of California, Berkeley 94720 Printed by Allen Press, Inc., Lawrence, KS 66044 MADRONO, Vol. 45, No. 4, pp. 11-111, 1998 DEDICATION Robert Folger (Bob) Thorne was born at Spring Lake, N. J., 13 July 1920, a community near the family home in Belmar. When he was two, the family moved to a sec- ond home at Gulfport, Florida, where they wintered, re- turning each summer to Belmar. Thus, he was reared with- in walking distance of saltwater, which accounts, no doubt, for his life-long addiction to sea food and marine aquatics! Learning to swim helped him later. Bob’s father, Harry, was a professional portrait photographer and long served as the official photographer of both the New York Yankees and Boston Red Sox during their spring training activities at St. Petersburg, Florida. One of Bob’s prized mementos of this era is a large photograph autographed by both Babe Ruth and Lou Gehrig and dedicated to his father. Bob graduated from St. Petersburg Senior High School in 1937 and in the same year entered Dartmouth College with aspirations of becoming a linguist; fortu- nately for botany, he discovered plants during his fresh- man year, and the rest, as they say, is history. He studied at Dartmouth with tuition scholarships where his academic prowess brought him membership in Phi Beta Kappa; he received his A.B. summa cum laude in 1941. Under the auspices of a Cramer Fellowship he moved on to Cornell University where he gained his M.S. in Botany in 1942. As so often happened in those days, Bob was invited to enter the service. Although he doubtless could have collected and observed numerous plants as a member of one of the several ground forces, he chose instead, follow- ing his enlistment in 1942, to fly. He was trained as a navigator in the U. S. Army Air Force and served in the European Theater of Operations. His tenure with the 15th Air Force, 461st Bomb Group Heavy stationed in Italy, was not without considerable excitement, and he still re- calls the names and nature of many of his flying comrades Robert E Thorne. Back from military service in Europe, October, 1944 (left). At the International Botanical Congress, Berlin, 1983 (right). | 1998] | and the often harrowing times they shared. While on his 11th bombing mission into Axis territory, his plane, a B24 Liberator, was shot up over Austria and went down off the coast of Yugoslavia. Fortunately, he and his crew par- ' achuted to ‘“‘safety’’ on the Dalmatian island of Vis and were rescued by partisans—Bob landed in the Adriatic and managed to swim ashore despite the burden of a par- achute and a wet uniform. Soon back with his unit in Italy, he was so “‘admired”’ for his efforts that he was ‘‘asked”’ to take part in 29 more missions before being allowed to return to the states in September 1944 and ultimately re- leased from service shortly after the war in Europe ended early in 1945. Bob resumed his formal training in botany under the guidance of Professors Walter Muenscher and Arthur Eames at Cornell, receiving his Ph.D. in 1949. He married his wife of 52 years, the former Mae Zukel, in 1947. His doctoral thesis was a floristic study of southwestern Geor- gia and resulted in a long paper that brought him a share of the first George Cooley Award presented at the Amer- ican Society of Plant Taxonomy meeting in 1955. Before graduation, Bob accepted an Assistant Profes- sorship in Botany at the University of Iowa, Iowa City, becoming Associate Professor in 1954 and Professor in 1961. During his years at Iowa, he distinguished himself not only as organizer and curator of a herbarium but also as a teacher and as a mentor of numerous graduate stu- dents. Not incidentally, he developed his basic philosophy regarding the principles underlying angiosperm phylogeny and biogeography, and laid the groundwork for his be- coming what he is today—a world leader in the study of plant geography, phylogeny, and floristics. Of great significance to western botany was Bob’s move from Iowa to the Rancho Santa Ana Botanic Garden in 1962 where he became Taxonomist and Curator of the Herbarium as well as Professor of Botany in the Clare- mont Graduate School. Within a relatively short time, through diligent work both in the herbarium and in the field, he mastered the California flora to a degree that amazed many long-time students of the region. During his long tenure at the garden not only did he have continued success as teacher and researcher but also, under his guid- ance, the herbarium grew and prospered, becoming one of the outstanding repositories of plants from the western United States and elsewhere. And his studies, alone or lil with others, have contributed greatly to our knowledge of the biogeography and flora of California. Bob’s writings on geography, floristics, and phylogeny have brought him international renown. The “‘Thorne sys- tem”’ of the higher-level classification of plants is accepted and used by many botanists throughout the world, and it is the standard against which much modern-day research is compared. More often than not in the light of today’s molecular and cladistic studies, it stands up to critical scrutiny. His botanical studies and explorations have in- volved not only diverse regions of the United States, Mex- ico, and Jamaica in the New World but many parts of the Old World as well—Australia (Fullbright Research Schol- ar, 1959-1960; NSF Senior Postdoctoral Fellow, 1960), Tasmania, New Caledonia, New Guinea, Indonesia, Tai- wan, India, China, Iran, and Scandinavia. Bob has been accorded many honors for his scientific studies and academic accomplishments. He has been elected Foreign Member of the Royal Danish Academy, Fellow of the Linnean Society of London, and Fellow of the French Societé de Biogeographie. He is a member of numerous societies, including The American Institute of Biological Sciences; The American Society of Plant Tax- onomists (Secretary, 1957-1958; Council, 1961-1967; President, 1968); The Botanical Society of America (Chmn. of the Central States Section, 1956, the Systematic Section, 1957—1958; and the Pacific Section, 1977); The California Botanical Society (Vice-President, 1966); The Southern California Botanists (President, 1966; Council, 1963-1987; Vice-President, 1975—1977). He is a member of Sigma Xi, Phi Kappa Phi, and was National Treasurer of Gamma Alpha, 1954—1957. In 1996, he received a Bo- tanical Society of America Merit Award in recognition of his contributions to the field of plant systematics. Although Bob formally retired in 1987, he has contin- ued his studies at the garden nearly full time. He is as enthusiastic as ever, relentlessly searching for new items to add to his unparalleled plant-stamp collection and con- tinuing his many and varied botanical pursuits. The latter includes a never-ending perusal of the botanical literature in search of all new information having valid bearings on his system of classification, which he does not hesitate to upgrade as needed from time to time, even at the expense of long-held opinions. For his distinguished service to the botany of California, and to the world for that matter, it is entirely fitting that this volume of Madrono be dedi- cated to Robert Folger Thorne. TABLE OF CONTENTS Alarcon, Ruben (see Marchant, T. Alejandro, Ruben Alarcon, Julie A. Simonsen, and Harold Koopowitz) Allen-Diaz, Barbara (see Jackson, Randall D.) Anderson, David C. (see Hinshaw, Jay M., Gary L. Holmstead, Brian L. Cypher, and David C. Anderson) Anderson, David C. (see also Holmstead, Gary L., and David C. Anderson) Anderson, R. Scott, and Brian E Byrd, Late-holocene vegetation changes from the Las Flores Creek coastal lowlands: San, Diego, C oumty,Calitor inital 2) assests pease pees Arnot, Mildred (see Titus, Jonathan H., Scott Moore, Mildred Arnot, and Priscilla J. Titus) Ayers, Tina (see Holiday, Susan) Banks, Darin L., and Steve Boyd, Noteworthy collections from California Barbour, M. G. (see Fernau, R. EF) Barbour, Michael (see Danin, Avinoam) Beidleman, Richard G., Obituary, Lauramay Tinsley Dempster Boyd, Steve (see Banks, Darin L.) Boyd, Steve, and Andrew C. Sanders, Noteworthy collections from California Byrd, Brian E (see Anderson, R. Scott) Burwell, Trevor, Successional patterns of the lower montane treeline, eastern California Buxton, Eva, and Robert Ornduff, Noteworthy collection from Califormia Caicco, Steven L., Current status, structure, and plant species composition of the riparian vegetation of the lrtickeeRiver,-C alatornia: atid NOV aC a2: 5 is cs 8 eh a a eee ee Campbell, Jonathan E. (see Lichvar, Robert W.) Chmielewski, Jerry G., Antennaria dioica (Asteraceae: Inuleae): Addition to the vascular flora of California _. Chu, Ge-Lin (see Stutz, Howard C.) Colin, Larry J. (see Jones, C. Eugene) Connors, Peter (see Danin, Avinoam) Cooper, Stephen V. (see Lesica, Peter, Peter Husby, and Stephen V. Cooper) Curto, Michael, and Douglass M. Henderson, A new Stipa (Poaceae: Stipeae) from Idaho Cypher, Brian L. (see Hinshaw, Jay M., Gary L. Holmstead, Brian L. Cypher, and David C. Anderson) Danin, Avinoam, Stephen Rae, Michael Barbour, Nicole Jurjavcic, Peter Connors, and Eleanor Uhlinger, Early primary succession on dunes at Bodega Head, California Day, Alva G. (see Shevock, James R.) del Moral, Roger (see Titus, Jonathan H., Roger del Moral and Sharmin Gamiet) Del Tredici, Peter, Lignotubers in Sequoia sempervirens: Development and ecological significance Dorsett, Deborah K. (see Jones, C. Eugene) Elliott-Fisk, Deborah L. (see Stephens, Scott L.) Ellstrand, Norman C. (see Lyman, Jennifer C.) Erickson, Richard A. (see Harrison, James E.) Ericson, Trudy R. (see Jones, C. Eugene) Espeland, Erin K. (see Pavlik, Bruce M.) Fernau, R. FE, J. M. Rey Benayas, and M. G. Barbour, Early secondary succession following clearcuts in red fir forests.of the Sierra. Nevada.-C aliformia 2.5. 52. Sie Deu hoe Ne ee Fletcher, Grant, and Margriet Wetherwax, Noteworthy collection from California Fulgham, Kenneth O. (see Jackson, Randall D.) Gale, Nathan (see Parikh, Anuja) Gamiet, Sharmin (see Titus, Jonathan H., Roger del Moral and Sharmin Gamiet) Gill, David S., and Barbara J. Hanlon, Water potentials of Salvia apiana, S. mellifera (Lamiaceae) and their hybrids in the coastal sage scrub of southern California Hanlon, Barbara J. (see Gill, David S.) Harrison, James E., Richard A. Erickson and Fred M. Roberts, Jr., Noteworthy collection from California _____.. Henderson, Douglass M. (see Curto, Michael) Hileman, Lena C. (see Markos, Staci, Lena C. Hileman, Michael C. Vasey, and V. Thomas Parker) Hinshaw, Jay M., Gary L. Holmstead, Brian L. Cypher, and David C. Anderson, Effects of simulated oil field disturbance and topsoil salvage on Eriastrum hooveri (Polemoniaceae) _______-----------------2----22---22 222 Holiday, Susan, and Tina Ayers, Noteworthy collection from Arizona _.------ nnn nnn Holmstead, Gary L. (see Hinshaw, Jay M., Gary L. Holmstead, Brian L. Cypher, and David C. Anderson) Holmstead, Gary L., and David C. Anderson, Reestablishment of Eriastrum hooveri (Polemoniaceae)following ou melds distirbance activities. 25.c. 35 eens he Aa 8 od ero see I re a a een pe tee Husby, Peter (see Lesica, Peter, Peter Husby, and Stephen V. Cooper) Jackson, Randall D., Kenneth O. Fulgham, and Barbara Allen-Diaz, Quercus garryana Hook. (Fagaceae) stand suucture in aréas, with different prazinohiStOries © 22.28 aac ees Jones, C. Eugene, Larry J. Colin, Trudy R. Ericson, and Deborah K. Dorsett, Hybridization between Cercidium floridum and C. microphyllum (Fabaceae) in California 22-22 ee Jurjavcic, Nicole (see Danin, Avinoam) IgA 85 88 326 12 184 17 271 57 101 259 134 326 141 85 290 184 295 275 — 1998] TABLE OF CONTENTS | Kittel, Timothy G. FE, Effects of climatic variability on herbaceous phenology and observed species richness in temperate montane habitats, Lake Tahoe Basin, Nevada __.. nnn _ Koopowitz, Harold (see Marchant, T. Alejandro, Ruben Alarcon, Julie A. Simonsen, and Harold Koopowitz) Kuijt, Job, Noteworthy collection from British Columbia _ Kuykendall, Keli (see Zika, Peter EF, Keli Kuykendall, Danna Lytjen, and Nick Otting) Kuykendall, Keli (see also Zika, Peter E, Keli Kuykendall, and Barbara Wilson) Lassoie, James P. (see Sheppard, Paul R.) Lesica, Peter (see Shelly, J. Stephen) ~ Lesica, Peter, Peter Husby, and Stephen V. Cooper, Noteworthy collections from Montana Levin, Anna L. (see McGraw, Jodi M.) Lichvar, Robert W., William E. Spencer, and Jonathan E. Campbell, Distribution of winter annual vegetation across environmental gradients within a Mojave Desert playa Little, R. John, President’s report for Volume 45 — Lyman, Jennifer C., and Norman C. Ellstrand, Relative contribution of breeding system and endemism to ge- notypic diversity: The outcrossing endemic Taraxacum californicum vs. the widespread apomict T. officinale OSCE SEL) CEO) ) eters tesa Sree Se an I en eI ea ae So ge ies se Lytjen, Danna (see Zika, Peter F, Keli Kuykendall, Danna Lytjen, and Nick Otting) Mansfield, Donald H., Noteworthy collection from Oregon Marchant, T. Alejandro, Ruben Alarcon, Julie A. Simonsen, and Harold Koopowitz, Population ecology of Dudleya multicaulis (Crassulaceae); a rare narrow endemic — Markos, Staci, Lena C. Hileman, Michael C. Vasey, and V. Thomas Parker, Phylogeny of the Arctostaphylos hookerit complex (Ericaceae) based On nrDNA Gata: cece a Martin, Bradford D., Flowering phenology and sex expression of Croton californicus (Euphorbiaceae) in coastal SAVE ESEHUD Ol SOULICIE Ce ali LOIIt a: sabe f ces 5 ac Se 2 ee a ee eee Mathiasen, Robert L., and David C. Shaw, Adult sex ratio of Arceuthobium tsugense in six severely infected ESTES LETEH OPTI Mess as Eo I, EF ROMO cana a sd ces Saad McBride, Joe (see Russell, William H.) McGraw, Jodi M., and Anna L. Levin, The roles of soil type and shade intolerance in limiting the distribution of the edaphic endemic Chorizanthe pungens var. hartwegiana (Polygonaceae) Mensing, Scott A., 560 years of vegetation change in the region of Santa Barbara, California = Moore, Scott (see Titus, Jonathan H., Scott Moore, Mildred Arnot, and Priscilla J. Titus) Ornduff, Robert (see Buxton, Eva) Oswald, Vernon H., Roger Raiche and Carol Whitham, Noteworthy collection from California Otting, Nick (see Zika, Peter EF, Keli Kuykendall, Danna Lytjen, and Nick Otting) Parikh, Anuja, and Nathan Gale, Coast live oak revegetation on the central coast of California Parker, V. Thomas (see Markos, Staci, Lena C. Hileman, Michael C. Vasey, and V. Thomas Parker) Patterson, Robert, Review of Mojave Desert Wildflowers by Jon Mark Stewart Patterson, Robert (see also Pritchett, Daniel W.) Pavlik, Bruce M., and Erin K. Espeland, Demography of natural and reintroduced populations of Acanthomintha duttonii, an endangered serpentine annual in northern California Pritchett, Daniel W., and Robert Patterson, Morphological variation in California alpine Polemonium species __ Rae, Stephen (see Danin, Avinoam) Raiche, Roger (see Oswald, Vernon H.) Rey Benayas, J. M. (see Fernau, R. F) Reynolds, Don R., Limaciniaseta gen. nov. A California sooty mold Roberts, Fred M., Jr., Noteworthy collection from California Roberts, Fred M., Jr. (see also Harrison, James E.) Rondeau, J. Hawkeye, Noteworthy collection from California Rowntree, Rowan (see Russell, William H.) Russell, William H., Joe McBride and Rowan Rowntree, Revegetation after four stand-replacing fires in the | Liga S08 (2) pan 2 bo 0 eee cn See Ma eR Ne eae POE So ra deer Gees eee: Seer MMAE meal ey oe ne Rundel, Philip W., and Shari B. Sturmer, Native plant diversity in riparian communities of the Santa Monica VIC UI EATS SC AiG ON ay ete ee ees cet ce sce eeepc neers Se. gusts east een hoe BA nie nate aie ae, oe Sanders, Andrew C. (see Boyd, Steve, and Andrew C. Sanders) Sanderson, Stewart C. (see Stutz, Howard C.) Schierenbeck, Kristina A., Editor’s report for Volume 45 2 Seagrist, Randy V., and Kevin J. Taylor, Alpine vascular flora of Halsey Basin, Elk Mountains, Colorado, USA Seagrist, Randy V., and Kevin J. Taylor, Alpine vascular flora of Buffalo Peaks, Mosquito Range, Colorado, Shaw, David C. (see Mathiasen, Robert L.) Shelly, J. Stephen, Peter Lesica, Paul G. Wolf, Pamela S. Soltis, and Douglas E. Soltis, Systematic studies and conservation status of Claytonia lanceolata var. flava (Portulacaceae) Sheppard, Paul R., and James P. Lassoie, Fire regime of the lodgepole pine forest of Mt. San Jacinto, Cali- SYST cl pe a re ee -Deeeet tire echo 2 A ces A m8 eM? vs naoicme et anon th eo alert 1g « SU el ee eel no ae Shevock, James R., and Alva G. Day, A new Gilia (Polemoniaceae) from limestone outcrops in the southern Sicrra Nevada ol California ..2 5 a Simonsen, Julie A. (see Marchant, T. Alejandro, Ruben Alarcon, Julie A. Simonsen, and Harold Koopowitz) Soltis, Douglas E. (see Shelly, J. Stephen) 330 330 251 331 283 a 200 250 326 184 40 fb, 352 310 519 vi MADRONO Soltis, Pamela S. (see Shelly, J. Stephen) Spencer, William E. (see Lichvar, Robert W.) Stephens, Scott, Review of The Once and Future Forest: A Guide to Forest Restoration Strategies by L. J. Sauer Stephens, Scott L., and Deborah L. Elliott-Fisk, Seqguoiadendron giganteum-mixed conifer forest structure in 1900—1901 from the southern Sierra Nevada, CA Sturmer, Shari B. (see Rundel, Philip W.) Stutz, Howard C., Ge-Lin Chu and Stewart C. Sanderson, Atriplex longitrichoma (Chenopodiaceae), a new species from southwestern Nevada and east-central California Taylor, Kevin J. (see Seagrist, Randy V., and Kevin J. Taylor, two papers) Titus, Jonathan H., Roger del Moral, and Sharmin Gamiet, The distribution of vesicular-arbuscular mycorrhizae on Mount St. Helens, Washington Titus, Jonathan H., Scott Moore, Mildred Arnot, and Priscilla J. Titus, Inventory of the vascular flora of the blast zone, Mount St. Helens, Washington Titus, Priscilla J. (see Titus, Jonathan H., Scott Moore, Mildred Arnot, and Priscilla J. Titus) Uhlinger, Eleanor (see Danin, Avinoam) Vasey, Michael C. (see Markos, Staci, Lena C. Hileman, Michael C. Vasey, and V. Thomas Parker) Wetherwax, Margriet (see Fletcher, Grant) Wilson, Barbara (see Zika, Peter E, and Barbara Wilson) Wilson, Barbara (see also Zika, Peter EF, Keli Kuykendall, and Barbara Wilson) Witham, Carol (see Oswald, Vernon H.) Wolf, Paul G. (see Shelly, J. Stephen) Zika, Peter E, Keli Kuykendall, Danna Lytjen, and Nick Otting, Noteworthy collection from Oregon Zika, Peter EK, Keli Kuykendall, and Barbara Wilson, Carex serpenticola (Cyperaceae), a new species from the Klamath Mountains of Oregon and California UNITED STATES POSTAL SERVICE. 1.Publication Title Statement of Ownership, Management, and Circulation (Required by 39 USC 3685) 2. Publication Number Madrono 4. Issue Frequency Quarterly 7. Complete Mailing Address of Known Office of Publication (Not printer) (Street, city, county, state, and Zip+4) California Botanical Society, Inc. Herbaria, Life Sciences Building hiversity of California 415-744-1985 8. Complete Mailing Address of Headquarters or General Business Office of Publisher (Not printer) California Botanical Society, Inc. Herbaria, Life Sciences Building eit OF Ca if 9. Full names and Complete Mailing Addresses of Publisher, Editor, and Managing Editor (Do not teave blank) Publisher (Name and complete mailing address) California Botanical Society, Inc. Herbaria, Life Sciences Building Editor (Name and complete mailing address) Kristina A. Schierenbeck, California State University, Chico, Dept. of Biology, Chico, CA 95427-0515 Managing Editor (Name and complete mailing address) 10. Owner (Do not leave blank. ff the publication is owned by a corporation, give the name and address of the corporation immediately followed by the names and addresses of all stockholders owning or holding 1 percent or more of the total amount of stock. If not owned by a corporation, give the names and addresses of the individual owners. if owned by a partnership or other unincorporated firm, give its name and address as well as those of each individual owner. Hf the publicaiton is published by a nonprofit organization, give its name and address.) Full Name ‘Complete Mailing Address , A : , Herbaria, Life Sciences Buildin California Botanical Society, Inc a University of California Berkeley, CA 94720 11. Known Bondholders, Mortgages, and Other Security Holders Owning or Holding 1 Percent or More of Total Amount of Bonds, Mortgages, or Other Securities, none, check box. ————_—________» [¥] None Full Name Complete Mailing Address 12. Tax Status (For completion by nonprofit organizations authorized to mail at spectal rates) (Check one) The purpose, function, and nonprofit status of this organization and the exempt status for federal Income tax purposes: Has Not Changed During Preceding 12 Months Ohas Changed During Preceding 12 Months (Publisher must submit explanation of change with this statement) PS Form 3526, September 1995 (See Instructions on Reverse) Computerized Facsimile Zika, Peter E, and Barbara Wilson, Noteworthy collections from Oregon 13. Publication Title 14. Issue Date for Circulation Data Below Madrono July-Sept. 1998 [Vol. 45 273 221 128 162 146 261 86 15. ‘Average No. Copies Each Issue Actual No. Copies of Single Issue Extent and Nature of Ceculation During Preceding 12 Months Published Nearest to Filing Date a. Total Number of Copies (Net press run) 1075 1100 : b. Paid and/or and Counter Sales (Not mailed) Requested = : : advertiser's proof copies and exchange copies) c. Total Paid and/or Requested Circulation > (Sum of 15b(1) and 15b(2)) 8 Z d. Free Distribution by mail (Samples, complimentary, and other free) e. Free Distribution Outside the Mail (Carriers or other means) {. Total Free Distribution (Sum of 15d and 15e) Vv g. Total Distribution Sum of 15¢ and 15f) Vv (1) Office Use, Leftovers, Spoiled NX = = h. Copies not Distributed (2) Retums from News Agents |. Total (Sum of 15g, 15h(1), and 1Sh(2)) Vv 1075 1100 Percent Paid and/or Requested Circulation (15e / 15g x 100) 16. Publication of Statement of Ownership — (uy Publication required. Will be printed in the. 45 C Publication not required. 17. Signature and Title of Editor, Publisher, Business Manager, or Owner Date QiMerL ) Treasucec Aa ust IF | certify that all information furnished on this form is true and complete. | understand that anyone who furnishes false or misleading Information on this form issue of this publication. ‘or who omits material of information requested on the form may be subject to criminal sanctions (including fines and imprisonment) and/or civil sanctions (including multiple damages and civil penalties). Instructions to Publishers 1. Complete and file one copy of this form with your postmaster annually on or before October 1. Keep a copy of the completed form for your records. 2. In Cases where the stockholder or security holder is a trustee, include in items 10 and 11 the name of the person or corporation for whom the trustee is acting. Also include the names and addresses of individuals who are stockholders who own or hold 1 percent or more of the total amount of bonds, mortgages, or other securities of the publishing corporation. In item 11, if none, check the box. Use blank sheets if more space is required 3. Be sure to furnish all circulation information called for in item 15. Free circulation must be shown in items 15d, e, and f. 4. Ifthe publication had second-class authorization as a general or requester publication, this Statement of Ownership, Management, and Circulation must be published; It must be printed in any issue in October or, if the publication Is not published during October, the first issue printed after October. 5. In item 16, indicate the date of the issue in which this Statement of Ownership will be published. 6. Item 17 must be signed Failure to file or publish a statement of ownership may lead to suspension of second-class authorization. PS Form 3526, September 1995(Reverse) Computerized Facsimile Qe a M\82 %o4 AIBUME "tg ey Va | Via by \ iy Si VOLUME 45, NUMBER 1 JANUARY-APRIL 1998 MADRONO A WEST AMERICAN JOURNAL OF BOTANY | ss = | ICONTENTS 560 YEARS OF VEGETATION CHANGE IN THE REGION OF SANTA BARBARA, CALIFORNIA SCOEE As MICS seo sue sass ati oat Hea he tines eases ee een eee cee ee 1 SUCCESSIONAL PATTERNS OF THE LOWER MONTANE TREELINE, EASTERN CALIFORNIA DEV OT RUIN Lite Sa sete vats otdce ees tute esate ease See ease 12 CURRENT STATUS, STRUCTURE, AND PLANT SPECIES COMPOSITION OF THE RIPARIAN VEGETATION OF THE TRUCKEE RIVER, CALIFORNIA AND NEVADA SLC VETU IE WS QUCCO vers c ses tisrtn ah wats nora dnstustoe Ae AU ea SaaS 17 DEMOGRAPHY OF NATURAL AND REINTRODUCED POPULATIONS OF ACANTHOMINTHA DUTTONI, AN ENDANGERED SERPENTINITE ANNUAL IN NORTHERN CALIFORNIA Bruce M. Pavlik and Erin K. Espeland .0.....cccccccccccssccsescseesscessccesecsessseesees 3] REVEGETATION AFTER FouR STAND-REPLACING FIRES IN THE LAKE TAHOE BASIN William H. Russell, Joe McBride and Rowan ROWNn 7 ee ...........000cceeeeeeeeeees 40 FIRE REGIME OF THE LODGEPOLE PINE FoREST OF MT. SAN JACINTO, CALIFORNIA Paul R. Sheppard and James P. LASSO1€ ......ccccsscccccccccesecccececscesessstsseseeeeees 47 A NEw S7IPA (POACEAE: STIPEAE) FROM IDAHO AND NEVADA Michael Curto and Douglass M. Henderson ......ccccccccccccccccccccceeeeeeeetttttenees 51! SYSTEMATIC STUDIES AND CONSERVATION STATUS OF CLAYTONIA LANCEOLATA VAR. FLAVA (PORTULACACEAE) J. Stephen Shelly, Peter Lesica, Paul G. Wolf, Pamela S. Soltis and Douglas. SOUS 55 EL pve ba esta ecse scenes 64 EFFECTS OF CLIMATIC VARIABILITY ON HERBACEOUS PHENOLOGY AND OBSERVED SPECIES RICHNESS IN TEMPERATE MONTANE Hapsitats, LAKE TAHOE BASIN, NEVADA Timothy Goh HG sake ese ete TS) OTEWORTHY GATIBORNIA: ehetere sear l ceecice sc ncde sa cate vodeceds Pawan sel Wace res whee enc 85 OLLECTIONS WOREGON Ssssocss sects caste Wa esaches aatenast ae aeaet aioe neice denea na eeuled eae tances eee 86 OBITUARY LAURAMAY TINSLEY DEMPSTER (1905-1997) .u.....cccccccccccesessssssssssssssseseeeseeees 88 OUNCEMENTS CALIFORNIA BOTANICAL SOCIETY 18TH GRADUATE STUDENT MEETINGS .........- 91 RESOLUTION BY THE CALIFORNIA BOTANICAL SOCIETY ON TRANSPLANTATION ... 92 PUBLISHED QUARTERLY BY THE CALIFORNIA BOTANICAL SOCIETY Mapbrono (ISSN 0024-9637) is published quarterly by the California Botanical Society, Inc., and is issued from the office of the Society, Herbaria, Life Sciences Building, University of California, Berkeley, CA 94720. Subscription information on inside back cover. Established 1916. Periodicals postage paid at Berkeley, CA, and additional mail- ing offices. Return requested. PostMASTER: Send address changes to MADRONO, ‘/ Mary Butterwick, Botany De- partment, California Academy of Sciences, Golden Gate Park, San Francisco, CA 94118. Editor—KrisTINA A. SCHIERENBECK California State University, Chico Department of Biology Chico, CA 95427-0515 kschierenbeck @csuchico.edu Editorial Assistant—Davin T. Parks Book Editor—Jon E. KEELEY Noteworthy Collections Editors—DEITER WILKEN, MARGRIET WETHERWAX Board of Editors Class of: 1998—FRrReDERICK ZECHMAN, California State University, Fresno, CA Jon E. KeELey, U.S. Geological Service, Biological Resources Division, Three Rivers, CA 1999—Timotny K. Lowrey, University of New Mexico, Albuquerque, NM J. MARK Porter, Rancho Santa Ana Botanic Garden, Claremont, CA 2000—Pame_a S. Sottis, Washington State University, Pullman, WA JOHN CALLAWAY, San Diego State University, San Diego, CA 2001—Robert PartTERSON, San Francisco State University, San Francisco, CA PAULA M. ScHIFFMAN, California State University, Northridge, CA 2002—NorMaN ELLSTRAND, University of California, Riverside, CA Carta M. D’ Antonio, University of California, Berkeley, CA CALIFORNIA BOTANICAL SOCIETY, INC. OFFICERS FOR 1998-1999 President: R. JoHN LittLe, Sycamore Environmental Consultants, 6355 Riverside Blvd., Suite C, Sacramento, CA 95831 First Vice President: SusAN D’ ALcamo, Jepson Herbarium, University of California, Berkeley, CA 94720 Second Vice President: Davip Kei, California Polytechnic State University, Biological Sciences Department, San Luis Obispo, CA 93407 Recording Secretary: ROXANNE BitTMaAN, California Department of Fish and Game, Sacramento, CA 95814 Corresponding Secretary: | SUSAN BAINBRIDGE, Jepson Herbarium, University of California, Berkeley, CA 94720. suebain @casnowy.qal.berkeley.edu Treasurer: Mary Butrerwick, Botany Department, California Academy of Science, Golden Gate Park, San Fran- cisco, CA 94118. butterwick.mary @caepamail.epa.gov The Council of the California Botanical Society comprises the officers listed above plus the immediate Past President, WAYNE R. FERREN, JR., Herbarium, University of California, Santa Barbara, CA 93106; the Editor of Maprono; three elected Council Members: MARGRIET WETHERWAX, Jepson Herbarium, University of California, Berkeley, CA 94720; James SHEvock, USDI National Park Service, Pacific West Region, 600 Harrison Street, Suite 600, San Francisco, CA 94107; DIANE ELaM, U.S. Fish and Wildlife Service, 3310 El Camino Avenue, Sacramento, CA 95825; Graduate Student Representative: DENNIS P. WALL, Jepson Herbarium, University of California, Berke- ley, CA 94720. This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). MApDRONO, Vol. 45, No. 1, pp. 1-11, 1998 560 YEARS OF VEGETATION CHANGE IN THE REGION OF SANTA BARBARA, CALIFORNIA ScoTT A. MENSING Department of Geography, University of Nevada, Reno, NV 89557 ABSTRACT Pollen evidence from two sites in the Santa Barbara region show evidence of vegetation changes following European settlement in California. In the Santa Barbara coastal region, oak woodland popula- tions (dominated by Quercus agrifolia) remained stable during the pre-European period; however, in the last century woodland densities have increased. At higher elevations along the oak woodland/pine forest ecotone, pines are becoming dominant. Reduction in fire frequency has probably been the main factor contributing to density increases. The pollen record does not show any evidence of an expansion of chaparral over the last 200 years; however, there is weak evidence for an increase in coastal-sage scrub since the early 1800’s. The transformation of the California grassland appears to have begun particularly early with the invasion of Erodium cicutarium in the region even before the first Spanish settlement in California. With the setthement of Spanish Missionaries, be- ginning in 1769, the California landscape has been radically altered by human-caused environmental change. The exact nature of these changes is not always readily apparent, since descriptions of the pre-European vegetation are sketchy at best, and by the time reliable botanical records were gathered, much of the landscape had already been altered. The vegetation we see today represents the dynam- ic result of two centuries of response to various changes, including changes in fire regime, intro- duction of livestock, invasion of alien species, and land clearing for agriculture and urban develop- ment. Understanding how vegetation has changed in response to these impacts can provide valuable information for present-day conservation efforts. Of particular concern recently has been the affect of human-caused environmental changes in oak woodlands (Muick and Bartolome 1987; Bolsinger 1988). To date, no high resolution paleoecologic studies have been conducted that document envi- ronmental change in regions dominated by Califor- nia oak woodland. A variety of methods have been used to reconstruct past change in California oak woodlands, including age structure studies (White 1966; Vankat and Major 1978; Anderson and Pas- quinelli 1984; McClaran 1986; McClaran and Bar- tolome 1989; Mensing 1992), historical records (Mayfield 1981; Rossi 1980), analysis of aerial photographs (Brown and Davis 1991), and resur- veying of permanent plots (Holzman 1993). These studies have contributed significant information on oak woodland history; however, they are generally restricted to the last 200 years and reveal little about the pre-European period. Pollen analysis from an estuarine sediment core on Santa Rosa Island in the Santa Barbara Channel has documented the invasion of exotic taxa and changes in native vegetation following the settle- ment of the island in the late 1800’s (Cole and Liu 1994). Radiocarbon dates from the top three meters of core however are influenced by old carbon ef- fects which suggest that the bulk sediment dates may be 1200 years too old. The chronology for the historic period therefore was extrapolated from ex- otic pollen types. Oak pollen was only a minor component of the pollen sum and is not discussed by the authors. In this paper I use pollen evidence from two sites in the Santa Barbara region to reconstruct vegeta- tion history for the last 560 years. The sites include the Santa Barbara Channel, just off the coast of Santa Barbara, and Zaca Lake in northern Santa Barbara County (Fig. 1). Oak woodland, chaparral, coastal-sage scrub, and grassland comprise the ma- jority of the vegetation in the region. This area was one of the earliest settled by Spanish missionaries and provides an opportunity to identify the effects of European impacts on vegetation. Repeat photog- raphy is used to help illustrate changes in the last century. STUDY AREA Santa Barbara Basin. The Santa Barbara Basin (34°11'—34°16'N; 120°01’—120°05'W), is located between mainland California and the Channel Is- lands (Fig. 1). The center of the basin has a max- imum depth of ca. 590 m and the bottom waters are normally anoxic. Because of the absence of a bottom fauna seasonal differences in sediment den- sity are preserved as varves, and these have been used to establish a high resolution chronology of the last 560 years (Soutar and Crill 1977; Schim- melman et al. 1990). The source region for pollen deposited in the Santa Barbara Basin is the coastal plain and Santa Ynez Mountains. Quercus agrifolia appear domi- nant on north-facing slopes, in canyons, and in me- sic sites along the coastal marine terrace. West and N MADRONO [Vol. 45 /N a) fe. Study sites Now © Cities A 50 150 kilometers Fic. 1. south facing slopes are dominated by chaparral, and coastal-sage scrub. Along the coastal plain intro- duced grasses and herbaceous plants are common (Ferren 1985). Pinus muricata grows in the Santa Ynez Mountains northwest of Santa Barbara al- though its distribution is restricted (Griffin and Critchfield 1976). Zaca_ Lake. Zaca Lake (34°36'36'N; 120°02'17"W, elev. 730 m) is in the San Rafael Mountains, within the Los Padres National Forest, approximately 50 km northwest of Santa Barbara. Two massive Quaternary landslides blocked Zaca Creek to form the lake (Hall 1981). The lake is steep-sided and slopes to a flat bottom 13 meters in depth. Surface area is 6.9 ha. and the maximum length is 350 m. The lake lies at the transition zone between oak woodland at lower elevations and pine forest at higher elevations. Quercus agrifolia, P. coulteri, P. ponderosa, and P. sabiniana are co-dominants around the lake. South-facing slopes are character- ized by Ceanothus spp., Arctostaphylo spp., Yucca spp., Salvia spp., and Artemisia californica. Can- yons include small stands of Q. douglasii, Q. chry- solepis, and Calocedrus decurrens. Small patches of exotic pines remain as a legacy of tree planting yO \ % Los Angeles > San }Diego (e) Elston 1.10.96 Location map of the Santa Barbara Basin core site (SB), and Zaca Lake (ZA). at the turn of the century (Blakely personal com- munication). Ornamental conifers planted near the lake include Sequoia sempervirens and Cedrus de- odara (Peterson 1980). METHODS Santa Barbara Basin cores SABA 87-1 and 88-1 were recovered by researchers from Scripps Insti- tute of Oceanography with a Soutar Box Corer and a Kasten Corer (Schimmelman et al. 1990). All cores were initially sampled at near annual resolu- tion, but the pollen analysis was based on sub-sam- ples which represented consecutive, ca. 5 year in- tervals. An equal weight of sediment was taken from each annual sub-sample. Fifty-nine samples were analyzed from the period 1425 to 1985. Two sediment cores were recovered from Zaca Lake in May, 1992 using a modified square-rod Livingstone piston corer fitted with a 5 cm diameter plastic tube liner; a 910 cm core (core C) and an overlapping 865 cm core (core D). While still in the plastic tube, cores were X-radiographed at the University of California Museum of Paleontology to record stratigraphy and density changes. Mag- netic susceptibility and Gamma-ray analyses were carried out at the United States Geologic Survey 1998] laboratories in Menlo Park. Sediment samples (0.5 cc) were then removed for pollen analysis. Standard techniques were used to concentrate pollen (Faegri and Iversen 1975). A known quantity of Lycopodium spores was introduced as a control to calculate absolute pollen concentration and ac- cumulation rate (Stockmarr 1971). A minimum of 400 pollen grains were counted for each level. For Zaca Lake, aquatic and riparian pollen types were counted but excluded from the pollen sum. RESULTS Santa Barbara Basin Chronology. The Santa Barbara Basin varve chronology has been corroborated by radiometric dating, cross-correlation with tree-rings and corre- lation with hydrological data (Soutar and Crill 1977; Koide et al. 1972; Krishnaswami et al. 1973; Hulsemann and Emery 1961). The chronology used here is that of Schimmelman et al. (1990), and Schimmelman et al. (1992). Varve counts were made using high quality X-radiographs and age as- signments were checked against distinctive marker layers of known events such as El Nifio periods, floods, and oil spills. The estimated precision of the time scale is +1 year for 1900 to 1987, +2 years from 1900 to 1840, +5 years from 1840 to 1750, and +10 years at the 1425 level. Taxonomy. Taxonomic nomenclature follows Hickman (1993). Forty-eight pollen and spore types were identified (Mensing 1993). Percentage abun- dance of the nine most important types is shown in Figure 2. Quercus probably represents Q. agrifolia, by far the dominant tree species in the region. Ad- ditional, but less important sources may include Q. lobata, which is important in the Santa Ynez drain- age, Q. durata and Q. dumosa which are found in association with chaparral, and Q. tomentella from the Channel Islands. Pinus would primarily be P. muricata, P. sabiniana, P. coulteri, and P. ponder- osa. Following Heusser (1978), the taxonomically difficult group including Rhamnaceae and Rosa- ceae are combined. These taxa include many chap- arral species and probably represent the genera Ce- anothus, Rhamnus, Adenostoma, Cercocarpus, Pru- nus, and Heteromeles. Artemisia is primarily Arte- misia californica. Other Asteraceae are difficult to identify below family level and have been com- bined in Figure 2. Poaceae and Polemoniaceae also are not identified below the family level. Several native taxa from the Brassicaceae are present in small quantities early in the record; however, alien taxa became important in California in the early 19th century and the pollen increase probably rep- resents introduced species. Erodium has been iden- tified to the species Erodium cicutarium, a Medi- terranean annual (Mensing and Byrne in press). Pollen analysis. Below 1760, Quercus shows few changes; however, after 1760 the record be- MENSING: 560 YEARS OF VEGETATION CHANGE ) comes increasingly variable. Beginning in 1870, Quercus steadily increases from 20% to 42%, twice as high as the average during the pre-European pe- riod. Pinus remains below 10% through most of the record, and shows virtually no change. Although Rhamnaceae and Rosaceae show high variability, no long term trends appear over the 460 year pe- riod. Artemisia averages 7% and shows little vari- ability for the first 400 years of the record, then, from 1820 to 1985 it increases to an average of 10%. Asteraceae averages 20% from 1435 to 1700, but then begins to decline, dropping to only 7% percent in 1970. Poaceae declines at a slow but fairly con- stant rate through most of the record, but clearly increases between 1945 and the present. The Po- lemoniaceae are primarily insect pollinated, con- sequently only small quantities of pollen reach the Santa Barbara Basin. Polemoniaceae is present in virtually every level between 1425 and 1795, av- eraging nearly 1% of the pollen sum. In the last two centuries, Polemoniaceae is commonly absent, Brassicaceae is infrequent prior to 1825, but in- creases substantially in the modern period, most likely due to the introduction of European Brassi- caceae. Erodium first appears in the pollen record in 1755 and is continuously present after that date. Zaca Lake Chronology. The Zaca Lake chronology was de- veloped using core D (O—0.5 m depth) and core C (0.5—2.75 m depth). Both cores clearly record a complex stratigraphy of laminations, dark silty lay- ers, and dense clay layers described in earlier stud- ies (Caponigro 1976; Peterson 1980). X-radio- graphs were used to correlate core stratigraphy with that described by Caponigro and Peterson. The chronology was developed using core-to-core cor- relation, radiocarbon dating, and the first appear- ance of two exotic pollen types (Erodium and Ced- rus). The base of the core section analyzed gave a ra- diocarbon age of 2510 + 70 BP (Beta-55301) (Cal- endar calibration B.C. 661 + 150, Stuvier and Rei- mer 1986) (Fig. 3). The first occurrence of Erodium at 110 cm depth is assigned a date of 1830 + 40. The data and error estimate are approximated from the Santa Barbara Basin data (Mensing and Byrne in press), and the species dispersal ability. The date of 1953 at the 47 cm depth is based on a '*’Cs peak identified by Caponigro (1976). The first appear- ance of Cedrus pollen at 35 cm dates to 1964 as- suming a 15 year maturation period following the first planting in 1949 (Peterson 1980). The disparity in sedimentation rate between the upper core (110 cm in 160 years) and the lower core (165 cm in 1500 years) suggests that the radiocarbon age may be artificially old. Even assuming no changes in sedimentation rate, the lower half of the core would span at least two centuries of vegetation history pri- or to Spanish settlement. MADRONO [Vol. 45 ‘(€66| Sulsuspy) Sutuueds Aq poynuapr sem ualjod wniposg a19YyM [edo] & syuasoidas (+) wnIposy JOY aduKYS JOS dy} DION “UONLIIBSeX9 soul] SAY sjUosoidor svooRIUOW[Od JopuN vore popeysuN oy ‘suiei13 poynuapriun paasasoid Aj10od Zurpnyoxa sasods pue usjj[od [e10} Jo asejusoJed ev se passaidxo oe SUING “UIseg Bleqieg BIURg sy} WOT sodAj uayjod paysayag §=*7 “DI L 0 Ol 0 Ol Ol OL 0 0c Ol O Ot 0 OV 0€ 02 Ol 0 eigen eee —> >, i seri aT s ee ewe ye —— - Bis eat us S8rl Se ee ie Oh (aire ee Lee eee ---------- BM-------- -==-- Ses} es i i in P00 irre ea eae ee Ce FO EL =5)3©|6|6m Sesl ae Ne le we — lc .l.Umlmlté“‘<éiéra se91 i ia meg | OR Rg Reena ce || Seo, ape NS Sa Se re Sh Se a aoe Me ee hia. y See ee ee See ay eee Tl ill VK ssol $8 es ae ee Eee pera fae ere a ae aes a a OOOO EEE llCT— EO EllmlmlmllmlCU aS > 5S Soni pa aes is ———E 6=h=hEMhCi RS os ROS » Co re) We) & Fe 2 x ulseg eieqieg Blues MENSING: 560 YEARS OF VEGETATION CHANGE 1998] ‘snapay pure ‘snjddjponyq ‘wniposq JOJ aBuvyo geOs oY) ION “UONeIasexa sou dAY syUdsoido1 ovsdROISSeIg JopuN vole popeYysUN sy ‘suIeIs paynuaplun paAsosoid Aj10od pue (snuvivjq pue x1j/0S) exe) uvuedu ‘(piddny pure ‘avaoeviadAD ‘vyd&z) sonenbe Surpnyoxo sasods pue usyjod [e10) Jo asejusdIod & se passoidxa oie sung “ayey] voVZ Woy sadAj uat[od payoa[a¢g O0¢ O¢ OF OOF O Ob Ok O T T Is s05t 0 7. O20b--<0 a re te ee cre nurs fey ee ae i ee ae EY eee eee 0 0c Ok O 0S Ov OF O2 Ol RY 0 vid iz SEP(199)P6L ‘O'"E 0S2 Goo | 002 GLI OSI Gol wo ul ujdeq me OV /+ 0¢8I AV OOl GZ 0S lel C -/+ EO6T “CV me C /+ 196 AV Go oye] eoeZ 6 MADRONO Taxonomy. Fifty-four pollen and spore types were identified (Mensing 1993), of which ten taxa are graphed in Figure 3. Species of oaks in the re- gion that most likely contribute pollen to the site include Q. agrifolia, Q. lobata, Q. durata, Q. chry- solepis, Q. wislezenni, and Q. douglasii. Of these species Q. agrifolia and Q. lobata dominate the re- gion today. Pinus includes P. ponderosa, P. sabi- niana, and P. coulteri as well as P. attenuata, plant- ed around the turn of the century. Rhamnaceae/Ro- saceae includes taxa similar to those described for the Santa Barbara Basin as well as Cercocarpus betuloides, Prunus ilicifolia, Heteromeles arbuti- folia. Asteraceae around Zaca Lake primarily rep- resent herbaceous taxa. Brassicaceae probably in- clude introduced mustards in the post-European pe- riod. Erodium was identified as Erodium cicutar- ium. Eucalyptus is no longer present at the site and the type is unknown. Cedrus is from Cedrus deo- dora planted near the lake. Pollen analysis. Twenty-six levels were analyzed at approximately 10 cm intervals. Six levels were excluded from the analysis because of extremely high percentages of Asteraceae pollen presumably associated with erosion events. Dense clay layers are present in the core at 130-150 cm and 200- 240 cm depth. These lenses are associated with above average magnetic susceptibility and gener- ally low organic content (Mensing 1993). High magnetic susceptibility readings generally result from deposition of iron bearing sediments. Peterson (1980) hypothesized that these layers may be as- sociated with periods of higher than average ero- sion. Asteraceae pollen is particularly resistant to biodegradation and during periods of high runoff, pollen accumulated on the soil surface may have been washed into the lake biasing the sample. For most of the record, Quercus pollen remains between 40-50%. Low percentages are seen at the 50-60 cm 120 cm and 170-180 cm depths. The declines are mirrored by increases in Asteraceae. These strata are not clay layers; however, they show above average magnetic susceptibility suggesting that they may also be associated with erosion events. Organic rich lake sediments (gytja) com- monly had 50% Quercus pollen. In the upper part of the record, from about 1970 to the present, Quer- cus averages 35%. Unlike previous decreases in Quercus, Asteraceae also declines during this time. Pinus values remain stable at about 10% through most of the core then begin to increase rapidly to 29% in the mid 1900’s. The Rhamnaceae/Rosaceae curve shows little variation averaging about 12%. Similarly, Artemisia varies little, reaching as high as 7% between 40—65 cm, but remaining at less than 4% for most of the record. Asteraceae is the most important herbaceous pol- len type. The record is highly variable, ranging from 1% to 34%. Poaceae, which remains fairly constant at about 2—3% for most of the core shows [Vol. 45 an interesting short term increase at the 60 and 65 cm levels (ca. 1920 to 1930), jumping up to 14%. Zea mays (corn) pollen was also found in the 65 cm level, suggesting a period of local cultivation. Brassicaceae is present at about 1% in the core sec- tion below 65 cm (ca. 1920). From 65 cm up to the surface level, Brassicaceae steadily increases from 1% up to 6%. Erodium first appears in the core at 110 cm depth and is present in ten levels. Eucalyp- tus 1S present in three levels with approximate dates of 1920, 1930, and 1960. Cedrus first appears in 1964 and increases in abundance in the surface samples. DISCUSSION Oak woodlands. The Santa Barbara pollen record shows no significant vegetation changes during the pre-European period. The evidence suggests that oak woodland populations remained stable for up to four centuries. Beginning around 1870 and con- tinuing until 1985, percent Quercus pollen steadily increases to its highest level in 560 years. Principal components analysis of pollen accumulation rates indicates that the abundance of Quercus pollen has indeed increased over the last century (Mensing 1993). The twofold increase of Quercus pollen dur- ing the last century strongly suggests an increase in oak woodlands in the Santa Barbara region. This increase may be from an increase in woodland den- sity, expansion of oak woodland habitat, or a com- bination of the two. The Zaca Lake record is less clear concerning oak woodlands. For most of the record, Quercus is the dominant pollen type with maxima averaging 50 percent. Periodic declines in Quercus consis- tently correspond with increases in Asteraceae. An increase in individuals of the Asteraceae, locally composed primarily of herbaceous annuals, would not displace oak woodlands. Since decreases in Quercus are not associated with increases in woody taxa, I suggest that oak populations in this area re- mained stable prior to the mid-1900’s. Since 1950 another woody taxon, Pinus, has increased substan- tially. The increase in the importance of Pinus re- corded in the pollen record is confirmed by repeat photography (Fig. 4). Scattered pine groves, visible on the distant slopes in the 1895 photograph, now appear as dense forest. Today, the understory sur- rounding the lake is thick with young pines and oaks, but pines over-top oaks in most places. Zaca Lake is located at the transition between oak wood- land and coniferous forest. Although oaks may be increasing to some extent at this site, the primary signal in the pollen record, supported by evidence from repeat photography, is an increase in the im- portance of pines. This study presents the first high-resolution pae- loecologic records to document changes in Califor- nia oak woodlands from the pre-European period to the modern period. Of significance here is that 1998] MENSING: 560 YEARS OF VEGETATION CHANGE al by a local Santa Barbara photographer ca., 1895 (courtesy Santa Barbara Historical Society Museum). The lower photograph is by the author, 1992. Q. agrifolia populations in the Santa Barbara area have increased in the recent century, after a long period with no apparent changes. Other studies, pri- marily stand age analyses of Q. douglasii, have also documented increases in woodland density (White 1966; Vankat and Major 1978; Mensing 1992); however, these studies do not extend to the pre- European period. A resampling of permanent plots in northern California found that Q. douglasii have increased over the last 60 years, with live oaks be- ginning to emerge as co-dominants (Holzman 1993): In general, there is concern that California oak woodlands are in decline. A recent assessment of 8 MADRONO Q. douglasii found that 87% of the study locations were experiencing a net loss in both tree density and canopy cover (Sweicki et al. 1993). Studies have documented negative impacts of human- caused environmental change to oak woodlands throughout the state including direct loss of wood- lands through clearing (Bolsinger 1988; Rossi 1980) and poor regeneration as a result of livestock grazing, invasion of annual grasses, and other changes (Griffin 1971; Bartolome et al. 1987; Borchert et al. 1989; Harvey 1989; Gordon and Rice 1993; Muick 1995). To understand the full extent of these changes on California oak woodlands, it is valuable to have data on how current populations compare with those from the pre-European period. In this regard, paleoecologic studies provide important informa- tion to understand the long term implications of hu- man-caused environmental change. This study sug- gests that Q. agrifolia populations in the Santa Bar- bara region have increased during the modern pe- riod, a time of significant human-caused environmental change. The increase in Q. agrifolia is most apparent in the 1900’s. I believe that the environmental change most likely to have resulted in an increase of oak woodland is a change in fire regime. In the absence of fire, Q. agrifolia tends to increase. Density and canopy cover for Q. agrifolia at Burton Mesa in Santa Barbara County was found to be highest on sites without recent fires (Davis et al. 1988). McBride (1974) examined plant succession in the Berkeley hills and suggested that in the absence of recurrent fires, Q. agrifolia and Umbellularia cali- fornica would succeed Baccharis pilularis. In a comparison of vegetation dynamics on burned and unburned plots at Gaviota State Park west of Santa Barbara, Callaway and Davis (1993) found that chaparral was being converted to oak woodland at a rate of 0.12% per year in the absence of fire. They predicted that with the absence of fire and grazing, oak woodland would dominate a larger proportion of the landscape. The Chumash regularly set fires along the coastal plain, and this practice continued even after estab- lishment of the missions (Timbrook et al. 1982). Many of these grass fires probably burned through the understory of adjacent oak woodlands, killing oak seedlings and saplings. Trees of less than 7.5 cm diameter breast height have bark approximately 0.6 cm thick and may be killed by low intensity fires (Plumb and Gomez 1983). This process would have maintained open oak woodlands similar to the oak parks typically described by early Spanish ex- plorers. Since the turn of the century, urban and agricultural development has concentrated in areas dominated by grassland and oak woodland. Al- though urban and agricultural development have been responsible for clearing oaks, fire protection in developed areas favors oaks in nearby wildland settings (Davis et al. 1988; Callaway and Davis [Vol. 45 1993). The Santa Barbara Basin pollen record sug- gests that the last 100 years have produced such an increase in oaks in the Santa Barbara region. Reduction in fire frequency may also be respon- sible for the recent increase in woodland and shrub cover at Zaca Lake. Fires have been systematically recorded in the Los Padres National Forest since 1911. Three fires have burned on the chaparral slopes to the northwest of the lake; however, no fires larger than a few acres have burned the wood- ed slopes in the upper Zaca Lake watershed (Los Padres National Forest Fire Statistical Database). Here, absence of fire appears to have favored pines over oaks. Zaca Lake is located at the pine/oak eco- tone. The tendency for pine to invade oak wood- land following fire suppression has been clearly demonstrated in Yosemite Valley where open oak meadows were converted to closed coniferous for- est after fire suppression (Reynolds 1959; Gibbens and Heady 1964; Anderson and Carpenter 1991). The Zaca Lake pollen record suggests that at upper elevation sites where coast live oak grows with pines, coniferous forest will replace oak woodland in the absence of frequent fire. Chaparral, coastal-sage scrub, and herbaceous vegetation. There is some debate concerning the impact of European settlement on chaparral, coast- al-sage scrub, and herbaceous vegetation. Dodge (1975) argued that grassland was much more ex- tensive during the pre-European period because fre- quent low-intensity fires cleared out young shrub seedlings. He concluded that heavy grazing and fire suppression have reduced low-intensity fires and permitted shrub invasion of vast areas formerly dominated by grasses. Timbrook et al. (1982) ech- oed this sentiment and concluded that chaparral has increased in density and extent over the last 200 years because of suppression of grassland burning. Furthermore, they suggested that a grassland which dominated the Santa Barbara coastal plain and foot- hills has been largely replaced by coastal-sage scrub as a result of fire suppression. The pollen record does not support the idea that chaparral has expanded over the last 200 years. The Rhamnaceae/Rosaceae curve from each site show virtually no consistent trends (Figs. 2, 3). If any- thing, the Santa Barbara Basin diagram shows a modest decline in chaparral taxa in the recent cen- tury. The pollen record suggests that chaparral has not expanded its range in response to European im- pacts. There is some evidence to suggest a modest in- crease in the importance of coastal-sage scrub over the last 200 years. Artemisia averages 7% of the pollen sum from the period between 1425 and 1820 (Fig. 2). However, beginning in 1820, it increases to 12%, and averages 10% between 1820 and 1985. Pollen percentage remains at the higher levels ex- cept for two brief declines centered on 1920 and 1980. At Zaca Lake Artemisia averages 2.7% prior 1998] to about 1800, then increases to an average of 5.0% in the upper core (Fig. 3). Although this may rep- resent a true increase, it is difficult to interpret too much from such a small change. Comparison of burned and unburned plots in Santa Barbara County found that coastal-sage scrub invaded grassland in the absence of fire, but fre- quent fire favored grassland (Callaway and Davis 1993). Westman (1976) also found that coastal-sage scrub replaced undisturbed grassland when fire was removed. In northern California, Baccharis pilular- is was found to invade grassland during periods of low fire frequency (McBride and Heady 1968). Re- duced fire frequency along the coastal plain may have favored a slight expansion of coastal-sage scrub; however, there is no evidence that this im- pact affected the distribution or abundance of chap- arral. Pollen evidence of herbaceous taxa shows that the invasion of alien species into grasslands began very early. Erodium first appears in the pollen rec- ord in 1760, nearly a decade prior to the first Span- ish settlement in San Diego and more than 20 years before the founding of the Mission Santa Barbara (Fig. 2). The pollen has been identified as Erodium cicutarium (Mensing and Byrne in press), a Medi- terranean native, and provides evidence that the in- vasion and transformation of herbaceous vegetation began prior to European settlement. Polemoniaceae averages nearly 1% and is consistently present through the 1700’s. Asteraceae, the dominant her- baceous pollen type, averages 18% in the pre-Eu- ropean period. Both taxa decline markedly in the modern period. The decline becomes particularly pronounced after the arrival of alien Brassicaceae which became widespread along the coastal plain in the early 1800’s (Cleland 1951). CONCLUSIONS The evidence from this study suggests that oak woodlands in the Santa Barbara region have in- creased during the last 100 years. The nature of this increase varies between sites. In the Santa Barbara area, Q. agrifolia appears to have increased begin- ning in the late nineteenth century. Fire suppression on the coastal plain has probably been the main factor contributing to this increase. The Chumash are reported to have periodically set fires; however, with an increase in settlement and development, burning has been suppressed. In the absence of fire, Q. agrifolia has increased in density. A policy of fire suppression appears to favor Q. agrifolia, and where fire return intervals are long, oaks would be expected to continue to increase in density. At higher elevations where Q. agrifolia grow alongside pine, such as at Zaca Lake, fire suppres- sion appears to have favored pine over oak. Here, coniferous forest is expanding into oak woodland. The taller pines may eventually shade out the slower growing oaks if the present trend continues. MENSING: 560 YEARS OF VEGETATION CHANGE 9 The pollen record shows that prior to European settlement, oak populations had been stable for at least three centuries. In the past two centuries, oak populations have changed in response to European impacts, including the introduction of grazing, sup- pression of fire, and a shift in understory compo- sition. In some cases, these changes appear to have favored oaks, creating woodlands more dense than during the pre-European period. Chaparral does not appear to have expanded sig- nificantly in response to European land use changes. Coastal-sage scrub may have expanded some; however, the evidence for this change is weak. Invasion and transformation of grassland ap- pears to have begun particularly early with the first alien taxa reaching the area even before the first Spanish settlement of California. ACKNOWLEDGMENTS I thank Roger Byrne and Eric Edlund for assistance in the field and laboratory and for reviewing earlier versions of this paper and three anonymous reviewers for their valuable comments. I am particularly indebted to Arndt Schimmelmann and Carina Lange for samples and devel- opment of the Santa Barbara Basin core chronology. This work was supported by the National Science Foundation, The California Department of Forestry Integrated Hard- wood Range Management Program, and Sigma Xi. LITERATURE CITED ANDERSON, R. S. AND S. L. CARPENTER. 1991. Vegetation change in Yosemite Valley, Yosemite National Park, California, during the protohistoric period. Madrofio 38:1-13. ANDERSON, M. V. AND R. L. PASQUINELLI. 1984. Ecology and management of the northern oak woodland com- munity, Sonoma County, California. M.A. thesis. Sonoma State University, Rohnert Park, CA. BARTOLOME, J. W., P. C. Muick, AND M. P. MCCLARAN. 1987. Natural regeneration of California hardwoods. Pp. 26-31 in T. R. Plumb and N. H. Pillsbury (tech. coords.), Proceedings of the Symposium on Multiple- Use Management of California’s Hardwood Re- sources. U.S. Department of Agriculture, Forest Ser- vice, Pacific Southwest Forest and Range Experiment Station, Berkeley, CA. BOLSINGER, C. L. 1988. The hardwoods of California’s timberlands, woodlands and savannas. U.S. Depart- ment of Agriculture, Forest Service, Pacific North- west Research Station, Portland, OR. BORCHERT, M. I., FE W. DAvis, J. MICHAELSEN, AND L. D. OYLER. 1989. Interactions of factors affecting seed- ling recruitment of blue oak (Quercus douglasii) in California. Ecology 70:389—404. Brown, R. W. AND E W. Davis. 1991. Historical mortality of valley oak (Quercus lobata, Nee) in the Santa Ynez Valley, Santa Barbara County, 1938-1989. Pp. 202-207 in Standiford, R. B. (tech. coord.), Proceed- ings of the symposium on oak woodlands and hard- wood rangeland management. U.S. Department of Agriculture, Forest Service, Pacific Southwest Forest and Range Experiment Station, General Technical Report PSW-126. CALLAWAY, R. M. AND E W. Davis. 1993. Vegetation dy- 10 MADRONO namics, fire, and the physical environment in Central California. Ecology 74:1567—1578. CAPONIGRO, M. A. 1976. The diatom stratigraphy and pa- leolimnology of Zaca Lake, California. M.A. thesis. University of California, Santa Barbara. CLELAND, R. G. 1951. The cattle upon a thousand hills. Huntington Library Publications, San Marino, CA. Cote, K. L. AND G. Liu. 1994. Holocene paleoecology of an estuary on Santa Rosa Island, California. Quater- nary Research 41:326—335. Davis, E W., D. E. HICKSON, AND C. O. DENNIS. 1988. Composition of maritime chaparral related to fire his- tory and soil, Burton Mesa, Santa Barbara County, California. Madrono 35:169-195. DopcE, J. M. 1975. Vegetational changes associated with land use and fire history in San Diego County. Ph.D. dissertation. University of California, Riverside. FAEGRI, K. AND J. IVERSEN. 1975. Textbook of Pollen Analysis, 3rd ed. Hafner Press, New York. FERREN, W. R. 1985. Carpinteria Salt Marsh: Environ- ment, History, and Botanical Resources of a Southern California Estuary. The Herbarium, Department of Biological Sciences, University of California, Santa Barbara. GIBBENS, R. P. AND H. A. HEADY. 1964. The influence of modern man on the vegetation of Yosemite Valley. University of California, Division of Agricultural Sci- ences Manual 36. Gorpbon, D. R. AND K. J. RIcE. 1993. Competitive effects of grassland annuals on soil water and blue oak (Quercus douglasii) seedlings. Ecology 74:68-82. GORDON, D. R., J. R. WELKER, J. W. MENKE, AND K. J. Rice. 1989. Competition for soil water between an- nual plants and blue oak (Quercus douglasii) seed- lings. Oecologia 79:533-541. GRIFFIN, J. R. AND W. B. CRITCHFIELD. 1976. The distri- bution of forest trees in California. U.S. Department of Agriculture, Forest Service, Pacific Southwest For- est and Range Experiment Station, Research Paper PSW-82. GRIFFIN, J. R. 1971. Oak Regeneration in the Upper Car- mel Valley, California. Ecology 52:862—868. HALL, C. A. 1981. Map of geology along the Little Pine Fault, parts of the Sisquoc, Foxen Canyon, Zaca Lake, Bald Mountain, Los Olivos, and Figueroa Mountain quadrangles, Santa Barbara County, Cali- fornia. U.S. Geologic Survey Miscellaneous Field Studies. Map MF-1285. HARVEY, L. E. 1989. Spatial and temporal dynamics of a blue oak woodland. Ph.D. dissertation, University of California, Santa Barbara. HEUuSSER, L. E. 1978. Pollen in Santa Barbara Basin, Cal- ifornia: A 12,000 year record. Geological Society of America Bulletin 89:673—678. HICKMAN, J. C. (ed.) 1993. The Jepson Manual: Higher Plants of California. University of California Press, Berkeley. HOLZMAN, B. A. 1993. Vegetation change in California’s blue oak (Quercus douglasii) woodlands, 1932-1992. Ph.D. dissertation. University of California, Berkeley. HULSEMANN, J. AND K. O. Emery. 1961. Stratification in recent sediments of Santa Barbara Basin as controlled by organisms and water character. Journal of Geology 69:229-—290. Kobe, M. A., A. SOUTAR, AND E. D. GOLDBERG. 1972. Marine geochronology with Pb-210. Earth and Plan- etary Science Letters 14:442—446. KRISHNASWAMI, S., D. LAL, B. S. AMIN, AND A. SOUTAR. [Vol. 45 1973. Geochronological studies in Santa Barbara Ba- sin. Limnology and Oceanography 18:763-—770. MAYFIELD, D. W. 1981. Ecology of a discovered land. Pa- cific Discovery. Pp. 12-20. McBripE, J. R. 1974. Plant succession in the Berkeley Hills, California. Madrofio 22:317—329. McBrIDE, J. R. AND H. EK HEADY. 1968. Invasion of grass- land by Baccharis pilularis. Journal of Range Man- agement 21:106—108. McCLaraN, M. P. 1986. Age structure of Quercus doug- lasii in relation to livestock grazing and fire. Ph.D. dissertation. University of California, Berkeley. McCLaran, M. P. AND J. W. BARTOLOME. 1989. Fire-re- lated recruitment in stagnant Quercus douglasii pop- ulations. Canadian Journal of Forestry 19:580—585. MENSING, S. A. 1992. The impact of European settlement on blue oak (Quercus douglasii) regeneration and re- cruitment in the Tehachapi Mountains, California. Madrono 19:36—46. MENSING, S. A. 1993. The impact of European settlement on oak woodlands and fire: Pollen and charcoal evi- dence from the Transverse Ranges, California. Ph.D. dissertation. University of California, Berkeley. MENSING, S. AND BYRNE, R. in press. Pre-mission invasion of Erodium cicutarium in California. Journal of Bio- geography. Mutck, P. C. 1995. Effects of shade and simulated herbiv- ory on seedling emergence, survival and growth of two California oaks: Coast live oak (Quercus agri- folia Nee) and blue oak (Quercus douglasii Hook & Arn). Ph.D. dissertation. University of California, Berkeley. Mulick, P. C. AND J. R. BARTOLOME. 1987. An assessment of natural regeneration of oaks in California. Final Report, California Department of Forestry: The For- est and Rangeland Assessment Program. PETERSON, S. 1980. Late Holocene sedimentation at Zaca Lake, Santa Barbara County, California. B.A. senior thesis, University of California, Berkeley. PLUMB, T. R. AND A. P. GOMEZ. 1983. Five southern Cal- ifornia oaks: identification and postfire management. U.S. Department of Agriculture, Forest Service, Pa- cific Southwest Forest and Range Experiment Station, General Technical Report PSW-71. REYNOLDS, R. D. 1959. Effect of natural fires and aborig- inal burning upon the forests of the central Sierra Ne- vada. M.A. thesis, University of California, Berkeley. Ross, R. S. 1980. History of cultural influences on the distribution and reproduction of oaks in California. Pp. 7-18 in T. R. Plumb and N. H. Pillsbury (tech. coords.), Proceedings of the Symposium on the Ecol- ogy, Management, and Utilization of California Oaks. U.S. Department of Agriculture, Forest Service, Pa- cific Southwest Forest and Range Experiment Station, General Technical Report PS W-44. SCHIMMELMAN, A., C. B. LANGE, AND W. H. BERGER. 1990. Climatically controlled marker layers in Santa Bar- bara Basin sediments, and fine-scale core-to-core cor- relation. Limnology and Oceanography 35:165—173. SCHIMMELMAN, A., C. B. LANGE, W. H. BERGER, A. SIMON, S. K. BURKE, AND R. B. DUNBAR. 1992. Extreme cli- matic conditions recorded in Santa Barbara Basin laminated sediments: the 1835-1840 Macoma Event. Marine Geology 106:279—299. SOUTAR A. AND P. A. CRILL. 1977. Sedimentation and cli- matic patterns in the Santa Barbara Basin during the 19th and 20th centuries. Geological Society of Amer- ica Bulletin 88:1161-—1172. 1998] STOCKMARR, J. 1971. Tablets with spores used in absolute pollen analysis. Pollen et Spores 13:615—621. STUVIER, M. AND P. J. REIMER. 1986. A computer program for radiocarbon age calibration. Radiocarbon 28: 1022-1030. SwEICKI, T. J., BERNHARDT, E. A., AND DRAKE, C. 1993. Factors affecting blue oak sapling recruitment and re- generation. California Department of Forestry, Sac- ramento, CA. Final Report on contract 8CA17358. TIMBROOK, J., J. R. JOHNSON, AND D. D. EARLE. 1982. MENSING: 560 YEARS OF VEGETATION CHANGE 11 Vegetation burning by the Chumash. Journal of Cal- ifornia and Great Basin Anthropology 4:163—186. VANKAT, J. L. AND J. L. MAsor. 1978. Vegetation changes in Sequoia National Park, California. Journal of Bio- geography 5:377—402. WESTMAN W. E. 1976. Vegetation conversion for fire con- trol in Los Angeles. Urban Ecology 2:119—137. WHite, K. L. 1966. Structure and composition of foothill woodland in Central Coastal California. Ecology 47: 229-237. MADRONO, Vol. 45, No. 1, pp. 12-16, 1998 SUCCESSIONAL PATTERNS OF THE LOWER MONTANE TREELINE, EASTERN CALIFORNIA TREVOR BURWELL Geography Department, University of Wisconsin—Madison, Madison, WI 53706 ABSTRACT Stand age patterns of pinyon woodlands along the lower montane treeline ecotone (LMTE) in eastern California suggest that prior to European settlement, woodlands were very open (41 trees/ha) and were mostly restricted to xeric topographic settings with shallow, coarse-textured soils. Since European settle- ment in about 1861, pinyon woodlands have rapidly increased in density, expanded downslope, and invaded more mesic topographic sites. The lower montane treeline ecotone (LMTE) forms the lower elevation extent of montane veg- etation communities in the Intermountain West. In Inyo and Mono Counties the LMTE is dominated primarily by open woodlands of Pinus monophylla Torr. and Frem. (single-needle pinyon pine) with an understory of predominantly Great Basin shrub species such as Artemisia spp. (sagebrush), bitter- brush (Purshia spp.), and rabbitbrush (Chrysotham- nus spp.) (Vasek and Thorne 1977). The pinyon woodland occurs mostly between 1900 m—2900 m elevation. In the eastern Sierra Nevada, Sweetwater Mountains, and Glass Mountain Ridge, the upper elevations of pinyon grade into montane forests and woodlands of Jeffrey pine (Pinus jeffreyi), lodge- pole pine (Pinus murrayana), and white fir (Abies concolor). In the White Mountains and Bodie Hills, a sagebrush shrubland dominates the cover imme- diately above the pinyon woodlands. In the White- Inyo Range, subalpine woodlands of bristlecone pine (Pinus longaeva) and limber pine (Pinus flex- ilis) occur above the upper sagebrush zone (St. An- dre et al. 1965; Spira 1991; Vasek and Thorne 1977). The pinyon woodland in eastern California is un- usual in the rarity of Utah juniper (Juniperus os- teosperma), which commonly codominates with pinyon to form a pinyon-juniper woodland through- out the central Great Basin (West et al. 1978). Utah juniper co-occurs with pinyon in eastern California primarily in the White-Inyo Range where it is largely restricted to sites on alluvial soils (St. Andre 1962; Mooney 1973). Of the 521 trees sampled in this study, only two are Utah juniper, and the re- mainder are single-needle pinyon. Throughout the Intermountain and Rocky Moun- tain west, observers have noted that the lower mon- tane woodlands have been expanding downslope and have become denser within their elevational range since the late 1800’s (Cottam and Stewart 1940; Blackburn and Tueller 1970; Burkhardt and Tisdale 1969, 1976; Rogers 1982; Vale 1975). Ear- ly observations, historic photographs, and stand age data suggest that lower montane woodlands were formerly savanna-like, with trees restricted to rocky outcrops or steep slopes with coarse-textured soils. Many of these topographic sites with existing trees have subsequently converted to fully-stocked stands while much of the range expansion of pin- yon and juniper has been onto gently sloping or valley bottom sites with deeper, finer-textured soils (Cottam and Stewart 1940; Burkhardt and Tisdale 1969, 1976; Miller 1921; Miller and Rose 1995; Phillips 1909; Woodbury 1947). While the ecology of pinyon and juniper wood- lands has been extensively investigated, patterns and processes along the LMTE have received little study. Most previous workers have been concerned with larger-scale relationships of mature wood- lands, and so have avoided sampling and describing ecotonal or early successional stands (Koniak 1986; Meeuwig and Cooper 1981; Tausch et al. 1981; Tueller et al. 1979; West et al. 1978). Blackburn and Tueller (1970) described and classified the range expansion of single-needle pinyon and Utah juniper in eastern Nevada, but did not investigate differences in stand age and dominance in relation- ship to the landscape topography. In this work I present pinyon age and dominance data in relation to topographic settings along the LMTE. METHODS The study areas are all within Inyo and Mono Counties, California; and sites were located in the Bodie Hills, Glass Mountain Ridge, the White Mountains, the Sherwin Summit area just east of the Sierra Nevada, and at Tuttle Creek, at the base of the southern Sierra Nevada. Sites were selected where the LMTE occurs in each area. Elevations ranged from 1900-2000 m elevation. Sampling ar- eas were usually small drainage basins. Plots were randomly placed on a variety of topographic posi- tions with a maximum of 50 m elevation difference within a sampling area. At this small scale, it was assumed that soil moisture variation is primarily a 1998] function of soil texture and topographic position, and that the influence of elevation is a constant. The LMTE is highly interdigitated with respect to topographic position. The woodland is more de- veloped and extends further downslope on topo- graphic convexities, such as ridges, and steep slopes. To measure the influence of topographic po- sition on stand variables, the sampling frame was stratified based on topographic position using the Topographic Relative Moisture Index (TRMI) (Par- ker 1982). The TRMI is a scalar index (O—60, 0 = most xeric topographic sites, 60 = most mesic) de- signed for mountainous terrain in the western Unit- ed States, and offers a way to stratify the landscape into topographic position based on slope angle, configuration, aspect, and position on the slope. In- dividual plots were placed randomly within each landscape unit, such as valley bottom, ridge top, toe slope, and upper slope. This stratified random sam- pling process sometimes resulted in the placing of field plots where trees were absent; however, a sam- pling frame based on vegetative characteristics would not reflect the topographic influence on stands. Each plot was checked visually to insure that it did not include anomalous vegetative cover relative to the rest of the landscape unit. To examine the hypothesis that pinyon existed as open woodlands primarily on xeric sites prior to European settlement, and then invaded onto mesic sites, I separated the plots into 22 xeric dow TRMI, <30) and 22 mesic (high TRMI, >30) sites. While there is a gradation between xeric and mesic sites, the combined weight of the slope variables (topo- graphic position, slope steepness, and slope config- uration) made them more important than aspect in determining whether a site was classified as mesic or xeric. In general, sites described as xeric are mostly located on topographic convexities and steeper slopes, and the mesic sites are located in draws, valley bottoms and other topographic con- cavities. Soil variables are not incorporated into the TRMI; however, within each study site, there is a strong positive correlation between shallow, coarse- textured soils and xeric sites as determined by the TRMI excluding aspect (Vaughn 1983). Plots were 30 m X 30 m in all areas except the White-Inyo Range, where 50 m X 20 m plots were used to be consistent with previous field work (St. Andre 1962). All lives trees rooted in the plot >1.5 m height were cored at 30 cm for aging, and all trees were measured for basal area, crown projec- tion area, and height. Understory cover was mea- sured by the line-intercept placement of a 30 m transect through the plot. Percent cover was cal- culated as the percent distance intersected vertically by the line. In each study area, cross-sections were cut from juvenile trees <1.5 m height to determine the age to coring height. The average age at 30 cm was found to be 12.24 yrs (n = 30), so 12 years were added to the ring count age of each tree. To gain a representation of the age patterns of juve- BURWELL: LOWER TREELINE SUCCESSION 13 niles <1.5 m height, a simple linear regression equation was used to predict their ages. The best equation to predict age was based on stem radius and tree height (both measured in centimeters) us- ing data from the cross-sectioned trees and cored trees <3 m height. The simple linear regression equation is Age = 8.19(radius) + 0.169(height); n = 122, r = 65%, p < 0.0005. Growth rates are highly variable for immature trees due to their greater sensitivity to soil moisture fluc- tuations (Barton 1993), variations in overstory can- opy shade, and the potential influences of nurse shrubs (Drivas and Everett 1988). Consequently, the predicted ages of juvenile trees are not reliable in determining actual dates of recruitment or seed- ling establishment, but they do offer a relatively accurate picture for interpreting overall reproduc- tive status of the stand. Finally, while there is no evidence of historic wood cutting on the sites, such as stumps, and the sampled stands are not within the known wood- sheds of historic mining centers in the area, the possibility exists that some mature trees may have been cut for wood use in the past, and are now absent from the stands. RESULTS AND DISCUSSION At the LMTE, topographic position explains more variation in pinyon age and dominance pat- terns than elevation or aspect. Figures la and b 1i- lustrate tree ages from the 44 plots, and Table 1 summarizes stand age and density data. Of 381 trees aged by ring counts and 140 juvenile ages predicted through linear regression, 23% on all sites, 29% on xeric and 13% on mesic sites, predate European settlement. The mean (and median) tree ages of all sites today is 97.2 (79) years, on xeric sites 106.6 (81) years, and on mesic sites 83.6 (78) years. Currently and in 1861, pinyon stands on xer- ic sites are significantly older than trees on mesic sites along the LMTE (two-tailed t-test, p < 0.005). There was no difference in statistical significance when ages from cored only, or cored and predicted juvenile ages, were used. Prior to the establishment of permanent Euro- pean ranching and mining settlements in Inyo and Mono Counties in 1861 (Chalfant 1933; Sauder 1994), the data suggest that pinyon woodlands were largely restricted to xeric habitats, such as topo- graphic convexities and steep slopes with thin, rocky soils. From stand age reconstructions, xeric sites supported an average of 45 trees/ha in 1861. Draws and other topographic concavities with deeper, finer-textured soils supported 13 trees/ha on average. Today, xeric sites support 156 trees/ha, and mesic sites 107 trees/ha. Estimated average tree ages in 1861, however, show stronger contrasts in tree dominance: 41 years in mesic sites, with one 14 MADRONO [Vol. 45 Figure 1a. Tree establishment dates, high TRMI (mesic) sites. 35 a6 MI Age by ring count oe LI Age predicted o = 20 @® JS o 15 LL 10 5 ell n x) n x) wn n x) x?) Nn x) A Oo oO (=) (jo) oO Oo j=) oO oO oO a” oO oO oO oO nS vt = foe] w N o = 2 ss a = = = Be = e © Figure 1b. Tree establishment dates, low TRMI (xeric) sites. by ring count predicted >) Oo Cc ® =) oO 2 Le n oO iS foe) bead a2) x2) nN Oo (2) oO oO a <2) ise) (>) for) (op) [o>) fo?) = bcd r r 1840's 1810's 1780's 1750's 1720's 1690's> Decade of establishment Fics. la, b. Tree establishment dates of pinyon by decade, mesic (TRMI > 30) and xeric (TRMI < 30) sites along the lower montane treeline ecotone, Inyo and Mono Counties, California. tree (0.5/ha) over 100 years old; and 92 years on xeric sites, with 33 trees (17 trees/ha) over 100 years. The average 41 year old pinyon today is about 125—150 cm tall, which is similar to the av- erage shrub height in most of the plots. With 17/45 trees per hectare over 100 years old on xeric sites in 1861, ridges and slopes may have visually ap- peared as open, savanna-like stands consisting of predominantly mature, widely-spaced trees. Valley bottoms and draws may have appeared to be wholly absent of trees in 1861. Both xeric and mesic sites have had considerable seedling recruitment since European settlement. In the last 135 years, estimated tree density has in- creased 400% on xeric sites and 800% on mesic sites. This is consistent with early observations that lower montane woodlands were primarily restricted to topographic convexities prior to 1861, and that existing stands were formerly more open and con- tained largely mature trees. As pinyon cover has increased on all topographic settings, understory shrub, grass and herb cover has decreased. Current understory total vegetation cover in xeric sites is 14% on average, whereas mesic sites support 31% understory cover. While these values of understory cover may also reflect different soil moisture re- gimes and edaphic settings, and not simply rela- tionships to woodland canopy cover or competition with pinyon, the negative relationship between un- derstory vegetation cover and tree density or can- 1998] PINYON STAND DATA FOR XERIC AND MEsIc SITES AT THE LMTE.* TABLE 1. Mature trees, 1861 Tree density Mean age (years) (trees >100 years) Number of trees Current 1861 No. Density Current 1861 1861 Current Xeric sites 17/ha 106.6 45/ha 156/ha il 309 (TRMI < 30) Mesic sites BURWELL: LOWER TREELINE SUCCESSION Es) 83.6 41 0.5/ha 8.8/ha 13/ha 29/ha 107/ha 28 212 (TRMI > 30) All sites 34 80 O72 130.3/ha 119 * For all measures, stands on xeric sites are significantly different than mesic stands (p < 0.01). opy cover is well-documented (Austin 1987; Ev- erett and Koniak 1981; Pieper 1990). At the larger scale of the entire elevational gra- dient of mountain ranges in the Great Basin, a coarse-grained view would show the altitudinal zo- nation of vegetation is largely influenced by effec- tive soil moisture as determined by climatic vari- ables, such as average precipitation and tempera- ture (Daubenmire 1943; Billings 1951). However, a finer-grained view of the small scale topographic relationships along the LMTE suggests an inverse relationship of pinyon to topographically influenced moisture patterns. Greater pinyon stand develop- ment on xeric topographic settings along the LMTE is contrary to previous reports that montane tree species become increasingly restricted to mesic to- pographic settings towards the lower ecotone (Par- ker 1980; Peet 1988; Rourke 1988; Whittaker and Niering 1965). Two factors may have produced this inverse pattern: shrub and grass competition prior to 1861 may have excluded trees from the sites with deeper, finer-textured soils and more favorable moisture status; and, the greater density and bio- mass of shrubs and grasses may have supported more frequent fires, causing the mortality of inva- sive trees (Burkhardt and Tisdale 1969, 1976; Cot- tam and Stewart 1940; Blackburn and Tueller 1970). The introduction of livestock grazing in the West reduced competitive cover and fuel to carry fires, which may have allowed trees to become es- tablished, and then survive into maturity. ACKNOWLEDGEMENTS The University of California White Mountain Research Station awarded research grants in 1994 and 1995 to fund fieldwork. Two manuscript reviewers, Terry Hicks of the Inyo National Forest, Anna Halford of the Bureau of Land Management in Bishop, and others at both agencies, pro- vided suggestions and assistance. LITERATURE CITED AusTIN, D. J. 1987. Plant community changes within a mature pinyon-juniper woodland. Great Basin Natu- ralist 47:96—99. BARTON, A. M. 1993. Factors controlling plant distribu- tions: drought, competition, and fire in montane pines in Arizona. Ecological Monographs 63:367-397. BILLINGS, W. D. 1951. Vegetational zonation in the Great Basin of western North America. /n Les bases ecol- ogiques de la regeneration de la vegetation des zones arides. International Colloquium of the International Union Biological Sciences, Series B 2:101-122. BLACKBURN, W. H. AND P. T. TUELLER. 1970. Pinyon and juniper invasion in black sagebrush communities in east-central Nevada. Ecology 51:841-—848. BURKHARDT, J. W. AND E. W. TISDALE. 1969. Nature and successional status of western juniper vegetation in Idaho. Journal of Range Management 22:264—270. . 1976. Causes of juniper invasion in southwestern Idaho. Ecology 57:472—484. CHALFANT, W. A. 1933. The Story of Inyo. Chalfant Press, Bishop, CA. CoTTAM, W. AND G. STEWART. 1940. Plant succession as 1G MADRONO a result of meadow dessication by erosion since set- tlement in 1862. Journal of Forestry 38:613-—626. DAUBENMIRE, R. F. 1943. Soil temperature versus drought as a factor determining lower altitudinal limits of trees in the Rocky Mountains. Botanical Gazette 105: 1-13. Drivas, E. P. AND R. L. EVERETT. 1988. Water relations characteristics of competing singleleaf pinyon seed- lings and sagebrush nurse plants. Forest Ecology and Management 23:27-37. EVERETT, R. L. AND S. KONIAK. 1981. Understory vege- tation in fully stocked pinyon-juniper stands. Great Basin Naturalist 41:467—476. KOoNnIAK, S. 1986. Tree densities on pinyon-juniper wood- land sites in Nevada and California. Great Basin Nat- uralist 46:178—184. MEEvuwiIG, R. O. AND S. V. CoOoPER. 1981. Site quality and growth of pinyon-juniper stands in Nevada. Forest Science 27:593-601. MILLER, F H. 1921. Reclamation of grasslands by Utah juniper on the Tusayan National Forest, Arizona. Journal of Forestry 19:647-651. MILLER, R. F AND J. A. Rose. 1995. Historic expansion of Juniperus occidentalis (Western Juniper) in south- eastern Oregon. Great Basin Naturalist 55:37—45. Mooney, H. A. 1973. Plant communities and vegetation. Pp. 7-17 in R. M. Lloyd and R. S. Mitchell (eds.), A flora of the White Mountains, California and Nevada. University of California Press, Berkeley, CA. PARKER, A. J. 1980. Site preferences and community char- acteristics of Cupressus arizonica Greene (Cupres- saceae) in southeastern Arizona. Southwestern Natu- ralist 25:9-—22. . 1982. The topographic relative moisture index: an approach to soil-moisture assessment in mountain terrain. Physical Geography 3:160—168. PEET, R. K. 1988. Forests of the Rocky Mountains. Pp. 64-101 in M. G. Barbour and W. D. Billings (eds.), North American terrestrial vegetation. Cambridge University Press, NY. PHILLIPS, FE J. 1909. A study of pifion pine. Botanical Ga- zette 48:216—223. PIEPER, R. D. 1990. Overstory-understory relations in pin- yon-juniper woodlands in New Mexico. Journal of Range Management 43:413-—415. RoGERS, G. F 1982. Then and now: a photographic history of vegetation change in the central Great Basin Des- ert. University of Utah Press, Salt Lake City, UT. RourkE, M. D. 1988. The tree community of the Golden Trout Wilderness. Pp. 148-160 in C. A. Hall, Jr. and V. Doyle-Jones (eds.), Plant biology of eastern Cali- [Vol. 45 fornia, natural history of the White-Inyo Range Sym- posium, Volume 2. University of California White Mountain Research Station, Los Angeles, CA. SAUDER, R. A. 1994. The lost frontier: water diversion in the growth and destruction of Owens Valley agricul- ture. University of Arizona Press, Tucson, AZ. SpIRA, T. P. 1991. Plant zones. Pp. 77-86 in C. A. Hall, Jr. (ed.), Natural history of the White-Inyo range. University of California Press, Berkeley, CA. ST. ANDRE, G. J. 1962. An ecological study-the pinyon- juniper zone of the White Mountains of California. M.A. thesis. University of California, Los Angeles, Los Angeles, CA. , H. A. MOOoNEy, AND R. D. WRIGHT. 1965. The pinyon woodland zone in the White Mountains of California. American Midland Naturalist 73:225—239. TAUuSCH, R. J. AND P. T. TUELLER. 1990. Foliage biomass and cover relationships between tree- and shrub-dom- inated communities in pinyon-juniper woodlands. Great Basin Naturalist 50:121-—134. , N. E. WEST, AND A. A. NABI. 1981. Tree age and dominance patterns in Great Basin pinyon-juniper woodlands. Journal of Range Management 34:259-— 264. TUELLER, P. T., C. D. BEESON, R. J. TAUSCH, N. E. WEsT, AND K. H. REA. 1979. Pinyon-juniper woodlands of the Great Basin: Distribution, flora, vegetal cover. U.S. Department of Agriculture Forest Service, Inter- mountain Forest and Range Experiment Station, Re- search Paper INT-229, Ogden, UT. VALE, T. R. 1975. Invasion of big sagebrush (Artemisia- tridentata) by white fir (Abies concolor) on the south- eastern slopes of the Warner Mountains, California. Great Basin Naturalist 35:319-—324. VASEK, E C. AND R. E THORNE. 1977. Transmontane co- niferous vegetation. Pp. 797-832 in M. G. Barbour and J. Major (eds.), Terrestrial vegetation of Califor- nia. John Wiley and Sons, NY. VAUGHN, D. E. 1983. Soil inventory of the Benton-Owens Valley Area, Inyo and Mono Counties, California. USDI, Bureau of Land Management. WEsT, N. E., R. J. TAUSCH, K. H. REA, AND P. T. TUELLER. 1978. Phytogeographical variation within the juniper- pinyon woodlands of the Great Basin. Great Basin Naturalist Memoirs 2:119-136. WHITTAKER, R. H. AND W. A. NIERING. 1965. Vegetation of the Santa Catalina Mountains, Arizona: a gradient analysis of the south slope. Ecology 56:429-—452. Woopsury, A. M. 1947. Distribution of pygmy conifers in Utah and northeastern Arizona. Ecology 28:113- 126: - Mapbrono, Vol. 45, No. 1, pp. 17—30, 1998 CURRENT STATUS, STRUCTURE, AND PLANT SPECIES COMPOSITION OF THE RIPARIAN VEGETATION OF THE TRUCKEE RIVER, CALIFORNIA AND NEVADA STEVEN L. CAICCO U.S. Fish and Wildlife Service, Sacramento Field Office', 3310 El Camino Avenue, Sacramento, CA 95825 ABSTRACT Although riparian areas are a critical component of biodiversity in arid lands, our knowledge of many major rivers of the western United Startes remains limited. The Truckee River of California and Nevada is typical, with a general lack of published data on its riparian vegetation. Cover type mapping of eight reaches shows that the relative proportions between natural vegetation and cultural land-use types vary. Despite impacts from logging, railroad and highway construction, and water resource development, ri- parian vegetation along the upper three reaches is currently dominated by native riparian species. In the remaining reaches large proportions of the floodplain have been converted to urban and industrial or agricultural uses, or have been disturbed and are dominated by introduced weeds. Downstream reaches have also been more affected by flow regulation, water diversions, and related impacts. The lower reaches also, however, offer the greatest opportunities for restoration and enhancement of the riparian corridor. Data is also presented on the current plant species composition and structure of natural riparian vegetation which, in conjunction with hydrological data, can help land managers and biologists to formulate strategies for wildlife habitat enhancement. Riparian areas are a critical component of bio- diversity within the arid lands of the western United States and their importance is amplified by the mi- nor proportion of the overall area which they oc- cupy (Carothers 1977). Despite an increasing em- phasis on the ecology and management of western riparian ecosystems over the past two decades (e.g., Johnson and Jones 1977; Warner and Hendrix 1984; Johnson et al. 1985; Knopf et al. 1988; Abell 1989; Clary et al. 1992), there remains a striking lack of published basic ecological data on many major western rivers. The Truckee River, one of three large rivers which drain eastward from the Sierra Nevada to sinks in western Nevada is typical, with refereed publications infrequent and of narrow focus. Papers have appeared regarding food pref- erences and demographics of beaver (Hall 1960; Busher et al. 1983), local flora (Savage 1973; Smith 1984), historical avifaunal changes along the lower Truckee River (Klebenow and Oakleaf 1984), and seed germination in Salix (Martens and Young 1992). The Truckee River drains a 3100 km? basin in the Sierra Nevada into Pyramid Lake (elevation 1160 m), Nevada (Fig. 1). From Lake Tahoe (ele- vation 1899 m), the river flows 174 km through steep mountain canyons and narrow valleys, passes highly urbanized areas near the city of Reno, and continues through high desert canyons and irrigated agricultural land in broader valleys to its terminus. Significant development of natural resources in the ' Current Address: Portland Eastside Federal Complex, 911 N.E. 11th Avenue, Portland, OR 97232-4181. Truckee River drainage began in the 1860’s with extensive logging to provide timber to the nearby mining boomtown, Virginia City, and for railroad ties and snowsheds along the route of the Central Pacific Railroad (California Department of Water Resources 1991). Much of the Lake Tahoe Basin and the surrounding area was stripped of trees and the logs were transported by flumes along the river, as well as down the river channel itself. Prior to the turn of the century, numerous dams had been built at the outlets of lakes, including one at the outlet of Lake Tahoe constructed in the early 1870’s (California Department of Water Resources 1991). Subsequent to the passage of the Reclama- tion Act of 1902, the U.S. Bureau of Reclamation became the major developer of water projects on the Truckee River. Derby Dam, downstream of Reno, was this agency’s first construction project (Fig. 1). Completed in 1905 as part of the Newlands Project, the dam and its conveyance canal were de- signed to transfer water from the Truckee River to arable land with little rainfall in the adjacent Carson River drainage. Irrigation water supplied by the project continues to be the single largest use of Truckee River water. As a result of this water diversion, a steep de- cline in the water surface elevation of Pyramid Lake began in about 1910. The lake elevation reached a low point in the late 1960’s at about 23 m below its pre-Derby Dam diversion level. In re- sponse, the lower eight km of the river channel widened and incised into its floodplain, stranding the adjacent river terraces (Born 1972; Water En- gineering & Technology, Inc. 1991). Channeliza- tion of the river during the 1960’s led to further 18 MADRONO 120° 00° 00" 40° 00° 00" “- Pyramid Lake eIUJOsIeD ePpeAON Truckee Fic. 1. the eight reaches are shown. incision downstream of Wadsworth (Glancy et al. 1972). Since about 1990, the lake elevation has been about 15 m below the pre-Derby Dam level. Federal reservoirs for water storage and flood control in the Truckee River watershed and their construction dates include Boca (1937), Prosser Creek (1962), and Stampede Reservoir (1970), all owned by the Bureau of Reclamation, and Martis Creek (1971), owned by the U.S. Army Corps of Engineers. These reservoirs combined provide about 317,000 acre-feet of usable storage. Numer- ous smaller non-Federal reservoirs and diversions are located in the Truckee River watershed. In an attempt to reduce water loss due to evapo- transpiration, beaver (Castor canadensis) were in- troduced to the drainage in the late 1940’s (Hall 15 30 45 S—“‘;:SC Location of the study area along the Truckee River, California and Nevada. Urban areas and boundaries of [Vol. 45 Marble Bluff Dam Numana Dam Dead Ox Wash Wadsworth 120° 30° 00" 39° 00" 00" 60 km 1960). Evidence of beaver activity, primarily gnawed trunks, is present throughout the river cor- ridor although serious impacts seem highly local- ized. This study was restricted to the riparian corridor along the 174 river kilometers of the mainstem of the Truckee River from Lake Tahoe Dam to Marble Bluff Dam. In this paper, I present baseline data on the vascular plant species composition, structure, and areal extent of the existing riparian vegetation of this area in order to provide a framework for future ecological research. METHODS Terminology. Throughout this paper, the terms “‘riparian”’ and “‘riparian corridor” are used in the 1998] CAICCO: TRUCKEE RIVER RIPARIAN VEGETATION functional sense of a “‘three-dimensional zone of interaction between terrestrial and aquatic ecosys- tems’’ (Gregory et al. 1991). As such, the riparian zone includes, at a minimum, the low-flow and ac- tive river channels and the inferred historic flood- plain. Because of the vertical component of the def- inition, the riparian zone upstream of State Line may also include, as a minor component, lower hil- Islopes supporting coniferous forests. It is important to note that “‘riparian’”’ as used in this paper is not synonymous with ‘‘wetland’’ in a jurisdictional sense. In most areas downstream of Reno, and es- pecially downstream of Derby Dam, flow regula- tion and water diversion have altered the natural hydrography of the river and contributed to lowered groundwater tables with a consequent shift toward a less hydrophytic vegetation. As noted earlier, this effect has been most profound in the lower eight km of the river upstream of Pyramid Lake. Corridor ratio (ha/km) 4.49 Cover type mapping. The 174 km river corridor was subdivided into eight reaches based on mor- phological, geological, and hydrological character- istics. The reaches vary in length, stream gradient, floodplain area, and the ratio between floodplain area and reach length (Table 1). The latter is an indicator of the local constraints imposed on the channel and valley floor by geomorphic features and, by extension, the width of the riparian corridor (Gregory et al. 1991). Natural vegetation and land use types within the riparian zone were mapped on acetate overlays of enlarged black-and-white aerial photography at a scale of 1:1200. Source photos for these enlargements were flown on November 4, 1991; the scale of the original photos was 1:12,000. Cover type polygons were manually delineated. The minimum mapping criterion for forested types was at least six trees (as defined below) in an area of 0.2 ha. The overlays were scanned into AUTO- CAD and areas were calculated for each cover type by polygon. Areas adjacent to, and enroute to, field sample sites were field checked for boundary and classification accuracy and revised accordingly. Be- cause of interpretation difficulties, some non-for- ested areas outside of the historic floodplain may have been incorrectly included. These inaccuracies are believed to have resulted in only minor over- estimates of the extent of the riparian zone down- stream of Reno. Area (ha) 166.37 Gradient (m/km) 6.00 Elevation change (m) 37.03 Length (km) 20.93 Data collection. Data on plant species composi- tion and abundance were collected only from nat- ural vegetation types, excepting marshes and ponds, along 31 transects at 11 sites. Sites were chosen to be representative of general conditions with the study reaches and to isolate, to the extent possible, river hydrology as a controlling variable for vege- tation from other hydrological influences (e.g., ir- rigated pastures above channel banks, unlined irri- gation canals, springs, sewage effluent runoff, and significant grading and fill placement). An addi- tional constraint was relative proximity to stream RIVER LENGTH, CHANGE IN ELEVATION, RIVER GRADIENT, RIPARIAN ECOSYSTEM AREA, AND THE RATIO OF FLOODPLAIN AREA TO RIVER LENGTH FOR THE EIGHT TRUCKEE Reach RIVER STUDY REACHES. The ratios are measures of stream gradient and width of the riparian corridor, respectively. Lake Tahoe outlet dam lies at 1899 m, Marble Bluff Dam at 1176 m. Lake Tahoe to Boca TABLE 1. 9.88 4.96 19.66 22.17 31.38 11.30 37.46 15.58 206.81 393.86 537.97 403.32 505.24 PPL 2708.51 2.46 3 1.62 4.16 4.97 152.40 183.88 60.96 30.48 39.62 15.24 723.00 37.03 2137 17.71 16.10 6.44 11.27 173.88 Numana to Marble Bluff Dam Total Derby Dam to Wadsworth Wadsworth to Dead Ox Dead Ox to Numana Dam Vista to Derby Dam Boca to State Line State Line to Vista 20 MADRONO gages so that channel hydraulics could be calculat- ed in order to relate flow rates to topographic in- undation. Areas with extensive recent activity by beavers were avoided, as were areas of recent log- ging or fuelwood cuttings. The transects were ori- ented perpendicular to the river channel and varied in length according to the width of the riparian zone. A total length of 3380 m was sampled. Data were collected by structural layer according to the following definitions and procedures: Tree layer.—Single- or multiple-stemmed woody plants = 6 m in height and =10 cm diameter-at- breast height (dbh). Data were collected on dbh and density within 15 m of the transect (1.e., a 30 m belt transect). Stand canopy coverage was estimated by species at random locations along the transect using a spherical densiometer. Average stand can- opy height was estimated using a clinometer. Shrub layer(s).—AlIl woody plant species < 6 m in height or < 10 cm dbh. Canopy coverage was estimated visually along each transect using the line-intercept method (Mueller-Dombois and Ellen- berg 1974). Data were collected in three height classes: tall shrubs (>3 m), medium shrubs (>1 m and <3 m), and low shrubs (<1 m). Herbaceous layer.—Non-woody species includ- ing herbs, grasses, and graminoids. Canopy cover- age visually estimated by species using the line- intercept method. Ground surface.—Ground surface data on brush- piles, litter, and bare ground was tallied using the line-intercept method. Bare ground was further re- corded as clay, sand, gravel, cobbles, or boulders. Other information collected included site eleva- tion (taken from topographic maps), transect ori- entation (measured from aerial photos), current land use, and evidence of recent disturbance (e.g., graz- ing, beaver activity). The taxonomic reference for all plant scientific names is Hickman (1993). Data analysis. The areal coverage of individual vegetation and land use types was calculated by reach. Vegetation data from the transects were sum- marized for transect segments stratified by physi- ognomic type (forest, shrub, herbaceous). Each seg- ment was treated as a sample of its physiognomic type. The total of 178 samples were subjected to Two-way Indicator Species Analysis (TWIN- SPAN), a hierarchical classification procedure (Hill 1979; Gauch and Whittaker 1981); the TWINSPAN output was further refined based on the field ex- perience and professional judgement of the inves- tigator. RESULTS Physical characteristics. Physical characteristics of each of the eight reaches are provided in Table 1. The disparity in reach lengths is due to the va- riety of geologic and topographic settings through which the Truckee River passes along its course from the Sierra Nevada to Pyramid Lake. This is [Vol. 45 also reflected in the river gradient which generally decreases from Lake Tahoe to Pyramid Lake. There iS a corresponding general increase in the width of the riparian corridor. Exceptions to this trend occur in the State Line-Vista and Dead Ox-Numana Dam reaches, where the river passes through narrow bed- rock canyons. Cover type mapping. Four major categories of aquatic ecosystem, natural vegetation, and cultural types were mapped (Fig. 2). These include: 1) the active channel of the river including the low-flow wetted channel; 2) riparian forest and riparian shrub communities on the floodplain; 3) cultural types on the floodplain; and, 4) upland forest and upland shrub communities. The boundaries of the low-flow wetted channel were based on the area covered by water on the November 4th date of the aerial pho- tographs. Also occurring within the active channel were sparsely-vegetated cobble bars and patches of graminoids and herbs, here referred to as the veg- etated streambed. There is a dynamic relationship among these three elements of the active channel. The boundaries of the low-flow wetted channel ex- pand and contract in response to annual climatic variation, and water regulation or diversion. Chan- nel scour during high flows leads to an increase in the amount of cobble bars. Conversely, the absence of scouring flows results in an increase in the total area of vegetated streambed. Because this study was conducted during the sixth consecutive year of drought, the ratio of vegetated streambed within the active channel was greater than normal, when com- pared to either the cobble bars or the low-flow wet- ted channel. Overall, the active channel comprises about 25% of the riparian corridor, although in the steep, narrow canyons which characterize the Lake Tahoe-Boca and Dead Ox Wash-Numana Dam reaches, this value increases to 51% and 38%, re- spectively (Fig. 2). Riparian forest and riparian shrub communities occur on the floodplain of the river in most reaches. Deciduous riparian forests comprise between 2% and 20% of the riparian corridor upstream of Reno, with the lowest percentage occurring above Boca (Fig. 2). Downstream of Reno, the range is narrow- er (6% to 18%), although no riparian forest occurs in the Dead Ox-Numana Dam reach (Figs. 2, 3). A similar pattern is seen in the riparian shrub com- munities. Upstream of Reno, riparian shrub com- munities comprise 22% to 28% of the riparian cor- ridor (Fig. 2). Downstream of Reno, riparian shrub communities comprise 5% to 14% of the riparian corridor (Figs. 2, 3). Cultural types were defined to include agricul- tural fields and facilities, urban and industrial areas, sites dominated by the noxious weed Lepidium la- tifolium, and other disturbed areas. The proportions of the riparian corridor occupied by these habitats is low to moderate (7%—29%) in the upper three reaches, high (45—60%) in the middle three reaches, 1998] Lake Tahoe to Boca (37 km/166 ha) Low-Flow Channel 39% Cobble Bars 6% Disturbed 3% Urban/Industrial 4% Mixed Pine 9% Upland Shrub 2% Black Cottonwood 2% State Line to Vista (37 km/394 ha) Vegetated Streambed/Cobble Bars 4% Urban/Industrial 7% Disturbed 6% 2% Mixed Pine/Upland Shrub Fremont Cottonwood 8% Black Cottonwood 12% Fic. 2. CAICCO: TRUCKEE RIVER RIPARIAN VEGETATION Alder-Willow Shrub 28% Low-Flow Channel 23% Willow Shrub 23% 2 Boca to State Line (21 km/207 ha) Low-Flow Channel 23% Vegetated Streambed/ Cobble Bars 1% Agricultural 16% Disturbed 9% Black Cottonwood 16% Vista to Derby Dam (27 km/538 ha) 6% Fremont Cottonwood Peppergrass 23% Upland Shrub 10% ~/ Disturbed 11% Willow Shrub 6% Cobble Bars 2% Urban/Industrial 7% {>} Low-Flow Channel 14 i ° Vegetated Streambeds 3% Agriculture 18% Percentages of vegetation, aquatic, and land-use types for the upper four study reaches along the Truckee River. The length and area of riparian corridor is provided for each reach in parentheses. and moderate (29-33%) in the lower two reaches (Figs. 2, 3). Mixed pine communities occur on lower hillslopes adjacent to the floodplain only along the upper three reaches where they comprise 2% to 9% of the riparian corridor (Fig. 2). Upland shrub communi- ties occur on the floodplain in all reaches where they account for 2% to 10% of the riparian corridor, except along the lower two reaches where they comprise 18% and 28% of the corridor (Fig. 3). Marshes and ponds occur in several reaches, but they never account for more than 1% in any reach in which they occur (Figs. 2, 3). Species composition and abundance. The results of the TWINSPAN analysis supported distinctions between groups of samples of riparian forest, ri- parian shrub, vegetated streambed and cobble bar communities based on their occurrence upstream or downstream of Reno (Tables 2, 3). Upstream of Reno samples correspond roughly to the upper three study reaches, while samples downstream of Reno correspond to the lower five study reaches. Upland shrub communities showed no such dis- tinction, perhaps due to the infrequent occurrence of this type along transects upstream of Reno. Up- land mixed pine forests occur only along the upper three reaches. Populus trichocarpa ssp. balsamifera, with 80% canopy coverage, is the dominant tree species in deciduous riparian forests along the upper three reaches (Table 2). A tall shrub layer with 15% cov- er, dominated by Salix lutea, is present. The only other riparian shrubs present are S. exigua and sap- lings of P. trichocarpa ssp. balsamifera, each with only a few percent cover. Minor amounts of upland shrubs also occur in this type. The understory is dominated by Elymus trachycaulus and Poa pra- tensis with 19% and 14% coverage, respectively. Conium maculatum and Urtica dioica are the dom- inant herbaceous species. Both Populus trichocarpa ssp. balsamifera and P. fremontii ssp. fremontii (Fremont cottonwood) dominate individual deciduous riparian forest patches in the State Line-Vista reach, although no mixed stands of these species as canopy dominants were observed. Downstream of this reach, P. fre- montii, with 70% canopy coverage, is the sole dom- inant tree in the riparian forests (Table 3). There is a conspicuous dearth of riparian shrubs in these for- ests, where tall shrubs of P. fremontii provide only about 8% cover. Artemisia tridentata ssp. tridentata iS present in small amounts, and there is a sparse understory of Elymus trachycaulus and Lepidium latifolium. 20 MADRONO Derby Dam to Wadsworth (18 km/403 ha) Vegetated Streambeds/Cobble Bars Agriculture 42% 2% Peppergrass 8% Fremont Cottonwood 11% k\Low-Flow Channel 10% [Vol. 45 Wadsworth to Dead Ox Wash (6.2 km/505 ha) Vegetated Streambeds/Cobble Bars 13% V/ : Low ow pe arshes/Ponds 1% Vs Vee Nee Ol 89 8 1c 8 6C vv xWunyofun] unipidaT Z sqioH vy) gs OS 6c! tc OV 9'l [RIO], Joke] a £0 10 — — — pivoids s1yousiq vw — QP Ll LO — SHINDIAYIDA] SNWUWALA ww — _ L'0 — — yWUNAIYIUDAYIDAG WNapséOH 1) — _ r'0 — — DAYIDISOAYID XAADD 5 — — Ol — — yDIIDUIPUNAD SIUDIDYd e i — Ul — — DID]NIIUIN XAADD . — — 07 60 —— ySNIDUD] SNIJOH O rom) 60 — snsnffa snoune = — — rT i bers — 4SISUaJaIdsuou UosodAjodg Po — — el cp —- SNUDIIAUD SNdA1IS 6 — — Ol V'67C -— SUDINIIID SIADYIOIA Oo — — OZ — ro 4S1SUaIDAd DOd v SO a) snouyjoqg snoune ss —_ aa ee cO TO Isp] snop Xaivy — — a — ‘0 SUDINIYUA] XAADI cE L'0 — Z0 £0 SNAAIUID SNUKIT 6'l — L':0 ll £0 yWUNAOJIA] SNUOAG SPIOUIWIRIDn puke SdsseIH LCC SIC 6 Oc/e ILC TZ/0° LOI 9°6/L 8ET te LdO srs (WI) YSU] Josue UROUT/[RIOL (L1=U) (eT =u) (ci =u) (¢c=U) (Oc=4) ouleN qniys JSOIOJ qniys poquiesys seq pueldgq ueledny uevledry poqejosoA, 91qg0D CGHNNILNO,) “¢ ATIEVE 1998] MADRONO [Vol. 45 | Shrub (n=17) DiS2712 7 51.1 263 5 33.9 0.3 89.1 Upland 1.0 1.3 252 12 Riparian forest (n=13) 271.3/20.9 65.3 94.0 Riparian shrub (n=15) 107.0/7.1 40.9 13.8 27.0 3.0 84.7 Vegetated steambed (n=25) 238.7/9.6 204 4.0 32.4 302 87.0 bars (n=20) 545.0/27.3 0.9 28.4 3.8 313 97.1 Cobble 30.8 Ground Surface Litter Boulders Cobbles Gravel Sand Brush piles Ground Surface Total Total/mean transect length (m) Silt/clay TABLE 3. CONTINUED Name was substantially reduced from its typical abun-- dance during wetter periods. | DISCUSSION The conversion of the riparian corridor of the | Truckee River to urban, industrial, and agricultural uses, water resource development, and other human activities has led to a significant decline in the bi- — ological resources of the riparian corridor. Although | some areas have recovered from earlier impacts, the type, degree, and extent of recovery from these ac- tions varies among the study reaches. For example, the Lake Tahoe-Boca reach was intensively logged in the last half of the 19th century. Today, much of the riparian corridor in this area has natural vege- tation with only 7% in either urban and industrial uses or otherwise disturbed. In the downstream reaches between Boca (CA) and Vista (NV), human activities continue to di- rectly impact between 25% and 29% of the riparian corridor. For the lowermost of these two reaches, which contains the city of Reno, this is a significant underestimate of the historic losses within the ri- parian corridor since the river in this area is con- tained between flood control levees. The riparian corridor in the two reaches between Vista and Wadsworth is dominated by agricultural and industrial uses, or is otherwise disturbed. Only about 40% of the riparian corridor in these reaches remains in natural condition. Downstream of Wads- worth, the proportion of natural vegetation increas- es to between 55% and 70%. Two aspects of the data are of particular interest in regard to ecological restoration within the ripar- ian corridor. These are: 1) the relative impacts of introduced plant species; and, 2) the potential amount of habitat available for restoration. With re- spect to introduced plant species, the habitats vary when compared to each other as well as among reaches. Along the upper Truckee River, introduced plants include only grasses and herbs (Table 2). They are most abundant in the understory of the riparian forests where they comprise 47% of the total vegetative cover. Three species, Poa pratensis, Hordeum brachyantherum, and Conium maculatum account for most of this cover. Along the lower Truckee River, the introduced shrub Eleagnus angustifolius is a minor component of the riparian shrub community, but here are nu- merous introduced grasses and herbs. Individual species of introduced grass comprise only a few percent of any community. The predominant intro- duced herb is Lepidium latifolium, which dominates the herbaceous layer of all natural vegetation but the upland shrub community. Overall, it accounts for 33% to 72% of the total understory cover, and is most abundant in the vegetated streambed and riparian shrub communities. If one assumes that the total area currently oc- cupied by agricultural and otherwise disturbed ar- 1998] eas (including areas dominated solely by Lepidium latifolium) represents the maximum amount of area potentially available for habitat restoration, the three lower river reaches between Vista and Dead Ox Wash offer the most potential for restoration, with 280 ha, 238 ha, and 227 ha of these habitats, respectively. This is not, however, to say that po- tential restoration opportunities are not available, or should not be pursued, in other reaches. Opportu- ‘nities are most limited in reaches where the flood- plain is restricted to a relatively narrow canyon bot- tom. Examples include upstream of the town of Truckee where the already narrow floodplain is constrained by a highway and an increasing number of residences, and in the narrow gorge between Dead Ox Wash and Numana Dam. Urban and in- dustrial areas in the vicinity of Reno also offer lim- ited restoration opportunities. The presence of suitable habitat is not the only factor constraining restoration opportunities. Aside from the obvious need for the cooperation of pri- vate landowner’s, ecological processes must also be considered. As noted in the introduction, down- stream of Numana Dam the river has incised deeply into floodplain terraces which formerly supported extensive stands of cottonwood forest. The scat- tered skeletons of dead trees and the few decadent living cottonwood trees which remain on these ter- races suggest that groundwater levels in this area have dropped beyond that necessary to maintain trees. Serious constraints on restoration potential also exist upstream of the area where significant channel incision has occurred. Flow regulation and water diversions have altered the magnitude, timing, fre- quency, and duration of flows in the river. These changes have had the greatest effect downstream of Derby Dam where the interbasin diversion of water to the Carson River drainage takes place. River ter- races which currently support cottonwood forests are no longer flooded with the historic frequency, so conditions conducive to cottonwood (and wil- low) seed germination are less frequent. This par- tially accounts for the paucity of replacement cot- tonwoods in the shrub layers of the cottonwood for- ests, as well as the absence of any associated ri- parian shrubs. This absence, and the fact that Artemisia tridentata ssp. tridentata is the only shrub present in these forests, suggests that ground- water levels are below the effective rooting depth of riparian shrubs for an insufficient length of time to allow the establishment of any seedlings that might germinate during wet springs. The existing riparian shrub communities along the lower river occur primarily on sandy deposits along the edge of the active channel. These com- munities, dominated by cottonwood saplings, were established after a period of high runoff in 1983 (Lisa Heki, personal communication, August 1998). In 1995, an estimated 50,000 new cottonwoods re- generated along the active channel in this area as a CAICCO: TRUCKEE RIVER RIPARIAN VEGETATION 29 result of experimental flows patterned on a natural flow regime (Christensen 1996). Flow management in 1996 and 1997 resulted in additional channel re- shaping and creation, and the establishment of an estimated 15 to 20 million cottonwood seedlings downstream of Wadsworth (L. Heki, personal com- munication, August 1998). I had earlier expressed concerns that high flows might remove saplings already established in the active channel (U.S. Fish and Wildlife Service 1993). These concerns now appear unwarranted. Although the 1997 flood reached 23,000 cubic-feet- second, and removed some saplings that had estab- lished in 1983, many of these uprooted trees were deposited downstream where they resprouted after being buried in sediment. In addition, the flood re- arranged channels and created side channels that provided additional habitat for cottonwood recruit- ment. This newly created habitat more than com- pensates for cottonwood regeneration lost to the flood flows (L. Heki, personal communication, Au- gust 1998). Such remarkable short-term successes make the long-range prospects for the restoration and en- hancement of the lower Truckee River appear high- ly favorable. With continued flow management di- rected toward maintenance of the regenerated cot- tonwoods in the active channel, riparian forests can be expected to develop and eventually provide suit- able habitat for riparian forest associated plant and animal species. In addition, the erosive action of floods will be decreased by the network of roots associated with these forests and their aboveground vegetation (Gregory et al. 1991). Increased sedi- ment deposition resulting from the enhanced reten- tion of material in transport may lead to an increase in the general elevation of the streambed, thereby restoring the hydrological connection to the upper- most river terraces. In time, the river valley may once again resemble the one described over 120 years ago as consisting of ‘“‘meadowlands .. . stud- ded with fine large cottonwood trees ... which were here and there grouped into delightful groves, sometimes unencumbered, but generally with a shrubby undergrowth, amongst which the “‘buffalo- berry” (Sheperdia argentea) was conspicuous” (Ridgeway 1877). CONCLUSION This paper provides basic information on the species composition, structure, and areal extent of riparian vegetation along the Truckee River. These data, along with an understanding of hydrological processes such as flow magnitude, frequency, tim- ing, and duration can guide land managers and bi- ologists in their efforts to restore and enhance these habitats. Such actions will become increasingly im- portant as the urban and rural populations of the west continue to grow. 30 MADRONO ACKNOWLEDGMENTS This work was supported by the Nevada State Office of the U.S. Fish and Wildlife Service. I thank Chris Willis of the Sacramento Field Office, who helped prepare the cover type maps, Marla Macoubrie of the Sacramento Field Office, and Betsy Whitehill and Robin Hamlin of the Nevada State Office, who helped collect field data. I also thank Michael Williams and an anonymous reviewer, each of whom provided an excellent critique of drafts of this paper. LITERATURE CITED ABELL, D. L. 1989. Proceedings of the California riparian systems conference: protection, management, and res- toration for the 1990’s. U.S. Department of Agricul- ture, Pacific Southwest Forest and Range Experiment Station, Davis, CA. General Technical Report PSW- 110. Born, S. M. 1972. Late Quaternary history, deltaic sedi- mentation, and mudlump formation at Pyramid Lake, Nevada. Center for Water Resources Research, Desert Research Institute, University of Nevada, Reno. BUSHER, P. E., R. J. WARNER, AND S. H. JENKINS. 1983. Population density, colony composition, and local movements in two Sierra Nevadan beaver popula- tions. Journal of Mammalogy 64:314-318. CALIFORNIA DEPARTMENT OF WATER RESOURCES. 1991. Truckee River Atlas. The Resources Agency, State of California, Sacramento. CAROTHERS, S. W. 1977. Importance, preservation, and management of riparian habitats: an overview. Pp. 2- 4 in R. R. Johnson and D. A. Jones (eds.), Impor- tance, preservation, and management of riparian hab- itat: a symposium. U.S. Department of Agriculture, Rocky Mountain Forest and Range Experiment Sta- tion, Fort Collins, CO. General Technical Report RM- 43. CHRISTENSEN, J. 1996. Helping a river help itself. Nature Conservancy Magazine, September/October: 8-9. CLARY, W. P., E. D. MCARTHUR, D. BEDUNAH, AND C. L. WAMBOLT, (eds.). 1992. Proceedings of the sympo- sium on ecology and management of riparian shrub communities. U.S. Department of Agriculture, Inter- mountain Research Station, Ogden, UT. General Technical Report INT-289. GaucH, H. G. AND R. H. WHITTAKER. 1981. Hierarchical classification of community data. Journal of Ecology 69:135-152. GLancy, D. S., A. S. VAN DENBURGH, AND S. M. Born. 1972. Runoff, erosion and solutes in the lower Truck- ee River, Nevada during 1969. Water Resources Bul- letin 8:1157—1172. GREGORY, S. V., EF J. SWANSON, W. A. MCKEE, AND K. W. CumMINS. 1991. An ecosystem perspective of riparian zones. Bioscience 41:540—551. HALL, E. R. 1960. Willow and aspen in the ecology of beaver on Sagehen Creek, California. Ecology 41: 484-494. HICKMAN, J. C., (ed.). 1993. The Jepson manual: higher [Vol. 45 plants of California. University of California Press, Berkeley. HILL, M. O. 1979. TWINSPAN—a FORTRAN program for arranging multivariate data in an ordered two-way table by classification of individuals and attributes. Cornell University, Ithaca, New York. JOHNSON, R. R. AND D. A. JONES. 1977. Importance, pres- ervation, and management of riparian habitat: a sym- posium. U.S. Department of Agriculture, Rocky Mountain Forest and Range Experiment Station, Fort Collins, CO. General Technical Report RM-43. JOHNSON, R. R., C. D. ZIEBELL, D. R. PATTON, P. EK FFOL- LIOTT, AND R. H. HAMRE, (eds.). 1985. Riparian eco- systems and their management: reconciling conflict- ing uses. U.S. Department of Agriculture, Rocky Mountain Forest and Range Experiment Station, Fort Collins, CO. KLEBENOW, D. A. AND R. J. OAKLEAF. 1984. Historical avifaunal changes in the riparian zone of the Truckee River, Nevada. Pp. 203-209 in R. E. Warner and K. M. Hendrix (eds.), California riparian systems: ecol- ogy, conservation, and productive management. Uni- versity of California Press, Berkeley, CA. Knopr, E L., R. R. JOHNSON, T. RICH, E B. SAMSON, AND R. C. SZARO. 1988. Conservation of riparian ecosys- tems in the United States. Wilson Bulletin 100:272- 284. MARTENS, E. AND J. A. YOUNG. 1992. Seed germination data for yellow willow at a Nevada riparian site. Pp. 142-144 in W. P. Clary, E. D. McArthur, D. Bedunah, and C. L. Wambolt (eds.), Proceedings of the sym- posium on ecology and management of riparian shrub communities. U.S. Department of Agriculture, Inter- mountain Research Station, Ogden, UT. General Technical Report INT-289. MUELLER-DoMBoIS, D. AND H. ELLENBERG. 1974. Aims and methods of vegetation ecology. John Wiley & Sons, New York. RIDGEWAY, R. 1877. Ornithology. Pp. 303—669 in C. King. Ornithology and paleontology. Report of the geolog- ical exploration of the fortieth parallel. Professional Papers No. 18 of the U.S. Army Engineer’s Depart- ment. SAVAGE, W. 1973. Annotated checklist of vascular plants of Sagehen Creek drainage basin, Nevada County, California. Madrono 22:115-139. SMITH, G. L. 1984. A flora of the Lake Tahoe Basin and neighboring areas and supplement. University of San Francisco, CA. U.S. FISH AND WILDLIFE SERVICE. 1993. Truckee River ri- parian vegetation and fluvial geomorphology study. Technical report prepared for the Reno Field Office, Reno, Nevada. WARNER, R. E. AND K. M. HENpDRIX, (eds.). 1984. Cali- fornia riparian systems: ecology, conservation, and productive management. University of California Press, Berkeley. WATER ENGINEERING & TECHNOLOGY, INC. 1991. Recon- naissance geomorphic investigation of Truckee River from Vista to Pyramid Lake, Nevada. Report to Sac- ramento District, U.S. Army Corps of Engineers. Contract No. DACWO05-91-P-1543. “MaproNno, Vol. 45, No. 1, pp. 31-39, 1998 DEMOGRAPHY OF NATURAL AND REINTRODUCED POPULATIONS OF ACANTHOMINTHA DUTTONIT, AN ENDANGERED SERPENTINITE ANNUAL IN NORTHERN CALIFORNIA BRUCE M. PAVLIK AND ERIN K. ESPELAND Department of Biology, Mills College, Oakland, CA 94613 ABSTRACT The purpose of this study was to 1) demographically monitor the only remaining natural population of the rare serpentinite annual plant Acanthomintha duttonii (Lamiaceae); 2) attempt to reintroduce a new, experimental population within historic range; and 3) evaluate the new population by com- paring its demographic characteristics with those of the natural population. The natural population of A. duttonii at Edgewood Park significantly and progressively increased in abundance and density between 1990 and 1994, then began a decline that lasted through 1997. In general, high density and high yield (reproductive plants produced from previous year’s nutlet production) were associated with average or below-average years of precipitation while low densities and yields were associated with above-average rainfall years. During the entire study period, survivorship to reproduction re- mained fairly high and consistent, indicating that population trends were due to variations in nutlet production and the influence of cryptic factors that operate in the seed bank. The experimental population at Pulgas Ridge differed in several critical respects from the natural population, including low germination, low and variable survivorship, low nutlet production and perhaps high nutlet mor- tality. These features reduced the potential for self-sustained growth in the experimental population, which is likely to be extirpated within the next few years. This failure to produce a self-sustaining population of A. duttonii emphasizes the urgent need for in situ preservation of self-sustaining natural populations of serpentinite species. Serpentinite endemics comprise the largest single edaphic category of rare plants in the native flora of California (Pavlik and Skinner 1994). They tend to occur in low elevation grassland and chaparral habitats on rocky outcrops, gravel colluvium and alluvial clays throughout the California Floristic Province, especially in the coast ranges and Sierra Nevada foothills (Kruckeberg 1984; Fiedler 1992). These same localities are also favored by land de- velopers who build houses at the edges of rapidly expanding cities. The increasing development pres- sure on serpentinite habitats requires more effective conservation strategies if we are going to maintain these species as wild populations in California. Among those strategies are the preservation of re- maining natural populations on managed reserves (Reznicek 1987; Lessica and Allendorf 1992; Pav- lik 1996) and the creation of new populations with- in historic range (i.e., reintroduction) to decrease the overall probability of species extinction (Pavlik 1994; Guerrant and Pavlik 1997). Reintroduction has also been used in a mitigation context, attempt- ing to ameliorate the destruction of natural popu- lations by transporting propagules onto protected sites. Reintroduction in any context is fraught with difficulties and uncertainties (see Falk et al. 1996), but is especialy so when conducted under the fiscal, temporal, and political constraints imposed by mit- igation (Howald 1996). The San Mateo Thornmint, Acanthomintha dut- tonii, is among California’s most endangered plant species (York 1987). This state- and federally-listed plant occurs in a single, fragmented population at Edgewood County Park on the San Francisco pen- insula (Sommers 1984; Skinner and Pavlik 1994). The population has a particularly high risk of ex- tinction because of its small areal extent, proximity to high density suburbs and altered water runoff patterns from upslope development. Once a popular spot for off-road motor vehicle recreation, the site now experiences only sporadic disturbance by hik- ers and mountain bikers. Other than being an annual plant restricted to ser- pentinite grasslands, relatively little is known about the demography, ecology, and genetics of A. dut- tonil. No demographic monitoring (sensu Pavlik 1987; Pavlik and Barbour 1988; Pavlik 1994) has been conducted on this or any other species of Acanthomintha. Surveys of A. duttonii at its only remaining natural population have recorded large fluctuations in population size. In 1984, fewer than 5000 individuals were found, while estimates in 1981 and 1986 were closer to 3000 (Sommers 1986; CNDDB 1989). An apparent high of 6000 during 1985 was reported by McCarten (1986). Al- though such fluctuations are to be expected in pop- ulations of annual plants, the responsible factors have yet to be identified and related to management of this endangered species. The purpose of this study was to 1) institute de- mographic monitoring to determine general trends and limiting factors in the only remaining natural population of A. duttonii at Edgewood County Park, 2) attempt to reintroduce a new, experimental pop- 32 MADRONO ulation within historic range and appropriate habi- tat, and 3) evaluate the new population by compar- ing its demographic characteristics with those of the natural population. MATERIALS AND METHODS Study Species and Study Sites The most recent taxonomic treatment of the ge- nus Acanthomintha (Lamiaceae) has been done by James Jokerst for the Jepson Manual (Hickman 1993). He recognized four species (A. ilicifolia, A. obovata, A. duttonii, and A. lanceolata) on the basis of style morphology, corolla morphology, stamen fertility, leaf morphology, geographic distribution, and substrate preference (Jokerst 1991). Acantho- mintha duttonii is an annual, frequently unbranched herb with dense glomerules subtended by spineless bracts (in contrast to the spine-tipped cauline leaves). The bilabiate corolla is white or tinged with lavender and contains 4 stamens with reddish an- thers. Each flower is capable of producing a max- imum of 4 nutlets. Nothing is known about the genetic structure of the population. Acanthomintha duttonii is primarily a self-pollinating species (Steeck 1995), although visits to the corolla by small insects are frequent during the spring. As inferred from electrophoretic studies of other inbreeding plants, intrapopulation allelic variation would probably be relatively low and interpopulation variation would have been high prior to extirpation (Hamrick et al. 1991). This pat- tern is typical of habitat specialists whose popula- tions are isolated from each other by significant barriers to gene flow. Ecologically, A. duttonii is restricted to mesic serpentine grasslands that receive an average 500 mm of precipitation per year. At Edgewood Park, site of the only remaining natural population, mean annual temperature is 15°C with a mean annual temperature range of 11.1°C. Frosts are rare, with freezing temperatures occurring in less than 0.5% of the hours of the year (estimated from a Bailey nomogram). It is associated with more widespread grassland dominants, such as Nasella pulchra, Lol- ium multiflorum, Delphinium hesperium and Hem- izonia congesta var. luzulifolia (nomenclature fol- lows Hickman 1993). McCarten (1986) conducted detailed surveys of ac- tual and potential habitat of the species in San Mateo County. He did extensive soil sampling and found that the deep serpentinite clay of the Edgewood site was moist, chemically unusual, and rather uncommon in the county. Using these data and a wealth of field experience, McCarten and others (especially Susan Sommers, Toni Corelli and Ken Himes) mapped sev- eral possible sites for creating new populations. Pul- gas Ridge, largely composed of serpentinite clay, lies along the eastern edge of the San Andreas Rift Zone and was identified as a general location suitable for reintroduction of A. duttonii. [Vol. 45 | The process of selecting microsites for new A._ duttonii populations took many factors into consid- — eration, including the ecological (macroclimate, — soil, exposure, community associates, habitat size and degree of disturbance), and the logistic (land © use history, road access, property ownership). A microsite on Pulgas Ridge was selected because of its apparently high quality habitat (mesic grassland on serpentinite clay soil), its public status as wa- tershed lands operated by the San Francisco Water Department, and because it is very close to, if not within, the historic range of the target species. In many ways Pulgas Ridge resembles the Edgewood Park site, although its serpentinite areas are much larger and less fragmented by intrusions of non- serpentinite vegetation (e.g., oaks and annual grass- land). The soil at Pulgas Ridge also compared fa- vorably with soils at Edgewood Park (Pavlik et al. 1992) because it was found to be rich in clay (high saturation percentage) and chemically typical of lo- cal serpentinite (low nitrogen, low calcium/mag- nesium ratio, high nickel). In addition to the ecological and logistic criteria discussed above, the microsite was selected to be; 1) large enough to allow a total of 24, 26 x 28 cm quadrats, separated by row and column spaces (ac- cess paths) 2) relatively homogeneous with respect to microhabitat factors (soil depth, slope, associated species, etc.), 3) accessible but reasonably con- cealed to reduce the potential for vandalism or oth- er human disturbance, and 4) surrounded by suit- able habitat so as not to constrain population growth in the future. We chose an east-west trend- ing channel of a small intermittent stream, with gently-sloping (25%) banks of serpentinite clay. Plant cover was relatively sparse and open and would not excessively shade or otherwise crowd the new A. duttonii plants. Monitoring the Natural Population at Edgewood Park Seedling density and survivorship to reproduc- tion in situ. Estimates of adult plant densities in the natural population were made in May—June in the years 1990-1997. A total of thirteen 0.125 m? cir- cular quadrats were used to map the population and to record the densities of other species (e.g., Lolium multiflorum and Avena fatua) on and off the ser- pentinite clay. Given the lack of security and high visitation at Edgewood Park, we decided to leave very few, cryptic markers for mapping locations within the population. Consequently, only five per- manent quadrats were randomly positioned during May 1990 in the belief that more markers would have increased the probability of vandalism. An X- Y grid was superimposed on the population and used to determine the positions from coordinates generated by a random numbers table. In addition, eight transient quadrats were distributed around the permanent quadrats at that time, increasing the 1998] number of density estimates for our maps. In fol- lowing years, the locations of transient quadrats were approximated. After flooding had promoted downslope colonization during the winter of 1991— 1992, two new permanent quadrats were estab- lished. The quadrats were used to determine the mean density of reproductive plants and to estimate total population size (when multiplied by the area of the population, about 42 m°). The same permanent plots were also used to es- timate survivorship to reproduction during the 1991 to 1997 growing seasons. Fifty seedlings of A. dut- tonii were marked within each permanent plot when germinules were at the 4—6 leaf stage (Jan- uary—February). The plots were revisited in June and the number of marked, reproductive plants were tallied. Plant size and nutlet production. During peak flowering period (May-June) of 1990 to 1996, whole plants of A. duttonii were non-randomly se- lected to represent the complete range of plant sizes within a variety of microhabitats. These plants were clipped at ground height, sealed in individual bags and taken back to the laboratory. Stem length was measured from the clipped point to the base of the lowest glomerule. Forty-three plants were collected in 1990 and 25 plants were collected in each of the following years. Correlations between stem length and reproductive output were established using methods developed during studies of other endan- gered plants (Pavlik and Barbour 1988; Pavlik et al. 1993). Linear and non-linear models were ap- plied to the data in each year, but the former pro- duced higher correlation coefficients in most years and was, therefore, consistently applied across all data sets. In June of each year, all plants that survived to reproduce within the permanent survivorship quad- rats were measured for stem length and number of glomerules. These parameters were used to estimate mean plant size and nutlet output for the natural population and to generate frequency distributions of plant size for comparison with similar data col- lected for the experimental population. Reintroducing the Experimental Population Characteristics of the founder nutlets. The pro- pagules (=nutlets) of A. duttonii used were col- lected from Edgewood Park in May—June of 1990 to 1994. Nutlets were taken from at least 40 indi- viduals that represented the complete size range and microenvironmental amplitude of the natural population. The collection would be likely to con- tain, therefore, a representative sample of the ex- isting genetic variation (Falk and Holsinger 1991). Nutlets were stored at 4°C in paper pouches within sealed plastic bags until they were sown in the field. Laboratory germination trials were conducted in 1991 to 1993. Nutlets from each year’s crop were tested the following January using three replicates PAVLIK AND ESPELAND: DEMOGRAPHY OF A SERPENTINITE ANNUAL 53 of 25 nutlets each. A replicate consisted of a plastic petri dish (5.5 cm diameter) containing a filter pa- per disk that was kept moist with distilled water. Nutlets were spread across the paper disks and kept in a dark room in which the temperature averaged 25°C. Replicates were checked every day for 12 days, noting germination (protrusion of the radicle through the pericarp) and removing germinules with a paintbrush. Installation. The population installed at Pulgas Ridge during the early winter of 1991 consisted of two sets of 12 plots each. A removable wooden frame containing a7 X 7 grid of 49 holes was used as a template to precisely sow A. duttonii nutlets. The holes allowed exact placement and subsequent monitoring of germinules and juvenile plants. This ‘“‘precision-sowing”’ technique has been successful- ly used by Pavlik and Manning (1993) to establish and monitor new subpopulations of the endangered Oenothera deltoides ssp. howellii and Erysimum capitatum var. angustatum and by Pavlik et al. (1993) to establish and monitor new populations of Amsinckia grandiflora. In 1992 a total of 1176 nutlets of Acanthomintha duttonii, half from the 1990 crop at Edgewood and half from the 1991 crop, were sown into the 24 plots at Pulgas Ridge. Six additional precision- sown plots were established in the late fall of 1992. We also used a streak method for sowing five plots with 250 more nutlets on a small clay lens 30 m away from the stream bank. A linear furrow was cut in the soil with a blunt nail and sown with 50 seeds before covering it over with native soil. Streak plots did not allow for strict demographic measurements, but they were easier to establish and quicker to monitor. Additional enhancements to the reintroduced population were added in the falls of 1993, 1994 and 1995 in the form of ten more streak plots per year within 10 m of the demographic plots. All plots were sown in September—November of each year and no supplements of water or nutri- ents were applied. A summary of the nutlet inputs to all plots is presented in Table 3. Monitoring and evaluation. The fate of each pre- cision-sown nutlet was followed during the January to June growing season by repositioning the wood- en frames on each plot and searching for seedlings. The condition of each seedling was recorded on plot-specific data sheets to allow calculation of crit- ical demographic parameters (Pavlik 1994). Those parameters included field germination, stress fac- tors (desiccation, etiolation, grazing by microher- bivores), mortality, phenology, survivorship to re- production, and plant size (number of glomerules and stem length). Streak plots were checked three times during the growing season for seedling emer- gence and each plant was measured at peak flow- ering period (May—June) for number of glomerules and stem length. During the early summer of 1994, ten whole 34 MADRONO Rainfall (mm) 1989 1990 1991 1992 Year Fic. 1. 1993 (9,) aanqesodwa | 1994 1995 1996 Patterns of monthly precipitation and air temperature near the natural population of Acanthomintha duttonii at Edgewood County Park, 1990-1997. Rainfall totals for a growing season (November to November) are shown. plants of A. duttonii were collected from the Pulgas Ridge experimental population. Stem length was measured from the clipped point (at soil surface) to the base of the lowest glomerule. Correlations be- tween stem length and reproductive output were then calculated. RESULTS AND DISCUSSION Environmental Patterns 1990—1997 Rainfall varied significantly from year to year, but the seasonal patterns and magnitudes of mean monthly air temperature seemed to be similar for all years of the study (Fig. 1). Pronounced drought occurred during 1990 (November 1989-—June 1990 growing season) when less than 300 mm was re- ceived in the vicinity of Edgewood Park. The years 1991, 1992 and 1994 were also dry relative to nor- mal (about 500 mm), while 1993 was wetter than average. The years 1995, 1996 and 1997 were ex- tremely wet, each exceeding 800 mm during the growing season. The overall pattern during the study, therefore, was one of low rainfall during the first five years (1989-1994) and extremely high rainfall during the last three years (1994-1997). Flooding of the natural population may have oc- curred during all wet years, but during the winter of 1991-1992 a single, intense storm had facilitated seed dispersal. This storm evidently caused soil erosion and downslope movement of nutlets into an adjacent patch of serpentinite clay that previously did not support A. duttonii. The newly-colonized area was subsequently included in the monitoring program. Flooding of some portions of the experi- mental population was prolonged during the ex- tremely wet 1994-1995 growing season. Another form of disturbance to the population was observed each late spring and summer. As the serpentinite clay soil dried and shrank, large surfi- cial cracks began to open, sometimes as much as 4 cm wide and 30 cm deep. Mature nutlets would drop from adjacent plants into the gaping crevices. These cracks fill with water and clay during the next winter rain, perhaps burying a large number of nutlets at depths too great for seedling emer- gence. It is likely that these cracks significantly re- duce potential population growth, but it is also like- ly that some nutlets remain viable for long periods of time within the seed bank. Observations made on a small, transient colony of A. duttonii in an adjacent portion of Edgewood Park have confirmed that nutlets can produce plants after 8 years of qui- escence in situ. That quiescence, during which no adult plants were observed, was associated with high rainfall, E/ Nino climatic events during the early 1980’s (Sommers 1986, Pavlik and Espeland 1994). Demography of the Natural Population at Edgewood Park Density and survivorship. During eight years of observation at Edgewood Park, mean density of the reproductive A. duttonii population could vary by a factor of 5 (Table 1), with a low of 230 plants/ m? (1991) and a high 1106 plants/m? (1994). A 64% decline in density occurred between 1994 and 1995, with stepwise decreases through 1997. Spa- 1998] PAVLIK AND ESPELAND: DEMOGRAPHY OF A SERPENTINITE ANNUAL a5 | TABLE 1. CHARACTERISTICS OF NATURAL AND EXPERIMENTAL POPULATIONS OF ACANTHOMINTHA DUTTONII, 1990-1997, | Population size for natural population was estimated using the mean density and population area. Survivorship and _ density shown as means + SD. Yield estimates were obtained dividing present population size by the previous year’s nutlet production (Table 2). Reproductive Reproductive Reproductive population Emergence survivorship density size (%) (%) (#plants/m2) (#plants/site) Yield Natural Population (Edgewood) 1997 — 36,0:2 01:5 63-2 58 5289 0.128 1996 — 303° 122 89 + 68 6885 O226 1995 — 561052253227 390 + 210 20,280 0.048 1994 — 52:0° 2218,0 1106 + 589 53,136 0.110 1993 — 6202 212 794 + 756 36,279 0.305 1992 — 59.4 + 29.4 302 + 294 18.772 0.025 199] — 54.8 + 14.9 2302718 9660 0.073 1990 — — 689 + 704 12,864 a Experimental Population (Pulgas) 1997 — 28.0 Zo 2 —- 1996 — 39.0 35 77 —- 1995 ley, 13.6 54 145 —- 1994 10.9 S312. 66 158 — 1993 34.0 63.0 81 181 _- 1992 27.0 38.0 68 120 — tial variations in density were high (Fig. 2), with forbs (Plantago erecta and Lotus subpinnata) from unique patterns found in each of the permanent serpentinite sites at nearby Jaspar Ridge (Arm- plots. Patterns of increased abundance during — strong and Huenneke 1992). drought years (Fig. 1), are supported by observa- In contrast, survivorship to reproduction was tions made on some species of common annual fairly constant over time (~50%, Table 1) for the plant densities /m2 ie # 1000-1500 1500-2000 2000-2400 1994 1995 1996 1997 Fic. 2. Spatial pattern of reproductive plant density of the only remaining natural population of Acanthomintha duttonii at Edgewood County Park, 1990-1997. The smaller lobe of the population’s outline (to the right) is the downslope addition colonized after flooding in the winter of 1991-1992. n/a = data not available. 36 TABLE 2. 1990-1997. Nutlet output estimated using #nudets/plant MADRONO [Vol. 45 REPRODUCTIVE CHARACTERISTICS OF NATURAL AND EXPERIMENTAL POPULATIONS OF ACANTHOMINTHA DUTTONIH, m(stem length) + b. Nutlet production is the product of | nutlet output and population size (Table 1). * = estimated using 1996 nutlet output correlation for the site > = estimated using 1994 nutlet output correlation for the site. Mean plant : Correlation between nutlet output size sear se elie ES HE UCS (y) and stem length (x) aoe ee Stem length output production m b r P n (cm) n (#/plant) (#/site) Natural Population (Edgewood) 1997 — — — — — 4.1 + 3.8 206 56? 3.0 X 10° 1996 5.23 42.8 0.30 ns 25 es aah 211 60 4.1 xX 10° 1995 351 8.1 0.39 <0.05 25 ro Nevo 198 26 5.4 X 10° 1994 Bao 213 0.51 <0.05 2D 4.0 + 3.2 155 25 N52 108 1993 2.86 36.1 0.45 <0.05 25 45+ 2.4 220 49 13-70: 1992 1.88 Sel 0.85 <0.01 25 Geral 150 16 3.0 x 10> 1991 9.00 =a 0.92 <0.01 6 9222 3.6 25 78 TS X 103 1990 2.83 211 0.71 <0.01 43 Af 23 25 188 34 4.4 x 105 Experimental Population (Pulgas) 1997 — — — — — 45 + 3.4 55 oUF 1560 1996 2.99. 16.5 0.52 =0).05 ] 94+ 6.1 TD 45 3375 1995 — — — — —- 5 Oia 2.8 145 23 3045 1994 259 22.4 0.74 <0.05 25 5.2 + 4.8 158 36 5688 1993 — — — — — 44+ 2.5 181 34° 8869 1992 — = — — — 3 Eels 6 hlhs) 42° 3150 population as a whole, tending to be slightly higher downslope in the newly colonized area and slightly lower upslope. Variations in plant density, there- fore, are probably influenced by cryptic factors that operate in the seed bank (nutlet density, nutlet mor- tality, germination, emergence) rather than more obvious factors that control seedling growth and mortality. Estimated total population size progressively in- creased from a low of 9660 reproductive individ- uals in 1991 to a peak of 53,136 in 1994. The ratio of reproductive population size to estimated nutlet production in the previous year (yield) was com- monly between 0.07 and 0.13, with a single peak of 0.30 in the spring of 1993. The peak in popu- lation size was followed by a 60% decline in 1995, with no evidence of catastrophic flooding or an- thropogenic disturbance. In years of decline there were intrusions of Hemizonia congesta var. luzuli- folia, Perideridia sp., and Lolium multiflorum into the body of the population. Therefore, the recent decline in population size of A. duttonii appeared related to decreasing density across the entire hab- itat, with losses of potential habitat to common, ser- pentinite-tolerant species during especially wet years. Plant size and nutlet production. The output of nutlets by individual A. duttonii plants in the natural population was linearly related to the sum of the stem lengths per plant in most years (Table 2). The slopes and intercepts of the relationship varied from year to year, but again there was no obvious cor- relation with environmental patterns or plant den- sity. Mean stem length was greatest in 1991 and 1992 (years of below normal rainfall), but again there was no correlation between plant size and to- tal yearly precipitation. Regardless of year, the large majority of plants in the population fell into the one glomerule or short-stem size categories and few were large and well branched (data not shown). The total nutlet production of the population could be estimated using the nutlet output correla- tion along with estimates of mean plant size, pop- ulation density, and population area for each year. Average plants usually output between 16 and 80 nutlets (Table 2), but the largest plants could make between 150 and 200 nutlets each. In June of 1992 an extremely large individual was found to produce 662 nutlets (from 232 flowers in 18 glomerules, with 66 cm of stem length in eight branches). Given the high density of the population in most years of the study, nutlet rain ranged between 10,300 nut- lets/m? (1990) and 36,800 nutlets/m’ (1993) (data not shown). Consequently, the total nutlet produc- tion of the natural population was in the range of 10°—10°. Reintroduction at Pulgas Ridge Laboratory germination of the founding nutlets. Nutlets of A. duttonii had moderate to high rates of germination in the laboratory. Germination aver- aged 87% for 1990 nutlets, 63% for 1991 nutlets and 71% for 1992 nutlets, even though all crops were approximately seven months old at the time the tests were conducted. There was a strong after- ripening requirement that prevented any germina- tion during the six months following collection (June through December). Late winter germination 1998] appears to be characteristic of this species, owing to a rigid endogenous control mechanism that strat- ification, pericarp scarification, fire, wet-dry cy- cling, and red light cannot override (Pavlik and Es- peland 1991). Perhaps such a mechanism prevents germination before a thorough saturation of the clay substrate takes place, thus avoiding the possi- bility of seedling desiccation during warm days in fall and early winter (also discussed in Armstrong and Huenneke 1992). Percolation is slower within clay substrates and so a higher proportion of the falling rain is likely to run off. Furthermore, clay particles require much more water than sands and gravels to bring soil water potentials into the tol- erable range of —0.1 to —1.5 MPa for most seed- lings. Emergence in the field. Total emergence (in situ germination) during the late December 1991 to ear- ly April 1992 period was low compared to concur- rent laboratory germination on the same seed lots. On average, only 27% of all sown nutlets emerged (Table 1), with the majority occurring by the end of January. In subsequent years, emergence was as high as 34.0% (1993) and as low as 1.7% (1995). It is likely, therefore, that nutlets remained dormant or died within the seedbank, and constituted a sig- nificant constraint on growth of the founding pop- ulation. Seedling establishment and mortality. A total of 315 live seedlings and established plants were found over the entire 1991—1992 growing season, corresponding to densities between 150 and 175 plants/m? (comparable to the natural population at Edgewood Park [Table 1]). Physical contact and shading between the seedlings and other plants were minimal because of the 1) virtual absence of annual grasses, 2) relatively large spaces between Nassella pulchra bunches and 3) the open or lax growth forms of the common herbs (e.g., Perider- idia kelloggii, Delphinium virgatum) in this serpen- tinite grassland. Despite this moderate production of seedlings at Pulgas Ridge, fewer than half survived to repro- duce by early June 1992 (Table 3). Only 120 in- dividuals completed fruit formation, or 38% of the total plants produced during the growing season and 10% of the total nutlets sown (the initial yield). Overall, survivorship to reproduction was low in the experimental population compared to that ob- served in the natural population at Edgewood Park. Mortality began early, with weekly mortality rates as high as 16.9% per week during the 28 January to 10 March period. The principle cause was dif- ficult to identify from observations of grazing, des- iccation, and etiolation stresses. Grazing by mi- croherbivores (insects, snails, etc.) was the most commonly observed stress, affecting 4—34% of all live plants within plots during the growing season. Other stresses, including pathogens, may also be important during the early phases of population PAVLIK AND ESPELAND: DEMOGRAPHY OF A SERPENTINITE ANNUAL 37 TABLE 3. NUMBER OF NUTLET SOWN IN EXPERIMENTAL PLoTs (INPUT) DURING REINTRODUCTION AT PULGAS RIDGE AND THE NUMBER OF REPRODUCTIVE PLANTS SUBSEQUENTLY PRODUCED IN THOSE PLOTS AT PULGAS RIDGE, 1991—1995. Initial yield is the ratio of first year reproductive plants in spring to the number of nutlets input during the previous fall. Number of reproductive : Spring Fall input plants produced in spring aaiaal Year Nutlets 1992 1993 1994 1995 yield L991 W776 120 64 17 5 0.102 1992 514 _ Ty 29 4 0.228 1993 2000 — — Li2 25 0.056 1994 6450 — —_ _— 111 0.017 growth, but these were not assessed during this study. Mean survivorship increased during 1993 and 1994 (63 and 53% respectively), but decreased to 13.6% in 1995. The decrease was caused by flood- ing in the ephemeral creek channel which inundat- ed some of the reintroduction plots. Such large variation in survivorship, biased towards the low end of the range, characterizes the experimental population at Pulgas Ridge and not the natural pop- ulation at Edgewood Park. Plant size and nutlet production. Mean plant size at Pulgas Ridge in 1992 was less than that mea- sured at Edgewood Park (3.5 + 1.5 cm vs. 6.9 + 7.09 cm), but only if the large colonizing plants of the natural population were included in the latter estimate (Table 2). Colonizing plants were those found downslope in a previously unoccupied, con- tiguous area (Fig. 2). By excluding the colonists, mean plant size of the introduced population com- pared favorably with that of the natural population (3.5 cm vs. 4.0 cm, respectively). During other years, mean plant size and nutlet output in the ex- perimental population at Pulgas Ridge was com- parable to those observed in the same years at Ed- gewood Park. Missing from Pulgas Ridge were the few, large, fecund individuals observed primarily as colonists in the new area at Edgewood Park. This indicates that although the general conditions for reproduction at Pulgas Ridge were suitable, they were not optimal. Perhaps optimal microhabitat patches do exist at the Pulgas Ridge reintroduction site, but they were not sown with nutlets during these reintroductions. Such patches can have a dis- proportional effect on population growth by pro- ducing a few, highly fecund individuals that gen- erate hundreds, rather than tens, of nutlets each. Persistence of reintroduced cohorts. Although the initial yields of reproductive plants could be as high as 23% at Pulgas Ridge (relative to the number of sown nutlets) and mean plant size was compa- rable to that of the natural population, none of the 38 MADRONO founding cohorts was able to increase in size during the study period (Table 3). Reductions in cohort population size between years were between 50 and 90% regardless of how the nutlets were sown (pre- cision or streak) or the patterns of yearly precipi- tation (dry or wet). Despite a total input of more than 10,000 founding nutlets, fewer than 150 re- productive plants were found during the springs of 1995, 1996 and 1997. Again, we believe that post- dispersal factors, including high nutlet mortality and poor conditions for germination, were probably responsible. Even after a year such as 1994, when the natural population at Edgewood Park was flour- ishing, the reintroduced population was not able to increase its numbers. Progressive declines during unfavorable years with high rainfall indicate that the reintroduction of A. duttonii to Pulgas Ridge is unlikely to have conservation value or evolutionary potential and must be considered a biological fail- ure. We do expect, however, that germination from a declining but persistant seed bank at Pulgas Ridge will continue for a few more years. Ex situ studies have shown that nutlets of this species are long- lived (data not shown), and in situ observations have confirmed that nutlets can produce plants after 8 years of quiescence during periods of extraordi- nary annual rainfall. Perhaps the persistence of this population, as viewed from the standpoint of its ability to produce reproductive plants, is better judged by the long-term activity of the seed bank. If that seed bank were much larger (in the range of 10°—10°), the adult population produced even in un- favorable years might be enough to provide repro- ductive recharge. Collecting many nutlets from the natural population in order to boost the long-term activity of the seed bank may be another approach towards spreading the risk of extinction between multiple populations. If the Edgewood seed bank is as large as production measures indicate (Table 2), as much as 50% of the nutlet crop in a “‘good year”’ (e.g., total averaged 1.6 million in 1993/1994) could be harvested and transferred because the re- maining crop would be equivalent to maximum production in a “‘bad year’ (750,000 in 1991). The effects of harvest on the natural population would have to be monitored, but the very large existing seed bank should be able to buffer the impact for a few, non-successive years (Guerrant 1996). Care- ful spreading of these nutlets across the Pulgas site may also increase the probability of contacting op- timal, but otherwise invisible, microsites that en- courage the growth and persistence of the seed bank. CONCLUSIONS The natural population of Acanthomintha dutto- nii at Edgewood Park significantly increased in abundance and density during the 1990-1994 pe- riod. There were no strong correlations between [Vol. 45 density with overall temperature or precipitation - patterns, but in general, high density and high yield were associated with average or below-average years of precipitation and the declines were asso- ciated with above-average rainfall years. Survivor- ship to reproduction remained fairly high and con- sistent, indicating that the trends were due to vari- ations in nutlet production and the influence of cryptic factors that operate in the seed bank (nutlet mortality, germination). Nutlet production was in the range of 10°—10° per year for the population as a whole and it is likely that the lens of suitable serpentinite clay habitat at Edgewood Park can be- come saturated with A. duttonii nutlets. The experimental population at Pulgas Ridge dif- fered in several critical respects from the natural population at Edgewood Park. First, the cryptic seed bank factors at Pulgas Ridge, especially low germination (emergence) and perhaps high nutlet mortality, may have placed a severe constraint on population growth. Secondly, survivorship to repro- duction was more variable and more likely to be low at Pulgas Ridge. Finally, total nutlet production at Pulgas Ridge was on the order of 10° per year, even though plant size and nutlet output compared favorably with the natural population. These three features combined to reduce the potential for self- sustained growth in the experimental population. The reintroduction plots established in the winter of 1991 had 120 plants in the spring of 1992, but only 64 reproductive plants in 1993, 17 in 1994 and 5 in 1995. Similar patterns of decline in abundance were observed in the plots in all other years. It is likely, therefore, that the experimental population at Pulgas Ridge will be extirpated within the next few years depending on the activity of the seed bank. This failure to produce a self-sustaining pop- ulation of A. duttonii, despite having taken great care in site selection, sowing and monitoring of the reintroduction, emphasizes the urgent need for in situ preservation of self-sustaining natural popula- tions of annual serpentinite species. ACKNOWLEDGEMENTS This paper is dedicated to the memory of Jim Jokerst, student of the California flora and field botanist extraor- dinaire. His knowledge, dedication, and good cheer are deeply missed. We gratefully acknowledge the assistance and field ex- perience of Toni Corelli (California Native Plant Society), David Christy and Roman Gankin (County of San Mateo, Parks and Recreation), Jack Kuhn and Nick Ramirez (Ed- gewood Park), Joe Naras (San Francisco Water Depart- ment), Susan Sommers (California Native Plant Society), and Francis Wittman. Ken Berg, Ann Howald, Sandy Morey and Diane Steeck of the Plant Conservation Pro- gram, California Department of Fish and Game made funding for the research possible and contributed many helpful suggestions. Finally, thanks to Drs. Peggy Fiedler and Edward Guerrant, Jr. for comments that decidedly im- proved the manuscript. 1998] LITERATURE CITED ARMSTRONG, J. K. AND L. F HUENNEKE. 1992. Spatial and temporal variation in species composition in Califor- nia grasslands: The interaction of drought and sub- stratum. /n Baker, A. J. M., J. Proctor, and R. D. Reeves (eds.), The Vegetation of Ultramafic (Serpen- tine) Soils. Intercept, Andover, UK. Pp. 213-233. CNDDB (CALIFORNIA NATURAL DIVERSITY DATA BASE). 1989. File information on Acanthomintha obovata ssp. duttonii. State of California, Department Fish and Game, Natural Heritage Program, Sacramento, CA. FALK, D. A. AND K. E. HOLSINGER (eds.). 1991. Genetics and Conservation of Rare Plants. Oxford University Press, New York. 272 pp. FALK, D. A., OLWELL, M., AND C. I. MILLAR (eds.). 1996. Restoring Diversity: Strategies for Reintroduction of Endangered Plants. Island Press, Washington, DC. 505 pp. FIELDLER, P. L. 1992. Cladistic test of the adaptational hy- pothesis for serpentine endemism. /n Baker, A. J. M., J. Proctor, and R. D. Reeves (eds.), The Vegetation of Ultramafic (Serpentine) Soils. Intercept, Andover, UK. Pp. 421-434. GUERRANT, E. O. JR. 1996. Designing populations: de- mographic, genetic and horticultural dimensions. /n Falk, D. A., Olwell, M. and C. I. Millar (eds.), Re- storing Diversity: Strategies for Reintroduction of En- dangered Plants. Island Press, Washington, DC. Pp. 171-207. GUERRANT, E. O. JR. AND B. M. PAVLIk. 1997. Reintro- duction of rare plants: Genetics, demography and the role of ex situ conservation methods. /n Fiedler, P. L. and P. M. Kareiva (eds.), Conservation Biology: For the Coming Decade. Second edition, Chapman and Hall, New York. Pp. 80-108. HAMRICK, J. L., M. J. W. Gopt, D. A. MURAWSKI, AND M. D. LOvELEss. 1991. Correlations between species traits and allozyme diversity: implications for conser- vation biology. Jn Falk, D. A. and K. E. Holsinger (eds.), Genetics and Conservation of Rare Plants. Ox- ford University Press, New York. Pp. 75-86. HICKMAN, J. C. (ed.). 1993. The Jepson Manual: Higher Plants of California. University of California Press, Berkeley. Pp. 713. Howa Lp, A. 1996. Translocation as a mitigation strategy: lessons from California. Jn Falk, D. A., Olwell, M. and C. I. Millar (eds.), Restoring Diversity: Strategies for Reintroduction of Endangered Plants. Island Press, Washington, DC. Pp. 293-329. JOKERST, J. D. 1991. A revision of Acanthomintha obovata (Lamiaceae) and a key to the taxa of Acanthomintha. Madrono 38:278-286. KRUCKEBERG, A. R. 1984. California serpentines: flora, vegetation, geology, soils and management problems. University of California Publications in Botany 78. 180 pp. Lesica, P. AND EF W. ALLENDORF. 1992. Are small popu- lations of plants worth preserving? Conservation Bi- ology 6:135-139. McCarten, N. F 1986. A study of the ecological aspects related to the reintroduction of Acanthomintha obov- ata ssp. duttonii. State of California, Department Fish PAVLIK AND ESPELAND: DEMOGRAPHY OF A SERPENTINITE ANNUAL 59 and Game, Endangered Plant Program, Sacramento, CA. 36 pp. PAVLIK, B. M. 1987. Autecological monitoring of endan- gered plants. /n Elias, T. (ed.), Conservation and Management of Rare and Endangered Plants. Cali- fornia Native Plant Society, Sacramento. Pp. 385-— 390. PAVLIK, B. M. 1994. Demographic monitoring and the re- covery of endangered plants. /n Bowles, M. and C. Whalen (eds.), Restoration of Endangered Species. Blackwell Scientific, London. Pp. 322-350. PAVLIK, B. M. 1996. Conserving plant species diversity: the challenge of recovery. Jn Szaro, R. C. (ed.), Bio- diversity in Managed Landscapes—Theory and Prac- tice. Oxford University. Pp. 359-376. PAVLIK, B. M. AND M. G. BARBouR. 1988. Demographic monitoring of endemic sand dune plants, Eureka Val- ley, California. Biological Conservation 46:217—242. PAVLIk, B. M. AND E. K. ESPELAND. 1991. Creating new populations of Acanthomintha duttonii. 1. Preliminary laboratory and field studies. State of California, De- partment of Fish and Game, Endangered Plant Pro- gram, Sacramento, CA. 26 pp. PAVLIK, B. M., E. K. ESPELAND, AND F. WITTMAN. 1992. Creating new populations of Acanthomintha duttonit. II. Reintroduction at Pulgas Ridge. State of California Department of Fish and Game, Endangered Plant Pro- gram, Sacramento, CA. 35 pp. PAVLIK, B. M. AND E. K. ESPELAND. 1994. Creating new populations of Acanthomintha duttonii. IV. Demo- graphic performance at Pulgas Ridge and Edgewood Park. State of California Department of Fish and Game, Endangered Plant Program, Sacramento, CA. 25 pp. PAVLIK, B. M. AND E. MANNING. 1993. Assessing limita- tions on the growth of endangered plant populations, I. Experimental demography of Erysimum capitatum ssp. angustatum and Oenothera deltoides ssp. how- ellii. Biological Conservation 65:257—265. PAVLIK, B. M., D. NICKRENT, AND A. M. HowALp. 1993. The recovery of an endangered plant. I. Creating a new population of Amsinckia grandiflora. Conserva- tion Biology 7:510—526. PAVLIK, B. M. AND M. W. SKINNER. 1994. Ecological char- acteristics of California’s rare plants. 7n Skinner, M. W. and B. M. Pavlik (eds.), California Native Plant Society’s Inventory of Rare and Endangered Vascular Plants of California, 5th ed. California Native Plant Society Special Publication No. 1, Sacramento, CA. REZNICEK, A. A. 1987. Are small reserves wortnwhile for plants? Endangered Species Update 5(2):1—3. SOMMERS, S. 1984. Edgewood Park. Fremontia 11(5):19- 20; SOMMERS, S. 1986. Survey report on Acanthomintha obov- ata ssp. duttonii. File Report, California Natural Di- versity Data Base, Sacramento, CA. STEECK, D. M. 1995. Reproductive biology of a rare Cal- ifornia annual, Acanthomintha duttonii, and its con- gener Acanthomintha obovata ssp. cordata. M.S. the- sis. University of California, Davis. 42 pp. YorkK, R. P. 1987. California’s most endangered plants. /n Elias, T. (ed.), Conservation and management of rare and endangered plants. California Native Plant Soci- ety, Sacramento. Pp. 109-120. MADRONO, Vol. 45, No. 1, pp. 40—46, 1998 REVEGETATION AFTER FOUR STAND-REPLACING FIRES IN THE LAKE TAHOE BASIN WILLIAM H. RUSSELL AND JOE MCBRIDE Department of Environmental Science, Policy and Management, University of California, Berkeley, CA 94720 ROWAN ROWNTREE USDA Forest Service, Pacific Southwest Research Station, Box 245, Berkeley, CA 94701 ABSTRACT Low and moderate intensity surface fires have been accepted as a natural and beneficial part of the upper montane mixed conifer forests in the northern Sierra Nevada, both in terms of reducing risk of crown fires and improving the ecological health of forests. Stand-replacing fires have not generally been considered ecologically significant, in part because they have been assumed to be historically unimportant. However, high intensity stand-replacing fires did occur prior to fire suppression and may have affected vegetation structure more than previously thought. The occurrence of an extensive, pre-suppression, stand-replacing fire at the south end of Lake Tahoe on Angora Ridge was supported using historical evidence, aerial photographs, and stand age analysis. Data was gathered on the structure and composition of current vegetation on Angora Ridge and three subsequent stand-replacing fires using randomly selected plots. The post fire vegetation on these sites were compared in regard to species diversity, fuel accumulation, stand density, and com- position. Results indicate that a long period of herb and shrub domination occurred on Angora Ridge after the 1890’s fire. The treeless period was followed by simultaneous recruitment of fir and pine. The sites of the three later fires have developed little forest canopy up to the present, and currently remain dominated by shrubs and small trees. While surface fuel accumulation and the density of standing dead trees were higher on Angora Ridge than on the other fire sites, species diversity was lower. The prevalence of small, low intensity, surface fires in the Sierra Nevada before European settle- ment has been widely accepted (Kilgore 1981; Boyce 1921; Show and Kotock 1924). The exclu- sion of fire through suppression has resulted in a number of structural and compositional changes to forests throughout the Sierra Nevada, including the Lake Tahoe Basin. These changes include an in- crease in fuel load (Lunan and Habeck 1973; Min- nich 1983, 1989; Wagel and Eakle 1979); a reduc- tion in species richness (Baker 1992; Bock et. al. 1978; Murray 1992); and changes in species com- position. Shade tolerant and non-fire resistant spe- cies such as Abies concolor (Gordon & Glend.) Lindley (white fir) tend to increase in dominance (Lunan and Habeck 1973; Phillips and Sure 1990), while species that require disturbance for germi- nation, and tend to be more fire resistant, such as Pinus jeffreyi Grev. & Balf. Geffrey pine) and P. ponderosa Laws. (ponderosa pine) decrease in fre- quency. The high density of forest stands that re- sults from total fire suppression may also be a con- tributing factor in the tree mortality presently oc- curring in the Lake Tahoe Basin. Low intensity surface fires are presently being used to reduce surface fuel and repress the devel- opment of understory in the forests around Lake Tahoe. The California Department of Parks and Recreation has been using prescribed surface fire on 10 to 40 ha per year in Lake Tahoe area parks to reduce fuel loading and enhance wildlife habitat (Rice 1988, 1990; Walter 1992). The USDA-For- est Service has used non-broadcast techniques such as machine and hard-pile burning for fuel re- duction on 80—360 ha per year in the Lake Tahoe Basin (Swanson 1993). The importance of fire is also being considered in its relation to manage- ment of Lake Tahoe Basin forest ecosystems by the Tahoe Regional Planning Agency (TRPA) For- est Health Consensus Group (Swanson 1993; Sweeney 1993). Concern about the ecological condition of the forested lands surrounding Lake Tahoe has been growing in recent years due to the visible decline of a large number of forest trees. This decline can be attributed to a number of factors, including an extended drought coupled with species conversion from drought tolerant pine to drought susceptible fir, high forest density, and the activity of bark bee- tles (Wenz and DeNitto 1983; Williams et al. 1992). Recent bark beetle outbreaks, though these beetles are a natural part of affected ecosystems, have re- sulted in unprecedented interest and commitment by both private citizens and public agencies toward the development of management policies and goals for the Lake Tahoe Basin forests. The TRPA Forest 1998] | TABLE 1. Forest type RUSSELL ET AL.: STAND-REPLACING FIRE 41 FIRE AREA CHARACTERISTIC. The forest type for all four fire areas was upper montane mixed conifer. The | designations ABCO and PIJE stand for the dominant tree species on those sites, Abies concolor and Pinus jeffreyi | respectively. Angora Ridge Cathedral Creek Cascade Lake Luther Fire Date ~1890 1937 1978 1987 Hectares 100 2A 6.5 8 Elevation 1950-2190 1950-2190 1950-2190 2010-2190 Soil type Meeks-Talac Meeks-Talac Meeks-Talac Meeks-Talac ABCO ABCO PIJE ABCO Health Consensus Group, which includes private land owners and representatives of public agencies, has concluded that the forests should be managed toward their pre-European state (Swanson 1993; Sweeney 1993). The historical occurrence of low and moderate intensity surface fires in Sierra Nevada montane forests is well accepted. However, due to their rarity larger and more intense stand-replacing fires, as de- fined by Romme (1980), have not been considered an important ecological factor in the evolution of these forests. Though it is true that surface fires were much more common than stand-replacing fires before European settlement and fire suppression, stand-replacing fires did occur (Kilgore and Taylor 1979) and their effect on forest structure may have been profound (Agee 1974). For example, recruit- denuded fire area is clearly visible. ment of species that have high light requirements for seedling establishment is increased by stand- replacing fires through the opening of large canopy gaps. Habitat heterogeneity and species diversity may also be increased by crown fires (Baker 1992; Minnich 1983, 1989). In addition, gaps exceeding 10 ha caused by crown fire may be important in determining the structure and composition of the Sierra Nevada forests. Therefore, stand-replacing fires must be considered in understanding pre-Eu- ropean disturbance regimes. This paper presents the findings in a study fo- cusing on four stand-replacing fires in the south Lake Tahoe area. Our purpose is to demonstrate the existence of a stand-replacing fire that occurred pri- or to systematic fire suppression and to determine the development of vegetation after such fire. 42 MADRONO Fic. 2. 4 ~ * ~~ > an % : gases i mS “fo. Se eg 22 . Sie Pa See rey» 0. LE Neh Re PE NE, NB Ae ee Tea This is the earliest available aerial photograph of Angora Ridge taken in July 1940. The fire area is clearly visible particularly on the south east side. The area appears to be dominated by shrubs. METHODS Location of study sites. The location and perim- eter of four stand-replacing fires (Table 1) were de- termined through interviews with USDA Forest Service fire management personnel (Swanson 1993) and long-term residents of the area (Craven 1993; Gwinn 1993; Hildinger 1993), and through interpretation of historical and aerial photographs (1917-1983) (USDA-For. Serv. aerial photos). The Angora Ridge Fire burned approximately 100 years ago, covered approximately 100 ha, and was located on both sides of Angora Ridge between the present location of the Angora lookout tower and lower Angora Lake. The fire ran from 1950 to 2190 m in elevation burning both the northwest and the southeast sides of Angora Ridge. The area’s vegetation is currently dominated by a Abies con- color-Pinus jeffreyi mixed conifer forest type with Calocedrus decurrens (Torrey) Florin (incense ce- dar) included at the lower elevations, Abies mag- nifica Andr. Murray (red fir) becoming more prom- inent at higher elevations, and occasional occur- rences of both Pinus monticola Douglas (western white pine) and P. contorta ssp. murrayana (Grev. & Balif.) Critchf. odgepole pine). The Cathedral Creek Fire burned in 1937, cov- ered approximately 21 ha, and was located in the Cathedral Creek drainage on the southeast slope above Fallen Leaf Lake, from 1950 to 2190 m in elevation. The same forest type dominates the area surrounding both fires, and both have soils of the Meeks-Tallac formation type (USDA 1974). The Cascade Lake Fire burned in 1978, covered approximately 6.5 ha and was located on the south- east slope above Cascade Lake, from 1950 to 2190 m in elevation. The forest type surrounding this fire is similar to those above with more abundant Pinus Jeffreyi. The Luther Fire burned in 1987, covered ap- proximately 8 ha, and was located on the northwest slope above Christmas Valley, from 2010 to 2190 m in elevation. Vegetation and soil types are similar to those in the other three fire areas. Sampling techniques. Stand structure data was collected using 10 m by 20 m rectangular plots that were randomly selected along elevational gradients and distance from center of site within each study site. Within each plot slope, aspect, elevation, and forest type were recorded as well as information on size and identification of all live and dead standing trees present. The number of seedlings (trees less than 61 cm in height) and saplings (trees less than 10 cm in diameter) of each species were also re- corded. Canopy cover was determined for the four cardinal directions at the center of each plot with a spherical densiometer. Fuel load was determined etuas re . FIG: 3: on both sides of the ridge. using a natural forest residue photo series (Blonski and Schramel 1981). Stand age was determined by taking a core sample from the largest tree of each species present on each plot. Three 1.8 m diameter circular nested plots were used within each 10 m by 20 m rectangular plot to sample herbaceous and shrub species. For all shrubs present, cover class was determined using a standard cover scale (Daubenmire 1959, 1968). The presence of all herbaceous and shrub species on 1.8 m plots was recorded. Stand structure analysis. Data collected on 65 timber plots and 195 herbaceous-shrub plots yield- ed the total and relative frequency and density of each tree species, the total and relative frequency and density of standing dead trees, the total and relative frequency and density of seedlings and sap- lings of each species, the average canopy cover for each study site, the average percent cover of each shrub species, the average fuel load (metric tons/ ha) for each study site (Blonski and Schramel 1981), and three measures of species diversity (spe- cies richness, species evenness, and the Shannon diversity index). RESULTS Interpretation of aerial photographs. Aerial pho- tographs of the Angora Ridge indicate that a long RUSSELL ET AL.: STAND-REPLACING FIRE 43 $z i) ne = NW At rs 77 Da c-™ — ety eS ikl Se — v 4 * ar, a Beet oe MSH 9} yore? f ; bs) This 1983 photograph shows Angora Ridge in a state similar to its present condition, with high density forests interval occurred before tree cover developed fol- lowing the fire in the 1890’s. There is little or no tree canopy visible in aerial photographs in the fire area before 1952. Historical photographs taken in 1917 (Fig. 1) and aerial photographs taken in 1940 (Fig. 2) show that the post-burn was dominated by brush fields. Aerial photographs taken in 1952 and 1966 show the gradual development of tree canopy on Angora Ridge. By 1976 (Fig. 3) aerial photographs indicate that a significant forest canopy has developed but there are still areas of shrub domination. Stand age. Tree ages for the dominant tree on each plot ranged from 21 to 95 years on Angora Ridge. The modal age was 70 years for all species with no significant difference in age between Abies concolor, A. magnifica, and Pinus jeffreyi. This is consistent with the existence of a 100 year old stand-replacing fire. The 30 year delay in tree re- cruitment is consistent with shrub domination dur- ing the same period. As sheep grazing was common in the study area during the period in question it may have influenced vegetation recruitment pat- terns, however, the occurance of shrub domination on the Cascade Lake fire and Luther fire areas sug- gests that grazing is not necessarily a factor. On the Cathedral Creek Fire area the dominant tree age on each plot ranged from 20 to 54 years 44 MADRONO [Vol. 45 Stand Density Individuals/hectare Cathedral Creek Angora Ridge Fic. 4. Stand density. with a modal age of 39 years. There was no sig- nificant difference in age between Abies concolor and Pinus jeffreyi. This result is consistent with a 56 year old fire. Neither the Cascade Lake Fire area nor the Lu- ther Fire area had enough tree recruitment to de- termine stand age through coring. Both areas are dominated by shrubs. Because of their recent oc- currence, however, the age of the burns were easily determined through fire records and interviews with USDA Forest Service fire suppression personnel (Johnson 1993). None of the sites studied received artificial re- planting treatments so that stand age reflects natural recruitment and development. Structural analysis. The density of mature trees was much higher on the Angora Ridge Fire site than on the Cathedral Creek, Cascade Lake, and Luther Fire sites (Fig. 4). This difference in the density of mature trees was associated with other structural differences such as canopy and shrub cover. Canopy cover averaged 72% on Angora Ridge and less than 10% on each of the other three fire sites. Shrub cover dominated these three sites (Cathedral Creek = 88%; Cascade Lake = 76%; Luther Fire = 87%) while Angora Ridge exhibited only 13% shrub cover. These results suggest that the balance between tree and shrub cover is a func- tion of the time interval since fire and the ability of trees, once they become established, to shade out shrubs. Shannon Diversity Index Luther Fire Cathedral Creek Cascade Lake Angora Ridge Fic. 5. Shannon diversity index. ie Mature LJ Sapling LJ Seedling Cascade Luther Lake Fire A difference in species diversity occurred be- tween the Angora Ridge and the other three fire areas as is indicated by species richness counts and the Shannon diversity index (Fig. 5) which were higher on the three younger sites. Though the num- ber of tree species is relatively consistent on all of the sites, the abundance of herbaceous and shrub species was lower in the older fire area. This is due primarily to increased shading resulting from great- er tree canopy on that site. The accumulation of fuel also varied between the older site and the three younger sites with the high- est fuel accumulation occurring on Angora Ridge (Fig. 6). Results indicate a pattern of vegetation devel- opment consistent with conventional models of post-fire succession in mixed conifer forests (Lyon and Stickney 1976; Kercher and Axelrod 1984). However, the proportional abundance of the two major genera of trees, Abies and Pinus, was about the same for Angora Ridge, Cathedral Creek, and the Luther fire (Table 2) with white fir clearly the dominant species. In the Cascade Lake Fire area, however, Pinus was clearly dominant. The relative dominance of Abies and Pinus on these sites is like- ly the result of seed availability rather than a result of post-fire competition and succession. This is fur- ther demonstrated by a comparison of the relative density of mature fir and pine to seedlings and sap- lings (Table 2). The earliest recruitment of fir ap- load in tons/hectar Fuel Angora Cathedral Cascade Luther Ridge Creek Lake Fire Fic. 6. Fuel load. 1998] RUSSELL ET AL.: STAND-REPLACING FIRE 45 TABLE 2. RELATIVE DENSITY. ABMA = Abies magnifica, ABCO = A. concolor, PIJE = Pinus jeffreyi, PICO = P. contorta (ssp. murrayana). Angora Ridge Cathedral Creek Cascade Lake Luther Fire Species Mat Sap Seed Mat Sap Seed Mat Sap Seed Mat Sap Seed ABMA 6 Fs 23 4 — 1 — — —- — — — ABCO 89 93 96 79 83 99 — 4.5 — -— 86 83 PIJE 8.5 4.7 — | 16 l — All 100 oo 14 17 PICO = = = a= — — — 4.5 — a — — pears to have coincided with the earliest recruit- ment of pine. There does not appear to be a req- uisite period of pine domination on these sites as was described for a similar site in the Sierra Nevada (Bock and Bock 1969; Bock et al. 1976). The con- tinuous regeneration of A. concolor found in this study is comparable to that found by Conard and Radosevich (1982). Mortality. The total density of dead trees and the relative mortality of individual species may be use- ful in predicting the future direction of vegetation change. Clearly the density of dead to live trees were much higher on the Angora Ridge than on the other three sites (Fig. 7) suggesting that mortality may be connected to stand density. The mortality of A. concolor was much higher than that of any other species resulting from the high density of this species. All trees that were recorded as dead on both fire sites showed evidence of bark beetle ac- tivity. CONCLUSIONS The Angora Ridge fire occurred at a time before the implementation of systematic fire suppression (Craven 1993; Gwinn 1993; Hildinger 1993). The existence of this fire and the ecological information collected within its perimeters indicates that not only did stand-replacing fires exist before fire sup- pression, but that the process of forest development after such fires can be lengthy, including a long period with minimal forest canopy cover. In addi- tion, results point to a number of other interesting conclusions including a lack of a post-fire pine domination period on the sites studied, higher spe- cies diversity on the more recent fire sites, and higher fuel accumulation on the oldest fire site. Simultaneous recruitment of pine and fir oc- curred on the four fire sites during their post-fire periods. Fir and pine did not follow the traditional successional sequence after large fires that usually includes a period of pine domination followed by increasing fir domination (Bock and Bock 1969; Bock et al. 1976; Lyon and Stickney 1976; Kercher and Axelrod 1984). It appears that there was no obligatory period of post-fire pine domination in the study areas, and that fir regenerated with as much facility as pine. The higher species diversity on the more recent fire sites is important in that diversity seems to de- cline as canopy cover increases. This supports the notion that in some communities diversity has an inverse relationship to secondary successional de- velopment of forests after fire (Shafi and Yarranton 1977). Higher fuel accumulation and a high density of standing dead on the older fire site suggests that the long fire-free interval has increased the probability of another high intensity fire. In the quest to determine the proper role of fire in the forest management scheme in the Tahoe Ba- sin, stand-replacing fires need to be understood as part of the pre-European fire regime. In addition to periodic surface fires, stand-replacing fires have been instrumental in the formation of canopy gaps, and the maintenance of habitat heterogeneity and species diversity. Including stand replacing-fire as a part of forest management in the Lake Tahoe Ba- sin may be unpopular socially and politically, and difficult to implement. However, suppression of all Density of Standing Dead trees/ha Cathedral Creek Fic. 7. Standing dead. Cascade Lake Luther Fire 46 MADRONO stand-replacing fires, in the long run, may be more costly than their careful management. ACKNOWLEDGMENTS We thank Dr. Lester Rowntree of the Environmental Studies Department at San Jose State University for his role as adviser on the Master of Science thesis project from which this study originated. We thank the USDA Forest Service for financial support and the technical sup- port of John Swanson and Katherine Culligan. For their comments on the manuscript we thank Dr. Robert Swee- ney and Dr. V. Thomas Parker (San Francisco State Uni- versity), Scott Stevens (University of California Berke- ley), and John Swanson (USDA Forest Service, Lake Ta- hoe Basin Management Unit). LITERATURE CITED AGEE, J. 1974. Environmental impacts from fire manage- ment alternatives. Nat. Park Serv., Western Regional Office, San Francisco, CA. BAKER, W. L. 1992. Effects of setthement and fire sup- pression on landscape structure. Ecology 73:1879. BLONSKI, K. S. AND J. L. SCHRAMEL. 1981. Photo series for quantifying natural forest residues: southern Cas- cades, northern Sierra Nevada. USDA Forest Service General Technical Report PSW-56. Bock, J. H. AND C. E. Bock. 1969. Natural reforestation in the northern Sierra Nevada-Donner Ridge burn. Proc. Annual Tall Timbers Fire Ecology Conf. 9:119— 129; Bock, J. H., C. E. Bock, AND V. M. HAWTHORNE. 1976. Further studies of natural reforestation in the Donner Ridge burn. Proceedings of the Annual Tall Timber Fire Ecology Conference, 14:195—200. Bock, J. H., M. RAPHAEL, AND C. E. Bock. 1978. A com- parison of planting and natural succession after a for- est fire in the northern Sierra Nevada. Journal of Ap- plied Ecology 15:597—602. Boyce, J. S. 1921. Fire scars and decay. The Timberman 22S CONARD, S. G. AND S. R. RADOSEVICH. 1982. Post fire succession in white fir (Abies concolor) vegetation of the northern Sierra Nevada. Madrono 29:42—56. CRAVEN, B. 1993. Fallen Leaf Lake. Personal communi- cation. DAUBENMIRE, R. E 1959. Canopy coverage method of veg- etation analysis. Northwest Science 33:43—64. DAUBENMIRE, R. E 1968. Plant communities: A textbook of plant synecology. Harper & Row, New York. Gwinn, M. H. 1993. Fallen Leaf Lake. Personal commu- nication. HILDINGER, J. 1993. Angora Lakes Resort. Personal com- munication. JOHNSON, M. 1993. USDA Forest Service, Lake Tahoe Ba- sin Management Unit. Personal communication. KERCHER, J. R. AND M. C. AXELROD. 1984. A process model of fire ecology and succession in a mixed co- nifer forest. Ecology 60:129-142. KILGORE, B. M. AND D. TAYLOR. 1979. Fire history of a Sequoia-mixed conifer forest. Ecology 60:129—-142. KiLGorE, B. M. 1981. Fire in ecosystem distribution and structure: western forests and scrublands. Pp. 58-89 in H. A. Mooney, et al (eds.). Fire regimes and eco- [Vol. 45 system properties. USDA General Technical Report WO-26. LunaNn, J. S. AND J. R. HABEcK. 1973. The effects of fire exclusion on ponderosa pine communities in Glacier National Park, Montana. Canadian Journal of Forest Research 3:574—-579. Lyon, L. J. AND P. FE STICKNEy. 1976. Early vegetal suc- cession following large northern Rocky Mountain wildfires. Proceedings of the Tall Timbers Fire Ecol- ogy Conference 14:355—375. MINNICH, R. A. 1983. Fire mosaics in southern California and northern Baja California. Science 219:1287— 1294. MINNICH, R. A. 1989. Chaparral fire history in San Diego county and adjacent Baja California: An evaluation of natural fire regimes and the effect of suppression management, Pp. 37—47 in S. C. Keeley (ed.), The California chaparral: paradigms re-examined. No. 34, Natural History Museum L. A. Co. Murray, M. 1992. Suppression threatens subalpine mead- ows. Inner Voice 4:10. PHILLIPS, D. L. AND D. J. Sure. 1990. Patch-size effects on early succession in southern Appalachian forests. Ecology 71:204-—212. Rick, C. L. 1988. Fire history of Emerald Bay State Park. California Department of Parks and Recreation. Wild- land Resource Management, Walnut Creek, CA. Rick, C. L. 1990. Fire history of the Sierra District of the California Department of Parks and Recreation. Wild- land Resource Management, Walnut Creek, CA. ROMME, W. 1980. Fire history terminology: Report to the ad hoc committee. In Proceedings of the fire history workshop, October 20—24, 1980, Tuscon, Arizona. USDA General Technical Report RM-81. Scott, E. B. 1973. The Saga of Lake Tahoe, Vol II. Sierra- Tahoe Publ. Co. South Lake Tahoe, CA. SHAFI, M. I. AND G. A. YARRANTON. 1977. Vegetational heterogeneity during a secondary (post fire) succes- sion. Canadian Journal of Botany 51:73-90. SHow, S. B. AND E. I. Kotock. 1924. The Role of fire in the California pine forests. USDA Bulletin 1294. Washington D.C. SWANSON, J. 1993. USDA Forest Service, Lake Tahoe Ba- sin Management Unit. Personal communication. SWEENEY, J. R. 1993. Personal communication. USDA Forest SERVICE. Aerial Photos, courtesy of Lake Tahoe Basin Management Unit U.S. Forest Service, South Lake Tahoe, CA. USDA Solt CONSERVATION SERVICE AND FOREST SERVICE. 1974. Soil survey: Tahoe Basin Area, California and Nevada. WAGEL, R. F AND T. W. EAKLE. 1979. A controlled burn reduces the impact of a subsequent wildfire in a pon- derosa pine vegetation type. For. Sci. 25:123-129. WALTER, G. 1992. Prescribed Burn Plan; Sugar Pine Point State Park. California Dept. of Parks and Recreation. WENZ, J. AND G. DENITTO. 1983. Evaluation of pest con- ditions in the Kiva/Taylor Creek Visitor Center Rec- reation Areas on the Lake Tahoe Management Unit. USDA Forest Service Forest Pest Management Re- port No. PSW 83-06. WILLIAMS, C. B., D. L. AZUMA, AND G. T. FERREL. 1992. Incidence and effects of endemic populations of for- est pests in young mixed conifer forests of the Sierra Nevada. USDA Forest Service Research Paper PS W- RP-212. MApRONO, Vol. 45, No. 1, pp. 47-56, 1998 FIRE REGIME OF THE LODGEPOLE PINE FOREST OF MT. SAN JACINTO, CALIFORNIA PAUL R. SHEPPARD AND JAMES P. LASSOIE Department of Natural Resources, Cornell University, Ithaca, NY 14853 ABSTRACT The objective of this study of the 1000 ha lodgepole pine (Pifus contorta ssp. murrayana) forest of Mt. San Jacinto, California, is to typify several components of the fire regime, including intensity, size, frequency, and the relationship of fire with weather and its seasonal timing. We analyzed several lines of evidence, including fire occurrence from suppression records, fire dates from scarred P. contorta ssp. murrayana, forest structure as it relates to past disturbances, forest fuel loading, dendroclimatological modeling as it relates to past fire occurrences, and fire-igniting weather characteristics. The lodgepole pine forest of Mt. San Jacinto has a regime of fires that are typically of low intensity and that typically burn small areas (<10 ha). Lightning ignites fire somewhere within this forest as frequently as every few years, but given that lightning is a spatially random process and that fires typically burn areas less than 10 ha in size, the mean fire interval for any 10 ha area is more likely on the order of 100 years. Atypically large fires (20 ha or larger) are possible, but they have burned infrequently in the past. Fires appear to be more common during years when surface and ground fuels are relatively dry (a result of below-average winter precipitation), and fires occur primarily late in or after the growing season. Mt. San Jacinto State Park managers should always prepare for lightning fires, but they can probably safely allow most lightning fires to burn naturally within this ecosystem. Because the possibility exists for fires to burn large areas, fires should probably be monitored closely to protect a unique resource of very old P. contorta ssp. murrayana trees. RESUMEN El objetivo de este estudio en un bosque de mil hectareas de Pinus contorta ssp. murrayana en Monte San Jacinto, California, es describir varios componentes del régimen de incendios, incluyendo la inten- sidad, el area, la frecuencia, las asociaciones entre incendios y el tiempo, y su sincronizaci6n estacional. Analizamos varias lineas de evidencia, incluyendo la ocurrencia de incendios segtin los registros de supresion, fechas de incendios segtin los pinos con heridas de incendio, la estructura forestal en relacion a las pasadas perturbaciones, la cantidad de combustible forestal, modelos dendroclimatoldégicos en re- lacidn a pasadas ocurrencias de incendios, y las caracterfsticas climaticas iniciadoras de incendios. El bosque de Pinus contorta del Monte San Jacinto tiene un régimen de incendios que tipicamente arden con una intensidad baja y en areas menores de 10 ha. Los rayos inician incendios en cualquier parte dentro de este bosque con una frecuencia de 4 a 7 anos, pero dado que los rayos que acompanian las tormentas son un proceso espacialmente al azar y que los incendios tipicamente queman areas menores de 10 ha, el intervalo promedio entre incendios en cualquier punto es probablemente del orden de 100 anos. Incendios atipicamente grandes (decenas de hectareas 0 mas) son posibles, pero han ocurrido con poca frecuencia. Los incendios parecen ser mas comunes durante anos cuando los combustibles forestales en la superficie y el suelo estan relativamente secos (a causa de precipitaci6n invernal menor que la normal), y los incendios ocurren principalmente hacia el final o después del perfodo vegetativo. Los guardabosques de Monte San Jacinto deben estar siempre preparados para incendios producidos por rayos, pero probablemente pueden permitir arder naturalmente la mayoria de estos incendios dentro de este ecosistema. Debido a que existe la posibilidad de incendios que queman 4reas amplias, los incendios deben ser vigilados atentamente para proteger este recurso unico de Pinus contorta longevos. INTRODUCTION The first step in formulating a fire management plan for a wilderness or natural area is determining the natural role of fire for the area (Agee 1974). Knowledge of the natural fire regime forms the ba- sis for predicting fire behavior and responding to fires, helps illustrate the inevitability of fire in nat- ural areas, and aids in the application and public interpretation of fire management plans (Mutch 1980). The (lodgepole pine) Pinus contorta ssp. murrayana, forest of the Mt. San Jacinto State Wil- derness, California, has ample evidence of past fires in the form of abundant living, fire-scarred P. con- torta ssp. murrayana located throughout the forest (Hamilton 1983). However, few of the details of the behavior of past fires can be inferred merely from the presence of fire-scarred trees. Our objective here is to typify several components of the regime of lightning-ignited fires of this forest, including in- tensity, size, frequency, the relationship of fire with weather, and seasonal timing of fires. To meet this objective, we analyzed multiple lines of evidence, including fire occurrence from suppression records, fire dates from scarred P. contorta ssp. murrayana, forest structure as it relates to past disturbances, forest fuel loading, dendroclimatological modeling 48 Peak San °1841L IV £18391 a 01798L 1793L e ae 01798L 18411 “1785L 4786L Southern CALIFORNIA Mt. San Jacinto I -- IV Fic. 1. MADRONO [Vol. 45 Long Valley Station °1757L 01881L 01778L) 91 784L 4797, * 1860E 61797L «1797L], © 1797L °1860E 0 ae SS N [____] Meadowed campground areas Contour interval: 200 m Compartment boundary lines Compartment numbers Lodgepole pine forest of Mt. San Jacinto, Southern California. Dots mark plot locations, year values indicate the years when fire-scarred P. contorta ssp. murrayana were burned, and “‘nd”’ indicates fire-scar samples that were not crossdateable. “‘L” indicates a fire that occurred late in or after the growing season, and “‘E”’ indicates a fire that occurred early in the growing season; dates with neither letter indicate that the seasonal timing of the fire could not be unequivocally determined. as it relates to past fire occurrences, and fire-ignit- ing weather characteristics. To accommodate the natural fire ecology of this lodgepole pine forest, Mt. San Jacinto State Park managers may use information from this study when formulating or amending policies for man- aging fire within their broader management goals of maintaining natural disturbance processes and protecting unique natural resources that include P. contorta Ssp. murrayana up to several hundreds of years old (Hamilton 1983). This study also adds to the understanding of the ecology of P. contorta ssp. murrayana Which occurs throughout the Sierra Ne- vada, Transverse, and Peninsular Ranges of Cali- fornia and in Baja California and is thought to have a stable ecological role and distribution that is not related to fire (Lotan and Critchfield 1990). MATERIALS AND METHODS Study site. Mt. San Jacinto (33°49’N, 116°41'W) is 160 km east of Los Angeles, California (Fig. 1, inset). Between 2550 to 3200 m elevation P. con- torta ssp. murrayana forms a xerophytic and de- pauperate forest (Thorne 1988) covering approxi- mately 1000 ha. This forest is a topoedaphic climax with relatively minor amounts of other species (Hamilton 1983). At its lower elevations (2550 to 2900 m), P. contorta ssp. murrayana is mixed with Abies concolor (white fir), which extends to lower elevations beyond the range of P. contorta ssp. murrayana. At its higher elevations (2900 to 3200 m), P. contorta ssp. murrayana is mixed with Pinus flexilis (limber pine) in a sparse forest with an open canopy. The understory consists predominantly of Ceanothus cordulatus (snow bush) and Chrysolepis sempervirens (bush chinquapin). Climate is mediterranean (cold, wet winters and warm, dry summers; Bailey 1966). Soils are de- rived from granitic parent material and are classi- fied as lithic Xerorthents: They are shallow, coarse- textured with a large volume of rock fragments and outcrops, well drained, and very low in water hold- ing capacity and fertility (Cohn, B. R. and J. G. Retelas, unpublished soil survey on file with the San Bernardino National Forest, California). The fire regime and structure of this forest prob- ably reflect little influence by Native Americans of the past. The Cahuilla Indians had the strongest presence in the study area (Smith 1957), but their main territory was the Cahuilla Valley at 1200 m elevation (Barrows 1900), considerably below the lodgepole pine forest. Summer and fall encamp- 1998] ments were common in lower canyons (Aschmann 1959) but less common on higher ridges (James 1960; Bean and Saubel 1972). The Cahuillas rarely if ever used fire as a hunting aid (Drucker 1937). Although there was some cattle grazing on Mt. San Jacinto, it ended by about 1940. Anecdotal ac- counts indicate that grazing was limited in the lodgepole pine forest and essentially restricted to the largest of the three meadow areas (Fig. 1; Ham- ilton 1983). While we recognize the potential im- pact of grazing on fire regimes of ecosystems in general (Savage and Swetnam 1990), we believe that the specific impact of grazing on the lodgepole pine forest of Mt. San Jacinto was not substantial. Field sampling took place during the summers of 1983 and 1984. We divided the 1000 ha lodgepole pine forest study site into compartments by aspect (Fig. 1): Compartment I faces east, II faces south, III is a relatively flat plateau, and IV faces west. Meadowed campground areas were excluded be- cause of the potential impacts of heavy recreational use on the fire regime, fuel loading, and forest structure. Within each compartment we located transects perpendicular to contours (Arno et al. 1993). Transects were 200 m apart, and sample points were 500 m apart along each transect. The fire-scarred P. contorta ssp. murrayana (Mitchell et al. 1983) nearest each sample point (usually no more than a few tens of meters away) was chosen as the center point for a 0.04 ha (20 m X 20 m) sample plot. Occasionally a plot contained more than one fire-scarred P. contorta ssp. murrayana up to five in one case. For each plot, we recorded elevation, slope, and aspect, and we estimated forest fuel loading using a photo series of woody residues as quantified by Blonski and Schramel (1981) for this forest type. Within each plot, we measured diameter at breast height of all trees and counted trees, saplings, and seedlings shorter than breast height. We extracted increment cores from scarred P. contorta ssp. mur- rayana to dendrochronologically date the fires that caused the scars (Sheppard et al. 1988). We also cored old, unscarred P. contorta ssp. murrayana throughout the study site to develop a site- and spe- cies-specific reference chronology for crossdating purposes. ANALYSIS Fire occurrence from suppression records. A\- though San Bernardino National Forest fire sup- pression records extend back to 1912, early records appear to be incomplete and we restricted this anal- ysis to the period since 1945. We tabulated infor- mation on the suppression of lightning fires in the lodgepole pine forest of Mt. San Jacinto. We chron- ologically listed all years with at least one recorded lightning fire and calculated the mean time period between years when fires occurred. SHEPPARD AND LASSOIE: FIRE REGIME OF LODGEPOLE PINE 49 Fire dates from scarred lodgepole pines. We crossdated (Stokes and Smiley 1968) cores from unscarred P. contorta ssp. murrayana (two cores per tree from twenty trees) and measured ring widths to the nearest 0.01 mm. We removed growth trends from each ring-width series by dividing each measurement by the corresponding value of a trend line estimated from either a modified negative ex- ponential or a linear fit (Fritts 1976). We then au- toregressively modeled the detrended series into re- sidual series of white noise (Cook 1985), which were averaged into the reference index chronology. We verified the crossdating of this chronology against a tree-ring chronology from bigcone Pseu- dotsuga macrocarpa (Douglas-fir) growing 20 km from the study site (Keen Camp Summit, 33°43'N, 116°05'W, 1432 m elevation; Drew 1972). Scarred trees were exclusively single-scarred, which allowed the effective use of increment cores to date the fires by crossdating ring-width series with the reference chronology (Sheppard et al. 1988). To establish the year of formation of the outermost ring before the fire, we crossdated non- disturbed ring growth that preceded the scar. Our crossdating of scarred samples was checked by an- other dendrochronologist. Additionally, we deter- mined the general seasonal timing of the fire from the relative completeness of the outermost ring be- fore the scar on samples for which this was possible (Baisan and Swetnam 1990). A complete or nearly complete ring (with at least some latewood cells) indicated a fire that burned after or late in the grow- ing season (late July through early September). An incomplete ring (no latewood cells) indicated a fire that burned early in the growing season (early June through mid-July). Pinus contorta ssp. murrayana structure as relat- ed to disturbance regime. We evaluated the histo- gram of diameters of sampled P. contorta ssp. mur- rayana from throughout the study site to determine the placement of this forest along the continuum of relatively even- to uneven- or all-aged structures (Hanley et al. 1975; Despain 1983; Lorimer and Frelich 1984). The use of diameters instead of ac- tual ages requires that diameter and age be related such that age need not be determined for all trees sampled. To test for this relationship, we compared ages to diameters for a subset of P. contorta ssp. murrayana located throughout the study site for which pith dates were determined (Fig. 2). Based on this relationship, size is sufficiently related to age to use a histogram of diameters to characterize the age structure of P. contorta ssp. murrayana in this forest. Relationship between fire occurrence and weath- er. We dendroclimatically modeled (Fritts 1976) the Keen Camp Summit tree-ring chronology with monthly weather data for the NOAA climatic di- vision #7 (southeastern deserts) of California. This tree-ring chronology correlates well with precipi- 50 MADRONO 800 age = 81 +5.5 - diameter 600 R?=71% 2 © = 400 ® {e)) —t 200 0 [Vol. 45 50 60 70 80 90 Diameter (cm) Fic. 2. tation of the winter prior to the growing season (Fig. 3a). When the December through March pre- cipitation totals are summed, the winter season cor- relation with the chronology is 0.60 (P < 0.01). We extended the Keen Camp Summit tree ring chro- nology, which ends in 1966, by appending to it an indexed series of winter (December through May) precipitation at Idyllwild, California (1700 m ele- vation, 8 km south of Peak San Jacinto). We then assessed the relationship between fire occurrence and winter precipitation by comparing the average chronology index value for years when fire oc- curred to the average index value for the chronol- ogy (1.0 by definition, Fritts 1976). Seasonal timing of fire risk. We searched the dai- ly weather records from the Long Valley Ranger Station (2585 m elevation; Fig. 1) and the Idyllwild Fire Station for evidence of lightning occurrence for the entire study site. Combining data from these two stations located on opposite sides of the study site accounted for the local and sporadic nature of summer convection storms. Mid-May through mid- September is the period of likely convection storms (Tubbs 1972), and any day within this period with data or comments about rain, thunder, or lightning was considered to have had lightning possible. We grouped counts of these days into half-month pe- riods and averaged across years for which complete records exist for both stations (1965 to 1974 and 1979 to 1984). RESULTS Fire occurrence from suppression records. The mean time period between years with at least one suppressed lightning fire is 5.2 + 1.1 years (all er- ror estimates are 95% confidence intervals) for the 1000 ha lodgepole pine forest from 1945 to 1987 (Table 1). Notably, 6 distinct fires were recorded for 1972, including two within the same survey sec- Age-diameter relationship for P. contorta ssp. murrayana of Mt. San Jacinto. tion (259 ha) on each of two different days. Most of the listed fires occurred after mid-July (81%). Furthermore, all of the listed lightning fires were declared as spot fires—less than 0.4 ha in size at the time they were suppressed. Fire dates from scarred P. contorta ssp. murray- ana. We were able to crossdate 56 (88%) of the scarring events that we sampled (Table 2). Within plots with more than one fire-scarred P. contorta ssp. murrayana, all crossdatable samples dated the fires to the same year. We were able to determine the season of scarring in 52 scar samples, most of which (92%) contain at least some latewood cells in the outermost prescar ring and thus indicate fires that burned late in or after the growing season (Fig. 1). Many of the fires dated from scarred P. contorta ssp. murrayana appear to be independent of each other, either by date and/or location (Fig. 1). For example, the year 1860 is represented by four plots that are widely separated from one another. To quantify the degree of this independence, we com- pared fire dates within each pair of adjacent plots, and most pairs (79%) have fire dates that differ from one another. Most adjacent pairs with the same fire date are located within three different clumps in the study site: five 1797 dates are located in the southeastern edge of Compartment I, nine 1881 dates in the northern half of Compartment I, and five 1752 dates in Compartment III (Fig. 1). Our collection of fire scars includes only one date in the 20th century (1910). The lack of fire scars dating in the 20th century contradicts the sup- pression record (Table 1), which indicates frequent natural fires during the 20th century, especially since 1945. We attribute this apparent inconsistency to a bias of sampling visually obvious, larger scars of fires that burned further back in time. These P. contorta ssp. murrayana respond to fire scarring SHEPPARD AND LASSOIE: FIRE REGIME OF LODGEPOLE PINE aA Keen Camp Summit with Precipitation Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Current year | t t] 4 ’ é é e ‘? au ” 1998] (a) 5 w 2 O O | Prior year | | (b) 2 SLM Na | £ 14 | iN ! | 0 1750 1800 1850 Fic. 3. —— mos — ~ = = se) = 5 1900 1950 (a) Correlations of the bigcone Douglas-fir index chronology from Keen Camp Summit to total monthly precipitation for NOAA climatic division 7 (southeastern deserts) of California. Reference lines indicate critical value for significance for alpha = 0.05, n = 71. (b) Winter (December through March) precipitation index chronology. The solid line (1750 to 1966) is the Keen Camp Summit chronology, the dashed line (1967-1986) is from the actual meteorological data from the Idyllwild station, and the reference line of 1.0 indicates average winter precipitation. Open dots indicate all years with evidence of fires, and closed dots indicate years with evidence of early-season fires. with very slow growth rates after the scar, and scars from fires during the 20th century are small and inconspicuous such that we probably under-sam- pled them relative to older fire scars. Because of this sampling bias, we restricted our interval anal- ysis of fire-scar dates to the period from 1752 to 1886, and the mean time period between years with fire anywhere in the 1000 ha forest is 5.0 + 2.2 years (Table 2). Pinus contorta ssp. murrayana structure as re- lated to disturbance regime. The histogram of di- ameters of all sampled lodgepole pines is adequate- ly fit by a negative exponential curve pattern (Fig. 4a). The three areas with several adjacent plots with the same fire dates (1752, 1797, and 1881; Fig. 1) warrant closer scrutiny, and we evaluated size his- tograms of P. contorta ssp. murrayana located within plots burned during those years. In the case of 1752, the size histogram is shaped broadly in a negative exponential pattern (Fig. 4b), and it has a slope coefficient that does not differ substantially from that of the P. contorta ssp. murrayana size histogram for the entire forest. In the case of 1797, the size histograms are not shaped in a negative exponential pattern (Fig. 4c). In the case of 1881, the size histogram is shaped broadly in a negative exponential pattern, but the number of individuals in the smallest size class is overestimated (Fig. 4d). Forest fuel loading. The fuel loading within this lodgepole pine forest averages 18.9 + 2.9 tonnes/ ha, similar to the estimate of 15.4 tonnes/ha re- ported for a single area near the smallest meadowed campground area (Fig. 1; Compartment IV; Hana- walt and Whittaker 1976). The general fuel type of the lodgepole pine forest of Mt. San Jacinto is best described as Model H of the National Fire-Danger Rating System (Deeming et al. 1977), in which co- nifer trees pre-dominate; duff, litter, and branch- wood are the primary ground fuels; and needles are less than 5 cm long. Additionally, as an indicator 52 MADRONO TABLE 1. [Vol. 45 SAN BERNARDINO NATIONAL FOREST SUPPRESSION RECORDS (1912 TO 1989) OF LIGHTNING FIRES WITHIN THE LODGEPOLE PINE FOREST OF MT. SAN JACINTO*. All cf these fires were listed as spots, smaller than 0.4 ha when suppressed. Mean time period between years with fire = 5.2 + 1.1 years‘. ? The reports included all fires that burned within the San Jacinto District, but only lightning fires within the lodgepole pine forest were considered pertinent to this study. > All sections are approximately 259 ha (1 mi’) and are within Township 4S, Range 3E of the San Bernardino Meridian. * This was not calculated for years prior to 1945 because of the incompleteness of the records for that time period. ¢ For the period 1945 to 1987 and for the 1000 ha lodgepole pine forest. Return Section interval location” Date (yrs)° ea August 4, 1926 29 September 4. 1931 zl September 8, 1945 6 21 September 28, 1951 4 23 August 14, 1955 5 23 September 7, 1960 4 30 July 26, 1964 of the living fuel loading, the average stocking den- sity of this forest is 375 + 73 stems/ha for all spe- cies. Relationship between fire occurrence and weath- er. For the period from 1752 (earliest fire-scar date) to 1986 (last year of our chronology of winter pre- TABLE 2. CHRONOLOGICALLY SORTED YEARS FOR WHICH A FIRE SCAR FROM PINUS CONTORTA SSP. MURRAYAMA WITHIN THE LODGEPOLE PINE FOREST OF MT. SAN JACINTO WAS DENDROCHRONOLOGICALLY DATED. Mean time period be- tween years with fire = 5.0 + 2.2 years>. * The number in parentheses for some years is the number of trees scarred during that year; years without parentheses have just one scarred tree. b For the period 1752 to 1886 and for the 1000 ha lodgepole pine forest. Re- Re- Re- Re- turn turn turn turn inter- inter- inter- inter- val val val val Year (yrs) Year (yrs) Year (yrs) Year (yrs) 17525)? l 2 3 5 1786 1822 1863 7ST 3 17 ] 21 1789 1839 1864 1778 4 pi 9 I, 1793 1841(2) 1873 1779 4 ] 8 1 1797(7) 1842 1881(9) 1780(2) ] 10 4 1 1798(2) E852 (2) 1885 1781 ©) 7 ] 3 1803 1859(2) 1886 1784 17 | 24 1 1820 1860(4) 1910 1785 Return Section interval location Date (yrs) 8 20 June 2, 1972 21 June 2, 1972 21 June 2, 1972 23 July 30, 1972 23 July 30; 1972 23 July 31, 1972 5 23 August 16, 1977 S) 22 July 25, 1982 3) Zl September 1, 1987 cipitation), fires have occurred only slightly more commonly during years of below-average winter precipitation (Fig. 3b). The average chronology in- dex value for years with fire is 0.887 + 0.120, which is not significantly different from the average index value of 1.0. Eight of the ten years for which we have more than one fire scar (Table 2) or fire- suppression evidence of multiple ignitions (Table 1) had below-average winter precipitation. Similarly, all 3 years for which we have evidence of early- season fires (1859, 1860, and 1972), had below- average winter precipitation (Fig. 3b). Seasonal timing of fire risk. On average, 78% of all days with evidence of lightning occur after mid- July (Fig. 5). Thus, the natural fire regime of this lodgepole pine forest includes lightning risk that occurs primarily late in or after the growing season. DISCUSSION Fire intensity. Given that the thin bark of P. con- torta SSp. murrayana cannot protect trees from fires of moderate or high intensity (Minore 1979; Ryan and Reinhardt 1988), a fire scar on living P. con- torta ssp. murrayana indicates low intensity for that fire. By extension, the abundance of living, fire- scarred P. contorta ssp. murrayana of Mt. San Ja- cinto suggests that fires typically burned with low intensity. The exclusively nonserotinous nature of P. contorta ssp. murrayana of Mt. San Jacinto also indicates a regime of low intensity fires (Brown 1975; Lotan 1976). At 375 stems/ha, the lodgepole pine forest of Mt. San Jacinto is sparsely stocked and therefore has a highly discontinuous living fuel load. Pinus con- torta SSp. murrayana generally has an open crown Study site: R? = 87% Lodgepole pine frequency (log10 count) SHEPPARD AND LASSOIE: FIRE REGIME OF LODGEPOLE PINE 53 (b) Plots with 1752 fire dates: R? = 85% 100 10 30 50 70 90 110 130 Diameter (20 cm classes, centered around midpoint) Fic. 4. Size histograms of P. contorta ssp. murrayc and plots burned in (b) 1752, (c) 1797, or (d) 1881. form and foliage with low flammability (Minore 1979), which results in discontinuous ladder fuel loading with little potential for crown fires (Fahn- stock 1970). The average ground fuel loading of 18.9 (+2.1) tonnes/ha for the lodgepole pine forest of Mt. San Jacinto is also light, and the Model H fuel type supports fires that typically spread slowly and are not intense except possibly in areas of con- centrated downed woody material (Deeming et al. 5 ie) w & Average number of days —_ 0 Late Early Late May June June Fic. 5. Frequency of weather with possible ligh Data are summarized from July ina within (a) the entire lodgepole pine forest of Mt. San Jacinto, 1977). Furthermore, rock outcrops (common throughout the lodgepole pine forest) act as natural firebreaks that isolate live and dead fuels. This light and discontinuous fuel loading indicates a regime of low intensity fires (Philpot 1977) within the lodgepole pine forest of Mt. San Jacinto. The forest-wide structure of abundant young trees and exponentially fewer old trees (Fig. 4a) indicates that the lodgepole pine forest of Mt. San Late July Early August August Sept. forest of Mt. San Jacinto. ble lightning somewhere over the lodgepole pine daily records (1965-1974 and 1979-1984) from the Idyllwild and Long Valley stations. 54 MADRONO Jacinto is all-aged and self-perpetuating (Lorimer 1980) and that it has not experienced intense dis- turbances in the past few to several hundred years (Schmelz and Lindsey 1965; Brown 1975; Renkin and Despain 1992). These traits are similar to some Sierra Nevada subalpine forests where P. contorta ssp. murrayana forms stable populations of zonal dominants on sites beyond the ecological range of more tolerant competitors (Parker 1988) and where canopy or crown fires are virtually absent (Parker 1986). These results and interpretations are also similar to those for some lodgepole pine forests of Yellowstone National Park that have been described aS nonpyrogenous, with fire as a disturbance pro- cess of low intensity (Despain 1983). There is some structural evidence for fires of rel- atively higher intensity. Notably, P. contorta ssp. murrayana size histograms for the 1797 and 1881 fires are not shaped in a negative exponential pat- tern, indicating unusually intense disturbances. This is consistent with the fact that many forested eco- systems have regimes of typically low intensity fires with occasional fires of higher intensity (Kil- gore 1981). Fire area. The fact that all suppressed lightning fires were listed as spot (0.4 ha or smaller) suggests natural fires do not typically expand in size quickly. It can be argued, of course, that suppressed fires remained small because they were suppressed. However, remote fires in wilderness areas are in- herently difficult to access, and fire crews may have taken more than a day to reach some of these fires, especially before the availability of airborne access. That none of the listed fires grew larger than 0.4 ha before being suppressed suggests that fires typ- ically burn small areas. Our strategy for field sampling fire scars (tran- sects 200 m apart with grid points every 500 m along each transect) limited our minimum resolva- ble fire area to approximately 10 ha (Arno et al. 1993). If past fires typically burned areas exceeding 10 ha, we would expect a majority of pairs of ad- jacent sampling plots to have the same fire date. In contrast to that interpretation, most adjacent pairs of sampling plots have different fire dates, indicat- ing that these fires typically burned areas of less than 10 ha. The three areas containing several plots with the same fire dates (1752, 1797, and 1881; Fig. 1) have at least two possible interpretations. One is that sev- eral small, spatially independent fires ignited and burned during those years. The fire suppression rec- ord for 1972 (six fires independent of one another by date or location; Table 1) indicates that this sce- nario is at least possible because of frequent light- ning ignitions. A valid alternative scenario is that a single, large fire burned during those years. Given that higher intensity disturbance processes gener- ally impact larger areas (Sousa 1984) and that for- est structure results for plots burned in 1797 and [Vol. 45 1881 indicate that those fires may have been rela- tively intense, it is probable that those fires were relatively large (20 ha or larger). Frequency. We cannot precisely quantify mean fire interval, defined as the average of all fire in- tervals between successive fires of a designated area (Romme 1981), because the fire-suppression records and our fire-scar data do not indicate more than one fire date for any particular area. However, if we define the designated area as any single area 10 ha in size (our minimum resolvable fire area), and if lightning ignites fires randomly throughout the entire 1000 ha forest, then a natural fire should burn any designated area only once every 100 years. At this spatial scale, this regime qualifies as one of infrequent (mean fire interval of >25 years) fires of low intensity, similar to other subalpine for- ests of the Sierra Nevada (Kilgore and Briggs 1972; Kilgore 1981). In contrast to the mean fire interval of at least 100 years, the mean time interval between fires anywhere in the lodgepole pine forest—not neces- sarily any particular area—since the mid 18th cen- tury is approximately 4 to 7 years (5.2 + 1.1 years for the fire suppression record of 1945 to 1989, Table 1; and 5.0 + 2.2 years for the fire scar record 1752 to 1886, Table 2); thus, natural fires occur somewhere throughout the lodgepole pine forest fairly frequently. If this interval analysis is restrict- ed to include only those fires that appear to have burned atypically large areas (as was possible for 1752, 1797, and 1881), then the mean time interval between large fires somewhere throughout the lodgepole pine forest is 64 years, much longer than the 4 to 7 years between fires of any size. The fact that large fires occur infrequently conforms to the general principle that large disturbance events are relatively rare (Sousa 1984). Relationship of fire with weather. For the lodge- pole pine forest of Mt. San Jacinto, evidence that fires, especially most large fires and all early-season fires, occur more commonly during years with be- low-average winter precipitation is consistent with results from other Southwestern forest ecosystems where natural fires occur more commonly during dry years (Swetnam and Betancourt 1990). These results support the intuitive notion that moisture content of ground and surface fuels is an important factor in fire occurrence and behavior. Seasonal timing of fire. As indicated by daily weather records and by both the suppression and scar records of fire, lightning and/or fires in the lodgepole pine forests of Mt. San Jacinto common- ly occur late in or after the growing season (after mid-July). This temporal pattern of lightning risk is similar to that of the entire Peninsular Range of southern California (Tubbs 1972), but it differs from the pattern found in Southwestern forest eco- systems where lightning fires occur more common- 1998] ly during June and early July (Baisan and Swetnam 1990). MANAGEMENT IMPLICATIONS Given that lightning fires occur fairly frequently somewhere within the lodgepole pine forest of Mt. San Jacinto, State Park Wilderness managers should always expect and be prepared for fires, es- pecially during the latter half of the growing season and during summers preceded by below-average winter precipitation. However, because fires typi- cally burn small areas with low intensity, managers can expect to safely allow most lightning fires to burn naturally, as has been suggested for other for- ests (Heinselman 1970). Larger fires are possible, of course, and monitoring fires will continue to be prudent, especially for Mt. San Jacinto because of its unique natural resource of very old P. contorta ssp. murrayana. ACKNOWLEDGMENTS This project was funded in part by grants from the An- drew Mellon Foundation and the MclIntire-Stennis Pro- gram to the Department of Natural Resources, Cornell University. We acknowledge administrators and managers of the California Department of Parks and Recreation. We thank several field workers for their assistance, and we thank J. S. Dean, T. J. Fahey, L. J. Graumlich, H. D. Gris- sino-Mayer, M. P. Hamilton, R. L. Holmes, M. Keifer, G. R. McPherson, T. W. Swetnam, and peer reviewers for analytical assistance and/or comments on early drafts of this work. LITERATURE CITED AGEE, J. K. 1974. Fire management in the national parks. Western Wildlands 1:27—33. ARNO, S. E, E. D. REINHARDT, AND J. H. Scott. 1993. Forest structure and landscape patterns in the subal- pine lodgepole pine type: a procedure for quantifying past and present conditions. United States Department of Agriculture, Forest Service, Intermountain Re- search Station General Technical Report INT-294. ASCHMANN, H. 1959. The evolution of a wild landscape and its persistence in southern California. Annals of the Association of American Geographers 49:34—56. BAILEY, H. P. 1966. The Climate of Southern California. University of California Press, Berkeley, CA. BAISAN, C. H. AND T. W. SWETNAM. 1990. Fire history on a desert mountain range: Rincon Mountain Wilder- ness, AZ. Canadian Journal of Forest Research 20: 1559-1569. Barrows, D. P. 1900. The Ethno-botany of the Cahuilla Indians of Southern California. Dissertation. The Uni- versity of Chicago, IL. BEAN, L. J. AND K. S. SAUBEL. 1972. Temalpakh: Cahuilla Indian Knowledge and Usage of Plants. Malki Mu- seum Press, Morongo Indian Reservation, CA. BLONSKI, K. S. AND J. L. SCHRAMEL. 1981. Photo series for quantifying natural forest residues: southern Cas- cades and northern Sierra Nevada. United States De- partment of Agriculture, Forest Service, Pacific Southwest Experiment Station General Technical Re- port PSW-56. Brown, J. K. 1975. Fire cycles and community dynamics SHEPPARD AND LASSOIE: FIRE REGIME OF LODGEPOLE PINE ae) in lodgepole pine forests. Pp. 429-456 in D. M. Baumgartner, (ed.), Management of lodgepole pine ecosystems symposium. Washington State University Cooperative Extension Service, Pullman, WA. Cook, E. R. 1985. A time series analysis approach to tree- ring standardization. Ph.D. dissertation. University of Arizona, Tucson, AZ. DEEMING, J. E., R. E. BURGAN, AND J. D. COHEN. 1977. The National Fire-Danger Rating System—1978. United States Department of Agriculture, Forest Ser- vice, Intermountain Experiment Station General Technical Report INT-39. DESPAIN, D. G. 1983. Nonpyrogenous climax lodgepole pine communities in Yellowstone National Park. Ecology 64:231—234. Drew, L. G., (ed.). 1972. Tree-ring chronologies of west- ern america. III: California and Nevada. Chronology Series 1, Laboratory of Tree-Ring Research, Univer- sity of Arizona, Tucson, AZ. DRUCKER, P. 1937. Culture element distribution: V, south- ern California. University of California Anthropolog- ical Records 1:1—52. FAHNSTOCK, G. R. 1970. Two keys for appraising forest fire fuels. United States Department of Agriculture, Forest Service, Pacific Northwest Experiment Station Research Paper PNW-99. Fritts, H. C. 1976. Tree rings and climate. Academic Press, New York, NY. HAMILTON, M. P. 1983. A floristic basis for the manage- ment of rare plants and their communities in the San Jacinto Mountains, California. Ph.D. dissertation. Cornell University, Ithaca, NY. HANAWALT, R. B. AND R. H. WHITTAKER. 1976. Altitudi- nally coordinated patterns of soils and vegetation in the San Jacinto Mountains, CA. Soil Science 121: 114-124. HANLEY, D. P., W. C. SCHMIDT, AND G. M. BLAKE. 1975. Stand structure and successional status of two spruce- fir forests in southern Utah. United States Department of Agriculture, Forest Service, Intermountain Exper- iment Station Research Paper INT-176. HEINSELMAN, M. L. 1970. Preserving nature in forested wilderness areas and national parks. National Parks and Conservation Magazine 44(276):8—14. JAMES, H. C. 1960. The Cahuilla Indians. Westernlore Press, Riverside, CA. KILGorE, B. M. 1981. Fire in ecosystem distribution and structure: western forests and scrublands. Pp. 58—89 in Fire Regimes and Ecosystem Properties, Proceed- ings of the Conference. United States Department of Agriculture, Forest Service, Washington Office Gen- eral Technical Report WO-26. KILGorE, B. M. AND G. S. BRIGGS. 1972. Restoring fire to high elevation forests in California. Journal of For- estry 70:266—71. LorRIMER, C. G. 1980. Age structure and disturbance his- tory of a southern Appalachian virgin forest. Ecology 61:1169-1184. LORIMER, C. G. AND L. E. FRELICH. 1984. A simulation of equilibrium diameter distributions of sugar maple (Acer saccharum). Bulletin of the Torrey Botanical Club 111:193-199. Lotan, J. E. 1976. Cone serotiny-fire relationships in lodgepole pine. Pp. 267-278 in Proceedings, 14th Tall Timbers Fire Ecology Conference. Tallahassee, FL: Tall Timbers Research Station. LoTAN, J. E. AND W. B. CRITCHFIELD. 1990. Pinus contorta Dougl. ex Loud.: Lodgepole Pine. Pp. 302-315 in 56 MADRONO Silvics of North America, Vol. 1. Conifers. United States Department of Agriculture, Forest Service, Ag- riculture Handbook 654. MINorE, D. 1979. Comparative autecological characteris- tics of northwestern tree species—a literature review. United States Department of Agriculture, Forest Ser- vice, Pacific Northwest Experiment Station General Technical Report PNW-87. MITCHELL, R. G., R. E. MARTIN, AND J. STUART. 1983. Catfaces on lodgepole pine—fire scars or strip kills by the mountain pine beetle? Journal of Forestry 81: 598-601, 613. MutTcu, R. W. 1980. Workshop summary: who cares about fire history? Pp. 138-40 in M. A. Stokes and J. H. Dieterich, (eds.), Proceedings of the fire history work- shop. United States Department of Agriculture, Forest Service, Rocky Mountain Experiment Station Gen- eral Technical Report RM-81. PARKER, A. J. 1986. Persistence of lodgepole pine forests in the central Sierra Nevada. Ecology 67:1560—1567. PARKER, A. J. 1988. Stand structure in subalpine forests of Yosemite National Park, California. Forest Science 34:1047—1058. PHILPOT, C. W. 1977. Vegetative features as determinants of fire frequency and intensity. Pp. 12—16 in Proceed- ings of the symposium on the environmental conse- quences of fire and fuel Management in mediterra- nean ecosystems. United States Department of Agri- culture, Forest Service, Washington Office General Technical Report WO-3. RENKIN, R. A. AND D. G. DESPAIN. 1992. Fuel moisture, forest type, and lightning-caused fire in Yellowstone National Park. Canadian Journal of Forest Research 22:37-45. ROMME, W. 1981. Fire History Terminology: report of the [Vol. 45 ad hoc committee. Pp. 135-137 in Proceedings of the fire history workshop. United States Department of Agriculture, Forest Service, Rocky Mountain Exper- iment Station General Technical Report RM-81. RYAN, K. C. AND E. D. REINHARDT. 1988. Predicting post- fire mortality of seven western conifers. Canadian Journal of Forest Research 18:1291—1297. SAVAGE, M. AND T. W. SWETNAM. 1990. Early 19th-cen- tury fire decline following sheep pasturing in a Nav- ajo ponderosa pine forest. Ecology 71:2374—2378. SCHMELZ, D. V. AND A. A. LINDSEY. 1965. Size-class struc- ture of old-growth forests in Indiana. Forest Science 11:258-264. SHEPPARD, P. R., J. E. MEANS, AND J. P. LASSOIE. 1988. Crossdating cores as a nondestructive method for dat- ing living, scarred trees. Forest Science 34:781—789. SMITH, D. L. 1957. A Study a Human Use in the Mount San Jacinto Wild Area in the San Jacinto Mountain Range and the Surrounding San Bernardino National Forest Area. Masters’ thesis. California State College, Long Beach, CA. Sousa, W. P. 1984. The role of disturbance in natural com- munities. Annual Review of Ecology and Systematics 15:353-391. STOKES, M. A. AND T. L. SMILEY. 1968. Tree-ring Dating. University of Chicago Press, Chicago, IL. SWETNAM, T. W. AND J. L. BETANCOURT. 1990. Fire— southern oscillation relations in the southwestern United States. Science 249:1017—1020. THORNE, R. FE 1988. Montane and subalpine forests of the Transverse and Peninsular Ranges. Pp. 537—557 in M. G. Barbour and J. Major, (eds.), Terrestrial vegetation of California. California Native Plant Society, Special Publication Number 9. Tusss, A. M. 1972. Summer thunderstorms over southern California. Monthly Weather Review 100:799-807. MApDRONO, Vol. 45, No. 1, pp. 57-63, 1998 A NEW STIPA (POACEAE: STIPEAE) FROM IDAHO AND NEVADA MICHAEL CURTO Biology Department, Utah State University Logan, UT 84322-5305 DOUGLASS M. HENDERSON! Department of Biological Sciences, University of Idaho Moscow, ID 83843 ABSTRACT Stipa shoshoneana is a new grass species principally from east-central Idaho, but with a disjunct population in the Belted Range of southern Nevada. Stipa shoshoneana is allied with Eurasian species of Stipa L. sect. Lasiagrostis (Link) Hackel, and with North American stipoids historically assigned to Oryzopsis sensu amplo. Vegetative features, panicles, glumes, anthoecia, and flowers approximate Stipa canadensis Poir., but the lemma callus and awn morphology resemble Oryzopsis pungens (Torr.) A.S. Hitchc. INTRODUCTION During 1978, three grass specimens were col- lected (D. M. Henderson 4432) within the Salmon River Canyon of Lemhi County, Idaho, that possess a unique combination of micro- and macromor- phological character-states, but with clear alliance to Eurasian species of Stipa L. sect. Lasiagrostis (Link) Hackel (=Achnatherum P. Beauv. sensu stricto), and to a group of North American stipoids historically assigned to Oryzopsis sensu amplo. Vegetative features, spikelet arrangement, glumes, anthoecia (lemma and palea without callus or awn) and flowers approximate Stipa canadensis Pott. (=Oryzopsis canadensis [Poir.] Torr.), but the lem- ma callus and awn parallel Oryzopsis pungens (Torr.) A. S. Hitchce. Over 200 subsequently collected specimens of this enigmatic grass confirmed the presence of more than just a few anomalous plants. Fourteen widely separated populations are known within the Salmon River Mountains and Lemhi Range of east- central Idaho, and a small disjunct population exists in the Belted Range of southcentral Nevada. The known geographical range of this unde- scribed species occurs beyond the spatial extent of any close tribal relative. Stipa canadensis and Ory- zopsis pungens, both species of Canada and the northeastern United States, are presently unknown within Idaho, although both occur nearby in south- ern British Columbia and Alberta. Oryzopsis mi- cranthum (Trin. & Rupr.) Thurber (=Piptatherum micranthum [Trin. & Rupr.] Barkworth) is locally frequent to common throughout the Great Basin, central Rocky Mountains, and northern Great Plains, but is known within Idaho from only one site in southwestern Clark County (Moseley and Henderson 1994) ca. 30 km by air northeast across Birch Creek Valley from the closest population of ‘Deceased 1996. During 1993 and 1994, Henderson participated in the writing and editing of previous drafts. the undescribed species. Oryzopsis exigua Thurber is present at several sites across the Salmon River Mountains and Lemhi Range, but has not been found with the undescribed species. Twelve other stipoids occur within the geograph- ical extent of the undescribed species: Stipa comata Trin. & Rupr. (=Hesperostipa comata [Trin. & Rupr.] Barkworth); S. viridula Trin. (=Nassella vir- idula [Trin.] Barkworth); S. hymenoides R. & S.; S. lettermanii Vasey; S. nelsonii Scribn.; S. nevadensis B. L. Johnson; S. occidentalis Thurber; S. pineto- rum M. E. Jones; S. richardsonii Link; S. thurber- iana Piper; S. webberi (Thurber) B. L. Johnson; and Oryzopsis swallenii C. L. Hitchcock & Spellenberg. Note that the last ten species were all recombined in Achnatherum by Barkworth (1993). The unde- scribed species is, however, ecologically segregated from all twelve, as it occurs only on or at the base of near-vertical cliffs not occupied by any other sti- poid. Although these populations from east-central Idaho and southcentral Nevada constitute a previ- ously undescribed species, ready assignment to Oryzopsis is now incongruous with recent generic realignments in the American Stipeae (Barkworth 1983, 1990, 1993; Barkworth and Everett 1987) that 1) excluded Stipa L. from North and South America; 2) treated Oryzopsis as a unispecific ge- nus comprising only O. asperifolia Michx.; 3) reas- signed several North American Oryzopsis to other genera; and 4) temporarily retained S. canadensis, O. exigua, and O. pungens in Oryzopsis as an in- formally recognized group, ‘“‘Boreobtusae”’, with uncertain generic affinity. Anthoecial morphology and lemma epidermal cell patterns expressed by the ‘‘Boreobtusae”’ are traceable to Miocene stipoids of present-day Nebraska (Thomasson 1980). For the past decade, we have delayed the formal naming of this new species while these generic re- alignments have transpired. Without question, Sti- peae systematics are very complex and likely in- volve reticulate evolution among ancestral ge- 58 MADRONO [Vol. 45 nomes, as first discussed by Johnson (1945, 1972). The recognition of Oryzopsis sensu amplo has long been problematic owing to the obvious heteroge- neity within the genus, and to the lack of consistent macromorphological difference from Stipa sensu amplo (Johnson 1945, 1972; Hoover 1966; Hitch- cock and Spellenberg 1968; Spellenberg and Meh- lenbacher 1971; Maze 1972; Kam and Maze 1974; Freitag 1975, 1985; Barkworth and Everett 1987). While it is clear that the North American ‘‘Boreob- tusae”’ are only partially similar in macromorphol- ogy to O. asperifolia, an alternative generic group- ing is not readily evident, as species of ‘‘Boreob- tusae”’ are macromorphologically dissimilar. Obvious qualitative differences in floral, spikelet, and vegetative morphology exist among the four species of ‘“‘Boreobtusae,’’ including the unde- scribed species. Both O. pungens (2n=22, Johnson 1945) and O. exigua (2n=22, Hitchcock and Spel- lenberg 1968) share with O. asperifolia (2n=46, Johnson 1945; 2n=48, Bowden 1960) the combi- nation of a fused style column bearing two or three stigmata, together with obovate glumes both shorter than or equal to the lemma body apex and divari- cate at fruit dissemination. However, the florets of O. pungens and O. exigua differ markedly from each other (Table 1). Both S. canadensis and the undescribed species have free styles along with el- liptical glumes longer than the lemma body apex and non-divaricate at fruit dissemination. But again, the florets of S. canadensis and the undescribed species are otherwise dissimilar (Table 1). To better evaluate taxonomic placement of both the enigmatic “‘Boreobtusae”’ and the undescribed species, we reassessed the central argument es- poused by Barkworth and Everett (1987) for ge- neric monophyly as based on putatively autapo- morphic lemma epidermal cell patterns first dis- cussed by Thomasson (1976, 1978a). Oryzopsis exigua 3-8 mm long, once-ge- niculate, proximal seg- ment spirally contorted < 1 revolution 2-3) berg 1968) apex long 3, fused proximally Elliptic 3, papillate, exserted apical- Shorter than lemma body dorsal, somewhat persistent, 22 (Hitchcock and Spellen- penicillate, 1.5—3.0 mm 2.5 mm long, + straight, spirally contorted < 1 (Bowden 1960; Love and Love 1981) Oryzopsis pungens revolution Shorter than lemma body glabrate, 1.5—2.0 mm long 2, papillate, exserted apical- subterminal, caducous, |— 2, fused proximally 22 (Johnson 1945) 24 apex Obovate Stipa canadensis METHODS mm long, twice-genicu- late, proximal segment spirally contorted > 2 revolutions berg 1970) ally 22 (Johnson 1945; Spellen- Longer than lemma body terminal, persistent, 10—20 glabrate, 1.5—2.0 mm long 2, plumose, exserted later- 2, free throughout apex Obovate Laminar, lemmatal, and paleal abaxial epidermal patterns were observed from material prepared fol- lowing the sodium hydroxide/chlorozol black clear- ing/staining method detailed by Thomasson (1978b). Descriptions of all laminar and lemmatal preparations follow Ellis (1976, 1979). Embryos were dissected from mature fruits first immersed in boiling water removed from a hot plate and then left to cool overnight. Mitotic chromosome counts were made upon root-tips removed from fruits ger- minated on filter paper in petri dishes, pretreated in distilled water vials kept on ice for 48 hours, and subsequently fixed/stained in aceto-orcein. Meiotic counts were obtained from anthers fixed in 3:1 eth- anol: acetic acid and stained in aceto-carmine. Ob- servations and drawings were made through a Zeiss Standard 18 microscope with a camera-lucida at- tachment. Voucher specimens for chromosome counts are deposited at CAS. Stipa shoshoneana 2.5 mm long, + straight spirally contorted < | revolution long 2, free throughout ally 2, plumose, exserted later- 20 (This paper) Longer than lemma body subterminal, caducous, 1— penicillate, 1.5—2.2 mm SPECIES OF ‘‘BOREOBTUSAE’’. apex Obovate TABLE 1. Proximal glume length Anthoecium pro- file in fruit Lemma awn Lodicules Anthers Stigmata Styles 2n = 1998] For our morphometric comparisons among “‘Bor- eobtusae’”’ species, we examined and measured 204 specimens of the undescribed species, 64 speci- mens of O. exigua at ID, IDS, and UTC, and 470 specimens of other ““Boreobtusae”’ (76 S. canaden- sis, 157 O. micrantha, and 237 O. pungens), bor- rowed from CAN, MIN, and RM. RESULTS AND DISCUSSION A discussion of generic relationships and rea- lignments will be presented in another paper. In brief, we conclude that the undescribed species and S. canadensis are probably best placed with species of Stipa L. sect. Lasiagrostis (Link) Hackel (=Achnatherum P. Beauv. sensu stricto). As for the placement of O. exigua and O. pungens, we remain at an impasse. If both are transferred to Stipa or to Achnatherum, then either genus would incorporate species combining fused styles and short glumes, making the exclusion of O. asperifolia, the gener- itype of Oryzopsis, even less-tenable. Perhaps both O. exigua and O. pungens should remain in Ory- zopsis despite their incongruities otherwise with O. asperifolia. Achnatherum sensu stricto, the greatly enlarged and heterogeneous Achnatherum sensu Barkworth (1993), and the newly recognized Australian seg- regate genus Austrostipa S.W.L. Jacobs & J. Everett (1996), form a macromorphological continuum with Stipa; and are not globally circumscribable Linnean genera with any greater coherence or pre- dictive utility than Stipa sensu amplo. Although others may emphasize differences among modes of variation by segregating smaller genera, we choose to recognize the continuum among these modes by using subgenera within Stipa sensu amplo, as did Freitag (1975), Clayton and Renvoize (1986), and recently Vazquez and Devesa (1996). We anticipate that others will likely recombine this new species in segregate genera. Thus, we have selected a specific epithet that presents no nomen- clatural barrier to direct transfers. Table 1, and a phenetic key, enable discrimination of the new spe- cies from similar regional stipoids. Stipa shoshoneana Curto & D.M. Henderson, sp. nov. (Fig. 1). —TYPE: USA, Idaho, Salmon River Mts, ca. 15 km N of Challis, Morgan Creek Canyon ca. 7 km NW of US Hwy 93, near 44°39'47"N, 114°13'19"W, Gooseberry Creek 7.5 min quadrangle, TISN RI9E S4 SW% of NE’, el. ca. 1675 m, aspect SW, along N side of road in cracks of near vertical cliffs with Cercocarpus ledifolius Nutt., Heu- chera grossulariifolia Rydb., Elymus spicatus (Pursh) Gould, and Poa interior Rydb., 30 June 1987, L. Eno 17 (holotype: CAS; iso- types: BRY, ID, K, MIN, MO, NY, RM, UC, US, UTC, WTU; all to be distributed). Stipa canadensis Poir. affinis, cujus habitum, an- CURTO AND HENDERSON: A NEW STIPA ape) thoecia, et flores habet, sed differt callis lemmatum brevibus, aristis curtis caducis, antheris penicillatis, et chromosomatum numero aequante 20. Plants rhizocarpic, iteroparous perennials. Culms herbaceous, 20—50 cm tall, densely tufted, slender, geniculate or ascending to erect; unbranched dis- tally; nodes 2—3, glabrate, internodes hollow, an- trorsely scabridulous; innovations intravaginal. Leaves mostly basal, few cauline; vernation con- volute. Sheaths exauriculate; margins free; cross- section rounded; transverse septae absent; abaxial surface scabridulous. Ligules adaxial; membranous throughout; 1.8—5.5 mm long, apex acute, often lac- erate. Lamina narrowly elongate, the length: width ratio > 30:1; planar or involute with drying, stiff, scabridulous along veins and margins. Synfloresc- ences terminal, ebracteate or with a solitary linear bract at the proximal nodes; paniculate, ultimately diffuse, rachis and branches persistent; rachis nodes 3-7; rachis 33—220 mm long; most-proximal rachis internode 0.5—63.0 mm long; branches persistent, terminating at spikelets, slender to subcapillary, ir- regularly quadrangular, scabridulous; primary branches 1—2(—4) per node, most-proximal primary branches 18—116 mm long, distance to initial sec- ondary branch 4—76 mm long; reflexed 90°—270°, bearing axillary pulvini; secondary and tertiary branching strictly dichotomous, branches bearing axillary pulvini and divaricate pre- and post-anthe- sis; penultimate and ultimate branches adpressed. Spikelets borne as distinctly pedicellate monads, adpressed in fruit; florets and flowers 1, bisexual; 3.3—5.3 mm long excluding lemma awn; pre-anthe- sis anthoecial profile elliptical, obovate in fruit; pre- anthesis anthoecial compression subterete, some- what dorsoventrally compressed in fruit; rachilla terminating at floret attachment; disarticulation dis- tal to the glumes. Glumes two, persistent, size and shape subequal, both extending beyond lemma body apex, rounded abaxially, membranous to char- taceous, evident veins 1—9, green and purple prox- imally, colorless distally, glabrate or with midvein scabridulous distally; proximal (first or lower) glume 3.2—5.1 mm long, profile asymmetrically lanceolate or oblongly lanceolate, apex acute to acuminate, awnless; distal (second or upper) glume 3.3—-5.3 mm long, profile asymmetrically ovate, apex acute to acuminate, awnless. Lemma Callus obconic relative to pedicel, =0.3 mm long abaxi- ally, blunt, glabrate, articulation scar round, slightly excavated, peripheral ring raised. Lemma Body 2.2— 3.8 mm long; profile broadly elliptical; margins symmetrically involute, juxtaposed parallely with spikelet axis prior to anthesis, gaping in fruit; apex emarginate about excurrent midvein; rounded abax- ially; germination flap absent; texture coriaceous at anthesis; veins 5—7; evenly antrorsely hirtellous throughout, trichomes simple, <0.5 mm long, col- orless initially, aging tawny. Lemma Awn terminal, unbranched, straight or slightly arcuate, antrorsely scabridulous, 1.0—2.5 mm long; caducous. Palea 60 MADRONO [Vol. 45 1 ents fF ‘» ro . 149 Valley Co. 50 km —_—_—_—— 1998] 2.1-3.6 mm long; length, texture and vestiture sub- equal to lemma; profile broadly elliptical; margins planar; apex bifid, minutely biaristate; rounded abaxially; veins 2(3); disarticulating with respective lemma; abaxial epidermal pattern similar to lemma pattern. Lodicules three, free, adaxial pair obovate, 0.75—1.25 mm long, abaxial linear, 0.5—-1.0 mm long. Stamens three; filaments free, evanescent to marcescent; anthers free, penicillate, 1.75—2.2 mm long, yellow. Ovularium obovate, glabrate; styles two, subterminal, free, short; stigmata two, exserted laterally, white. Caryopsis obovoid, compression dorsoventral, length ca. 2 mm long; enclosed with- in, but free from anthoecium; exocarp smooth and glossy; hilum linear, ca. % caryopsis length; endo- sperm solid; embryo F+FE % to % caryopsis length. Seedling mesocotyl lengthy; first leaf lamina narrow, erect, 7- to 15-veined. Chromosomes rela- tively small, 2-4 pw long, n=10, 2n=20 (fruits from Eno 17, CAS), 2n=20 (fruits from Eno 18, CAS). Lamina Abaxial Epidermis without microhairs or papillae; costal/intercostal zonation conspicuous; costal regions: short-cells solitary, paired or in short rows, silica bodies round, square, horizontally nodular-elongate, irregulary dumbbell- or saddle- shaped, hooks infrequent, central, medium, prickles infrequent, single file, small to medium, barb point- ing toward blade apex; intercostal regions: long- cells elongated, mostly 75—150 w long, walls mod- erately thickened, side-walls parallel, undulations moderate, U-shaped, end-walls angled or interlock- ing, distributed in alternating long-cell/short-cell files with occasional short-cell pairs, intercostal short-cells square, rectangular, or irregular, silica body shape similar to cell shape, stomata low- dome-shaped, arranged in single or double rows along costal zones, one or two interstomatal long- cells between successive stomata, these occasion- ally separated by square to rectangular short-cells; transverse section exhibiting prominent adaxial ribs, midrib generally indistinguishable from oth- ers; sclerenchyma abundant interior to both epi- derms, forming ab- and adaxial vascular bundle girders. Lemma Abaxial Epidermis achnatheroid, costal/ intercostal zonation absent; microhairs absent; pa- pillae absent; stomata absent; long-cells ca. 10—30 w long, walls moderately thickened, side-walls ir- regularly undulating, end-walls irregular, arranged in files as long-cell/short-cell/long-cell, long-cell/ short-cell/suberin-cell, or long-cell/prickle/long- cell; silica bodies horizontally oblong or squarish, ca. 5—10 p long; silico-suberose couples occasional, suberin-cells crescentic, ca. 5 w long; prickles 20— 25 w long, barb short. <_ Fic. 1. Stipa shoshoneana. a) habit; b) spikelet; c) floret abaxial view; g) distribution of largest Idaho populations. CURTO AND HENDERSON: A NEW STIPA 61 Paratypes. USA, Idaho: Butte Co., Lemhi Range, ca. 35 km (air) NNW of Howe, Bunting Canyon above Badger Mine, 44°06’21”N, 113°07'48"W, TON R28E S16 SW%, 2255 m, 12 July 1979, S. & P. Brunsfeld 1132 (1D); loc. cit., 16 June 1981, J. Civille 251d (ID); loc. cit., 13 June 1987, L. Eno 6 (ID); loc. cit., 26 June 1987, L. Eno 11 (ID); Lemhi Range, ca. 12 km (air) NNE of Howe, Middle Can- yon, 43°53'30"N, 112°57'29"W, T7N R29E S26/35, 1950-2255 m, 17 June 1978, D. M. Henderson 4629, S. & P. Brunsfeld (ID, UTC); loc. cit., 16 June 1981, J. Civille 260 (ID); loc. cit., 10 June 1987, L. Eno 4, ID); loc. cit., 25 June 1987, L. Eno 9 (ID); Custer Co., Salmon River Range, Loon Creek ca. 500 m N of Bennett Creek Bridge, 44°47'34"N, 114°48'02”W, TI7N RI4E S22 SE%, 1400 m, 18 May 1988, L. Eno 25 (ID); ca. 4 km (air) SE of Cougar Creek Ranch along Hood Creek, 44°43'38"N, 114°52'38"W, T16N R13E S16 NE, 2000 m, 20 June 1982, J. Civille 309 (ID); Loon Creek ca. 1.5 km NE of Tin Cup Campground, 44°36'41’"N, 114°47'50”W, TIS5N R14E S26 NW%, 1700 m, 2 July 1987, L. Eno 18 (ID); ca. 14 km (air) N of Challis, Morgan Creek Road ca. 5 km NW of US Hwy 93, 44°38’39"N, 114°12'32’W, TISN RISE S10 NE“% SW%, 1600 m, 14 June 1987, L. Eno 7 (ID); Lemhi Co., Salmon River Range, Middle Fork Salmon River Canyon ca. 4.5 km N of Bernard Creek Guard Station, ca. 800 m S of Jack Creek confluence, 45°00'22’N, 114°43'14”"W, TION R14E S11 NW%, 1150 m, 19 July 1982, J. Civille 335 (ID); ca. 55 km (air) NW of Challis, Middle Fork Salmon River Canyon at Camas Creek confluence, 44°53'31"N, 114°43'25”W, TI8N RISE S16 SW%, 1160 m, 15 July 1982, J. Civille 332 (ID); ca. 55 km (air) NW of Challis, Middle Fork Salmon River watershed, Camas Creek ca. 400 m E of Macarte Creek Camp, 44°53'03’N, 114°41'36"W, T18N RISE S22, 1200 m, 10 June 1982, J. Civille 289 (ID); loc. cit., 19 May 1988, L. Eno 26 (ID); ca. 55 km (air) NW of Challis, Middle Fork Salmon River watershed, Camas Creek between Dry Gulch and Forage Creek, 44°53’30"N, 114°34'57’W, T18N R16E S15 SW%, 1700 m, 7 July 1981, D. M. Henderson 5990 (ID); ca. 13 km WNW of US Hwy 93, jct Salmon NF Roads 045 (Iron Creek Rd) & 088 [sic, 0467], 44°55'20"N, 114°06'43”’W, T18N R20E S4 NW%, on quartzite cliffs, 3 July 1987, P. M. Peterson 4764 & C. R. Annable (ID, UTC); Salmon River Canyon, ca. 55 km N of Challis, US Hwy 93 ca. 500 m N of mile 279, 400 m SE of highway op- posite Iron Creek, 44°53'00"N, 113°57'56"W, T18N R21E S15 SE% SW%, 1400 m, 14 June 1978, D. M. Henderson 4432 (ID); loc. cit., 12 July 1981, J. ; d) lemma abaxial epidermis; e) lemma callus; f) caryopsis, 62 MADRONO Civille 276 (ID); loc. cit., 21 June 1987, L. Eno 8 (ID); loc. cit., 28 June 1987, L. Eno 13 (ID); Salm- on River Range, Middle Fork Salmon River water- shed, Loon Creek ca. 3 km NW of Falconberry Guard Station, between Mearney Creek and Burn Creek, 44°42'07"N, 114°46'41”"W, T16N R1I4E S24 SW% NW%, 1500 m, 6 July 1982, J. Civille 330 (ID); Valley Co., Middle Fork Salmon River, W side ca. 1.55 km N of Golden Creek, Tombstone Rock, 45°09'44"N, 114°43’26”"W, T21N R14E S15 SE% NE’, 1025 m, 20 May 1988, L. Eno 27 (ID); Nevada: Nye Co., Belted Range, N of Cliff Spring, 37°30'45’"N, 116°05'15”W, T5S R53E S8, 2170 m, infrequent at cliff base, 18 June 1995, F. J. Smith 3936 & J. Heers (UNLV). Distribution. Stipa shoshoneana is known prin- cipally from canyons of the Middle Fork of the Salmon River and from its eastern tributaries, Cam- as and Loon Creeks, extending ca. 160 km by air southeast to the southern Lemhi Range (Fig. 1). Sti- pa shoshoneana is also curiously disjunct near Cliff Spring in the Belted Range of south-central Nevada about 750 km by air southwest of the most southern Idaho population. This disjunction suggests possi- ble presence in the intercalary ranges of eastern Ne- vada. Searches by Curto at some potential sites in the Jarbidge, Independence, Ruby, Schell Creek, Snake, White Pine, and Quinn Canyon Ranges of eastern Nevada found populations of O. exigua or O. micrantha, but no S. shoshoneana populations. Habitat. Stipa shoshoneana is nearly always found within moist crevices of intrusive or extru- sive igneous, metamorphic, or sedimentary cliffs and rock walls. Typical associate species include: Heuchera_ grossulariifolia Rydb., Ribes cereum Dougl., Potentilla glandulosa Lindl., Elymus spi- catus (Pursh) Gould, Poa interior Rydb., and Poa secunda K. B. Presl, with other taxa, such as Pseu- dotsuga menziesii (Mirb.) Franco, Cercocarpus led- ifolius Nutt., Artemisia tridentata Nutt., Amelan- chier alnifolia Nutt., Glossopetalon spinescens A. Gray, Mimulus cusickii (Greene) Piper, Petrophyton caespitosum (Nutt.) Rydb., and the east-central Ida- ho endemics, Astragalus amnis-amissi Barneby, Cryptantha salmonensis (Nels. & Macbr.) Pays., or Draba oreibata Macbr. & Pays., being locally com- mon at some sites. Chromosome number significance. Stipa shosh- oneana plants possess the fewest chromosomes (2n=20) of all North American Stipeae counted to date, and the second-lowest somatic number ever reported for the tribe; Prokudin et al. (1977) re- ported 2n=18 for S. bromoides (L.) Doerfler (as Achnatherum bromoides [L.] Nevski). Chapanov and Yurtsev (1976) reported 2n=20 for the Asian species Piptatherum vicarium (Grig.) Roshev., al- though all other reports indicate 2n=24 for this spe- cies, the common number of Piptatherum sect. Pip- tatherum. [Vol. 45 Epithet etymology. The specific epithet refers to the Northern and Western Shoshone people whose ancestral lands encompass the entire known distri- bution of this species. PHENETIC KEY 1 Styles fused proximally, persisting as a centric “‘beak”’ upon caryopsis; proximal glume length shorter than or equal to anthoecium length. 2 Lemma awn 1-2.5 mm long, exserted subapically, straight or weakly arcuate, often absent on herbar- ium specimens; anthers glabrate or rarely penicillate distally; stigmata 2 ........ Oryzopsis pungens 2’ Lemma awn 3-8 mm long, exserted abaxially, re- curved or geniculate at midlength, some usually present on herbarium specimens; anthers penicillate distally; stigmata 3 Oryzopsis exigua 1’ Styles free throughout, persisting as lateral ‘‘horns”’ upon caryopsis, or with no visible persistence; proxi- mal glume length longer than anthoecium length. 3 Anthoecium averaging <3 mm long, glabrate or sparsely antrorsely puberulent, trichomes adpressed Oryzopsis micrantha 3’ Anthoecium averaging > 3 mm long, evidently evenly antrorsely hirtellous to hirsute, trichomes de- flexed. 4 Ligule <1.0 mm long, longer laterally than me- dially. 5 Palea = % lemma body length; lemma awn <7 mm long, contorted < 1 revolution, weakly once-geniculate, caducous; primary panicle branches short, erectly adpressed to rachis at maturity .... Oryzopsis swallenii 5’ Palea =% lemma body length; lemma awn 18-30 mm long, contorted > 1 revolution proximally, twice-geniculate, persistent; pri- mary panicle branches elongate, deflexed from rachis at maturity inns pect idae, be orate ene Stipa richardsonii 4’ Ligule = 1.5 mm long, acute to attenuate. 6 Lemma awn persistent until fruit maturity or thereafter, 8-20 mm long, stout, once- or twice-geniculate, proximal segment distinctly spirally contorted; anthers glabrate ..... eee Oe nee ae Stipa canadensis 6’ Lemma awn caducous, 1—2.5 mm long, straight or weakly arcuate; anthers penicillate Stipa shoshoneana 2 6 © © © © © © © © 8 ee ee le ll ACCKNOWLEDGMENTS Special thanks to LeAnn Eno and Janey Civille for the arduous fieldwork effected in collecting the bulk of the Stipa shoshoneana specimens. All fieldwork by Civille, Curto, Eno, and Henderson was funded through the C. R. Stillinger Trust, University of Idaho. Anita Cholewa pro- vided some additional field collections and herbarium sup- port. Mary Barkworth allowed access to her lab and to all of the histological preparations made for her previous analyses. Dave Keil corrected our Latin diagnosis and as- sisted with our chromosome counts. Dieter Wilken pro- vided a much appreciated thorough review that improved this paper. Rhonda Riggins offered helpful comments about an earlier draft. Marianne Filbert prepared the fine illustrations of S. shoshoneana macromorphology. Dave Fross and staff of Native Sons Nursery propagated plants 1998] from seed. Lastly, thanks to the herbaria cited for their provision of loaned specimens. LITERATURE CITED BARKWORTH, M. E. 1983. Ptilagrostis in North America and its relationship to other Stipeae (Gramineae). Sys- tematic Botany 8:395-—419. . 1990. Nassella (Gramineae: Stipeae): revised in- terpretation and nomenclatural changes. Taxon 39: 597-614. . 1993. North American Stipeae (Gramineae): tax- onomic changes and other comments. Phytologia 74: 1-25. AND J. EVERETT. 1987. Evolution in the Stipeae: Identification and relationships of its monophyletic taxa. Pp. 251-264 in T. R. Soderstrom, K. W. Hilu, C. S. Campbell, and M. E. Barkworth (eds.), Grass Systematics and Evolution. Smithsonian Institution Press, Washington D.C. CHAPANOV, P. AND B. N. YURTSEV. 1976. Chromosome numbers of some grasses of Turmenia. I. Botanices- kij Zurnal SSSR 61:1240-1244. CLAYTON, W. D. AND S. A. RENVOIZE. 1986. Genera Gra- minum: Grasses of the World. Kew Bulletin Addi- tional Series 13. Her Majesty’s Stationery Office, London, U.K. ELLis, R. P. 1976. A procedure for standardizing compar- ative leaf anatomy in the Poaceae. I. The leaf-blade as viewed in transverse section. Bothalia 12:65—109. . 1979. A procedure for standardizing comparative leaf anatomy in the Poaceae. II. The epidermis as seen in surface view. Bothalia 12:641—671. FREITAG, H. 1975. The genus Piptatherum (Gramineae) in southwest Asia. Notes From the Royal Botanic Gar- den, Edinburgh 33:341—408. . 1985. The genus Stipa (Gramineae) in southwest and south Asia. Notes From the Royal Botanic Gar- den, Edinburgh 42:335-489. HiTcucock, C. L. AND R. W. SPELLENBERG. 1968. A new Oryzopsis from Idaho. Brittonia 20:162—165. Hoover, R. FE 1966. Notes on some California plants. Leaf- lets of Western Botany 10:340. CURTO AND HENDERSON: A NEW STIPA 63 JACOBS, S. W. L. AND J. EVERETT. 1996. Austrostipa, a new genus, and new names for Australasian species for- merly included in Stipa (Gramineae). Telopea 6:579-— 595: JOHNSON, B. L. 1945. Cyto-taxonomic studies in Oryzop- sis. Botanical Gazette 107:1-—32. . 1972. Polyploidy as a factor in the evolution and distribution of grasses. Pp. 18-35 in V. B. Younger and C. M. McKell (eds.), The Biology and Utilization of Grasses. Academic Press. New York, NY. KAM, Y. K. AND J. MAZE. 1974. Studies on the relation- ships and evolution of supraspecific taxa utilizing de- velopmental data. II. Relationships and evolution of Oryzopsis hymenoides, O. virescens, O. kingii, O. mi- crantha, and O. asperifolia. Botanical Gazette 135: 227-247. LovE, A. AND D. Love. 1981. Chromosome number re- ports LXXI. Poaceae. Taxon 30:510. MAZE, J. 1972. Notes on the awn anatomy of Stipa and Oryzopsis (Gramineae). Syesis 5:169-171. MOSELEY, R. K. AND D. M. HENDERSON. 1994. Noteworthy Collection of Piptatherum micranthum (Trin. & Rupr.) Barkworth (Poaceae). Madrono 41(2):149. PROKUDIN, Y. N., A. G. Vovk, O. A. PETROVA, E. D. ER- MOLENKO, AND Y. V. VERNICHENKO. 1977. Zlaki Ukrai- ny. Kiev. SPELLENBERG, R. 1970. IOPB Chromosome number re- ports XXV. Gramineae. Taxon 19:112. THOMASSON, J. R. 1976. Tertiary grasses and other Angio- sperms from Kansas, Nebraska, and Colorado. Ph.D. dissertation, Iowa State University. . 1978a. Epidermal patterns of the lemma in some fossil and living grasses and their phylogenetic sig- nificance. Science 199:975—977. . 1978b. Clearing, cuticle removal, and staining for the fertile bracts (lemmas and paleas) of grass an- thoecia. Stain Technology 53:233-—236. . 1980. Paleoeriocoma (Gramineae, Stipeae) from the Miocene of Nebraska: taxonomic and phyloge- netic significance. Systematic Botany 5:233-240. VAZQUEZ, E M. AND J. A. DEVESA. 1996. Revision del género Stipa L. y Nassella Desv. (Poaceae) en la Pen- insula Ibérica e Islas Balearas. Acta Botanica Mala- citana 21:125—189. MADRONO, Vol. 45, No. 1, pp. 64-74, 1998 SYSTEMATIC STUDIES AND CONSERVATION STATUS OF CLAYTONIA LANCEOLATA VAR. FLAVA (PORTULACACEAE) J. STEPHEN SHELLY ' Montana Natural Heritage Program, State Library, 1515 East 6th Avenue, Helena, MT 59620 PETER LESICA Division of Biological Sciences, University of Montana, Missoula, MT 59812 PAUL G. WOLF Department of Biology, Utah State University, Logan, UT 84321 PAMELA S. SOLTIS AND DOUGLAS E. SOLTIS Department of Botany, Washington State University, Pullman, WA 99164 ABSTRACT A biosystematic study of Claytonia lanceolata and related taxa in the Rocky Mountains was undertaken to evaluate the taxonomic status of C. lanceolata var. flava. This study was part of a broader assessment to determine the need for protection of the latter taxon under the federal Endangered Species Act. Elec- trophoretic and morphological studies revealed that C. lanceolata var. flava in southwestern Montana and northwestern Wyoming represents a distinct diploid species (n=8) whose populations consist of yellow- and/or white-flowered plants. Morphological, allozyme, and cytological data all indicate that this taxon does not belong in the C. lanceolata complex, but is best placed in the group of narrow-leaved species that includes C. rosea, C. tuberosa, and C. virginica. Numerous populations of C. lanceolata var. flava, most often consisting of the white-flowered phenotype, were found in Montana and Wyoming, and legal protection is not warranted at this time. In some cases, actions to conserve endangered plant taxa must be preceded by an evaluation of their taxonomic status; this study illustrates the utility of biosystematic techniques in conducting such evaluations. INTRODUCTION A need for accurate taxonomic evaluations of rare plant species has frequently arisen as conser- vation of biological diversity has become a priority on the part of government agencies and private or- ganizations. Such evaluations are critical to ensur- ing that the limited funding available for plant con- servation is devoted to taxa that are deserving from a biosystematic perspective. Claytonia lanceolata Pursh (Portulacaceae) is a common, wide-ranging species of western North America (Hitchcock et al. 1964). Claytonia lanceo- lata var. flava (A. Nels.) C. L. Hitchce. has been applied to yellow-flowered populations in the northern Rocky Mountains (Hitchcock et al. 1964; Davis 1966). The type collection of this variant was made in 1899 by Aven and Elias Nelson (5488, RM), near the northwest corner of Henry’s Lake in Fremont County, Idaho (Nelson 1900). From 1911 to 1988, it was collected at five additional stations in southwestern Montana (Shelly 1989) and one station in northwestern Wyoming (Marriott 1986). It was rediscovered at the type locality in 1986 (D. Atwood personal communication). The infrequency ' Present address: U.S.D.A. Forest Service, Watershed, Wildlife, Fisheries and Rare Plants Unit, PO. Box 7669, Missoula, MT 59807. of collection and the relatively restricted geograph- ic range of these yellow-flowered populations led to the designation of C. lanceolata var. flava as a candidate for listing under the federal Endangered Species Act (U.S. Fish and Wildlife Service 1985, 1993). The taxon was initially described as C. aurea (Nelson 1900). Rydberg (1922) reduced this name to a synonym of C. chrysantha Greene (=C. lan- ceolata var. chrysantha (Greene) C. L. Hitchc., a yellow-flowered form of the latter species occurring in western Washington (Douglas and Taylor 1972)), undoubtedly based on the shared flower color. Claytonia aurea was later renamed C. flava, the former name having already been used by Kuntze in 1891 (Nelson 1926). Rydberg (1932) also sub- sequently recognized it as C. flava. Since that time, C. flava has been reduced to a variety of C. lan- ceolata on two separate occasions (Hitchcock et al. 1964; Davis 1966). The latter revision was perhaps an oversight of the Hitchcock treatment, and Davis has occasionally been cited as the author of this change. Boivin (1968) placed C. flava as a variety of C. caroliniana Michx. More recently, the taxon has again been treated as a species (Dorn 1984). We conducted two studies to evaluate the need for listing of C. lanceolata var. flava under the fed- eral Endangered Species Act. In the first study, we 1998] used isozyme electrophoresis and field morpholog- ical analyses to compare C. lanceolata var. flava with sympatric populations of C. lanceolata var. lanceolata. The purpose of this study was to assess the current taxonomic treatment of the yellow-flow- ered populations as a variety of the latter, common taxon. During initial field work, surveys of known Montana populations of C. lanceolata var. flava re- vealed the presence of narrow-leaved, white-flow- ered plants that were morphologically very similar to the yellow-flowered individuals. These white- flowered plants did not fit the descriptions of typi- cally broader-leaved C. lanceolata var. lanceolata. Thus, we also examined the degree of isozyme dif- ferentiation between white- and yellow-flowered in- dividuals of these narrow-leaved plants, and wheth- er any other morphological differences aside from petal color exist between them. Yellow- and white- flowered individuals of these narrow-leaved plants are biotically sympatric, occurring in intermixed populations, in four of the five study locations. Fur- thermore, these narrow-leaved plants are either biotically or neighboringly sympatric (occurring in closely adjacent but non-overlapping populations) with C. lanceolata var. lanceolata in all five study locations. In the second study, we undertook herbarium morphological analyses in an initial attempt to place the narrow-leaved Claytonia populations of the northern Rocky Mountains in a broader context with respect to other congeneric taxa. In addition to C. lanceolata vars. flava and lanceolata, other taxa included in this herbarium study were C. lan- ceolata var. chrysantha (Greene) C. L. Hitchc., C. lanceolata var. multiscapa (Rydb.) C. L. Hitchc., and C. rosea Rydb. This second study did not in- clude electrophoretic analyses, as it was intended to be a preliminary assessment of the wider affin- ities of C. lanceolata var. flava within the narrow- leaved Claytonias. The taxonomy of Claytonia is currently being re- vised for the Flora of North America project (Miller and Chambers in mss.). Pending publication of this treatment, throughout this paper the name C. l/an- ceolata var. flava will refer to populations of both the white and yellow flower color phenotypes of the narrow-leaved taxon, except when citing pre- vious alternative treatments. MATERIALS AND METHODS Five populations of C. lanceolata var. lanceolata and seven of var. flava (four consisting of plants with both yellow and white flowers, two including only white-flowered plants, and one consisting of only yellow-flowered plants) were sampled for morphological and isozyme electrophoretic studies. All five populations of var. lanceolata were includ- ed in both studies. For var. flava, five of the seven populations were included in both studies; the two exceptions were the Boulder and Burton Park pop- SHELLY ET AL.: CLAYTONIA SYSTEMATICS 65 ulations (consisting of only the white-flowered phe- notype in both cases), which we were unable to include in the electrophoretic analysis. The study populations are located in southwestern Montana and northwestern Wyoming (Fig. 1, Table 1). Morphological studies. Morphological studies were conducted with live plants in the field and with herbarium specimens. We emphasized char- acters that are easily examined on living plants and pressed specimens, and that have been used in past keys treating some or all of the taxa of interest. In the field, morphological data were collected from 720 living plants, representing five popula- tions of C. lanceolata var. lanceolata and seven populations of var. flava (four including both yel- low- and white-flowered plants, two with white- flowered plants only, and one with yellow-flowered plants only). In each population (and for each color phenotype in the mixed populations of var. flava), 45 plants were examined for the following char- acters: stem height, leaf length and width, petal length and width, and sepal length. Stem height was measured in centimeters, from ground level to the point of attachment of the uppermost pedicel; all other lengths were measured in millimeters. For statistical analyses, length/width ratios of the leaves and petals were also calculated. One hundred eighty-four herbarium collections, representing C. lanceolata vars. lanceolata, flava, multiscapa, and chrysantha, as well as C. rosea, were examined from the following herbaria: MON- TU, OSC, RM, UA, UAL, WS, WTU. In addition to the characters listed above, the petal/sepal length was calculated, and petal apex outline and cauline leaf venation were scored for the herbarium speci- mens (see Table 4 for scoring criteria). Isozyme Electrophoresis. A total of 679 individ- uals from 10 populations (five of C. lanceolata var. lanceolata, four of var. flava consisting of both the yellow- and white-flowered phenotypes, and one strictly yellow-flowered population of the latter) was sampled. Both color phenotypes were sampled in the mixed populations of var. flava. Whole flow- ering stems, including the cauline leaves, were col- lected by clipping the plants at ground level. These were kept chilled in the field for one to several days until placement in ultracold storage (— 80°C). Leaves were ground immediately upon removal from the ultracold freezer, in the Tris HCI-PVP crushing buffer of Soltis et al. (1983) with 6% PVP. Nineteen putative loci, coding for twelve enzymes, were resolved using three electrophoretic buffers. A morpholine buffer, pH 6.4 (Odrzykoski and Gott- lieb 1984) was used to resolve glyceraldehyde-3- phosphate dehydrogenase (G3PDH), malate dehy- drogenase (MDH), and _ phosphoglucomutase (PGM). Buffer 8 of Soltis et al. (1983), as modified by Haufler (1985), was used to resolve alcohol de- hydrogenase (ADH), aspartate aminotransferase (AAT), leucine aminopeptidase (LAP), phospho- 66 MADRONO 0 20 40 60 80 100 Kilometers aS NN “7 45°N \ Montana y \ Idaho wN Type Locality @ Study Sites Fic. 1. © Other Occurrences [Vol. 45 Wyoming Distribution of Claytonia lanceolata var. flava in Montana, Idaho and Wyoming, and locations of study populations (A = Anaconda; B = Boulder; BP = Burton Park; CP = Champion Pass; HL = Hebgen Lake; VP = Vipond Park; W = Wyoming; YNP = Yellowstone National Park). Open circles indicate additional occurrences of yellow- and/or white-flowered populations, as recorded by the Montana Natural Heritage Program; the type locality, in Idaho, has only been observed to contain yellow-flowered individuals. glucoisomerase (PGI), and triosephosphate isom- erase (TPI). Buffer 11 of Soltis et al. (1983) was used to resolve isocitrate dehydrogenase (IDH), menadione reductase (MNR), shikimate dehydro- genase (SkDH), and 6-phosphogluconate dehydro- genase (6PGD). The stain recipe for ADH was that described by Wendel and Weeden (1989). All other staining protocols were those of Soltis et al. (1983). Data analysis. We assessed the morphological distinctiveness of the taxa using principal compo- nents analysis (PCA) and discriminant analysis of the characters listed above. These analyses were performed using SYSTAT (Wilkinson 1986), and were based on log-transformed values for the char- acters listed above. Electrophoretic data were analyzed using the computer program BIOSYS-1 (Swofford and Se- lander 1981). Two separate analyses were per- formed: 1) allele frequencies at 19 loci were en- tered for all ten populations (14 samples total, since both color phenotypes were included from the four mixed populations of var. flava) and analyzed for genetic variability statistics and Nei’s genetic iden- tity between populations of C. lanceolata vars. lan- ceolata and flava, and between the nine populations of var. flava (five yellow- and four white-flowered); and 2) eight C. lanceolata var. flava populations, from all localities except Hebgen Lake (yellow- flowered phenotype only), were entered as geno- type numbers and analyzed for population substruc- turing, to examine differences between color phe- notypes within localities. An unweighted pair group method (UPGMA) was used for cluster analysis of Nei’s genetic identity relationships. RESULTS Field studies. Taxon means, ranges, and standard deviations for the eight quantitative characters mea- sured on living plants are given in Table 2. 1998] TABLE 1. SHELLY ET AL.: CLAYTONIA SYSTEMATICS 67 POPULATIONS OF CLAYTONIA LANCEOLATA VARS. FLAVA AND LANCEOLATA ANALYZED IN ISOZYME AND FIELD MORPHOLOGICAL STUDIES. Flower color phenotypes of var. flava were sampled as separate “‘populations”’ where they are biotically sympatric (Anaconda, Champion Pass, Vipond Park, and Wyoming). Vouchers are deposited at MONTU; * = duplicates deposited at OSC. + = the Boulder and Burton Park populations of white-flowered flava were not included in the electrophoretic study. Taxon Abbreviation Claytonia lanceolata var. flava ANACONDA YELLOW BOULDER WHITE+ BURTON PARK WHITE+ CHAMPION YELLOW CHAMPION WHITE HEBGEN YELLOW VIPOND YELLOW VIPOND WHITE WYOMING YELLOW WYOMING WHITE Claytonia lanceolata var. lanceolata CHAMPION LANCEO HEBGEN LANCEO VIPOND LANCEO WYOMING LANCEO PCA of the living-plant morphological characters other than flower color revealed that white-flowered and yellow-flowered forms of C. lanceolata var. fla- va are indistinguishable from each other but are easily separable from C. lanceolata var. lanceolata (Fig. 2). The first principal component accounted for 46% of the variation and had strong contribu- tions by petal width, leaf length, stem height, leaf length/width ratio, sepal length, and petal length/ width ratio. The second component had strong loadings by leaf width and petal length and ac- counted for 20% of the variation (Table 3). The cross-validation error rate for the discrimi- nant analysis comparing white- and yellow-flow- ered individuals of C. lanceolata var. flava was 0.42; there is only a 58% chance of correctly iden- tifying the two flower color phenotypes of C. lan- ceolata var. flava based on the morphological char- acters used in the analysis. Thus, the two pheno- types cannot be reliably discriminated on characters other than flower color. ANACONDA WHITE ANACONDA LANCEO Collection data Montana, Deer Lodge Co. Shelly & Lesica 1412* Montana, Deer Lodge Co. Shelly & Lesica 1413* Montana, Sweet Grass Co. Shelly 1617 Montana, Silver Bow Co. Shelly, Schassberger & Schitoskey 1504 Montana, Jefferson Co. Shelly 1417* Montana, Jefferson Co. Shelly & Lesica 1423* Montana, Gallatin Co. Shelly & Lesica 1419* Montana, Beaverhead Co. Shelly & Scow 1444* Montana, Beaverhead Co. Shelly & Scow 1445* Wyoming, Fremont Co. Shelly & Lesica 1446* Wyoming, Fremont Co. Shelly & Lesica 1447* Montana, Deer Lodge Co. Shelly & Lesica 1411 Montana, Jefferson Co. Shelly & Lesica 1422 Montana, Madison Co. Shelly & Lesica 1420* Montana, Beaverhead Co. Shelly 1201 Wyoming, Teton Co. Shelly & Lesica 1448* Herbarium studies. Taxon means, ranges, and standard deviations for the eight quantitative and two qualitative characters examined on the herbar- ium collections are given in Table 4. PCA of the herbarium morphological characters other than flower color also revealed that white- flowered and yellow-flowered forms of C. lanceo- lata var. flava are indistinguishable from each other, and are very similar to var. multiscapa and C. ro- sea, but that specimens of all three latter taxa are easily separable from C. lanceolata var. lanceolata (including C. lanceolata var. chrysantha; Fig. 3). The first principal component accounted for 33% of the variation and had strong contributions by leaf venation, petal apex outline, leaf length/width ratio, leaf width, petal/sepal length ratio, and _ sepal length. The second component had strong loadings by petal width and length and accounted for 22% of the variation (Table 5). Isozyme electrophoresis. Coding of populations of C. lanceolata var. flava was straightforward, as 68 MADRONO TABLE 2. TAXON MEANS, RANGES AND STANDARD DEVI- ATIONS FOR FIELD MORPHOLOGICAL DATA, CLAYTONIA LAN- CEOLATA VARS. FLAVA AND LANCEOLATA. Yellow flava White flava lanceolata No. of specimens 225 270 225 Height (cm) Mean 7.6 9.6 4.1 Range 3.5-16.9 4.0—27.2 1.4-10.8 SD 2.1 1.3 js: Leaf length (mm) Mean 36.0 42.8 26,3 Range 13.0—76.0 14.0-111.0 14.0—46.0 SD 11.7 18.4 6.8 Leaf width (mm) Mean 5.4 5.9 9.1 Range 2.5-11.5 3.0-—13.5 4.0-19.0 SD 1.6 1.9 oe | Sepal length (mm) Mean 5.0 a | 4.0 Range 3.0-8.5 3.5-8.0 2.0—6.0 SD 0.98 0.83 0.79 Petal width (mm) Mean 5.3 a. 4.2 Range 3.0-8.5 3.0-9.0 2.5-9.0 SD 0.96 0.97 0.84 Petal length (mm) Mean 8.6 o2 8.8 Range 6.0—12.0 6.5-13.5 4.5-12.5 SD L2 1.1 L.3 Leaf length/width ratio Mean 6.8 Ta 3.0 Range 3.3-14.2 2.6-18.5 1.6-6.3 SD 2.0 2.6 0.9 Petal length/width ratio Mean 1.7 1.6 pA Range 1.1-2.3 1.1-2.3 0.5-3.0 SD 0.2 0.2 0.3 simple diploid expression was observed in all cases. However, C. lanceolata var. lanceolata ex- pressed more complex banding patterns indicative of tetraploidy. To make comparisons among varie- ties and flower-color phenotypes at each locality, it was necessary to code allele frequencies for C. lan- ceolata var. lanceolata. This was done by assuming that each individual was tetraploid and possessed four allelic doses per locus. Some individuals, therefore, expressed more than two alleles at a lo- cus. Relative staining intensities were used to de- termine dosage effects (Wolf 1988). Allele frequen- cies are given in Table 6. Differences between varieties. The UPGMA cluster analysis of Nei’s genetic identity values is shown in Figure 4. All five populations of C. lan- ceolata var. lanceolata were completely separated from the nine populations of C. lanceolata var. fla- va (represented by samples of both white- and yel- [Vol. 45 low-flowered plants). The mean genetic identity be- tween populations of these two taxa was 0.69. Differences among populations of C. lanceolata var. flava. The UPGMA cluster analysis also indi- cates the level of differentiation among populations of C. lanceolata var. flava (Fig. 4). With the color phenotypes pooled within localities, genetic iden- tity values among the five study localities ranged from 0.913 to 0.979. The genetic identities corre- spond to geographic proximity; the more southerly populations (Hebgen Lake and Wyoming) clustered together, as did the northern populations (Anacon- da, Champion, and Vipond). Differences between color phenotypes within lo- calities of C. lanceolata var. flava. In the four cases where they are biotically sympatric, yellow- and white-flowered “‘populations”’ of C. lanceolata var. flava were always more similar allozymically to each other than to allopatric populations of the same flower color (Fig. 4). The Nei’s genetic iden- tity values between color phenotypes within local- ities were high, ranging from 0.995 (Vipond Park) to 1.00 (Anaconda). By contrast, interpopulation genetic identity values within color phenotypes ranged from 0.935 to 0.987 for the yellow form, and from 0.910 to 0.989 for the white form. DISCUSSION Morphological studies and isozyme electropho- resis revealed that populations ascribed to C. lan- ceolata var. flava represent a diploid species (2n=16; Marriott 1986) that is distinct from the C. lanceolata complex. Claytonia lanceolata dis- played banding patterns suggestive of autopoly- ploidy in the populations we sampled. Tetraploid populations (n=16) of C. lanceolata have been re- ported from Utah (Halleck and Wiens 1966; Stew- art and Wiens 1971), and populations with n=8, 12, 18, 22, 24, and 32 have been found in other Rocky Mountain populations of this species (Davis and Bowmer 1966; Halleck and Wiens 1966). In past treatments, petal color, described as ‘“‘golden yellow” by Davis (1952, 1966) and “‘deep yellowish-orange”’ by Hitchcock et al. (1964), was the primary character used to distinguish C. lan- ceolata var. flava from related taxa at the level of species (as C. flava; Davis 1952) or variety (Hitch- cock et al. 1964; Davis 1966). However, PCAs of our morphological data indicated that the characters most important for distinguishing C. lanceolata var. flava from typical C. lanceolata are related to leaf morphology (length/width ratio and venation) and petal shape (length/width ratio and apex outline). Davis (1952) described the leaves of C. lanceo- lata var. flava as “‘linear or lance-linear,’’ as com- pared to ‘‘stem leaves lanceolate”’ in C. lanceolata. Similarly, Hitchcock et al. (1964) described the stem leaves of C. lanceolata var. flava as “‘lanceo- late or narrowly oblong, several times longer than broad’’ and those of C. lanceolata (represented by 1998] SHELLY ET AL.: CLAYTONIA SYSTEMATICS 69 3 ° © ° Ss * } ° = . ° ‘ ee ° e° ° by ° ° : . i ° ° : x o* ° ° * ° ° ° * * o ° ° ta ‘ate *4 * rol a * °, "i i °° ° oe 1 * * 1 4 ee oe & Ms €o CW si Ks alia oe FF 6 eo % 8m | > Pp | vytettete rt 8 “ey. e+ Oo 4 yaa . ee. ES Gate, %5 = is ss et od 6° Boe e @. Be } = 0) : ae 3 og a grrode "e ¥..3 - % a * * * © ove ° ° ee ©, 5 oe 1 oT ore VER, Bo es ot 1 ce gk ee hoe PT eet ag es og 4 er es a as oa eee a ee ge = : oe* ed be Soy ? 6 © : Cars Fa : oi acien tas, o oe o FF 1 ° é » var. lanceolata Ad bd = 27 e yellow flava : g white flava — Bo -3 ca a 1 a 3 Factor 1 Fic. 2. Scatter diagram of individuals of Claytonia lanceolata var. flava (yellow- and white-flowered) and C. lanceo- lata var. lanceolata on principal components | and 2. Based on morphological data collected from living plants in the field. var. chrysantha) as “‘broadly elliptic to ovate, (1)1.5—2.5(4) cm long, usually over % as broad.” Our study showed that the leaves of C. lanceolata var. flava average approximately seven times longer than wide, whereas those of typical C. lanceolata average three times as long as wide (Table 2). These numeric differences are in accordance with the earlier, largely qualitative leaf shape differences described for these taxa. Rydberg (1922) recognized the patterns in leaf venation that distinguish the narrow-leaved species of Claytonia (1.e., C. virginica, C. rosea, and C. multiscapa) from C. lanceolata. He observed that the former group has leaves ‘‘1-ribbed or indistinct- ly 3-ribbed,’’ whereas the latter species has leaves that are “‘distinctly triple-ribbed.’’ Our results up- hold this as a valid and important means of distin- guishing the narrow-leaved Claytonia populations from those of C. lanceolata in the northern Rocky TABLE 3. LOADINGS OF THE FIRST Two PRINCIPAL COM- PONENTS FOR THE QUANTITATIVE CHARACTERS MEASURED IN THE FIELD MORPHOLOGY STUDIES. Component Character 1 2 Petal width 0.832 0.114 Leaf length 0.807 O77 Height 0.789 0.033 Leaf length/width 0.755 —0.426 Sepal length 0.700 0.202 Petal length/width —0.646 0.395 Leaf width —0.189 0.847 Petal length 0.474 0.702 Mountains; the leaves of C. lanceolata var. flava have only the distinct midvein, whereas populations of typical C. lanceolata have leaves with two prom- inent lateral veins in addition to the midvein (Table 4). Davis (1952) also distinguished C. lanceolata var. flava from typical C. lanceolata by petal apex outline, describing the former as having petals ‘rounded at the apex,”’’ and the latter with petals ‘“‘retuse or emarginate.’’ Our studies confirmed that the petals of C. lanceolata var. flava are rounded at the apex, while those of typical C. lanceolata are usually retuse or emarginate (Table 4). In addition, the results of both the field and herbarium morpho- logical studies confirmed that the petals of C. lan- ceolata var. flava are more nearly oval in shape, whereas those of C. lanceolata are most often ob- ovate, and frequently narrowly so. These results also concur with the descriptions by Davis (1952). Isozyme electrophoresis also clearly indicated that C. lanceolata var. flava is distinct from typical C. lanceolata and warrants recognition as a distinct species. The mean genetic identity between C. lan- ceolata var. lanceolata and populations represent- ing C. lanceolata var. flava (I = 0.69) is close to the mean between congeneric species (I = 0.67) presented in several reviews (Gottlieb 1981; Craw- ford 1983). This value contrasts greatly with mean values for conspecific populations of var. flava (I = 0.91 to 0.98) and for populations of typical var. lanceolata (I = 0.89 to 0.99). The diploid taxon represented by populations as- signable to C. lanceolata var. flava includes con- specific yellow- and white-flowered plants. In this 70 MADRONO [Vol. 45 TABLE 4. TAXON MEANS, RANGES AND STANDARD DEVIATIONS FOR QUANTITATIVE AND QUALITATIVE MORPHOLOGICAL CHARACTERS FROM HERBARIUM SPECIMENS, CLAYTONIA LANCEOLATA VARS. FLAVA, LANCEOLATA, AND MULTISCAPA, AND C. ROSEA. For some characters, the number of accessions was less than that shown in the first line; exceptions are given in parentheses after the means. * Petal apex outline scores: 0O—retuse/emarginate, 1—rounded; ** Leaf venation scores: 0Q—lateral veins inconspicuous or absent, 1—lateral veins conspicuous. flava lanceolata multiscapa rosea No. of accessions 17 124 8 35 Leaf length (mm) Mean 41.6 32.6 43.4 44.0 Range 18-71 13-59 29-56 17.5-84 SD 3.1 0.9 3.7 2.8 Leaf width (mm) Mean a2 10.4 59 Sl Range 2.4-8.4 2.8-26 2.3-10.6 1.3-14 SD 0.4 0.4 el OS Leaf length/width ratio Mean 8.3 35 8.7 10.8 Range 4.7-15.1 1.7-12.5 5.2—-13.0 4.0—32.8 SD 0.6 0.1 |e 12 Sepal length (mm) Mean 4.4 3.8 4.8 4.7 Range 3.3—-5.7 2.0-6.6 4.0-5.9 2.9-7.0 SD 0.2 0.1 0.3 0.2 Petal width (mm) Mean 4.3 4.0 (123) 4.6 4.1 (32) Range 3.0-5.4 1.8-6.2 2.9-6.0 2.7-5.5 SD 0.2 0.1 0.3 0.1 Petal length (mm) Mean 9.5 9.1 9.2 9.3 (34) Range 6.8-11.8 5.2—14.0 7.5-11.3 5.8-12.7 SD 0.3 0.1 0.6 0.3 Petal length/width ratio Mean 23 2.4 (123) 21 25 (32) Range L5=3.1 1.6-3.7 1.3-—2.6 15-32 SD 0.1 0.1 0.2 0.1 Petal/sepal length ratio Mean 22 2.4 1.9 2.1 (34) Range 1.6—2.8 1.2—3.7 1.3—2.5 1.2—3.7 SD 0.1 0.1 0.1 0.1 Petal apex outline* Mean 0.9 (16) 0.1 (118) 1.0 0.9 Leaf venation** Mean 0.0 0.9 (123) 0.1 0.1 (34) and other cases, flower color has been found to be of limited use in delineating true phylogenetic re- lationships within Claytonia. Elsewhere in North America, several other predominantly white- or pink-flowered taxa in Claytonia include named or unnamed yellow-flowered forms. Examples include C. lanceolata var. chrysantha (Douglas and Taylor 1972), C. virginica L. var. hammondiae (Kalm- bacher) Doyle, Lewis and Snyder (Snyder 1992), and a recently discovered population of C. caroli- niana in Maryland that contains yellow-flowered plants in addition to typical white- to pink-flowered plants (Snyder 1992). Such color forms are proba- bly best viewed as minor variants within their re- spective taxa. They probably do not typically war- rant taxonomic recognition, except in cases where their populations are correlated with ecological, ge- netic, geographic, and/or further morphological segregation (as is the case for C. virginica var. ham- mondiae) (Snyder 1992). In the case of C. lanceo- lata var. chrysantha, Douglas and Taylor (1972) found that, based on morphological, ecological, and biochemical analyses, ‘‘... there is no significant difference between the yellow and white forms of Claytonia lanceolata, other than petal color,’’ and that ‘‘(t)he difference in petal color is most likely 1998] SHELLY ET AL.: CLAYTONIA SYSTEMATICS vi) ao ° 31 ° ; s : t 5 ate a O -1- * +f * = * Ps Lo ; , E | * * . %e x flava | » lanceolata » multiscapa 4 + OSseG -3 T T T T T T T ae T T T 1 PaCwOr 7 Fic. 3. Scatter diagram of individuals of Claytonia rosea and C. lanceolata vars. flava, lanceolata (including var. chrysantha), and multiscapa on principal components | and 2. Based on morphological data collected from herbarium specimens. due to one or very few genes, as evidenced by the virtual lack of intermediate color forms.’’ They concluded that “*... there is no basis for the rec- ognition of var. chrysantha ...’’ (Douglas and Tay- lor 1972). Our results indicate the same situation with respect to the yellow and white flower color phenotypes of “C. lanceolata var. flava.” Plants of the two color phenotypes are biotically sympatric in at least four populations in the northern Rocky Mountains, and these phenotypes reflect little or no morphological or isozyme differentiation within or among those populations. While there was some genetic differentiation among populations of C. lanceolata var. flava, plants of the two flower color phenotypes are undoubtedly conspecific; at the four sites where they are biotically sympatric, individ- TABLE 5. LOADINGS OF THE FIRST Two PRINCIPAL COM- PONENTS FOR THE QUANTITATIVE AND QUALITATIVE CHAR- ACTERS USED IN THE HERBARIUM MORPHOLOGY STUDY. Component Character ] 2 Venation —0.840 0.039 Petal apex outline 0.830 —0.051 Leaf length/width 0.775 —0.061 Leaf width —0.630 0.472 Petal/sepal length ratio —0.607 0.121 Sepal length 0.588 0.450 Petal width 0.069 0.872 Petal length —0.095 0.778 Leaf length 0.344 0.500 Petal length/width —0.174 —0.409 uals of the two phenotypes are nearly or completely identical genetically. This suggests that in such cases they are part of the same breeding population, that flower color represents simple genetic differ- ences (i.e., determined by one or a few genes), and that flower color does not warrant taxonomic rec- ognition. The allozyme data also suggest separate origins of the yellow flower phenotype in each lo- cality where it occurs. The morphological comparison among species of Claytonia revealed a strong similarity between C. rosea and C. lanceolata vars. flava and multiscapa (Fig. 3). The latter variety, all collections of which are white-flowered, is reported by Hitchcock et al. (1964) as occurring in *‘Yellowstone National Park and vicinity.”” The morphological similarity of var. multiscapa to var. flava, and its complete geograph- ic overlap with stations of white- and/or yellow- flowered populations of the latter entity, support the notion that var. multiscapa is the same taxon as the white-flowered form of var. “‘flava.’’ The more southerly white- to pink-flowered Claytonia rosea probably represents a similar, closely related nar- row-leaved taxon (J. Miller personal communica- tion). Like the populations of C. lanceolata var. fla- va sampled in Montana and Wyoming, numerous Colorado populations of C. rosea are diploid (n=8; Halleck and Wiens 1966). In summary, electrophoretic and morphological data clearly revealed that C. lanceolata var. flava does not belong in the C. lanceolata complex. Rather, its affinities lie with the narrow-leaved group of species that includes C. rosea, C. tuberosa Pallas ex Willd. and C. virginica. Furthermore, C. a) - roo o6¢0 Lc0 cO'0 08°0 £60 vs'0 SoO0 cS'0 0) OS 0 eS 0 0) v6'0 q e 68°0 0) 0) 860 0c 0 LO'0 9V'0 0) cO'0 00'1 O 0) 00'I O eB JUN = OOT OOT ODOT 00'T 00'l 00'I 00'l 00'l 00'1 00'1 00'l 00'I 001 00'l e e-Idy, OOT OOT OOT 00'T 00'l 001 00'I 00'| 001 00'| 00'1 00'I 00'l 00'l e c-1d 0) 0) O O 0) 0) 0 0) 0 0) 0) 0) 0) £00 P rOO0 roo coo 100 vIO 80°0 vO'O e10 60°0 810 cO'0 £00 8c 0 SIO 3 960 960 ¢60 66°0 98°0 c6'0 960 L380 160 1380 c60 160 ILO [80 q 0) O s00 0) 0 O 0) 0) 0) 100 90°0 90°0 100 100 eB [-1d], IcO v60 860 Sc O 19°0 69°0 tv 0 £60 88°0 vr0 ILO 89°0 cS'0 96°0 2) ‘vO 900 coo OL'0 6c°0 [e'0 c¢o'0 LOO clO SSO 6c'0 ce 0 vr'0 vO'0 q 90°0 0) 0) SO'0 0) O cO'0 0) 0) 100 0) 0) vO'O 0) e T-18d 0) 0) O 0) 0) 0) 0) cO'0 0) 0) 10'0 O 0) O q OOT OOT OOT 00'1 00'T 00'T 00'l 860 00'1 00'I 660 O0'l 001 00'l eB [-18d OoOT vod coOO0 06'0 0) 0) 86°0 0) 0) IZ°0 cO'0 0) 00'T 0) ) 0 60 860 O10 79°0 vs'0 cO'0 00 860 8c 0 860 O0'I O 00'l q 0 coo 0) 0) 9¢°0 9V'0 O 0) cO'0 10'0 O 0) 0) 0) e 4PV OOT OOT OOT 00'I 00'l 00'T 00°! 00'l 00'l 00'T 00'| 00'1 00'l 00'1 eB eV 6c°0 0 coo 8V'0 0) 0) £00 O 100 O10 S00 O £9°0 0) 3 SSO OOT 860 vr'O 00'I 00'1 L6°0 00'l 660 L380 S60 00'l 9¢°0 00'l q 90°0 0) 0) 80°0 0) 0) 0) 0) 0) £0°0 0) 0) 100 0) eB dey O OOT OOT OOT 00'T 00'l 00'I 00'I 00'l 00'l 00'l 001 00'1 00'l 00'1 e ypden a ITO Ov0O Ic0 S00 0) 0) c0'0 S00 £00 90°0 LO'0 61°0 €lO 0) P ry 6r0 810 cso0 LLO cr'0 6c 0 8r'0 cO'0 £00 LS‘0 £00 90°0 IcO 9c 0 3 Z, Oro cro L7c0 810 90°0 90°0 9r'0 ©60 88°0 LeO 060 SLO $90 O10 q > 0) 0) 0) 0) vs'0 co O0 vO'O 0) 0) O 0 0) 100 v9'0 e c-Wsd 0 £00 0) 0) 0) 0) 0) Lv'0 8c0 0) 0) 0) 0) O10 q OOT L60 OOT 00'T 00'l 00'l 00'l eo 0 c9'0 00'I 00'I 00'I 00'I 060 eB [-Wsq 0 £00 S00 90°0 0) 0) 0) 0) 0) 0) £00 vO'0 0) 0) P 0 L60 S60 ITO 00'l 00'T SIO 00'l 00'| 0) L6°0 96°0 0) 00'I 3 00'l O 0) S90 0) 0) 790 0) 0) 001 0) 0) v6 0 0) q 0) 0) 0) 810 O 0) [70 0) 0) 0) 0) 0) 90°0 0) Le ¢-4PIN OOT OOT OOT 00°! OOI 001 00'l 00'l 00'1 66°0 00 00'1 00'I 00° q 0) 0) 0) O 0) 0) O 0) 0) 100 0) 0) 0 0) e c-4UPIN 0 v60 060 0) 00'l 00'T cO'0 06'0 c60 O 66 °0 S60 61°O 00'l 3 OOT 900 O10 00'T 0) 0) 860 O10 80°0 86°0 100 S0'0 [80 0) q 0) O 0) 0) 0) O 0) 0) 0) cO'0 0) 0) 0 0) e T-4UPIN Oc OV Oc OS OS OS 6 Oc OS OS OS OS OS eOS S1FI1V snso’y] ouP] M A our] MN K ouP] M K ouP] M A ouP] 2A yieg puodi, SUILLOA AA, sstg uoidweyy) epuooeuy jayey uasqoyH suoneso, Apmis ‘UOT}VIOT YORI ye poyduies ydiour JO[OO JOMOY IO UOXe}] Jod S[eNpPIAIpUI JO JOqUINU = , ‘DIDJOAIUD] “eA DIDJOJIUD] “QD = Jue] ‘vaAvYf “tea vIvJOaIuY] “Dd Jo adAjouayd pasaMmoy-ayM = M ‘vAvd}f eA YIDJOaIUD] ‘dD Jo adAjousyd paramMoy-MoOT[IA = A z ‘YeT uUIsqoH 3e pojussoidos c qy JOU SI DAD “eA DIDJOaIUY] “~D Jo sdAjousyd poiaMoy-aYyM = , ‘VLIWIOSONVI VINOLAVIJ AO SNOILVINdOg pl YOA IOOT AWAZNY 6] LV SHIONANOTA ATATTY ‘9 ATAVLE Lt 1998] TABLE 6. CONTINUED Study locations Vipond Park Champion Pass Wyoming Anaconda Hebgen Lake! lanc Ww lanc 50 lanc 49 lanc 50 lanc 30 0.73 50 30 50 50 0 0.61 0.45 0.46 0.50 0.47 0.06 0.07 1.00 0) 1.00 0.58 1.00 0.58 1.00 1.00 0.20 0.79 0.01 1.00 0.23 0.76 0.01 1.00 1.00 1.00 0.63 0.37 1.00 0.41 0.59 1.00 1.00 0.55 0.45 1.00 0.51 0.49 1.00 Skdh 0.22 0.78 6pgd-1 SHELLY ET AL.: CLAYTONIA SYSTEMATICS q3 O53 0.47 1.00 0.40 0.42 0.10 0.90 0.94 0.06 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 6pgd-2 0.07 0.93 0 0 0.05 0.95 0.13 0.87 0.03 0.97 Idh 1.00 O31 0.69 0.35 0.65 0.77 0.23 O75 0.25 0.95 0.05 0.91 0.09 0.99 1.00 0.01 0.98 0.02 lanceolata vars. flava and multiscapa would best be treated conspecifically as C. multiscapa (J. Miller personal communication); such a proposed treat- ment is supported by the results of our herbarium morphological study, as well as the entirely over- lapping geographic ranges, of the plants currently bearing these names from the Yellowstone and sur- rounding areas. Formal nomenclatural changes are not made here, but left for publication of a com- plete revision of the genus (Miller and Chambers in mss.). Conservation status. In the northern Rocky Mountains, narrow-leaved populations of Claytonia consisting wholly or partially of yellow-flowered individuals remain relatively uncommon (ten such populations are now known from Idaho, Montana, and Wyoming). However, the morphologically and allozymically highly similar white-flowered popu- lations are more common and widespread. These white-flowered populations occur over a larger area in northwestern and north-central Wyoming, and south-central to southwestern Montana; a popula- tion has also recently been confirmed in the Sweet- grass Hills of north-central Montana (B. Heidel per- sonal communication). Populations of both flower color phenotypes are usually very large in size and areal extent, and at least 30 populations consisting of one or both forms have been documented in Montana (Montana Natural Heritage Program un- published data). Because these yellow and white flower color phenotypes are “‘contaxonomic,”’ C. lanceolata var. flava is not in need of protective listing, regardless of its eventual taxonomic dispo- sition. When necessary, legal protection and manage- ment of putatively endangered taxa should be pre- ceded by accurate evaluations of their phylogenetic relationships and taxonomic status (Avise and Nel- son 1989). Biochemical and molecular techniques will continue to be increasingly useful for ensuring that the limited funds available for endangered spe- cies conservation are correctly focused on evolu- tionarily deserving taxa in the endeavor to maintain biological diversity. ACKNOWLEDGMENTS Funding for these studies was provided to the Montana Natural Heritage Program by the U.S. Forest Service and the U.S. Fish and Wildlife Service. Duane Atwood, An- gela Evenden, Robert Moseley, Jan Nixon, Lisa Schass- berger, Frank Schitoskey and Ken Scow assisted with the field work. Karen Bothel, Eric Collier, Trace Collier and Tracy Tucker assisted with the electrophoretic analyses. Jeff Doyle, Bonnie Heidel, Hollis Marriott, John Miller, and David Snyder provided valuable information or com- ments. Cedron Jones produced the map. Kenton Chambers and Bruce Pavlik reviewed the manuscript, and Elizabeth Painter guided it to publication. We gratefully acknowl- edge the curators and their assistants from the herbaria cited for making loans available. 74 MADRONO 0.60 0.67 0.73 0.80 0.87 0.93 [Vol. 45 HEBGEN YELLOW WYOMING YELLOW WYOMING WHITE ANACONDA YELLOW ANACONDA WHITE VIPOND YELLOW VIPOND WHITE CHAMPION YELLOW CHAMPION WHITE HEBGEN LANCEO ANACONDA LANCEO WYOMING LANCEO CHAMPION LANCEO VIPOND LANCEO 1.00 Fic. 4. Phenetic relationships among populations of Claytonia lanceolata var. flava (yellow- and white-flowered) and C. lanceolata var. lanceolata, based on cluster analysis (UPGMA) of Nei’s genetic identity values. LITERATURE CITED AVISE, J. C. AND W. S. NELSON. 1989. Molecular genetic relationships of the extinct dusky seaside sparrow. Science 243:646—648. BoIvIN, B. 1968. Flora of the prairie provinces. Phytologia 16:265-339. CRAWFORD, D. J. 1983. Phylogenetic and systematic in- ferences from electrophoretic studies. Pp. 257—287 in S. D. Tanksley and T. J. Orton (eds.), Isozymes in Plant Genetics and Breeding, Part A. Elsevier Science Publishers, Amsterdam, Netherlands. Davis, R. J. 1952. Flora of Idaho. Wm. C. Brown Com- pany, Dubuque, IA. . 1966. The North American perennial species of Claytonia. Brittonia 18:285-—303. and R. G. BOwMER. 1966. Chromosome numbers in Claytonia. Brittonia 18:37-38. Dorn, R. D. 1984. Vascular Plants of Montana. Mountain West Publishing, Cheyenne, WY. DoucG.Las, G. W. AND R. J. TAYLOR. 1972. The biosyste- matics, chemotaxonomy, and ecology of Claytonia lanceolata in western Washington. Canadian Journal of Botany 50:2177—2187. GoTTLigB, L. D. 1981. Electrophoretic evidence and plant populations. Progress in Phytochemistry 7:1—45. HALLECK, D. K. AND D. WIENS. 1966. Taxonomic status of Claytonia rosea and C. lanceolata (Portulacaceae). An- nals of the Missouri Botanical Garden 53:205-—212. HAUFLER, C. H. 1985. Enzyme variability and modes of evolution in Bommeria (Pteridaceae). Systematic Bot- any 10:92-—104. HitTcHcock, C. L., A. CRONQUIST, M. OWNBEY, AND J. W. THOMPSON. 1964. Vascular Plants of the Pacific Northwest, Part 2. University of Washington Press, Seattle, WA. MarrioTT, H. 1986. Status report, Claytonia lanceolata var. flava. Unpublished report to U.S. Fish and Wild- life Service, Denver, Colorado. Rocky Mountain Her- itage Task Force, Laramie, WY. NELSON, A. 1900. New plants from Wyoming—xII. Bul- letin of the Torrey Botanical Club 27:258-274. . 1926. Taxonomic studies. 2, Miscellaneous new species. University of Wyoming Publications in Bot- any 1:122-143. ODRZYKOSKI, I. J. AND L. D. GOTTLIEB. 1984. Duplications of genes coding 6-phosphogluconate dehydrogenase (6-PGD) in Clarkia (Onagraceae) and their phyloge- netic implications. Systematic Botany 9:479—489. RYDBERG, P. A. 1922. Flora of the Rocky Mountains and Adjacent Plains, 2nd ed., 1954 reprint. Hafner Pub- lishing Co., New York. . 1932. Claytonia in: North American Flora 21(4): 279-313. New York Botanical Garden, New York. SHELLY, J. S. 1989. Status review of Claytonia lanceolata var. flava, Beaverhead, Deerlodge and Gallatin Na- tional Forests, Montana. Unpublished report to U.S. Forest Service, Region 1, Missoula, Montana. Mon- tana Natural Heritage Program, Helena, MT. SNYDER, D. B. 1992. A new status for New Jersey’s yellow spring beauty. Bartonia 57:39—49. SoLtis, D. E., C. H. HAUFLER, D. C. DARROW, AND G. J. Gastony. 1983. Starch gel electrophoresis of ferns: A compilation of grinding buffers, gel and electrode buffers, and staining schedules. American Fern Jour- nal 73:9—27. STEWART, D. AND D. WIENS. 1971. Chromosome races in Claytonia lanceolata (Portulacaceae). American Jour- nal of Botany 58:41—47. SWOFFORD, D. L. AND R..B. SELANDER. 1981. BIOSYS-1: A Fortran program for the comprehensive analysis of electrophoretic data in population genetics and sys- tematics. Journal of Heredity 72:281—283. U.S. FISH AND WILDLIFE SERVICE. 1985. Notice of review. Federal Register 50:39525-—39584. . 1993. Notice of review. Federal Register 58: 51144-51190. WENDEL, J. EF AND N. EF WEEDEN. 1989. Visualization and interpretation of plant isozymes. Pp. 5—45 in D. E. Soltis and P. S. Soltis (eds.), Isozymes in Plant Bi- ology. Dioscorides Press, Portland, OR. WILKINSON, L. 1986. SYSTAT: The System for Statistics. Evanston, IL. WoLF, P. G. 1988. Analysis of electrophoretic variation in Claytonia lanceolata vars. lanceolata and flava. Un- published report to Montana Natural Heritage Pro- gram, Helena, Montana. Washington State University, Pullman, WA. MADRONO, Vol. 45, No. 1, pp. 75-84, 1998 EFFECTS OF CLIMATIC VARIABILITY ON HERBACEOUS PHENOLOGY AND OBSERVED SPECIES RICHNESS IN TEMPERATE MONTANE HABITATS, LAKE TAHOE BASIN, NEVADA TIMOTHY G. FE KITTEL National Center for Atmospheric Research, Box 3000, Boulder, CO 80307-3000 and Natural Resource Ecology Laboratory, Colorado State University, Fort Collins, CO 80523 ABSTRACT Surveys of herbaceous flora found in flower in the eastern central Sierra Nevada (Nevada) demonstrated the influence of climate variability on herbaceous phenology and observed species richness. Mid-summer surveys conducted in two climatically different years showed differences in number of flowering species (greater in the wetter year) and their characteristic phenological class (more early- vs. late-flowering species in the wetter year). Between year variation in phenology and species richness appeared to be keyed to snowpack duration, spring and summer moisture availability, and early growing season temper- atures. Responses in mesic habitats were greater than in xeric sites. Phenological responses to climate variability can be interpreted in terms of displacement, compression, and amplification of the phenological cycle. Comparison of the two surveys suggests that both displacement and amplification generated ob- served differences. In the wetter year, late snowpack and lower spring temperatures delayed flowering times and greater spring and early summer moisture resulted in greater observed species richness. Sen- sitivity of herbaceous species in these montane habitats to year-to-year differences in climate highlights the importance of considering effects of climate variability on phenology and population dynamics in short-duration plant biodiversity surveys and long-term monitoring programs. INTRODUCTION Climate variability strongly affects population, community, and ecosystem processes in herbaceous systems (e.g., Tilman and El Haddi 1992; Walker et al. 1994). Stand phenological responses to inter- annual variability in thermal, moisture, and light regimes can be dramatic, reflecting seasonal dy- namics of species reproductive cycles, community composition, and plant production. Understanding the herbaceous response to interannual climate vari- ability and identification of key surface climate fac- tors controlling this response can reveal the sensi- tivity of the herbaceous component of plant com- munities to short-term climate variation and direc- tional climate change. This understanding can, in addition, aid in evaluating the representativeness of rapid biodiversity assessments and design of long- term monitoring studies. In temperate regions, soil and air temperature, moisture availability, photoperiod, and, in areas re- ceiving snow, snowpack duration are common con- trols over growth initiation and timing of subse- quent phenophases (Holway and Ward 1963, 1965; Lieth 1974; Dickinson and Dodd 1976; Pitt and Heady 1978; Schemske et al. 1978; Bertiller et al. 1990). In mountain environments of western North America, one of the strongest interannual climatic signals is that of winter snowpack. The strength of this signal is derived from the effect of quasi-pe- riodic interannual variations in tropical and mid- latitude Pacific sea surface temperatures on the strength and tracks of winter storms, such as as- sociated with El Nifio and the North Pacific Oscil- lation (Ropelewski and Halpert 1986; Sheaffer and Reiter 1985). The objective of this study was to understand the phenological response of the herbaceous compo- nent of several montane communities in the eastern central Sierra Nevada (Nevada) to interannual snowpack variation and associated variation in tem- perature. The approach was to analyze changes be- tween two plant surveys, one conducted in a low snowpack year and the other in a high snowpack year, in terms of (1) differences in the number of species found in flower and predominant phenolog- ical class of these species and (2) whether these responses varied by habitat. METHODS Study area. The plant surveys were conducted in the Sierra Nevada montane forest belt on the east- ern side of the Lake Tahoe Basin, Nevada. The study area covered four watersheds: Skunk Harbor, Slaughterhouse Canyon, Marlette Creek, and Spooner Lake (Fig. 1). These sites were chosen to capture a wide variety of vegetation types charac- teristic of this montane region. The predominant type is East-slope Jeffrey Pine Forest as described by Rundel et al. (1980), with dominance shared by Pinus jeffreyi, P. ponderosa, Abies concolor, and Calocedrus decurrens. The forest is generally open, reflecting a relatively xeric environment. 716 MADRONO [Vol. 45 I2O0°W IS | | 4 2676m _ Marlette CALIFORNIA | NEVADA } =X Lake ae S Vall now Vaile Washoe Co. Creek “i ~ Peak y ras 2808 m Spooner ‘i —_ Douglas Co. ‘\ ‘\ “< C Survey Areas \ ~ Fic. 1. Location of the survey area in the Lake Tahoe Basin, California-Nevada. The four survey sites are indicated by enclosed short-dashed lines on the detailed map. Each watershed also contains varying amounts of other vegetation types. Skunk Harbor (elevation range 1900—2000 m) includes narrow riparian com- munities and montane chaparral with Arctostaphy- los patula and Chrysolepis sempervirens (see Smith 1973; Ornduff 1974). The Slaughterhouse Canyon site (1920-1990 m) has extensive montane mead- ows and adjacent stands of Pinus contorta ssp. murrayana. Vegetation of the Marlette Creek wa- tershed (1970—2440 m) is predominantly Jeffrey Pine Forest with a transition at higher elevations to Red Fir (Abies magnifica) Forest (Rundel et al. 1980). The Spooner site (2120-2150 m) includes a large meadow surrounding Spooner Lake, that grades from wet meadow dominated by Carex spp. to dry meadow (Rundel et al. 1980; Smith 1973), and an adjacent sagebrush scrubland (Smith 1973) dominated by Artemisia tridentata. Surveys. Gordon and Dossman (1968) undertook a presence-absence survey of non-graminoid her- baceous species across the study area in 1968. In the following year, I repeated their survey to eval- uate how the findings of such efforts might vary between two climatically-different years (Kittel 1969). Both 1968 and 1969 surveys spanned the first week of July. One observer from the first sur- vey (J. Gordon) assisted in the second. This facil- itated consistency in both diversity and areal cov- erage of vegetation types sampled within each wa- tershed. The surveys recorded presence or absence of non-graminoid herbaceous species that were in flower during repeated walk-throughs in each area. As an indicator of species habitat, the surveys re- corded vegetation type (e.g., wet meadow, Jeffrey Pine Forest, montane chaparral) for each species recorded. The original taxonomic authority was Munz and Keck (1973); taxonomy presented in this paper was updated to Hickman (1993). Differences between years in total area sampled across watersheds and by vegetation type could not be quantified, potentially limiting the significance of results. However, coverage of the study area was extensive, widely sampling each watershed and vegetation type in both years. In general, sampling of each vegetation type was continued until no or very few additional species were found on visits to different parts of a watershed. As a result, between- year differences in the total area surveyed within a watershed or by vegetation type likely had only small systematic effects on the number and phe- nological class of species found. Climate data. As representative measures of cli- mate for the region, I used monthly and daily tem- perature and precipitation data for Glenbrook, Ne- vada (elevation 1915 m; Fig. 1) and precipitation and snow depth data from Spooners Station (2142 1998] m) near Spooner Lake (U.S. Environmental Data Service 1967, 1968, 1969a, b, 1970, 1973; U.S. Weather Bureau 1965). I calculated growing degree days from mean daily temperatures with a base temperature of 0°C. Analyses. I classified the species into early- and late-flowering taxa based on flowering times given by Munz and Keck (1973). While the given flow- ering times span those observed over the elevation- al extent of a species, they indicate a species’ ten- dency toward early- vs. late-flowering. Because the surveys were made in early July, I defined early- flowering species as those with flowering periods centered on June or earlier and late-flowering spe- cies as those with flowering periods centered on July or later. I included species with June—July flowering periods in the early-flowering class. I omitted from the classification 21 species with broadly specified flowering times overlapping this division. I compared the 1968 and 1969 frequency distri- butions of early- vs. late-flowering classes for in- flower species that were unique to a year for each of the sites. I used a chi-square (x’) test for inde- pendence (Conover 1980; SPSS and NoruSis 1986) to test the null hypothesis that the distribution of unique species between early- and late-flowering categories was the same for the two years. In addition to the site-by-site analysis, I em- ployed three analyses on data combined across sites. First, in a combined-flora analysis, I deter- mined whether a species was unique to a year based on 1968 and 1969 species lists combined across all sites. I evaluated flowering class distributions of these unique species in a x’ analysis as in the site- by-site analysis. This test was more conservative than the site-by-site analysis because species unique to a year at one site were omitted if they were present for the other year at other sites. Second, in a pooled-distribution analysis, I summed the distributions of flowering time classes across all sites and evaluated the pooled distribution with a x’ test. This was a more powerful test than the combined-flora analysis because the response patterns of each site were represented in the anal- ysis. In a third cross-site analysis, I evaluated whether habitat influenced phenological responses to cli- mate variability across the study area. I stratified unique species into two habitat classes based on broadly-defined ‘‘mesic”’ and ‘‘xeric”’ groupings of vegetation types. This stratification reflects a simple but clear distinction in habitats across the study area. I analyzed pooled flowering class distributions for each habitat class with a x? test. Habitat classes based on vegetation type were as follows: mesic habitats included forest-meadow edge, wet and dry meadow, and riparian areas, while xeric habitats in- cluded conifer forest, chaparral, and sagebrush scrubland. Although vegetation type is only one de- KITTEL: CLIMATE VARIABILITY AND HERBACEOUS SURVEYS 77 scriptor of a species’ habitat, vegetation type tended to co-vary across the landscape with other site char- acteristics (slope, aspect, drainage, and soil). Mesic vegetation types tended to be on less well drained sites with more developed finer soils. Xeric types were found on steep slopes with poorly developed coarse soils (e.g., decomposed granite). A benefit of this analysis is that stratification by habitat class helped to remove possible systematic differences between years in areal coverage of vegetation types. RESULTS Mean climate. The long-term mean maximum temperature for the warmest month (July) at Glen- brook was 28.2°C and mean minimum for the coldest month (January) was —5.0°C (Fig. 2a). Long-term mean annual precipitation for Glen- brook was 487 mm. The seasonal cycle of precip- itation has a winter maximum characteristic of Mediterranean climates (Fig. 2a). Winter precipi- tation is generated by mid-latitude cyclonic storms from the Pacific and generally falls as snow at this elevation. There is a second, smaller precipitation peak in late spring (around May) that is evident in records for individual years (Fig. 2b). This peak is not distinguishable in the mean pattern (Fig. 2a) because of variability in its timing. This secondary maximum is common at higher elevations of the central and southern Sierra Nevada and is largely due to convective precipitation associated with greater surface heating and conditionally unstable air associated with late spring weak cyclones from the Pacific (Pyke 1972). Summer climate is domi- nated by dry subsiding air from the eastern Pacific subtropical high pressure center, with only occa- sional rainfall from convective storms. Climate variability. There was high variability in monthly precipitation and temperatures among years 1967, 1968, and 1969 at Glenbrook (Fig. 2b). For the 1967—1968 water year (July—June), that is, for the 12 months prior to the 1968 survey, precip- itation totalled 443 mm, or 91% of the long-term mean. In contrast, the 1968-1969 water year pre- cipitation was 771 mm, or 158% of the mean. This interannual difference was also reflected in Spoon- ers Station precipitation and snow depth records (Fig. 3). Snowpack was greater in 1969 than in 1968 (e.g., 152 cm on 15 February 1968 vs. 221 cm on the same date in 1969; Fig. 3a). Snow cover lasted at Spooners through early April in 1969 ver- sus through the end of February in 1968 (Fig. 3a). Timing of the spring precipitation maximum at Glenbrook was one month later in 1969 than in 1968 (June vs. May, Fig. 2b). The month shift in 1969 spring precipitation was not an artifact of di- viding the record into months when constructing means: daily precipitation data showed that the spring minimum and maximum in each year were well centered in the indicated months. Timing of 78 MADRONO [Vol. 45 (a) GLENBROOK (I95Im) 8.3° 487mm JFMAMJJASOND oO o 300 S oe 8.0° 789mm 8.2° 459mm 8.1° 722mm roe = = < P a ie 80 2 = m i fe) S v a 4 2 = = oO 2 = = 3 Z 3 2) — = JFMAMJJASONDVJVFMAMJJASOND:-JFMAMJJASOND Fic. 2. Walter (1985) climate diagrams for Glenbrook, Nevada. (a) Long-term (1945-1970) mean temperature (T; left axis) and monthly precipitation (ppt; right axis). (b) Monthly mean temperature and monthly precipitation for 1967, 1968, and 1969. Arrows mark survey periods (first week in July) and asterisks (*) indicate the spring precipitation maxima preceding each survey. Vertically-hatched areas indicate periods of water surplus and stippled areas those of water deficit as defined by Walter (1985). Bars under the x-axis show months with mean daily minimum temperature < O°C (solid bar) and months with absolute monthly minimum < 0°C but mean daily minimum > 0°C (diagonally- hatched bar). Solid area under the precipitation curve in (b) denotes a scale change in the precipitation axis. The two values above each year’s plot in (a) and (b) are the mean annual temperature (°C) and annual precipitation (mm/y). In (a), the value given in square brackets is the number of years in the record, and values to the left of the plot are, from top to bottom, highest temperature in the record, mean daily maximum of the warmest month, mean daily minimum of the coldest month, and lowest temperature in the record. the spring maximum is critical because it generally marks the end of the precipitation season. Between year variability in early season temper- atures was reflected in growing degree-days. 1968 had more growing degree-days from January through June than did 1969 (1136 vs. 1019°C- days). Average temperature in June, the month just prior to the surveys, was greater in 1968 than in 1969 (15° vs. 13.3°C). In summary, the water year prior to the 1968 survey was warmer, drier, and snow-free earlier than that preceding the 1969 sur- vey. Survey comparison. The surveys identified 105 species in 30 families (Table 1). The survey in 1969 found significantly more species (88 species) than that in 1968 (75 species; P < 0.0015, df = 3, 1- tailed paired Student’s t-test). The number of both early- and late-flowering species also increased sig- nificantly from 1968 to 1969 (P < 0.0001 and 0.05, respectively). Of the total number of species found combined from both years, a higher portion was unique to the 1969 survey than to the 1968 survey (29% vs. 16%). At individual sites, species unique to the 1969 survey accounted for up to more than half of the species found (Table 1). Early-flowering species dominated both surveys, but more so in 1969 than in 1968. The ratio of early- to late-flowering species was 1.15 in 1968 and 1.56 in 1969. For species unique to each sur- vey, the frequency distribution of flowering times differed between years (Fig. 4a). At each site, unique species were predominantly late-flowering in 1968 and early-flowering in 1969. The strength of this relationship varied between sites, ranging from statistically significant at the 0.05 level (Spooner Lake) to not significant (Marlette Creek) (Fig. 4a). The shift in dominance across sites from late to early species from 1968 to 1969 was strong- ly supported by the combined-flora analysis (x? = 4.3, P < 0.05, df = 1; Table 2) and the pooled- distribution analysis (x? = 10.6, P < 0.0015, df = 1), 1998] SPOONERS STN 0 oO io) Le 1968-69 , 1 SNOW DEPTH (cm) _ _ rae) r=) on ro) o o o oO o MONTHLY PPT (mm) OCT NOV DEC JAN FEB MAR APR MAY JUN MONTH Fic. 3. (a) Biweekly snow depth and (b) monthly pre- cipitation at Spooners Station for 1967-1968 and 1968— 1969 winters. X-axis ticks in (a) mark the 15th of each month. Precipitation data for June 1969 were not avail- able. Weak differences at the site level may have aris- en from the variety of habitats within sites and be- cause different habitats responded differently to cli- mate variability. Consequently, phenological shifts may have been stronger than observed, but were blurred because sampling at each site ranged across habitats. In the analysis by species’ habitat pooled across sites, both habitat classes showed a signifi- cant difference in flowering class frequencies be- tween years (Fig. 4b). The response was stronger for mesic habitats (yx? = 11.4, df = 1, P < 0.001) than for more xeric habitats (x? = 6.0, df = 1, P< 0.015). In addition, both habitats showed an in- crease from 1968 to 1969 in the number of unique species found, with xeric habitats having a larger increase (from 12 to 50 species pooled across sites) than mesic habitats (9 to 26). The effect of habitat on flowering class response was also reflected at the site level. Slaughterhouse Canyon and Spooner Lake sites, dominated by wet- ter habitats (meadow and riparian), had greater KITTEL: CLIMATE VARIABILITY AND HERBACEOUS SURVEYS 79 TABLE 1. NUMBER OF SPECIES, TOTAL AND UNIQUE, FOR THE 1968 AND 1969 SURVEYS AND FOR BOTH YEARS COM- BINED. The percent of unique species relative to the com- bined total is given in parentheses. 1968 Survey 1969 Survey Com- bined Unique Unique Site total Total to year ‘Total to year Skunk Harbor 49 25 6 (12%) 43 24 (49%) Slaughterhouse Canyon 71 41 7(13%) 64 30 (42%) Spooner Lake 7 45 14 (19%) 61 30 (40%) Marlette Creek 57 26 5 (9%) 52 31 (54%) All sites 105. 75 «17 (16%) ~=—s 88-—s- 30 (29%) flowering class shifts, as shown by the magnitude of corresponding x’’s (4.5 and 3.7, respectively; Fig. 4a), than Skunk Harbor (x? = 2.6) and Marlette Creek (x? = 0.5), dominated by conifer forest and montane chaparral. DISCUSSION Modes of phenological response to climate vari- ability. Phenological response of herbaceous spe- cies to variation in climate can be in terms of al- tered phenophase timing (displacement), duration (phase compression or extension), and amplitude (phase amplitude modulation) (Fig. 5). In displace- ment, timing of phenophases is delayed or ad- vanced, so that the flowering season occurs later or earlier in the year (Fig. 5a). On the other hand, if the growing season is shortened or lengthened, the duration of phenophases may be compressed or ex- tended, respectively (Fig. 5b). With compression, for example, the phenological cycle is completed more rapidly so that the flowering period starts later and ends earlier. In amplitude modulation, the tim- ing and length of the phenological cycle are not altered, but the number of individuals per species and number of species observed in a given pheno- phase is amplified or reduced (Fig. 5c). In ampli- fication, favorable conditions enhance plant growth, flower production, or other physiological and pop- ulation factors that increase the number of species observed in flower at a given time. Comparison of the 1968 and 1969 surveys sug- gests that both displacement and amplitude modu- lation were important in generating observed dif- ferences. In 1969, prolonged snowpack and later spring precipitation delayed plant phenologies so that substantially more early-flowering than late- flowering species were in flower during July. In 1968, below normal winter and spring precipitation and more growing degree-days advanced plant phe- nologies and increased the number of late-flowering species in blossom in July. Other studies have ob- served phenological displacement in winter-snow environments, where delayed snowpack melt de- layed initiation of growth and flowering in early 80 MADRONO FREQUENCY OF FLOWERING CLASSES [Vol. 45 a BY SITE SKUNK SLAUGHTER- SPOONER MARLETTE i HARBOR HOUSE CANYON LAKE CREEK c X?=2.6, p<.11 X?=3.7, p<.06 X?=4.5, p<.05 X?=0.5, n.s. oO. 7) Lu > e] z 5 LL. e) ow Lu a = = z 68 69 68 69 68 69 68 69 YEAR b BY HABITAT 30 X?=11.4, p<.001 POOLED # OF UNIQUE SPECIES 8 oe re ae 1968 1969 X?=6.0, p<.015 YEAR EARLY FLOWERING [fl LATE FLOWERING Fic. 4. Frequency distribution of early- and late-flowering classes for species unique to either 1968 or 1969 surveys (a) by site and (b) by mesic vs. xeric habitats pooled across sites. Chi-square values (x?) and significance levels (p) are given (df = 1). ns = not significant. In (b), the y-axis is number of unique species pooled (summed) across sites. season species (Holway and Ward 1965; Owen 1976). The greater number of both early- and late-flow- ering species found in flower in 1969 may have resulted from phenophase amplification. Better spring and early summer moisture conditions in that year possibly resulted in greater population sizes or greater flower production, increasing the likelihood of less common species being included in the 1969 survey. Jackson and Bliss (1984) also found phen- ophase amplification in a subalpine meadow where greater mid-season precipitation increased recruit- ment, growth, flowering, and seed set. Phenophase duration responses may have also occurred but cannot be demonstrated by the 1968— 1969 comparison because of the short, single sam- pling period used by the surveys. This response has been observed in other winter-snow environments including alpine communities (Holway and Ward 1965; Billings and Bliss 1959). In the extremes, a severely shortened growing season can lead to trun- cation of phenologies for late season species (Hol- way and Ward 1963), while an unusually prolonged growing season can lead to early season species completing all or part of a second cycle in a year (Holway and Ward 1965; Dickinson and Dodd 1976). While analysis of phenological responses in terms of these three modes is a useful approach for understanding observed variation as a function of KITTEL: CLIMATE VARIABILITY AND HERBACEOUS SURVEYS 81 1998] d ‘Yue sn1ofilusijpD snjnounuvy d ‘yOOH MuMosg DIUOaDg V Kern -V 2 AdO]L, Visasuod v1jazjuap V yooy (ARID “Y) Duunu DIpD| d duIOUL (NNN) Vadiuasojoy ‘dss "TJ vI101p DIA V qudaIn “Fy (CuUsg) snyjn1j1I9 snyjuvulT d DIYIPIU “IA QUIdID “AY VOYIpiu DsvAfIXVS d UOSUIGOY (ABID “V) SISUAPDAIU DISIMIT d ARID “Y Snjuap1gvAs UOWAIsSuad (DV "Ty aulivdp wnyvy ‘ped Cqpxsy) d DUddID “YY (Jasuaids) 1ossiuvys vDIUOW Vv Dipautsgjul ‘dss uoyug (JaT[e@AA) VIDUUId DIUIDINISAG Vv DUIDID “FY SnNIyIydaw snjnuip d ‘dey wnuviyjojinu wniuiydjag Vv "AOD (QUdIdID A) Mamasq sninuipy V Aa|pury vsopiaivd visuijjoy Vv ABID “YW wnuljnssads wnuosoiuy V UdARY (SP[SNOC) VJj4OJUOI DIUOSSIUDy) Aeq (491[9H “VW'V) Wwnaovjoia Vv [Tews (si9g) vIpofi]]Adsas adksapuvyy Vv ‘dss quvein ‘A 23 IRID ‘GV (WWsg) sapio1jis wnjjAydo]]y A[UO 6961 a KeAD “VY 1YIISND DIIMOLAA d AeIDn “YW wnyjuduow wn1jofis I I3A00H d 2 uosdor ppisi4s “eA ‘YWUag Saplosnip SKYyIDIS d "JT soul DjIOAKg d ‘NN Vapawosapuv vAOdSOsald ds ueulajoD winaonjndis “rea “T wnigiydup wnuosdAjog pieyooH d (puvig) pjopdwoos ‘dss ‘wya] vivispy DIl]aIvDYd /d BID “A (USING) DIDSaA83D Sisdowod] (Dad "T snyofissod uosodosvA] d DUIDID “FY Vasod DIADUUaJUy V ‘T]9U,.L (QUD0ID “q) suadsaqnd “1eA "YT wnoiuis841A wnipidaT d KID “VY Mamadg DI1asUY () V "YT vdva DIISSDAG AJUO Q96| 11qeH sarseds 11qeH sarsadg Ieok SULIOMOY We] SULIBMOY Ape plied eel “poyeoIpul Osye oie satoads (ZJ) paonponut ATqissod Jo (J) psonpoul ‘soUu0 SULIaMOY Joye soip ‘jetuudiod = /q pure ‘[etuusied = g ‘JeluUaIq = q ‘[enuUe = VY ‘:UDAIS SI (CHGH[ URUYOTY WO’) WQeY YIMOID (CHG) UCU SMOT[OJ sINJe[OUSWION “PoyTULO 219M spoliod ZuLIaMOY pouyap A[peoig YIM saloads “SALIS TTIW SSOYOV SSVID ONIAAMOTI-ALV] “SA ATAVY AD SATANNS 6961 YO SVG YAHLIA OL ANOING) SAIadS “7 ATAVL $2 MADRONO Displacement Compression/Extension Amplitude Modulation NUMBER OF SPECIES IN FLOWER TIME Fic. 5. Idealized phenological response curves. Pheno- phase (a) displacement, (b) compression and extension, and (c) amplitude modulation are illustrated in terms of number of species in flower during the course of a grow- ing season. climate, responses may be complex combinations of these modes arising from multiple and interac- tive effects of temperature and moisture on plant physiology (Walker et al. 1994; Holway and Ward 1965; Jackson and Bliss 1984). Additional com- plexity arises at the community level because re- sponses are an admixture of species-specific re- sponses moderated by species interactions (Jackson and Bliss 1984). In addition, climate conditions in one year have important carry-over effects to the next through seed set, belowground shoot meristem production in perennial herbs, and, in non-herba- ceous taxa, flower primordia production (Scott 1977; Jackson and Bliss 1984; Billings and Moo- ney 1968; Holway and Ward 1965). Habitat effects. Habitat type influenced the mag- nitude of phenological response to climate vari- ability. Phenologies were more delayed in less well drained sites, where higher soil moisture (and cool- er soil temperatures) likely persisted longer into the 1969 summer following heavy snowpack and late spring precipitation. In contrast, forest and chap- arral areas on slopes with poorly developed soils (primarily decomposed granite) likely dried out ear- lier, i.e. by mid-summer, than did wetter sites in both years. At these drier sites, greater winter snowpack, late spring rainfall, and associated cool- [Vol. 45 er temperatures in 1969 appear to have resulted in a large increase in the number of flowering species (indicating phase amplification), but a smaller delay in herbaceous phenologies (Fig. 4b). Key abiotic factors. Differences in survey floras and climate for 1968 and 1969 suggest that herba- ceous phenology and observed species richness in the Sierra Nevada montane region are controlled by snowpack amount and duration and early growing season precipitation and temperature. Studies of al- pine systems indicate that thawing of soil is the key control over initiation of growth and timing of sub- sequent phenophases (Billings and Bliss 1959; Bill- ings and Mooney 1968; Holway and Ward 1965; Ram et al. 1988). Increased soil moisture in high snowfall years is an additional control over phe- nology in these systems, increasing production and delaying senescence on drier sites (Holway and Ward 1965). Interaction among abiotic factors is critical to un- derstanding seasonal dynamics of montane herba- ceous communities because direct effects of changes in a factor may be either amplified or countered by indirect effects. For example, years with greater winter precipitation tend to have great- er and more persistent snowpack, cooler early sea- son temperatures, and greater growing season soil moisture. The negative impact of spring snowpack on growing season length can be more than offset by the positive effect of greater soil moisture (Walker et al. 1994). Implications for plant biodiversity surveys. Be- tween-year changes observed in this study dem- onstrate that surveys of herbaceous species need to be designed to account for phenological variability and population and community dynamics driven by climate variability. In rare and endangered plant surveys, effects of climatic variability must be con- sidered to avoid misleading assessments of popu- lations (Gruber et al. 1979; Clark and Dorn 1981; Tilman and Wedin 1991). Weather driven within- season shifts in phenology and population sizes can limit the utility of rapid “‘snapshot”’ biodiversity as- sessments. Likewise, because climate variability can force year-to-year shifts in community com- position (Tilman and El Haddi 1992; Pitt and Heady 1978; Borchert et al. 1991), such effects must be considered in long-term studies monitoring the status of plant communities facing threats from local to global environmental change. These responses emphasize the need to under- stand the role of climate variability in population and community dynamics that influence the pres- ervation of rare species and maintenance of com- munity diversity. Such environmental variation can play a key role in maintaining species diversity (Jackson and Bliss 1984; Grime 1973) or, in the case of extreme climate events, cause local extinc- tions of individual species, reducing community richness in the long term (Tilman and El Haddi 1998] 1992). The sensitivity of the central Sierra Nevada montane herbaceous species to interannual climate variability suggests that population, community, and ecosystem processes will respond rapidly to climate fluctuations and directional climate change. ACKNOWLEDGEMENTS Both surveys were supported by National Science Foundation Summer Science Training Program grants to Foresta Institute for Ocean and Mountain Studies, Carson City, Nevada. Thanks go to James V. A. Conkey, Richard Miller, James Gordon, and Karen Jensen for guidance and to the Executors of the George Whittell Estate for granting access for research on estate lands at Lake Tahoe. Special gratitude goes to Maya Miller, Richard Miller, and family (Washoe Pines Ranch, Franktown, Nevada) for their gen- erosity. Thanks also go to James A. Neilson for reviewing species identifications, Michael E Miller (Div. of Statis- tics, Univ. of California, Davis) for help with statistical design, and to Alan Carpenter, Tom Coon, Mel Knapp, Chris Knud-Hansen, Jack Major, Mark Otto, Rex Palmer, Buck Sanford, and Susan Spackman for valuable com- ments on analyses and earlier versions of the manuscript. Kay McElwain, Gaylynn Potemkin, Michele Nelson, and Suzanne Whitman assisted with manuscript preparation. Thanks to Frank Davis and Anna Sala for constructive comments in their reviews. The National Center for At- mospheric Research is sponsored by the National Science Foundation. LITERATURE CITED BERTILLER, M. B., M. P. IRISARRI, AND J. O. ARES. 1990. Phenology of Festuca pallescens in relation to topog- raphy in north-western Patagonia. Journal of Vege- tation Science 1:579—584. BILLINGS, W. D. AND L. C. BLIss. 1959. An alpine snow- bank environment and its effects on vegetation, plant development, and productivity. Ecology 40:388—397. BILLINGS, W. D. AND H. A. Mooney. 1968. The ecology of arctic and alpine plants. Biological Reviews 43: 481-529. BORCHERT, M., EF W. DAVIS, AND B. ALLEN-DIAZ. 1991. Environmental relationships of herbs in Blue Oak (Quercus douglasii) woodlands of central coastal Cal- ifornia. Madrono 38:249-—266. CLARK, T. W. AND R. D. Dorn. (eds.). 1981. Rare and endangered vascular plants and vertebrates of Wyo- ming, 2nd ed. Committee on Rare and Endangered Species of Jackson, WY. CONOVER, W. D. 1980. Practical non-parametric statistics, 2nd ed. John Wiley and Sons, New York. DICKINSON, C. E. AND J. L. Dopp. 1976. Phenological pat- tern in the shortgrass prairie. American Midland Nat- uralist 96:367—378. Gorpbon, J. AND C. M. DossMan, JR. 1968. An overall botanical survey of the Skunk Harbor area at Lake Tahoe, 1968. Unpublished report. Foresta Institute, Carson City, NV. GrimM_E, P. J. 1973. Control of species richness in herba- ceous vegetation. Journal of Environmental Manage- ment 1:151-—167. GRUBER, E. H., S. C. SEYER, M. A. STERN, AND C. E. WRIGHT. 1979. Rare, threatened, and endangered plant survey. Bureau of Land Management, Burns District, OR. HICKMAN, J. C. (ed.). 1993. The Jepson manual. Higher KITTEL: CLIMATE VARIABILITY AND HERBACEOUS SURVEYS 83 plants of California. University of California Press, Berkeley. Hotway, J. G. AND R. T. WARD. 1963. Snow and mel- twater effects in an area of Colorado alpine. Ameri- can Midland Naturalist 69:189-—197. Ho.tway, J. G. AND R. T. WARD. 1965. Phenology of al- pine plants in northern Colorado. Ecology 46:73-83. JACKSON, L. E. AND L. C. BLIss. 1984. Phenology and water relations of three plant life forms in a dry tree- line meadow. Ecology 65:1302-—1314. KITTEL, T. 1969. Comparison of the 1968 and 1969 wild- flower surveys. Pp. 57-64 in K. Jensen (ed.), NSF Summer Science Training Program 1969 yearbook. Foresta Institute, Carson City, Nevada. LIETH, H. (ed.). 1974. Phenology and seasonality model- ing. Springer-Verlag, New York. Munz, P. A. AND D. D. Keck. 1973. A California flora and supplement. University of California Press, Berkeley. OrRNDuFF, R. 1974. Introduction to California plant life. University of California Press, Berkeley. Owen, H. E. 1976. Phenological development of herba- ceous plants in relation to snowmelt date. Pp 323- 341 in H. W. Steinhoff, and J. D. Ives (eds.), Ecolog- ical impacts of snow augmentation in the San Juan Mountains, Colorado. U.S. Dept. of Interior, Bureau of Reclamation, Denver, CO. Pitt, M. D. AND H. E HEApy. 1978. Responses of annual vegetation to temperature and rainfall patterns in northern California. Ecology 59:336—350. PykE, C. B. 1972. Some meteorological aspects of the seasonal distribution of precipitation in the western United States and Baja California. Water Resources Center, University of California, Los Angeles. Con- tribution no. 139. Ram, J., S. P. SINGH, AND J. S. SINGH. 1988. Community level phenology of grassland above treeline in central Hima- laya, India. Arctic and Alpine Research 20:325-332. ROPELEWSKI, C. EF AND M. S. HALPERT. 1986. North Amer- ican precipitation and temperature patterns associated with the El Nino/Southern Oscillation (ENSO). Monthly Weather Review 114:2352—2362. RUNDEL, P. W., D. J. PARSONS, AND D. T. GORDON. 1980. Montane and subalpine vegetation of the Sierra Ne- vada and Cascade ranges. Pp 559-599 in M. G. Bar- bour and J. Major, Terrestrial vegetation of California. John Wiley and Sons, New York. SCHEMSKE, D. W., M. E WILLSON, M. N. MELAmpPY, L. J. MILLER, L. VERNER, K. M. SCHEMSKE, AND L. B. BEST. 1978. Flowering ecology of some spring wood- land herbs. Ecology 59:35 1-366. Scott, D. 1977. Plant ecology above timber line on Mt. Ruapehu, North Island, New Zealand. II. Climate and monthly growth of five species. New Zealand Journal of Botany 15:295-—310. SHEAFFER, J. D. AND E. R. REITER. 1985. Influence of Pa- cific sea surface temperatures on seasonal precipita- tion over the western plateau of the United States. Archives for Meteorology, Geophysics, and Biocli- matology, Series A 34:111—130. SMITH, G. L. 1973. A flora of the Tahoe Basin and neigh- boring areas. Wasmann Journal of Biology 31:1—231. SPSS AND M. J. NoruSis. 1986. SPSS/PC+ FOR THE IBM PC/XT/AT. SPSS, INc., CHICAGO. TILMAN, D. AND D. WEDIN. 1991. Oscillations and chaos in the dynamics of a perennial grass. Nature 353:653— 655; TILMAN, D. AND A. EL Happr. 1992. Drought and biodi- versity in grasslands. Oecologia 89:257—264. 84 MADRONO U.S. ENVIRONMENTAL DATA SERVICE. 1967, 1968, 1969a. Climatological data. United States by sections: Ne- vada. Dept. of Commerce, Washington, D.C. Vols. 82, 83, and 84. U.S. ENVIRONMENTAL DATA SERVICE. 1969b, 1970. Stor- age-gage precipitation data for western United States, Vols. 13 and 14. Dept. of Commerce, Asheville, NC. U.S. ENVIRONMENTAL DATA SERVICE. 1973. Monthly nor- mals of temperature, precipitation, and heating and cooling degree-days 1941-1970: California. Clima- tography of the United States no. 81. Dept. of Com- merce, Asheville, NC. [Vol. 45 U.S. WEATHER BUREAU. 1965. Decennial census of United States climate. Climate summary of the United States, supplement for 1951 through 1960. Nevada. Clima- tography of the United States no. 86-22. Dept. of Commerce, Washington, D.C. WALKER, M. D., P. J. WEBBER, E. H. ARNOLD, AND D. EBERT-MAy. 1994. Effects of interannual climate variation on aboveground phytomass in alpine vege- tation. Ecology 75:393—408. WALTER, H. 1985. Vegetation of the Earth and ecological systems of the geo-biosphere, 3rd English ed. Spring- er-Verlag, New York. MaprRONO, Vol. 45, No. 1, pp. 85-87, 1998 NOTEWORTHY COLLECTIONS CALIFORNIA QUERCUS ENGELMANNII Greene (FAGACEAE).—Orange Co., San Joaquin Hills, upper Coyote Canyon, ca. 4 km from coastline along a drainage bottom in a small valley just north of Signal Peak. Found in association with Quer- cus agrifolia var. agrifolia, Q. berberidifolia, Rhus inte- grifolia, and Heteromeles arbutifolia in southern oak ri- parian woodland, adjacent to annual (non-native) grass- land. Hybrids involving Q. berberidifolia and Q. engel- mannii were observed in close proximity to the solitary Q. engelmannii (Engelmann oak); ca. 33°37'N, 117°49'W, elev. ca. 250 m, 11 June 1991, R. A. Erickson s.n. (RSA), verified by E M. Roberts, Jr. Additional specimens ob- tained on 17 Oct 1996 (J. E. Harrison 500, RSA). Previous knowledge. Recognized distribution is gener- ally from southern base of San Gabriel Mts. in eastern Los Angeles Co. south to northwestern Baja California (F M. Roberts, Jr., Illustrated Guide to the Oaks of the Southern Californian Floristic Province, pp. 60-63, FE M. Roberts Publ., Encinitas, CA, 1995). Overall distribution is patchy, with several disjunct populations of varying size and configuration; the vast majority of individuals are found in interior foothills and valleys of western, cismon- tane San Diego Co. Scattered occurrences in eastern Or- ange Co. are known from Casper’s Regional Park and on private lands (1.e., Rancho Mission Viejo) (KE M. Roberts, Jr., Rare and endangered plants of Orange County, Cros- sosoma 16(2):3—12, 1990). Extent to which distribution of this species may have been influenced by Native Ameri- cans is unknown, but is potentially significant. Significance. First record in San Joaquin Hills of Or- ange Co; next closest population ca. 20 km east. This specimen documents the San Joaquin Hills distribution of Q. engelmannii mapped by Roberts (1995). Following the October 1993 Laguna Canyon Fire and recent construction activities, significant hybrids are apparently no longer present. Remaining Engelmann oak is charred, with the trunk split to the base, and the entire tree has fallen. Re- sprouting up to 2.5 m in height has occurred at base of remaining trunk (J. E. Harrison 500). Roberts (1990) not- ed that non-hybrid Engelmann oaks are very rare in Or- ange Co. Currently, Q. engelmannii has no federal or State status, but is designated as “‘List 4: Plants of Limited Dis- tribution—A Watch List’? by the California Native Plant Society. —JAMES E. HARRISON and RICHARD A. ERICKSON, LSA Associates, Inc., One Park Plaza, Suite 500, Irvine, CA 92614; FRED M. ROBERTS, JR., U.S. Fish and Wildlife Ser- vice, 2730 Loker Avenue West, Carlsbad, CA 92008. CALIFORNIA MADIA MADIOIDES (Nutt.) E. Greene (ASTERACEAE).—San Diego Co., NW Palomar Mountains, Agua Tibia Moun- tains, Cleveland National Forest, Agua Tibia Wilderness Area, W slope of Agua Tibia Mountain, SSW of the Cros- ley Saddle, N branch of upper Agua Tibia Creek, E of large drainage coming from the W before the confluences of branches of the Creek, N of the Wilderness boundary, T9S RIW, NE/4, SW/4, sect. 15, elev. 3120 ft, 13 June 1995, Darin L. Banks & Steve Boyd 0618 (RSA), verified by Dr. B. Baldwin (JEPS). Previous knowledge. Known from redwood forest, north coastal coniferous forest and mixed coniferous for- est of the outer southern coast ranges and northern high Sierra Nevada, north to British Columbia (J. C. Hickman [ed.], The Jepson manual: higher plants of California, 1993; P. A. Munz and D. D. Keck, A California flora, 1959). Significance. First report for California south of San Luis Obispo Co., and first record for San Diego Co. Plants are associated with a relictual Arbutus menziesii popula- tion in the northern Palomar Mountains. SENECIO ASTEPHANUS E. Greene (ASTERACEAE).—San Di- ego Co., NW Palomar Mountains, Agua Tibia Mountains, Cleveland National Forest, Agua Tibia Wilderness, E slope of Agua Tibia Mountain, along the Arroyo Seco Drainage, SE of the Crosley Trail, where the two forks of upper Arroyo Seco converge in Section 11, T9S R1W, SE/ 4, NE/4, sect. 11, elev. 3120 ft, 27 June 1995, Darin L. Banks & Steve Boyd 0723 (RSA); Palomar Range, Agua Tibia Wilderness Area, northeastern flank of Agua Tibia Mountain, Arroyo Seco tributary at the north base of Ea- gle Crag, from the Wildhorse Trail east to the main trunk of Arroyo Seco, T9S RIW S'’% NY sect. 11, Near 33°24'37"N 116°57'30"W, elev. ca. 3200 ft, local in moist shaded side draw on north-facing slope in understory of Quercus agrifolia, near center of section 11, 6 April 1995, Steve Boyd & Darin L. Banks 8448 (RSA). Previous knowledge. Known from chaparral and steep rocky slopes of the San Gabriel and San Bernardino Mountains, west to Ventura Co. and north to San Luis Obispo Co. (J. C. Hickman 1993, loc. cit.; P. A. Munz, A flora of Southern California, 1974). Significance. First reports for California south of the Transverse Ranges and first records for San Diego County and the Palomar Mountains. FESTUCA CALIFORNICA Vasey var. PARISH (Piper) A. Hitchc. (POACEAE).—San Diego Co., NW Palomar Moun- tains, Cleveland National Forest, Agua Tibia Wilderness Area, N peak of Agua Tibia Mountain on the NE corner of the peak, just S of the Riverside County boundary, in a very steep bowl shaped depression on the N flank of Agua Tibia Mountain, T9S RIW, SE/4, NW/4, sect. 4, 33°25'18"N—116°57'24", elev. 4200 ft, 1 June 1995, Darin L. Banks & Steve Boyd 0509 (RSA); Palomar Range, Agua Tibia Wilderness Area, western crest of Agua Tibia Mountain at the head of a steep draw in the Pechanga Creek watershed, just northwest of the large Quercus agri- folia woodland about the junction of the Palomar Divide and Dripping Springs trails, T9S RIW, SW%4, NW, sect. 4, near 33°25'10"N 116°59'38’W, elev. ca. 4500 ft, locally common on more mesic exposures with some afternoon 86 MADRONO sun, 25 April 1995, Steve Boyd 8508 (RSA); Cleveland National Forest, Agua Tibia Wilderness Area, E face of Eagle Crag, S of upper Arroyo Seco, W of the Cutca Valley, along the Palomar-Magee Trail, T9S R1W, NE/4, SW/4, sect. 14, elev. 4600 ft, 10 May 1995, Darin L. Banks & Steve Boyd 0429 (RSA); NW Palomar Moun- tains, Agua Tibia Mountains, Cleveland National Forest, Agua Tibia Wilderness Area, NE face of Eagle Crag, SE of the Crosley Saddle, S of the Cutca Trail along drainage that parallels the trail, E of upper Arroyo Seco, T9S R1IW, SE/4, SE/4, sect. 14, elev. 4520 ft, 15 June 1995, Darin L. Banks & Steve Boyd 0684 (RSA). Previous knowledge. Considered endemic to the San Bernardino Mountains, where known from dry chaparral and yellow pine forest. (P. A. Munz 1974, loc. cit.). Significance. First report for California south of the San Bernardino mountains and first record for San Diego County and the Palomar Mountains. POLYPOGON MARITIMUS Willd. (POACEAE).—Riverside Co., Magnesium Canyon [possibly Magnesia Springs Canyon], north base of Santa Rosa Mountains, 29 May 1955, C. K. Buechner C73 (RSA); NW Palomar Moun- tains; Agua Tibia Mountains; Cleveland National Forest, Agua Tibia Wilderness Area, along lower Arroyo Seco, S of the Dripping Springs Campground, along the narrow benches where Arroyo Seco turns to the E, just W of the Metasedimentary hills, T8S RIW, SW%4, NE, sect. 27, elev. 1720 ft, 17 October 1995, Darin L. Banks & Steve Boyd O816B (RSA).—San Diego Co., NW _ Palomar Mountains, Agua Tibia Mountains, Cleveland National Forest, Agua Tibia Wilderness Area, E slope of Agua Tib- ia Mountain, SW of the Crosley Homestead, W of Arroyo Seco Drainage, along the Wildhorse Trail, T9S R1W, NW/ 4, SE/4, sect. 2, 33°25'15”N 116°57'02"W, elev. 2790 ft, 12 June 1995, Darin L. Banks & Steve Boyd OS588 (RSA); NW Palomar Mountains, Agua Tibia Mountains, Cleve- land National Forest, along the Cutca Trail, E of Cutca Valley, approximately 1.3 km W of the Aguanga Trail junction, T9S RIE, SE/4, NW/4, sect. 16, 33°23'39"N 116°53'27"W, elev. 3540 ft, 27 October 1995, Darin L. Banks & Steve Boyd 0838 (RSA). Previous knowledge. Known from moist areas of north- western California, east to the Sierra Nevada foothills, and south to the San Francisco Bay area and throughout the San Joaquin Valley. Introduced from Europe and Africa (J. C. Hickman [ed.] 1993, loc. cit.; P. A. Munz and D. D. Keck 1959, loc. cit.). Significance. First reports for California south of the Transverse Ranges and first records for Riverside and San Diego Counties. Attempts to identify this taxon using A flora of Southern California, Munz (1974), will result in misidentification as Polypogon monspeliensis (L.) Desf. —DARIN L. BANKS and STEVE Boypb, Herbarium, Rancho Santa Ana Botanic Garden, 1500 N. College Avenue, Claremont, CA 91711. OREGON LOMATIUM FOENICULACEUM (Nutt.) J. M. Coult. & Rose SSp. FIMBRIATUM Theob. (Apiaceae).—Malheur Co., Rome ash beds, 4 miles west of Rome on Hwy 95. T31S R41E sect. 32 SEY%, (42°49'5"N, 117°42'45”"W), growing on heavy clay outwash south of a barren gray ash outcrop, [Vol. 45 1120 m, 26 April 1997, EF Wernette 28 with D. Mansfield and Field Botany class (CIC, OSC). Previous knowledge. Well distributed across southcen- tral Nevada from eastern California to western Utah. Significance. First record for Oregon. Population is ca. 400 km north of northern most Nevada population. —DONALD H. MANSFIELD, Department of Biology, Al- bertson College of Idaho, 2112 Cleveland Blvd., Caldwell, ID 83605. OREGON CAREX CRAWFORDII Fernald (Cyperaceae)—Coos Co., OR, SW of Floras Lake, 2.5 km W of U.S. Route 101, elev. 15 m, T31S RIS5W S20, 21 August 1997, P. F. Zika 13357 & B. Wilson (OSC; MICH), naturalized weed in cranberry crop fields, on sandy banks and ditch margins, with Rubus ursinus, Spergula arvense, and Vaccinium ma- crocarpon; Jackson Co., OR, Spruce Lake, 13 air km WNW of Wizard Island (Crater Lake), Crater Lake Na- tional Park, W slope of Cascade Mts., elev. 1450 m, T30S R4E S12 NW¥%, 30 August 1995, P. F. Zika 12703 (Crater Lake National Park Herbarium [hereafter abbreviated CLNP]; OSC), grassy receding shorelines on moist sunny soil, with Agrostis hyemalis var. scabra and Carex lentic- ularis var. impressa. Previous knowledge. A transcontinental boreal species, C. crawfordii ranges south in Washington State to Sno- qualmie Pass (Hitchcock et al. Vascular Plants of the Pa- cific Northwest, Part 1., Univ. Washington Press, Seattle, 1969.), 500 km N of Crater Lake. Significance. First records of Crawford’s sedge for Or- egon. The coastal site appears to be a weed introduced with rooted cranberry stock from a Great Lakes or New England source of commercial Vaccinium. FESTUCA OVINA L. 8s. str. (Poaceae)—Klamath Co., OR, Mazama Campground, 5 air km NE of Union Peak, Crater Lake National Park, E slope of Cascade Mts., elev. 1830 m, T31S RS5E S13 SW%, 24 August 1994, Wilson (7526), Zika & Kuykendall (CLNP, OSC, det. B. Wilson), dry sun- ny gravel roadsides and parking lot margins, weed among Sitanion hystrix, Carex inops. Previous knowledge. At one time, a broad F. ovina spe- cies concept included what are now considered five native Oregon taxa (F. brachyphylla Schultes & Schultes, F. ida- hoensis Elmer var. idahoensis, F. idahoensis var. roemeri Pavlick, F. occidentalis Hook., and F. saximontana Rydb.). However, F. ovina s. str. (sheep fescue) is a Eur- asian bunchgrass cultivated in the Pacific Northwest. It occasionally escapes and may become naturalized. The species is probably under reported due to identification difficulties. Significance. First record as an escaped plant for Ore- gon. FESTUCA TRACHYPHYLLA (Hackel) Krajina—Klamath Co., OR, Route 62, 3.9 air km S of Crater Peak, Crater Lake National Park, E slope of Cascade Mts., elev. 1570 m, T32S R6E S10 SE%, 9 July 1995, Zika 12497 (CLNP, OSC), sunny dry gravelly roadside weed, with Carex sub- fusca, Collomia tinctoria, Madia minima, Poa secunda, and Sitanion hystrix. 1998] NOTEWORTHY COLLECTIONS 87 Previous knowledge. A bunchgrass native to Eurasia, F. trachyphylla (hard fescue) is cultivated in the Pacific Northwest and occasionally escapes. This taxon should not be confused with F. rubra var. trichophylla Gaudin. See Darbyshire and Pavlik (Phytologia 82:73-78, 1997) for justification for use of the name F. trachyphylla for this taxon rather than the later names F. longifolia Thuill. or F. brevipila Tracey. Significance. First record for Oregon. ACKNOWLEDGMENTS We are grateful to the Native Plant Society of Oregon, National Park Service, the Oregon Nature Conservancy, and the Crater Lake Natural History Association for par- tial funding of field work in Crater Lake National Park. —PETER F. ZIKA and BARBARA WILSON, Carex Working Group, Herbarium, Dept. of Botany and Plant Pathology, Oregon State Univ., Corvallis, OR, 97331. MAbDRONO, Vol. 45, No. 1, pp. 88-90, 1998 OBITUARY LAURAMAY TINSLEY DEMPSTER (1905-1997) Lauramay Tinsley Dempster, who died at home in Or- inda, California, on November 14, 1997, started her bo- tanical career at the University of California, Berkeley as a freshman when she was sixteen and finished this career at the same institution almost eighty years later! Her ac- tive life—a balancing of personal and professional—is in- deed a salutory reflection of distinguished women in sci- ence during the twentieth century. Born on May 11, 1905, in El Paso, Texas, daughter of creative parents (her mother a writer, her father a telescope builder), she grew up in the San Francisco Bay area, with an early interest in natural history. Encouraged by her mother and grandmother, she became a botany student at Berkeley in 1921 and soon was active both in and out of class, joining enthusiastic fellow students to create the University’s first field botany club, Calypso. The club would draw faculty and pupils together for memorable junkets into the natural bounty of California for almost two decades, an antidote for what William Morton Whee- ler once called ‘“‘the dry rot of biology.” Lauramay’s undergraduate course in ‘“‘phaenogamic botany’? from Professor Willis Jepson about 1923 launched an academic relationship which would flourish on and off for more than two decades. Upon graduation in 1925, she was urged by Jepson, who was working on the mustard family for his monumental Flora, to consider doing a master’s with him, focusing on Lepidium for a thesis, undoubtedly not the most exciting genus in the botanical world. But the challenge resulted in a superb, comprehensive monograph, which contained many mati- culous line drawings (photographed for the thesis by her father) and was a portent for this budding botanical artist. Said Jepson, when Lauramay completed her master’s work, “‘If it were up to me, I would give you a doctor’s degree.’ Lauramay’s lifetime regret was that she didn’t published that thesis—*‘a terrible mistake’’; but Jepson’s regret would be rectified by Lauramay’s eventual profes- sional accomplishments. Meanwhile, while a university senior Lauramay had met and fallen in love with her future husband, Everett Ross Dempster, an engineering student and native of San Francisco, two years her elder. The Dempster clan had a retreat at Inverness, and the Dempster family young peo- ple shared Lauramay’s love for the out-of-doors. Indeed, part of the summer before she started her graduate work had been spent on an extended Dempster hiking hagiera in the southern Sierra. Miss Tinsley received her master of arts degree early in 1927 and then served as a research assistant for the rest of the academic year. Everett would not complete his en- gineering program until 1928, but already the couple was looking forward to a wedding in the autumn of 1927. Lauramay decided to explore the science teaching profes- sion in the interim. During the summer of 1927 she taught biology at a Fresno State College field session at Hunting- ton Lake (a job which had been turned down by Herbert Mason). She was the youngest teacher there, expressed some concerns about her abilities, but at least reveled in the botany of the mountain area, adding to her personal herbarium and wondering if she could do some collecting for Professor Jepson. On October 8, 1927, Lauramay Tinsley and Everett Dempster were married at Berkeley’s First Unitarian Church; and Lauramay, who had always had a distinct disgust for housework, began the combination of being a housewife and working in biology, continuing to assist Jepson with his research as well as creating a new house- hold. By early May of 1928, she was perceptibly ready to conclude, in a note to Dr. Jepson: ‘“‘How could a married lady be a botanist even if so inclined?” The ensuing academic-year (1928-1929), as Everett pursued an engineering job with Magnavox in Oakland, Lauramay tackled teaching again, this time a beginning botany course for seven young people at Cora Williams Institute on College Avenue in Berkeley. She flunked six of the seven students both semesters, incurring the wrath of administrators and parents alike. But she was later vin- dicated, since the six who flunked went on to the Univer- sity of California, where they also flunked. Lauramay’s ultimate reaction to pedigogy after these teaching experi- ences: ‘“‘teacherly ambitions do not in any way harmonize with my plan of life.” Everett’s electrical engineering position with Magnavox Company now took the young couple to Chicago in late May of 1929, where they were “‘installed”’ in a tiny apart- ment in the heart of a big city. Lauramay complained to Jepson that “‘There is very little here to tease the eye of a botanist, though the new weeds in vacant lots might prove interesting.’’ She thought of visiting the University of Chicago and Northwestern University Botany Depart- ments in the event that there might be some botanical occupation, and Jepson suggested she go to the Depart- ment of Botany at the Field Museum, where Jepson’s ca- sual acquaintance Paul Stanley was curator. Alas, although she enjoyed the Field Museum exhibits, she was bluntly told that ‘“‘there was no room for me...” The only appealing Middle West experience came dur- ing a vacation trip with Everett by second-hand canoe northward in Wisconsin and into Lake Michigan, where the canoe eventually swamped and the couple had to re- turn to Chicago by train. Also, Lauramay visited New York and New Jersey, finding the countryside in the latter state superior in every way to IIliois. But, as she reflected, ‘Tam more than ever impressed with the uniqueness of California, geographically and climatologically. Com- pared with it, all the rest of the country that I have seen is nearly monotonously alike. Simultaneously, my longing for California increases .... Hoping to come back some day.” Everett’s engineering job would shift him from Chicago to Fort Wayne, Indiana, back to Chicago, and to England for a time. But eventually he and Lauramay did return to Berkeley, in 1933, where, of all unexpected things, Everett gave up his engineering profession and decided to become a biologist like his wife. He began taking the necessary background courses and started his pursuit of a Ph.D. in genetics, meanwhile commencing his own teaching career as a botanical assistant in 1935. Dempster received his Ph.D. in 1941 and in his botanical career would go on to become chairman of the Department of Genetics at Berke- ley for many years. 1998] Thus it was that Lauramay Tinsley Dempster again be- came a research assistant to Professor Willis Lynn Jepson in 1933. To be specific, she was his “‘botanical dissector and preparateur of details for drawings,” working part- time for him on the Flora and to a greater extent as some- what of a personal secretary, especially when Jepson was off campus, handling his correspondence, reading proof, toiling over the index of Madrojno, visiting the library and herbarium, coping with students, visitors, and other assis- tants. Incidentally, Jepson was continually piqued beyond measure that he, supposedly unlike his academic col- leagues, never had a full-time secretary. Unfortunately, this working arrangement which was beneficial to both Jepson and Lauramay came to an end by early 1936 as Lauramay anticipated the birth of her second child (she had lost her first several years earlier). Although over ensuing years, until World War II, Laur- amay did keep in contact with Jepson, and indeed casually did little jobs for him such as proof reading sections of the Flora as they appeared, she increasingly devoted her life to Everett and their growing family. As she wrote Dr. Jepson in 1939, ‘‘My botany is at present petty and quite avocational ... but has not been dropped altogether.”’ While Everett Dempster was advancing at the Univer- sity of California from instructor in 1941 to assistant pro- fessor of genetics in 1944, Lauramay again commenced helping Jepson with his Flora, ‘‘far from an expert typist but improving,”’ spending about an hour every evening, typing and proof-reading. As she informed him, ‘‘Wel- come though the money is [from Jepson’s university grant], I have enjoyed feeling that I was helping you in a way that relatively few people would be able to do. Indeed it is a pleasure to know that I have contributed, however slightly to so great a work.” A friend noted at this time that although Lauramay did “have her hands full’? with her family, she “‘wouldn’t do anything for anybody else but for Jepson, she’d give him whatever time she could manage.”’ Everett, incidentally, had not been too happy about his wife working for Jepson, even part-time, be- cause her “‘job at home is more than man-sized already.” Willis Lynn Jepson died in 1946, his great Flora un- completed; but provisions of his will provided for con- tinuing pursuit of the project. And on October 8, 1951, Lauramay was employed by the University of California as Herbarium Botanist, part-time, fittingly on the Jepson Endowment Fund, a position which she would occupy un- til the summer of 1963. With the establishment of the Jepson Herbarium in 1951, she became the assistant of the first curator, Dr. Rimo Bacigalupi, and was initially charged with organizing Jepson’s botanical books into the herbarium library. Later, between 1959 and 1967, she also received an appointment as Research Geneticist at 60% of full time. During these years, Lauramay and Everett’s own family was growing up, but an “‘extended family” and befriended friends (Everett was always known for putting students “first’), initially at the house in Berkeley and then in the new oak-woodland suburban home at Orinda and the In- verness retreat, made continuing demands on the Demps- ters Their residence, as a friend once commented, was “like Grand Central Station!” Nevertheless, despite familial distractions, Lauramay was increasingly able to pursue what through the years she had longed to pursue, the professional life of a taxo- nomic botanist. She had been continuing her work on re- vision of Umbelliferae, and, thinking ‘‘there is great dan- ger in devoting all of one’s attentions to a single group,”’ was also pursuing Hydrophyllaceae, had worked on OBITUARY 89 Scrophulariaceae, Solanaceae, and Gilia. In 1958 she pub- lished her first major paper, ‘““Dimorphism in the fruits of Plectritis [a genus in the Verbenaceae family] and its tax- onomic implications”, in Brittonia. It was Dr. Bacigalupi who first suggested that Lauramay tackle the revision of that difficult Rubiaceae genus Galium, and she would eventually become world famous for this research. Shortly before Jepson died, Lauramay had written him that ‘“‘All my life I have craved land, as many people do the seas. In fact there is no type of land that I do not love, and the plants upon it are only its most charming mani- festations.”’ Starting in the 1950’s, alone or with other companions, she was to see lands—and their natural his- tory—that encompassed the Biosphere. It was across Af- rica, Europe from France to Sweden, the Alaskan highway through Canada to Fairbanks and Mt. McKinley, the di- versity of Australia from the tropics of Darwin to the Red Center at Alice, the alpine tundra of the Snowy Mountains and the coastal rainforests of Queensland; New Zealand, Malayasia ... and of course North, Central, and South America. In 1988 the continent of Antarctica was added to Lauramay’s geographical roster. At home in California her botanical research and writ- ing accelerated. From 1959-1964, with National Science Foundation grants and especially in close association with Ledyard Stebbins, she was almost continually in the field spring-summer-and-fall the length of the Golden State and elsewhere, working on Galium. Publications began ap- pearing, in Madrono, University of California Publica- tions in Botany, Allertonia, Phytologia, Boletin de la So- ciedad Botanica de Mexico, Great Basin Naturalist, Field- iana, Brittonia, Leaflets of Western Botany, Sida. Her bi- ographical sketch appeared in the 1965 edition of American Men of Science. In 1968 Lauramay Dempster was designated as a Research Associate with the Jepson Herbarium and Library, an appointment which she would hold until her death. Commented Robert Ornduff with re- spect to this unpaid appointment: “‘Mrs. Dempster holds an M.A. degree from this institution but does not have Ph.D. Nevertheless, I believe that the quality and volume of her research work are equivalent to that of a person holding a doctorate, and exceed that of a number of doc- toral products of our department.’’ Ornduff added that much of her research was carried out without financial support. When the new Jepson Manual appeared in 1993, Lauramay was author for not only Rubiaceae, but also Apocynaceae, Caprifoliaceae, most of Valerianaceae and Convolvulaceae, not to mention the genus Lewisia. Her definitive treatment of Galium for the Flora of North America was completed in 1996, seventy years after she was doing research on Lepidium for her master’s thesis. The shy young co-ed in the photograph of a Calypso Club excursion in the mid-1920’s, Lauramay Tinsley then, would over the decades become more than a distinguished taxonomic botanist. For one who detested housework, Lauramay learned to put on a memorable Thanksgiving spread for 30 people. But on the other hand, when Dr Jepson came to her parents’ home for dinner, Lauramay spilled a platter of bisquits on his head. Her tenacity at driving her car from Orinda through the tunnel to the Uni- versity when over ninety years old was probably a latter- day manifestation of her youthful love for motorcycles. Yet despite spending much leisure time along the northern California coast, she was not partial to boats. No wonder that when she went overboard while floating the Colorado River through the Grand Canyon, her husband had to pull her back aboard by the hair. Although quiet spoken and self-contained, she was chairman of the Society of Amer- 90 MADRONO ican Geographers for three years, an active member of University Women’s Club, California Botanical Society, American Society of Plant Taxonomists, and the Amphion Club ... not to mention a founding member of the Ca- lypso Club. Lauramay was accomplished at playing the recorder and was remembered for her duets on the oboe with a dog. Her plant anatomy drawings were detailed and ex- quisite; in the Orinda home her colorful wall-sized murals of tropical rainforest and of flowers were overwhelming. Although many an hour and day were spent indoors over a typewriter, a manuscript, a dried plant specimen, Laur- amay was an outdoorswoman, a world traveler, a com- mitted conservationist of the natural world. Although she refused to snoop around another botany professor’s office for Jepson, on the other hand Jepson rarely offered her [Vol. 45 credit in publication for her research contributions. But she forgave him. “‘He was always ready to give me atten- tion if I asked for it, but for the most part he left me alone. Left me to use my own judgement. And I think he couldn’t have done better!” When Lauramay Dempster died this past November, her surviving contemporaries were few. Many who had re- ceived a coveted Ph.D. diploma during those bygone days, and were entitled to spend a lifetime at research and/or teaching in the hallowed Halls of Ivy, are now both gone and forgotten. Miss Tinsley need no longer regret that she couldn’t have been one of them. Dividing her many years between a pursuit of botany and a commitment to family and society, she succeeded in transcending them all. —RICHARD G. BEIDLEMAN ANNOUNCEMENT CALIFORNIA BOTANICAL SOCIETY 18™ GRADUATE STUDENT MEETINGS SATURDAY, 20 FEBRUARY 1999 CALIFORNIA POLYTECHNIC STATE UNIVERSITY SAN LUIS OBISPO These meetings provide an opportunity for current and recent graduate students in all aspects of plant science to meet and exchange ideas regarding proposed research, research in progress, or completed research. Interested parties (not limited to graduate students) are encouraged to attend the presentations. The California Botanical Society Annual Banquet will be held following the graduate student meetings. Guest Speaker: Dr. Sherwin Carlquist Santa Barbara Botanic Garden For further information, contact: Mary Lea Biological Sciences Department California Polytechnic State University San Luis Obispo, CA 93407 (805) 756-2950 email: mlea@polymail.calpoly.edu OR: Dennis Wall - Graduate Student Representative Department of Integrative Biology Valley Life Sciences Building University of California, Berkeley Berkeley, CA 94720 (510) 643-7008 email: dpwall @socrates.berkeley.edu November 10, 1998 RESOLUTION BY THE CALIFORNIA BOTANICAL SOCIETY ON TRANSPLANTATION Whereas: I. native plants and their habitats have experienced increased threat from development and other land use practices on public and private lands and, II. relatively few permanently protected areas exist that support natural populations of native plants and their habitats, and, III. studies have shown that transplant attempts of native plants back into habitat are mostly unsuccessful and long-term monitoring of such transplants are currently completely inadequate. The California Botanical Society strongly urges all appropriate agencies, organizations, and individuals involved with the protection of native plants, including both common and rare taxa, for the purpose of maintaining plant diversity to: 4. develop and implement policies that make protection of natural populations of native plants the first and foremost priority for mitigation and other preservation activities, and, restrict the use of transplantation for mitigation of native plants and populations as a last resort and least preferred option for protection, and, when transplant mitigation is chosen as the last resort option, that a scientific study of known habitat conditions and species biology be completed prior to transplantation, and, a quantitative demographic monitoring study of mitigation transplants be implemented annually for a period of not less than seven years. Volume 45, Number 1, pages 1—92, published 8 February 1999 SUBSCRIPTIONS—MEMBERSHIP Membership in the California Botanical Society is open to individuals ($27 per year; family $30 per year; emeritus $17 per year; students $17 per year for a maximum of 7 years). Late fees may be assessed. Members of the Society receive Maprono free. Institutional subscriptions to MADRONO are available ($60). Membership is based on a calen- dar year only. Life memberships are $540. Applications for membership (including dues), orders for subscriptions, and renewal payments should be sent to the Treasurer. Requests and rates for back issues, changes of address, and undelivered copies of MADRONO should be sent to the Corresponding Secretary. INFORMATION FOR CONTRIBUTORS Manuscripts submitted for publication in MAprRoNo should be sent to the editor. It is preferred that all authors be members of the California Botanical Society. Manuscripts by authors having outstanding page charges will not be sent for review. Manuscripts may be submitted in English or Spanish. English-language manuscripts dealing with taxa or topics of Latin America and Spanish-language manuscripts must have a Spanish RESUMEN and an English ABSTRACT. Manuscripts and review copies of illustrations must be submitted in triplicate for all articles and short items (NOTES, NOTEWORTHY COLLECTIONS, POINTS OF VIEW, etc.). Follow the format used in recent issues for the type of item submitted. Allow ample margins all around. Manuscripts MUST BE DOUBLE-SPACED THROUGHOUT. For articles this includes title (all caps, centered), author names (all caps, centered), addresses (caps and lower case, centered), abstract and resumen, 5 key words or phrases, text, acknowledgments, literature cited, tables (caption on same page), and figure captions (grouped as consecutive paragraphs on one page). Order parts in the sequence listed, ending with figures. Each page should have a running header that includes the name(s) of the author(s), a shortened title, and the page number. Do not use a separate cover page or ‘erasable’ paper. Avoid footnotes except to indicate address changes. Abbreviations should be used sparingly and only standard abbrevia- tions will be accepted. Table and figure captions should contain all information relevant to information presented. All measurements and elevations should be in metric units, except specimen citations, which may include English or metric measurements. ; Authors of accepted papers will be asked to submit an electronic version of the manuscript. Microsoft Word 6.0 or WordPerfect 6.0 for Windows is the preferred software. Line copy illustrations should be clean and legible, proportioned to the MADRONO page. Scales should be in- cluded in figures, as should explanation of symbols, including graph coordinates. Symbols smaller than | mm after reduction are not acceptable. Maps must include a scale and latitude and longitude or UTM references. In no case should original illustrations be sent prior to the acceptance of a manuscript. Illustrations should be sent flat. No illustrations larger than 27 X 43 cm will be accepted. Presentation of nomenclatural matter (accepted names, synonyms, typification) should follow the format used by Sivinski, Robert C., in Maprono 41(4), 1994. Institutional abbreviations in specimen citations should follow Holmgren, Keuken, and Schofield, Index Herbariorum, 8th ed. Names of authors of scientific names should be abbreviated according to Brummitt and Powell, Authors of Plant Names (1992) and, if not included in this index, spelled out in full. Titles of all periodicals, serials, and books should be given in full. Books should include the place and date of publication, publisher, and edition, if other than the first. All members of the California Botanical Society are allotted 5 free pages per volume in MApRONO. Joint authors may split the full page number. Beyond that number of pages a required editorial fee of $40 per page will be assessed. The purpose of this fee is not to pay directly for the costs of publishing any particular paper, but rather to allow the Society to continue publishing MApDRONO on a reasonable schedule, with equity among all members for access to its pages. Printer’s fees for illustrations and typographically difficult material @ $35 per page (if their sum exceeds 30 percent of the paper) and for author’s changes after typesetting @ $4.50 per line will be charged to authors. At the time of submission, authors must provide information describing the extent to which data in the manu- script have been used in other papers that are published, in press, submitted, or soon to be submitted elsewhere. a : WW SSS SS. SNS S SS SSN Z a VOLUME 45, NUMBER 2 APRIL-JUNE 1998 MADRONO A WEST AMERICAN JOURNAL OF BOTANY midi ie Mis: cts: CONTENTS NATIVE PLANT DIVERSITY IN RIPARIAN COMMUNITIES OF THE SANTA MONICA MOouNTAINS, CALIFORNIA Philip W. Rundel and Shari B. Sturmer ..........cccccccccceeeeeeetesteteeeseeeeeeeeeeees 93 EARLY PRIMARY SUCCESSION ON DUNES AT BODEGA HEAD, CALIFORNIA Avinoam Danin, Stephen Rae, Michael Barbour, Nicole Jurjavcic, Peter Connors, and Eleanor UNALINger ........ccccccccccccecseesceeeeeeeeeeeeeeeseeeseaees 101 HYBRIDIZATION BETWEEN CIRCIDIUM FLORIDUM AND C. MICROPHYLLUM (FABACEAE) IN CALIFORNIA C. Eugene Jones, Larry J. Colin, Trudy R. Ericson, and DDODOLAIIARE DOTS CLE ere ri a Se 110 THE ROLES OF SoIL TYPE AND SHADE INTOLERANCE IN LIMITING THE DISTRIBUTION OF THE EDAPHIC ENDEMIC CHORIZANTHE PUNGENS VAR. HARTWEGIANA (POLYGONACEAE) Jodi M. McGraw and Anna L,Levin 2... .icieiiccvveccs shiek se sioseasveesccssvceceoeess 119 ATRIPLEX LONGITRICHOMA (CHENOPODIACEAE), A NEw SPECIES FROM SOUTHWESTERN NEVADA AND EAST-CENTRAL CALIFORNIA Howard C. Stutz, Ge-Lin Chu, and Stewart C. Sanderson ..........00..00..00+ 128 EARLY SECONDARY SUCCESSION FOLLOWING CLEARCUTS IN RED FIR FORESTS OF THE SIERRA NEVADA, CALIFORNIA R. F. Fernau, J. M. Rey Benayas, and M. G. Barbour ............000c0ceseeeeeeees 131 A NEw GILIA (POLEMONIACEAE) FROM LIMESTONE OUTCROPS IN THE SOUTHERN SIERRA NEVADA OF CALIFORNIA James:R. ShevocK Gnd Alv@ Go DGY 035. .11000 Sinks upeeseateeeeneneacneceeees 137 WATER POTENTIALS OF SALVIA APIANA, S. MELLIFERA (LAMIACEAE) AND THEIR HYBRIDS IN THE COASTAL SAGE SCRUB OF SOUTHERN CALIFORNIA David S. Gill and: Barbara Je TOMO oi csc.ccccecs0 bs ce eevcccssocnsevecsessstecs 141 INVENTORY OF THE VASCULAR FLORA OF THE BLAST ZONE, Mount St. HELENS, W ASHINGTON Jonathan H. Titus, Scott Moore, Mildred Arnot, and Priscilla J. Titus.... 146 THE DISTRIBUTION OF VESICULAR-ARBUSCULAR MyYCORRHIZAE ON MOUNT St. HELENS, WASHINGTON Jonathan H. Titus, Roger del Moral, and Sharmin Gamicet ...............00006+ 162 LaTE-HOLOCENE VEGETATION CHANGES FROM THE LAS FLORES CREEK COASTAL LOWLANDS, SAN DiEGo County, CALIFORNIA R. Scott Anderson and Brian F. By1d .......ccccccccccccccssssssesececeeesceeeceeeaaaaaaaes 171 REVIEW MOJAVE DESERT WILDFLOWERS, By JON MARK STEWART KR ODETE POMECESON ea ae se ee 183 NOTEWORTHY TNIRULOIN Bige oottee aoe see cette ar Snare een aa ces Seer eats ese esata eter eee ener eee ree 184 COLLECTIONS COATIBORINTAS oo eecceat ccc eic ee occuctneesrck rovnsue de ok een auaaneuee es eeaainaa sel endieget es eam ate ice kot 184 OREGON cetera 88 sags ee a ie ues ee cere ee Hee ne S ns Led rest Met cone rr eee ena 185 PUBLISHED QUARTERLY BY THE CALIFORNIA BOTANICAL SOCIETY ~ Maprono (ISSN 0024-9637) is published quarterly by the California Botanical Society, Inc., and is issued from the office of the Society, Herbaria, Life Sciences Building, University of California, Berkeley, CA 94720. Subscription information on inside back cover. Established 1916. Periodicals postage paid at Berkeley, CA, and additional mail- ing offices. Return requested. PosTMASTER: Send address changes to MApRONO, ‘/ Mary Butterwick, Botany De- partment, California Academy of Sciences, Golden Gate Park, San Francisco, CA 94118. Editor—KRISTINA A. SCHIERENBECK California State University, Chico Department of Biology Chico, CA 95427-0515 kschierenbeck @csuchico.edu Editorial Assistant—Davip T. PARKS Book Editor—Jon E. KEELEY Noteworthy Collections Editors—DIETER WILKEN, MARGRIET WETHERWAX Board of Editors Class of: 1998—FREDERICK ZECHMAN, California State University, Fresno, CA Jon E. KeeLey, U.S. Geological Service, Biological Resources Division, Three Rivers, CA 1999—Timotny K. Lowrey, University of New Mexico, Albuquerque, NM J. MARK Porter, Rancho Santa Ana Botanic Garden, Claremont, CA 2000—Pame_a S. Sottis, Washington State University, Pullman, WA JOHN CALLAWAY, San Diego State University, San Diego, CA 2001—RobertT PATTERSON, San Francisco State University, San Francisco, CA PauLa M. ScuHIFFMAN, California State University, Northridge, CA 2002—NorMAN ELLSTRAND, University of California, Riverside, CA Cara M. D’ Antonio, University of California, Berkeley, CA CALIFORNIA BOTANICAL SOCIETY, INC. OFFICERS FOR 1998—1999 President: R. JoHN LittLe, Sycamore Environmental Consultants, 6355 Riverside Blvd., Suite C, Sacramento, CA 95831 First Vice President: Susan D’ ALcAmMo, Jepson Herbarium, University of California, Berkeley, CA 94720 Second Vice President: Davip KetL, California Polytechnic State University, Biological Sciences Department, San Luis Obispo, CA 93407 Recording Secretary: ROXANNE BitTMAN, California Department of Fish and Game, Sacramento, CA 95814 Corresponding Secretary: SUSAN BAINBRIDGE, Jepson Herbarium, University of California, Berkeley, CA 94720. suebain @casnowy.qal.berkeley.edu Treasurer: Mary BuTTerwick, Botany Department, California Academy of Science, Golden Gate Park, San Fran- cisco, CA 94118. butterwick.mary @caepamail.epa.gov The Council of the California Botanical Society comprises the officers listed above plus the immediate Past President, WAYNE R. FERREN, JR., Herbarium, University of California, Santa Barbara, CA 93106; the Editor of Maprono; three elected Council Members: MARGRIET WETHERWAX, Jepson Herbarium, University of California, Berkeley, CA 94720; James SHEvock, USDI National Park Service, Pacific West Region, 600 Harrison Street, Suite 600, San Francisco, CA 94107; DIANE ELaM, U.S. Fish and Wildlife Service, 3310 El Camino Avenue, Sacramento, CA 95825; Graduate Student Representative: DENNIS P. WALL, Jepson Herbarium, University of California, Berke- ley, CA 94720. This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). SMITHSONTZ W may 171999 LIBRARIES NATIVE PLANT DIVERSITY IN RIPARIAN COMMUNITIES OF THE SANTA MONICA MOUNTAINS, CALIFORNIA Mapbrono, Vol. 45, No. 2, pp. 93-100, 1998 PHILIP W. RUNDEL AND SHARI B. STURMER Department of Biology, University of California, Los Angeles, CA 90095 ABSTRACT Riparian ecosystems in the mountains of southern California are characterized by unusually dynamic conditions of fluvial disturbance, fire, landslides, and other physical processes. Riparian communities and associated wetland habitats make up less than 1% of the land area of the Santa Monica Mountains of southern California but are the primary habitat for nearly 20% of the native vascular plant flora. While the conditions responsible for such high biodiversity has not been well investigated, the dynamic distur- bance regime and seasonal availability of water in riparian ecosystems are two critical factors. In com- parison to the total native flora of the Santa Monica Mountains, riparian specialists (1.e., those species with their primary ecological occurrence in such habitats) showed a higher relative frequency of herba- ceous perennials, and a lower relative frequency of woody shrubs, geophytes, and annuals. Winter decid- uous growth habit characterizes nearly 80% of the woody riparian specialists, a far higher level than in comparable non-riparian species. Compound to the total flora, riparian specialists were found to have relatively broader geographic patterns of distribution within California. No rare or endangered species are included in the riparian-specialist flora of the mountains. Biological diversity of native plant species in riparian communities appears to be negatively impacted by human disturbance. Habit modification, weedy exotic species introductions, stream channel modification, and heavy recreational use all appear to lead to sharp reductions in plant species diversity. Without additional knowledge of the demography and ecology of potentially keystone riparian species, it is impossible to accurately model the impact of an- thropogenic disturbance regimes on the structure and stability of these riparian ecosystems. Riparian ecosystems play a critical role in a va- riety of ecosystem processes. Situated at the inter- face between terrestrial and aquatic ecosystems, these ecosystems act to buffer hydrologic and ero- sional cycles, control and regulate biogeochemical cycles of nitrogen and other key nutrients, limit fire movements, and create unique microclimates for animal species (Gregory et al. 1991; Clary and McArthur 1992; Naiman et al. 1993; Malanson 1993). Both terrestrial and aquatic wildlife depend on riparian ecosystems with their year-round avail- ability of water, nutrients, food sources, and organic sediments. In addition to these critical components of food resources, riparian ecosystems provide wildlife with a structural complexity that includes mosaics of shade and sun, shelter, and protected corridors between adjacent plant communities. It is not surprising, therefore, that riparian ecosystems are centers of high biodiversity (Nilsson et al. 1989; Gregory et al. 1991). The ecological significance of riparian zones is accentuated in the semi-arid mediterranean-climate regions of southern California where water re- sources are strongly limiting for plant growth. Ri- parian communities of the Santa Monica Mountains occur in an unusually dynamic geomorphic and cli- matic environment. The southern California moun- tains are subject to impacts of fluvial disturbances from flooding and associated sediment deposition and/or erosion (Rice and Foggin 1971), as well as nonfluvial impacts from adjacent upland areas in the form of fire, landslides, and other alterations to their physical structure. Wildfires burn through the chaparral and woodlands of the Santa Monica Mountains at intervals of 10—30 years or more (Radke et al. 1982), with near total removal of up- land vegetation cover and litter layers, as well as significant impacts on the canopy cover of woody species in riparian habitats. Flood cycles frequently follow fires when vegetation cover has been strong- ly altered, causing sharp increases in runoff with associated erosion, soil slippage, and mudflows. Fluvial disturbances from flooding can be particu- larly severe in irregular years with strong El Nino/ Southern Oscillation events when rainfall of re- markable intensity occurs over short periods of time. Added to this are a geologically young land- scape, steep terrain, and a high frequency of geo- logical faulting and associated earthquakes as fur- ther factors in destabilization of hydrologic drain- age zones. Thus the riparian ecosystems of the Santa Monica Mountains are subject to extreme but unpredictable events of catastrophic disturbance. Do dynamic environmental disturbance regimes and the associated structural diversity of riparian ecosystems produce habitats with high diversity of native species? Despite the importance of such a question in understanding the environmental signif- icance of these zones and thus developing effective and objective policies of resource management for riparian habitats, there has been little study of the significance of native plant species diversity within riparian ecosystems of southern California. In this paper we address this question of biodiversity of 94 MADRONO native vascular plant species in riparian communi- ties of the Santa Monica Mountains, and evaluate this diversity in relation to the life-forms and bio- geographic distribution of the species present. We further discuss the potential significance of distur- bance regimes, both natural and anthropogenic, as important factors in influencing the structural and ecological diversity of riparian communities. Non- native species, an increasingly important compo- nent of riparian community structure, are not in- cluded in this study. MATERIALS AND METHODS The Santa Monica Mountains form the most southwestern component of the east-west trending Transverse Ranges of southern California. They ex- tend for 73 km from the Oxnard Plain on the west to the Los Angeles River on the east. Elevations range from sea level to 948 m. This small eleva- tional range and a generally semi-arid mediterra- nean-climate regime restricts overall plant species diversity because of the relative rarity of permanent water supplies in the summer months (Raven et al. 1986). Chaparral, coastal sage scrub, oak wood- lands, and valley grasslands cover most of the range. Significant wetlands are present, however, in the form of riparian woodlands and limited areas of freshwater marsh, salt marsh, and aquatic com- munities. A checklist of wetland and/or riparian species existing in the Santa Monica Mountains was pre- pared from field observations and from habitat in- formation presented in Raven et al. (1986). This checklist was restricted to native species whose characteristic habitat was a wetland type. Riparian habitat specialists were defined as those species whose characteristic habitat was riparian zones, flood plains, ravine bottoms, and springs. Taxa present in wetland habitats but equally or more characteristic of other habitats were not included. Thus a species such as Quercus agrifolia Nee which is abundant along riparian corridors was not included since it is widespread on slopes and flats away from true wetland habitats. The life-form and biogeographic distribution of each wetland species was classified using infor- mation in Hickman (1993). Life-forms used were annual (therophyte), herbaceous perennial (hemi- cryptophyte), suffrutescent subshrub (most cham- aephytes), and woody shrub and tree (phanerophy- tes). Biogeographic distributions in California for these species were divided into eight regions: Northwestern, Cascade Ranges, Sierra Nevada, Great Valley, Central Western, Southwestern, Great Basin, and Desert. The distribution and extent of wetland types in the Santa Monica Mountains and adjacent area of the Santa Monica Bay drainage basin have been described by Josselyn et al. (1993) in a study for the Santa Monica Bay Restoration Project. Restrict- [Vol. 45 ing the Santa Monica Mountains to those areas con- tained within the Newbury Park, Triunfo Pass, Thousand Oaks, Point Dume, Calabasas, Malibu Beach, Canoga, Topanga, Van Nuys, Beverly Hills, Burbank and Hollywood 7.5 minute USGS quad- rangles, these investigators calculated the existence of 1066 ha of wetlands. This estimate was made using an analysis method based on digitized quad- rangle maps to determine aerial coverage or length of each wetland type. Length of stream systems was converted to area using a standardized width for each stream type. Wetland communities in the Santa Monica Mountains were divided into six types by Josselyn et al. (1993). Channelized streams with little or no vegetation were classified as riverine habitats and included 16.9% of all wetland areas. Intermittent streams with vegetation cover and surface flow for only part of the year but permanently wet soils comprised 23.8% of the wetlands. Streams in To- panga, Little Sycamore, and Solstice Canyons are examples of such intermittent streams. Perennial streams with woody vegetation corridors, such as Malibu Creek and Las Virgenes Creek, made up 18.0% of the wetland area. Freshwater marshes and permanently moist flood plains away from stream channels are relatively uncommon in the Santa Monica Mountains, and totaled only 6.5% of exist- ing wetlands. Lakes, ponds, and reservoirs com- prised the largest area of wetlands covering 357 ha or 33.5% of the total. These areas, however, are largely made up of Metropolitan Water District res- ervoirs above Hollywood and Beverly Hills, and artificial lakes near Thousand Oaks. These reser- voirs and lakes support little natural vegetation or species diversity. The final category of wetlands are salt marshes and estuaries which are largely formed by Malibu Lagoon, with a small area near Point Dume. Salt marshes make up only 13.3 ha, or 1.2% of all wetlands. Taxonomic names used in this study follow Hickman (1993). RESULTS Based on this analysis by Josselyn et al. (1993), the total wetland area of 1066 ha in the Santa Mon- ica Mountains is only 1.1% of the total area of 98,500 ha (Raven et al. 1986). Excluding salt marsh and lake/reservoir habitats, riparian areas of all types including freshwater wetlands comprise only 0.7% of the area of the mountains. There are cer- tainly small areas of springs and first and second order streams that contain riparian species but could not be mapped at the scale used in this study. Nev- ertheless, it seems doubtful that the addition of such habitats would raise the total area of riparian hab- itats in the Santa Monica Mountains beyond 1—2% of total land area. Despite the small relative area of wetland habi- tats in the Santa Monica Mountains, the signifi- 1998] on o fo) o A -S (oe) Ls) (2) Relative frequency of occurence ~ rf) ° ° P Ch Fic. 1. RUNDEL AND STURMER: RIPARIAN SPECIES DIVERSITY 95 H G T Life-forms Relative distribution of plant life-forms within the riparian (dark bars) and total vascular plant floras (open bars) of the Santa Monica Mountains. Only native species are included in these summaries. Abbreviations for life- forms are as follows: P = phanerophytes (woody plants above 50 cm height), CH = chamaephytes (woody or semi- woody plants below 50 cm in height), H = hemicryptophytes (perennial plants dying back to ground level annually), G = geophytes (perennial herbs dying back to below-ground fleshy tissues), and T = therophytes (annuals). cance of these habitats as centers of biodiversity 1s surprisingly high. Of the total native vascular plant flora of 644 species within the mountains (Raven et al. 1986), 161 are characteristically wetland spe- cies. This amounts to 25% of the flora. Salt marsh habitats are relatively poor in vascular plant diver- sity, with only 24 species in the Santa Monica Mountain region. Excluding obligate salt marsh species, three marine angiosperms, and aquatic spe- cies that occur in freshwater lakes and pools, ri- parian habitats and associated moist flood plains and freshwater marshes form the primary habitats for 125 species (19.4% of the flora) in only a tiny fraction of the surface area of the region (Appen- dix). Although woody species form the dominant as- pect of riparian vegetation, trees and shrubs made up only a small part of the overall diversity of ri- parian species. There are ten riparian-specialist trees in the flora of the Santa Monica Mountains. These, together with four characteristically riparian shrub species, form 11% of the riparian flora (Fig. 1). This group includes ten deciduous tree species (Acer macrophyllum Pursh, Alnus rhombifolia Nutt., Fraxinus velutina Torrey, Platanus racemosa Nutt., Populus fremontii S. Watson ssp. fremontii, P. balsamifera L. ssp. trichocarpa (Torrey & A. Gray) Brayshaw, Salix exigua Nutt., S. laevigata Bebb, S. lasiolepis Benth. and S. lucido Muhlenb. ssp. lasiandra) and one evergreen tree (Umbellu- laria californica (Hook. & Arn.) Nutt.). The fac- ultative riparian tree Quercus agrifolia, one of the most abundant and ecologically significant trees in riparian habitats, is also an evergreen species. Two deciduous shrubs (Cornus glabrata Benth. and Hol- odiscus discolor (Pursh) Maxim.) and two ever- green shrubs (Baccharis salicifolia (Ruiz Lopez & Pavon) Pers. and Myrica californica Cham. & Schldl.) are the additional woody plants character- istic of riparian habitats. Herbaceous perennials (hemicryptophytes) form the largest single group (58%) of riparian special- ists. Next are annuals with 28%, followed by suf- frutescent subshrubs (chamaephytes) with 3% of the riparian flora. This relative distribution of life- forms among the riparian flora of the Santa Monica Mountains differs notably from the life-form spec- tra of the total native flora. Woody shrubs, suffru- tescent subshrubs, and annuals are proportionally less common among the riparian specialist flora than the total flora, while herbaceous perennials are far more common (Fig. 1). No riparian-specialist geophytes occur in the Santa Monica Mountains. The biogeographic range of distribution for most riparian species in the Santa Monica Mountains is broad. Two-thirds or more of this riparian flora can also be found in riparian areas of the Northwest Coast, Sierra Nevada, Central Valley, and Central Coast phytogeographic regions of California. (Fig. 2, Appendix). Even the arid Great Basin and Mo- jave/Sonoran Desert regions share more than 40% of the riparian species from the Santa Monica Mountains. Only 4% of riparian species from the Santa Mon- ica Mountains are restricted in distribution in Cal- ifornia to the Southwestern phytogeographic region (SW), while 56% of the riparian species occur in six or more of the regions (Fig. 3). Nearly a quarter of the riparian species occur in all eight phytogeo- graphic zones. These data can be compared with data for the total flora of the large and ecologically diverse family Asteraceae in the Santa Monica 96 MADRONO CAR - 59% GB - 42% NW - 77% GV - 70% CW - 86% SW - 100% Fic. 2. Relative presence of riparian species from the Santa Monica Mountains in each of the eight major phy- togeographic regions of California (sensu Hickman 1993). Each value shown is the percent of the 132 riparian/wet- land species from the Santa Monica Mountains that also occur naturally in each region. NW = northwestern, CaR = Cascade Ranges, SN = Sierra Nevada, GV = Great Valley, CW = Central Western, SW = Southwestern, GB = Great Basin Province, and D = Desert Province. Mountains, where 5% of the species are restricted to the southwestern region in California, but only 40% occur in six or more regions. DISCUSSION The small percentage of total land area occupied by riparian habitats in the Santa Monica Mountains is wholly consistent with relative cover of such communities elsewhere in California, as well as the western United States. Omart and Anderson (1986) have estimated that riparian habitats make up ap- proximately 0.5% of land area in the West. The ability of such small areas to support high levels of biodiversity for both plant and animal species has been attributed to a variety of reasons (Hubbard 1977; Brode and Bury 1984; Klebenow and Oak- leaf 1984; Nilsson et al. 1989; Gregory et al. 1991; Clary and Medin 1992). Primary among these caus- es is structural diversity in plant above-ground ar- chitecture, derived from the multiple selective pres- sures operating in the dynamic riparian landscape. Resource availability in riparian habitats throughout the year is also important. The relative differences in life-form distribution between riparian species and the total flora of the Santa Monica Mountains may provide clues to the significance of disturbance regimes in maintaining high plant species diversity within riparian com- munities. Despite the availability of water resources [Vol. 45 and high plant biomass, riparian habitats have low- er diversity in woody shrubs than the overall flora of the Santa Monica Mountains. This lower diver- sity in shrubs may reflect the impacts of intense flooding and attendant streambank and streambed scouring on the establishment and long-term sur- vival of low-growing woody species. However, the association of a tree growth form with mesic habitats in the Santa Monica Mountains is highly pronounced. Ten of the 14 native tree spe- cies in the Santa Monica Mountains (71% of the total number) are riparian specialists. Of the re- maining four tree species, Quercus agrifolia, Quer- cus lobata Nee, and Juglans californica S. Watson all favor conditions with high moisture availability as indicated by the deep root systems of the two oak species which tap underground water supplies (see Griffin 1967; Rundel 1980) and the association of J. californica with areas of seepage or high soil moisture availability. Only the rare Juniperus cali- fornica Carriére occurs on seemingly xeric sites among the tree species in the Santa Monica Moun- tains. It is interesting to note that 11 of the 14 species of riparian-specialist woody shrubs and trees (10 trees and four shrubs) are winter deciduous. This phenological pattern is relatively uncommon among woody shrubs in the Santa Monica Moun- tains where the majority of shrubs are evergreen chaparral species or drought-deciduous species as- sociated with coastal sage scrub. The success of such a winter deciduous phenology suggests that water availability throughout the dry summer months is a critical element of the success of woody riparian species, while cool winter temperatures may limit potential photosynthetic gain during the leafless period for these species. Annual plants (therophytes), another group poor- ly represented in the riparian-specialist flora, typi- cally require soil seed pools for continued estab- lishment in mediterranean-climate regions. Flood and erosion cycles along streams and water chan- nels may not favor native species in this group, although non-native annual species can be wide- spread. Herbaceous perennials (hemicryptophytes) form the most diverse life-form of riparian special- ists, with a relative frequency well above that of the group for the total flora of the Santa Monica Mountains. Individuals of these species have the ability to persist for many years along streambanks or streambeds, and to rapidly resprout or reestablish themselves after disturbance. Adaptations for survival in dynamic riparian communities appears to result from generalist strat- egies that work well in riparian zones across broad mesoclimatic gradients. Patterns of biogeographic distribution for the riparian-specialist flora of the Santa Monica Mountains clearly indicate that these species are commonly widespread in distribution. This is not surprising given that the relatively high availability of water in riparian habitats decouples RUNDEL AND STURMER: RIPARIAN SPECIES DIVERSITY oF 4 3 2 1 Number of phytogeographic regions 1998] o oOo ec 2 3 (4) © ° r) > 0 ec $s og © © 2 s o fed 8 7 6 5 Fic. 3. Relative biogeographic distribution of riparian species of the Santa Monica Mountains among the eight major phytogeographic regions of California. The values shown are the percent of these riparian species that are restricted to each number of phytogeographic regions. many plant species from regional rainfall regimes. None of the specialist riparian flora of the Santa Monica Mountains is included as rare or endan- gered for the California flora (Skinner and Pavlik 1994). Despite generalist adaptations to disturbance re- gimes, riparian zone plants are highly sensitive to human impacts. Vegetation clearance, trampling, stream channel modifications, altered fire regimes, grazing, and recreational activities have significant impacts on the structure and diversity of riparian communities in the Santa Monica Mountains. These impacts come about through physical changes in the environment as well as secondarily by the in- troduction of exotic species which choke out the growth of natives. We have noted that riparian plant species diversity is commonly inversely correlated with levels of disturbance, while frequency of ex- otic plant occurrences is directly related to distur- bance (Rundel and Sturmer unpublished). We know very little, however, about the relationship between biodiversity and ecological function in riparian eco- systems. Non-native species appear to be increasing in diversity and abundance in riparian habitats in the Santa Monica Mountains. Some of these spe- cies, particularly Arundo donax L., have the poten- tial to profoundly impact these ecosystems. Woody riparian species of Salix spp., Platanus racemosa Nutt. and Alnus rhombifolia Nutt., as well as Quercus agrifolia are keystone species pro- viding the structural stability and biological pro- ductivity of riparian zones in the Santa Monica Mountains. Yet our knowledge of the demographic pattern of seed dispersal, seedling establishment, and sapling growth in relation to natural and an- thropogenic disturbance regimes remains poorly Studied. Process-based models of the dynamics of flood and fire cycles impacting the structure and function of riparian ecosystems in the mediterra- nean-climate regions of California are critically needed if we are to effectively manage riparian community resources for the future. Given the dynamic changes of fire and flood cy- cles that alter the physical environment of riparian ecosystems in southern California, it is not surpris- ing that riparian plant communities are highly ir- regular in structural and compositional diversity. Riparian communities reflect not only the effects of both individual and cumulative disturbance regimes along their stream channels, but also the impacts of landscape processes affecting adjacent chaparral and woodland communities. Structural and com- positional complexity presents problems in devel- oping workable classification systems based on spe- cies dominance for riparian plant communities in California (Holland 1986; Sawyer and Keeler-Wolf 1995). The impacts of irregular disturbance re- gimes, geomorphic history and structure are now being considered in developing new classification systems for riparian communities in the coastal mountains of Southern California (Ferrin et al. 1994). ACKNOWLEDGMENTS This project was made possible by a grant DEB- 9412224 from the National Science Foundation to the se- nior author, a grant from the Southwest National Parks and Monuments Association, and a Patricia Roberts Harris Fellowship to the second author. We thank Drs. Karen Esler, M. Rasoul Sharifi, Ray Sauvajot, and Qinfeng Guo for helpful discussions and review of earlier drafts of this manuscript. LITERATURE CITED BrRopbE, J. M., AND R. B. Bury. 1984. The importance of riparian ecosystems to amphibians and reptiles. Pp. 98 MADRONO 30—36 in R. E. Warner and K. M. Hendrix (eds.), California riparian systems: ecology, conservation, and production management. University of California Press, Berkeley. CLARY, W. P. AND E. D. MCARTHUR. 1992. Introduction: ecology and management of riparian shrub commu- nities. Pp. 1—2 in Proceedings of the symposium on ecology and management of riparian shrub commu- nities. USDA Forest Service General Technical Re- port INT-289. CLARY, W. P. AND D. E. MEDIN. 1992. Vegetation, breeding bird, and small mammal biomass in two high-eleva- tion sagebrush riparian habitats. Pp. 100—110 in Pro- ceedings of the Symposium on Ecology and Manage- ment of Riparian Shrub Communities. USDA Forest Service General Technical Report INT-289. FERRIN, W., P. L. FIEDLER, R. A. LEmDY, K. D. LAFFERTY, AND L. A. K. MERTES. 1995. Wetlands of California. Part I. Classification and description of wetlands of the central and southern California coast and coastal watersheds. Madrono 43:125—182. FERRIN, W., P. L. FIEDLER, R. A. LeEtpy, K. D. LAFFERTY, AND L. A. K. MERTES. 1995. Wetlands of California. Part II. Key to and catalogue of wetlands of the cen- tral and southern California coast and coastal water- sheds. Madrono 43:183—233. GREGORY, S. V., FE J. SWANSON, A. MCKEE, AND K. W. CUMMINS. 1991. An ecosystem perspective of riparian zones. BioScience 41:540—551. GRIFFIN, J. R. 1967. Xylem sap tension in three woodland oaks of central California. Ecology 54:152—159. HICKMAN, J. (ed.). 1993. The Jepson manual: higher plants of California. University of California Press, Berke- ley. HOLLAND, R. FE 1986. Preliminary description of the ter- restrial natural communities of California. California Department of Fish and Game, Sacramento, CA. HUBBARD, J. P. 1977. Importance of riparian ecosystems: biotic considerations. Pp. 14—18 in Importance, pres- ervation and management of riparian habitat: a sym- posium. USDA Forest Service General Technical Re- port RM-43. JOSSLEYN, M., S. CHAMBERLAIN, P. GOODWIN, AND K. CuFFE. 1993. Wetland inventory and restoration po- tential: Santa Monica Bay watershed. Public report. [Vol. 45 Santa Monica Bay Restoration Project, Monterey Park, CA. KLENBOW, D. A. AND R. J. OAKLEAF. 1984. Historical avi- faunal changes in the riparian zone of the Truckee River, Nevada. Pp. 203—209 in R. E. Warner and K. M. Hendrix (eds.). California riparian systems: ecol- ogy, conservation, and production management. Uni- versity of California Press, Berkeley. MALANSON, G. P. 1993. Riparian Landscapes. Cambridge University Press, Cambridge. NAIMAN, R. J., H. DECAMPS, AND M. POLLOcK. 1993. The role of riparian corridors in maintaining regional bio- diversity. Ecological Monographs 3:209-—212. NILSSON, C., G. GRELSSON, M. JOHANSSON, AND U. SPER- ENS. 1989. Patterns of plant species diversity along riverbanks. Ecology 70:77-—84. OHMART, R. D. AND B. W. ANDERSON. 1986. Riparian hab- itat. Pp. 2—4 in Strategies for protection and manage- ment of floodplain wetlands and other riparian eco- systems: proceedings. USDA Forest Service General Technical Report WO-12. RADKE, K. W.-H., A. M. ARNDT, AND R. H. WAKIMOTO. 1982. Fire history of the Santa Monica Mountains. Pp. 438—443 in C. E. Conrad and W. C. Oechel (eds.). Dynamics and Management of Mediterranean-Type Ecosystems. USDA Forest Service General Technical Report PSW-58. RAVEN, P. H., H. J. THOMPSON, AND B. A. PRIGGE. 1986. Flora of the Santa Monica Mountains. 2nd ed. Southern California Botanists Special Publication No. 2. RicE, R. M. AND G. T. FoGaIn. 1971. Effect of high in- tensity storms on soil slippage on mountainous wa- tersheds in southern California. Water Resources Re- search 7:1485—1496. RUNDEL, P. W. 1980. Adaptations of mediterranean-climate oaks to environmental stress. Pp. 43—54 in Ecology, management, and utilization of California oaks. USDA Forest Service General Technical Report PSW-44. SAWYER, J. O. AND T. KEELER-WOLF. 1995. A manual of California vegetation. California Native Plant Soci- ety, Sacramento, CA. SKINNER, M. W. AND B. M. PAvVLIk. 1994. Inventory of rare and endangered vascular plants of California. California Native Plant Society Special Publication No. 1. Sacramento, CA. APPENDIX. WASCULAR PLANT SPECIES CHARACTERISTIC OF RIPARIAN AND ADJACENT FRESHWATER MARSH HABITATS IN THE SANTA MONICA MOUNTAINS, CALIFORNIA, AND THEIR LIFE FORMS AND BIOGEOGRAPHIC DISTRIBUTIONS. Taxonomy follows Hickman (1993). See text for discussion of life forms and distributions. Life forms are phanerophytes (Ph), hemicryp- tophytes (H), therophytes (Th), and chamaephytes (Ch). Phytogeographic regions are as illustrated in Figure 2. | Also found in fresh water marshes. * Also found in salt marshes. Life form Phytogeographic region Ph: HL. th Ch Species Family Genus Species Aceraceae Acer macrophyllum Alismataceae Alisma plantago-aquatica Echindorus berteroi Amaranthaceae Amaranthus californicus Apiaceae Berula erecta Hydrocotyle umbellata Oenanthe sarmentosa Apocynaceae Apocynum cannabinum Aspleniaceae Asplenium vespertinum NW CaR SN GV CW SW GB D + + + + + + + + + + + + = + + + + + + + + + + + + + + = + + + + + + + + + «+ + + + + + + + + + - + + + + + + + + + =+ +. 1998] RUNDEL AND STURMER: RIPARIAN SPECIES DIVERSITY 99 APPENDIX. CONTINUED Species Life form Phytogeographic region Family Genus Species Ph H Th Ch NWCaR SN GV CW SW GB D Asteraceae Ambrosia scanthicarpa ate ae a Artemisia douglasiana | a: awe I ce Se Aster subulatus a ahs a Baccharis douglasti =e Ss oa: aa i Baccharis salicifolia + | LS es = ca + Bidens frondosa + + + es = Bidens laevis a 5 aes Se ae Euthamia occidentalis! sie oo Sa fae Ge. ae EE Gnaphalium leucocephalum + + Gnaphalium palustre Se Sa es Helenium puberulum 5 ae i 5 5 ns ee a Hemizonia pungens = =P a Sas Lepidospartum — squamatum + + + + Madia elegans at ns 5c Ses Madia exigua so Ss GA Sts aioe: PE Pluchea odorata sa 5 a oes eS = + Psilocarphus tenellus + 2s es a a Gs COS + Betulaceae Alnus rhombifolia os epee + =P) =o =e Blechnaceae Woodwardia fimbriata a = pn es + + Brassicaceae Barbarea orthoceras =F Le: es Cee: Ce a Guillenia lasiophylla a | ad Oe a A ac Rorippa curvisiliqua - 5 Sa OU ae: LG a Campanulaceae Lobelia dunnii + a ae Chenopodiaceae Chenopodium macrospermum a “+ ee Cornaceae Cornus glabrata + = eee Sas os x Cyperaceae Carex barbarae + SG ee + Carex praegracilis + sn Sc a oR ee Carex senta “F Sa a, en Carex SpISSa a3 | eG Cyperus eragostis al +o + FS HS Cyperus erythrorhizos Ss 2 se os as en Cyperus esculentus te 2 Is pa: i oR: SE aS: a Cyperus niger = ee ee, Cyperus odoratus a + + + Eleocharis macrostachya “F se is a gy a SS I Eleocharis montevidensis as + oy Sag Scirpus acutus ete ee ee ee Scirpus americanus + 5 ets sac ae ras ec I oe Scirpus californicus a a ee a Scirpus cernuus ale a 7 a a Sa Scirpus maritimus? + an + + + + =+ Scirpus microcarpus = 5 a Ss ts ga I Datiscaceae Datisca glomerata ae ts aes Euphorbiaceae Chamaesyce serpyllifolia + = Ns ee i Ps Fabaceae Glycyrrhiza lepidota + See ae Hoita macrostachya 6 7 ae: Rs = Lotus oblongifolius = eee ee + Rupertia physodes sl a3 ; aS Trifolium obtusiflorum + = si oe: Trifolium variegatum 4 7 tae Se ee ee Hydrophyllaceae Nama stenocarpum + + Juncaceae Juncus balticus = 5 Gs i: ac Ms Juncus bufonius aE 5 a OS as 2 Os on Juncus macrophyllus i ee + Juncus mexicanus” se - a = ee Juncus patens =f Sm cy. ae Juncus phaeocephalus + = ne Sc Juncus textilis 5 = a Juncus torreyi “tr es Juncus xiphioides at ne: a A ae ei i See 100 Family Lamiaceae Lauraceae Lemnaceae Liliaceae Lythraceae Myricaceae Oleacea Onagraceae Orchidaceae Platanaceae Poaceae Polygonacae Polypodiaceae Primulaceae Pteridaceae Ranunculaceae Rosaceae Salicaceae Saxifragaceae Scrophulariaceae Solanaceae Thelypteridaceae Typhaceae Urticaceae Vitaceae Species Genus Mentha Stachys Stachys Stachys Umbellularia Lemna Lemna Lilium Ammannia Lythrum Lythrum Myrica Fraxinus Epilobium Ludwigia Oenothera Epipactis Platanus Agrostis Andropogon Hordeum Leptochloa Leptochloa Panicum Phragmites Poa Polypogon Polygonum Polygonum Polygonum Polygonum Rumex Rumex Polypodium Samolus Adiantum Ranunculus Holodiscus Populus Populus Salix Salix Salix Salix Boykinia Boykinia Castilleja Mimulus Mimulus Petunia Thelypteris Typha Typha Urtica Vitis APPENDIX. CONTINUED Species arvensis ajugoides albens bullata californica gibba triscula humboldtii coccinea californicum MADRONO Life form Ph hyssopifolium californica velutina brachycarpum peploides' laeta gigantea racemosa exarata glomeratus brachyantherum fascicularis uninervia capillare australis palustris interruptus amphibium hydropiperoides lapathifolium punctatum Crassus salicifolius californicum parviflorus capillus-veneris cymbalaria discolor balsamifera fremontii exigua laevigata lasiolepis lucida occidentalis rotundifolia spiralis cardinalis floribundus parviflora puberula domingensis latifolia dioica girdiana ++++4+4+4++4+ H ++4++ +++ +++ +++4+t++4+4+4+ ++4++ Th Ch +++ +++ + + + +t+++et+4++4+44 +++ +++ ++++t+4+4+4+4++4+ +++ +++ [Vol. 45 Phytogeographic region + + +++4+ + ++ + ++++4+4+4++4+ +++4+4++4+ ++ + + + ++++4+4++4+ +++t+4++4+4+4+ +++++ 444 +t+++t+4+4+4+4++4+ ~ +++ + + + +++ ++++4+4++4+ + + t+ett +H+ettett+ +44 +++ +++ ++ + +t+4++4++ 4444 ++++44 +t+++++4+4+4++4 ++t++tt¢+4+¢+4+4+4+ +++ tH HHH HHH tHe tte teeteeteetee tee tee ttettettetteettee tts + + + ++ +++4+4++ ++ +++4+4++ ++4++ NW CaR SN GV CW SW GB D ++ ++ 44+4++4+ 4 +++ ++ ++ ++ +++ | 1 MapRONO, Vol. 45, No. 2, pp. 101-109, 1998 EARLY PRIMARY SUCCESSION ON DUNES AT BODEGA HEAD, CALIFORNIA AVINOAM DANIN Evolution, Ecology, and Systematics, Hebrew University of Jerusalem, Israel 91904 STEPHEN RAE MUSCI, 1130 Cayetano Court, Napa, CA 94559 (srae @ musci.com) MICHAEL BARBOUR AND NICOLE JURJAVCIC Environmental Horticulture, University of California, Davis, CA 95616 PETER CONNORS AND ELEANOR UHLINGER Bodega Marine Laboratory, Bodega Bay, CA 94923 ABSTRACT Field examination of dune hillocks (nebkas) showed that changes in nebka topography and spread, sand texture, vascular plant and cryptogram cover, and species prescence were correlated with nebka ages between 15 and 135 yr. We determined age by a series of aerial photographs dating back to 1955 and topographic maps drawn as early as 1862. Development of a cryptogamic crust was significant during this range of time; on the oldest nebka it contributed 43 g biomass m * of nebka surface, which represented 3% of total above-ground biomass. Nebkas at Bodega grew in height 4 cm yr ' during the past century, whereas non-vegetated areas were deflated at the same rate. Sucession is driven by sand- stilling attributes of Ammophila arenaria, introduced to northern California in the mid-ninteenth century and additionally planted at Bodega in the mid-twentieth century. By reference to the 1862 map we concluded that A. arenaria has built a prominent, continuous foredune and hinddune since the time of its arrival. Some of the earliest American concepts about succession came from studies of coastal dunes (Cowles 1901), and the Pacific coast dunes were one of the first Californian ecosystems to be mono- graphed (Cooper 1936, 1967). Despite this history, most published information about primary succes- sion on California dunes is largely anecdotal and inferrential (Barbour and Johnson 1988; McBride and Stone 1976; Barbour et al. 1981; Holton and Johnson 1979). California dunes have been invaded in this century by Ammophila arenaria (L.) Link (Eu- ropean beachgrass) to the point that the prior dominant, Leymus mollis (Trin.) Pilger ssp. mollis (American dunegrass), has been virtually elimi- nated from many dune locations, with consequent changes in dune topography, species richness, and (probably) primary succession. Furthermore, no attention has been paid to the role of crypto- gams in California dune succession, even though the broad importance of biological crusts to eco- Systems with sand substrates is well known (e.g., Danin 1996; Dor and Danin 1996; St. Clair and Johansen 1993). Our objective was to begin a study of dune suc- cession, with a focus on changes in dune topogra- phy, sand texture, vegetation cover, species rich- ness, and growth forms. MATERIALS AND METHODS Study area. Our study location is the Bodega Marine Reserve (Fig. 1), property managed by the University of California for the purpose of biolog- ical research and teaching. The physical and bio- logical setting of Bodega Head have been described in detail by Barbour et al. (1973), Koonig (1963), Light et al. (1967), Lipps and Moores (1971), Standing et al. (1975), and Diamond and Kennedy (1975). The reserve consists of 132 ha in the middle of Bodega Head Peninsula, located approximately 80 km north of San Francisco at 38°19’N latitude and 123°04'W longitude. The rocky, granitic headland is abruptly truncated within the Reserve’s bound- aries by the San Andreas Fault. Northeast of the fault is a 0.7—1.7 km by 3.5 km swath of relatively epen dunes jointly administered by the University of California and the California Department of Parks and Recreation. Climate type is maritime- moderated-mediterranean. Annual precipitation is 78 cm, about 80% of which falls in the months November through March. Daily and seasonal am- plitudes of temperature are modest and frosts are rare. Mean annual air and surface water tempera- tures are both 12°C. The majority of summer days have morning and evening fog. Prevailing winds are from the northwest and average 15 kph; spring 102 MADRONO BODEGA MARINE RESERVE he 200m BODEGA MARINE LABORATORY ——————_——_—_———_=_— smn ee — ———————— — RESERVE BOUNDARY scale: 1°-400° .—— + — LABORATORY ENCLAVE eer "ene TRAILS [Vol. 45 Fic. 1. Paired contour maps of a portion of Bodega dunes. The 1986 map (top) has 5 ft contours; the 1862 map (bottom) has 20 ft contours. The intersection of lines on the 1862 map, just above (north of) Horseshoe Cove, corre- sponds to 38°19'N latitude 123°03'W longitude. The approximate location of our 30 X 150 m study plot is shown by the arrow (top). Scale for both maps is shown on the 1986 map. is the windiest season, with afternoon northwest winds averaging 30-50 kph on many days. Dune sand originates as material eroded from northern watersheds and carried south by ocean currents. Bodega Head and the adjacent dunes were oc- cupied by Coast Miwok for at least the past 3000 yr, judging from well-documented artifacts includ- ing shell middens (Colley 1970; Greengo 1951). Russian settlers established a farming settlement early in the nineteenth century, but retreated to Alaska in the 1840’s. American settlers since that time have used the area mainly for dairy cattle grazing, farming, and sheep pasture. It is probable that dune plant cover declined as a result of grazing 1998] and farming, and that the dunes became more mo- bile than they were during Indian occupation. Sand encroachment into Bodega Harbor this century has required periodic dredging to keep shipping lanes open. Between the 1920’s and 1950’s, extensive portions of the dunes were planted to European beachgrass in an attempt to stabilize the sand and minimize the need for repeated dredging. Now, the dunes are a mosaic of densely vegetat- ed areas, barren blowouts, swales, and scattered hillocks which have some degree of vegetative cov- er. We will hereafter call hillocks ‘‘nebkas,’”’ a north-African term meaning a topographic feature created by the sand-stilling nature of vegetation rooted in sandy substrate. Our focus is on a rela- tively open, 500-m-deep region, between a 5 m tall foredune and a 45 m tall hinddune. The foredune and hinddune vegetation cover is dominated by A. arenaria and the shrub Lupinus arboreus Sims (bush lupine, a California native here near its nat- ural northern limit; Davidson and Barbour 1977). Aerial photographs and historical information both confirm that this study area was not included in the A. arenaria plantings of this century. An 1862 topographic map of the dune area (Fig. 1; Rodgers and Kerr 1862; original scale 1:10,000), compiled before any A. arenaria was _ present, shows neither a foredune nor a hinddune, and in- stead details a series of low, broken dune ridges running perpendicular to the coast. Evidently, Bo- dega dune topography has been modified by the past century’s displacement of native vegetation by A. arenaria in much the same way Cooper (1967) concluded A. arenaria had changed other parts of the Pacific coast. Field sampling methods. Our intention was to se- lect a series of adjacent nebkas of varying age. Us- ing a series of aerial photographs of the dunes taken in 1955, 1971, 1977, 1988, 1990, and 1991, we identified four adjacent nebkas (Fig. 2A—D) that ap- peared at different times. These four occurred in a 30 X 150 m rectangular area with long dimension perpendicular to the shore, parallel to prevailing winds and remnants of nineteenth century dune ridges. Prevailing wind direction is from nebka A to nebka D. Nebka D is the windward portion of a ridge, hence differs in its present form from the island-like nebkas A, B, and C. In the late spring of 1995, we mapped each neb- ka’s edges to scale. The edge was identified by a congruence of: significant change in slope, plant cover, and surface coarseness. The high point of each nebka was marked with a pole and transect tapes were laid out in the four cardinal compass directions (N, E, S, W). We measured transect lengths to nebka edges. Using an inclinometer, we calculated nebka height relative to the edge. We sampled plant and cryptogamic cover by taking a point sample along north and east transects, the di- rections typically with highest plant cover due to DANIN ET AL.: DUNE SUCCESSION 103 Sag Ph gad aL sey janis ees (4 mR i) : ap avers Bes os “ eae Rik : at Y3 i pitas Baan aa Byte Pode hee an » ne , . wh Fic. 2. Paired aerial photographs of our study plot taken in 1991 (top) and 1955 (bottom). The 1991 photograph shows nebkas A—D and an exposed midden (m), whereas the 1955 photograph shows only nebkas B—D. The shore (west) lies to the left, just out of view. The area of nebkas A-—D is approximately 30 * 150 m. aspect or protection from prevailing winds. If no species covered a point the point was recorded as bare, and total percent cover was calculated as the percent of points with vegetation. Cryptogams were sent to experts for determina- tion: mosses to Bruce Allen, Marshall Crosby, and Alan Whittemore of the Missouri Botanical Garden, and to Dan Norris of the University of California, Berkeley; and lichens to Clayton Newberry of the University of California, Berkeley. Reference spec- imens are deposited in those herbaria. Determined vascular plant taxa and specimens are deposited in both the University of California, Davis and the Bodega Marine Laboratory herbaria. Vascular plant nomenclature follows Hickman (1993). We collected sand samples of 500—1000 g size at two depths and nine locations for each nebka: from the surface 2 cm and the subsurface 10—12 104 ee aA oe “? ‘ni * 7} ey x iM Fu A iw Mari ao J Fic. 3. Metal frame device used to extract plugs of cryp- togamic crust 2 cm thick and 5 X 5 cm in area. cm at the nebka center, midway along each of the four transects, and at the end of each transect. An equal number of random samples were collected from internebka locations. Samples were oven- dried at 105°C for 48 hr, then shaken on a recip- rocating shaker for 3 min through a sequence of standard sieves: 2.0, 0.5, 0.25, and 0.05 mm. These sizes separated five fractions as: gravel, coarse sand, medium sand, very fine sand, and silt + clay, respectively (Gee and Bauder 1986; USDA 1951). We also sampled the biotic crust for organic mat- ter content and sand texture. We constructed a met- al sampling device (Fig. 3) to extract a plug of crust TABLE 1. PHYSICAL TRAITS OF NEBKAS A, B, C, AND D, superscript letters are statistically different at P < 0.5. Trait A Area (m7?) 378 Height (m) 0.4 Coarse sand (%) 34. 1a Very fine sand (%) 6.88 MADRONO [Vol. 45 5 X 5 cm square and 2 cm deep. The top 3—4 mm of the plug consisted of green moss or lichen mat- ter, whereas the rest consisted of living, pale-to- nongreen moss material. The top of the device was criss-crossed with fine wires, giving 16 intersec- tions which could be used as points for making cover estimates. We chose sample locations subjec- tively to provide single-species samples of crusts dominated by the most abundant cryptogams: the fruticose lichens Cladonia fimbriata (L) Fr. and C. macilenta Hoffm. and the mosses Brachythecium albicans (Hedw.) BSG and Didymodon vinealis (Brid.) Zander. The only other cryptogam in the study site, the moss Bryum capillare Hedw., was restricted to a single nebka and therefore we did not sample its biomass. We also required that sam- pled crusts be exclusively comprised of one taxon only and that it covered 100% of a minimum area of 6 X 6 cm. We extracted three samples of each species of crust from each of three nebkas. The samples were oven-dried at 105°C for 48 hr, at which time they ranged in weight between 50 and 140 g. Organic matter was chemically deter- mined by potassium dichromate reduction and spectrophotometric measurement (a modified Wilk- ley-Black method; Nelson and Sommers 1982) and the samples were sieved to determine sand texture. We measured changes in areas of nebkas A—D Over time by examining sequential aerial photo- graphs. We projected photographs onto a screen, traced nebka boundaries on graph paper, measured virtual areas, and converted them to actual areas by reference to the photograph’s scale. We estimated our range of error to be <10%. RESULTS Nebka size, topography, and age. Present nebka height increased from <1 m for nebka A to >4 m for nebka D (Table 1). Present nebka area in gen- eral increased from A to D, although nebka B was larger than nebka C. Recall that nebka D was on the windward end of an old, elongate ridge; we arbitrarily defined the sample area as the most northwestern 4000 m? portion. Nebka area thus dif- fered at the A and D extremes by an order of mag- nitude. In 1955, nebkas B, C, and D were present, but B and C were very small (30—40 m? each), two orders of magnitude smaller than at present (Table 2 and Fig. 2). We estimate that the year of origin AS OF SPRING, 1995. Within a row, numbers with different B C D 908 4000 2.0 29 4.6 283° 23 5) 10.8° 10.0° LO 3? | WAG hg 1998] TABLE 2. APPROXIMATE AREA (m?) OF EACH NEBKA AND OF A NEARBY MIDDEN OVER TIME. Based on interpretation of aerial photographs for all years prior to 1995, and on- the-ground-measurements in 1995. Nebka D, which re- mained constant at >4000 m7’, is not shown. Year Nebka A Nebka B Nebka C Midden 1955 O 30 40 O 1971 0) 160 115 100 1977 0 240 110 80 1988 290 1410 515 180 1990 470 2090 840 305 1991 440 2120 1120 375 1995 380 2205 910 400 for B and C did not long precede the 1955 photo- graph, and certainly was not equal to that of well- developed nebka D. Provisionally, we assigned an age of 45 yr to nebkas B and C. Nebka D—that is, the ridge to which it is the windward front—ap- pears to have been present in the 1862 topographic map. Provisionally, we assigned an age of 135 yr to nebka D. Nebka A did not appear until the 1988 photo- graph, at which time it measured nearly 300 m? in DANIN ET AL.: DUNE SUCCESSION 105 area. We interpret this relatively large size as in- dicating that its year of origin was closer to 1977 than to 1988. Provisionally, we assigned an age of 15 yr to nebka A. The present star-shaped appear- ance of nebka A may indicate that it was created by a single establishment event; that is, the star’s rays are the results of A. arenaria rhizomatous spread from the single establishment locale. The rate of growth in nebka area was unequal among the four nebkas (Table 2). It was essentially zero for nebka D. Nebkas A, B, and C grew in geometric fashion, slowly at first and then more rapidly over time. Within intervals, however, growth rates differed, nebkas A and C typically growing more slowly than nebka B (Table 2). Neb- ka B has increased in area 70-fold since 1955, whereas nebka C has increased only 25-fold. The current height of B, however, is nearly 1 m lower than that of C, thus aerial and vertical accretion seem to be uncoupled in this system. Growth for all three nebkas was slow in 1991-1995 compared to earlier periods of time. Vegetative cover and richness of species and growth forms differed appreciably among the four nebkas (Table 3). Nebka A, the presumed pioneer TABLE 3. WEGETATIVE COVER, IN 1995, ON THE FouR NEBKAS. Data, in percent, are averages of two line transects, with point data taken every 50 cm. T = <0.1%; — = not present. Growth form and species Nebka A Nebka B Nebka C Nebka D Shrubs Lupinus arboreus — 11.6 22.4 10.3 Baccharis pilularis --- -— 7.8 D2 Herbs Ammophila arenaria 40.1 67.7 41.8 90.9 Cardamine oligosperma 4.9 a D2 10.8 Claytonia perfoliata oy be P| 1.9 Lond Lotus heermannii 2.4 — — — Camissonia cheiranthifolia T 0.7 1.4 — Carpobrotus chilensis T os —- — Conyza canadensis T —- T —_ Brassica rapa — 0.9 — aa Gnaphalium purpureum — Li? — — Ambrosia chamissonis a 0.9 — — Crassula connata — 3.0 Lg — Spergularia macrotheca —- 0.9 del — Erechtites minima — 0.7 T 1A Carpobrotus edulis — — 4.1 — Daucus pusillus — _ T T Juncus bufonius — T — — Galium aparine = — --- ar Poa douglasii -—— T — — Solanum americanum — -— T — Cryptogams Bryum capillare — — — 0.8 Didymodon vinealis T 4.7 T 0.8 Brachythecium albicans T 0.9 | 10.5 Cladonia fimbriata and C. macilenta T T — —= Total number of species 10 16 15 11 Bare ground (%) 49.0 17.9 25.5 8.2 Number of samples 61 135 63 Bis 106 MADRONO TABLE 4. ORGANIC MATTER IN THE Top 2 cm OF BIOTIC Crust. Data are means of three samples, one sample per nebka. Within a column, data with a different superscript statistically differ at P < 0.5. Dominant cryptogam OM as % OM as gm” Cladonia spp. 30.34 497° Brachythecium albicans 29.63 408° Didymodon vinealis 31.9% 269° phase, had the fewest species and the lowest vas- cular plant and cryptogamic cover. Nebka A lacked the presence of shrub taxa. Nebkas B and C had the most species and an intermediate amount of plant cover. Nebka D had almost as few taxa as nebka A, yet exhibited the most plant and crypto- gam cover. The contribution of cryptogams to plant cover increased from trace, to 5.6, to 7.1, and to 11.3% on nebkas A-—D, respectively. Crytogamic cover was typically patchy, covering the entire surface in local 0.2—0.5 m? areas, while virtually absent else- where. The organic matter content of the topmost 2 cm in such patches averaged 30% by weight, or 391 g m~ crust. There was significantly less bio- mass per unit area for the moss D. vinealis among the three cryptogams sampled (Table 4), but the other taxa were not Statistically different from each other. From Table 3, we can estimate that 11% of a mature nebka’s surface would contain a biotic crust such as we sampled, and therefore the standing bio- mass would be (0.11 X 391 = 43 gm’). Sand texture differed among nebkas, coarse sand (2.0—0.5 mm) and very fine sand (0.25—0.05 mm) fractions showing the most difference. Gravel (>2 mm) and silt + clay (<0.05 mm) fractions com- bined never accounted for >1% of any sample’s weight. Surface and subsurface textures, however, were not statistically different, hence we have pooled those data in this paper. Textures at the neb- ka center and midway along the transects were in- significantly different among the nebkas, but they were different from texture at the nebka edge. Con- sequently, each nebka’s texture data in this paper represent average data for ten samples: surface and subsurface samples at the nebka’s center and mid- way along each of its four transects. Nebka sand [Vol. 45 texture was significantly finer than sand at nebka edges or at random internebka locations (Table 5). One layer of coarse sand, approximately 2 m be- low the current surface of several ridges and of nebka D, was uniquely coarsest of all samples. We think this stratum was the result of a period of in- tense erosion during which finer particles were blown away. Our examination of the series of aerial photographs lead us to conclude that this coarse layer was at the surface just after 1955. One reason for our choice of 1955 was the presence of a mid- den mound adjacent to nebka D (Fig. 2 and Table 2) in the 1971 photograph, but not visible in the 1955 photograph. By 1995, the midden’s exposed area had quadrupled from that in 1971, indicating that scouring of the unvegetated surface is continu- ing. The location of the coarse sand layer today, 2 m below the vegetated surfaces of ridges and nebka D, indicates an average sand accretion rate of 5 cm per year. This rate of accretion is modest compared to known rates for A. arenaria dunes elsewhere that approach 100 cm per yr (Barbour et al. 1985). At the same time, there must have been an equal amount of deflation between the ridges because the height of the coarse stratum is about 2 m above present internebka swales. DISCUSSION AND CONCLUSIONS The vegetation on nebkas A-—D fits within the definition of dune scrub dominated by A. arenaria formally classified as “‘European beachgrass”’ se- ries by Sawyer and Keeler-Wolf (1995) or as “‘Am- mophila-Erechtites”’ and “‘Ammophila-Baccharis”’ communities by Parker (1974). Ammophila arenaria today has a central impor- tance to nebka formation, vegetation succession, and the composition of dune scrub on stabilized substrates at Bodega Head. We can presume that modern dune succession and dune scrub differ from that of the last century (when Leymus mollis was the dominant sand-stilling grass) because published studies show that A. arenaria grows more densely than L. mollis and that it suppresses species rich- ness and cover from other taxa (Barbour and John- son 1988; Barbour et al. 1976; Pavlik 1983). Remnant areas of northern California dune scrub still dominated by L. mollis exist at Point Reyes National Seashore, Lanphere-Christensen Dunes TABLE 5. SorL TEXTURE. Coarse = a buried stratum found repeatedly beneath several nebkas and dune ridges. Inter- nebka data come from five random samples within the 50 X 150 m study area. Edge = the surface and subsurface samples combined from 16 samples at nebka outer edges. Within = the surface and subsurface samples combined from 20 samples taken within nebkas. Data to the far right represent soil texture just beneath biotic crusts of Cladonia, Brachythecium, and Didymodon. Within a row, numbers with different superscripts are statistically different at P < 0.5: Texture class Coarse Internebka Edge Coarse sand Lu" ie ale 38.5) Very fine sand 1 Plz 8.94 No. samples 2 S) 32 Within Clad Brach Didy PPA 18.4° 9.0° Wase 10:0: 153° 21.4° 26.3° 40 3 ) 5 1998] Preserve, just south of the mouth of Ten Mile River in Mendocino County, and in very local patches along Bodega Bay. This vegetation has been vari- ously described as “‘native dunegrass,” “‘yellow bush lupine,”’ and “‘coyote brush”’ series by Sawyer and Keeler-Wolf (1995), as “‘Baccharis-Scrophu- laria,”’ ‘‘Poa-Lathyrus,”’ and ‘Solidago spathula- ta-Lupinus arboreus dune mat’? communities by Parker (1974), or as “‘Lupinus arboreus-Haplopap- pus ericoides”’ association by Holton and Johnson (1979). Vegetation on nebkas A-D are related by pat- terns commonly reported for progressive succes- sion in general (Barbour et al. 1987). For example, total cover increases through the entire sere, but species richness peaks at an intermediate phase. We presume that the late-seral decline in species rich- ness is due to growing dominance by A. arenaria and its homogenization of the microenvironment, reducing niche diversity. The few studies of plant zonation and putative succession on California dunes suggest that the most highly correlated abiotic factors (possible drivers of succession) are: coarseness of the sub- strate, amount of soil organic matter, wind speed near the surface, intensity of salt spray, and degree of surface sand movement (Barbour 1992; Barbour et al. 1973; Barbour and DeJong 1977; Barbour 1978; Holton and Johnson 1979; Johnson 1963; McBride and Stone 1976; Parker 1974). Our results corroborate and add detail to the pub- lished pattern of increasingly fine sand texture ac- companying the progression of primary succession. The percent of coarse sand monotonically declined nearly an order of magnitude from internebka sand to sand at the edge of nebkas, to sand well within nebkas, and to sand within a cryptogamic crust on nebkas. At the same time, the percent of very fine sand more than doubled. Although we collected no data on the ecological effect of this textural cline, we can easily imagine that increasingly fine texture improves soil moisture retention. We can also infer that the driving factor for increasingly fine texture is declining surface wind speed, which allows smaller particles to settle out (Danin and Yaalon 1982; Danin 1996). Thus, gradients of sand move- ment, salt spray, and wind speed are all undoubt- edly interdependent in this ecosystem. Our observation of a widely distributed layer of exceptionally coarse sand gave us a possible datum from which to estimate rate of nebka building or internebka deflation. The 15-cm thick layer (Table 5; Fig. 4) had 39% more coarse sand and 34% less very fine sand than average internebka samples. It was visible at a consistent depth of about 2.0 m beneath stabilized ridges or large nebkas, and at a similar height above internebka swales and blow- outs. The coarse layer may represent surface ma- terial prevalent prior to the widespread planting of A. arenaria and the creation of a foredune which has since slowed surface winds. It could have been DANIN ET AL.: DUNE SUCCESSION 107 Fic. 4. the current vegetated surface of ridges and mature nebkas. Stratum of visibly coarse sand about 2 m below present on the surface as recently as 1955 based on our interpretation of aerial photographs. If so, its burial has been the result of sand-stilling vegeta- tion, trapping 2 m of sand in a span of 40 yr (av- erage annual rate of accumulation = 5 cm). We know that short-term sand accumulation rates driv- en by the presence of A. arenaria may approach 1 m yr! along the foredune (Barbour and Johnson 1988; Barbour et al. 1985), but there are no pub- lished records of long-term sand accumulation rates in the lee of the foredune with which we can com- pare our deduced rate of 5 cm yr“'. Our study has shown that the role of cryptogams in nebka formation and vegetative biomass accu- mulation is significant. The percent cover of a cry- togamic crust increased monotonically with pre- sumed nebka age and physical complexity. On the presumed oldest nebka D, the 2-cm-thick crust con- tained an average of 391 g organic matter (biomass) per square meter of crust, or 47 g m~ of nebka surface. This is not a trivial contribution to above- ground biomass. According to a more extensive California coast study by Barbour and Robichaux (1976) vegetation dominated by A. arenaria has a biomass of 1819 g m °’ dune surface when its cover is 100%. Nebka D had 91% cover by A. arenaria; ignoring any other plant cover, this would translate as (0.91 X 1819) 1655 g m? nebka surface. Thus, the cryptogamic biomass, 47 g m~’, is 3% of total above-ground biomass. We expect that the contri- bution of biotic crust to biomass is even higher be- yond the hinddune, where older and more stabilized dunes have a more continuous crust. However, there are no data yet collected for that part of the Bodega dunes ecosystem. 108 MADRONO We have also been able to conclude that a con- tinuous hinddune and foredune did not exist at Bo- dega Head prior to the introduction of A. arenaria. An 1862 survey map shows a gradual increase in base sand surface to the east and inland from tide- line, reaching about 25 m elevation and—super- imposed on this basal surface—a series of broken ridges running perpendicular to shore with maxi- mum heights of 40 m. The distribution of peak ridge heights resembled a string of islands roughly parallel to shore and situated about 470-500 m in- land. These peaks were approximately 100-150 m more shoreward of today’s 45-m-tall hinddune that lies 600—650 m inland. The development of the hinddune was already substantially complete by 1958 according to a map prepared by Pacific Gas and Electric Company (PGE 1958), only a decade after wide-spread planting of A. arenaria. The im- pact of Ammophila on foredune topography thus matches that described recently for the northwest coast of North America in general by Wiedemann and Pickart (1996). Finally, the rate of spread of A. arenaria on Bo- dega nebkas parallelled a pattern reported for more extensive dunes near Arcata, CA over the period 1939-1989 (Buell et al. 1995). In that study, A. arenaria first spread slowly, then entered a phase where the square root of area was linearly related to time in years. Bodega nebkas B and C (Table 2) showed a similar slow phase of increasing area be- tween 1955 and 1971, then a linear increase phase thereafter. The linear phase exhibited an average increase of 100 m? yr“! per nebka. LITERATURE CITED BaARBouR, M. G. 1978. Salt spray as a microenvironmental factor in the distribution of beach plants at Point Reyes, CA. Oecologia 32:213—224. BARBOUR, M. G. 1992. Life at the leading edge: the beach plant syndrome. Pp. 291-307 in U. Seeliger (ed.), Coastal plant communities of Latin America, Aca- demic Press, New York, NY. BARBOUR, M. G., J. BURK, AND W. D. Pitts. 1987. Ter- restrial plant ecology, 2nd ed. Benjamin-Cummings, Palo Alto, CA. BARBOUR, M. G., R. B. CRAIG, EF R. DRYSDALE, AND M. T. GHISELIN. 1973. Coastal ecology, Bodega Head. University of California Press, Berkeley. BaARBouR, M. G. AND T. M. DEJONG. 1977. Response of west coast beach taxa to salt spray, seawater inun- dation, and soil salinity. Bulletin of the Torrey Bo- tanical Club 104:29—34. BARBOUR, M. G., T. M. DEJONG, AND B. M. PAVLIK. 1985. Marine beach and dune plant communities. Pp. 296— 322 in B. EK Chabot and H. A. Mooney (eds.), Phys- iological ecology of North American plant commu- nities, Chapman and Hall, New York, NY. BARBOUR, M. G. AND A. F JOHNSON. 1988. Beach and dune. Pp. 223—261 in M. G. Barbour and J. Major (eds.), Terrestrial vegetation of California, 2nd ed., California Native Plant Society, Sacramento, CA. BARBOUR, M. G. AND R. H. ROBICHAUX. 1976. Beach phy- [Vol. 45 tomass along the California coast. Bulletin of the Tor- rey Botanical Club 103:16—20. BARBOUR, M. G., A. SHMIDA, A. EK JOHNSON, AND B. HOL- TON, JR. 1981. Comparison of coastal dune scrub in Israel and California: physiognomy, association pat- terns, species richness, phytogeography. Israel Jour- nal of Botany 30:181-198. BUELL, A. C., A. J. PICKART, AND J. D. STUART. 1995. Introduction history and invasion patterns of Ammo- phila arenaria on the north coast of California. Con- servation Biology 9:1587—1593. COoLLey, C. C. 1970. The missionization of the Coast Mi- wok Indians of California. The California Historical Society Quarterly 49:143—162. Cooper, W. S. 1936. The strand and dune flora of the Pacific coast of North America: a geographic study. Pp. 141-187 in T. H. Goodspeed (ed.), Essays in geo- botany, University of California Press, Berkeley. Cooper, W. S. 1967. Coastal dunes of California. Memoirs of the Geological Society of America, No. 104. Den- ver, CO. Cow Les, H. C. 1901. The physiographic ecology of Chi- cago and vicinity: a study of the origin, development, and classification of plant societies. Botanical Gazette 31:73—108 and 145-182. DANIN, A. 1996. Plants of desert dunes. Springer, New York, NY. DANIN, A. AND D. H. YAALON. 1982. Silt plus clay sedi- mentation and decalcification during plant succession in sands of the Mediterranean coastal area of Israel. Israel Journal of Earth Sciences 31:101—109. DAVIDSON, E. D. AND M. G. BARBovuR. 1977. Germination, establishment, and demography of coastal bush lupine (Lupinus arboreus) at Bodega Head, California. Ecol- ogy 58:592—600. DIAMOND, L. AND J. KENNEDY. 1975. Bodega Marine Re- serve: an environmental inventory and trail siting analysis. University of California Natural Reserve System, Berkeley, CA. Dor, I. AND A. DANIN. 1996. Cyanobacterial desert crusts in the Dead Sea Valley, Israel. Algological Studies 83:197—206. GEE, G. W. AND J. W. BUADER. 1986. Particle-size analy- sis. Pp. 383-411 in A. Klute (ed.). Methods of soil analysis, part 1, physical and mineralogical methods, 2nd ed., American Society of Agronomy, Madison, WI. GREENGO, R. E. 1951. Molluscan species in California shell middens. Reports of the University of California Archeological Survey, No. 13, Berkeley, CA. HICKMAN, J. C. (ed.). 1993. The Jepson Manual: higher plants of California. University of California Press, Berkeley. Hotton, B., JR. AND A. E JOHNSON. 1979. Dune scrub communities and their correlation with environmental factors at Point Reyes National Seashore, California. Journal of Biogeography 66:317—328. JOHNSON, J. W. 1963. Ecological study of dune flora, Hum- boldt Bay. Master’s thesis, Humboldt State Univer- sity, Arcata, CA. Koonic, I. B. 1963. The geologic setting of Bodega Head. Mineral Information Service, State of California, Di- vision of Mines and Geology 16(7):1-9. Lipps, J. H. AND E. M. Moorgs (eds.). 1971. Geologic guide to the northern Coast Ranges, Point Reyes re- gion, California. Geological Society of Sacramento, Sacramento, CA. McBribg, J. R. AND E. C. STONE. 1976. Plant succession 1998] on the sand dunes of the Monterey Peninsula, Cali- fornia. American Midland Naturalist 96:118—132. NELSON, D. W. AND L. E. SOMMERS. 1982. Total carbon, organic carbon, and organic matter, Pp. 539-579 in A. L. Page, R. H. Miller, and D. R. Keeney (eds.), Methods of soils analaysis, part 2, chemical and mi- crobiological properties, Monograph 9, 2nd ed., American Society of Agronomy, Madison, WI. PGE. 1958. Bodega Head topographic map, 1:24,000, based on aerial photography dated 28 May 1958. San Francisco, CA. PARKER, J. 1974. Coastal dune systems between Mad Riv- er and Little River, Humboldt County, California. Master’s thesis, Humboldt State University, Arcata, CA. PAVLIK, B. M. 1983a. Nutrient and productivity relations of the dune grasses Ammophila arenaria and Elymus mollis. 1. Blade photosynthesis and nitrogen use ef- ficiency in the laboratory and field. Oecologia 57: 227-232. PAvVLIK, B. M. 1983b. Nutrient and productivity relations of the dune grasses Ammophila arenaria and Elymus mollis. 11. Growth and patterns of dry matter and ni- trogen allocation as influenced by nitrogen supply. Oecologia 57:233—238. DANIN ET AL.: DUNE SUCCESSION 109 PAVLIk, B. M. 1983c. Nutrient and productivity relations of the dune grasses Ammophila arenaria and Elymus mollis. Ul. Spatial aspects of clonal expansion with reference to rhizome growth and the dispersal of buds. Bulletin of the Torrey Botanical Club 110:271-— 219; RODGERS, A. E AND D. KERR. 1862. Map of a part of the coast from Tomales Bay to Salmon Creek, California. United States Coast Survey Office, Washington, DC. ST. CLAIR, L. I. AND J. R. JOHANSEN. 1993. Introduction to the symposium on soil crust communities. Great Basin Naturalist 53:1—4. SAWYER, J. O. AND T. KEELER-WOLF. 1995. A manual of California vegetation. California Native Plant Soci- ety, Sacramento, CA. STANDING, J., B. BROWNING, AND J. W. SPETH. 1975. The natural resources of Bodega Harbor. Coastal Wetlands Series 11, California State Department of Fish and Game, Sacramento, CA. USDA. 1951. Soil survey manual. USDA Soil Survey, Handbook No. 18, Washington, DC. WIEDEMANN, A. M. AND A. PICKART. 1996. The Ammo- phila problem on the northwest coast of North Amer- ica. Landscape and Urban Planning 34:287—299. MADRONO, Vol. 45, No. 2, pp. 110-118, 1998 HYBRIDIZATION BETWEEN CERCIDIUM FLORIDUM AND C. MICROPHYLLUM (FABACEAE) IN CALIFORNIA C. EUGENE JONES, LARRY J. COLIN, TRUDY R. ERICSON, AND DEBORAH K. DORSETT Department of Biological Science, California State University, Fullerton, CA 92834 ABSTRACT Cercidium floridum A. Gray ssp. floridum and C. microphyllum (Torrey) Rose & I. M. Johnston (Fa- baceae) hybridize where they have overlapping distributions in an area between Earp, and Parker Dam, CA. In this area, the two species have substantially overlapping blooming times, but are normally eco- logically separated by habitat requirements with C. floridum ssp. floridum preferring the sandy washes and C. microphyllum preferring the volcanic rocky slopes of the Whipple Mountains immediately adjacent to the washes. The hybrids tend to be found only on the sandy terraces between the wash and the mountains or on the sand dunes in the area. The best diagnostic traits for distinguishing the parental species and their hybrids include leaflet length, banner petal width and color, legume shape in cross section, the degree of constriction between seeds in the legume, and the seed shape in cross section. Introgression from C. microphyllum into C. floridum ssp. floridum may be occurring, but there is presently only limited evidence of any reciprocal introgression. The taxonomic and evolutionary significance of hybridization between these two taxa is discussed. RESUMEN En un area entre Earp, California, y la represa Parker se han encontrado hibridos entre Cercidium floridum A. Gray ssp. floridum y C. microphyllum (Torrey) Rose & I. M. Johnston. Estas dos especies se encuentran separadas ecologicamente ya que el habitat preferido por C. floridum ssp. floridum esta re- presentado por lechos arenosos mientras que C. microphyllum prefiere las laderas montanosas de la cadena Whipple formadas por roca volcanica que bordean dichos lechos de rio. Sin embargo, en esta zona la distribuci6n de estas dos especies tienen una distribuciOn superpresta y ademas comparten un periodo de florecimiento comun. Los hibridos se encuentran primordialmente en las terrazas arenosas localizadas entre las montafas y los lechos arenosos o en las dunas que se encuentran en el area. Los razgos caracteristicos mas apropiados para distinguir entre las especies paternas y los hibridos incluyen el largo de las hojuelas foliolos, el ancho y color del petalo central superior, la forma del corte transversal de la vaina, el grado de constricci6n la vaina entre las semillas, y la forma del corte transversal de la semilla. Aunque se observa una aparente introgresion de C. microphyllum en C. floridum ssp. floridum, evidencia introgresiOn reciproca es limitada. El significado entre la taxonomia y la evolucion de hibridaci6én de estas taxas se discuten. While examining specimens of Cercidium flori- dum A. Gray ssp. floridum and C. microphyllum (Torrey) Rose and I. M. Johnston in the herbarium at Rancho Santa Ana Botanic Garden for a study of differences in ultraviolet (UV) floral patterns that act as a pre-pollination isolating mechanism (Jones 1978), we discovered a specimen collected by G. Wolf in 1940 that was annotated by him as a pos- sible hybrid between these two species. The spec- imen (Wolf 9722, RSA) was collected about 11 miles (18 km) north of Earp, along the road to Par- ker Dam in San Berdardino County, CA. A field examination of this area in May 1972 revealed some interesting plants that had leaf characters that appeared to be intermediate between C. floridum ssp. floridum and C. microphyllum. The area was reinvestigated in April 1973 when both species were in full flower and we identified several pos- sible hybrid individuals. The vegetative and reproductive characters of the hybrids and their parental species have been de- scribed elsewhere (Jones 1978) and the parental species have been discussed and described by Car- ter (1974a, b) and by Carter and Rem (1974). Car- ter (1974a), in her work on the genus Cercidium in the Sonoran Desert of Mexico and the Southwest- ern United States, indicated that she had found very few hybrids between C. floridum ssp. floridum and C. microphyllum. She cites only two herbarium specimens as examples, one of which was the Wolf 9722 (RSA) and the other was Kamb 2014 from Molina Crater, northwest of Sierra Pinacate in So- nora, Mexico. Because we had found several plants that appeared to be morphologically intermediate between these two species in the area around Earp, we decided to investigate the possibility that these individuals represented several additional examples of hybrids. Herein we describe our determination of these plants as hybrids based on distributional overlap and sympatry of the two parental species (Carter 1974a; Jones 1978), morphological inter- mediacy, ultraviolet floral pattern differences (as previously reported in Jones 1978), and reproduc- tive potential as determined by pollen stainability, 1998] JONES ET AL.: HYBRIDIZATION IN CERCIDIUM 111 TABLE 1. COLLECTION DATA FOR THE FIVE SAMPLE POPULATIONS OF CERCIDIUM. # of plants Popula- sam- tion Taxon represented pled Locality 1 C. floridum ssp. floridum 10 Base of the Whipple Mountains, adjacent to the Colorado River, 16.8 km north of Earp, on State Hwy. 62 toward Parker Dam, elevation 60.96 m, San Bernardino Co., CA. 2 C. floridum ssp. floridum 10 4.83 km south of the Iron Mountains, along north side of State Hwy. 62, elevation 610 m, Riverside Co., CA. 3 C. microphyllum 1a Same location as no. 1. 4 C. microphyllum ila 47.15 km south of Mexicali, Baja California, on Mex. Hwy. 5 near El Faro, in wash on side of road. 5 C. floridum ssp. floridum X | he Same location as no. 1. C. microphyllum (hybrid) seed germination, and artificial hybridization stud- ies. MATERIALS AND METHODS Field studies were conducted during spring and early summer months from May 1972 through July 1976. Five populations were selected for study (Ta- ble 1) and every Cercidium plant within these pop- ulations was examined. The Earp populations (1, 3 and 5) were selected to represent sympatric indi- viduals of C. floridum ssp. floridum and C. micro- phyllum and possible hybrids. Populations 2 and 4 were selected to represent allopatric individuals of the parental species. The distributions of the paren- 30°.N Tropic of Cancer ~~__ 7 7s -- te fe) 200 400 600 KILOMETERS C. floridum floridum Populations 1, 3 and 5 C. floridum peninsulare Population 2 Population 4 .. |C. microphyllum Fic. 1. Study sites and distribution of Cercidium flori- _ dum ssp. floridum, C. floridum ssp. peninsulare and C. microphyllum. Map based on Carter (1974a). tal species and the locations of the population sites are shown in Figure 1. Herbarium specimens of the 54 plants analyzed from the five populations are deposited in the Faye A. MacFadden Herbarium (MACFP) at California State University, Fullerton. After an extensive examination of field samples and a careful review of the detailed descriptions and analyses provided by Carter (1974a, b) and Carter and Rem (1974), 12 quantitative and 13 qualitative characters were chosen for study (Table 2). Dial calipers, accurate to 0.01 mm, were used to make the 12 quantitative measurements illustrat- ed in Figure 2. Five measurements per plant were taken for each quantitative character. Ultraviolet floral patterns were determined using techniques described by Jones (1978). Student’s t-tests or one- way ANOVA were completed to determine if sig- nificant differences exist between character states in allopatric versus sympatric populations. Varia- tion in the morphological characters for all plants from the five populations were analyzed graphically using a pictorialized scatter diagram and we used the hybrid index technique (Anderson 1949) to de- termine which plants to use for our artificial cross- es. Characters and index values assigned to each are presented in Table 2. Nectar samples taken from both parental types and the hybrids were sent to Drs. Irene (now de- ceased) and Herbert Baker at the University of Cal- ifornia, Berkeley for analysis. Pollen viability was estimated for each sample plant by staining the pollen with Cotton Blue (1% aniline blue in lactophenol) for at least 24 hours. Pollen from at least five separate flowers for each plant was used and a minimum of 400 grains were counted per plant. Those grains that stained dark blue and were of normal size and shape were con- sidered viable, whereas those that stained faintly or not at all or were misshapen were considered in- viable (Lawrence 1963). Seeds were collected from every plant sampled. At least 100 seeds from each plant were scarified by soaking in concentrated sulfuric acid for three 112 MADRONO [Vol. 45 TABLE 2. COMPARATIVE CHART OF MORPHOLOGICAL CHARACTERISTICS ILLUSTRATING THE INTERMEDIACY OF THE PUTATIVE HYBRIDS BETWEEN CERCIDIUM FLORIDUM Spp. FLORIDUM AND C. MICROPHYLLUM. The mean and the range, in parentheses, are given for each quantitative character. Characters used to construct the morphological hybrid index are indicated with an asterisk. Values assigned to the hybrid index characters are zero (O) for C. floridum ssp. floridum, two (2) for C. microphyllum and one (1) for the intermediate state. C. floridum ssp. floridum Hybrid C. microphyllum Hybrid index value 0) 1 2 Character Branchlet tips* without spinescent tips variable spinescently tipped Branchlet surfaces* glabrous variable pubescent Axillary spines* present present or absent absent Branch color* blue-green green to yellow-green yellow-green Leaflet color* blue-green green to yellow-green yellow-green Banner petal color* yellow cream white Banner petal orange dots* present present or absent absent Ovary* glabrous variable pubescent Apex of legume* broadly acute intermediate long accuminate Pod shape* flattened somewhat rounded rounded Fruit* not constricted between intermediate constricted between seeds Seed shape* oblong, flattened intermediate elliptic, rounded Leaflets, # of pairs* 2.8 (2-4) 3.6 (2-5) 6 (5-9) Leaflets, length (mm)* 4.1 (2.0—7.0) 3.3 (1.8—6.0) 1.6 (0.05—2.5) Banner petal length (mm) 11.9 (9-15) 11.1 (8-14) 8.7 (4-10) Banner petal width (mm)* 9.5 (7-12.5) 7.3 (4-14) 4.9 (3.1-7) Flower diameter (mm)* 22.5 (18-26) 20.2 (16—27) 16.0 (12-21) Anther length (mm) 2.0 (1.5—2) 2.1 (1.9-3) 2.0 (1.8—2.2) Rachis length (mm) 27.7 (14-45) 221 (LI—35) 31.1 (12-52) Sword length (mm) 9.0 (3-19) 14.1 (4.5—37) 16.3 (7-34) Length of pedicel, base of sepals to abscission layer (mm)* 3.3 (1.5—5.5) 2.1 (1.5—4) 1.5 (1.0-3) Length of pedicel, abscission layer to rachis (mm) 7.0 (4-10) 7.1 (4.5-14) 5.0 (3-8) Seed length (mm) 10.3 (8.8—12.2) 9.8 (8.2—12.4) 9.2 (7.0-11.9) Seed width (mm) 7.2 (5.1-8.4) 6.5 (5.3-8.2) 6.4 (5.4-7.6) Seed thickness (mm) 3.9 (3.3-4.9) 4.0 (3.1-5.1) 4.9 (3.9-6.1) Pollen stainability (%) 77.4 (39-97) 77.3 (54-93) 85.7 (21-98) Seed germination (%) 83.8 (58-100) 84.7 (46-99) 80.7 (32-96) hours, then rinsed with distilled water and planted in small plastic pots filled with standard potting soil mix. The percent germination was calculated for each plant. Artificial crosses were conducted in the field us- ing pollinator exclusion bags from Wards (Wards No. 20W-7300, which are no longer available). Both crosses between the parental species and backcrosses to the putative hybrid were attempted. Branches having numerous unopened floral buds were bagged and the buds allowed to open. Flowers to be used as the female parent, or pollen recipient, were emasculated in the bud. Pollination was ac- complished by removing pollen from mature an- thers with a dissecting needle and placing it on the stigma of an emasculated flower. After pollination, the pollinator exclusion bags were replaced. The bags were periodically examined and all fruits that developed were harvested at maturity. The seeds from these crosses were treated, and percentage germination was determined, as described for the wild-collected seed. RESULTS Comparative morphology. Twenty-two of the 25 morphological characteristics studied (Table 2) were shown to be of some value in the delineation of the parental taxa and in the establishment of the intermediacy of their hybrids. However, certain characters such as leaflet length, banner petal width and color, legume shape in cross section, the degree of constriction between seeds in the legume, and seed shape in cross section proved to be more di- agnostic. Analysis of population samples. Variation exhib- ited in allopatric and sympatric populations of the parental species is represented in Figure 3. Squares represent allopatric individuals of each parental species, and circles represent sympatric individuals of each. Group (A) represents C. microphyllum (populations 3 and 4) and Group (B) represents C. floridum ssp. floridum (populations 1 and 2). Based | on the morphological characteristics employed in | this study the two species are quite distinct, al- 1998] ie q FRUIT endview 7 | BANNER ; PETAL INFLORESCENCE FLOWER Fic. 2. Quantitative measurements of morphological characteristics as follows: a—a’, leaflet length; b—b’, le- gume sword length; c—c’, length of pedicel from base of sepals to abscission layer; c—c”, length of pedicel from abscission layer to rachis; d—d’, length of rachis; e—e’, length of seed; f—f’, width of seed; g—g’, thickness of seed; h-h’, length of banner petal; i-i’, width of banner petal; jj’, diameter of flower; and k—k’, length of anther. though there is greater range of variability in C. floridum ssp. floridum in most of the quantitative characteristics analyzed than in C. microphyllum. The differences between the allopatric and the sym- patric populations of C. floridum, ssp. floridum as reflected in Figure 3 and based on leaflet length and banner petal width, are in marked contrast to the lack of variation between the allopatric and the sympatric populations of C. microphyllum. Using a Student’s t-test, the leaflet length is significantly different at the 0.05 level (t = 2.78, df = 9) when the allopatric and the sympatric populations of C. floridum ssp. floridum are compared. Significant differences were not found when a comparison of allopatric and sympatric populations of C. micro- phyllum was completed. Similar influence of C. microphyllum on C. flor- idum ssp. floridum in sympatry is exhibited in ban- ner petal length (comparison of allopatric vs. sym- patric populations of C. floridum ssp. floridum were significant (P = 0.05, t = 2.38, df = 9). Although not significant at the 0.05 level, the diameter of the flowers showed a similar trend. In Figure 4, leaflet length and banner petal width are plotted for the hybrids, (represented by trian- _ gles) along with the parental types. This figure Clearly demonstrates the intermediacy, in leaflet length and banner petal width, of these hybrids be- tween the parental species. JONES ET AL.: HYBRIDIZATION IN CERCIDIUM 113 Nectar analysis. Table 3 summarizes the analysis of the nectar from both parental species and a pu- tative hybrid. These data indicate that in amino acid content the putative hybrid sampled (# 1016) is identical to the plant of C. microphyllum sampled (# 4010). Cercidium floridum ssp. floridum differs in lacking lysine and tryptophan and in having a somewhat reduced concentration of threonine. In sugar content, the putative hybrid has more fructose and glucose, but considerably less sucrose than ei- ther parental species. Fertility, experimental crosses and seed germi- nation. The range and average pollen stainability and seed germination for both parental species and the hybrids were so similar that no statistical tests were done. These results are included in Table 2. All of the F, and backcrosses attempted between C. floridum ssp. floridum and C. microphyllum or their hybrid produced viable seed. The results of at- tempted crosses are presented in Table 4. The hy- brid index value (HI) is given for each parent and the pollen stainability is reported for the male par- ent in each cross. DISCUSSION Cercidium floridum ssp. floridum and C. micro- phyllum have many obvious distinguishing charac- ters (Carter 1974a; Carter and Rem 1974; Jones 1978; Siemens et al. 1994), illustrated in Figure 5, including size and number of leaflets; position and size of branch spines; size and coloration of flower parts; degree of constriction between seeds in fruits; and shape, size, and color patterns of seeds. The hybrids tend to be intermediate in most of these characters (see also Siemens et al. 1994). It should be noted that some of the variation in flower color attributed to C. microphyllum in pre- vious studies (Carter 1974a, p. 48) is probably due to post-pollination changes in banner petal color and is not typical of unpollinated flowers. Carter (1974a) notes, “‘Flowers of C. microphyllum differ also in having the limb of the long-clawed upper petal [banner petal] white, or occasionally cream or pale yellow ...’’ Upon completion of pollination, the flowers of both C. microphyllum and C. flori- dum ssp. floridum undergo significant changes in floral color or symmetry that result in these flowers no longer being visited by the pollinating bees. In C. microphyllum, the change in banner petal color from white to cream or even yellowish is the result of a simple pH change. This change was duplicated by placing white banner petals in a weak sodium hydroxide solution and was reversed by immersing these treated petals in a weak solution of hydro- chloric acid. In C. floridum, ssp. floridum on the other hand, the color of the banner petal does not change, but the banner petal folds down over the stamens, thus changing the symmetry of these spent flowers. This latter type of post-pollination change was previously reported for Caesalpinia eriostach- 114 Leaflet Length (mm) Leaflet Length (mm) MADRONO O O 2 O O gO O Puig Banner Petal Width (mm) A O ; ee oS PHB Banner Petal Width (mm) O plants from allopatric pops 2 and 4 Seed shape in cross section Legume shape in cross section 0 plants from sympatric pops 1 and 3 © founded O rounded 4 hybrids, pop 5 O-_ intermediate Q somewhat rounded OA flattened fe) flattened Banner petal color Apex of legume Fruit constriction between seeds oO white O long, acuminate O constricted @ cream o intermediate =) somewhat constricted @ yellow } broadly acute FO not constricted [Vol. 45 1998] JONES ET AL.: HYBRIDIZATION IN CERCIDIUM iS TABLE 3. A BREAKDOWN OF THE MAJOR COMPONENTS OF THE NECTAR OF CERCIDIUM MICROPHYLLUM, C. FLORIDUM Ssp. FLORIDUM AND THEIR PUTATIVE HYBRID. + indicates presence, ++ indicates abundance, +/— indicates traces and — indicates absence. ' Osmic acid test. ? 2-6 dichlorophenol-indophenol test. * Dragendorff test. * Brom-phenol test. > p- nitraviline test. C. microphyllum C. floridum ssp. floridum Collection no. 4010 Sugars Fructose 16% Glucose 25% Sucrose 59% Maltose — Amino acids mg/ml 1 Alanine + Arginine a Asparagine se ho Aspartic acid + Glycine eb Histidine ++ Lysine 5 Proline + Serine ++ Threonine ++ Tryptophan + Lipids! + Antioxidants? _— Alkaloids? — Protein* + Phenolics? deep yellow ys (Jones and Buchmann 1974). In the hybrids, the banner petal in some flowers folds down over the stamens; and in other flowers, on the same plant, it stays erect and changes in color from cream to yel- lowish. As noted in Jones (1978), the two parental spe- cies differ in corolla size and in their response to ultraviolet light. The petals and stamens of the small-flowered Cercidium microphyllum are entire- ly absorbtive, whereas only the banner petals and stamens of the large-flowered C. floridum ssp. flor- idum absorb UV while the lateral petals are reflec- tive. The hybrid plants display large corollas simi- lar to those of C. floridum ssp. floridum but the flowers are entirely absorbtive (Fig. 5). Although not considered in this paper, it should be pointed out to other taxonomists who might con- sider the use of UV floral patterns as a taxonomic characteristic, that such patterns, if determined from herbarium specimens and depending on which flavonoid pigments are involved in the UV absorp- Hybrid 1016 4011 27% 22% 32% 23% 37% 50% 4% 5% 1217 0.78 + + ++ ++ +/— +/— + + ++ ++ ++ ++ + — - + ++ ++ ++ + oo —_ 7 + - + deep yellow pale yellow tion portion of the pattern, may change when the flowers are dried. These changes appear to com- monly occur when the UV absorbing pigments are anthochlors (Scogin, Young, and Jones 1977). Since anthochlor pigments are known to be asso- ciated with UV floral patterns in the genus Cerci- dium (Hiegel and Jones unpublished), caution should be exercised in making taxonomic judg- ments based on these traits; such changes can result in spurious variation. When extensive variation in UV floral patterns is detected from herbarium spec- imens, such as Carter (1974b) found in Cercidium Praecox and C. sonorae, a thorough investigation of living material should be undertaken to deter- mine if such variation exists in living plants. The flowers of C. floridum ssp. floridum have smaller banner petal widths when in sympatry with C. microphyllum and more closely resemble those found in C. microphyllum. The differences found in the banner petal width of C. floridum ssp. flori- dum collected from different populations may be <_ FIG. 3. (B) C. floridum. ssp. floridum. Pictorialized scatter diagram for Cercidium populations 1—4. Grouping (A) represents C. microphyllum and Fic. 4. Pictorialized scatter diagram for Cercidium populations 1—5 with hybrids plotted as triangles. Key to symbols the same as in Fig. 3. 116 MADRONO [Vol. 45 TABLE 4. RESULTS OF ARTIFICIAL CROSSES ATTEMPTED BETWEEN CERCIDIUM FLORIDUM, C. MCIROPHYLLUM AND THEIR PUTATIVE HysrRIDS. * +Female parent is listed first. ** Fruits and seeds eaten by rodents. Female parent Taxa Coll. # HI Coll. # C. floridum* X 1020 4 C. microphyllum 1019 C. floridum X 5011 B) C. microphyllum S012 C. floridum X 1020 4 hybrid 1018 C. floridum X S011 =) hybrid 5010 Hybrid Xx 5010 18 C. floridum 5011 Hybrid x 1018 15 C. floridum 2011 Hybrid X 2010 19 C. floridum 2011 Hybrid X 2010 19 C. microphyllum 1019 Hybrid x 5010 18 C. microphyllum 5012 C. microphyllum X 1019 33 hybrid 2010 C. microphyllum X 5012 33 hybrid 5010 C. microphyllum X 1019 33 C. floridum 2011 C. microphyllum X 5031 32 C. floridum 5001 best attributed to the influence of C. microphyllum on C. floridum ssp. floridum as mediated through the hybrids. The differences in leaflet length might be attributed to habitat moisture differences be- tween the allopatric and the sympatric populations of C. floridum ssp. floridum, which were not pres- ent in C. microphyllum populations. Cercidium flor- idum ssp. floridum exhibits an ecological prefer- ence for desert washes, whereas C. microphyllum tends to be found up out of the washes on the plains or hillsides (Carter 1974a; Jones 1978). The sym- patric site (population 1) for C. floridum ssp. flor- idum is along the Colorado River and although no exact quantification was undertaken, these washes appear to be more mesic than the washes where the allopatric population of C. floridum ssp. floridum (population 2) was sampled. Although there is no evidence of hybrid Cerci- dium progeny successfully outcompeting the paren- tal species in the sympatric zone in California (Jones 1978), the differences in floral morphology described above suggest the remote possibility of introgression through the hybrids of C. microphyl- lum and C. floridum ssp. floridum (see also Siemens et al. 1994). This interpretation, although the an- tithesis of an earlier study (Jones 1978), is sup- ported by the observation that the peak flowering of the hybrids shows greater synchronism with C. floridum spp. floridum than with C. microphyllum. This greater synchronism of flowering seems to be Male parent # fruits Percent % pollen # flowers devel- # seeds germina- HI _ stainability pollinated oped produced tion 97.5 7 2 3 —_—* 33 93.0 35 9 17 81.3 33 OLS 41 > 8 62.5 15 74.0 16 1 1 0.0 18 71.5 58 5 4 50.0 5 96.5 23 + 6 —_—* 2 96.5 19 1 0 — 2 OTD 29 4 7 42.9 255) 93.0 21 3 5 60.0 33 $5.5 24 2 3 33.0 19 74.0 oT 6 8 50.0 18 96.5 a2 6 9 44.4 2 81.0 17 pe 5 66.7 5 of greater importance than the fact that the hybrid | and C. floridum ssp. floridum have distinct UV and © visible floral patterns. The sugars, amino acids, and other components | present in the nectar of the hybrid appear to rep- resent simply a summation of those present in the two parental species (Baker and Baker 1976). The hybrid resembles C. microphyllum in amino acid and phenolic content but presents novel proportions of the simple sugars, and the presence of maltose reveals the genetic contribution of C. floridum ssp. floridum. Further evidence of hybridization was demonstrated by Siemens et al. (1994) using two- dimensional flavonoid spot patterns of both species and hybrids. Viable seed can be produced from interspecific | crosses involving C. floridum ssp. floridum and C. | microphyllum and from artificial backcrosses (Table 4). It is of interest to note that although there is © little evidence of any influence of backcrossing on © C. microphyllum, such backcross progeny are pos- — sible. In both parental species and the hybrids a | large range of pollen stainability percentages was | encountered. This may indicate an otherwise cryp- | tic influence of introgression on C. microphyllum, | as well as further evidence for backcrossing be- | tween C. floridum spp. floridum and the hybrids, | because most of the lower pollen stainability counts — were derived from sympatric individuals of the two parental species. 1998] Cercidium floridum Fic. 5. Hybridization between C. microphyllum and C. floridum ssp. floridum either is an uncommon event or environmental pressures prevent the hybrids from becoming established in greater numbers. The average frequency of hybrids found in the various populations studied in the area of sympatry in Cal- ifornia was only 3.2% (range 1.0—9.97%). This fre- quency of hybrids is particularly low given that these species are long-lived perennials living up to 400 years in age (Benson and Darrow 1944). Al- though the frequency of hybrids compares favora- bly with the frequency of ‘‘mistakes’’ made by pol- linating bees in areas of sympatry (Jones 1978), it _ Should be emphasized that in species with such slow replacement rates, minor differences in eco- logical requirements exhibited by these two species may result in significant selective pressures for the JONES ET AL.: HYBRIDIZATION IN CERCIDIUM 117 C. floridum ATES: re Bees C. microphyllum U.V. floral patterns Distinguishing characteristics of Cercidium floridum ssp. floridum, C. microphyllum and their hybrid. maintenance of certain genotypes in the population. As a result of such pressures, even fertile hybrid progeny would be eliminated from all but the rather extensive, naturally disturbed, floodplain, interme- diate habitats. Most of the hybrids were found in this habitat. This may help explain the very cryptic evidence of introgression found in the population of C. floridum spp. floridum that was sympatric with C. microphyllum. Although our work has substantially increased the knowledge about the number of known, natu- rally-occurring hybrid progeny between C. floridum spp. floridum and C. microphyllum, it appears that hybridization is limited to some peripheral sympat- ric zones; and as such is probably of little evolu- tionary consequence for these species. However, it would be of interest to thoroughly examine other 118 zones of sympatry to see if hybridization is occur- ring. For example, the specimen collected by Kamb (Kamb 2014) and cited by Carter (1974a) as a hy- brid from Sonora, Mexico, would serve as a start for more detailed studies on the extent over the vast area of sympatry of these two species. A collection trip in 1997 revealed another possible area of hy- bridization, between Quartzite, AZ and Blythe, CA where individuals of both species were simulta- neously in bloom. Isozymic studies of the two pa- rental species and their hybrids are currently un- derway and we anticipate the agreement of these data with our morphometric analyses (Mary Samp- son unpublished). Additionally, our work comple- ments a monographic study of Cercidium and Par- kinsonia completed by Dr. Julie Hawkins that sug- gests the possibility of widespread hybridization. At the present time, we agree with Carter (1974a) and Hawkins (1996) that the species are well-delimited entities and should be recognized as separate taxa. ACKNOWLEDGMENTS Special thanks to Paula McKenzie for Figures 2 and 5, to Susan Smith for Figures 3 and 4, and to Chirag Shah for computer assistance. Thanks to Drs. Irene (now de- ceased) and Herbert Baker for the nectar analysis. The Spanish abstract was provided by Jeffrey P. Colin, Dr. Rodrigo Lois, Dr. Marcelo Tolmasky, and Rosaedith Vil- lasenor. Thanks also to Dr. Gene Hiegel for his demon- stration of the pH nature of the banner petal color change in C. microphyllum. This study was supported in part by the National Science Foundation Grant GB-40082. LITERATURE CITED ANDERSON, E. 1949. Introgressive hybridization. John Wi- ley and Sons, NY. BAKER, I. AND H. G. BADER. 1976. Analyses of amino acids in flower nectars of hybrids and their parents, MADRONO [Vol. 45 with phylogenetic implications. New Phytologist 76: 87-98. BENSON, L. AND R. A. DARROW. 1944. A manual of south- western desert trees and shrubs. Bulletin no. 6, Uni- versity of Arizona Press, Tuscon, AZ. CARTER, A. M. 1974a. The genus Cercidium (Legumino- sae: Caesalpinioideae) in the Sonoran Desert of Mex- ico and the United States. Proceedings of the Cali- fornia Academy of Science 4th Series 40:17—57. CARTER, A. M. 1974b. Evidence for the hybrid origin of Cercidium sonorae (Leguminosae: Caesalpinioideae) of north-western Mexico. Madrofio 22:266—272. CARTER, A. M. AND N. C. REM. 1974. Pollen studies in relation to hybridization in Cercidium and Parkinson- ia (Leguminosae: Caesalpinioideae). Madrofio 22: 303-311. HAWKINS, J. A. 1996. Systematics of Parkinsonia L. and Cercidium Tul. (Leguminosae, Caesalpinioideae). Ph.D. dissertation. University of Oxford, U.K. JONES, C. E. 1978. Pollinator constancy as a pre-pollina- tion isolating mechanism between sympatric species of Cercidium. Evolution 32:189—198. JONES, C. E. AND S. L. BUCHMANN. 1974. Ultraviolet floral patterns as functional orientation cues in Hymenop- terous pollination systems. Animal Behavior 22:48 1— 485. KEVAN, P. G. 1978. Floral coloration, its calorimetric anal- ysis and significance in anthecology. Jn The pollina- tion of flowers by insects. A. J. Richards (ed.), Lin- nean Society Symposium Series 6:51—78. LAWRENCE, G. H. M. 1963. Taxonomy of vascular plants. The MacMillan Co., NY. Munz, P. A. 1968. A flora of California. University of California Press, Berkeley, CA. ScoGIN, R., D. A. YOUNG, AND C. E. JONES. 1977. An- thochlor pigments and pollination biology. II: The UI- traviolet floral pattern of Coreopsis gigantea (Aster- aceae). Bulletin of The Torrey Botanical Club 104: 155-159. SIEMENS, D. H., B. E. RALSTON, AND C. D. JOHNSON. 1994. Alternative seed defence mechanism in a palo verde (Fabaceae) hybrid zone: effects on bruchid beetle abundance. Ecological Entomology 19:381—390. MaproNno, Vol. 45, No. 2, pp. 119-127, 1998 THE ROLES OF SOIL TYPE AND SHADE INTOLERANCE IN LIMITING THE DISTRIBUTION OF THE EDAPHIC ENDEMIC CHORIZANTHE PUNGENS VAR. HARTWEGIANA (POLYGONACEAEBE) JopI M. McGRAw! AND ANNA L. LEVIN?? Board of Environmental Studies, University of California at Santa Cruz, Santa Cruz, CA ABSTRACT Understanding the ecological factors that cause narrow geographic range and habitat specificity is essential for the conservation of rare species of edaphic endemic plants. Here we investigated the relative roles of soil and light in limiting the distribution of Chorizanthe pungens Benth. var. hartwegiana Rev. & Hardham (Polygonaceae), an annual plant endemic to open patches of low nutrient soils in the sandhills habitat of the Santa Cruz Mountains, Central Coastal California. Seedlings were grown in a controlled pot experiment under three light conditions and five soil treatments. The growth, survival, and reproduc- tion of individual plants were compared. Plants were least successful when grown on their native low nutrient soil, suggesting that soil type is not a limiting factor in the taxon’s distribution. However, when grown under high shade, survivorship, growth, and reproduction of individuals were low. This suggests that shade intolerance is the major cause of this taxon’s restriction to open, sandy areas. Thus management to preserve this federally endangered species should include artificial or natural disturbances to prevent populations from being extirpated due to encroachment of taller, shade-producing species. The geographic distributions of plant species are determined by many factors. While history, geo- graphic barriers, and isolation all influence the dis- tribution of species, the ultimate determinant of where a taxon can be found is its inherited toler- ance to environmental factors (Kruckeberg and Ra- binowitz 1985). Climatic, biotic, and edaphic char- acteristics of a given environment determine the constraints or opportunities an individual plant fac- es, depending on its genetically determined physi- ologic capabilities (Mason 1946; Baskin and Bas- kin 1988). The question of what limits plant distri- butions is important both to biogeography and in assessing the threats to populations of rare species. Among the most striking of geographic factors affecting plant distribution are unique edaphic con- ditions, including unusual bedrock outcrops (e.g., serpentinite), nutrient poor soils, and varying water regimes (e.g., vernal pools). Areas with these un- usual conditions, which are considered inhospitable to the growth and reproduction of most plant spe- cies, are often inhabited by unique assemblages of specialized, morphologically distinct plant taxa. Due to their narrow habitat specificities and small geographic ranges, these endemic taxa are often among the rarest of plants (Rabinowitz 1981). Fur- ' Department of Integrative Biology, University of Cal- ifornia, Berkeley, CA 94720 e-mail: jmmcgraw @socrates.berkeley.edu * Department of Environmental Sciences, Policy, and Management, University of California, Berkeley, CA 94720 e-mail: alevin@nature.berkeley.edu > An equal time production; order of authorship deter- mined by coin toss. thermore, small population sizes, limited geograph- ic distributions, and generally poor competitive abilities render edaphic endemic plants highly vul- nerable to extinction due to stochastic events, hab- itat degradation, and invasion by weedy species (Harper 1981; Kruckeberg and Rabinowitz 1985; Janzen 1986; Falk 1991; Schemske et al. 1994). Here we report the results of an experiment de- signed to determine the relative roles of soil and light in limiting the distribution of a rare taxon en- demic to the Zayante soils of Central Coastal Cal- ifornia. Recent studies on edaphic endemic plants have attempted to determine the specific factors that al- low them to inhabit their typically harsh habitats, as well as the factors that confine them to those habitats (Latham 1983; Baskin and Baskin 1988; Buchele et al. 1989; Snyder et al. 1994). Three gen- eral hypotheses have been advanced to explain the habitat restriction of edaphic endemic plants. First, edaphic endemics may have specific chemical, physical, or biological requirements that are met only on a particular substratum (Walker 1954). Sec- ond, edaphic endemics may tolerate, though not re- quire, the inimical conditions where they typically occur, while they are excluded from more hospita- ble habitat due to competitive interactions with oth- er species. In particular, these often diminutive spe- cies may be limited to open communities where they are unshaded by taller, more competitive spe- cies that thrive in adjacent habitats (Baskin and Baskin 1988; Collins et al. 1989; Ware and Pinion 1990). Finally, edaphic endemics may be highly susceptible to soil pathogens found in nutrient-rich soils, and thus limited to low productivity areas in 120 which their exposure to pathogens is decreased (Tadros 1957; Latham 1983). In addition to these three ecological factors (soil requirements, shade intolerance, and susceptibility to soil pathogens) low genetic diversity has been suggested as the root cause of the poor competitive abilities hypothesized for edaphic endemics (Stebbins 1942). Little evidence supports the hypothesis that edaphic endemics are restricted to their habitats due to specific chemical, physical, or biological require- ments. Most edaphic endemic plants can be culti- vated more successfully in soils which do not con- tain the unique edaphic material in which they nor- mally grow (Walker 1954; Kruckeberg 1954; Hart 1980; Ware and Pinion 1990). In studies of edaphic endemic plants, many species were found to have relatively high levels of genetic variation, suggest- ing that a lack of genetic diversity is also unlikely to be a common factor restricting many edaphic endemic plants to edaphically inhospitable and sparsely vegetated areas (Baskin and Baskin 1988; Collins et al. 1989; Menges 1992). Although previous work has implicated light competition as a common limitation of edaphic en- demics, the relative importance of competition for light and the influence of various soil factors, in- cluding competition for nutrients and the limiting effects of soil pathogens, has not been thoroughly evaluated for any single species. In this study, we chose to examine the composite effects of soil types in combination with varying shading levels on a highly restricted edaphic endemic. To compare the effects of different limiting factors and to under- stand the importance of interaction among these factors, we grew individuals of the federally endan- gered edaphic endemic, Chorizanthe pungens Benth. var. hartwegiana Rev. & Hardham (Poly- gonaceae) (the Ben Lomond Spineflower) in a con- trolled growth experiment in which we varied shade levels and soil types. By measuring growth, survi- vorship, and reproduction of this rare annual plant, we were able to evaluate the combined importance of shading and soil factors in limiting performance of this highly restricted species, thus providing a functional understanding of its distributional limits. We chose a controlled growth experiment as a complementary approach to a previous study ex- amining the demographic performance of C. pun- gens var. hartwegiana through a reciprocal trans- plant experiment in the field (J. Kluse and D. Doak personal communication). Experimental manipula- tion of environmental conditions allowed us to ex- amine the specific effects of soil and shade while controlling for other factors. We note at the onset that we did not attempt to control or measure in- dividual aspects of soils. Instead, given that each soil type is very distinct, our goal was to evaluate the multiple effects of soil characteristics in their entirety. MADRONO [Vol. 45 METHODS AND MATERIALS The study species. Chorizanthe pungens var. hartwegiana is a diminutive annual plant found on many of the ‘islands’ of sandhills soils of the Santa Cruz Mountains, including the Bonny Doon Eco- logical Reserve, located at 37°03'N latitude, 122°08'W longitude. The taxon is further restricted to open, sandy, and frequently disturbed areas of | this soil type. Due to its limited population size, narrow geographic range, and habitat degradation, C. pungens var. hartwegiana was listed as federally endangered in February of 1994 (Federal Register | 1994). Though not noted as a separate species in The Jepson Manual, both the United States Fish and Wildlife Service and the California Native Plant Society recognize C. pungens var. hartwegiana as distinct. The population biology of C. pungens var. hart- wegiana previously has not been described in a peer-reviewed publication. Although little is known about the phenology of this rare plant, the basic seasonal pattern is similar to that of other winter- spring annuals. Seeds germinate in late fall after the first substantial rain in this region. The plants ma- | ture through the wet winter, then bolt and produce | branches, flower in April and May, and die soon | after seed production in June. To date, researchers have considered habitat de- struction to be the predominant threat to the taxon’s existence (Morgan and Marangio 1987). Sand quar- rying has already greatly reduced population num- bers, and at least half of the sandhills habitat cur- | rently occupied by C. pungens var. hartwegiana is | on property owned by sand and gravel companies | with plans to expand mining operations. Residential | development on smaller parcels also has eliminated | populations and fragmented the remaining habitat | (Federal Register 1994). The reduction in habitat | area and the resulting decrease in population size | greatly increases the likelihood of extinction due to environmental and demographic stochasticity (Falk | 1991). Changes in the disturbance regimes (e.g., fire suppression) appear to be further reducing pop- | ulations of C. pungens var. hartwegiana by allow- | ing for the encroachment of larger, more competi- | tive species such as Arctostaphylos silvicola Jepson | & Wiesl. (personal observation). The study site. The vegetation of the Santa Cruz | Mountains of the Central California Coast is com- ; prised of two main plant communities: Redwood | Forest and Hardwood Forest-Oak Woodland (Bar- | bour and Major 1977). Interspersed within these | two vegetation types are small islands of deep, | coarse sand derived from the highly erodible Zay- ante soil series, which were formed from weathered | marine sediment of Santa Margarita sandstone (Soil | Conservation Survey 1980). Over fifty separate is- | lands, comprising approximately 8000 acres, of © sandhills habitat have been mapped in the Santa | Cruz Mountains (Marangio 1985). Managed by the 1998] California Department of Fish and Game, the Bon- ny Doon Ecological Reserve (BDER) contains ap- proximately 120 acres of sandhills habitat set aside for preservation and ecological research. The well-drained, low nutrient soil of the san- dhills habitats supports the sandhills plant com- munities—unique assemblages of species found primarily or exclusively on these soils (Marangio 1985). Analogous to the distinct communities found on serpentinite soil, the sandhills flora con- sists of many diminutive annuals and shrub species, several of which are state or federally listed (Mor- gan and Marangio 1987). Three distinct types of sandhills vegetation are found on the BDER: maritime coast range ponder- osa pine forest, the endemic Arctostaphylos silvi- cola (manzanita) mixed chaparral, and sparse as- semblages of low-growing herbaceous species pop- ulating the open, sandy areas. Chorizanthe pungens var. hartwegiana occur almost exclusively in this last community type. The soils of the BDER are comprised mainly of the Zayante coarse sands derived from weathered marine sediment of Santa Margarita sandstone. However, the soil within each of the three plant communities is distinct in color, content, and humus level. In open areas of the reserve, the sands have remained relatively free of organic addition. How- ever, in the manzanita and ponderosa pine areas, organic matter from the overstory has created sandy loams. The redwood soil is composed of slightly acidic loams with high levels of organic matter characteristic of the Lompico-Felton complex. Sim- ilarly, the oak soil is composed of sandy loams of the Ben Lomond-Felton complex with a subsoil of clay loams (Soil Conservation Survey 1980). The upper layers of both soils are rich with decaying redwood and oak duff, respectively. Experimental design. To determine the relative effects of edaphic factors and light levels on C. pungens var. hartwegiana, we conducted a con- trolled growth experiment at the University of Cal- ifornia, Santa Cruz Arboretum. For this experiment, the plants were grown in five soil types. We chose to test performance in the three most predominant soils of the sandhills habitat (pure sand, manzanita, and pine) as well as the two most widespread soils bordering the BDER (oak and redwood forest). On 22 January 1994, we collected approximately 57 liters of each soil type from three sites within a single 10 X 10 m area, homogenizing it before use. The redwood soil was collected in a grove 3 km from the reserve’s northern border and the oak soil from an oak woodland 5 km east of the reserve. The sand, pine, and manzanita soils were collected within the reserve. To collect all soils, we cleared away any leaves, duff, or herbaceous ground cover and then removed the top 10 cm of soil. On 25 January 1994 we collected C. pungens var. hartwegiana seedlings from the open, sandy McGRAW AND LEVIN: ENDEMIC PLANT CONSERVATION A areas of the BDER by removing blocks of sand containing 10 to 30 seedlings each. The plants were transported in plastic bags and transplanted into standard 4 liter (15.2 cm rim diameter) pots within 4 h after removal from the reserve. We planted 5 seedlings in each of 12 pots of each soil type for a total of 300 individuals. Five seed- lings were planted in a circle within each pot such that each individual was 6 cm from the nearest neighbor and 3 cm from the edge of the pot. Plants for each pot were taken from several of the col- lected blocks of sand in order to mix the seedling sources. Extra plants were planted in separate pots for later use as replacements. The pots were placed in two protective wire cag- es in an unshaded field at the University of Cali- fornia at Santa Cruz Arboretum. On 1 February 1994, shade treatments were established by sus- pending neutral shade cloth 30 cm above the top of the pots. Two different densities of shade cloth were used to create the low and high shade treat- ments. Each cage contained one block of each of the three shade treatments (high, low, and no shade) and two pots of each soil type were placed under each shade treatment. Both the placement of the pots within each shade treatment and the placement of the shade treatments within each cage were ran- domized. Each plant was assigned a number so that we could record and track the fate of each of the 300 plants individually (20 plants per treatment combination). In order to prevent water from being a limiting factor in growth, plants were kept moist by water- ing with tap water whenever any one pot showed significant drying. In a second transplant on 3 February 1994, one week after initial transplanting, we replaced any seedlings which were dead or dying. Five days fol- lowing the second transplant initial size measure- ments of all 300 seedlings were taken. For each plant, we recorded the length of the longest leaf, a measure shown to have a strong positive correlation with total aboveground biomass in this taxon (Kluse 1994). Intermediate measurements were taken on 12 February, 13 March, and 11 April 1994. We recorded deaths and observations of plants infected by an unidentified rust common among C. pungens var. hartwegiana at the BDER. Plants in our experiment flowered between 21 April (78 days after transplant) and 12 June (118 days after transplant). As each plant flowered, we recorded a final measurement of longest leaf length and then harvested it by cutting the plant at ground level, taking care not to disturb the other plants in the pot. Plants were dried to a constant mass then weighed to obtain final aboveground biomass. Since the closed cages in our experiment pre- vented natural pollinator visitation, we could not measure seed production directly. However, Chor- izanthe spp. are known to produce a single, one- seeded flower within each involucre (Hickman 122 1993). Thus, we counted the number of involucres produced by each individual and used this as our measure of reproductive output. In order to compare the light levels under our shade treatments with those in the field, we mea- sured photosynthetically active radiation (400—700 nm) using an LAI-2000 Plant Canopy Analyzer. Photon flux measurements above and below each canopy type were considered in our experiment. These values were compared to percent transmit- tance measurements in our three shade treatments. To test for significant soil or shade effects on growth, we performed a 2-way analysis of variance (ANOVA) on longest leaf length measurements re- corded throughout the experiment. Similarly, we used 2-way ANOVA to test for soil and shade ef- fects on the reproduction and final biomass of C. pungens var. hartwegiana using (number of invo- lucres)* and In (final mass) as dependent variables (transformations to achieve normality). Due to the significant effect of initial seedling size on final biomass (P = 0.02), we performed all subsequent analyses of final biomass as ANCOVAs, using ini- tial leaf length measurements as a covariate. Pair- wise post-hoc comparisons (Tukey-Kramer) were then conducted to determine significant differences between means of final biomass and reproduction in the various shade and soil treatments. Finally, we used G-tests to test for treatment effects on sur- vivorship to final harvest. RESULTS Light measurements. Percent light transmittance values for sand, manzanita, pine, redwood, and oak habitats were 99.5%, 26.2%, 59.4%, 1.6%, and 3.2% respectively (SD = 1.2, 16.4, 30.2, 0.4, and 1.8, respectively). The no shade treatment allowed complete light transmittance, the low shade 38.6% transmittance, and the high shade 19.7% (SD = 2.4 and 0.5, respectively). Morphological observations. Throughout the growth period, substantial variation in plant mor- phology and time to flowering among the different treatments was observed. Plants in full sun grew prostrate with no basal stem and showed much sec- ondary branching. Their leaves were short, wide, had crinkled edges and turned deep orange to red with age. Low shade plants also were prostrate, yet had fleshier, smoother (lacking crinkled edges) leaves which did not redden with age. Low shade plants showed less secondary branching. The high shade conditions produced etiolated plants with highly elongated basal stems. These plants lacked the three-branched pattern typical of the species’ natural growth pattern. Their leaves were elongat- ed, thin, fleshy, and very light in color. Although there was overlap among flowering times in all treatments, plants that received more sunlight flowered earlier. Plants grown in full sun flowered first, as early as 21 April, 78 days after MADRONO [Vol. 45 initial transplantation. The low shade plants began to flower a week later. However, most of the high shade plants did not flower until the last week of harvest. Survivorship. Shading level had a significant ef- fect on the survival (G = 30.6, 2 df, P < 0.005) with 48% of the plants surviving under full sun, 55% under low shade, and 20% under high shade conditions (Fig. la). In contrast, soil type did not significantly affect survival (G = 6.6, 4 df, P = 0.127), but certain treatment combinations of soil and shade yielded incongruent results. For example, plants in oak soil under high shade conditions had the lowest survivorship (10%), while those in oak soil in low shade had one of the highest survival rates overall (65%; Fig. 1c). While this pattern sug- gests a possible interaction between soil and shade treatments, there was no significant interaction ef- fect (G = 11, 8 df, P = 0.202). In addition, high mortality of plants in oak soil under high shade resulted in a small sample size (1.e., two plants) for this treatment, thus caution must be taken in inter- preting these data. Reproduction. Shade had a highly significant ef- fect on the number of flowers produced (P < 0.001; Fig. 2a). The low shade plants averaged signifi- cantly higher flower output than both high shade and no shade plants (P = 0.001 and P = 0.049, respectively). While plants grown under no shade produced a greater average number of flowers than the high shade plants, this result was not significant. There was no significant effect of soil on flower output (P = 0.109; Fig. 2b) when the interaction effect between soil and shade is included in the analysis. However, when the interaction effect, which was not significant, was removed from the model, there was a significant effect of soil on flow- er output (P = 0.012). Plants in oak soil produced dramatically more flowers than those in any other treatments while plants in sand produced the fewest flowers; however pairwise comparisons showed no significant difference between flower output in any of the soil types. In all five soil types, low shade plants generated the greatest number of flowers, high shade plants the fewest, and no shade plants an intermediate number (Fig. 2c), paralleling the result that shade and soil did not significantly interact in their effects on reproduction. Final biomass. Analysis of the In(mass) by AN- COVA, using initial leaf length of transplanted seedlings as a covariate, showed that final mass was affected by shade treatment (P = 0.036, Fig. 3a). Plants in the low shade accumulated the greatest mass while no shade plants reached moderate mass- es and plants in the high shade had the lowest mass- es. However, the only significant difference by pair- wise comparison was between the low shade and the high shade (P = 0.027). Final biomass also was strongly affected by soil 1998] Percentage of Plants Survivng Shade Treatment Percentage of Plants Surviving Sand Manzanita oD 33 o's oD a3 o 2 & & es Soil Treatment Fic. 1. Survivorship of plants: a) in the three shade treat- ments; b) in the five soil treatments; c) grown in the 15 different soil and shade treatment combinations. All val- ues plotted are the percentages of C. pungens var. hartwe- giana in each treatment which survived until harvest. type (P = 0.004, Fig. 3b). Plants grown in the non- sandhill soils had the greatest final biomass with the biomass for plants grown in pine and manzanita soils intermediate, and plants in sand soil remain- ing, On average, very small. Both pine and redwood soils yielded plants with significantly higher bio- mass than sand soil (P = 0.008 and P = 0.015, respectively). Although plants grown in oak soil at- McGRAW AND LEVIN: ENDEMIC PLANT CONSERVATION 100 a. 3 ho} > 75 J 5 Dy 50 : 25 [o) a 0 {e) Zz E 3 s a Shade Treatment a b. S 3 > J & a. 5 3 (o) aa no} 8 o “4 a q = 3 o) io: Soil Treatment 150 C. 3 MS} 2 100 fs [] No Shade a, 5 Low Shade z 505 ee High Shade 0 ee us) 8 2 g ze 8 E 3 s ~ Soil Treatment Fic. 2. Mean number of flowers per plant: a) in the three shade treatments; b) in the five soil treatments; c) in the 15 different soil and shade combinations. All values plot- ted are the mean number of flowers produced by each plant + one standard error. Sample sizes ranged from two to fifteen among the different treatment combinations. tained the highest average mass, pairwise compar- isons showed no significant differences in biomass between oak and the other soil treatments. This is most likely due to the low sample size which re- sulted from high mortality in oak soil. There was no significant interaction effect be- 124 a. ae Ae | $3 Cet (o) (o) Z, 3 " 8 af Shade Treatment 15 2 a & ae o) 2 &: 2 < 2 2 & 3 6 E g Soil Treatment C. q [] No Shade A, S Low Shade 7] a = High Shade 5 2 Soil Treatment Fic. 3. Mean final aboveground biomass of C. pungens var. hartwegiana: a) in the three shade treatments; b) in the five soil treatments; c) in the 15 different soil and shade treatment combinations. Values plotted are the mean dry weights of plants + one standard error. Sample sizes ranged from two to fifteen among the different treatment combinations. tween soil and shade treatments on biomass. How- ever, when grown under low shade, plants in oak soil averaged the largest biomass; yet, when grown under high shade, plants in oak soil had the second to smallest final biomass (Fig. 3c). MADRONO [Vol. 45 Table 1 shows the percentage of the variance (measured as percentage of total sum of squares from a two way ANOVA) in the longest leaf length explained by the variables soil, shade, and the in- teraction of soil and shade at each of the three in- termediate measurements times in our study. By the eleventh day of the experiment, the interaction ef- fect between soil and shade treatment accounted for the greatest amount of variance not due to error (16%). Although they accounted for a small amount of variance, the separate effects of shade (7.8%) and soil (5.5%) were statistically significant. By the forty-fifth day of the experiment, the amount of variance caused by the interaction effect had de- creased to 10%, while the influence of both shade (8.7%) and soil (7.6%) had increased. This trend continued until the final measurement 113 days af- ter transplanting, when the effect of soil became the most prominent cause of variance (18.8%), and shade had increased slightly, to 8.9%, leaving the interaction effect to account for only 2% of the variance. DISCUSSION Ecological implications. Our results implicate shade as a primary factor in limiting the distribu- tion of C. pungens var. hartwegiana. Soil treatment also had significant effects on performance but, be- cause atypical soils were most conducive to growth, survival, and reproduction, soil per se cannot be viewed as a limiting factor in the distribution of C. pungens var. hartwegiana. Instead, results showing that high shade levels correlated with relatively low survivorship, reproduction, and final biomass im- plicate shade intolerance as the primary cause of the species’ restriction to open sandy areas. The relative importance of the shade regime, soil treatment, and their interaction upon individual per- formance varied throughout the growing season. We note here that while the initial sample size was 100 plants for each shade treatment and 60 plants for each soil type, only 20 plants comprise each soil/shade treatment combination. Therefore, cau- tion should be used when considering results of in- teraction tests. The interaction between soil and shade diminished during the course of the study. Conversely, the variation in plant size accounted for by both the shade and soil treatments increased throughout the experiment, with the soil effect more than tripling by the last measurement (Table 1). However, changes in the percentage of the vari- ance explained by the soil, shade, and interaction effect may be an artifact of the change in plant growth form that occurred throughout the experi- ment. From the time that the plants bolted, addi- tional growth in the form of increased branch length, and subsequently flowers, was observed, while leaf length remained constant. Our results show that the distribution of C. pun- gens var. hartwegiana is not restricted by soil char- 1998] TABLE 1. McGRAW AND LEVIN: ENDEMIC PLANT CONSERVATION 125 PERCENTAGE OF VARIANCE IN SIZE OF THE PLANTS EXPLAINED BY SOIL, SHADE, AND THEIR INTERACTION EFFECT AT THREE TIME INTERVALS DURING THE EXPERIMENT. Variance is measurement of the percentage of the total sum of squares accounted for by each independent variable in a two-way ANOVA on the longest leaf length, a measure of plant size. Both of the main effects and the interaction effect were significant at each sampling date (p<0.05). Time elapsed since transplant (percent) Source of variance Day 11 (3/13/94) Soil a5 Shade 7.8 Soil by Shade 16.2 Interaction Error 70.5 acteristics. In fact, all of our measures of perfor- mance were highest for individuals grown in the four soils where C. pungens var. hartwegiana does not naturally occur. The higher mass and flower number of plants in the manzanita, oak, pine, and redwood soils is perhaps due to the comparatively higher organic contents and greater water retention capacities of these soils relative to the sand soil. Many previous studies have used the addition of specific nutrients to controlled soil conditions to test the performance of edaphic endemics (Kruck- eberg 1954; Baskin and Baskin 1988). However, we chose to use actual soils from the habitat adjacent to the naturally occurring populations in order to infer plant performance in these habitats. The suc- cess of C. pungens var. hartwegiana in the four test soils does not support the hypothesis that chemical, physical, or biological requirements of this taxon are met only on the sandy soil. Although the soils with higher organic content, which have an inherently greater water holding ca- pacity, significantly increased final biomass (Fig. 3b), growth in these soils did not result in strong or consistent increases in reproduction over the plants grown in the sand soil (Fig. 2b). Instead, only shade level had a significant effect on flower number through its effect on plant morphology. Plants grown in low shade and no shade treatments had multiple branches, and this translated into more inflorescences. Chorizanthe pungens var. hartwe- giana growing naturally at the Bonny Doon Eco- logical Reserve were morphologically similar to those in the no shade treatment (personal observa- tion). Plants grown under the low shade treatment did consistently better in all three measurements of per- formance (biomass, survivorship, and reproduction) than plants grown in the no shade control treatment, which closely resembled natural light conditions. Low shade probably decreased water loss from evaporation and evapotranspiration while the cor- responding decrease in light did not significantly reduce plant performance. In contrast, plants in the high shade showed drastically reduced success compared with individuals in the full sun treatments as high shade conditions caused etiolation, poor survival (Fig. la), low flower output (Fig. 2a), and small biomass (Fig. 3a). Day 45 (4/21/94) Day 113 (5/20/94) 7.6 18.8 8.7 3.9 10.1 2.6 Wel 69.7 These results support our hypothesis that C. pun- gens var. hartwegiana is shade intolerant. Increased growth, survival and fecundity of plants in the man- zanita, oak, pine, and redwood soils, indicates that they are restricted from these areas because they are unable to compete for light with robust, com- mon species that naturally occur on these soils. This helps to explain the distribution of C. pungens var. hartwegiana in the sandhills communities of the Bonny Doon Ecological Reserve, where the taxon is restricted to otherwise unvegetated, pure sand soils. Management implications. The combination of habitat specificity and narrow geographic range of C. pungens var. hartwegiana renders the species vulnerable to extinction from both habitat degra- dation and the encroachment of other more com- petitive species. Habitat destruction by sand quar- rying and residential development have been con- sidered the major threats to the persistence of pop- ulations. Our results indicate additional threats to Species’ survival even within protected areas. In particular, shade intolerance increases vulnerability to extinction due to chaparral community succes- sion and the encroachment of alien species. The taxon’s inability to compete for light could explain its restriction to monospecific stands or open areas where it grows in association with other diminutive annual plants such as WNavarretia hamata E. Greene. At the Bonny Doon Ecological Reserve, the dis- tribution of C. pungens var. hartwegiana currently is limited to open, physically disturbed areas, such as trails and old roads, where chaparral species have not become established. Such open habitats are transitory and disappear in the absence of dis- turbance by fire, wind, or sand movement. Histor- ically, the open areas of the sandhills were prob- ably maintained by frequent fires (Potts 1993). However, fire suppression in the Santa Cruz Mountains may have allowed for the spread of chaparral species such as Eriodictyon californicum (Hook & Arn.) Torrey and Arctostaphylos silvi- cola, a sandhills specialist species, into formerly open, unshaded areas. Even in areas with partial shrub cover, C. pungens var. hartwegiana popu- lations are generally sparse. We speculate that, in 126 the absence of fire or other physical disturbance, succession will result in widespread extirpation of C. pungens var. hartwegiana. The invasion of alien weedy species poses a sim- ilar threat. These exotics not only create shade, but could also alter soil composition through nutrient addition (Vitousek 1990; Janzen 1986). Huenneke et al. (1990) found that increasing nutrient avail- ability through fertilizing classically low-nutrient serpentinite soils resulted in increased biomass of the serpentinite endemic vegetation. By the second year, however, this addition of nutrients had al- lowed for the invasion and eventual dominance of non-native vegetation. As native or exotic plants invade the pure sand areas of the sandhills, the deposition of nutrients can further facilitate the establishment of shade- producing species that were previously excluded due to low nutrient availability. The current inva- sion of smaller grasses (e.g., Vulpia myuros (L.) C. Gmelin) on the open sand areas of the Bonny Doon Ecological Reserve may not outcompete C. pun- gens var. hartwegiana for light; however, the cu- mulative alteration of the soil by this denser vege- tation could facilitate the invasion of larger, shade- producing species such as Genista monspessulanus (L.) L. Johnson. Encroachment by invasive species is especially threatening to the sandhills because of their small size and island-like geography. In par- ticular, the large edge to area ratio of small habitat fragments increases their susceptibility to invasion by aggressive species in surrounding habitats (Schierenbeck 1995). Given the shade intolerance of C. pungens var. hartwegiana and current threats of invasion, we suggest a management strategy aimed at the main- tenance of unvegetated, open areas in the sandhills through controlled burning and intentional mechan- ical disturbance. Because little is known about the fire ecology of C. pungens var. hartwegiana, small scale field tests and seed viability analyses should be conducted prior to implementing controlled burn management. However, due to its close association with chaparral vegetation and the frequent fire his- tory of the sandhills prior to suppression, it is likely that the taxon is fire tolerant. Alternatively, a mechanical method of distur- bance (e.g., bulldozing or hand clearing) could maintain open, unshaded areas required by C. pun- gens var. hartwegiana. Indeed, the largest popula- tions of C. pungens var. hartwegiana currently at the Bonny Doon Ecological Reserve exist on old road cuts and human trails. However, disturbance is a double-edged sword, as it often allows for the invasion of weedy species (Parker et al. 1993; Schierenbeck 1995). Local invasive exotics such as G. monspessulanus, and V. myuros are often found along the roads that border the reserve, apparently thriving because of disturbance. Future disturbance within the reserve could allow their spread into C. pungens var. hartwegiana habitat. To insure the vi- MADRONO [Vol. 45 ability of a disturbance-dependent plant, it is im- portant to match the intensity and timing of novel disturbances with the natural disturbance regime to which the species is adapted (Pavlovik 1994). Therefore, we recommend further study on the re- sponses of C. pungens var. hartwegiana to new dis- turbances prior to the widespread implementation of mechanical biomass removal or controlled burn- ing. In summary, the results of our study strongly suggest that a laissez faire approach to the conser- vation of the sandhills flora is not sufficient. Estab- lishing reserves free of habitat conversion and de- struction is the first step toward conservation of C. pungens var. hartwegiana. However, it is also es- sential to prevent the additional alterations in hab- itat caused by native chaparral succession in the absence of fire or alien species invasion from sur- rounding areas. Fortunately, the implementation of such management practices is feasible due to the small size of the sandhills habitats. Active conser- vation efforts aimed at controlling encroaching veg- etation are likely to protect other similarly adapted endemics of the sandhills community in addition to preserving the remaining populations of C. pungens var. hartwegiana. ACKNOWLEDGMENTS We thank J. DeWald and D. Hillyard of the California Department of Fish and Game, members of the Winter 1994 Conservation Practicum, and B. Hall for assistance with this research. Thanks to our families, David Ranney, and Adam Sullivan for their varied and abundant support. We also acknowledge the helpful suggestions from two anonymous reviewers. Finally, we give infinite thanks to Dan Doak whose contributions span a wide spectrum. Fa- cilities supporting preparation of this manuscript were provided by National Science Foundation DEB-9424566 to D. F Doak. LITERATURE CITED BARBOUR, M. G. AND J. Mayor. (eds.). 1877. Terrestrial vegetation of California. New York, John Wiley and Sons. BASKIN, C. C. AND J. M. BASKIN. 1988. Endemism in rock outcrop plant communities of the unglaciated eastern United States: an evaluation of the roles of edaphic, genetic, and light factors. Journal of Biogeography 15:829—840. BUCHELE, D. E., J. M. BASKIN, AND C. C. BASKIN. 1989. Ecology of the endangered species Solidago shortit: V. Plant associates. Bulletin of the Torrey Botanical Club 119:208—213. CoLLins, S. L., G. S. MITCHELL, AND S. C. KLAR. 1989. Vegetation-environment relationships in a rock out- crop community in southern Oklahoma. American Midland Naturalist 122:339-348. FaLK, D. S. 1991. From conservation biology to conser- vation practice. Pp. 338—431 in D. A. Falk and K. E. Holsinger (eds.), Genetics and conservation of rare plants. Oxford University Press, New York. FEDERAL REGISTER. 1994. Vol. 59, No. 24, Friday February 4, 1994, pp. 5499-5509. 1998] Harper, J. L. 1981. The meanings of rarity. Pp. 189-203 in H. Synge (ed.), The biological aspects of rare plant conservation. John Wiley & Sons, New York. Hart, R. 1980. The coexistence of weeds and restricted native plants on serpentine barrens in southeastern Pennsylvania. Ecology 61:668—701. HICKMAN, J. C. 1993. Chorizanthe. Pp. 856-860 in J. C. Hickman (ed.), The Jepson manual: higher plants of California, University of California Press, Berkeley, CA. HUENNEKE, L., S. P HAMBURG, R. KOIDE, H. A. MOONEY, AND P. M. ViTousEK. 1990. Effects of soil resources on plant invasion and community structure in Cali- fornian serpentine grassland. Ecology 71:478—491. HURLBERT, S. H. 1984. Pseudoreplication and the design of ecological field experiments. Ecological Mono- graphs 54:187-211. JANZEN, D. H. 1986. The eternal external threat. Pp. 286— 303 in M. E. Soule (ed.), Conservation biology: the science of scarcity and diversity. Sinauer Associates, Inc. Sunderland, MA. KLusE, J. 1994. The effects of habitat on the demographic performance of Chorizanthe pungens var. hartwegi- ana (Polygonaceae). Senior thesis. University of Cal- ifornia, Santa Cruz. KRUCKEBERG, A. K. 1954. The ecology of serpentine soils: plant species in relation to serpentine soils. Ecology 35:267-274. KRUCKEBERG, A. K. AND D. RABINOWITZ. 1985. Biological aspects of endemism in higher plants. Annual Review of Ecological Systematics 16:447—479. LATHAM, R. 1983. Are the piedmont serpentine barrens fire dependent communities? Unpublished paper, Swarthmore College Dept. of Biology, 33 pages. Mason, H. L. 1946. The edaphic factor in narrow endem- ism: the nature of environmental influences. Madrono 8:209-226. MccarTEN, N. FE 1987. Ecology of the serpentine vege- tation in the San Francisco Bay region. Pp. 335-339 in H. Synge (ed.), The biological aspects of rare plant conservation. California Native Plant Society, Sacra- mento, CA. MENGES, E. S. 1992. Habitat preferences and response to disturbance for Dicerandra frutescens, a Lake Wales Ridge (Florida) endemic plant. Bulletin of the Torrey Botanical Club 119:308—313. MarRANGIO, M. S. 1985. Preservation study: sandhills bi- otic communities of Santa Cruz County, California. Professional Project, Department of Landscape Ar- chitecture, University of California at Berkeley. MoraGan, R. AND M. S. MaRANGIO. 1987. The endangered McGRAW AND LEVIN: ENDEMIC PLANT CONSERVATION 128 sandhills plant communities of Santa Cruz County. Pp. 267-273 in T. S. Elias, (ed.), Conservation and management of rare and endangered plants. Califor- nia Native Plant Society, Sacramento, CA. PARKER, I. M., S. K. MERTENS, AND D. W. SCHEMSKE. 1993. Distribution of seven native and two exotic plants in a tallgrass prairie in southeastern Wisconsin: the importance of human disturbance. American Mid- land Naturalist 130:43—55. PAvLOvik, N. B. 1994. Disturbance-dependent persistence of rare plants: anthropogenic impacts and restoration implications. Pp. 159-193 in M. L. Bowles and C. J. Whelan (eds.), Restoration of endangered species. Cambridge University Press, Great Britain. Potts, K. 1993. Vegetation survey and mapping of the Bonny Doon sandhills ecological reserve. Senior the- sis, Environmental Studies, University of California at Santa Cruz. RABINOWITZ, D. 1981. Seven forms of rarity Pp. 205—217 in H. Synge (ed.), The biological aspects of rare plant conservation. John Wiley and Sons, Ltd., New York. SCHEMSKE, D. W., B. C. HUSBAND, M. H. RUCKELSHAUS, C. GOODWILLIE, I. M. PARKER, AND J. G. BISHOP. 1994. Evaluating approaches to the conservation of rare and endangered plants. Ecology 75:584—606. SCHIERENBECK, K. A. 1995. The threat to California flora from invasive species; problems and possible solu- tions. Madrono 42:168—174. SNYDER, K. M., J. M. BASKIN, AND C. C. BASKIN. 1994. Comparative ecology of the narrow endemic Echin- acea tennesseensis and two geographically wide- spread congeners: relative competitive ability and growth characteristics. International Journal of Plant Science 155:57-—65. SOIL CONSERVATION SERVICE. 1980. Soil Survey of Santa Cruz County. United States Department of Agricul- ture and University of California Agriculture. Pp. 16— 85. STEBBINS, G. L. 1980. The genetic approach to problems of rare and endemic species. Madrono 6:241-—272. Tapros, T. M. 1957. Evidence of the presence of an eda- pho-biotic factor in the problem of serpentine toler- ance. Ecology 38:14—23. VITOUSEK, P. M. 1990. Biological invasions and ecosystem processes: towards an integration of population biol- ogy and ecosystem studies. Oikos 57:7-13. WALKER, R. B. 1954. The ecology of serpentine soils: in- troduction. Ecology 35:275—288. WarRE, S. AND G. PINION. 1990. Substrate adaptation in rock outcrop plants: eastern United States Talinum (Portulaceae). Bulletin of the Torrey Botanical Club 117:284—290. MADRONO, Vol. 45, No. 2, pp. 128-130, 1998 ATRIPLEX LONGITRICHOMA (CHENOPODIACEABE), A NEW SPECIES FROM SOUTHWESTERN NEVADA AND EAST-CENTRAL CALIFORNIA HOWARD C. STUTZ Department of Botany and Range Science, Brigham Young University, Provo, UT 84602 GE-LIN CHU Institute of Botany, Northwest Normal University, Lanzhou, Gansu, China 7300700 STEWART C. SANDERSON USDA Forest Service Shrub Sciences Laboratory, Provo, UT 84601 ABSTRACT Atriplex longitrichoma, a newly reported annual species from southwestern Nevada, and east-central California is described and illustrated. It is most abundant in Pahrump Valley, NV, in abandoned agri- cultural fields, in roadside borrow pits; and, in favorable years, occurs in contiguous, undisturbed sites. It is a tetraploid species, apparently most closely related to Atriplex hillmanii (Jones) Standley. INTRODUCTION In Pahrump Valley, Nye Co., NV and in neigh- boring Stewart Valley, Inyo Co., CA, there are sev- eral populations of a previously undescribed annual species: Atriplex longitrichoma sp. nov. It is most abundant within and around the community of Pah- rump, NV, in abandoned cultivated areas, alongside roadways, and in other disturbed sites. In favorable years, it extends into neighboring, undisturbed, des- ert areas. In some localities, A. longitrichoma oc- curs aS a near monoculture, in others it grows in association with several other species. Atriplex longitrichoma Stutz, Chu, & Sanderson sp. nov. (Fig. 1)—TYPE: USA, Nevada: Nye Co., W side of Pahrump, T20S R53E S14, abundant. S. Sanderson & G. L. Chu 95303 26 May 1990 (holotype, BRY; isotypes, CAS, DAV, GH, MO, RENO, RSA, UC, UCR). Herba annua. Caulis erectus, 10—20 cm altus, Sparsim ramosus; rami basales plerumque decum- bentes, 10—30 cm longi, teretes, leviter flexuosi, tri- chomatibus clavatis elongatis dense tectus. Folia sessilia vel inferiora aliquando petiolata; lamina an- guste elliptica usque ovato-elliptica, 25-35 mm longa, 10-13 mm lata, integra, acuminato apice, cuneato base anguste, utrinque dense furfurascentes trichomatibus elongatis fragilibus, subtus cinereo- viridia, supra viridia; anatomia foliaris Kranz-typi. Staminati et pistillati flores in glomerulis mixti, ax- illares; perianthium floris staminalis hemisphaeri- cum vel infundibuliforme, 1.5—2.0 mm diam., pler- umque 5-partitum, apicibus segmentorum incurva- tibus; antherae oblongae, ca. 0.6 mm longae, saepe purpureo-rubellae, leniter exsertae sub anthesi; fi- lamenti filiformes, compressi, ca. 1.5 mm longi; bracteolae floris pistillati marginibus connatae infra medium; stigmata 2, filiformia, ca. 2 mm longa; stylus minus quam 1 mm, inconspicuus. Bracteae fructiferae oblongae usque late ovatae, 5-6 mm longae, 5—6 mm latae, trichomatibus elongatis fra- gilibus dense tectus, stipite brevi, basibus et areis centralibus induratis, marginibus dentibus lanceo- latis irregularibus leviter curvis, utrinque appendi- cibus numerosis mollo-spinescentibus 2—3 mm lon- gis ferentibus. Utriculus ovatus, 2—2.5 mm latus, pericarpio membranaceo. Semen flavo-brunneolum, perispermate duro; radicula supera. Annual herb. Stem erect, 10—20 cm tall, sparsely branched, lower branches 10—30 cm long, usually decumbent, frequently longer than stem, terete, slightly flexuous, densely covered with clavate, sin- gle-cell, elongate trichomes (Figs. 2, 3). Leaves ses- sile or lower leaves sometimes petiolate; leaf blade narrowly elliptical to ovate-elliptical, 25-35 mm long, 10-13 mm wide, apex acuminate, base nar- rowly-cuneate, entire, becoming densely furfura- ceous on both surfaces with elongate, fragile tri- chomes, gray-green abaxially, green adaxially; Kranz-type anatomy. Flowers in mixed axillary glomerules; male flowers most abundant in glom- erules on the upper branchlets, perianth half-glo- bose or funnel-shaped, 1.5—2.0 mm in diam., usu- ally 5-parted, segment apices incurved, stamens as many as perianth segments, anthers oblong, ca. 0.6 mm long, frequently reddish purple, slightly ex- serted when flowering, filaments filiform, com- pressed, ca. 1.5 mm long; female flowers present in mixed glomerules throughout all branches, margin of bractlets united below the middle, stigmas 2, fi- liform, ca. 2 mm long, style very short, less than 1 mm. Fruiting bracts oblong to broad-ovate, 5—6 mm long with short stalk, 5-6 mm wide, center part and base indurate, with several soft-spiny 2-3 mm long 1998] 20.0 mm MARKS vince wr Fic. 1. Vincent.) appendages on both surfaces, margins with lanceo- late, irregular, slightly curved teeth, densely furfu- raceous with elongate, fragile trichomes. Utricle ovate, 2—2.5 mm broad, pericarp membranaceous. Seed yellow-brown, with solid perisperm; radicle Fic. 2. and A. hillmanii (right). Bar = 3 mm. Fruiting bracts of Atriplex longitrichoma (left) STUTZ ET AL.: ATRIPLEX LONGITRICHOMA SP. NOV. 9 2.0 mm Atriplex longitrichoma. a. Habit. b. Fruiting bract. c. Seed. d. Embryo. e. Male flowers. (Drawings by Marcus superior. Flowering and fruiting period: April—June. Chromosome number: 2” = 36. Paratypes. USA, California: Inyo Co., Stewarts Valley, Shoshone Road, on alkaline area, R. S. Fer- ris 7365 26 April 1928 (CAS, US); Stewarts Valley, E of Resting Spring Range, S of Dry Lake, ca. 2500 feet, M. DeDecker 5463 23 June 1983 (RSA); 20 mi E of Shoshone, H. C. Stutz 95534 3 June 1991 (BRY); Nevada, Nye Co., 5 mi W of Pahrump, S. Sanderson & G. L. Chu 95301 25 May 1990 (BRY); 5 mi. W of Pahrump, S 7st., H. C. Stutz 95473 17 April 1991 (BRY); W side of Pahrump, dense population covering ca. 10 acres but all ripe and dead, H. C. Stutz & G. L. Chu 9760 27 June 1995 (BRY). Distribution and habitat. Atriplex longitrichoma is currently known only from a small area in Pah- rump Valley, Nye Co., NV, and neighboring Stew- art Valley, Inyo Co., CA. It is abundant in and around Pahrump, NV, occurring in open desert communities, alongside roadways, and in agricul- tural fields. The soils are gypsiferous clays with pH of ca. 6.5. 130 MADRONO [Vol. 45 Fics. 3, 4. wm). 3. A. longitrichoma. 4. A. hillmanii. Taxonomic relationships. Atriplex longitrichoma appears to be most closely related to A. hillmanii (Jones) Standley but differs in several conspicuous characteristics including much larger fruiting bracts (S—6 mm long, 5—6 mm wide, vs. 3—4 mm long, 3-4 mm wide), bearing large, curved, marginal teeth (Fig. 2); decumbent branching habit in which the lower lateral branches are longer than the cen- tral branches; and the presence of a copious coating of elongate trichomes on all stems, leaves, and fruits (Fig. 3). Trichomes of A. hillmanii are spher- ical (Fig. 4). So abundant and conspicuous are the trichomes of A. longitrichoma that when plants are collected and placed in a container such as a paper sack, the deciduous trichomes accumulate in the bottom of the sack in quantities sufficient to permit them to be picked up with the fingers or by the spoonful. Atriplex longitrichoma also differs from A. hill- manii in having narrow-elliptic to ovate-elliptic vs. oval-deltoid leaves. The leaf margins of A. longi- trichoma are always entire whereas the leaf margins of A. hillmanii are often, but not always, sparingly toothed. Atriplex longitrichoma differs from A. argentea Nutt. by its copious deciduous trichomes, narrower leaves (10-13 mm vs. 25—35 mm) and terete vs. quadrate branches. Fruiting-bract appendages of A. longitrichoma are flat or conical, whereas those of A. argentea are often folded. Chromosomally, A. argentea is diploid (2n = 18), A. longitrichoma is tetraploid (2n = 36), and A. hillmani is mostly tetraploid (2n = 36), although some plants are diploid (2n = 18) and others are hexaploid (2n = 54). Scanning electron micrographs of trichomes of Atriplex longitrichoma and A. hillmanii (scale bars = 100 Associated species. Atriplex longitrichoma some- times grows as a near monoculture but most often grows in association with Atriplex canescens (Pursh) Nutt., Atriplex confertifolia (Torrey & Fré- mont) S. Watson, Bromus madritensis L. ssp. rub- ens (L.) Husnot, Hordeum marinum Hudson, Lar- rea tridentata (DC.) Cov., Prosopis glandulosa Torrey, Salsola tragus L., or Suaeda moquinii (Tor- rey) E. Greene. Phenology. Flowering and fruiting of Atriplex longitrichoma is in early spring (April and May). By mid-June the fruits are fully mature and the plants are mostly dead. In early June, 1996, no liv- ing plants of A. longitrichoma could be found any- where. However, many dead plants, left over from 1995, were present in most areas where they had been found earlier. Apparently they were unable to grow during the severe drought of 1996. Since such droughts occur quite often in these deserts, seeds very likely remain dormant in seed banks during unfavorable years. Plants grown in the greenhouse and nursery at Brigham Young University, Provo, UT, from seed collected from populations near Pahrump, NV, showed the same characteristics as plants growing in the native populations, indicating high heritabil- ity of the distinctive features. ACKNOWLEDGMENTS We thank BHP and Brigham Young University for fi- nancial assistance, Timothy S. Ross for his numerous helpful suggestions with the manuscript, and the curators of the following herbaria for loans of specimens and ac- cess to their collections: CAS, RENO, RSA, UC, UCR. Maprono, Vol. 45, No. 2, pp. 131-136, 1998 EARLY SECONDARY SUCCESSION FOLLOWING CLEARCUTS IN RED FIR FORESTS OF THE SIERRA NEVADA, CALIFORNIA R. E Fernau,! J. M. REY BENAYAS,”? AND M. G. BARBOUR Environmental Horticulture Dept., University of California, Davis, CA 95616 ABSTRACT Vegetation was quantified for clearcuts, age 4—32 yr, of Abies magnifica Andr. Murray old-growth forests along the west face of the central Sierra Nevada. TWINSPAN analysis of 113 sites < 87 common taxa generated six ecofloristic units mainly related to each other on a time-since-harvest basis, but also exhibiting correlations with slope, elevation, latitude, soil depth, harvest area, and ratio of edge-to-area. Nearly half the taxa showed non-random distributions among the TWINSPAN units, but only a minority of those could be related to time since harvest. The herbs Gayophytum diffusum Torrey & A. Gray, Phacelia hydrophylloides A. Gray, and Sidalcea glaucescens E. Greene were most signif- icantly associated with pioneer sites age 4—10 yr since harvest, whereas the shrubs Ceanothus cordu- latus Kellogg and Ribes roezlii Regel and the herb Viola pinetorum E. Greene were most significantly associated with later seral sites 16-32 yr since harvest. CCA ordination diagrams arranged the 113 sites along a continuum, rather than breaking them up into units, and this approach also revealed a strong relationship between site vegetation and time since harvest. The general path of early succession did not show dramatic floristic shifts nor was there any significant change in species richness over time. The first 32 yr of secondary succession probably represents only one-seventh the time necessary to attain old-growth status. Abies magnifica Andr. Murray (California red fir) is a dominant of upper montane conifer forests of northern California at elevations between 2000 and 3000 m (Barbour and Minnich 1999; Barbour et al. 1991; Laacke 1990; Potter 1994; Rundel et al. 1988). The range of A. magnifica extends into southern Oregon and western Nevada but is largely within the boundaries of California, and its range accounts for 2% of the state’s area. The autecology of A. magnifica and the dynamics of forests it dom- inates are poorly known, although A. magnifica has recently become a subject of research (e.g., Chap- pell and Agee 1996; Taylor 1993; Taylor and Hal- pern 1991). The pattern and rate of succession from logging are unreported in the general literature, even though A. magnifica has been harvested this century at in- creasing intensities over time. A study by the Sierra Nevada Ecosystem Project Science Team (SNEP 1996) estimated that 65% of historic old-growth red fir forest acreage has been harvested by a combi- nation of clear-, seed tree-, and select-cut methods. Some clearcuts have been slow to be invaded by A. magnifica, |-3 decades passing before young red firs, 1-2 m tall, are present (Donald Potter personal communication; Barbour et al. 1998). The objective of this study is to describe the course of floristic Succession during those first three decades follow- ing clearcut harvest. ‘Centre for Population Biology, Imperial College at Silwood Park, Ascot, Berkshire SL5 7PY UK. * Present address: Departmento Ecologia, Facultad de Sciencias, Universidad Alcala de Henares, 28872 Madrid, Spain. METHODS Over the course of three field seasons, we inter- viewed virtually every district silviculturalist in five national forests (hereafter, NF) in the central Sierra Nevada, regarding the location of clearcuts in old- growth red fir forests that met the following crite- ria: 1) the clearcut was 1-35 yr in age and >2 ha in area; 2) the stand harvested was >80% A. mag- nifica in tree composition; 3) the clearcut was lo- cated on the west side of the range, to keep climate uniform; 4) the stand harvested had not been pre- viously disturbed by crown fire, blowdown, or se- lective logging; and 5) there was no record of post- harvest shrub suppression or tree thinning. (Some clearcuts had records of planting fir or pine seed- lings after harvest, but mortality was consistently 100% (Barbour et al. 1998) allowing us to safely accept this form of post-harvest treatment as having no effect.) Once acceptable clearcuts were identi- fied on maps and aerial photographs, they were vis- ited and sampled. A total of 113 clearcuts are in- cluded in our analysis. Each clearcut was described as to slope steep- ness, slope face, topographic profile (convex, con- cave, uniformly level, or mixed), elevation, lati- tude, soil depth, longitude, and national forest; and then a photograph was taken of the site. Data on clearcut area, shape, year of harvest, and subse- quent post-harvest treatments were obtained from stand record cards (or whatever fragmentary re- cords did exist, including the memory of long-term employees; typically clearcuts older than 35 yr had too few records for us to be sure that they satisfied our selection criteria). Unless the clearcut was a narrow strip, it was 132 MADRONO [Vol. 45 TABLE 1. MEANS AND RANGES OF ENVIRONMENTAL VARIABLES AMONG THE 113 SITES. Variable Mean Minimum Maximum Slope (°) 14.7 3.0 30.0 Aspect (°) 227 11.0 360.0 Soil depth (cm) 28.9 10.0 58.0 Elevation (m) 2273 1969 2656 Latitude (°(N) 38.6 37.0 40.1 Area (ha) 6.7 2.0 27.9 Perimeter (m) 1185 580 3610 Edge (m/m7) 0.023 0.006 0.066 Time (yr) 19.9 4.0 32.0 Herb cover (%) 5.8 0.0 525 Shrub cover (%) 3.8 0.0 9.4 Non-red-fir tree cover (%) 3.9 0.0 54.1 Red fir tree cover (%) 5:5 0.0 All Red fir density (per 25 m7’) 39 0.0 63.0 Red fir average age (yr) 12.4 0.0 41.1 divided into four equal subareas and each was sam- pled with one 5 X 5 m plot, located randomly with- in the subarea in such a way that it fell halfway between the margin of the clearcut and the center of the subarea. If the clearcut was a strip, the plots fell along the center line of the strip. In each plot, we recorded cover of all herb, shrub, and tree sapling species to the nearest tenth of a percent. Soil depth to rock was estimated at each plot corner by pounding a length of rebar into the ground until it reached an obstruction. In one 500 m°? area, subjectively located in the center of the clearcut, all stumps >30 cm at the cut surface were tallied by species as a check on our criteria that A. magnifica was dominant, and also as a means of identifying the stand according to a clas- sification of upper montane Sierran forests by Pot- ter (1994). Additional information, to aid this clas- sification process, was provided from an adjacent, uncut stand that we presumed was similar to the stand harvested. We sampled a 20 X 20 m area of that stand, noting the composition and relative spe- cies importance in overstory and understory strata. We used Potter’s classification because it is much more detailed than the series descriptions of Sawyer and Keeler-Wolf (1995) and community/habitat de- scriptions of Robert Holland (unpublished, but see Barbour and Major, 1988). Voucher specimens were collected for all herbaceous species at each site. Nomenclature follows the Jepson Manual (Hickman 1993) and vouchers are housed in the Tucker Herbarium at the University of California, Davis. Our analysis of successional dynamics followed two approaches. We used both the classicial ap- proach of searching for discrete seral stages and the continuum approach of relating each site individu- alistically to gradients of environmental factors. For the first approach, we utilized TWINSPAN, a com- monly used divisive computer program which sorts species into maximally different, but minimum in number, groups (Hill 1979; van Groenewoud 1992). These groups were then statistically correlated with site factors by ANOVA (analysis of variance) ta- bles. For the second approach, we utilized an or- dination technique called canonical correspondence analysis (CCA), first described by Ter Braak (1987). RESULTS AND DISCUSSION Overview of all sites. Our sample of 113 stands was nearly exhaustive, rather than random. We could have sampled more replicate clearcuts clus- tered within the same harvest unit, but we sam- pled all harvest units that fit our criteria. We as- sume that these sites are representative of mature, old-growth A. magnifica stands elsewhere in the range, stands which by chance have not yet been harvested. We have no evidence to support the possibility that only the most productive or un- usual sites were harvested; rather, it is more like- ly that the logged stands were simply closest to already existing roads, making their exploitation less costly. The distribution of our sites was neither uni- formly distributed across the range of red fir forest nor was it random, because the use of clearcuts as a harvest technique has not been uniformly ap- plied across all five forests, nor has the distribu- tion of harvests been constant over time. Of our 113 clearcuts, 46 were in Tahoe NE 42 in Stan- islaus, 16 in Sierra, 7 in Eldorado, and 2 in Plu- mas. Thus, the latitudinal extremes (Sierra NF in the south, Plumas NF in the north) were the least intensively sampled. The latitudinal span of sam- ples extended from 37 to 40°N (Table 1). Because of the NW-SE trend of the mountain range, lon- gitude (119—-121°W) is strongly linked with lati- tude and can’t be treated as an independent vari- able in our analysis. The span of clearcut age extended from 4 to 32 yr (Table 1). Younger clearcuts tended to fall to- ward the north, in Tahoe NE whereas older clear- 1998] FERNAU ET AL.: RED FIR FOREST SUCCESSION I33 TABLE 2. TWINSPAN Ecor_oristic UNITS. N = number of the 113 sites which fell in each unit. Assignment of the percentage of N sites to early, mid, and late is by age of the clearcut: 4—10 yr old is early-successional, 11—25 is mid- successional, and 26—32 is late-successional. Percentage of N sites by clearcut age Unit N Early Mid Late 1 12 17 33 50 2 10 0 40 60 3 35 11 43 46 4 26 15 62 23 5 16 60 28 12 6 14 100 0) 0) cuts tended to fall toward the south, in Stanislaus and Sierra NFs. The correlation of clearcut age with latitude was highly significant (r —0.38, P <0.0001). More than half the Tahoe clearcuts were <11 yr old, whereas two-thirds of the Stainslaus clearcuts were >24 yr old. Clearcuts in the other three forests were more evenly distributed among the range of ages. Clearcut areas varied more than an order of mag- nitude, from 2 to 28 ha, averaging nearly 7 ha (Ta- ble 1). Clearcut shapes ranged from long, narrow strips to roughly rectangular, square, and circular. ‘““Edge’’—the ratio of perimeter to area—varied by more than an order of magnitude, from 0.006 to 0.066 m/m?’. Clearcut elevations were between 1768 and 2591 m, increasing to the south somewhat fast- er than the Sierra-wide ecological displacement of 172 m per degree latitude (Parker 1994). Slopes were generally modest, 3—30°, and their topograph- ic profiles were diverse (Table 1). All aspects were represented. Soil depth varied nearly six-fold, from 10 to 58 cm. We did not determine soil series or geologic substrates for use as variables in this study, as they were overwhelmingly Inceptisols on granitic parent material. Herb cover was 0—52%, shrub cover was 00-80%, and tree cover (mostly A. magnifica) was 0—-54% (Table 1). A. magnifica sapling density ranged from 0.5 to 387 per 25 m* (equivalent to 200—155,000 ha~', averaging 82,000 ha~'). Average sapling age on all 113 sites was 13 yr, younger than the average site age of 20 yr (Table 1), which is another indi- cation that there is a many-year lag between the time of harvest and the time of red fir invasion. The average age of other tree species in the sites was insignificantly different from that of A. magnifica: 12 yr for Pinus contorta Loudon ssp. murrayana (Grev. & Balf.) Critchf., 11 yr for P. jeffreyi Grev. & Balf., and 9 yr for Abies concolor (Gordan & Glend.) Lindley. Stumps of harvested trees (all taxa combined) averaged 160 ha”! and the average age of the largest A. magnifica stumps was 350 yr (the oldest was 425 yr). Uncut forests adjacent to the clearcuts fell into Six associations described by Potter (1994): 70 sites fit his red fir association, 29 fit his red fir-white fir association, 9 fit his red fir-western white pine as- sociation, 2 fit his mountain hemlock association, and 1 each fit his red fir-western white pine-lodge- pole pine and his white fir-sugar pine-red fir asso- ciations. Given the overwhelming preponderance of sites within one association, we did not attempt to analyze succession by separating out original as- sociations harvested. TWINSPAN analysis. A total of 208 vascular species were present in the 113 clearcuts. In order to statistically analyze patterns of change over time, a given species had to occur in at least five sites; only 87 of those taxa fulfilled this require- ment. Included among the 87 were five tree spe- cies (A. magnifica, A. concolor, Pinus monticola, P. jeffreyi, P. contorta ssp. murrayana) and ten shrub species (Arctostaphylos nevadensis A. Gray, A. patula E. Greene, Ceanothus cordulatus Kel- logg, C. velutinus Hook., Prunus emarginata (Hook.) Walp., Ribes cereum Douglas, R. roezlii Regel, R. viscosissimum Pursh, Sambucus race- mosa L., and Symphoricarpos mollis Nutt.). Shrub Species contributing the highest average cover were Ribes roezlii, R. viscosissimum, and Ceano- thus cordulatus, in that order. The remaining 72 herb taxa were almost evenly split between mon- ocots and dicots. Those contributing the highest average cover were Phacelia hydrophylloides A. Gray, Elymus elymoides (Raf.) Swezey, Gayoph- ytum diffusum Torrey & A. Gray, and Sidalcea glaucescens, in that order. TWINSPAN analysis of the 113 sites < 87 taxa resulted in six ecofloristic units, which we num- bered 1—6 (Table 2). As a first approach towards understanding the units, we assigned their member sites to several discrete time classes. We subjec- tively created three equal groups of sites by divid- ing the 113 sites into three age spans: 4-10 yr (35 sites), 11-25 yr (35 sites), and 26—32 yr (43 sites). It is apparent, from Table 2, that the six units do represent a time gradient, with the sequence from youngest to oldest: 6 to 5 to 4 to 1 to 3 and 2. Unit 6, for example, has 100% of its 14 sites in the youngest age category of 4—10 yr, whereas unit 2 has none of its 10 sites in that youngest category and 60% of its sites are in the oldest category of 26-32 yr. 134 In an attempt to refine our interpretation of the units, we performed ANOVA’s of units against abiotic variables. Only four variables showed a Statistically significant pattern (P = <0O.05): slope, elevation, latitude, and time. Two biotic variables (shrub cover and tree sapling cover) also showed statistically significant differences among the six units. Three other abiotic variables just missed the significant cutoff (P = 0.06): soil depth, clearcut area, and edge. Thus, the six TWINSPAN units do reflect successional time, but they also reflect non-seral microenvironmen- tal differences. A cursory examination of the raw unit X< taxa table (not included here) showed the following taxa to be most characteristic and abundant in the “‘old- er’ units 1, 2, and 3: the shrubs Ceanothus cor- dulatus, Ribes cereum, and R. roezlii; the forbs Hackelia nervosa (Kellogg) I. M. Johnston, Viola bakeri E. Greene, and V. purpurea Kellogg, and the grasses Achnatherum occidentalis (Thurber) Back- worth, and Poa bolanderi Vasey. No species, how- ever, appeared to be characteristic of the younger units 4, 5, and 6. A more powerful search for relationships be- tween taxa and units is to subject each of the 87 common taxa to ANOVA, in essence asking the question: are they distributed among the six units randomly, by chance, or do any of them exhibit a non-random pattern? Nearly half exhibited a non- random distribution (P = <0.05): A. magnifica, A. concolor, Achnatherum nelsonii (Scribner) Back- worth ssp. dorci (Backworth & J. Maze) Back- worth, Antennaria rosea E. Greene, Arabis sparsi- flora Torrey & A. Gray, Arctostaphylos nevadensis, A. patula, Bromus carinatus Hook & A. M., Carex fracta Mackenzie, Ceanothus cordulatus, Chamae- saracha nana (A. Gray) A. Gray, Cirsium ander- sonit (A. Gray) Jepson, Collinsia torreyi A. Gray, Cryptanthus glomeriflora E. Greene, Elymus glau- cus Buckly, Eriogonum spergulinum A. Gray, Hackelia nervosa, Lotus purshianus (Benth) Cle- ments & E. G. Clements var. purshianus, Lupinus andersonit S. Watson, Monardella odoratissima, Pedicularis semibarbata A. Gray, Phacelia hydro- phylloides, Pinus contorta ssp. murrayana, P. jef- freyi, P. monticola, Poa bolanderi, Polygonum douglasii E. Greene, Potentilla glandulosa Lindley, Ribes cereum, R. roezlii, Sidalcea glaucescens, Stephanomeria lactucina A. Gray, Symphoricarpos mollis Nutt., Viola bakeri E. Greene, V. pinetorum E. Greene, V. praemorsa Douglas, and V. purpurea Kellogg. None of these, however, exhibited a sta- tistically significant least-significant difference (Tu- key’s test, P = 0.05) between any two of the six units. Ordination analysis. The ordination diagram (Fig. 1) shows the 113 sites distributed along three axes. Each site is represented as an open circle. The ordination program adds lines radiating from the MADRONO [Vol. 45 center of the clusters which statistically relate to site factors. The direction of the line shows the di- rection of the relationship, relative to the axes, and the length of the line shows the statistical strength of the relationship (the longer the line, the stronger the relationship). Axes | and 2 reveal that latitude and time were the abiotic factors which had the most effect on site similarity (circles close to each other) or difference (circles far apart). Axes 1 and 2 also show that the most important biotic factors for site similarity or difference were herb, shrub, and tree cover. The same biotic and abiotic variables were highlighted by axes | and 3 and 2 and 3 (Fig. 1). The outliers in the ordination diagram are all unique in that these sites had very rapid, massive invasion by A. magnifica seedlings, and in that they had a concave topography. Taxonomic and functional species change over time. When we performed ANOVA’s on the 87 common taxa X three age classes of sites, 15 ex- hibited significantly non-random distributions (Ta- ble 3). Those which peaked in cover in early suc- cession (yr 4—10) were: Cryptantha glomeriflora, Gayophytum diffusum Torrey & A. Gray, Phacelia hydrophylloides, Sidalcea glaucescens, and Sym- Phoricarpos rotundifolius A. Gray. Those which peaked in yr 11—25 were: Arabis platysperma A. Gray, Elymus elymoides (Rat.) Swezey, Hieracium albiflorum Hook. Pinus monticola, and Ribes vis- cosissimum Pursh. Those which peaked in yr 26— 32 were: Ceanothus cordulatus, Pinus jeffreyi, Ri- bes roezlii, and Viola pinetorum. Even though these species trends are statistically significant, they in- volve relatively small amounts of cover, and the ecological significance or impact of the changes on community function are uncertain. The general path of succession during the first 32 yr after harvest did not reveal dramatic changes, nor did it reveal trends widely reported elsewhere. For example, when species richness was regressed against site age, there was no significant linear or curvilinear relationship. Our data showed a flat relationship, with an average of 23 taxa per site, plus considerable ‘‘noise’’ (maximum richness was 49 taxa for a site 27 yr after harvest and minimum was 3 for a site 8 yr after harvest). In his recent summary of succession, Robert Peet (1992) concluded that species diversity and species richness both continually rise during suc- cessional time, peaking either at or just prior to the climax/equilibrium phase. Such functional attributes as growth forms, “strategies” (e.g., C, S, R of Grime 1979), and metabolic specialists (N-fixers, helophytes vs sciophytes, etc.; see Barbour et al. 1998) also typ- ically change during succession, but our clearcut sites only showed an increase in shrub cover and total plant cover. There were no trends over time, for example, of annuals to perennials, N-fixers to non-fixers, grass- es to forbs, nor in pollination or seed dispersal syn- dromes. 1998] AXIS 2 AXIS 1 outlier 45 outlier 29 AXIS 1 Fic. 1. AXIS 3 FERNAU ET AL.: RED FIR FOREST SUCCESSION 135 outlier site 29 outlier 58 outlier 45 ° outlier 29 , AXIS 2 Ordination of 113 sites by canonical correspondence analysis along three axes. The lines show correspondence with the site variables latitude (LAT), herb cover (HERBC), tree cover (TREEC), shrub cover (BC), and age of the clearcut (TIME). The longer the line, the stronger the correspondence. Possibly the magnitude of change is modest be- cause tree canopy (sapling canopy) cover is modest, even 32 yr after harvest (refer to Table 1, which shows that maximum tree cover was only about 50%). Our companion publication on A. magnifica regeneration for these sites (Barbour et al. 1998) showed that A. magnifica percent cover, when re- gressed against time, was = (0.21) (time, in yr) + 1.42. The positive slope was statistically significant (P = 0.006) even though the relationship accounted for only 7% of the variance. If the entire course of succession were similarly linear, 270 yr would be required to attain mature tree cover of 60% (aver- age for old-growth red fir forests as reported by Barbour and Woodward, 1985). No doubt, there is an eventual curvilinear relationship between cover and time, where it rises more steeply, but there are no data yet to show when this might occur, or how steeply the cover might increase. Based on the general age of trees in old-growth stands as reported in the literature, and stand data from our own unpublished data, secondary succes- sion for red fir forests probably requires more than 200 yr. The portion of secondary succession inves- tigated in this paper, then, is a rather small fraction, about one-seventh, of seral time. 136 MADRONO [Vol. 45 TABLE 3. SPECIES WHICH EXHIBITED A NON-RANDOM DISTRIBUTION (ANOVA, P < 0.05) AMONG AGE CLASSES OF STANDS. Early-successional = 4—10 yr old sites, mid-successional = 11—25, late-successional = 25—32. Data are percent cover. Figures in the same row which share the same superscript do not differ at P < 0.05. Species Early Arabis platysperma 0.02 Ceanothus cordulatus 0.8? Cryptantha glomieriflora 0.048 Elymus elymoides 0.08 Gayophytum diffusum 0.952 Hieracium albiflorum 0,02” Pedicularis semibarbata 0.018 Phacelia hydrophylloides aN Pinus jeffreyi 0.042 Pinus monticola 0.03% Ribes roeczlii 1.7* Ribes viscosissimum 0.28 Sidalcea glaucescens O35? Symphoricarpos rotundifolius 0.58 Viola pinetorum 0.014 ACKNOWLEDGMENTS Donald Potter, Zone Ecologist for the USDA Forest Service, provided valuable advice about the biology and ecology of Abies magnifica as well as the identity of help- ful agency silviculturalists and timber sale specialists. Fred Hrusa identified many of our voucher specimens. Fi- nancial support was provided by the USDA Competitive Grants Program, award No. 92-87101-7420. LITERATURE CITED BarRBour, M. G., J. A. BURK, AND W. D. Pitts. 1998. Terrestrial plant ecology, 3rd ed. Benjamin-Cum- mings, Palo Alto, CA. BARBOUR, M. G., R. FE FERNAU, J. M. REY BENAYAS, N. JURJAVCIC, AND E. B. Royce. 1998. Tree regeneration following clearcut logging in red fir forests of the Sierra Nevada, California. Forest Ecology and Man- agement 104:101-111. BARBOUR, M. G. AND J. MAJor (eds.). 1988. Terrestrial vegetation of California, 2nd ed. California Native Plant Society, Sacramento. BARBOUR, M. G. AND R. A. MINNICH. 1999. Californian upland forests and woodlands, in M. G. Barbour and W. D. Billings (eds.), North American terrestrial veg- etation, 2nd ed., Cambridge University Press, New York. BARBOUR, M. G. AND R. A. WOODWARD. 1985. The Shasta red fir forests of California. Canadian Journal of For- est Research 15:570-—576. CHAPPELL, C. G. AND J. A. AGEE. 1996. Fire severity and tree seedling establishment in Abies magnifica forests, southern Cascades, Oregon. Ecological Applications 6:628—640. GRIME, P. J. 1979. Plant strategies and vegetation process- es. Wiley, New York, NY. Mid Late 0.02° 0.01% 8.92 133° 0.0° 0.022» O31 0272 0.22° 0.28> 0.062 0.02° 0.0° 0.01% 0.1° 0.2° 1.6° 1.9° 13° 0.1° O72 45° 1.9> O72 0.202 0.05° 0.0 0.02 0.03? 0.5> HICKMAN, J. C. (ed.). 1993. The Jepson manual: higher plants of California. University of California Press, Berkeley. Hitt, M. O. 1979. TWINSPAN: a FORTRAN program for arranging multivariate data in an ordered two-way table by classification of individuals and attributes. Cornell University, Ithaca, NY. LAACKE, R. J. 1990. Abies magnifica A. Murr., California red fir. Pp. 71-79 in R. M. Burns and B. H. Honkala (eds.), Silvics of North America, Volume 1, USDA Forest Service, Washington, D.C. PEET, R. K. 1992. Community structure and ecosystem function. Pp. 103-151 in D. C. Glenn-Lewin, R. K. Peet, and T. T. Veblen (eds.), Plant succession, Chap- man and Hall, New York, NY. SAWYER, J. O. AND T. KEELER-WOLF. 1995. A manual of California vegetation. California Native Plant Soci- ety, Sacramento. SNEP. 1996. Status of the Sierra Nevada, Volume 1. Cen- ters for Water and Wildland Resources, Report No. 36. University of California, Davis. TayLor, A. H. 1993. Fire history and structure of red fir (Abies magnifica) forests, Swain Mountain Experi- mental Forest, Cascade Range, northwestern Califor- nia. Canadian Journal of Forest Research 23:1672-— 1678. TAYLOR, A. H. AND C. B. HALPERN. 1991. The structure and dynamics of Abies magnifica forests in the south- ern Cascade Range, USA. Journal of Vegetation Sci- ence 2:189—200. TER BRAAK, C. J. E 1987. The analysis of vegetation- environment relationships by canonical correspon- dence analysis. Vegetatio 69:69-—77. VAN GROENEWOUD, H. 1992. The robustness of correspon- dence, detrended correspondence, and TWINSPAN analysis. Journal of Vegetation Science 3:239—246. MAabRONO, Vol. 45, No. 2, pp. 137-140, 1998 A NEW GILIA (POLEMONIACEAE) FROM LIMESTONE OUTCROPS IN THE SOUTHERN SIERRA NEVADA OF CALIFORNIA JAMES R. SHEVOCK USDI National Park Service, Pacific West Region 600 Harrison Street, Suite 600, San Francisco, CA 94107 ALVA G. DAY Department of Botany, California Academy of Sciences, San Francisco, CA 94118 ABSTRACT Gilia yorkii, a new species discovered in the southern Sierra Nevada, is restricted to limestone outcrops. From its morphology G. yorkii is identified as a member of Gilia sect. Saltugilia, and within the section it appears most like G. scopulorum M. E. Jones, a desert species. The two species differ in several characters, including different corolla proportions and some contrasting trichome details. From recent explorations on xeric limestones (marbles) in the southern Sierra Nevada a new spe- cies of Gilia has been discovered. Gilia yorkii is scattered primarily over the southern exposures of a large north-south trending limestone ridge. At its highest point, this sheer-walled limestone formation rises more than 1000 m above the South Fork of Kings River. It occurs just east of Horseshoe Bend along both sides of the river canyon. This interest- ing new species was discovered by Dana York on 31 July 1995 in Monarch Wilderness, while ex- ploring the steep and rugged limestone terrain. Gilia yorkii Shevock & A. G. Day, sp. nov. (Fig. 1)—TYPE: USA, CA, Fresno Co, 86 km E of Fresno, Sequoia National Forest, Monarch Wil- derness, 1 km S of Boyden Cave on S side of Kings River Canyon, 36°48’20"N, 118°48'40’"W; T13S, R29E, S10, NW % of SW %, 1290 m. 27 June 1996, York & Shevock 949 (holotype CAS; isotypes JEPS, FSC). Herba annua inter species sect. Saltugilia V. Grant & A. D. Grant trichomatibus a foliis multi- septatis et translucidis, calyce glanduloso, corollae lobulis lavandulis, capsula late ovoidea, inclusa, et seminibus in loculis 2—3, ad G. scopulorum M. E. Jones accedens sed differt foliis basalibus paucis, foliis paucilobis, vel interdum foliis integris, tri- chomatibus curvus, eglandulosis (nec patentibus nec glandulosis), trichomatibus pedicello et calyce glandulosis minutus incoloribus (nec magnus nec nigris), corolla minor, tubo incluso (nec exserto nec exserto longissimo). Annual herb 10—25 cm high, with one or two Somewhat spreading branches arising near base of main stem, sometimes vigorous plants becoming diffusely branched and to 4 dm high. Stems pubes- cent below middle or near base, with multiseptate, translucent, eglandular trichomes; upper stems densely glandular-puberulent. Basal and lower cau- line leaves 2-8, with basal mostly drying and fall- ing early, thus only vigorous, immature plants with unexpanded internodes show a definite basal ro- sette. Lower leaves pubescent on ventral side, densely so in their axils, trichomes very fine, mul- tiseptate, eglandular, broad at base, fine-tapered near tip, shining, translucent, often + bent at septa (Fig. 1D). Basal and lower cauline leaves 1.5—2.5 cm long, 1-pinnate or entire, oblanceolate, rachis linear below, broadened between lobes, lobes 5—7 in ascending, subopposite pairs, 3—5(7) mm long, elliptic, acute, entire or with 1-2 teeth near apex. Upper cauline leaves glandular-puberulent, elliptic, generally with two narrow, spreading, basal lobes, uppermost leaves reduced, entire. Upper stems, pedicels and calyces densely glandular-puberulent, trichomes stipitate, glands minute, (0.05 mm diam.), colorless, stipe narrower than gland, ca. twice as long as diameter of gland. Inflorescence loose, cymose, pedicels 3—10 mm in flower, in sub- equal to unequal pairs, or in 3’s, elongating and divergent spreading in fruit to 5-33 mm. Calyx 3— 3.6 mm long in flower, accrescent, ribs, lobes and membrane glandular-puberulent or rarely in earliest flowers calyx trichomes longer, non-glandular, as on lower leaves, ribs 0.2 mm wide, green, or in age red-streaked, connected by sinus membrane in low- er half, membrane at maturity splitting between ribs to near base. Corolla 7—8.5 mm long, tube and throat together +4 mm long, included in calyx, or upper throat exserted, tube and lower throat white, sometimes lavender-tinged, upper throat and lobes lavender to white; lobes spreading, 3-5 mm long, 2.5—4 mm wide, elliptical to broadly elliptical, apex rounded. Stamens inserted equally 0.5 mm below sinuses, filaments 0.2—1 mm long, anthers 0.9—1.5 mm long, maturing in or slightly above orifice, pol- len blue. Style exserted beyond anthers, stigmas 1— 1.5 mm long, spreading, tips curling downward to- ward, or touching anthers. Capsule broadly ovoid, 3—4.5 mm long, 1.4—1.7X as long as wide, equal to, or slightly exceeded by calyx, dehiscing by 138 MADRONO [Vol. 45 D Fic. 1. Habit of Gilia yorkii Shevock & A. G. Day, with details shown in comparison with G. scopulorum M. E. Jones. A—G. Gilia yorkii, from holotype collection: A—B. Plant habit; A. young plant in early flower, with cotyledons present; B. Mature plant in flower and fruit; C. Basal and lower cauline leaves; D. leaf trichomes; E. *Flower with detail of calyx trichomes; F Dissected corolla with stamens and style; G. Calyx with mature, included capsule. H—-K. G. scopulorum: H. Variation in basal leaves, (1 to r) Darrow s.n. (CAS 336894), Prigge 994 (CAS 615000), Raven 11790 (CAS 517748); I. leaf trichomes; J. *Flowers from three different collections, showing variation in corolla, and detail of calyx trichomes, (1 to r) Pinzl 2611 (CAS 637910), Raven 11790 (CAS 517748), Prigge et al. 769 (CAS 606023); K. Calyx with mature, included capsule, Prigge 994 (CAS 615000). Note: Corolla lobes are shown erect for comparisons of size and form. In real life, as in other Gilias, the lobes are spreading. This is shown in views of two flowers in the habit drawing of G. yorkii (B). 1998] splitting between valves from apex to near base. Seeds 0.8-1.5 mm long, brown, verrucate, + ob- long, 2—3 per locule, mucilaginous when wet. Pol- len grains + spheroidal, zonocolporate, colpi 7-8, exine striate. Paratypes. USA, CA, Fresno Co., Monarch Wil- derness, vicinity of Boyden Cave, 31 Jul 1995, York & Shevock 107 [CAS]; 31 Jul 1995, York & Shevock 112 [CAS, FSC, JEPS]; 13 Oct 1995, York 207 [CAS]. DISTRIBUTION, HABITAT AND PHENOLOGY Gilia yorkii is known only from the southern Si- erra Nevada in Fresno Co., California, from lime- stone outcrops in the vicinity of Boyden Cave in the Kings River Canyon. Plants grow in fissures, ledges, and terraces in sandy or gravelly soils de- veloped from weathered limestone, at elevations from 1290 to 1830 m. Within its habitat G. yorkii can easily be overlooked because the pale lavender flowers blend in with the surrounding limestone. Plant size may vary widely, depending on the tim- ing and amount of late spring to early summer rain. In a year with early as well as late rains, plants may become tall and spreading, diffusely branched and abundantly flowered. In a more normal year, however, plants are apt to be smaller at maturity and to have fewer flowers. The exposed limestone habitat occupied by G. yorkii is very arid, and during summer plants are subjected to daily ambient temperatures in excess of 34°C. The time of anthesis, depending upon the particular seasonal conditions, appears to range from May to July. While the number of G. yorkii plants has not been estimated, the population could number in the thousands under optimum conditions. Only a por- tion of the potential habitat has been surveyed be- cause of the rugged nature of the area. This habitat is dominated by many petrophilous taxa along with species generally located within chaparral and foothill woodland plant communities. Prominent among these are woody associates in- cluding three Cercocarpus species: C. intricatus S. Watson, C. betuloides Torrey & A. Gray, and C. ledifolius Nutt. var. intermontanus N. Holmgren; and also Garrya flavescens S. Watson, Pinus mon- ophylla Torrey & Frémont, Rhamnus tomentella Benth., Umbellularia californica (Hook. & Arn.) Nutt., and Yucca whipplei Torrey. Other common associates, including annuals, ferns, and other perennial herbs are the following: Argyrochosma jonesii (Maxon) M. D. Windham, Asclepias fascicularis Decne., Astragalus congdon- it S. Watson, Avena sativa L., Bromus madritensis L. ssp. rubens (L.) Husnot, Cheilanthes cooperae D. Eaton, Cirsium occidentale (Nutt.) Jepson var. californicum (A. Gray) Keil & C. Turner, Clarkia rhomboidea Douglas, Eriogonum nudum Benth. (sensu lato), Erysimum capitatum (Douglas) E. SHEVOCK AND DAY: GILIA YORKII 139 Greene, Heterotheca monarchensis York, Shevock & Semple, Heuchera rubescens Torrey var. alpicola Jepson, Mentzelia laevicaulis (Hook.) Torrey & A. Gray, Mimulus floribundus Lindley, Nemacladus interior (Munz) G. Robb., Petrophyton caespitosum (Nutt.) Rydb. ssp. acuminatum (Rydb.) Munz, Se- laginella asprella Maxon, S. hansenii Hieron., and Streptanthus fenestratus (E. Greene) J. Howell. The general aspect of the vegetation is markedly different at the borders of the limestone habitat, where reddish metamorphic rock outcrops support canyon live oak woodland with scattered Torreya californica Torrey and chaparral elements, but not Gilia yorkii. RELATIONSHIPS From its morphology Gilia yorkii fits readily into Gilia sect. Saltugilia, and within the section it ap- pears rather like G. scopulorum M. E. Jones, shar- ing characters concerning pubescence, seed-num- ber, and capsule form. They have a similar leaf type, as seen in the venation pattern, with the lobes, and their veins ascending. The lobes are toothed on both sides (Fig. 1C, H), but the leaves of G. yorkii are smaller, more delicate, and with lobes sparingly toothed. The two species also show some similarity in habitat, both commonly occurring in limestone areas. However, Gilia scopulorum has also been found on volcanic substrates. In the Jepson Manual key to Gilia (Day 1993) G. yorkii would fall near the G. scopulorum posi- tion. A segment of that key, shown below, has some added details, and a new dichotomy to include G. yorkii. The important differences between G. yorkii and G. scopulorum are illustrated in Figure 1. Gilia scopulorum leaf lobes tend to be large, and many-toothed, but sometimes are + reduced, as in G. yorkii, entire or with 1—2 teeth. A stout, erect, leading stem above a well-developed rosette of basal leaves is the usual condition in G. scopulorum. However, this is not found in mature plants of G. yorkii, which lack a definite basal rosette of leaves, and have a shorter, generally several-branched central stem. Other important differences are the shorter, in- cluded corolla tube of G. yorkii (Fig. 1E), and dif- ferences in size and gland-type of trichomes borne on the calyx and lower leaves and stems of the two species (Fig. 1E, J). The importance of trichomes in differentiating species and larger taxonomic units of Polemoni- aceae has been well demonstrated, as in Gilia sect. Arachnion, named for a characteristic trichome type, making up a fine arachnoid pubescence of all the member species (Grant and Grant 1956), and which is not found elsewhere in the genus. Within sect. Saltugilia all species have multisep- tate trichomes on the lower leaves, but as the key indicates, they are not all alike. Usually the multi- septate trichomes are coarsely translucent, often + glandular, and straight or variously curving; but in 140 G. stellata A. A. Heller they are opaque-white, eglandular and markedly geniculate. In G. yorkii the multiseptate trichomes (Fig. 1D) have yet another variation: although translucent, they are very fine, eglandular, and often bent at the septa, or curving, but not markedly geniculate as in G. stellata. The densely glandular-puberulent calyx of G. yorkii is unique in the section. The calyx in other species is either glabrous (G. splendens H. Mason & A. D. Grant group) or coarsely glandular-dotted (G. scopulorum and G. stellata). The scattered, large black glands on the calyx of G. scopulorum (Fig. 1J) contrast with the closely-spaced, minute, colorless calyx glands in G. yorkii (Fig. 1E). Corolla proportions are variable in G. scopulo- rum (Fig. 1J). Usually the corolla tube is long-ex- serted; but a variant with a short, slightly-exserted corolla tube was collected several times in Mohave Co., Arizona and in south-western Nevada (Fig. 1J, flower no. 1). This approaches the form of the G. yorkii corolla, but the tube is exserted. In G. yorkii the tube, as well as the throat, or part of it, are included in the calyx (Fig. 1E). In all significant characters the above variant of G. scopulorum dif- fers from G. yorkii. KEY TO GILA SECT. SALTUGILIA, (IN PART) (Modified from The Jepson Manual, Key to Gilia (Day 1993)) 1. Calyx glabrous; capsule narrowly ovoid, exceeding ca- lyx; capsule valves 2—3 X longer than wide; seeds 7=29-Pet 1OCUle: a.c.cqsbo8e 2.5 eo e.g a ae 28 G. splendens, G. australis and G. caruifolia. (For key to this species group see Day, 1993: 830.) 1’ Calyx glandular-dotted; capsule broadly ovoid, includ- ed in calyx, capsule valves <2 longer than wide; seeds 2—6 per locule. 5. Trichomes on lower leaves opaque-white, eglandular, geniculate; corolla throat with purple spot below each lobe; seeds 3—6 per locule ...G. stellata 5’ Trichomes on lower leaves translucent, straight or + curved, glandular or eglandular; corolla throat not spotted; seeds 2—3 per locule. 6. Corolla 7—-8.5 mm long, tube and throat (or part of throat) included in calyx; calyx and pedicels densely glandular-puberulent, glands minute, col- orless; trichomes on lower leaves and stems few, scattered, denser in leaf axils, very fine, eglan- dular, variously curving. Rare endemic. Sierra Nevada, CA., Kings River Canyon .. G. yorkii 6’ Corolla 9-17 mm long, tube exserted, generally 2X calyx or longer; calyx and pedicels glandular-dot- ted, glands large, black; trichomes on lower leaves and stems dense, coarse, gland-tipped, + straight, spreading. Desert mountains E of Sierra Nevada, CA to AZ, UT G. scopulorum DISCUSSION Gilia yorkii and G. scopulorum are allopatric, be- ing geographically isolated. Gilia scopulorum oc- curs in desert mountains of the Great Basin, Mo- MADRONO [Vol. 45 jave, and Sonoran deserts from Utah and Arizona to California, but it does not extend to the Sierra Nevada. Gilia yorkii occurs only in the southern Sierra Nevada, west of the summit divide. How- ever, this particular Sierran habitat is arid and des- ert-like, which is also indicated by the dominant species, and is quite unlike surrounding areas in the Kings River Canyon. The two species show relationship in the numer- ous morphological characters that they share, but their recognition as distinct species is justified by their differences with respect to a number of other characters: trichome types, corolla proportions, plant habit, etc., and because no intermediates be- tween them have been found. RARITY Gilia yorkii, previously unknown and uncollect- ed, is an extremely rare species due to its litho- phytic nature on limestones, a relatively uncommon substrate in the southern Sierra Nevada. While ad- ditional occurrences may well be discovered in steep canyons and rocky slopes, expansion of the distribution is unlikely to extend beyond the exist- ing limestone outcrops in the Kings River basin. This species, therefore, is expected to remain a rare and localized endemic worthy of conservation ef- forts. Fortunately this new species, limited to its limestone habitat, is located within the Monarch Wilderness in very steep and rugged terrain. For these reasons anthropogenic impacts are likely to be few. It is a pleasure to name this species for a col- league and friend who is actively exploring the Kings River Basin for the purpose of developing a floristic treatment for the region. His field work has already led to the discovery of three additional new taxa: Heterotheca monarchensis York, Shevock and Semple, and two as yet unnamed species of Carex and Eriogonum. ACKNOWLEDGMENTS We are grateful to Linda A. Vorobik for her excellent illustration comparing the species. We also thank Diana L. Lusk for her kind assistance in translating the diagnosis to Latin; and Dana York for sending photos and specimens of young G. yorkii plants that he grew from seed. We thank the reviewers, Robert W. Patterson and Dieter H. Wilken for their numerous helpful comments and sugges- tions. We appreciate the courtesies extended by the her- barium staff of the California Academy of Sciences (CAS/ DS), and of the University of California at Berkeley (UC). LITERATURE CITED Day, A. G. 1993. Gilia (Polemoniaceae). Pp. 828-839 in J. C. Hickman, (ed.), The Jepson Manual: higher plants of California. University of California Press, Berkeley. GRANT, A. D. AND V. GRANT. 1956. Genetic and taxonom- ic studies in Gilia VIII. The cobwebby gilias. Aliso 3(3):203—287. MADRONO, Vol. 45, No. 2, pp. 141-145, 1998 WATER POTENTIALS OF SALVIA APIANA, S. MELLIFERA (LAMIACEAB), AND THEIR HYBRIDS IN THE COASTAL SAGE SCRUB OF SOUTHERN CALIFORNIA DAvID S. GILL AND BARBARA J. HANLON Department of Biological Science, California State University, Fullerton, CA 92634 ABSTRACT Shrubs in the genus Salvia are often dominants in Coastal Sage Scrub communities throughout southern and central California. Salvia mellifera E. Greene (black sage) tends to be more abundant near the coast with a more northerly distribution, whereas S. apiana Jepson (white sage) tends to occur farther inland and ranges much farther south. In areas of local sympatry in southern California, S. apiana is more frequently observed on south-facing slopes. It has been suggested to be better able to withstand drought conditions than S. mellifera, because S. apiana occurs in what appear to be more xeric sites. In addition, in areas of sympatry, somewhat fertile, morphologically intermediate hybrids form. We tested the hy- pothesis that S. apiana is better able to withstand drought than S. mellifera, and that hybrids are physi- ologically intermediate, by measuring predawn and midday water potential every month over 15 months at a site in the Santa Ana Mountains. Salvia apiana had statistically higher water potentials than S. mellifera during summer drought, and the hybrid exhibited intermediate values. The ability of S. apiana to maintain water potentials of 3 to 4 MPa greater than S. mellifera supports the hypothesis that S. apiana can better avoid summer drought. This difference in water relations may account in part for the distri- butional patterns of these two species. Salvia apiana Jepson (White Sage) and S. mel- lifera E. Greene (Black Sage) are widespread spe- cies of summer deciduous shrubs that often domi- nate Coastal Sage Scrub (CSS) communities of cen- tral and southern California (Epling 1938; Harrison et al. 1971; Axelrod 1978; Kirkpatrick and Hutch- inson 1980; Westman 1981) [botanical nomencla- ture follows Hickman (1993)]. The center of the geographical range of S. apiana, however, is over 300 km south of that of S. mellifera; where their ranges overlap in southern California, S. apiana can occur about 15-30 km farther inland extending into the western margins of the Mojave and Colorado deserts (Fig. 1). Over much of its range, therefore, S. apiana experiences a hotter, drier climate than S. mellifera. In areas where they are locally sympatric, S. apiana exhibits greater relative abundance on more xeric sites (interior, south-facing, well- drained, etc.); whereas, S. mellifera is often rela- tively more abundant in somewhat less xeric sites (Epling 1947; Anderson and Anderson 1954; Kirk- patrick and Hutchinson 1980; Westman 1981; DeSimone and Burk 1992). These distributional differences suggest that S. apiana can somehow better withstand summer drought conditions than S. mellifera, but this idea has not been tested. Somewhat fertile hybrids will form where S. api- ana and S. mellifera are sympatric (Epling 1938; Grant and Grant 1964) and the hybrids exhibit Clearly intermediate reproductive and vegetative characteristics (Epling 1947; Anderson and Ander- son 1954; Meyn and Emboden 1987). The large leaves of S. apiana are covered with dense, white pubescence, the smaller, narrower leaves of S. mel- lifera are glabrous and green, and those of the hy- brid are intermediate in size, shape and degree of pubescence (Webb and Carlquist 1964). Survival of these hybrid populations primarily at disturbed sites has been used to infer that the hybrids are less com- petitive than either parent in the absence of distur- bance (Epling 1947; Anderson and Anderson 1954; Meyn and Emboden 1987), however, very little is known of the physiological ecology of hybrid plants growing in natural environments (Rieseberg 1995). Much of the original CSS habitat has been ad- versely impacted by agriculture and urbanization and is being reduced to increasingly smaller habitat islands (Westman 1981; O’Leary 1990). In spite of these environmental concerns, very little data exists on the physiological ecology of the summer decid- uous plants composing this rapidly disappearing community (Mooney 1977; Westman 1981). The purpose of this study is to investigate seasonal pat- terns of water potential of Salvia apiana, S. melli- fera and their hybrids in a CSS site where they co- occur. Studying these taxa in a single site where they all experience essentially the same rainfall and environmental conditions is a useful investigative approach, since differences in water potentials will indicate ecophysiological differences in drought re- sponse. The specific null hypothesis we tested was that water potentials of these three taxa would not differ (H,: S. apiana = hybrid = S. mellifera). If S. apiana can maintain significantly more positive water potentials while receiving the same amount of precipitation, this would support the hypothesis 142 MADRONO Fic. 1. The geographical ranges of Salvia apiana (solid line) and S. mellifera (dashed line) based on Figure | in Epling (1947), and data in Epling (1938) and Jepson (1939). The primary study site was located at Starr Ranch National Audubon Sanctuary (SR), while secondary sites were located in the San Joaquin Hills (SJ) and in San Gabriel Canyon (SG). that S. apiana is better able to avoid summer drought than S. mellifera. METHODS The study site was located at Starr Ranch Na- tional Audubon Sanctuary in southeastern Orange Co., CA (Fig. 1) which has large stands of rela- tively intact CSS (DeSimone and Burk 1992). The mediterranean climate is characterized by warm (21°C mean air temperature), dry summers and cool [Vol. 45 (12°), wet winters with an annual precipitation of 36 cm with most falling from November to April. The primary research site (ca. 1.5 ha) was located at 370 m elevation on a west-facing slope (290° azimuth, 19° slope). The well-drained soils through- out the site are from the Gabino gravelly clay loam series (DeSimone and Burk 1992). The vegetation at the time of the study was about 15 years old and was dominated by Artemisia californica Less. and codominated by both species of Salvia, several oth- er species of shrubs, and by a fairly high cover of native, perennial bunchgrass (Nassella sp.). The Salvia taxa were distributed throughout the site with no apparent differences in microhabitat selec- tion. In early February 1995, 10 sets (=blocks) of plants were selected, with each block containing one individual each of Salvia apiana, S. mellifera, and their hybrid all growing within 10 m of one another. Individuals within each block of plants were thus experiencing environmental conditions as similar as is possible in the field. Plants were as- signed to respective taxa based on inflorescence structure, and on the size, shape and extent of pu- bescence of leaves (see Epling 1947; Anderson and Anderson 1954; Webb and Carlquist 1964; Meyn and Emboden 1987). Based on these criteria the hybrids appeared to be F,’s or first generation back- crosses toward S. mellifera. A random sample of 5 of these sets was selected for water potential mea- surements. Predawn and midday water potentials (WP) were measured monthly on 3 shoots per plant using a Scholander-style pressure chamber (Ritchie and Hinckley 1975). Every 2-3 months we switched to the other 5 sets of plants to minimize possible effects of harvesting shoots on plant vigor. Rainfall data were obtained from a gauge at the Starr Ranch headquarters about 1 km away and 70 m lower in elevation. To confirm the generality of WP responses ob- tained in the primary site at Starr Ranch, WP was also measured in the summer of 1996 (23-25 June, 10 weeks since last spring rain) at two additional sites in southern California. The San Joaquin Hills site was located about 20 km SW near the Pacific Ocean and the San Gabriel Canyon site was about 60 km NNE in the interior foothills of the San Ga- briel Mountains. Mean midday WP of Salvia api- ana, S. mellifera, and their hybrids was measured on 3 shoots from five plants of each taxa except at the San Gabriel Canyon site where only three hy- brid individuals could be located. To test for differences among species (=taxon effects) and blocks (=microsite effects), data were analyzed using two-way analysis of variance (GLM procedure) on raw data since the data met the as- sumptions for homogeneity of variance (Levene’s test, Minitab 1995). Microsite effects were not sig- nificant so all subsequent analysis was based on one-way analysis of variance (one-way procedure) and when significant (P < 0.05) taxon-level differ- 1998] Predawn Water Potential (MPa) S. mellifera Weekly rainfall (cm) 1995 Fic. 2. GILL AND HANLON: WATER POTENTIAL OF SALVIA 143 1996 Rainfall and predawn water potential of Salvia apiana, S. mellifera, and their hybrid at the Starr Ranch site. Rainfall is expressed as weekly total (cm), and water potential as the mean (+1 pooled SD) of five plants of each taxa, from January 1995 through April 1996. When no error bars are observed the pooled SD is smaller than the plotted symbol. ences were detected, Fisher’s aposteriori test was used to compare means (Minitab 1995). Data are expressed as means (+1 pooled standard deviation; 1 SD,ooiea = Vmean square error from the ANO- VA). RESULTS From February through early June of 1995 there were no significant differences in predawn water potential (WP) among taxa, with WP of all taxa being greater than —0.5 MPa (Fig. 2). After the last rain of the season, WP of all taxa started to decline as water loss began to exceed uptake. During sum- mer drought conditions, S. mellifera experienced more negative WP than S. apiana and the hybrids were intermediate. On all 9 sampling dates from mid-July through early January 1996, WP differed significantly (P < 0.01) among taxa and the mean values always ranked in the order of S. apiana > hybrid > S. mellifera. Following an early autumnal rain of 0.46 cm on 2 November, WP of all taxa rapidly increased within 36 hours (Fig. 2). With no additional rain for two months, WP again declined showing the same pattern of differences among taxa. By early December, WP of S. mellifera was 3.6 MPa less than S. apiana. This same pattern of increasing and decreasing WP occurred again after a mid-December rain and a subsequent month-long dry period. After the middle of January 1996, rel- atively large rainstorms occurred almost every week and by early February WP of all taxa in- creased to above —0.5 MPa as had occurred the previous spring. There were no significant differ- ences among taxa during the wet period from Feb- ruary through the end of the study in late April 1996 (Fig. 2). The seasonal pattern in midday water potentials among these three taxa throughout the study was 144 MADRONO TABLE 1. [Vol. 45 MEAN MID-DAY WATER POTENTIALS (MPa) OF SALVIA APIANA, S. MELLIFERA, AND THEIR HYBRID IN LATE JUNE 1996 AT THREE DIFFERENT SITES: STARR RANCH, SAN JOAQUIN HILLS, AND SAN GABRIEL CANYON (SEE METHODS). Overall significance levels (P) for taxon effects were determined by ANOVA; means within a site sharing the same superscript did not differ significantly (P < 0.05) as determined by Fisher’s aposteriori comparison of means. Starr Ranch S. apiana ==] 867 Hybrid —2,90° S. mellifera =3.46° Pooled SD +0.49 le <0.001 the same as for predawn values, with all taxa ex- periencing midday depressions in WP of about —0.4 to —0.6 MP relative to predawn (data not pre- sented). In addition, to confirm the generality of the results obtained at the Starr Ranch site, WP was measured during the early summer drought (23-25 June 1996, 10 weeks since last spring rain) at two additional sites in southern California. At all three sites there were significant differences among taxa, with higher WP exhibited by S. apiana, interme- diate values by the hybrids, and the most negative values by S. mellifera (Table 1). DISCUSSION When these taxa co-occur at the same site, S. apiana maintains a significantly higher WP during summer drought than S. mellifera (Fig. 2, Table 1). The values of WP of S. mellifera obtained here are similar to other published reports of decreases from —5 to —9 MPa by the end of the summer drought (Mooney 1977; Gill and Mahall 1986; Kolb and Davis 1994). In San Diego Co., WP for S. apiana had decreased to —3.7 MPa at the end of summer drought, whereas WP in Artemisia californica, an- other dominant at many CSS sites, decreased to < —6.5 MPa (Poole and Miller 1975). This limited data from the literature on WP of CSS species also suggest that S. apiana does not experience WP as negative as S. mellifera or A. californica. Salvia apiana appears to be able to ‘avoid’? summer drought relative to S. mellifera, which appears to be able to “‘tolerate’’ prolonged periods of very low WP (avoidance/tolerance sensu Mooney and Dunn 1970). The ability of S. apiana to maintain much higher WP during summer drought could account in part for distribution patterns of these species, ap- parently permitting survival of S. apiana in drier sites. The functional basis for these different responses to summer drought does not appear to be related to summer deciduousness. By the end of summer, S. apiana retains ca. 50% of its spring-time leaf area, whereas S. mellifera becomes close to fully decid- uous (Gill and Mahall 1986). Another possible ex- planation could be rooting depth. While most CSS species are apparently shallowly rooted, very little quantitative data is available. Hellmers et al. (1955) present limited data which indicate that S. apiana San Joaquin Hills San Gabriel Canyon = 1872 “hele? =3.07? —2.64° =4./6° = 3102 +0.82 +0.42 <0.001 <0.001 (n = 1) may have slightly greater rooting depth than S. mellifera (n = 2), 1.5 m compared to 0.6 m deep, respectively. The greater rise in WP im- mediately after the first rain of the season by S. mellifera (Fig. 2) also suggests that at least some of the roots of S. mellifera are shallower than those of S. apiana; however, it seems unlikely that small differences in rooting depth alone could account for the large differences in WP between the species during summer drought. Perhaps S. apiana has | more conservative rates of transpirational water — loss. Whole-leaf absorptance of photosynthetically active radiation (400—700 nm) is about 0.60 for the whiter leaves of S. apiana and almost 0.85 for the green leaves of S. mellifera (Ehleringer and Com- stock 1989). Lower leaf absorptance in S. apiana — (pubescent leaves, interior distribution) than S. mel- lifera (glabrous, more maritime) may help reduce | leaf temperatures and transpirational demand in > ways analogous to pubescent versus glabrous spe- | cies of Encelia (Ehleringer and Cook 1990). During summer drought, the hybrids exhibited WP intermediate between the parental species (Fig. — 2, Table 1). We are aware of few studies on com- parative WP of natural hybrids and parental spe- — cies. In a study of Arctostaphylos patula E. Greene | (generally more mesic), A. viscida C. Parry (gen- © erally more xeric), and their hybrids in the Sierra | Nevada, the hybrids tended to have WP interme- | diate between parental species (Ball et al. 1983). The tissue osmotic potentials and elastic modulus © of hybrids of two species of Dubautia in Hawaii also exhibited values intermediate between parental | species, but WP was not measured (Robichaux 1984). The hypothesis that these hybrids are competi- | tively inferior to the parental species except in dis- _ turbed sites (Epling 1947; Anderson and Anderson 1954; Meyn and Emboden 1987) is difficult to test since necessary levels of disturbance were not spec- — ified. However, based on seasonal changes of WP at our study sites (Fig. 2, Table 1) it appears that © hybrid individuals are not inferior to parental spe-_ cies in terms of competition for soil moisture at least among adult plants. Current hybridization be- tween S. apiana and S. mellifera has produced in- dividuals that are intermediate relative to parental species in terms of ability to avoid experiencing 1998] very low WP during summer drought. Controlled crosses could help elucidate the genetic basis for the different responses to summer drought of these two important CSS species. ACKNOWLEDGMENTS We gratefully acknowledge access to Starr Ranch Na- tional Audubon Sanctuary, the support and suggestions of Peter and Sandra DeSimone, and use of rainfall data col- lected at the Sanctuary. We also acknowledge financial support from California State University, Fullerton, Intra- mural Faculty Research Grants to DSG and Department of Biological Sciences Graduate Research Grants to BJH. Careful reviews of the manuscript by Jack Burk and C. Eugene Jones were also gratefully appreciated. LITERATURE CITED AXELROD, D. I. 1978. The origin of coastal sage vegeta- tion, Alta and Baja California. American Journal of Botany 65:117-131. ANDERSON, E. AND B. R. ANDERSON. 1954. Introgression of Salvia apiana and Salvia mellifera. Annals Mis- souri Botanical Garden 41:329-—338. BALL, C. T., J. KEELEY, H. MOONEY, J. SEEMAN, AND W. WINNER. 1983. Relationship between form, function, and distribution of two Arctostaphylos species (Eri- caceae) and their putative hybrids. Acta Oecologica 4:153-164. DESIMONE, S. A. AND J. H. BuRK. 1992. Local variation and distributional factors in Californian coastal sage scrub. Madrofio 39:170-188. EHLERINGER, J. H. AND J. P Comstock. 1989. Pp. 21—24 Stress tolerance and adaptive variation in leaf absorp- tance and leaf angle. in S. Keeley (ed.), The Califor- nia Chaparral: Paradigms revisited. Natural History Museum of Los Angeles County, Los Angeles, CA. EHLERINGER, J. H. AND C. S. Cook. 1990. Characteristics of Encelia species differing in leaf reflectance and transpiration rates under common garden conditions. Oecologia 82:484—489. EPLING, C. 1938. The California Salvias. A review of Sal- via, section Audibertia. Annals Missouri Botanical Garden 25:95-188. EPLING, C. 1947. Natural hybridization of Salvia apiana and S. mellifera. Evolution 1:69—78. GILL, D. S. AND B. E. MAHALL. 1986. Quantitative phe- nology and water relations of an evergreen and a de- ciduous chaparral shrub. Ecological Monographs 56: 127-143. GRANT, K. A. AND V. GRANT. 1964. Mechanical isolation GILL AND HANLON: WATER POTENTIAL OF SALVIA 145 of Salvia apiana and Salvia mellifera (Labiate). Evo- lution 18:196—212. HARRISON, A. T., E. SMALL, AND H. A. Mooney. 1971. Drought relationships and distribution of two medi- terranean-climate California plant communities. Ecol- ogy 52:869-875. HICKMAN, J. C., (ed.). 1993. The Jepson manual. Univer- sity of California, Berkeley, CA. HELLMERS, H., J. S. HORTON, G. JUHREN, AND J. O’ KEEFE. 1955. Root systems of some chaparral plants in south- ern California. Ecology 36:667—678. KIRKPATRICK, J. B. AND C. EK HUTCHINSON. 1980. The en- vironmental relationships of California coastal sage scrub and some of its component communities and species. Journal of Biogeography 7:23-28. KOLB, K. J. AND S. D. Davis. 1994. Drought tolerance and xylem embolism in co-occurring species of coastal sage and chaparral. Ecology 75:648—659. MEyn, O. AND W. A. EMBODEN. 1987. Parameters and con- sequences of introgression in Salvia apiana X S. mel- lifera (Lamiaceae). Systematic Botany 12:390-—399. MINITAB INC. 1995. Minitab reference manual, release 10Xtra. State College, PA. Mooney, H. A. 1977. Southern coastal scrub. Pp. 471— 478 in M. G. Barbour and J. Major (eds.), Terrestrial vegetation of California. Wiley-Interscience, New York. Mooney, H. A. AND E. L. DUNN. 1970. Convergent evo- lution of mediterranean climate evergreen sclero- phyllous shrubs. Evolution 24:292-—303. O’LgEary, J. E 1990. Californian coastal sage scrub: gen- eral characteristics and considerations for biological conservation. Pp. 24—41 in A. A. Schoenherr (ed.), Endangered plant communities of southern Califor- nia. Southern California Botanists Special Publication #3, Rancho Santa Ana Botanic Garden, Claremont, CA. PooLe, D. K. AND P. C. MILLER. 1975. Water relations of selected species of chaparral and coastal sage com- munities. Ecology 56:1118—1128. RIESEBERG, L. H. 1995. The role of hybridization in evo- lution: Old wine in new skins. American Journal of Botany 82:944-953. RITCHIE, G. A. AND T. M. HINCKLEyY. 1975. The pressure chamber as an instrument for ecological research. Ad- vances in Ecological Research 9:165—254. ROBICHAUX, R. H. 1984. Variation in the tissue water re- lations of two sympatric Hawaiian Dubautia species and their natural hybrids. Oecologia 65:75-81. WESTMAN, W. E. 1981. Diversity relations and succession in Californian coastal sage scrub. Ecology 62:170-— 184. WEBB, A. AND S. CARLQUIST. 1964. Leaf anatomy as an indicator of Salvia apiana-mellifera introgression. Al- iso 5:437—449. MabpDrONO, Vol. 45, No. 2, pp. 146-161, 1998 INVENTORY OF THE VASCULAR FLORA OF THE BLAST ZONE, MOUNT ST. HELENS, WASHINGTON JONATHAN H. Titus! Department of Botany, Box 355325, University of Washington, Seattle, WA 98195 SCOTT MOORE 7009 23rd Avenue NW, Seattle, WA 98117 MILDRED ARNOT 717 E. First Street, Arlington, WA 98223 PRISCILLA J. TITUS Department of Biological Sciences, University of Nevada, Las Vegas, Las Vegas, NV 89154-4004 ABSTRACT Mount St. Helens is an active volcano located in the Cascade Range of southwestern Washington. The volcano erupted in 1980 and created a wide array of devastated landscapes. Since the eruption, vegetation has been colonizing these new landscapes. In the summers of 1993 and 1994 we inventoried plant species and their relative abundance on the Pumice Plain, Plains of Abraham, Toutle Debris Avalanche, Toutle Ridge, and the crater. We distinguished plants as those found in primary successional uplands and wetlands and in refugia. Refugia are defined as habitat where plants survived the eruption as rootstock. The principal refugia are located between the Pumice Plain and Plains of Abraham. The current flora is dominated by wind-dispersed invasive species, mainly those in the families Aster- aceae, Poaceae, Cyperaceae, and Onagraceae. Large-seeded, late-successional understory species are com- mon in refugia and, to a limited extent, have spread into primary substrates. The species documented comprise 341 species in 178 genera and 53 families. These species comprise 4 Sphenophyta, 6 Pterophyta, 9 Coniferophyta, and 322 Anthophyta (221 Dicotyledonae and 101 Monocotyledonae). Fifty-seven of the species are non-native. Species were surveyed for relative abundance on a three way scale—widespread, locally common, and infrequent. This checklist provides a baseline to judge future composition of the flora and plant invasion patterns. Mount St. Helens is an active volcano located in the Cascade Range of southwestern Washington in the southern Cascade physiographic Province (46°12'N, 122°11'W). Pre-eruption forests were typ- ical of the montane Abies amabilis (Douglas) James Forbes zone (Franklin and Dyrness 1988). After 130 years of inactivity, a major series of eruptions oc- curred beginning with a violent eruption on 18 May 1980. These eruptions created a diversity of devas- tated landscapes, and since the eruption, vegetation has been slowly colonizing these new landscapes. The many studies that have been conducted on Mount St. Helens have dramatically increased our understanding of primary successional processes, i.e., the patterns of revegetation of a devastated land- scape (del Moral and Bliss 1993; del Moral et al. 1995; del Moral and Wood 1993a, b; Frenzen et al. 1994; Tsuyuzaki and Titus 1996; Wood and del Mor- al 1987, 1988). However, since no baseline inven- tory of the taxa present on the mountain has been available for use in assessments of the vegetation, we inventoried the vegetation of the primary-suc- ' Present address: Department of Biological Sciences, University of Nevada, Las Vegas, Las Vegas, NV 89154- 4004. cessional substrates on the mountain to provide a > base-line species list for future studies. | Cascade Range volcanoes are isolated from other | areas of similar elevation. These volcanoes are high | elevation islands surrounded by mountain ranges which are more than 1000 m below the volcanic — peaks. Thus, a species list of Mount St. Helens veg- © etation may also be useful in understanding the bio- | geography of the Cascades volcanoes. | Plant inventories of primary successional envi- | ronments on volcanoes are infrequent. Tsuyuzaki — (1995) provides a species list for Mt. Usu in Japan, which erupted in 1977 and 1978. Other Cascade | volcanoes, which erupted thousands of years ago, ° have been surveyed and the results are contained | in Burnett (1985) for Mt. Hood, Wynd (1936) and | Applegate (1939) for Mt. Mazama (Crater Lake), | Cooke (1940) for Mt. Shasta, Gillett et al. (1961) | and Oswald et al. (1995) for Mt. Lassen, Ireland | (1968) for the Three Sisters, Jones (1938) for Mt. . Rainier, and St. John and Hardin (1929) for Mt. | Baker. Other primary successional environments | are occasionally inventoried, such as the Thompson — and Wade (1991) checklist of the vegetation of a 12-year-old surface-mined coal area in Kentucky. Plants that successfully invade primary succes- . sional landscapes usually have seeds which are 1998] dispersed over long distances. These small, aerial- ly-dispersed seeds, however, are only marginally capable of developing viable seedlings under harsh conditions. Large-seeded species, although more adapted to surviving as seedlings under harsh con- ditions, are slower to reach these devastated land- scapes (Wood and del Moral 1987; del Moral 1993; del Moral and Bliss 1993). Plants which invade these landscapes fall into groups of functional traits (Tsuyuzaki and del Moral 1995). Functional traits are attributes or life history characteristics that groups of species exhibit in response to similar en- vironmental pressures, and can be used to infer constraints imposed by the environment. Thus, an inventory of the species present on Mount St. Hel- ens provides an opportunity to study ecological traits of invading species. ENVIRONMENTAL CONDITIONS ON MOUNT ST. HELENS Climate. The climate of the Mount St. Helens area is montane maritime, i.e., cool, wet winters with a significant snowpack at higher elevations, and warm dry summers. The climate is character- ized by abundant yearly precipitation and short pe- riods of drought, typically in July and August. An- nual precipitation averages 2373 mm, yet often less than 5% of this falls between June and August. The snow-free growing season extends from April or May until late September, but it usually begins in June and ends by early September. Temperatures range from mean monthly minima of —4.2°C in January and 7.3°C in August to maxima of 0.5°C in January and 22.2°C in July (Spirit Lake Ranger Station (987 m), Anon. 1969; Reynolds and Bliss 1986). Summer temperatures range from 0 to 35°C with a mean ca. 12°C (Reynolds and Bliss 1986). Considerable variation is an important aspect of the Mount St. Helens climate. Precipitation shows dra- matic annual variation for the summer months of July and August. Surface soil temperatures are of- ten high on the devastated landscapes, approaching 50°C on tephra surfaces near Spirit Lake (Reynolds and Bliss 1986). The eruption. The eruption of May 18, 1980, cre- ated a wide range of disturbances including tephra (all airborne materials ejected from a volcano, in this case pumice of different sizes), pyroclastic flows (masses of hot dry rock flowing suspended in air), hot airblasts, and lahars (unsorted mud and debris flows). Some of the features created by these disturbances include a pyroclastic flow into Spirit Lake, a 550 km? area of blown-down trees bordered by 96 km? of scorched trees, a 60 km? debris ava- lanche, and massive mudflows down the major _ Streams draining the area (Lipmann and Mullineaux _ 1981). Most areas in the perimeter of the devasta- tion were salvage logged and replanted. Because _ the major force of the eruption was directed later- ally to the north, vegetation on the southern flank TITUS ET AL.: MOUNT ST. HELENS CHECKLIST 147 of the volcano was not destroyed except in a few locations by lahars. The eruption reduced the height of the cone from 2950 m to 2550 m. The focus of this study is on those areas where the vegetation was essentially eliminated and re- vegetation was dependent upon the dispersal of propagules from outside the site. These areas are: the Pumice Plain directly north of the cone; the Plains of Abraham to the east of the cone; refugia areas with surviving plants between the Pumice Plain and the Plains of Abraham; the crater breach area where the crater wall was demolished by the lateral eruption; the crater proper, which contains the lava dome; the Toutle Debris Avalanche; and Toutle Ridge between the North and South forks of the Toutle River (Fig. 1). The Pumice Plain, Toutle Debris Avalanche, and the breach contain a variety of both upland and wetland habitats. The Toutle Ridge contains refugia where the vegetation was shielded from the blast. The refugia between the Pumice Plain and Plains of Abraham contains many species which survived the eruption as rootstock. Pumice Plain The Pumice Plain ranges in elevation from 1150 to 1300 m and occupies an area of approximately 20 km? immediately north of the crater. This area was formed by the deposit of over 100 m of ma- terial from the debris avalanche, subsequent pyro- clastic flows, and incandescent pumice depositions. It was also repeatedly impacted by later lahars (Lip- mann and Mullineaux 1981). The current surface is generally flat or gently sloping with numerous gul- lies formed by water-erosion dissecting the surface. Before the eruption, the Pumice Plain was vegetat- ed by open coniferous forest (Kruckeberg 1987) with timberline 600—800 m below its climatic limit (Lawrence 1938). Now the area is blanketed by pumice ranging in depth from 10 to 200 m (Lip- mann and Mullineaux 1981). Surface pumice par- ticles range in size from 1 mm to 10 cm. Surface colors range from light to dark gray with high sur- face albedo (Reynolds and Bliss 1986). The surface layer of pumice acts as a mulch that impedes evap- oration from below; thus, considerable moisture may be present at lower depths while surface layers may be very dry. In the summer, the surface pum- ice dries quickly between rains. Therefore, most Pumice Plain habitats probably do not remain moist long enough to allow seedling establishment. At greater depths, however, the soil remains moist so that adult plants rarely suffer from drought (Reyn- olds and Bliss 1986; Chapin and Bliss 1988; Pfitsch and Bliss 1988; del Moral and Bliss 1993). Sub- strates with fine particles contain more moisture than areas with coarse particles, and the erosion rills were slightly moister than other microsites (del Moral and Bliss 1993). The Pumice Plain soils are pedologically immature with very low concentra- tions of N, K, P, organic matter, and microbial bio- mass (Nuhn 1987; Halvorson et al. 1991, 1992). 148 MADRONO Coldwater Lake ~S . 4 NN ‘Debris Avalanche \7 i) NINA pce NIN y Ac RAWAL Fes A wan ae Castle Lake Toutle Ridge Refugia Fic. 1. Although seed sources for the Pumice Plain are distant (>3 km), the seed rain is dense with the seeds of species commonly found in upwind clear- cuts (del Moral and Bliss 1993). In the post-erup- tion landscape, vegetation is sparse and was, until recently, confined to rills, but has now spread across the Plain. High surface soil temperatures and frequent summer droughts may explain why seed- ling establishment has been very low in many plac- es on the Pumice Plain since the eruption (del Mor- al and Bliss 1993). Four upland habitat types were recognized on the Pumice Plain by del Moral et al. (1995). These were refugia, pumice barrens, pyroclastic surfaces, and drainages. In this study refugia are treated as a separate habitat type, and the other three upland habitats were combined due to high floristic simi- [Vol. 45 Harry's Ridge Spirit Lake Pumice Plain se ae! _. Plains of ._J Abraham ok 47 Map of the study areas on Mount St. Helens, Washington. larity. Scattered on the pyroclastic and stable drain- age surfaces are sites that developed dense patches of Lupinus lepidus Douglas within two years after | the eruption. Initially, these patches were nearly | monospecific, but eventually they created biologi- | cal oases that facilitated invasions of open-site spe- | cies (Wood and Morris 1990; del Moral 1993; del Moral and Wood 1993a). Since the large seeds of L. lepidus are too heavy to be wind dispersed, the | patches apparently originated from a few individ- uals which survived the eruption (Bishop 1996). The wetland areas of the Pumice Plain are found | primarily along permanent and seasonal creeks and | in or near springs. Both hot and cold water springs are found near Spirit Lake and both support lush | wetland vegetation. The waters of Spirit Lake itself | Support aquatic vegetation. 1998] Refugia Refugia exist on north-facing slopes close to the cone. During the eruption, these snow-covered ar- eas provided refuge for many species and enabled them to survive the eruption as rootstock, particu- larly late-successional forest understory species. These areas were usually steep, and thus erosion quickly removed the pumice layer and exposed the rootstock for resprouting. Refugia have also been invaded by open-site species characteristic of the Pumice Plain (del Moral et al. 1995). Soils of re- fugia are richer in nutrients than the Pumice Plain (J. H. Titus unpublished) and contain mycorrhizae, which are essential to the survival of many of the late-successional forest understory species (Titus 1995). Plains of Abraham The Plains of Abraham is at an elevation of 1400-1450 m in an area of gentle topography, and was barren prior to the eruption (Kruckeberg 1987). Although the main fury of the 1980 eruption was directed northward, the Plains of Abraham, located to the east of the cone, was also devastated by the blast and the resulting pyroclastic flows and lahars (Foxworthy and Hill 1982). In 1980 the Plain was covered by tephra and pockets of silt, and vascular plants were absent. By 1987 wind erosion had re- moved all fine materials so that the surface now consists of coarse pumice generally 2—5 cm in di- ameter. Water erosion initiated numerous shallow rills in the pumice-dominated landscape. The pum- ice on the Plains of Abraham is of pre-1980 origin. Thus soil nutrient levels are higher than on the Pumice Plain because the substrate was formed pri- or to the most recent eruption (del Moral and Bliss 1993). As on the Pumice Plain, rills were moister than surrounding flat sites, and fine textured soils had twice the moisture content of coarse soils (del Moral and Bliss 1993). Over time, the pumice weathers to a fine sand. Wind removes much of this material so that a “‘desert pavement’’ is formed. Areas not eroded by water are essentially smooth, lack sites where seeds might lodge, and dry rapidly. Primary succession on the Plains of Abraham is proceeding under highly stressful conditions, and plant cover is still relatively sparse. Microsite vari- ation is distinct with plants much more prevalent on rill edges (del Moral and Wood 1993a). Toutle River Debris Avalanche The debris avalanche was formed by a massive landslide of the north flank of the mountain that occurred during the eruption. The debris material averages 45 m in depth, 2 km in breadth, and ex- tends 25 km from the crater along the North Fork of the Toutle River (Lipman and Mullineaux 1981). Plant devastation on the debris avalanche was vir- tually complete (Adams et al. 1982); however, rare TITUS ET AL.: MOUNT ST. HELENS CHECKLIST 149 individuals of at least 20 species survived on the debris deposit, the most widespread being Epilo- bium angustifolium L. ssp. circumvagum Mosgq., Cirsium arvense (L.) Scop., and Lupinus latifolius J. Agardh. These individuals apparently resprouted from plant fragments that had been transported by the debris slide. The most common woody species to survive were willows, which regenerate readily from root and stem fragments. These willows were found primarily in the western portion of the ava- lanche and farthest from the crater (Dale 1986). Re- covery of the vegetation was initiated within the first year after the eruption, and rapid colonization by Alnus rubra Bong. and other riparian species has occurred since that time (Adams et al. 1987). Re- covery has been slow in other areas, however, es- pecially in areas closer to the mountain. The pri- mary seed source for the debris avalanche is seed rain from adjacent scorched pre-eruption clearcuts, heavily vegetated by herbaceous perennials that produce copious light, wind-dispersed plumed seeds (Franklin et al. 1985; Dale 1989). Debris avalanche soils have an adequate balance between moisture retention and aeration properties, and percent organic matter is low but adequate for plant growth (Adams and Dale 1987). The soils are sandy or silty sand with the greater than 2 mm frac- tions comprising about 65% of soil samples. A massive reseeding effort was conducted in 1980 and 1981 on the Toutle River debris ava- lanche using non-native grasses and forbs to reduce erosion. Many of the seeded species did not grow well on the new substrates, and there is no evidence that erosion was reduced (Dale 1989; Tsuyuzaki 1995). However, many other seeded species sur- vived and now dominate much of the mudflow. This dominance by invasive non-native species may be slowing natural succession. Our survey was conducted on the eastern portion of the mudflow and terminated at a line roughly connecting the south side of Coldwater Lake to the northwest side of Castle Lake. This boundary line is at an eleva- tion of approximately 760 m (Fig. 1) and was se- lected so as to avoid heavily artificially seeded ar- eas. Much of the habitat west of our study area is dominated by non-native species, particularly leg- umes. Wetlands on the Toutle debris avalance appear to be unique and have not been studied. Four distinct types can be distinguished based on hydrological and physical characteristics. Deep potholes created by large masses of entrained ice are now occupied by wetlands composed of temporary snow melt and rain water ponds or permanent spring-fed ponds. Broad cattail-dominated marshes spread across the mudflow. Riparian vegetation along many creeks is abundant, and numerous springs on the mudflow support a diverse flora. Two types of riparian hab- itat are found on the debris avalanche: unstable ri- parian areas that support a sparse herbaceous veg- 150 etation, and more stable riparian areas dominated by dense willow thickets. Toutle Ridge Toutle Ridge is an elevated area located between the North and South Forks of the Toutle River. This area was devastated by the eruption, although in a somewhat peripheral manner similar to that of the Plains of Abraham. There are two small refugia on the Ridge, but no wetlands. The study area ends to the west of the Ridge where blown-down trees are common and the devastation was less intense. Crater Since the creation of a crater by the eruption of 1980, a lava dome has been growing in its center. The crater can be separated into two parts: the breach, which is a wide area where the lateral erup- tion removed the north face of the mountain, and the high crater, which contains the lava dome. The breach contains both upland and wetland habitats from small springs. Creek banks are too unstable to support vegetation. The high crater is mostly barren upland. Very small wetland areas were created on the lava dome where stream from fumaroles con- tinues to condense. These condensation areas prin- cipally support moss and algae, but a few wind dis- persed vascular plants have also colonized these unusual sites. METHODS Census technique. To complete the current in- ventory, surveys were conducted during the sum- mers of 1993 and 1994. The entire inventory area was examined several times over the season, and locations were recorded for all species present. The species list was compiled from observations noted during these surveys. Species were categorized into three abundance categories: widespread, locally common, and infrequent. These are qualitative rankings and are not based on quantitative data al- though in some cases quantitative data assisted in assigning the abundance rank. “‘Widespread”’ in- dicates that the plant was abundant throughout the inventory area. “‘Locally common’’ indicates a plant that was only locally common or occasionally a plant with a more scattered distribution. “‘Infre- quent”’ describes plants which were difficult to de- tect. Determinations and nomenclature. Species de- terminations were made using the Flora of the Pa- cific Northwest (Hitchcock and Cronquist 1973) with updated nomenclature from The Jepson Man- ual: Higher Plants of California (Hickman 1993). Voucher specimens were deposited at the Univer- sity of Washington herbarium. Pre-eruption Conditions Before 1980, the 2950 m Mount St. Helens was surrounded by a patchwork of forested and clearcut MADRONO [Vol. 45 land in varying stages of reforestation. The forests were typical of the Abies amabilis zone (Franklin and Dyrness 1988), composed primarily of Abies amabilis, Abies procera Rehder, Pseudotsuga men- ziesit (Mirbel) Franco var. menziesii, and Tsuga het- erophylla (Raf.) Sarg. (Lawrence 1938). Clearcut land was generally replanted with P. menziesii and A. procera seedlings and had a lush cover domi- nated by herbs (e.g., Epilobium angustifolium and Anaphalis margaritacea (L.) Benth. & Hook.) and shrubs (e.g., Acer circinatum Pursh and Rubus ur- sinus Cham. & Schldl.). The riparian vegetation along the Toutle River was comprised primarily of deciduous trees (Alnus rubra, Populus balsamifera L. ssp. trichocarpa (Torrey & A. Gray) Brayshaw, Salix scouleriana Hook. and Salix sitchensis Bong.). Before the eruption the vegetation on the Plains of Abraham was sparse (Kruckeberg 1987), and timberline was 600-800 m below its climatic limit. The forest that occurred was composed only of scattered conifers dominated by Abies lasiocarpa (Hook.) Nutt. var. lasiocarpa, Polygonum newber- ryi Small, and Penstemon cardwellii (Lawrence 1938). Previous botanical exploration. The pre-eruption | Mount St. Helens flora was considered to be de- | pauperate in species richness in comparison with | other Northwestern Pacific volcanic summits due to geologically recent eruptions and mudflows © (Kruckeberg 1987; del Moral and Wood 1986, | 1988). This geological activity caused the volcano © to have a suppressed timberline (Lawrence 1954). | Since the eruption, researchers have sought an ac- | curate picture of the pre-eruption assemblage of in- | dividual plant species, the plant communities in which they resided, and how those plant commu- © nities related to other nearby volcanic massif com- munities. Unfortunately, no comprehensive flora — exists. A relatively complete pre-eruption flora is inferred by inventories contained within three pub- lications, Flora of the State of Washington (Piper | 1906), The Flora of Mt. St. Helens (St. John 1976), | and Plant Life on Mount St. Helens before 1980 | (Kruckeberg 1987). The first organized botanical exploration of | Mount St. Helens and vicinity was conducted in | 1898 by Dr. E V. Colville, principal botanist of the U.S. Department of Agriculture. However, no flora | was published until Harold St. John researched Col- | ville’s journals and published a brief accounting | (St. John 1976). According to the journal, Colville | traveled and collected throughout Oregon and | Washington. His collections of the Mount St. Hel- | ens area appear to be limited to the south side of | the mountain since he approached from the south | via the Lewis River, camped at Merrill Lake, and then proceeded up the Kalama River to Three | Buttes Camp (1220 m) at the southwest base of Mount St. Helens on 19 August 1898. From there | 1998] he ascended and collected on the mountain. He list- ed 77 species; a partial collection of these are de- posited at the Smithsonian Institution. Colville’s flora remains unpublished, but his collection was used by C. V. Piper to construct a flora of the state of Washington, published by the U.S. National Her- barium in 1906 (Piper 1906). The most comprehensive inventory of the Mount St. Helens flora was conducted in 1925 by Harold St. John and students including C. S. English, Jr. During an eleven-day visit in August 1925, they collected 315 plant species. Eight days were spent botanizing the region surrounding Spirit Lake and the north slope of the mountain, and one day climb- ing to the summit. An additional two days were spent collecting on the south side, exploring the upper valley of the Lewis River and the area around the present town of Cougar. The result of the in- ventory was not published until 1976, fifty years after the fieldwork was completed (St. John 1976). The collection is stored at the Washington State University Marion Ownbey Herbarium in Pullman, Washington. The last documented botanical exploration of Mount St. Helens prior to the eruption was con- ducted in 1979 by Dr. Arthur Kruckeberg and eight- een members of the Washington Native Plant So- ciety (Kruckeberg 1979). They identified approxi- mately 86 plant species. They limited their inves- tigation to timberline and above, beginning at Timberline Camp and the adjacent parking lot (1340 m—this area is now part of the Pumice Plain), and ascending to just below Sugar Bowl on the NE face (2075 m). Additional observations were made on a second day’s traverse of Windy Pass (1495 m) and across the Plains of Abraham to the head of Ape Canyon (1280 m) on the southeast flank of the mountain. DESCRIPTION OF CURRENT VEGETATION Mount St. Helens is surrounded by thousands of hectares of recent clearcuts that are thickly vege- tated with weedy wind-dispersed invasive species. The clearcuts provide a seed source for the recently created landscapes found on the volcano (del Moral and Bliss 1993). Thus, weedy invasive species typ- ical of clearcuts were found at all sites examined during our surveys. Upland primary successional landscapes were similar in vegetation, except for the Toutle River Debris Avalanche which had ad- ditional dense stands of non-native legumes. The refugia have a vegetation distinct from the primary ‘successional landscapes. Tree species were infre- quent or only locally common. However, dense groves of trees did not occur on the Toutle River Debris Avalanche. _ An unusual species that occurred in several of the areas was Salix exigua Nutt. This species generally occurs to the east of the Cascades. The probable cause for the presence of this species is that the open TITUS ET AL.: MOUNT ST. HELENS CHECKLIST 151 TABLE 1. NUMBER OF NATIVE AND NON-NATIVE SPECIES IN HABITATS OF THE Mount ST. HELENS BLAST ZONE. Number Number of of non- native native Habitat species species Pumice Plain uplands 151 20 Pumice Plain wetlands 110 i Refugia 160 10 Plains of Abraham 65 4 Toutle Debris Avalanche uplands 146 44 Toutle Debris Avalanche wetlands 114 42 Toutle Ridge uplands 86 3 Toutle Ridge refugia 93 6 Breach uplands 47 5 Breach wetlands 27 5 High Crater 14 3 primary successional habitats of Mount St. Helens provide colonization sites for widely dispersed wind- dispersed species such as Salix species. Pumice Plain Upland. The barren areas of the Pumice Plain are floristically consistent across the Plain and have low cover. Anaphalis margaritacea, Hypochaeris radicata L., Lupinus lepidus, Epilobium angustifol- ium, Penstemon cardwellii Howell, Penstemon ser- rulatus Menzies, Hieracium albiflorum Hook., Car- ex mertensii Prescott, Carex spectabilis Dewey, Agrostis pallens Trin., Agrostis scabra Willd., and Juncus parryi Engelm. are typically among the leading dominants in all barren sites. Densely veg- etated areas of the Pumice Plain tend to be domi- nated by these same species. Across the barren ar- eas of the Pumice Plain, L. lepidus also forms ex- tensive densely vegetated patches. One hundred and seventy-three species were found on Pumice Plain uplands (Tables 1 and 2). Wetland. Juncus and Salix species often domi- nate springs and wet areas, especially Salix sitch- ensis Bong., Juncus bufonius L., and Juncus acu- minatus Michaux. Equisetum arvense L. also oc- cupies broad wet areas. Epilobium watsonii Barbey, Salix sitchensis, and Mimulis guttatus DC. are com- mon along creeks. The waters of Spirit Lake con- tain Potamogeton natans L., Myriophyllum sibiri- cum V. Komarov, and Ranunculus aquatilis L. Ex- tensive algal mats occur in both thermal and cold water springs. Wetland areas are usually more thickly vegetated and diverse than upland areas. One hundred and twenty-one species were found in Pumice Plain wetlands. Refugia Refugia are dominated by woody species such as Alnus viridis (Chaix) DC. ssp. sinuata (Regel) A. Love & D. Love, Ribes laxiflorum Pursh, Rubus 152 MADRONO [Vol. 45 TABLE 2. CHECKLIST OF WASCULAR PLANT SPECIES IN PRIMARY SUCCESSIONAL HABITATS ON MounrtT ST. HELENS. w = widespread; c = locally common; i = infrequent; 1 = Pumice Plains uplands; 2 = Pumice Plain wetlands; 3 = Refugia; 4 = Plains of Abraham; 5 = Toutle Debris Avalanche uplands; 6 = Toutle Debris Avalanche wetlands; 7 = Toutle Ridge uplands; 8 = Toutle Ridge refugia; 9 = Breach uplands; 10 = Breach wetlands; 11 = High Crater; E = exotic non-native species. Species 9 Cupressaceae Thuja plicata Pinaceae Abies amabilis Abies lasiocarpa Abies procera Pinus contorta var. latifolia Pinus monticola Pseudotsuga menziesii Tsuga heterophylla Tsuga mertensiana Equisetaceae Equisetum arvense Equisetum fluviatile Equisetum hyemale ssp. affine Equisetum palustre Blechnaceae Blechnum spicant Dennstaedtiaceae Pteridium aquilinum var. pubescens Dryopteridaceae Athyrium filix-femina Cystopteris fragilis Polystichum munitum Pteridaceae Cryptogamma cascadensis Aceraceae Acer circinatum Acer glabrum var. douglasii Acer macrophyllum Apiaceae Heracleum lanatum Lomatium martindalei Oenanthe sarmentosa Osmorhiza purpurea Araliacene Oplopanux horridum Asteraceae Achillea millefolium Agosteris aurantiaca Agoseris glauca var. glauca Agoseris grandiflora Agoseris heterophylla Anaphalis margaritacea Antennaria rosea Antennaria umbrinella Arnica cordifolia var. cordifolia Arnica latifolia var. gracilis Arnica mollis Arnica nevadensis Aster brachyactis Aster frondosus Aster ledophyllus var. ledophyllus Aster modestus Cirsium arvense var. horridum (E) Cirsium vulgare (E) ee Cee ee ee —s ms O eee ee SS pete pete pte te eee © OC fe) meee ee Ss — OQ pmo pie Q meee es OO wd ie ete ete O jalicc pes r= KO 0 € QrQgr = Hee ee S 10 11 1 i 1 c c i 1 Ww Cc 1998] TITUS ET AL.: MOUNT ST. HELENS CHECKLIST TABLE 2. CONTINUED 153 Species 1 2 3 N Conyza canadensis Crepis capillaris (E) Erigeron aliceae Erigeron peregrinus var. callianthemus Eriophyllum lanatum var. lanatum Gnaphalium canescens ssp. thermale Gnaphalium purpureum Gnaphalium uliginosum (E) Hieracium albiflorum Hieracium gracile Hypochaeris radicata (E) Lactuca muralis (E) Lactuca serriola (E) Lapsana communis (E) Leontodon taraxacoides spp. taraxacoides (E) Leucanthemum vulgare (E) Luina hypoleuca Petasites frigidus var. palmatus Senecio fremontii var. fremontii Senecio jacobaea (E) Senecio sylvaticus (E) Senecio triangularis var. triangularis Senecio vulgaris (E) Solidago candensis ssp. elongata Sonchus arvensis (E) Sonchus asper ssp. asper (E) Sonchus oleraceus (E) Taraxacum officinale (E) Trimorpha lonchophylla Berberidaceae Achlys triphylla ssp. triphylla Vancouveria hexandra Betulaceae Alnus rubra Alnus viridis ssp. sinuata Boraginaceae Myosotis laxa Brassicaceae Cardamine oligosperma var. oligosperma Cardamine pensylvanica Draba verna Rorippa curvisiliqua var. lyrata Rorippa nasturtium-aquaticum Rorippa palustris var. occidentalis Callitrichaceae — Callitriche Stagnalis (E) -Campanulaceae Campanula rotundifolia Campanula scouleri Caprifoliaceae Linnaea borealis var. longiflora Lonicera ciliosa Sambucus racemosa vat. arborescens Caryophyllaceae Arenaria serpyllifolia ssp. serpyllifolia (E) Cerastium arvense Cerastium nutans Moehringia macrophylla Sagina saginoides CQ ote ete pete - dodge pedis: Jami’ = esis pee? fee. ed fk, > te) jade, eae fads eas! Q ee ee ee ee S pets pee pm ep mms Cote ete me ete te OC —_ < pete pete ee < pie pete me Ce ee ee 154 MADRONO [Vol. 45 | TABLE 2. CONTINUED Species f 2. 8 of GS 96:7) 2-85 39°7a0 ae Silene parryi 1 | Spergula arvensis ssp. arvensis (E) i Spergularia marina Spergularia rubra (E) 1 1 1 1 Stellaria borealis ssp. stichana Stellaria calycantha 1 1 i Stellaria crispa 1 1 Cc Stellaria nitens QO = =O =: QA 0O0F0O ee Be: Celastraceae Paxistima myrsinites 1 Cornaceae Cornus canadensis i c Crassulaceae Sedum oreganum 1 1 i Ericaceae Arctostaphylos nevadensis 1 1 i Cc w Arctostaphylos uva-ursi i Gaultheria ovatifolia 1 i Gaultheria shallon Cc Menziesia ferruginea var. ferruginea w Orthilia secunda 1 Phyllodoce empetriformis Pyrola asarifolia Rhododendron albiflorum Vaccinium membranaceum 1 Vaccinium ovalifolium Vaccinium parvifolium 1 BQ See: = 5 Fabaceae Cytisus scoparius (E) 1 Lotus corniculatus (E) Lotus purshianus var. purshianus Lupinus latifolius var. latifolius Lupinus lepidus var. lepidus Medicago lupulina (E) Melilotus alba (E) Trifolium microcephalum Trifolium pratense (E) Trifolium repens (E) 1 1 << Zfr-f e282 8 ££ FEO =< Gentianaceae Centaurium erythraea (E) 1 w Grossulariaceae Ribes bracteosum Ribes howellii Ribes lacustre i Ribes laxiflorum Ribes sanguineum var. sanguineum 1 Ribes viscosissimum var. viscosissimum 1 220 Zee Halogaraceae Myriophyllum hippuroides i 1 Myriophyllum sibiricum 1 i Hydrophyllaceae Hydrophyllum fendleri var. albifrons Phacelia leptosepala Cc Cc 1 Phacelia mutabilis Phacelia nemoralis ssp. oregonensis Cc Cc 1 — _ _ jai Shiipedle jade tee Q Q Hypericaceae Hypericum perforatum (E) 1 1998] TITUS ET AL.: MOUNT ST. HELENS CHECKLIST 155 TABLE 2. CONTINUED Species l 2 3 4 5 6 e 8 9 10° 11 Onagraceae Circaea alpina ssp. pacifica Epilobium anagallidifolium Epilobium angustifolium Epilobium brachycarpum Epilobium ciliatum ssp. ciliatum Epilobium clavatum Epilobium glaberrimum ssp. fastigiatum Epilobium glaberrimum ssp. glaberrimum Epilobium hornemannii ssp. hornemannii Epilobium lactiflorum Eiplobium luteum Epilobium minutum Cc pete eke ee te KO 20 pe ee ee ee ee Se yee ee ee pete ete pee ee _ _ Oxalidaceae Oxalis oregana i Papaveraceae Dicentra formosa 1 1 Plantaginaceae Plantago lanceolata (E) 1 1 Plantago major var. major (E) 1 Polemoniaceae Collomia heterophylla Ww Cc Collomia larsenii (S Collomia tinctoria 1 1 Phlox diffusa 1 Ww Phlox gracilis 1 1 Polygonaceae Eriogonum pyrolifolium var. coryphaeum w 1 w w 1 Polygonum douglasii ssp. douglasii Polygonum minimum Ww 1 Cc 1 Cc Cc Polygonum newberryi 1 Polygonum persicaria (E) Rumex acetosella (E) i 1 Rumex crispus (E) Rumex obtusifolius (E) —_ wy << pee pete pe pe Portulacaceae Calyptridium umbellatum var. caudiciferum WwW WwW Ww 1 Ww 1 Claytonia lanceolata var. lanceolata 1 Claytonia sibirica i iC 1 1 Cc Montia parvifolia var. parvifolia Ww Cc 1 1 1 Primulaceae Trientalis latifolia 1 1 - Ranunculaceae Actaea rubra 1 1 Aquilegia formosa w Ranunculus aquatilis var. capillaceus c Cc Ranunculus aquatilis var. hispidulus Ranunculus sceleratus var. sceleratus Cc 1 Trautvetteria caroliniensis var. occidentalis 1 Q at) Rosaceae Amelanchier alnifolia Aruncus dioicus var. pubescens Fragaria virginiana ssp. platypetala Holodiscus discolor Luetkea pectinata Potentilla anserina ssp. anserina Ww z£=0 O eae = = wer < O < 156 MADRONO [Vol. 45 TABLE 2. CONTINUED Species 1 2 3 4 5 6 7 8 9 10) 1 Potentilla glandulosa ssp. glandulosa i Prunus emarginata Rosa gymnocarpa i i 1 Rubus discolor (E) Rubus idaeus var. gracilipes Rubus laciniatus (E) Rubus lasiococcus 1 w Rubus leucodermis i Rubus parviflorus i w i Rubus spectabilis 1 Ww Sibbaldia procumbens Sorbus sitchensis var. grayi w i i i ee ee ee ee ee ee ee gone = Ome mee Spiraea betulifolia Spiraea densiflora var. densiflora Spiraea douglasii Rubiaceae Galium asperrimum 1 Galium triflorum i Salicaceae Populus balsamifera ssp. trichocarpa c 1 1 1 w (e i i Populus tremuloides Cc Salix commutata 1 i Salix exigua ssp. exigua 1 Salix geyeriana Salix lucida ssp. lasiandra 1 C 1 Salix scouleriana 1 1 1 Salix sitchensis w Ww w Cc < feaic penile a pele = ms Ome me Saxifragaceae Heuchera glabra Heuchera micrantha Mitella breweri Mitella pentandra Saxifraga arguta Saxifraga ferruginea var. macounii w Saxifraga tolmiei var. tolmiei 1 Tellima grandiflora Tiarella trifoliata var. unifoliata i Tolmiea menziesii 1 1 a ee ee rr ee ee ee ee Scrophulariaceae Castilleja miniata ssp. miniata c WwW i 1 w ow Digitalis purpurea (E) 1 Mimulus floribundus 1 Mimulus guttatus Cc Cc Mimulus lewisii c Ww Nothochelone nemorosa Cc Parentucellia viscosa i Cc Pedicularis racemosa var. racemosa i i Penstemon cardwellii Ww Ww w Ww Ww Ww i Penstemon confertus i Penstemon rupicola 1 i Penstemon serrulatus Ww c 1 Cc i c c Verbascum thapsus (E) i Veronica americana c Veronica officinalis (E) 1 Cc Veronica serpyllifolia ssp. humifusa 1 1 Q _ Q ae Q Q Q Valerianaceae Valeriana sitchensis ssp. sitchensis w w Violaceae Viola sempervirens 1 (fe i 1 1 1 1998] TITUS ET AL.: MOUNT ST. HELENS CHECKLIST TABLE 2. CONTINUED Species 1 2 3 4 5 Cyperaceae Carex canescens 1 i Carex deweyana ssp. leptopoda 1 Carex illota 1 Carex lenticularis var. lipocarpa i Carex leporinella Carex mertensii Carex microptera Carex ovalis Carex pachystachya Carex paysonis Carex phaeocephala Carex praticola Carex preslii Carex proposita Carex rossii Carex spectabilis Carex stipata var. stipata Carex subfusca c Eleocharis macrostachya w Scirpus acutus var. occidentalis c i c i Cc << a = ars eee Oot Se oS Se: Scirpus americanus Scirpus maritimus Scirpus microcarpus Scirpus tabernaemontani Juncaceae Juncus acuminatus Juncus articulatus Juncus bolanderi Juncus bufonius 1 Juncus effusus var. gracilis Juncus ensifolius var. montanus Juncus mertensianus w Juncus nevadensis var. badius Juncus parryi w Juncus regelii 1 Juncus tenuis var. tenuis 1 1 i Ww 1800 m) western portion of the Ridge. Unlike the other spe- cies on the Ridge it is a subalpine species (Hitch- cock and Cronquist 1973). This species may have invaded from less disturbed ridges to the south or survived the eruption under a thick snow pack in a protected site. The small refugia are dominated by a typical forest understory flora similar to the larger refugia between the Pumice Plain and Plains of Abraham. Ninety-nine species were found in the Toutle Ridge refugia. Crater The crater is sparsely vegetated by herbaceous wind-dispersed species similar to those of upland areas of the Pumice Plain and Plains of Abraham. Vegetation, dominated by Juncus and Epilobium species, is dense in the seepage wetlands. In 1994 the first individual of Lupinus lepidus was found in the breach. This was the first large-seeded, non- wind-dispersed species found in the crater since the eruption. The crater proper, which surrounds the lava dome, is very barren. Only a few sparsely dis- tributed, wind-dispersed species have become es- tablished. The highly unstable nature of the lava dome limits vegetation over most of the dome. Ar- eas where steam from fumaroles condenses rarely support vascular plants, yet a few individuals of Hypochaeris radicata and Salix sitchensis were ob- served. Fifty-two species were found in the breach uplands, 32 species in the breach wetlands, and 17 species in the high crater. Non-native Species The second most widespread invader of upland successional landscapes on the Pumice Plain, Plains of Abraham, and Toutle Ridge (the most wide- spread is Anaphalis margaritacea) is the non-native Hypochaeris radicata. It is possible that the rate of succession has changed substantially because of the presence of this non-native species. Most of the other exotics, such as Cirsium, Senecio, and Son- chus species, do not dominate the landscape. Non- natives are less common in wetlands. The only Cy- tisus Scoparius (L.) Link plant was located on the west edge of the Pumice Plain and was chopped down in 1994. Cytisus scoparius forms dense thick- ets on the Toutle Debris Avalanche outside of the range of this study. In all, 57 non-native species were found. The Toutle River Debris Avalanche has vast ar- eas completely dominated by non-native species to the complete exclusion of natives presumably due to the artificial seeding that was undertaken after the eruption for erosion control. Species such as Lotus corniculatus, Medicago lupulina, Melilotus alba, Trifolium pratense, Trifolium repens, and Vul- pia myuros cover large areas of the Debris Ava- lanche. These species create extensive monocul- 160 tures which exclude native species and prevent nat- ural successional processes from occurring. For ex- ample, the entire mouth of Coldwater Lake is completely dominated by non-native species. Thus, the possibility of natural succession to a scrub- shrub or forested wetland lakeshore is reduced. Other non-natives on the lahar, such as Hypericum perforatum L., Rubus discolor Weihe & Nees, Cir- sium arvense (L.) Scop., and Cirsium vulgare (Savi) Ten., probably arrived on their own via wind or animal dispersal. CONCLUSION The focus of this study was to investigate the recently devastated areas of Mount St. Helens. Spe- cies richness is difficult to compare with studies of other Cascade Range volcanoes in which the entire volcano, including relatively undisturbed areas, was surveyed. These other Cascade Range volcanoes have not been subjected to recent volcanic erup- tions. Because areas investigated during the study var- ied greatly in terms of current physical features, environmental conditions and size of area available for colonization, species richness varied across the landscape. The highest richness was observed in the Toutle Debris Avalanche uplands, probably due to the area’s great size, diversity of physical fea- tures and proximity to seed sources of potential col- onizers. Richness on Mount St. Helens has increased greatly since 1980 when richness equaled zero (ex- cept in the refugia). Although most of the species currently dominant on the landscape are invasive species that characteristically have small wind-dis- persed seeds, some large seeded plants such as Lu- pinus lepidus are common. No threatened, endan- gered or sensitive species or regional endemics were detected during our survey of the volcano. An interesting finding was the presence of Salix exigua. This species is not thought to occur west of the Cascades. The open primary successional habitats created by the volcanic eruption allow colonization by wind-dispersed species from great distances. This study creates a baseline to judge changes to the landscapes of Mount St. Helens, and should also facilitate future studies of primary successional processes on the mountain. ACKNOWLEDGMENTS Thanks to all the following people who contributed to this project. Roger del Moral and Dave Wood provided useful comments on the species list and manuscript. Paul Yurky contributed time and effort to all aspects of this project. Michele Glazer improved the manuscript. Sarah Gage of University of Washington herbarium and Scott Sundberg of Oregon State University herbarium checked the species list for taxonomic correctness. Shiro Tsuyu- zaki, Mandy Tu, John Bishop and Peter Chilson provided companionship in the field and assisted in plant collec- tions. JoEllen VanDeMark and Betsy Lyons assisted with MADRONO [Vol. 45 plant identification. Difficult taxa were verified by Alan Yen (Cyperaceae), John Christy (aquatics), Ed Alverson (ferns), George Argus (Salicaceae), Thomas Hinckley (Pinaceae), Joe Arnett (Silene), and Sarah Gage (problem id’s). Brian Haber created the figure used in this publi- cation. Also, thanks to two anonymous reviewers. A grant from the Washington Native Plant Society made this pro- ject possible. NSF Grants BSR-89-06544 and DEB- 9406987 to Roger del Moral were also of assistance. LITERATURE CITED ADAMS, A. B., V. D. ADAMS, AND E. MERRILL. 1982. Vas- cular plant species impacted by 80/05/18 lateral blast of Mt. St. Helens. Unpublished. AND V. H. DALE. 1987. Vegetative succession fol- lowing glacial and volcanic disturbances in the Cas- cade Mountain Range of Washington, USA. Pp. 70— 147 in D. E. Bilderback (ed.), Mount St. Helens 1980: botanical consequences of the explosive eruption. University of California Press, Berkeley, CA. , V. H. DALE, E. P. SMITH, AND A. R. KRUCKEBERG. 1987. Plant survival, growth form and regeneration following the 18 May 1980 eruption of Mount St. Helens, Washington. Northwest Science 61:160—170. ANON. 1969. Climatological Handbook Columbia Basin States, Vol. 1, Part A and Vol. 2. Meterology Com- mittee, Pacific Northwest Rivers Commission. APPLEGATE, E. I. 1939. Plants of Crater Lake National Park. American Midland Naturalist 22:225—314. BisHop, J. G. 1996. Demographic and population genetic variation during colonization by the herb Lupinus lep- idus on Mount St. Helens. Ph.D. dissertation. Uni- versity of Washington, Seattle, WA. BURNETT, R. E. 1985. Flowering plants of the Mt. Hood area. Mazama 47:45—56. CHAPIN, D. M. AND L. C. BLIss. 1988. Soil-plant water relations of two subalpine herbs from Mount St. Hel- ens. Canadian Journal of Botany 66:809-8 18. CookE, W. B. 1940. Flora of Mt. Shasta. American Mid- land Naturalist 23:497—5772. DALE, V. H. 1986. Plant recovery on the debris avalanche at Mount St. Helens. Pp. 208-214 in S. A. C. Keller (ed.), Mount St. Helens: five years later. Eastern Washington State University Press, Cheney, WA. . 1989. Wind dispersed seeds and plant recovery on the Mount St. Helens debris avalanche. Canadian Journal of Botany 67:1434—-1441. DEL Mora_, R. 1993. Mechanisms of primary succession on volcanoes: a view from Mount St. Helens. Pp. 79-— 100 in J. Miles and D. H. Walton (eds.), Primary succession on land. Blackwell Scientific Publications, London. AND L. C. BLIss. 1993. Mechanisms of primary succession: insights resulting from the eruption of Mount St. Helens. Advances in Ecological Research 24: 1-66. , J. H. Titus, AND A. M. Cook. 1995. Early pri- mary succession on Mount St. Helens, Washington, USA. Journal of Vegetation Science 6:107—120. AND D. M. Woop. 1986. Subalpine vegetation re- covery five years after the Mount St. Helens erup- tions. Pp. 215—221 in S. A. C. Keller (ed.), Mount St. Helens: five years later. Eastern Washington State University Press, Cheney, WA. AND D. M. Woop. 1988. The high elevation flora of Mount St. Helens. Madrofio 35:309-319. AND D. M. Woop. 1993a. Early primary succes- 1998] sion on a barren volcanic plain at Mount St. Helens, Washington. American Journal of Botany 80:981-— 991. AND D. M. Woop. 1993b. Early primary succes- sion on the volcano Mount St. Helens. Journal of Vegetation Science 4:223—234. Foxwor Tuy, B. L. AND M. HILL. 1982. Volcanic eruptions of 1980 at Mount St. Helens: the first 100 days. U.S. Geological Survey Professional Paper 1249. USS. Government Printing Office, Washington, D.C. FRANKLIN, J. EK AND C. T. DyRNEss. 1988. Natural Vege- tation in Oregon and Washington. Oregon State Uni- versity Press, Corvallis, OR. , J. A. MACMAHON, FE J. SWANSON, AND J. R. SE- DELL. 1985. Ecosystem response to the eruption of Mount St. Helens. National Geographic Research |: 198-216. FRENZEN, P. M., A. M. DELANO, AND C. M. CRISAFULLI. 1994. Mount St. Helens, biological research follow- ing the 1980 eruptions: an indexed bibliography and research abstracts (1980—1993). General Technical Report PNW-GTR-342. U.S. Department of Agricul- ture, Forest Service, Pacific Northwest Research Sta- tion, Portland, OR. GILLETT, G. W., J. T. HOWELL, AND H. LESCHKE. 1961. A flora of Lassen Volcanic National Park, California. Wasmann Journal of Biology 19:1—185. HALvorson, J. J., E. H. FRANZ, J. L. SMITH, AND R. A. BLAcK. 1992. Nitrogenase activity, nitrogen fixation, and nitrogen inputs by lupines at Mount St. Helens. Ecology 73:87-98. , J. L. SMITH, AND E. H. FRANZ. 1991. Lupine in- fluence on soil carbon, nitrogen and microbial activity in developing ecosystems at Mount St. Helens. Oec- ologia 87:162—170. HICKMAN, J. C. (ed.). 1993. The Jepson Manual: higher plants of California. University of California Press, Berkeley, CA. Hitcucock, C. L. AND A. CRONQUIST. 1973. Flora of the Pacific Northwest. University of Washington Press, Seattle, WA. IRELAND, O. L. 1968. Plants of the Three Sisters region, Oregon Cascade Range. Bulletin of the Museum of Natural History, University of Oregon 12:1—130. JONES, G. N. 1938. The flowering plants and ferns of Mt. Rainer. University of Washington Publications in Bi- ology 7:1—192. KRUCKEBERG, A. R. 1979. Mt. St. Helens field trip. Doug- lasia 3:9—10. . 1987. Plant life on Mount St. Helens before 1980. Pp. 3—23 in D. E. Bilderback (ed.), Mount St. Helens 1980. University of California Press, Berkeley, CA. LAWRENCE, D. G. 1938. Trees on the march: notes on the recent volcanic and vegetational history of Mount St. Helens. Mazama 20:49—54. 1954. Diagrammatic history of the northeast slope of Mt. St. Helens, Washington. Mazama 36:41— 44. TITUS ET AL.: MOUNT ST. HELENS CHECKLIST 161 LIPMAN, P. W. AND D. R. MULLINEAUX. 1981. The 1980 Eruptions of Mount St. Helens. U.S. Geological Sur- vey Professional Paper 1250. NuHN, W. W. 1987. Soil genesis on the 1980 pyroclastic flows of Mount St. Helens. Master’s thesis. University of Washington, Seattle, WA. OSWALD, V. H., D. W. SHOWERS, AND M. A. SHOWERS. 1995. A revised flora of Lassen Volcanic National Park. California Native Plant Society, Sacramento, CA. PIPER, C. V. 1906. Flora of the State of Washington. Con- tributions from the United States National Herbarium, Vol. 11. Bulletin of the United States National Mu- seum, Smithsonian Institution, Washington, D.C. PFITSCH, W. A. AND L. C. BLIss. 1988. Recovery of net primary production in subalpine meadows of Mount St. Helens following the 1980 eruption. Canadian Journal of Botany 66:989—997. REYNOLDS, G. D. AND L. C. BLIss. 1986. Microenviron- mental investigations of tephra covered surfaces at Mount St. Helens. Pp. 147—152 in S. A. C. Keller (ed.), Mount St. Helens: five years later. Eastern Washington State University Press, Cheney, WA. ST. JOHN, H. 1976. The flora of Mt. St. Helens, Washing- ton. The Mountaineer 70:65—77. AND E. HARDIN. 1929. Flora of Mt. Baker. Ma- zama 11:52—102. THOMPSON, L. R. AND G. L. WADE. 1991. Flora and veg- etation of 12 year-old coal surface-mined area in Rockcastle County, Kentucky. Castanea 56:99-116. Titus, J. H. 1995. The role of Mycorrhizae and Microsites in Primary Succession on Mount St. Helens. Ph.D. dissertation. University of Washington, Seattle, WA. TSUYUZAKI, S. 1995. Vegetation recovery patterns in early volcanic succession. Journal of Plant Research 108: 241-248. AND R. DEL MorAL. 1995. Species attributes in early primary succession on volcanoes. Journal of Vegetation Science 6:517—522. AND J. H. Titus. 1996. Vegetation development patterns in erosive areas on the Pumice Plains of Mount St. Helens. American Midland Naturalist 135: 172-177. M., J. H. Titus, S. TsuyYUZAKI AND R. DEL MORAL. 1998. Composition and dynamics of the wetland seedbank on Mount St. Helens, Washington, USA. Folia Grobotanica and Phytotaxonomica 33:3—16. Woop, D. M. AND R. DEL MorRAL. 1987. Mechanisms of early primary succession in subalpine habitats on Mount St. Helens. Ecology 68:780—790. AND R. DEL MorRAL. 1988. Colonizing plants on the Pumice Plains, Mount St. Helens, Washington. American Journal of Botany 75:1228—1237. AND W. FE Morris. 1990. Ecological constraints to seedling establishment on the Pumice Plains, Mount St. Helens, Washington. American Journal of Botany 77:1411-1418. WYND, E L. 1936. The flora of Crater Lake National Park. American Midland Naturalist 17:881—949. TU, MAbRONO, Vol. 45, No. 2, pp. 162-170, 1998 THE DISTRIBUTION OF VESICULAR-ARBUSCULAR MYCORRHIZAE ON MOUNT ST. HELENS, WASHINGTON JONATHAN H. Titus! Department of Botany, Box 355325, University of Washington, Seattle, WA 98195 ROGER DEL MORAL Department of Botany, Box 355325, University of Washington, Seattle, WA 98195 SHARMIN GAMIET Department of Botany, Box 355325, University of Washington, Seattle, WA 98195 ABSTRACT Vesicular-arbuscular mycorrhizae (VAM) occur in most terrestrial ecosystems and are crucial to un- derstanding community structure and function. However, their role in primary succession is poorly un- derstood. This study examined the distribution of VAM propagules, spores, and plants across the Pumice Plains of Mount St. Helens. VAM colonized plants and propagules were common in sites with thick vegetation, such as areas of relict pre-eruption vegetation and lupine patches, but were very infrequent in barren areas which comprise nearly all of the Pumice Plain. The vegetation of the Pumice Plain is composed primarily of facultatively mycotrophic species which are currently nonmycorrhizal. Mycorrhizal plants occur in refugia and thickly vegetated areas. VAM spore density and richness was low and spores are essentially restricted to densely vegetated habitats. The focus of this study is the distribution of VAM plants and VAM propagules across the Pum- ice Plain of Mount St. Helens, and their relationship with microsites. The relationship between VAM and microsites is of interest because microsites are crucial to the colonization dynamics on Mount St. Helens (del Moral and Bliss 1993, Titus 1995). During primary succession on volcanic sub- Strates, it is unlikely that pioneer species would de- pend on mycorrhizae (Allen 1991). Only non-host and facultatively mycotrophic species could invade these sites. Obligately mycotrophic species would be prevented from establishing until a population of VAM fungi was present in the soil, presumably having arisen in association with facultatively my- cotrophic species. Seral sequences may reflect the mycorrhizal dependence of the colonizing species (Allen 1991). Thus, the pattern of VAM distribu- tion across the primary successional landscape of the Pumice Plain may regulate plant invasion pat- terns (Allen 1988). However, previous to this study, the distribution of VAM propagules across the Pumice Plain was unknown. VAM propagules are composed of spores, hy- phae and VAM colonized roots. The two indices of VAM density, spore counts and degree of root col- onization, are not necessarily correlated (Louis and Lim 1987; Johnson et al. 1991). Spore counts as- sess only one type of propagule while colonization indirectly estimates all types of propagules. There- ' Present address: Jonathan H. Titus, Department of Bi- ological Sciences, University of Nevada, Las Vegas, Las Vegas, NV 89154-4004. fore, root colonization is a more accurate measure © of total VAM density. In this study, the distribution | of VAM has three facets: 1) The presence of VAM | fungal propagules in the soil, i.e., the mycorrhizal inoculum potential (MIP), which is determined through a root bioassay; 2) the presence of VAM | plants; and 3) the presence of VAM fungal spores in the soil. This study examines the distribution of these three components of VAM across the Pumice © Plain. METHODS Study site. The Pumice Plain was formed by the | 18 May, 1980 eruption of Mount St. Helens (46°12'N, 122°11'W). The Pumice Plain covers 20 , km? immediately north of the crater between 1150- | 1300 m elevation. It was formed by the deposit of up to 200 m of material from a debris avalanche, subsequent pyroclastic flows, air-fall pumice, and © was repeatedly impacted by later lahars. The Pum- | ice Plain is blanketed by pumice that ranges in depth from 10 to 200 m. The surface is flat or gent- — ly sloping, with numerous gullies created by ero- | sion dissecting the surface. Surface pumice parti- cles range in size from 1 mm to 10 cm (del Moral | and Bliss 1993). The climate is maritime, with cool wet winters and warm, dry summers. Periods of drought often | occur in July and August. Annual precipitation av- | erages 2373 mm, yet usually less than 5% of this — falls between June and August. The growing season begins in June and ends by early September. Tem- | peratures range from mean monthly minima of | —4.2°C in January and 7.3°C in August to maxima 1998] of 0.5°C in January and 22.2°C in July (Spirit Lake Ranger Station (987 m a.s.l.), Pacific Northwest River Basins Commission 1969). Summer temper- atures range from 0 to 35°C with a mean ca. 12°C. Surface soil temperatures are often very high in the summer approaching 50°C on tephra surfaces (Reynolds and Bliss 1986). Pumice Plain soil is immature, with very low concentrations of carbon, nitrogen, and microbial biomass (del Moral and Bliss 1993). Considerable variation in soil moisture values has been recorded between and within microsites. Substrates with fine particles contain more moisture than areas with coarse particles and erosion rills are more moist than other microsites (del Moral and Bliss 1993). In summer the surface tephra dries quickly between rains, thus, most Pumice Plain habitats do not re- main moist for periods sufficient to allow seedling establishment. However, the surface layer of tephra acts as a mulch to impede evaporation and is ca- pable of holding considerable moisture at lower depths so that adult plants rarely suffer from drought (Reynolds and Bliss 1986). Microsites. The seven types of microsites in this study appear to differ in environmental character- istics on the spatial scale of an individual seed or seedling. They were chosen because personal ob- servation and the literature both suggest them to be important to revegetation processes on the Pumice Plain. These sites are: Flat—sites which have homogeneous gravel, sand or silt substrates in which the topography is lev- el. Pumice particles are less than 5 cm in di- ameter. Flat sites occupy most of the Pumice Plain and are sparsely vegetated. Rill—small gullies formed by erosive water action. These are linear habitats that marginally protect seedlings from wind, collect more snow, and have lower solar radiation (del Moral and Bliss 1993). Rill edges are more stable than rill bot- toms and drainages. Near-rock—adjacent to rocks larger than 25 cm in diameter. On exposed surfaces rocks protect seedlings from direct solar exposure, reduce wind and surface temperatures, and are more | likely to trap seeds. Ridges—sites located on small ridgetops where | there is evidence of extensive wind erosion. Lupinus patch—sites associated with dense patches | of living and dead Lupinus lepidus Douglas. These sites contain higher levels of soil nitro- gen and lupines effectively trap seeds and or- ganic matter. Lupinus patches are described in Halvorson et al. (1992), del Moral and Bliss (1993), del Moral et al. (1995), Bishop (1996). Crowded vegetation—-sites located in thick vege- tation on new volcanically emplaced surfaces | which are not dominated by L. lepidus. _ Refugia—sites with pre-eruption soil exposed by erosion in which some belowground plant or- TITUS ET AL.: DISTRIBUTION OF VAM ON MOUNT ST. HELENS 163 gans survived and subsequently sprouted. Re- fugia are densely vegetated and are confined to the eastern fringe of the Pumice Plain on steep north facing slopes. Refugia vegetation is de- scribed by del Moral et al. (1995). These sites were investigated to determine the dis- tribution of VAM propagules and plants across the Pumice Plain. The first study looks at the distribu- tion of VAM propagules, the second at the distri- bution of VAM plants and the third at the distri- bution of VAM spores. Corn bioassay for assessing VAM propagule dis- tribution. Soil samples were collected at 20 repre- sentative locations within each site (except refugia) in July 1991. Four 250 g samples from the upper 8 cm of soil were collected at each location and combined to form two composite samples. Soil was sifted to remove all particles larger than 4 mm, amended with 20% sterile perlite to increase po- rosity, and 650 g was placed into 10 cm by 10 cm freely draining plastic pots. Bioassays were con- ducted with non-fungicide treated Zea mays seeds. All pots were watered daily with tap water. Fertil- izer was applied in 50 ml aliquots per pot of Col- well’s solution minus phosphorus at planting and at weekly intervals throughout the experiment. Col- well’s solution mimics natural proportions of nutri- ents in typical temperate soils (Colwell 1943; R. B. Walker personal communication). The control con- sisted of 20 pots of sterile greenhouse soil placed randomly among the treatment pots and planted with corn to determine if contamination by green- house VAM propagules occurred. Previous work showed that VAM propagules, if present, rapidly colonize corn in the greenhouse (Titus personal ob- servation). Pots were randomized and maintained at the University of Washington Botany Green- house at 20—25°C, and rotated every 10 days. Bio- assay plants were grown for 35 days from 20 July to 14 August 1991. Plants were harvested, roots washed, and frozen at — 18°C until October 1991 at which time roots were assayed for VAM coloni- zation. The quantity of inoculum in the soil, my- corrhizal inoculum potential (MIP), was estimated by percent colonization of corn roots (Moorman and Reeves 1979; Doerr et al. 1984; Johnson and McGraw 1988). Staining. Roots were washed, cleared and stained with trypan blue (Brundrett et al. 1994; E. Cazares, Oregon State University, personal communication). Percent colonization was estimated by placing a grid of 1 cm squares below a petri plate which con- tained the root sample under a dissecting micro- scope. One hundred locations where a root crossed a line on the grid were scored for mycorrhizae. Many samples were examined under higher power to ascertain that the fungus was indeed VAM. Root segments containing vesicles, arbuscles or intercel- lular hyphal coils or hyphae were recorded as being colonized. The number of mycorrhizal “‘hits”’ is an 164 estimate of the percent of the root colonized (Brun- drett et al. 1994). Mycorrhizal colonization of pioneer species. The roots of 14 plants of six major pioneer species were collected from each of the seven site types at dif- ferent locations across the Pumice Plain. Most root samples were collected during July, 1992, and the remainder during July, 1993. All sampled plants were at least 4 m from their nearest neighbor, ex- cept for those in lupine patches, crowded vegetation and refugia. Species sampled were Anaphalis mar- garitacea (L.) Benth. & Hook., Carex mertensii Prescott, Epilobium angustifolium L. ssp. circum- vagum Mosq., Hieracium albiflorum Hook., Hypo- chaeris radicata L. and Penstemon cardwellii. In addition to these target species, roots were collected from several other naturally occurring species where they occurred. Roots of these species were not collected in all microsite types because they only occurred in certain ones. Roots were washed when harvested and stored in alcohol until they were cleared and stained to assess VAM coloniza- tion as above. Spore isolation. Soil samples were collected us- ing two sampling regimes. For the first regime soil was collected from 20 representatives of each of the seven site types. Four 100 ml soil samples were collected from the top 8 cm of soil and combined to form a single sample during August 1993. The second sampling regime was part of a larger study. Percent cover of each plant species in 150 100 m? circular plots was assessed across the Pum- ice Plain during summer, 1993. Vegetation of these plots were grouped into five habitat types based on substrate and vegetation (del Moral et al. 1995). In conjunction with vegetation sampling, 100 ml of soil were collected at each of four locations within each plot and combined into a single sample. Soil samples were dried at room temperature and stored at 3°C in sealed plastic bags. Spores were extracted from two subsamples of the soil from each plot by the wet-sieving and decanting tech- nique (Gerdemann and Nicolson 1963; Pacioni 1992; Brundrett et al. 1994). One hundred and fifty ml of soil were placed into a 1.651 mm mesh sieve above 0.417 and 0.052 mm mesh sieves. The soil was washed vigorously with water. Roots in the top sieve and soil from the fine mesh bottom sieve were examined in a petri dish under a dissecting micro- scope at 40 power for VAM spores. In order to compare spore extraction techniques, the soil from ten samples with two replicates each were analyzed using both the wet-sieving with de- canting technique and the differential water/sucrose centrifugation method (Ianson and Allen 1986). Se- lected soil samples were those likely to contain VAM spores. Spore isolation efficiency was not 1m- proved using the differential water/sucrose tech- nique. Although Ianson and Allen (1986) found better spore isolation with the differential water/ MADRONO [Vol. 45 TABLE 1. VAM Corn BIOASSAY FOR THE DETERMINATION OF MYCORRHIZAL INOCULUM POTENTIAL (MIP) OF PUMICE PLAIN SorL. Soil MIP is shown by percent VAM coloni- zation of corn roots. % plants colonized is the percentage of the plants of each species which were colonized by VAM. (mean = standard deviation, n = 20 paired sam- ples). % plants Microsite MIP colonized Flat 0) 0) Near Rock 0) 0) Ridge 0) 0) Rill 0.3 + 0.6 15 Lupine Patch 3.0°S 3.3 60 Crowded Vegetation 4.3 + 3.0 70 sucrose technique, the extremely low organic mat- ter content of Pumice Plain soils obviated the need for improved resolution in the case of these soils. Spores were isolated and stored dry on filter pa- per. Spore types were determined from the experi- ence of the third author and the use of spore iden- tification guides (Mosse and Bowen 1968; Gerde- mann and Trappe 1974; Trappe 1982; Morton 1988; Schenck and Perez 1990). Data analysis. Mean percent mycorrhizal colo- nization was determined to yield MIP (Experiment I) and mean colonization (Experiment II). In Ex- periment III, spore density was averaged and rich- ness determined for each site and for each habitat type. The preponderance of zeros precluded statis- tical data analysis, so values are reported only as observational data. Although both parametric and non-parametric techniques are robust for violations of their respective assumptions, the statistical tech- niques appropriate for analysis of this experimental design (e.g., one-way ANOVA or the Kruskal-Wal- lace test (Zar 1984)) are invalid for the analysis of data with many zeros. Even non-parametric statis- tics require homoscedastic variances. This aside, the patterns in the data are clearly apparent without statistical tests. Frequency of VAM colonization or spores are also reported. RESULTS Corn bioassay for assessing VAM propagule dis- tribution. Ridge, flat, and near-rock substrates con- tained no detectable MIP. Rill microsites occasion- ally contained VAM propagules, whereas lupine patch and densely vegetated site soils usually con- tained mycorrhizal inoculum (Table 1). Mycorrhizal colonization of pioneer species. An- — aphalis margaritacea, Hieracium albiflorum and | Hypochaeris radicata were without mycorrhizal — colonization in flat, ridge and near-rock sites, but — were mycorrhizal in rill, lupine patch, crowded | vegetation and refugia (Table 2). Carex mertensit was without mycorrhizal colonization in all sites, 1998] TITUS ET AL.: DISTRIBUTION OF VAM ON MOUNT ST. HELENS 165 TABLE 2. VAM COLONIZATION OF PLANT SPECIES COLLECTED FROM MICROSITES ON THE PUMICE PLAIN. (mean + standard deviation, n = 14). Flat, ridge and near rock microsites contained no VAM plants and are not shown. Microsite Rill Lupine patch Crowded vegetation Refugia % % % % plants plants plants plants % VAM _ colo- % VAM colo- % VAM colo- % VAM colo- Species colonization nized colonization nized colonization nized _ colonization nized Anaphalis margaritacea LA 26 36 3.3 2°15;3 ad 10.2 2 1226 86 6.4 + 7.8 63 Carex mertensii O 0) 0 0 0) O 0.1 = 0.4 14 Epilobium angustifolium 0) 0 2.0 + 5.6 29 0.4 + 1.6 14 4.2 + 6.6 36 Hieracium albiflorum 0.2 + 0.8 14 4.9 + 9.3 64 io nea LORS 64 8.9 + 10.5 50 Hypochaeris radicata ODE 25 29 8.1 + 11.8 64 $.0°2 155 33 FO. 25 Sad. 50 Penstemon cardwellii 0.8 + 2.8 7 3.2 + 4.2 43 7.0 + 7.8 79 15.7 + 9.0 q9 except for a trace of VAM in refugia. Epilobium DISCUSSION angustifolium was not colonized in rill microsites, but was occasionally colonized in lupine patch, crowded vegetation and refugia. The species with the highest mycorrhizal colonization was A. mar- garitacea, and the site with the most VAM plants was crowded vegetation. Most of the VAM fungal hyphae observed were of the fine endophyte type. Non-target species were all nonmycorrhizal in flat, ridge and near-rock sites (Table 3). Only Juncus parryi Engelm. was mycorrhizal in rill microsites. Juncus parryi and Lupinus lepidus were mycorrhi- zal in lupine patches. Many species were mycor- rhizal in crowded sites and refugia, and species re- stricted to refugia were usually mycorrhizal. Spore distribution. No spores were found in flat, rill, or near-rock sites. Densities were variable where spores were found in dead lupine, crowded vegetation, and refugia (Table 4). Dead lupine and crowded vegetation microsites with VAM spores often were located far from refugia. Most refugia samples contained many spores. Pumice barrens, pyroclastic surfaces and drain- ages (del Moral et al. 1995) rarely contained VAM fungal spores (Table 5). Samples containing spores ' were usually located near refugia. The only excep- tion was an isolated barren pumice site which also _ contained a large willow. Lupine patches occasion- ally contained spores which, when present, were in ' high densities. Lupine patches which contained | Spores were widely spread across the Pumice Plain. Refugia almost always contained spores. Three spore types were found: Glomus macro- _carpum (Tul. and Tul.), Glomus mosseae (Nicol. and Gerd.) Gerd. and Trappe, and an Acaulospora _type. The most common spores found were dead (which are empty), dark brown, brassy, or black. _There was usually only one spore type present in a Sample (not including dead spores which were usu- ally present), however, all three spore types oc- curred in microsites and habitat types where spores _ Were common. VAM distribution. Based on observational data only, microsites differ in MIP and spore density, and pioneer species differ in mycorrhizal coloni- zation depending on the microsite it inhabits. Sites with thicker vegetation contained more VAM pro- pagules and VAM plants. After the 1980 eruption, the Pumice Plain was free of VAM fungal propagules (Allen 1987). The VAM propagules detected in this study show that dispersal forces, most likely animals (Allen 1987), have been returning VAM propagules to this land- scape. The invasion of VAM propagules is sporadic aS some microsites contain more mycorrhizal pro- pagules and more heavily colonized plants than do other microsites. This supports the patch-dynamic model which proposes that the pattern of VAM fun- gal propagules dispersed by animals searching among patches for food and cover in sparsely veg- etated landscapes creates a patchy distribution of inoculum (Allen 1987, 1988). This patchiness may also result from the ability of certain microsites to effectively trap windborne or waterborne propagu- les. VAM spores were uncommon across the Pumice Plain, but they are present in crowded and lupine patch microsites. The bulk of the Pumice Plain ap- pears to remain VAM spore free. Evidence for a non-patchy landscape level spread of VAM fungi was observed in the plots adjacent to refugia in which VAM propagules were found. Since refugia are on steep slopes, VAM propagules could immi- grate to adjacent barren and drainage habitats by erosion. Plant diversity is also slightly higher in ar- eas adjacent to refugia (del Moral et al. 1995). However, one pumice barren plot distant from sites with high levels of VAM spores contained VAM spores. This is evidence for a patchy distribution of VAM spores and adds support to the patch-dynamic model (Allen 1988). This pumice barren site con- tained a large willow which is probably a locus for small mammal activity in a barren landscape. In 166 MADRONO [Vol. 45 TABLE 3. WAM COLONIZATION OF PLANT SPECIES COLLECTED FROM MICROSITES ON THE PUMICE PLAIN. n = sample size. (mean + standard deviation). Microsite Flat Ridge Near rock % % % % J VAM plants VAM plants J% VAM plants coloni- colo- coloni- colo- coloniza- colo- Species n zation nized n zation nized n tion nized Achillea millefolium Agrostis pallens 4 0 0) 2 0 0 4 0) 0) Agrostis scabra 7 23-64 14 2 0) 0) 7 0) 0) Blechnum spicant Calyptridium umbellatum 4 0) 0) 3 O 0 Carex pachystachya 3 0) 0 3 0 0 3 0 0 Carex phaeocephala 3 0 0 3 0) 0 3 0) 0) Cirsium arvense 2 0) 0) Epilobium anagallidifolium 5) 0 0) 2 0) 0) 4 0) 8) Epilobium brachycarpum Epilobium ciliatum Eriogonum pyrolifolium Fragaria virginiana Gnaphalium uliginosum Juncus mertensianus Juncus parryi 10 0 0) 2 0) 0) Luetkea pectinata Lupinus latifolius 4 0) 0) 4 0) 0 Lupinus lepidus 17 0) 0 6 0 0 16 0 0 Luzula parviflora Phacelia hastata Polygonum minimum Ribes laxiflorum Rubus lasiococcus Rubus spectabilis Saxifraga ferruginea 5 0 0) 4 0) 0 Sambucus racemosa Senecio sylvaticus 8 0) 0 4 0) 0) 6 O:7 +t 1.6 17, Smilicina racemosa Spergularia rubra 2 0) 0 0 0) 0) 2 0) 0) Vaccinium membranaceum Vancouveria hexandra addition, several of the crowded vegetation and lu- pine patch microsites which contained VAM spores were isolated across the Pumice Plain far from re- fugia, adding further support to the patch-dynamic model. The differences in spore counts for lupine patch and refugia between Tables 4 and 5 were not unexpected due to the large standard deviations and patchy nature of spore distribution (Anderson et al. 1983; St. John and Koske 1988). Mycotrophic Status of Colonizing Species Glomus tenuis. Glomus tenuis (Greenall) Hall is distinguished by hyphal diameters in the 0.5-1.5 wm range (Hall 1987). Other Glomus species have coarse (5—30 ym in diameter) hyphae (McGonigle and Fitter 1990; Wang et al. 1993). Therefore, only root colonizations caused by G. tenuis can be iden- tified confidently in the absence of sporulating structures (Carling and Brown 1982; Hall 1987). Colonization by G. tenuis has been found to be highest in dry very low phosphorus environments (Rabatin 1979), low pH soils (Wang et al. 1993), and in alpine environments (Read and Haselwand- ter 1981; Mullen and Schmidt 1993). Glomus tenuis is also often the dominant VAM fungal species in pioneer species and disturbed environments (Daft and Nicolson 1974). In this study, fine endophyte hyphae were observed frequently, but no spores of G. tenuis were detected. Glomus tenuis spores may be too small (7-12 wm) to be extracted by the wet sieving technique (Hall 1987; Wang et al. 1993). Thus the possibility exists that these spores are common but were not detected. Although spores are the only way to identify Glomus species, they are not indicative of the actual infectivity of a soil and should be used only in conjunction with other indices. For example, no spores were detected in rill microsites, but there was some VAM coloni- zation of corn roots and the target species in rills were occasionally VAM. 1998] TITUS ET AL.: DISTRIBUTION OF VAM ON MOUNT ST. HELENS 167 TABLE 3. EXTENDED Microsite Rill Lupine patch Crowded vegetation % VAM % plants % VAM % plants % VAM % plants n colonization colonized n colonization colonized n colonization colonnized 2 6 100 4 0) 0) 4 0 0) 12 8.0 + 21.8 phe, 7 O 0) 7 O 0 6 0.4 + 0.9 17 2 O 0 1 6) 0 3 0) 0 3 0) 0) 0) O 3 0) 0) 3 0 0 0) 0) 3 O 0) 2 0) 0) 3 5.7 + 4.0 100 4 0 0 2 0 O 2 0 0 2 15.0 + 7.1 100 2 0) 6) 1 6) 0) 6 6) 0 11 OS 15 9 8 9.0 + 7.0 88 16 0.8 + 2.6 13 1 0 0) 5 0 0 16 6) 11 19 0.2 + 0.6 11 6 0.7 + 1.6 Tq 14 6) 6) 2 AO 23.7 50 2 6) 0 5 0 0) 1 6) 0 2 102+ 1.4 50 4 0 0 5 0 O 4 LS = 19 50 2 6) 6) 6) 0) 6) Carex spp. are considered to be non-hosts (Pow- ell 1975; Anderson et al. 1984), although mycor- rhizal Carex spp. have been found in the alpine (Read and Haselwandter 1981; Allen et al. 1987) and in grasslands (Read et al. 1976). VAM Carex mertensii plants were only observed in two indi- viduals in this study. Juncus parryi is generally thought to be in a non-host genus (Powell 1975). However, in this case it was heavily colonized by VAM in rill, lupine patch and crowded sites. Gen- era-wide generalizations of mycorrhizal depen- dence may be inaccurate and extensive examina- tions of the species must be conducted to ascertain mycotrophy (Read et al. 1976; Newman and Red- dell 1987). Annuals are often considered to be non-hosts (Trappe 1987; Boerner 1992; Peat and Fitter 1993), but in this survey Senecio sylvaticus was frequently mycorrhizal. Allen et al. (1992) found the annual Epilobium paniculatum to be mycorrhizal in a lu- pine patch. In this survey the species was found to be nonmycorrhizal. Allen et al. (1992) found Lupinus latifolius J. Agardh. and L. lepidus to be mycorrhizal, while this survey found L. lepidus, but not L. latifolius, to be mycorrhizal. Avio et al. (1990) observed Lupinus to be a strongly non-host genus. O’ Dell and Trappe (1992) found both L. lepidus and L. latifolius to be occasionally mycorrhizal. They located a mycor- rhizal L. latifolius on Mount St. Helens but did not find a mycorrhizal L. lepidus on the volcano. O’ Dell and Trappe (1992) suggested that VAM fun- gi may need to be established on a companion host before colonizing roots of lupines. Most plant species now colonizing Mount St. Helens barren sites appear to be facultatively my- cotrophic (Titus 1995). This status supports a broad range of tolerance to VAM, from rarely mycorrhi- zal to nearly always colonized depending upon the species, neighboring species and site conditions (Allen 1991; Boerner 1992). VAM fungal species. VAM fungal richness was low, with only three spore types, but greater than the single species (Glomus macrocarpum) found in the blast zone by Allen et al. (1984), Allen and 168 TABLE 3. CONTINUED Microsite Refugia % plants % VAM colo- Species n colonization nized Achillea millefolium Agrostis pallens Agrostis scabra Blechnum spicant Calyptridium umbellatum Carex pachystachya Carex phaeocephala Cirsium arvense Epilobium anagallidifolium Epilobium brachycarpum Epilobium ciliatum Eriogonum pyrolifolium Fragaria virginiana Fe) Gnaphalium uliginosum Juncus mertensianus Juncus parryi 0 O:5°2 15 9 Luetkea pectinata 4 0) 0 Lupinus latifolius Lupinus lepidus Luzula parviflora Phacelia hastata Polygonum minimum Ribes laxiflorum Rubus lasiococcus Rubus spectabilis Saxifraga ferruginea Sambucus racemosa Senecio sylvaticus Smilicina racemosa Spergularia rubra Vaccinium membranaceum Vancouveria hexandra N 50°20 100 15.0 = 91 RNA D297 12.8 2.8 14.1 Kye Seat: So aa sane Pe UA N oh S So {+ co) MacMahon (1988), and Allen et al. (1992). This indicates that VAM fungal species are invading the blast zone or at least proliferating into detectable densities. The preponderance of inviable spores found in this study is not unusual (Read et al. 1976; MADRONO [Vol. 45 TABLE 4. NUMBER AND RICHNESS OF VAM _ FUNGAL SPORES IN 150 ML SOIL SAMPLES FROM MICROSITES ON THE PUMICE PLAIN. (mean + standard deviation for spore counts, n = 20 for each microsite type). ' Mean richness is based only on samples which contained spores. % sam- ples Mean Mean number’ with _ rich- Microsite of spores spores ness! Flat 0 0) — Near Rock 0 0) — Ridge 0) 0) — Rill O 0 —- Lupine Patch 13:6-= 29:2 55 1.4 Crowded Vegetation 18.4 + 41.1 70 1.3 Refugia 20.7 + 49.6 85 1.8 Berliner and Torrey 1989). The patchy nature of VAM species distribution is evidenced by the large variance in spore densities and by the presence of different spore types in different sites with little overlap. However, each species was present in the microsites and habitat types which had detectable spore populations. It is important to note the dif- ference in sampling intensity between above- and belowground environments. Plot size in del Moral et al. (1995) was 100 m/?, where as the surface area of the belowground sampling effort was only ap- proximately 400 cm’, which is 0.0004 as large as the aboveground sampling area. Therefore, state- ments about patchy spore distributions must be re- garded in the light of the small belowground sam- pling area (Anderson et al. 1983). In the few studies which address VAM species distribution, richness is usually low and density variable. It is therefore difficult to draw conclusions about successional patterns in VAM fungal types from the results pre- sented here. CONCLUSION This study assessed both VAM colonization and VAM fungal propagules. The results are comple- TABLE 5. NUMBER AND RICHNESS OF VAM FUNGAL SPORES IN 150 ML SormL SAMPLES FROM HABITAT TYPES ON THE PUMICE PLAIN. Habitat types based on del Moral et al. (1995). ‘‘Near’’ indicates a site adjacent to a refugia, “‘far a9 indicates a site distant from a refugia. n = sample size. (mean + standard deviation for spore counts). ' Mean richness is based only on samples which contained spores. Habitat type n Pumice Barrens—near 11 Pumice Barrens—far a2 Pyroclastic Surfaces 15 Drainages—near 4 Drainages—far 15 Wetlands 23 Lupine Patches 16 Refugia 26 Mean number % samples Mean of spores with spores richness! 15e22" 320) 4 1 0.03 + 0.2 3 1 0) 0 sis 0.3 + 0.5 25 1 0) ) ae O 0) — 10S = 26.2 25 L5 [4.522 3351 62 he] 1998] mentary and converge to the conclusion that the Pumice Plain remains essentially VAM free, except for the few isolated lupine patch and crowded sites. Refugia contain VAM fungal propagules and my- corrhizal plants. The sparse vegetation of the Pum- ice Plain is composed largely of facultatively my- cotrophic species which are at present nonmycor- rhizal. ACKNOWLEDGMENTS Thanks to Doug Ewing, Elaine Bassett, and Jeanette Milne for attending to the plants in the Greenhouse and for good advice. Extensive lab work was contributed by JoEllen VanDeMark. Thanks to Joseph Ammirati, Shan- non Berch, Thomas Odell, and David Ianson for advice. Thanks to Priscilla Titus for editing the manuscript and for her generous encouragement and support. NSF Grants BSR-89-06544 and DEB-9406987 to Roger del Moral helped support this research. LITERATURE CITED ALLEN, E. B., J. C. CHAMBERS, K. E CONNER, M. F. ALLEN, AND R. W. BRown. 1987. Natural reestablishment of mycorrhizae in disturbed alpine ecosystems. Arctic and Alpine Research 19:11—20. ALLEN, M. FE 1987. Re-establishment of mycorrhizae on Mount St. Helens: migration vectors. Transactions of the British Mycological Society 8:413—417. . 1988. Re-establishment of VA mycorrhizae fol- lowing severe disturbance: comparative patch dynam- ics of a shrub desert and subalpine volcano. Proceed- ings of the Royal Society of Edinburgh 94B:63-—71. . 1991. The Ecology of Mycorrhizae. Cambridge University Press, New York. , C. CRISAFULLI, C. E FRIESE, AND S. L. JEAKINS. 1992. Re-formation of mycorrhizal symbioses on Mount St. Helens, 1980—1990: interactions of rodents and mycorrhizal fungi. Mycological Research 96: 447-453. AND J. A. MACMAHON. 1988. Direct VA mycor- rhizal inoculation of colonizing plants by pocket go- phers (Thomomys talpoides) on Mount St. Helens. Mycologia 80:754—756. , J. A. MACMAHON, AND D. C. ANDERSEN. 1984. Reestablishment of Endogonaceae on Mount St. Hel- ens: survival of residuals. Mycologia 76:1031—1038. ANDERSON, R. C., A. E. LIBERTA, AND L. A. DICKMAN. 1984. Interactions of vascular plants and vesicular- arbuscular mycorrhizal fungi across a soil moisture- nutrient gradient. Oecologia 64:111—117. , A. E. LIBERTA, L. A. DICKMAN, AND J. A. KATZ. 1983. Spatial variation in vesicular-arbuscular my- corrhiza spore density. Bulletin of the Torrey Botan- ical Club 110:519—525. Avio, L., C. SBRANA AND M. GIOVANNETTI. 1990. The re- sponse of different species of Lupinus to VAM en- dophytes. Symbiosis 9:321—323. BERLINER, R. AND J. G. TORREY. 1989. Studies on mycor- rhizal associations in Harvard Forest, Massachusetts. Canadian Journal of Botany 67:2245—2251. BisHop, J. G. 1996. Demographic and population genetic variation during colonization by the herb Lupinus lep- idus on Mount St. Helens. Ph.D. dissertation. Uni- versity of Washington, Seattle, WA. BOERNER, R. E. J. 1992. Plant life span and response to TITUS ET AL.: DISTRIBUTION OF VAM ON MOUNT ST. HELENS 169 inoculation with vesicular-arbuscular mycorrhizae. I. Annual versus perennial grasses. Mycorrhiza 1:153— 161. BRUNDRETT, M. C., L. MELVILLE, AND L. PETERSON. 1994. Practical methods in mycorrhiza research. Mycologue Publications, Guelph, Ontario. CARLING, D. E. AND M. FE Brown. 1982. Anatomy and physiology of vesicular-arbuscular and nonmycorrhi- zal roots. Phytopatology 72:1108—1114. COLWELL, W. E. 1943. A biological method for determin- ing the relative boron content of soils. Soil Sciences 56:71—94. Dart, M. J. AND T. H. NICOLSON. 1974. Arbuscular my- corrhizas in plants colonizing coal wastes in Scotland. New Phytologist 73:1129—1138. DEL MorAL, R. AND L. C. BLIiss. 1993. Mechanisms of primary succession: insights resulting from the erup- tion of Mount St. Helens. Advances in Ecological Re- search 24:1—66. DEL Mora., J. H. Titus, AND A. M. Cook. 1995. Early primary succession on Mount St. Helens, Washing- ton, USA. Journal of Vegetation Science 6:107—120. Doerr, T. B., E. F REDENTE, AND FE B. REEVES. 1984. Effects of soil disturbance on plant succession and levels of mycorrhizal fungi in a sagebrush-grassland community. Journal of Range Management 37:135— 139. GERDEMANN, J. W. AND T. H. NICOLSON. 1963. Spores of mycorrhizal Endogone species extracted from soil by wet sieving and decanting. Transactions of the British Mycological Society 46:235—244. GERDEMANN, J. W. AND J. M. TRAPPE. 1974. Endogonaceae of the Pacific Northwest. Mycologia Memoir 5:1-—76. HALL, I. R. 1987. Taxonomy and identification of vesic- ular-arbuscular mycorrhizal fungi. Angewandte Bo- tanik 61:145—152. HALvorson, J. J., E. H. FRANZ, J. L. SMITH, AND R. A. BLACK. 1992. Nitrogenase activity, nitrogen fixation, and nitrogen inputs by lupines at Mount St. Helens. Ecology 73:87—98. IANSON, D. C. AND M. FE ALLEN. 1986. The effects of soil texture on extraction of vesicular-arbuscular mycor- rhizal fungal spores from arid sites. Mycologia 78: 164-168. JOHNSON, N. C. AND A. McGraw. 1988. Vesicular-arbus- cular mycorrhizae in taconite tailings. I. Incidence and spread of Endogonaceous fungi following recla- mation. Agriculture, Ecosystems and Environment 21:135-142. , FE L. PFLEGER, R. K. CROOKSTON, S. R. SIMMONS, AND P. J. COPELAND. 1991. Vesicular-arbuscular my- corrhizae respond to corn and soybean cropping his- tory. New Phytologist 117:657—663. Louis, I. AND G. Lim. 1987. Spore density and root colo- nization of vesicular-arbuscular mycorrhizas in trop- ical soil. Transactions of British Mycological Society 88:207-212. MCGONIGLE, T. P. AND A. H. FITTER. 1990. Ecological specificity of vesicular-arbuscular mycorrhizal asso- ciations. Mycological Research 94:120—122. MoorMan, T. AND FE B. REEVES. 1979. The role of endo- mycorrhizae in revegetation practices in the semi-arid west. II. A bioassay to determine the effect of land disturbance on endomycorrhizal populations. Ameri- can Journal of Botany 66:14—18. Morton, J. B. 1988. Taxonomy of VA mycorrhizal fungi: classification, nomenclature, and identification. My- cotaxon 32:267—324. 170 MossE, B. AND G. D. BOWEN. 1968. A key to the recog- nition of some Endogone spore types. Transactions of the British Mycological Society 51:469—483. MULLEN, R. B. AND S. K. SCHMIDT. 1993. Mycorrhizal infection, phosphorus uptake, and phenology in Ra- nunculus adoneus: implications for the functioning of mycorrhizae in alpine systems. Oecologia 94:229— 234. NEWMAN, E. I. AND P. REDDELL. 1987. The distribution of mycorrhizas among families of vascular plants. New Phytologists 106:745—751. O’DELL, T. E. AND J. M. TRAPPE. 1992. Root endophytes of lupin and some other legumes in Northwestern USA. New Phytologist 122:479—485. PAcIFIC NORTHWEST RIVER BASINS COMMISSION. 1969. Cli- matological Handbook Columbia Basin States, Vol. 1, Part A and Vol. 2. Meterology Committee, Pacific Northwest River Basins Commission, Portland, OR. PACIONI, G. 1992. Wet-sieving and decanting techniques for the extraction of spores of vesicular-arbuscular fungi. Pp. 317-322 in J. R. Norris, D. J. Read, and A. K. Varma (eds.), Methods in Microbiology, Vol. 24. Academic Press, London. PeaT, H. J. AND A. H. Fitter. 1993. The distribution of arbuscular mycorrhizas in the British flora. New Phy- tologist 845-854. POWELL, C. L. 1975. Rushes and Sedges are non-myco- trophic. Plant and Soil 42:481—484. RABATIN, S. C. 1979. Seasonal and edaphic variation in vesicular-arbuscular mycorrhizal infection of grasses by Glomus tenuis. New Phytologist 83:95—102. READ, D. J. AND K. HASELWANDTER. 1981. Observations on the mycorrhizal status of some alpine plant com- munities. New Phytologist 88:341—352. MADRONO [Vol. 45 , H. K. KOUCHEKI, AND J. HODGSON. 1976. Vesic- ular-arbuscular mycorrhiza in natural vegetation sys- tems. I. The occurrence of infection. New Phytologist 76:641—653. REYNOLDs, G. D. AND L. C. BLIss. 1986. Microenviron- mental investigations of tephra covered surfaces at Mount St. Helens. Pp. 147-152 in S. A. C. Keller (ed.), Mount St. Helens: five years later. Eastern Washington State University Press, Cheney, WA. ST. JOHN, T. V. AND R. E. KoskE. 1988. Statistical treat- ment of Endogonaceous spore counts. Transactions of the British Mycological Society 91:117—121. SCHENCK, N. C. AND Y. PEREZ. 1990. Manual for the iden- tification of VA mycorrhizal fungi, 3rd ed. Synergis- tic Publications, Gainesville, FL. Titus, J. H. 1995. The role of mycorrhizae and microsites in primary succession on Mount St. Helens. Ph.D. dissertation. Department of Botany, University of Washington, Seattle, WA. TRAPPE, J. M. 1982. Synoptic Keys to the genera and spe- cies of zygomycetous mycorrhizal fungi. Phytopa- thology 72:1100—1108. . 1987. Phylogenetic and ecologic aspects of my- cotrophy in the angiosperms from an evolutionary standpoint. Pp. 2—25 in G. R. Safir (ed.), Ecophysi- ology of VA mycorrhizal plants. CRC Press, Boca Raton, FL. WANG, G. M., D. P. STRIBLEY, P. B. TINKER, AND C. WALK- ER. 1993. Effects of pH on arbuscular mycorrhiza I. Field observations on the long-term liming experi- ments at Rothamsted and Woburn. New Phytologist 124:465—472. ZAR, J. H. 1984. Biostatistical analysis, 2nd ed. Prentice- Hall, Inc., Englewood Cliffs, New Jersey. MapRONO, Vol. 45, No. 2, pp. 171-182, 1998 LATE-HOLOCENE VEGETATION CHANGES FROM THE LAS FLORES CREEK COASTAL LOWLANDS, SAN DIEGO COUNTY, CALIFORNIA R. SCOTT ANDERSON Center for Environmental Sciences & Education, and Quaternary Studies Program, Box 5694, Northern Arizona University, Flagstaff, AZ 86011 BRIAN F ByrD ASM Affiliates, Inc., 543 Encinitas Blvd., Suite 114, Encinitas, CA 92024 and Department of Anthropology, University of California, San Diego, CA 92093 ABSTRACT The vegetation history of coastal southern California is incompletely known, due primarily to a lack of suitable sites for preservation of subfossil organic remains, and destruction of sites by recent human impact. One coastal site, along Las Flores Creek in San Diego County, has yielded a pollen record, enabling us to reconstruct vegetation and climatic changes for the last ca. 4300 radiocarbon years. Pollen from riparian plants, including Typha (cattail) and Cyperaceae (sedges), are common between ca. 4000 and 2600 years ago. Cupressus, or a closely related tree, may have grown along this riparian corridor, suggesting formerly larger range for this plant. By ca. 2600 years ago, a vegetation mosaic including elements of the coastal sage, chaparral, and grassland communities, were established near the site, per- sisting until the introduction of exotic weed and tree species within the last century. The pollen record from Las Flores Creek provides information on the environment of the Native American population during the Late Holocene. Local plant species undoubtedly provided resources for seasonal occupation of the coast, which supplemented extensive shellfish collecting, especially after ca. 2000 years ago. An increase in pollen from the Chenopodiaceae at that time may correspond to local human-caused disturbances. Analysis of pollen assemblages from stratigraph- ic profiles in southern California has lagged signif- icantly behind studies in other parts of the state (Adam 1985). The primary reasons for this include a perceived lack of suitable sites, as well as gen- erally poor pollen preservation from sites that have been analysed. Because traditional sites for pollen profiles (i.e., lakes or bogs) in this arid region are generally missing at low elevations in southern Cal- ifornia, alternative sites, such as alluvial sections or estuaries, must be investigated in order to provide information on local vegetation changes for the coastal region. Pollen recovery from alluvial and colluvial sec- tions at locations in California and the American Southwest has been highly to moderately success- ful. For example, Anderson and Smith (1994) and Koehler and Anderson (1994) reported on montane alluvial/colluvial profiles at middle elevations in the Sierra Nevada. Pollen preservation was excellent in these sections, which occurred as low as 1509 m. Holocene pollen profiles have been investigated as well from several coastal estuary sites in south- ern California, including Los Penasquitos Lagoon and Mission Bay in San Diego (Mudie and Byrne 1980), San Joaquin Marsh in Orange County (Davis 1992), and Abalone Rocks Marsh on Santa Rosa Island (Cole and Liu 1994). Pollen deposited in es- tuarine sediments contains assemblages from ter- restrial, freshwater, and marginal marine plants, and has been used to infer changes in sealevel as well as the terrestrial record from the adjoining uplands. A late Pleistocene pollen record was also obtained from a rockshelter on San Miguel Island (West and Erlandson 1994). In January 1994, we discovered a thick alluvial section near the mouth of Las Flores Creek, on Camp Pendleton Marine Base, San Diego County. Our investigations were part of ongoing archaeo- logical excavations on the Base (Byrd 1996). Pollen preserved in this 4.75 m section has allowed us to reconstruct the vegetation history of the upland and riparian habitats near the shoreface, in contrast to those from the sub- and near-tidal marshes of south- ern San Diego, Orange, and Ventura Counties named above. The data from the Las Flores Arroyo not only provide a Late Holocene record from a region with inadequate paleoenvironmental cover- age, but also provide an environmental context use- ful in archaeological reconstructions of the region. Location and Characteristics of Site The Las Flores Arroyo site is located along Las Flores Creek, at the mouth of Las Pulgas Canyon, on Camp Pendleton Marine Base (Las Pulgas Can- yon USGS 7.5’ Quad; 33°17'30"N latitude; 117°27'30"W longitude; Fig. 1). The top of the sec- tion occurs at an elevation of ca. 6 m. The deposit is a deeply incised Holocene alluvial terrace, with stacked silts and sands comprising both alluvial de- posits and paleosols (Fig. 2A, personal observa- tions, and Waters 1996). 12 MADRONO PACIFIC OCEAN ST renee _ J L wence So 121° 120° 119° 118° 117° 116° 115° 114° Y an Luis G Obis; Kern oan 7 anta Bernardino Barbara |Ventura Los 05 Angeles . QJ Orange” Pendleton 2 Fic. 1. 5] Oceanside Harbor ‘an [Vol. 45 Location of the Las Flores pollen site, San Diego County, California, as well as other nearby sites mentioned in the text. | = Las Flores profile (this paper); 2 = Los Penasquitos Lagoon and Mission Bay (Mudie and Byrne 1980); 3 = San Joaquin Marsh (Davis 1992); 4 = Abalone Rocks Marsh (Cole and Liu 1994); 5 = Santa Barbara Basin (Huesser 1978; Byrne et al. 1977). The vegetation of Camp Pendleton and the ad- jacent coastal uplands was mapped by Pacific Southwest Biological Services, Inc. (1986); Zedler et al. (unpublished) has studied the community composition of the Base. At least four distinct plant communities occur within 500 m of the Las Flores site (Fig. 2B). Community composition and vege- tation distribution on the Base is heavily influenced by past land use activities, including farming and grazing, fire, construction activities, and military training (Zedler et al. unpublished). The arroyo itself presently supports Cottonwood/ Willow Riparian Woodland. Common tree species in the riparian zone include Populus fremontii S. Watson ssp. fremontii, P. balsamifera L. ssp. tri- chocarpa (Torrey & A. Gray) Brayshaw, Salix gooddingii C. Ball, S. lasiolepis Benth. and S. lae- vigata Bebb. Typical associates include Platanus racemosa Nutt. (California sycamore), as well as Artemisia douglasianna Besser (mugwort), Bac- charis_ salicifolia (Ruiz Lopez & Pavon) Pers. (mule-fat), Conium maculatum L. (poison hem- lock), Xanthium strumarium L. (cockle burr), Ur- tica dioica L. (stinging nettle) and Vitis girdiana Munson (wild grape) (Zedler et al. unpublished; Beauchamp 1986). The alluvial deposit itself is covered by Southern Willow Scrub. This vegetation unit includes the willows mentioned above, plus Salix exigua Nutt. Associated herbaceous species include Toxicodendron diversilobum (Torrey & A. Gray) E. Greene (poison oak), Ambrosia psilos- tachya DC. (ragweed), U. dioica, X. strumarium, Artemisia douglasianna, and the non-natives of C. maculatum, and Foeniculum vulgare Miller (sweet fennel). On the headland immediately south of the arroyo is found a small patch of Diegan Coastal Sage Scrub, dominated by low (<2 m), soft-leaved, drought-deciduous shrubs (Zedler et al. unpub- lished). Plant composition is variable across gradi- ents of exposure, elevation and soil type. However, dominant species include Artemisia californica Less. (California sagebrush), Salvia mellifera E. Greene (black sage), Malosma laurina (Nutt.) Abrams (laurel sumac), Baccharis pilularis DC (coyote brush), Lotus scoparius (Nutt.) Ottley (deerweed) and Eriogonum fasciculatum Benth. (California buckwheat). Chaparral species poten- tially occurring within this community include Ce- anothus crassifolius Torrey (hoaryleaf ceanothus), Rhamnus ilicifolia Kellogg (holly-leaf redberry), Rhus integrifolia (Nutt.) Brewer & S. Watson (le- monadeberry), Heteromeles arbutifolia (Lindley) Roemer (toyon), and Cercocarpus betuloides Tor- rey & A. Gray (mountain-mahogany). Non-Native Annual Grassland presently occupies a large area south of Las Flores. Common intro- duced grasses include Bromus diandrus Roth, B. madritensis L. ssp. rubens (L.) Husnot, B. hordea- ceus (ripgut, foxtail chess and soft chess), Avena barbata Link, A. fatua L. (slender and wild oat), Hordeum spp. (wild barley), and Lolium multiflo- rum Lam. (Italian ryegrass). METHODS The Las Flores Arroyo wall face was sampled on 5 May 1994. Using an extension ladder to reach the ANDERSON AND BYRD: LAS FLORES CREEK POLLEN ' } - —) ae eatin te mpeg ae Fic. 2. A. Exposed section of the Las Flores Creek alluvial deposit. The man on the ladder is ca. 2 m tall for scale. B. Vegetation communities in the vicinity of Las Flores Creek. Most of the foreground and middle of the photo is coastal sage scrub and introduced species. Beyond Interstate-5 is coastal grassland; chaparral and oak woodland [En- gelmann (Quercus englemanni) and coast live (Q. agrifolia) oaks] blanket the hills in the background. Sample site in Figure 2A is at center-right. 174 profile top, the arroyo face was scraped clean with a trowel, and samples of the sediment were col- lected with a hammer and chisel from the face. Samples were taken at ca. 25 cm intervals from the surface of the deposit to 475 cm depth (total of 20 pollen samples). Samples were placed in 1-gallon bags for transport back to the laboratory. At Northern Arizona University’s Laboratory of Paleoecology (LOP), the pollen samples were pro- cessed by a technique which included suspension of a 20 cc subsample of sediment in dilute HCl to dissolve carbonates. Lycopodium tracer tablets are added at this step to allow for the calculation of pollen concentration. Subsequent steps included ov- ernite suspension in HF to dissolve silicates, and flotation of the pollen in ZnBr,. The resulting pol- len residue was washed with distilled water and placed in glycerol for examination on microscope slides. Each sample contained abundant charcoal, so a final step was added in which we suspended the pollen residue in sodium pyrophosphate and fil- tered it through an 8 pw mesh. This eliminated most of the charcoal, making the sample more easily counted. The resulting pollen samples were counted to a fixed pollen sum as necessary on a Leitz mi- croscope, with reference to the modern pollen ref- erence collection at the LOP. The pollen sum in- cluded all terrestrial pollen types (plus degraded and unknowns), and excluded aquatics (Cypera- ceae, Typha), riparian trees (Tamarix, cf. Platanus, Populus, Juglans, Alnus and Salix), and spores. Three samples were taken for radiocarbon dat- ing, at 100, 170 and 370 cm depth. Radiocarbon dates were performed by Beta Analytic, Miami. Conventions for pollen identifications. Forty-one pollen and four spore types were recognized, ex- clusive of deteriorated and unknown grains. Where possible, grains were assigned to the generic level, more rarely to the family level. Several groupings were further subdivided, including the Asteraceae, the genus Pinus, the Malvaceae, and the spores. The following conventions were used for those groups that were subdivided: Asteraceae. Pollen of the Asteraceae were sepa- rated into 11 groups. Main groupings included Am- brosia (ragweed), Artemisia (sage), Cirsium (this- tle), Liguliflorae (chicory), and other Asteraceae. However, since a majority of grains were in the other Asteraceae group, six morphotypes were rec- ognized, based upon the size and shape of the spines, as well as the size of the grain itself. Based upon an unpublished list of plants growing on Camp Pendleton Marine Base, the following genera were included in the 11 Asteraceae subgroups. Am- brosia-type included Ambrosia, Iva, and Xanthium. Artemisia-type included only Artemisia. Cirsium- type included Cirsium and Centaurea. Liguliflorae included Cichorium, Lactuca, and Taraxacum. Each of these species has been introduced to the Base, thus ancient pollen grains must include ad- MADRONO [Vol. 45 ditional species not presently near the site. Bac- charis-type included Baccharis, Brickellia, Conyza, Cotula, Encelia, Erigeron, Eriophyllum, Gutierre- zia, Helianthus, Heterotheca, Pluchea, and Senecio. Coreopsis-type included only Coreopsis. Chaen- actis-type included Bebbia, Chaenactis, and Vigui- era. Solidago-type included Euthamia, Solidago, and Gnaphalium. Two additional members of the Asteraceae include Type 4 and Type 6, not refer- able to other members of the family presently iden- tified. Pinus. Pine pollen was separated into diploxylon- type (Pinus coulteri D. Don and P. ponderosa Laws.; both presently occur on the Base), and un- differentiated pine-type. Brassicaceae. Only one pollen type of mustard was recovered primarily in the most recent sedi- ments. This was assigned to the genus Brassica, because of the ubiquity of the plant. Apiaceae. Only one pollen type of umbel was recovered, primarily contemporaneous with Bras- sica. The pollen type is most similar to Foeniculum, which is also the most common umbel at the site © today. Malvaceae. Specimens of Sphaeralcea and an unknown Malvaceae occurred in the pollen assem- blages. Spores. Four types of trilete spores occurred in the assemblages. These were assigned to Cheilan- thes-type, Ophioglossum-type and two unknowns. RESULTS Sediments. Sediments of the pollen profile were de- _ scribed in detail by Waters (1996). The column consists of six major units. Unit I (bottom-most © unit) is a brown sandy clay with a paleosol at the © top, containing calcium carbonate nodules and root — casts. Unit II is a silty sand, also containing calcium carbonate nodules and root casts, capped by another © paleosol. Unit III is a fining upward sequence with © sand at the base, fining upward to sandy silty clay | at the top. A third paleosol caps the sequence. Cal- | cium carbonate was also present in these sediments. Unit IV, a silt and silty clay, again contained pedo- genic features, including calcium carbonate. Unit V © was a gray, bioturbated silt, with archaeological de- _ bris and marine shells throughout the deposit. The upper unit, VI, is highly disturbed, and may consist | of spoil sediments. Chronology. Three radiocarbon dates were obtained — from this profile (Table 1). Beta-75375 (1800 + 80 yr BP) came from Unit IVb, Beta-76432 (2610 + 80; collected and submitted by Mike Waters) came from Unit IIIb, and Beta-75376 (4230 + 60 yr BP) came from near the top of Unit Ib. The first and © third samples contained low carbon content, and © required extended counting. The low standard error suggests that little if any contamination by older or | 1998] TABLE 1. RADIOCARBON DATES FROM THE LAS FLORES POLLEN PROFILE. Depth Conventional Calibrated Lab # (cm) C14 age calendar yrs Beta-75375 100 1,800 + 80 AD 60-420 Beta-76432 170 2,610 + 80 BC 905-515 Beta-75376 370 = 4,230 + 60 BC 2920-2610 younger carbon was present, even though each sample consisted of bulk sediments. Pollen. Forty-one pollen types were identified from the Las Flores pollen assemblages (Appendix Table 1). Pollen preservation varied from very good in levels near the surface, to very poor near the bot- tom. This was reflected both in the number of grains counted to achieve a rational pollen sum, as well as the pollen concentration from each of the samples. Pollen sums exceeded 300 grains in sam- ples above 150 cm depth, were ca. 200 grains be- tween 175 and 225 cm depth, dropped to below 100 grains from 250 to 300 cm, and were essentially barren below 330 cm depth (Appendix Table 1). Similarly, pollen concentrations (grains/cc) aver- aged 25,400 grains/cc (range = 888 to 83,700) in the top 150 cm, but dropped to 1815 grains/cc (range 521 to 6060) from 175-225 cm. Pollen con- centrations below 225 cm were generally a couple of hundred grains/cc or less (Appendix Table 1). Pollen zones. Based upon composition of the pollen assemblages, as well as pollen concentration val- ues, five pollen zones are recognized. Pollen zones correspond largely to major breaks in sedimentary composition, as determined by Waters (1996). The pollen assemblages are described below: Pollen Zone I (bottom of profile [475 cm] to 315 cm). Pollen is essentially absent from this section of the core (Fig. 3; Appendix Table 1). Individual grains of Cupressus-type, Baccharis-type, Brassi- caceae, and Apiaceae were found, but their num- bers were insufficient to warrant an interpretation. These sediments correspond largely to Waters’ (1996) Units I and II. The single radiocarbon date from this unit is middle Holocene (4230 + 60 yr BP). Pollen Zone II (ca. 315—140 cm). Pollen concentra- tions increase substantially throughout the zone, from negligible amounts at the bottom to several thousand near the top. This trend is paralleled by an increase in degraded grains (see below). The dominant pollen type in the lower portions of the zone is Cupressus-type, but above ca. 225—250 cm Cupressus declines as several pollen types of the Asteraceae increase (Fig. 3). Prominent among these is the Baccharis-type, though Solidago-, Co- reopsis-, and Chaenactis types also become impor- tant. Members of the Ambrosia group, as well as the Liguliflorae, are abundant. Artemisia and grass- ANDERSON AND BYRD: LAS FLORES CREEK POLLEN 175 es also increase toward the end of the zone. Spores of several species are prominent during this zone, including Cheilanthes, Ophioglossum, and two un- identified types. Aquatic herbs are most abundant during the early portion of the zone. Pollen Zone II largely corresponds to Waters’ (1996) Unit III. Waters collected a radiocarbon date near the top, dating 2610 + 80 yr BP. Pollen Zone III (ca. 140-100 cm). Pollen in this zone is little changed from Zone II below it, except in the relative absence of Cupressus-type and the increase for the first time in pollen of Chenopodi- aceae-Amaranthus plants (Cheno-Am) (Fig. 3). The pollen assemblage is still dominated by members of the Asteraceae, though Coreopsis- and Chaenac- tis-types are less important. Other than a few grains of sedge (Cyperaceae) no aquatic or riparian pollen is found. This zone corresponds to Waters’ (1996) Unit IV. The single radiocarbon date near the top of this unit is 1800 + 80 yr BP. Pollen Zone IV (ca. 100-40 cm). Major changes in the pollen profile begin in this zone. Pollen con- centrations are the highest of any pollen zone, av- eraging 67,750 grns/cc. The pollen assemblage con- tinues to be dominated by Baccharis- and Solidago- types, but Cheno-Am, Artemisia, and grasses re- main prominent. Cupressus and Ambrosia are absent. One pollen type, Eriogonum, becomes most abundant in this zone, while a second, Brassica- type, is first recognized here. Riparian indicators (Alnus, Salix, and Cyperaceae) are also important once again. Pollen zone IV corresponds to Unit V of Waters (1996). Pollen Zone V (ca. 40 cm to the profile top). This pollen assemblage is the most distinctive of any in the profile. The assemblage is dominated by Bras- sica- and Foeniculum-type pollen, both indicative of severe disturbance. Cheno-Am pollen, another disturbance indicator, reaches its maximum here. Tree pollen includes Pinus, and the introduced spe- cies Eucalyptus, Olea, and Tamarix. Riparian spe- cies include cf. Platanus, Juglans, Salix, Alnus, and Cyperaceae (Fig. 3). This zone corresponds to Wa- ters’ (1996) Unit VI. DISCUSSION We examined several sites along the Camp Pen- dleton coastline for potential reconstruction of Ho- locene paleoenvironments. Our reconnaissance, as well as those of Waters (1996), suggested that anal- ysis of alluvial deposits provides the best opportu- nity for paleoecological reconstruction there. Wa- ters (1996) and Waters et al. (in press) identified two types of stream systems operating along the coast. Some streams, like Santa Margarita Creek, have large drainage basins with through-flowing discharge capable of maintaining an open channel to the ocean. Other fluvial systems have consider- ably smaller drainage basins with smaller discharg- 176 MADRONO [Vol. 45 1800 + 80- 2610 + 80- Radiocarbon Dates 400 425 450 e e 475 pone e 10000 20 20 40 20 40 60 20 40 20 20 grains/cc 450 e@ @ 475 ea = 2 ee 20 20 20 Fic. 3. Pollen percentage diagram for the Las Flores Creek pollen profile. Silhouettes are the pollen percentages X10. Dots represent single-grain occurrences of the individual pollen types. es, and do not terminate at the ocean. Instead, these drains. Subsequently, shoreline processes restore drainages most often terminate in lagoons or the beach and the slough forms once again. The sloughs, dammed by beach sand. During large Las Flores Creek deposit falls in the latter category. storms when stream discharge increases, a channel Differences of opinion exist between researchers cuts through the beach deposit, and the slough regarding the interpretation of alluvial pollen as- 1998] semblages. Hall (1985) reviewed the literature from alluvial sites in the Southwest. Though local veg- etation changes can be interpreted from the data, Hall noted that many profiles contain unconformi- ties, with pollen sequences typically beginning or ending during the late Pleistocene or middle Ho- locene, due to changes in sedimentation, erosion, or soil formation. Fall’s (1987) work in Arizona suggested to her that alluvial pollen was unreliable for reconstruction of regional paleoenvironments. For most contexts, it should be assumed that allu- vial pollen integrates vegetation occurring near the stream or channel. At a minimum, alluvial pollen records local vegetation changes, often associated with paleohydrologic changes, and tied to climate. The record exposed along Las Flores Creek rep- resents at least the last ca. 4300 years, perhaps lon- ger. It is an important record since records from only three other coastal sites (Davis 1992; Cole and Liu 1994; Mudie and Byrne 1980) have together defined the Late Holocene sequence for southern California. Each of these studies record vegetation changes within coastal salt marshes, while the rec- ord from Las Flores Creek is largely from a coastal bluff. In addition, the Las Flores site exists along- side a prehistoric archaeological site occupied by Native Americans (during Pollen Zone IV). The pollen changes presented here record some startling vegetation changes along the Las Flores corridor, though unfortunately, the record cannot be interpreted with any precision below ca. 315 cm depth. Pollen Zone I (sedimentary Units I and II), deposited primarily before ca. 4200 yr BP, record a basal and succeeding unit of deposition, each cap- ped by soil development. Pollen preservation was largely unfavorable in these sediments. Poor pres- ervation and low concentrations could result from intense decomposition during soil formation, or rapid sedimentation and dilution of pollen. Existing evidence cannot exclude either hypothesis. However, during Pollen Zone II time (deposi- tional Unit IID), pollen preservation increased from bottom to top. Waters (1996) described this unit as a fining upward sequence, implying channel infill- ing and overbank deposition. The resulting pollen assemblage provides some support for this inter- pretation, as well as insight into the vegetation communities that existed along the banks of the an- cient Las Flores Creek. Riparian indicators (Typha, Cyperaceae, Ophioglossum-type) are most promi- - hent in the coarser sediments deposited near the opening of Zone II, and become less important with time. The increase in degraded grains near the top of the unit are indicative of intense soil formation. A third period of soil formation commenced some- time around 2600 years ago. The occurrence of Cupressus-type pollen is of great interest, yet becomes problematic, for several reasons. Cupressus does not occur on Camp Pen- dleton today. The abundance of Cupressus-type pollen during this time is suggestive of a riparian ANDERSON AND BYRD: LAS FLORES CREEK POLLEN ATs gallery forest of Cupressus, or, alternatively, trans- port of the pollen from nearby upstream locations. The disappearance of Cupressus-type pollen during the waning stages of Zone II time could then rep- resent movement of the stream away from the pres- ent profile site, or local extermination of the spe- cies, perhaps due to more frequent drying of the slough. Problematic is the fact that today Cupressus guadalupensis ssp. forbesii and C. arizonica ssp. E. Greene stephensonii are both found in chaparral foothills, though the former presently occurs as low as 150 m (Beauchamp 1986). A remnant population of the former grows in the Santa Ana mountains of Orange County (Vogl et al. 1988). The morphology of the grain is not referable to Juniperus, and is less like Calocedrus, both of which occur at higher elevations and away from the site today. All mem- bers of the Cupressaceae produce copious amounts of pollen, and the absence of Cupressus-type pollen higher in the profile suggests local extirpation of the plant. Increases in pollen of Artemisia, Eriogonum and Poaceae, as well as the establishment of a large number of species in the Asteraceae family, suggest establishment of a vegetation mosaic, including coastal sage, chaparral and grassland communities, near the site by ca. 2600 years ago. This mosaic has largely persisted until the present, with allow- ances for recent introductions of non-native plants and changes associated with human disturbance. Pollen spectra from Santa Barbara Basin cores show similar increases in species of coastal sage scrub (Artemisia, Eriogonum, Labiatae) and chap- arral (Rosaceae-Rhamnaceae-Anacardiaceae) dur- ing the Late Holocene (Heusser 1978). Since the source of pollen extracted from this marine core is the area drained by the Ventura and Santa Clara Rivers, we conclude that the record from Las Flores is a local example of a large-scale development of coastal vegetation in southern California. The Las Flores record is important in confirming this theory, and fixing the timing of the vegetation change. The pollen assemblage of Zone III time largely represents continuation of conditions during Zone II, and a transition into Zone IV times. Sediments deposited during Zone IV time (Unit V of Waters, 1996) include a prehistoric archaeological occupa- tion represented in the section by marine shells (dominated by Donax gouldii), fire-affected rock, and charcoal; a radiocarbon date suggests deposi- tion centered around 1800 years ago. Archaeolog- ical excavations of this site directly to the north revealed an extensive coastal shell midden also ra- diocarbon dated between 1800 and 1500 years ago (Byrd 1996). This prehistoric occupation is part of a regional trend toward intensive exploitation of pe- riodic Donax resurgences during the last 2000 years (Byrd, in press; Reddy 1996a; Laylander and Saun- ders 1993). These sediments are darker colored and considerably more organic than those below. This may have contributed to the higher concentration 178 MADRONO of pollen found during Zone IV time. Though the pollen is not significantly different from below, in- dicators of local disturbance (e.g., Brassica-type and Cheno-Am pollen) suggest human activity. (Brassica-type pollen probably includes several species or genera, and probably does not represent the same species found in Zone V above). The first occurrence of willow pollen suggests that the mod- ern willow scrub community originated on the de- posit surface during this time. No pollen types indicative of aboriginal cultiva- tion were found in these sediments. Recent inves- tigations have demonstrated that aboriginal occu- pation patterns were probably not characterized by agriculture or horticulture (see Shipek 1989), but instead entailed intensive exploitation of wild plant and animal resources, of which shellfish, fish, small terrestrial mammals, grasses, and nuts were dietary staples (Byrd 1996; Glassow and Wilcoxon 1988; Jones 1991, 1992). The slowing of sea-level rise during the late Holocene converted the region’s mainly rocky shorelines into larger stretches of sandy beach (Inman 1983), and caused a decline in shellfish productivity (Warren 1964). However, pe- riodic exploitation of massive resurgences of sandy beach swelling Donax and offshore fishing during the last 2000 years provided a niche in which this area’s coastline could be occupied for an extended period each year. Based on seasonality analysis of fish otoliths and paleoethnobotanical remains, this site along Las Flores Creek was occupied at least from March through October (Hudson 1996; Reddy 1996b). If abandoned during the winter, inland sites situated among oak groves would have provided an alternative seasonal niche to exploit. The uppermost sediments and pollen are certain- ly recent in age, dating to the occupation of Euro- peans. Several pollen types confirm this interpre- tation. Dominating the pollen assemblage are Foen- iculum-type (Apiaceae) and Brassica-type (Brassi- caceae). Foeniculum vulgare and Brassica nigra, which today grow on the deposit surface, are both native to Europe (Beauchamp 1986). Eucalyptus trees, native to Australia, were imported to North America and planted in the San Francisco region prior to 1860 (Ingham 1908); they were first plant- ed in San Diego County during the period 1902— 1910 (Stanford 1970; Mudie and Byrne 1980). Ta- marix, a native of the Middle East and Old World, was widely planted in the southwestern U.S., and has escaped along watercourses. Most of the plants in cultivation here were part of a clone established by J. J. Thornber in Arizona, during the early part of the 20th century (Benson and Darrow 1981; see also Baum 1967). Arroyo cutting probably inten- sified during Zone V time as witnessed by the in- crease in riparian pollen types (Fig. 3). This may have allowed the deposit surface to dry somewhat, causing weedy annuals to be favored. Though timing of the Pollen Zone I/II transition is uncertain, interpolation between radiocarbon [Vol. 45 dates suggests that the major palynological change occurred ca. 3800 years ago, consistent with an in- crease in effective precipitation, following a drier middle Holocene. Abundant evidence from the Si- erra Nevada and other locations suggest that effec- tive precipitation increased during the late Holo- cene. Anderson and Smith (1994) reported wet meadow deposition originated in the Sierra by ca. 4500 years ago, indicating rising groundwater lev- els. Treeline was higher (warmer conditions) in the Sierra and adjacent White Mountains prior to ca. 3700 years ago (LaMarche 1973; Scuderi 1987). Along the coast, Cole and Liu (1994) inferred in- creased precipitation on Santa Rosa Island begin- ning ca. 3250 yr BP, while Davis (1992) inferred drier conditions. Cole and Liu (1994) suggested that the disparity might be due to the proximity of the Santa Rosa site to the ocean, which, like the Las Flores site, is less than 100 m from the shore- front, while San Joaquin Marsh (Davis 1992) is ca. 7 km inland. Thompson et al. (1993) presented maps of the effective moisture for the western United States, depicting the maximum effective moisture for the entire Pacific Coast during the last ca. 3000 years. Unfortunately the only blank spot on the map lies in southern California. The data presented here sug- gest that southern coastal California has responded to climatic change in a manner similar to the rest of coastal California and the Pacific Northwest. CONCLUSIONS Analysis of pollen from a stratigraphic section along Las Flores Arroyo has allowed us to recon- struct the local paleoenvironment over the last sev- eral thousand years. The Las Flores data are im- portant because they represent one of only a hand- ful of Holocene paleobotanical sites along the southern California coast, and are the first attempt to reconstruct paleoenvironments from Camp Pen- dleton Marine Base. In most cases, pollen changes can be correlated with sedimentary changes, as determined by Waters (1996). Declines in pollen concentration with depth | are indicative of the increased length of time that the pollen has been subject to decomposition. Al- | kaline conditions, as shown by petrocalcic nodules, are unfavorable for pollen preservation; in general, the greater oxidation potential of sandy sediments also contributes to increased pollen decomposition. | Even so, the pollen evidence tells us that the pa- leoenvironment during Zone II time (near the end of the Middle Holocene) was considerably different — from the modern environment. Cupressus (or a sim- ilar tree) probably grew along the watercourse. This | suggests wetter conditions than occur today during | the early part of Zone II, allowing the tree to grow _ at lower elevations. Subsequent stabilization of the surface and colonization by herbs occurred by about 2600 years ago. These conditions were main- tained until the present. 1998] During the archaeological occupation of the val- ley floor, Pollen Zone IV, the vegetation growing locally was little changed from earlier, but in- creased disturbance may be seen by elevated Cheno-Am pollen. No distinctly identified cultivars were noted, but pollen evidence suggests semi-per- manent water nearby probably provided at least a seasonal freshwater supply. The historic period is recorded in the top 30—40 cm of the profile by the occurrence of several pol- len types of introduced species. We do not know the source of these sediments, but the high riparian pollen content suggests that the sediments either came from a different alluvial source of recent de- position, or that incision of the deposit occurred within the most recent 100 years or so. ACKNOWLEDGMENTS This research was conducted as part of a project for the Army Corps of Engineers through a Legacy Resources Management Program award to Camp Pendleton Marine Corps Base. We thank Pamela Maxwell, Army Corps of Engineers, and Stan Berryman, Base Archaeologist, Camp Pendleton, for their support, Seetha Reddy for assistance in collection of pollen samples, and Susie Smith for con- structing the pollen diagram. We also thank Paul Zedler and Dawn Lawson for allowing us to use unpublished information on the vegetation of Camp Pendleton. Con- tribution Number 57 of the Laboratory of Paleoecology. LITERATURE CITED ADAM, D. P. 1985. Quaternary pollen records from Cali- fornia. Pp. 125-140 in V. M. Bryant, Jr. and R. G. Holloway (eds.), Pollen records of Late-Quaternary North American Sediments. American Association of Stratigraphic Palynologists, Dallas, TX. ANDERSON, R. S. AND S. J. SMITH. 1994. Paleoclimatic interpretations of meadow sediment and pollen stra- tigraphies from California. Geology 22:723—726. Baum, B. 1967. Introduced and naturalized tamarisks in the United States and Canada. Baileya 15:19—25. BEAUCHAMP, R. M. 1986. A flora of San Diego County, California. Sweetwater River Press, National City, California. BENSON, L. AND R. A. DARROW. 1981. Trees and shrubs of the southwestern deserts. University of Arizona Press, Tucson. ByrbD, B. FE In Press. Harvesting the littoral landscape dur- ing the Late Holocene: new perspectives from North- ern San Diego County. Journal of California and Great Basin Anthropology. Byrp, B. E 1996. Coastal archaeology of Las Flores Creek and Horno Canyon, Camp Pendleton, Califor- nia. Report submitted to the U.S. Army Corps of En- gineers, Los Angeles District, California. ASM Affil- iates, Inc., Encinitas, CA. (Report on file, South Coast Information Center, San Diego State University). BYRNE, R., J. MICHAELSEN, AND A. SOUTAR. 1977. Fossil charcoal as a measure of wildfire frequency in Southern California: a preliminary analysis. Pp. 361— 367 in Proceedings of the Symposium on the envi- ronmental consequences of fire and fuel management in Mediterranean ecosystems. General Technical Re- port WO-3, USDA Forest Service, Washington, D.C. ANDERSON AND BYRD: LAS FLORES CREEK POLLEN AS CoLe, K. L. AND G. Liu. 1994. Holocene paleoecology of an estuary on Santa Rosa Island, California. Quater- nary Research 41:326—335. Davis, O. K. 1992. Rapid climatic change in coastal southern California inferred from pollen analysis of San Joaquin Marsh. Quaternary Research 37:89—100. FALL, P. L. 1987. Pollen taphonomy in a canyon stream. Quaternary Research 28:393—406. GLassow, M. A. AND L. R. WILCOXON. 1988. Coastal ad- aptation near Point Conception, California, with par- ticular regard to shellfish exploitation. American An- tiquity 53:36—51. HALL, S. A. 1985. Quaternary pollen analysis and vege- tational history of the Southwest. Pp. 95-124 in V. M. Bryant, Jr. and R. G. Holloway (eds.), Pollen re- cords of Late-Quaternary North American sediments. AASP, Dallas. HEusser, L. E. 1978. Pollen in Santa Barbara Basin, Cal- ifornia: a 12,000 year record. Geological Society of America Bulletin 89:673—678. Hupson, J. 1996. Vertebrate remains from SDI-811, SDI- 4538, and SDI-10,726 at Camp Pendleton. Pp. 241— 274 in Byrd, B. EF (ed.), Coastal archaeology of Las Flores Creek and Horno Canyon, Camp Pendleton, California. Report submitted to U.S. Army Corps of Engineers, Los Angeles District, California. ASM Af- filiates, Inc., Encinitas, CA. INGHAM, N. D. 1908. Eucalyptus in California. University of California Experiment Station Bulletin 196. INMAN, D. L. 1983. Application of coastal dynamics to the reconstruction of paleocoastlines in the vicinity of La Jolla, California. Pp. 1-49 in P. M. Masters and N. C. Flemming (eds.), Quaternary coastlines and ma- rine archaeology. Academic Press, London. JONES, T. L. 1991. Marine-resource value and the priority of coastal settlement: a California perspective. Amer- ican Antiquity 56:419—443. JONES, T. L. 1992. Settlement trends along the California coast. Pp. 1-38 in T. L. Jones (ed.), Essays on the prehistory of maritime California. Center for Archae- ological Research at Davis Publications No. 10, Da- vis, CA. KOEHLER, P. A. AND R. S. ANDERSON. 1994. The paleo- ecology and stratigraphy of Nichols Meadow, Sierra National Forest, California, U.S.A. Palaeogeography, Palaeoclimatology, Palaeoecology 112:1—17. LAMARCHE, V. C. JR. 1973. Holocene climatic variations inferred from treeline fluctuations in the White Moun- tains, California. Quaternary Research 3:632—660. LAYLANDER, D. AND D. SAUNDERS. 1993. Donax exploi- tation on the Pacific Coast: spatial and temporal lim- its. Proceedings of the Society of California Archae- ology 6:313-325. Mupigz, P. J. AND R. ByRNE. 1980. Pollen evidence for historic sedimentation rates in California coastal marshes. Estuarine and Coastal Marine Science 10: 305-316. PACIFIC SOUTHWEST BIOLOGICAL SERVICES, INC. 1986. Veg- etation of Camp Pendleton, Las Pulgas Canyon Quad- rangle. Map. Reppy, S. N. 1996a. Experimental ethnoarchaeological study of Donax gouldii exploitation. Pp. 231—240 in Byrd, B. F (ed.), Coastal archaeology of Las Flores Creek and Horno Canyon, Camp Pendleton, Califor- nia. Submitted to U.S. Army Corps of Engineers, Los Angeles District, California. ASM Affiliates, Inc., En- cinitas, CA. (Report on file, South Coast Information Center, San Diego State University). 180 MADRONO Reppy, S. N. 1996b. Paleoethnobotanical investigations at Las Flores Creek and Horno Canyon, Camp Pendle- ton. Pp. 275—304 in Byrd, B. E (ed.), Coastal archae- ology of Las Flores Creek and Horno Canyon, Camp Pendleton, California. Submitted to U.S. Army Corps of Engineers, Los Angeles District, California. ASM Affiliates, Inc., Encinitas, CA. (Report on file, South Coast Information Center, San Diego State Universi- ty). SCUDERI, L. A. 1987. Late Holocene upper treeline vari- ation in the southern Sierra Nevada. Nature 325:242-— 244. SHIPEK, EF C. 1989. An example of intensive plant hus- bandry: the Kumeyaay of southern California. Pp. 159-170 in D. R. Harris and G. C. Hillman (eds.), Foraging and farming: the evolution of plant exploi- tation. Unwin Hyman, London. STANFORD, L. G. 1970. San Diego’s eucalyptus bubble. Journal of San Diego History 16:11—-19. THOMPSON, R. S., C. WHITLOCK, P. J. BARTLEIN, S. P. HAR- RISON, AND W. G. SPAULDING. 1993. Climatic changes in the western United States since 18,000 yr B.P. Pp. 468-513 in H. E. Wright, Jr., J. E. Kutzbach, T. Webb Il], W. EF Ruddiman, FE A. Street-Perrott, and P. J. Bartlein (eds.), Global climates since the last glacial maximum. University of Minnesota Press, Minneap- olis. [Vol. 45 VOGL, R. J., W. PB. ARMSTRONG, K. L. WHITE, AND K. L. CoLe. 1988. The closed-cone pines and cypress. Pp. 295-358 in M. G. Barbour and J. Major (eds.), Ter- restrial vegetation of California. California Native Plant Society, Special Publication 9. WARREN, C. N. 1964. Cultural Change and Continuity on the San Diego Coast. Ph.D. Dissertation. University of California, Los Angeles. WATERS, M. R. 1996. Geoarchaeological investigations at Horno Canyon and Las Flores Creek on Camp Pen- dleton, California. Pp. 47-56 in Byrd, B. E (ed.), Coastal archaeology of Las Flores Creek and Horno Canyon, Camp Pendleton, California. Submitted to U.S. Army Corps of Engineers, Los Angeles District, California. ASM Affiliates, Inc., Encinitas, CA. (Re- port on file, South Coast Information Center, San Di- ego State University). WartERS, M. R., B. EK BYRD, AND S. N. REDDY. 1999. Late Quaternary geology and geoarchaeology of two streams along the southern California coast. Geoar- chaeology 14:289—306. WEST, G. J. AND J. M. ERLANDSON. 1994. A late Pleisto- cene pollen record from San Miguel Island, Califor- nia: preliminary results. AMQUA Abstracts 13:256. ZEDLER, P., J. GiESSow, S. DESIMONE, D. LAwson, J. ELSE, AND S. BLIss. 1996. A guide to the plant communities of Camp Pendleton Marine Corps Base, California. Unpublished Manuscript. 181 ANDERSON AND BYRD: LAS FLORES CREEK POLLEN 1998] 00 exe) 00 OO 00 00 a0) a0) 00 00 0'O ox) 0'0 00 exe) snjndog 0'0 exe) 00 00 00 00 00 0'O 0'0 0'0 00 00 00 0) ox) SNUDID] d “Jo 0'0 00 00 OO oxe) 00 00 00 00 00 0-0 00 00 €°O 90 XIADUD J 0'0 00 00 ve 00 00 00 00 00 00 00 00 00 00 00 IVIOVIBLUO 0'0 00 00 O00 00 00 0'0 0) 00 00 0'0 00 Oxe) 00 00 vaglps1avyds 00 00 0O0 OO 00 00 00 O'l 00 00 0'O 00 0'O 00 00 SBIOBATRIN 0'0 Ox) 00 OO 00 00 00 0'0 00 00 00 £0 0-0 00 0'O wuni]OY) 00 00 0O0 OO 00 0-0 00 00 0'0 0:0 £0 90 00 00 00 SBIORTUIP’T 00 00 0O0 OO 00 00 00 00 Ox) 00 90 &0 00 0-0 0-0 svaoRIUOULa[Od 00 0°0 O11 200 00 00 00 00 00 0:0 0'O 00 00 £0 0°0 adAj-auayig aevaoel[Aydodre) 0'0 00 00 OO 00 00 00 00 0'0 00 00 00 00 LC 0°0 ava0RUOSA[Od 0'0 00 OO OO 00 00 00 00 0) om) 00 90 om) LO 0'0 wnuosolg 0'0 00 00 VC 00 £9 Ce Ge 0'O el O'l 90 LO LO 0'0 seIOYI[NSTT] 0'0 00 OO OO 00 00 00 00 00 00 00 00 0-0 ou 6) 00 seoa0RlIT] x 0'O Ooo OO 0'O 0'O 0'0 0'O 0'O OT me) 0'O €'O e721 rol odAj-winjnoiuaoy sesoeidy 0'0 00 00 OO 00 00 00 00 0'0 00 00 0-0 00 00 £0 avaoelepnydos19g 0-0 00 00 OO 00 00 00 00 00 00 Ox) 00 £0 00 c0 WUNISA1) x 00 00 OO 00 00 00 00 Ol L‘0 9°0 Ol VT i | odA}-DoIsspAg 0'0 00 00 OO 00 Tl c'0 Ol O'l 6v Ol 90 9 00 £0 9BIIVOd 0'0 00 OO. 00 00 0°0 00 a0) O'l ce Ol 91 Ol 00 90 DISIUIJAY 0'0 00 00 OO 00 0'0 0'0 0) 0'O OT 0'O 90 00 LO £0 DEIDVSOY/SNAIVIOILID 0'0 00 00° 00 0'0 00 Ol CO CC 00 I'L 87 I'8 LOL 89 wy-ousy,) 0'0 00 OO: EL x v3 Cl Cl O'l Stl 6 Ol LO LO 61 DISOAQUY 00 00 0'0>«00 00 00 Cl OC O'l 00 | Ol rc Ox) aaa g-adA} aeaovlaisy 00 00 O02 Te 00 00 Vol ee £0 eS CoOL OTe ay 84 | Lil 9'Cl adAj-OdDp1j0g Iade1I\SV 00 00 00 OO x |e c'0 VS 0'O LO om 0) ag | 00 cil exe) p-adA} avaoviaysy 00 00 00 OO 00 £9 Ol 00 CO Ol 00 exe) 00 £0 00 adAj}-sijapuavy) aea0e121SV 00 00 00 OO 00 Ge Ol CO Ox) £0 00 00 ec 00 00 adAj-sisdoaio) 3839e19)SY x x » 4 88 x 8 9¢ COS OPV 66S LL 6°CS CLV ¢ ov Ls 9'OV adA}-sDYyIIVG BeIdVI1I\SV 0'0 00 0O0 OO 00 00 00 0'O 0'0 00 CQ 00 00 LO 00 SBIIBUBLOG “fo 0'0 00 00 00 00 Ll 00 00 00 0-0 0'0 £0 00 00 00 snyjouvay) 0'0 00 00 OO 0'0 00 00 00 0'0 0-0 00 00 00 LO 0'0 D21O “J? 00 00 00 O00 00 00 00 00 0'0 00 00 00 00 £0 00 snjdajvongq 00 00 OO Ct x 00 Ol 0) a0) 00 00 00 00 £0 90 poyenuoloffIpun SMuld 00 ox) 00 O00 00 00 0) 00 00 00 00 00 00 O'€ Ol psosapuod/14aj]nor snuld xX 00 OO 88 x VLC ec EL Cl CL 90 0'0 00 00 00 adAj-snssasdny 0'0 00 00 OO 0'0 |Get Ol 00 0) O'l £0 cl 91 OV 6'l suMOUAUL) 0'0 00 00 VC 00 €9 Ol v9 6'8 SIT EO L'6 6S CL cS poyetloltojod Test 6¢6l OO v6IT CLT 6906 PITS O886 00909 ISL9I O8SIv OOOLES F¢908IS C888 SIVIOI 90/SUTBIS “QUOD UIT][Od 8 I 0) C8 rey C6 COT COC COT pO 80¢ OTE LOE OO 80¢ ung Ue[[Od 6S L I €CS 101 EST €cS 9OLT CV SVC OOT ¢ 8 9S Iv SIOOPIL OSV OOV Otte OO SLC OSC STC OOT CLI OSI OOT CL OS CC 0) (wd) yidop ajdures ‘SHOVLNAOUAd NATIOd NOILLOAS IWIANTIV MIAN SAYOT SWI “| AIAVL XIGNadd Vy MADRONO [Vol. 45 182 00 00 00 O00 00 00 c'0 Ol 00 00 00 00 00 00 00 p-adA} o10dg a}9] 1) 00 00 00 O00 00 00 00 00 00 rome) 00 00 00 00 00 ¢-odAy a1odg aio] 00 00 0O0 O00 00 00 OT me) me) © 0 rome) £0 00 00 0°O adAj-winssojso1ydoC 00 00 00 cl 00 VC Vol OC Ol LO 0 0 LO 00 00 odAj-sayjuppiay) 00 00 O00 v¢ 00 00 00 00 00 00 00 00 00 oo £0 pyds 00 00 00 O00 00 es 00 € 0 =.0 9°0 © 0 rome) 00 Ol 00 avaorIadéD) 00 00 00 O00 00 00 00 00 00 00 00 00 Ol oO 9°0 x11D8' 00 00 00 O00 00 00 00 00 00 00 00 0 00 00 00 SnuTy 00 00 00 OO 00 00 00 00 00 00 00 00 00 L‘0 00 supjéne OSV OOF Ofte ODF SLC OST SCC 007c SLI OSI OOT SL OS SC (wid) yydop ojduies GaNNILNOO “T AIgdV] XIGNdddV Mapbrono, Vol. 45, No. 2, p. 183, 1998 REVIEW Mojave Desert Wildflowers. Jon Mark Stewart. 1998. vi + 210 pages. $14.95. Published by the author. ISBN 0- 9634909-1-5. Mojave Desert Wildflowers is the second offering of photographer Jon Mark Stewart, following his Colorado Desert Wildflowers. Stewart fell under the spell of desert botany while a student, and he has maintained his interests through his photography. Mojave Desert Wildflowers is a recent addition to a wide array of photographic guides to plants of desert regions. The goals of most such guides include 1) provide an easy means to identify plants via photographs and/or drawings, and 2) display the beauty of form and color in the plants of the desert. Attainment of both of these goals, and the quality of the reproduction, varies from guide to guide, but Stewart’s new addition is a notable success. Here is produced a set of very good photographs of many of the commoner, and several less common species that inhabit the Mojave Desert. What caught my attention immediately was the exceptional quality of most of the photos. They are in focus with great depth of field, brightness and contrast are very good, and the color of the plants is true. Too often photos in field guides, whether due to inadequate originals or diminished reproduction, appear over- or underexposed. Bright yellow or white flowers are especially difficult to photograph for publication without seeming glary. Not so in this book. Field guides, to be useful, must show clearly the details that distinguish and differentiate species, and in this aspect Stewart succeeds. Mojave Desert Wildflowers covers 195 species orga- nized by flower color, with the exception of the cacti, which are contained in their own chapter. The taxonomy follows The Jepson Manual (1993), the most recent com- prehensive flora of the region. In the back of the book there is a useful cross-list of names used by Munz’ Flora of California (1959) and Kearney and Peebles’ Flora of Arizona (1960), for those of us who learned the names of desert plants prior to 1993. I didn’t notice a reference to where Stewart got his common names, but I suspect that many of them came from Jaeger’s (1940) Desert Wild Flowers. As is common with wildflower guides, common names of the better known species are well-known and widely accepted, while those of lesser known species of- ten appear to be forced on the species. For example, Lin- anthus parryae is called ‘‘Parry gilia’’, dating from the 19th century when all species of Linanthus were recog- nized as species of Gilia. I doubt if (m)any botanists today would actually call L. parryae by that name. Nonetheless, as there are no rules by which use of common names are followed, any common name appears fair game. At the end of the book there is a page describing the film, equipment, and methods used in the photography. I enjoyed reading about this—not so much because I use essentially the same materials as Stewart, but because I can have hope that someday I might be able to achieve the high standard of photography that is present in this book. I took this book with my plant taxonomy class to the Mojave Desert, and the consensus among these advanced students of botany was that it was a keeper. Among the wide assortment of picture books Mojave Desert Wild- flowers stands out. I recommend the book to all who col- lect and use field guides to desert plants. —ROBERT PATTERSON, Department of Biology, San Francisco State University, San Francisco, CA 94132. MaprOoNno, Vol. 45, No. 2, p. 184-185, 1998 NOTEWORTHY COLLECTIONS ARIZONA CYMOPTERUS BECKI! Welsh & Goodrich (Apiaceae).— Navajo Co., Tsegi Canyon, UTM E. 545000, UTM N. 4066000, near small spring, July, 1996, verified by Stan Welsh of Brigham Young University. Voucher specimen on file at Deaver Herbarium (ASC), Northern Arizona University, Flagstaff. Previous knowledge. Listed as an endemic in San Juan and Wayne counties, Utah. First described in 1981 by Stan Welsh. Significance. This is a first report for Arizona and rep- resents a range extension of about one hundred ten kilo- meters. This plant is a candidate for the federal listing of rare and endangered species. It is found only near seeps and springs in the area and its occurrence is fairly rare. —SUSAN HOLIDAY and TINA AYERS, Northern Arizona University, Box 5640, Flagstaff, AZ 86011. CALIFORNIA LIMNANTHES MACOUNII Trel. (LIMNANTHACEAE).— Abundant on ca 18 acres of a seasonally fallow field along the east side of Highway 1 just south of Moss Beach and opposite the Half Moon Bay airport, San Mateo Co. 24 March 1998. (UC); 12 April 1998. EF. Buxton s. n. R. Orn- duff 10168 (UC). This large population was discovered by the first author in early February, 1998; flowering plants were present on the site until late May, 1998, when the field was plowed prior to planting cabbage. Previous knowledge. Limnanthes macounii is otherwise restricted to a small portion of southern Vancouver Island and offshore islets in and near Victoria, British Columbia, Canada. Elsewhere the genus occurs in California and southwestern Oregon (L. alba Benth. has become locally established in Linn County, Oregon, where it is cultivated as an oilseed crop). Signficance and comment. There were doubtless many more individuals of L. macounii in the Moss Beach pop- ulation in 1998 than in all the British Columbia popula- tions combined. The Moss Beach plants are unusually ro- bust for L. macounii, producing decumbent fruiting stems that are up to 60 cm long. In certain foliar characters they differ somewhat from British Columbia specimens (A. Ceska, personal communication). Because the California population occurs in a field that is adjacent to and easily visible from a well traveled highway and is opposite an airport, we suspect that it is not native to the site but originated via an accidental introduction. The species is autogamous and thus successful establishment of a new population requires the introduction of only a single nut- let. We have no idea how long this population of L. ma- counii has occupied the field, but its large size suggests that it has been present since well before 1998. We thank Adolf Ceska for his helpful comments. —Eva Buxton, LSA Associates, Inc., 157 Park Place, Point Richmond, CA 94801; ROBERT ORNDUFF, Depart- ment of Integrative Biology, University of California, Berkeley, CA 94720-3140. CALIFORNIA SAGITTARIA RIGIDA Pursh (ALISMATACEAE).—Marin Co., Pt. Reyes Peninsula, dunes at SE end of Abbotts La- goon, very abundant in small farm pond N of radio tower facility, associated with Polygonum amphibium L. var. emersum Michaux, Hydrocotyle ranunculoides L. f., Co- tula coronopifolia L., etc., 38°06'30"N, 122°56'40’W, alt. ca. 6 m, 20 Jul 1987, R. Raiche 70477 (JEPS). Tehama Co., in large stock pond on an unnamed tributary of Inks Creek ca. 1.3 air miles N of Dales Lake, 40°21’'N, 122°3'55"W, T29N R2W NE% of SE% S22, alt. 185 m, ca. 1000 individuals growing in association with Sagittar- ia latifolia Willd, S. sanfordii E. Green, Eleocharis ma- crostachya Britton, and Scirpus acutus Bigelow var. oc- cidentalis (S. Watson) Beetle, 26 May 1992, Dean Wm. Taylor 12649 (UC!) & 21 Jul 1992, C. Witham 450 (JEPS). Plumas Co., east side of Last Chance Marsh lo- cated at the north end of Lake Almanor, 40°20'5’N, 121°12'25"W, T29N R7E NW% of NE% S33, alt. 1365 m, ca. 1000 individuals in colonies scattered along 300 m of marsh in association with Menyanthes trifoliata L., Nu- phar lutea (L.) Sibth & Sm. ssp polysepala (Engelm.) E. Beal, Potamogeton natans L., Utricularia vulgaris L., and other marsh vegetation, vegetative plant on 6 Sep 1994, V. Oswald 6476 (CHSC) & flowering and fruiting plants on 22 Jul 1997, V. Oswald 8768 (CHSC). Previous knowledge. A plant of brackish and saline wa- ters of eastern North America (Que. to MN south to KS, MO, VA). Significance. First records for California and western North America. S. rigida can be separated from all other species of Sagittaria in California based on the three pis- tillate flowers and fruiting heads, which appear to be ses- sile in the lowest whorl of the inflorescence. All other Sagittaria in California have obviously pedicelled pistil- — late flowers and fruiting heads. Plants from all three pop- ulations have been annotated by Robert Haynes, The Uni- versity of Alabama, who is coordinating the treatments of aquatic plants for the Flora of North America project. The three California populations of S. rigida are in artificial ponds and lakes separated by distances of from 72 to 282 km. How the plant arrived in California is open to con- jecture, but it can now be expected to become more wide- ly dispersed through the movements of waterfowl. —VERNON H. Oswa Lp, Herbarium, Department of Bi- ological Sciences, California State University, Chico, CA | 95929-0515; ROGER RAICHE, | Maybeck Twin Dr., Berke- ley, CA 94708-2037; CAROL WITHAM, 1028 Cypress Lane, | Davis, CA 95616-1364. CALIFORNIA UTRICULARIA OCHROLEUCA R. Hartman (U. occidentalis Gray) Lentibulariaceae. Plumas Co., northern end of Lake | Almanor, east side of lake north of Hwy. 36 bridge, T29N- R7E-sw sec. 28, 1437 m. Selected associate species: Utricularia macrorhiza, U. minor, Nuphar polysepalum, 1998] Eriophorum gracile; 24 June 1994, J. H. Rondeau 5169 (SJSU). Previous knowledge. This species is very rare through- out the western U.S. with only three citations in Oregon (Rondeau, 1995) and two in Washington (Ceska & Bell, 1973). The nearest known location is 470 km northward at Gold Lake in central Oregon, although it may exist as far south as Bull Swamp in Klamath County (Rondeau, #995). Special thanks to Goran Thor (Swedish Univ. of Agric. Sciences, Uppsala, Sweden) for taxonomic assistance via quadrifid gland analysis. Significance. First collection for California. LITERATURE CITED Ceska, A. and M. A. Bell. 1973. Utricularia (Lentibular- iaceae) in the Pacific Northwest. Madrono, 22:74—84. Rondeau, J. Hawkeye. 1995. Carnivorous plants of the west, Volume II: California, Oregon, and Washington. San Jose, CA 95127. Thor, G. 1988. The Genus Utricularia in the Nordic Coun- tries, with special emphasis on U. stygia and U. och- roleuca. Nord. Jr. Bot. 8(3):213—225. NOTEWORTHY COLLECTIONS 185 —J. HAWKEYE RONDEAU, ““mybog@aol.com’’, 37 Sun- nyslope Ave., San Jose, CA 95127. OREGON CAREX DIANDRA Schrank (Cyperaceae).—Lake Co., Dog Lake, 4.8 air km SSE of Dog Mountain, E of crest of Barnes Rim, Fremont National Forest, T40S R17E S22 SW%, alt. 1583 m, floating mat in mid-lake, with Carex utriculata, Typha latifolia, Scirpus acutus, 20 July 1996, Zika et al. 12917 (OSC). Previous knowledge. Circumboreal and recorded spo- radically south in our region, with collections from north- ern California, northern Washington, Idaho and Montana. An earlier Oregon report by Peck (A Manual of the Higher Plants of Oregon, 1961) stated: “‘bogs in the high moun- tains of eastern Oregon’’, but was unsubstantiated by her- barium collections. Significance. First verified record for Oregon. —PETER F. ZIKA, KELI KUYKENDALL, DANNA LYTIJEN, and Nick OTTING, Herbarium, Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR 97331. Volume 45, Number 2, pages 93-185, published 5 May 1999 SUBSCRIPTIONS—-MEMBERSHIP Membership in the California Botanical Society is open to individuals ($27 per year; family $30 per year; emeritus $17 per year; students $17 per year for a maximum of 7 years). Late fees may be assessed. Members of the Society receive Maprono free. Institutional subscriptions to MADRONO are available ($60). Membership is based on a calen- dar year only. Life memberships are $540. Applications for membership (including dues), orders for subscriptions, and renewal payments should be sent to the Treasurer. Requests and rates for back issues, changes of address, and undelivered copies of MADRONO should be sent to the Corresponding Secretary. INFORMATION FOR CONTRIBUTORS Manuscripts submitted for publication in MADRONO should be sent to the editor. It is preferred that all authors be members of the California Botanical Society. Manuscripts by authors having outstanding page charges will not be sent for review. Manuscripts may be submitted in English or Spanish. English-language manuscripts dealing with taxa or topics of Latin America and Spanish-language manuscripts must have a Spanish RESuMEN and an English ABSTRACT. Manuscripts and review copies of illustrations must be submitted in triplicate for all articles and short items (NOTES, NOTEWORTHY COLLECTIONS, POINTS OF VIEW, etc.). Follow the format used in recent issues for the type of item submitted. Allow ample margins all around. Manuscripts MUST BE DOUBLE-SPACED THROUGHOUT. For articles this includes title (all caps, centered), author names (all caps, centered), addresses (caps and lower case, centered), abstract and resumen, 5 key words or phrases, text, acknowledgments, literature cited, tables (caption on same page), and figure captions (grouped as consecutive paragraphs on one page). Order parts in the sequence listed, ending with figures. Each page should have a running header that includes the name(s) of the author(s), a shortened title, and the page number. Do not use a separate cover page or ‘erasable’ paper. Avoid footnotes except to indicate address changes. Abbreviations should be used sparingly and only standard abbrevia- tions will be accepted. Table and figure captions should contain all information relevant to information presented. All measurements and elevations should be in metric units, except specimen citations, which may include English or metric measurements. Authors of accepted papers will be asked to submit an electronic version of the manuscript. Microsoft Word 6.0 or WordPerfect 6.0 for Windows is the preferred software. Line copy illustrations should be clean and legible, proportioned to the Maprono page. Scales should be in- cluded in figures, as should explanation of symbols, including graph coordinates. Symbols smaller than 1 mm after reduction are not acceptable. Maps must include a scale and latitude and longitude or UTM references. In no case should original illustrations be sent prior to the acceptance of a manuscript. Illustrations should be sent flat. No illustrations larger than 27 X 43 cm will be accepted. Presentation of nomenclatural matter (accepted names, synonyms, typification) should follow the format used by Sivinski, Robert C., in MADRONO 41(4), 1994. Institutional abbreviations in specimen citations should follow Holmgren, Keuken, and Schofield, Index Herbariorum, 8th ed. Names of authors of scientific names should be abbreviated according to Brummitt and Powell, Authors of Plant Names (1992) and, if not included in this index, spelled out in full. Titles of all periodicals, serials, and books should be given in full. Books should include the place and date of publication, publisher, and edition, if other than the first. All members of the California Botanical Society are allotted 5 free pages per volume in MAprono. Joint authors may split the full page number. Beyond that number of pages a required editorial fee of $40 per page will be assessed. The purpose of this fee is not to pay directly for the costs of publishing any particular paper, but rather to allow the Society to continue publishing MApRONOo on a reasonable schedule, with equity among all members for access to its pages. Printer’s fees for illustrations and typographically difficult material @ $35 per page (if their sum exceeds 30 percent of the paper) and for author’s changes after typesetting @ $4.50 per line will be charged to authors. At the time of submission, authors must provide information describing the extent to which data in the manu- script have been used in other papers that are published, in press, submitted, or soon to be submitted elsewhere. f . VOLUME 45, NUMBER 3 JULY-SEPTEMBER 1998 MADRONO A WEST AMERICAN JOURNAL OF BOTANY WK \ MND S Se. CONTENTS PHYLOGENY OF THE ARCTOSTAPHYLOS HOOKERI COMPLEX (ERICACEAE) BASED ON NRDNA Data Staci Markos, Lena C. Hileman, Michael C. Vasey, and V. Thomas OTIC, Seas ese Re ete Och DEN RO, is PP ta 187 MOorRPHOLOGICAL VARIATION IN CALIFORNIA ALPINE POLEMONIUM SPECIES Daniel W. Pritchett and Robert Patterson .......cccccccccccccccceseccccsesccceneecceeas 200 ADULT SEX RATIO OF ARCEUTHOBIUM TSUGENSE IN SIX SEVERELY INFECTED TSUGA HETEROPHYLLA Robert L. Mathiasen and David C. SHAW vicccccccccccccsccccccseeccccssccccseescceueees 210 POPULATION ECOLOGY OF DUDLEYA MULTICAULIS (CRASSULACEAE); A RARE NARROW ENDEMIC T. Alejandro Marchant, Ruben Alarcon, Julie A. Simonsen, and FLAPOld KOODOWILZ . iiadica ies laveoces caicc REPRO ays soe hose eed ona asics: DNS SEQUOIADENDRON GIGANTEUM-MIXED CONIFER FOREST STRUCTURE IN 1900-1901 FROM THE SOUTHERN SIERRA NEVADA, CA Scott L. Stephens and Deborah L. Elliott-Fisk .....cccccccccceeeeesssseeceeeeeeeeeees 221 DISTRIBUTION OF WINTER ANNUAL VEGETATION ACROSS ENVIRONMENTAL GRADIENTS WITHIN A MOJAVE DESERT PLAYA Robert W. Lichvar, William E. Spencer, and Jonathan E. Campbell ....... 231 FLOWERING PHENOLOGY AND SEX EXPRESSION OF CROTON CALIFORNICUS (EUPHORBIACEAE) IN COASTAL SAGE SCRUB OF SOUTHERN CALIFORNIA Bradford IGM EA od ASG fos oe Nachos bss 1aekle se bbs Sea lp sannsoastensceesess 239 LIMACINIASETA GEN. NOV. A CALIFORNIA SOOTY MOLD DOWR, Reynolds 7.20 of ssc VicGE AGN occas sone ss coca 0M outa oe ate AD a seasawasseeceoeas 250 LIGNOTUBERS IN SEQUOIA SEMPERVIRENS: DEVELOPMENT AND ECOLOGICAL SIGNIFI- CANCE PETEr DCU TT CAIC eee ee ose tee SON RE See EEE 259 CAREX SERPENTICOLA (CYPERACEAE), A NEW SPECIES FROM THE KLAMATH MOunN- TAINS OF OREGON AND CALIFORNIA Peter F: Zika, Keli Kuykendall, and Barbara WiISON ............:::0000eeeeeeeeeee 261 ANTENNARIA DIOICA (ASTERACEAE: INULEAE): ADDITION TO THE VASCULAR FLORA OF CALIFORNIA CII G GUUIVEl EW SKU. eaasde eee tree ane aimee eee eae nee se eer ae eee eee DAA REVIEW THE ONCE AND FUTURE ForREST: A GUIDE TO FOREST RESTORATION STRATEGIES, By L. J. SAUER SC OUEST CDINCIIS wm ns teres iene recesses etn sey neue Ausan ee Neen Cos tase «ase ese aes wee a 273 PUBLISHED QUARTERLY BY THE CALIFORNIA BOTANICAL SOCIETY Maprono (ISSN 0024-9637) is published quarterly by the California Botanical Society, Inc., and is issued from the office of the Society, Herbaria, Life Sciences Building, University of California, Berkeley, CA 94720. Subscription information on inside back cover. Established 1916. Periodicals postage paid at Berkeley, CA, and additional mail- ing offices. Return requested. Postmaster: Send address changes to MApDRONO, ‘/ Mary Butterwick, Botany De- partment, California Academy of Sciences, Golden Gate Park, San Francisco, CA 94118. Editor—KrisTINA A. SCHIERENBECK California State University, Chico Department of Biology Chico, CA 95427-0515 kschierenbeck @csuchico.edu Editorial Assistant—Davip T. PARKS Book Editor—Jon E. KEELEY thy Collections Editors—DIETER WILKEN, MARGRIET WETHERWAX Board of Editors Class of: 1998—FREDERICK ZECHMAN, California State University, Fresno, CA Jon E. KegLey, U.S. Geological Service, Biological Resources Division, Three Rivers, CA 1999—Timotny K. Lowrey, University of New Mexico, Albuquerque, NM J. MARK Porter, Rancho Santa Ana Botanic Garden, Claremont, CA 2000—Pame-a S. Sottis, Washington State University, Pullman, WA JOHN CaLLAway, San Diego State University, San Diego, CA 2001—Robert PATTERSON, San Francisco State University, San Francisco, CA PAULA M. ScHIFFMAN, California State University, Northridge, CA 2002—NorMAN ELLSTRAND, University of California, Riverside, CA Carta M. D’ Antonio, University of California, Berkeley, CA CALIFORNIA BOTANICAL SOCIETY, INC. OFFICERS FOR 1998-1999 President: R. JoHN LittLe, Sycamore Environmental Consultants, 6355 Riverside Blvd., Suite C, Sacramento, CA 95831 First Vice President: SusAN D’ ALcamo, Jepson Herbarium, University of California, Berkeley, CA 94720 Second Vice President: Davip KetL, California Polytechnic State University, Biological Sciences Department, San Luis Obispo, CA 93407 Recording Secretary: ROXANNE BITTMAN, California Department of Fish and Game, Sacramento, CA 95814 Corresponding Secretary: | SUSAN BAINBRIDGE, Jepson Herbarium, University of California, Berkeley, CA 94720. suebain @casnowy.qal.berkeley.edu Treasurer: Mary Butterwick, Botany Department, California Academy of Science, Golden Gate Park, San Fran- cisco, CA 94118. butterwick.mary @caepamail.epa.gov The Council of the California Botanical Society comprises the officers listed above plus the immediate Past President, WAYNE R. FERREN, JR., Herbarium, University of California, Santa Barbara, CA 93106; the Editor of Maprono; three elected Council Members: MARGRIET WETHERWAX, Jepson Herbarium, University of California, Berkeley, CA 94720; James SHEvock, USDI National Park Service, Pacific West Region, 600 Harrison Street, Suite 600, San Francisco, CA 94107; DIANE ELam, U.S. Fish and Wildlife Service, 3310 El Camino Avenue, Sacramento, CA 95825; Graduate Student Representative: DENNIS P. WALL, Jepson Herbarium, University of California, Berke- ley, CA 94720. This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). MaprONOo, Vol. 45, No. 3, pp. 187-199, 1998 PHYLOGENY OF THE ARCTOSTAPHYLOS HOOKERI COMPLEX (ERICACEAE) BASED ON NRDNA DATA STACI MARKOS!, LENA C. HILEMAN?, MICHAEL C. VASEY, AND V. THOMAS PARKER Department of Biology, San Francisco State University, San Francisco, CA 94132 ABSTRACT The Arctostaphylos hookeri G. Don complex is composed of five subspecies whose classification has been problematic. We investigated the monophyly of A. hookeri using sequence data from the ITS and 26S regions of nuclear rDNA. Several of the individual plants sequenced contained ITS sequence poly- morphism. An investigation of 34 Arctostaphylos taxa, using RFLP data, demonstrated that polymorphism in the ITS region exists in several members of the genus. Collectively, our results indicate 1) the sub- species of A. hookeri are not monophyletic, 2) the current understanding of many species relationships within the genus and the circumscription of subgenera and sections need to be further investigated, and 3) a complex pattern of ITS sequence evolution is suggestive of either hybridization or sorting of ancestral polymorphism. Arctostaphylos (Ericaceae: Arbutoideae) is a large and taxonomically complex genus composed of over 100 taxa (Wells 1991). Geographically, the distribution of Arctostaphylos is circumboreal but most of that distribution is accounted for by only one species, A. uva-ursi (L.) Sprengel. Most species are restricted to the California Floristic Province where approximately half of the taxa are considered rare, threatened, or endangered (Skinner and Pavlik 1994). Species diversity is highest in the coast ranges of California where more than 30 species occur (Fig. 1). Arctostaphylos is considered to have originated in the Miocene, approximately 15 million years ago (Stebbins and Major 1965; Raven and Axelrod 1978). Based on the fossil record, radiation of the genus is considered to have begun during the Pleis- tocene, approximately 1.5 million years ago (Raven and Axelrod 1978). Diversification in the genus has been attributed to a number of abiotic and biotic influences which include: the development of di- verse topography and soils, edaphic restriction (Wells 1993; Raven and Axelrod 1978), the pres- ence of fire regimes characteristic of the California Floristic Province (Raven and Axelrod 1978), dif- ferent life history strategies (Wells 1969; Raven and Axelrod 1978), polyploidy (Stebbins 1980), and hybridization (Shapin 1966; Gottlieb 1968; Kruckeberg 1977; Schierenbeck et al. 1992). Determining evolutionary relationships in Arc- tostaphylos has challenged botanists. Currently, taxa thought to be closely related most often are Classified as subspecies, as within A. glandulosa ' Present address: Jepson Herbarium and Department of Integrative Biology, University of California Berkeley, Berkeley, CA 94720. . * Present address: Harvard University Herbaria, 22 Di- vinity Ave., Cambridge, MA 02138. Eastw., A. stanfordiana C. Parry, A. (Pursh) Lindley, and A. hookeri G. Don. We tested the monophyly of A. hookeri (sensu Wells 1993). This complex of five subspecies was chosen for investigation because the subspecies have been allied with different species in earlier treatments (Table 1). Also, the subspecies of A. hookeri are all limited in distribution and are threat- ened by development. An understanding of their relationships is essential for their conservation. tomentosa An overview of the systematics of Arctostaphylos. Based on taxonomic treatments (Drude 1897; Busch 1952; and Thorne 1992) Arctostaphylos is one of six allied genera in the Ericaceae. These genera (Arctostaphylos, Arbutus, Comarostaphylis, Arctous, Ornithostaphylos, and Xylococcus) have been considered to constitute the tribe Arbuteae within the subfamily Vaccinioideae (Stevens 1971), but recent molecular studies support the placement of these genera as the subfamily Arbutoideae (Cull- ings 1994, 1996; Kron and Chase 1993; Kron 1996): Current understanding of the genus Arctostaphy- los has been enhanced greatly by the work of Jep- son (1922, 1939), Eastwood (1934), Adams (1940), Roof (1976, 1978), and Wells (1968, 1988, 1992, 1993) who undertook systematic studies of the en- tire genus. Several subgeneric groups have been recognized in Arctostaphylos. These groups are re- flected in the classification proposed by Wells (1992). He proposed two subgenera, subgenus Mic- rococcus and subgenus Arctostaphylos. Subgenus Micrococcus, once elevated to generic standing by Eastwood (1937), comprises only 4 taxa. There are three sections in the subgenus Micrococcus, two of which are monotypic. There are three sections in subgenus Arctostaphylos: Foliobracteata, Arcto- staphylos, and Pictobracteata. Arctostaphylos hook- eri is a member of subgenus Arctostaphylos sect. Arctostaphylos. 188 MADRONO 00 bb sass Sennen by e 4 e wieiesetetes [Vol. 45 eee a > 30 species 1-2 species 3-10 species LPs a) Rreseter peteceeee — ~200 miles Fic. 1. distribution which is not shown. Systematics of Arctostaphylos hookeri. Arctosta- phylos hookeri C. Don ssp. hookeri occurs in the hills, dunes, and forests near Monterey Bay. Roof (1980) stated that *‘... efforts to erect an ‘Arcto- staphylos hookeri complex’ seem unwise and im- practical, since the synthesis requires the cross- placing of individuals of the A. pungens Kunth. al- liance with those of the A. uva-ursi alliance.”” Ac- cordingly, in Roof’s (1976, 1980) treatments of the five taxa treated as subspecies of A. hookeri by Wells (1993), none fell within the circumscription of A. hookeri. Taxa that were previously treated as members of the hookeri complex were classified as infraspecific taxa of either A. uwva-ursi or A. pun- gens (Table 1). Arctostaphylos hookeri was allied with A. uva-ursi, as A. uva-ursi ssp. hookeri. Wells (1993) recognized A. hookeri ssp. hookeri, reaf- firming his concept of the A. hookeri complex. Arc- tostaphylos hookeri ssp. hookeri is listed as rare, Distribution of Arctostaphylos taxa in the western United States. Arctostaphylos uva-ursi has a circumboreal threatened, or endangered in California by the Cal- ifornia Native Plant Society (CNPS) (Skinner and Pavlik 1994). Arctostaphylos hookeri ssp. hearstiorum (Hoover and Roof) P. Wells was described as A. hearstiorum by Hoover and Roof (1966). Wells (1968) reduced A. hearstiorum to a subspecies of A. hookeri. Arc- tostaphylos hookeri ssp. hearstiorum is endemic to grassy hills and mesas of the Arroyo de la Cruz area of San Luis Obispo County. It is listed as rare, threatened, or endangered in California by CNPS (Skinner and Pavlik 1994) and is State-listed as en- dangered. Arctostaphylos hookeri G. Don ssp. franciscana (Eastw.) Munz was described as A. franciscana by Eastwood (1905) and was reduced to a subspecies of A. hookeri by Munz (1958). Arctostaphylos hookeri ssp. franciscana was known to occur on serpentine outcrops at three locations in San Fran- 1998] TABLE 1. A SUMMARY OF THE TAXONOMIC HISTORY OF THE A. HOOKERI COMPLEX (SENSU WELLS 1993). A. hookeri ssp. franciscana A. franciscana Eastw. (1905) Uva-ursi franciscana Heller (1914) A. hookeri G. Don ssp. franciscana (Eastw.) Munz (1958) A. uva-ursi (L.) Spreng. var. franciscana (Eastw.) Roof (1980) A. hookeri ssp. hearstiorum A. hearstiorum Hoover & Roof (1966) A. hookeri G. Don ssp. hearstiorum (Hoover & Roof) Wells (1968) A. uva-ursi (L.) Spreng. var. hearstiorum (Hoover & Roof) Roof (1980) A. hookeri ssp. hookeri A. hookeri G. Don (1834) Uva-ursi hookeri (G. Don) Abrams (1914) A. uva-ursi (L.) Spreng. ssp. hookeri (G. Don) Roof (1980) A. hookeri ssp. montana A. montana Eastw. (1897) Uva-ursi montana Abrams (1914) A. pungens HBK var. montana (Eastw.) Munz (1958) A. hookeri G. Don ssp. montana (Eastw.) Wells (1968) A. pungens HBK ssp. montana (Eastw.) Roof (1976) A. hookeri ssp. ravenii A. pungens HBK var. ravenii (Wells) Roof (1976) A. hookeri G. Don ssp. ravenii Wells (1968) cisco (Roof 1976). It is now extinct in the wild and persists only in cultivation. Arctostaphylos hookeri G. Don ssp. ravenii P. Wells was considered extinct in the wild until 1952 when it was rediscovered by Peter Raven at a pre- viously unknown location, a serpentine outcrop in the San Francisco Presidio. Arctostaphylos hookeri ssp. ravenii has been treated as A. pungens var. rav- enii (Roof 1976). Arctostaphylos hookeri ssp. rav- enii is listed as rare, threatened, or endangered in California by CNPS (Skinner and Pavlik 1994) and is also listed as state and federally endangered. Arctostaphylos hookeri G. Don ssp. montana (Eastw.) P. Wells was described as A. montana by Eastwood (1897). Munz (1958) later placed it with- in A. pungens as A. pungens var. montana. In 1968, Wells classified A. pungens var. montana as a sub- species of A. hookeri because of the “‘resemblance ...1n morphology, ecology, and chromosome num- ber”? to A. hookeri ssp. ravenii. Roof (1976) fol- lowed Munz (1958) and recognized A. pungens ssp. montana. In this same publication Roof disagreed with an observation made by Jepson (1939) that A. pungens ssp. montana **... has an affinity with A. hookeri and might be referred to as that species save for their thick leaves,’’ and stated that this tax- on should not be included in the A. hookeri alli- ance. Wells (1993) transferred A. pungens ssp. montana to A. hookeri. Arctostaphylos hookeri ssp. MARKOS ET AL.: PHYLOGENY OF THE ARCTOSTAPHYLOS HOOKERI COMPLEX 189 montana is restricted to serpentine-derived soils on Mount Tamalpais, Marin County. It is listed as rare, threatened, or endangered in California by CNPS (Skinner and Pavlik 1994). MATERIALS AND METHODS Terminal taxa in the phylogenetic analyses. ITS Region.—The ITS region of nine Arctostaphylos taxa was sequenced and subjected to phylogenetic analysis. The taxa included were all five members of the A. hookeri complex, the two taxa to which the subspecies of hookeri have been previously al- lied (A. pungens and A. uva-ursi), and two taxa presumably distantly related to A. hookeri: A. to- mentosa ssp. tomentosa (a member of section Fo- liobracteata) and A. nummularia A. Gray (a mem- ber of subgenus Micrococcus) (Wells 1987). At least two individuals per taxon, from disparate parts of its range, were sequenced for all taxa except A. tomentosa, A. nummularia, and the outgroup (Ar- butus menziesii Pursh) (Table 2). 26S Region.—The taxa included in the 26S study were two members of the A. hookeri complex (ssp. hookeri and ssp. montana), and 14 species repre- senting five sections of Arctostaphylos (Table 2). One individual per taxon was sequenced. In both analyses, the outgroup, Arbutus menzie- sil, was chosen based on results of a recent molec- ular phylogenetic study that placed Arbutus in a basally divergent position in the Arbutoideae (Hil- eman et al. 1994). The choice of outgroup was also well supported by morphological data and fossil ev- idence (Stevens 1971; Wehr and Hopkins 1994; Schorn personal communication). For all taxa sam- pled, taxonomic positions, collection localities, voucher information, and accession numbers are provided in Table 2. Taxa included in the ITS-RFLP analysis. Thirty- four taxa were included in the ITS-RFLP study. These taxa represent all members of section Arc- tostaphylos and select taxa from sections Folio- bracteata and Pictobracteata, and subgenus Micro- coccus (Table 2). DNA extraction. Total DNA’s were isolated from dried leaves of individual plants. DNA extraction followed a simplified Doyle & Doyle CTAB ex- traction (1987). The protocol used is detailed in Cullings (1992). Amplification and sequencing. ITS Region.— Double-stranded PCR products were amplified us- ing the primers c28KJ (Cullings 1992) and “ITS 5”’ (White et al. 1990). The 25 wl reactions con- tained 14.9 wl water, 2.5 wl Tag enzyme buffer, 2.5 wl 25 mM MgCl, 0.475 pl 40 mM dNTP’s, 1.25 wl 10 pm c28KJ, 1.25 pl 10 wm “ITS 5”, 0.125 wl Tag polymerase (5 units/w1), and 2 wl genomic DNA (dilutions of genomic DNA ranged from 1: 10 to 1:1000). Amplification parameters followed —Y y LOS9OLAV ro 2. 61 8901AV LL61604V —«[S6160AV 90890IAV.—- 8ST 890TAV LIS90IAV 91890IAV © = SO890IAV 'Z O ov a < >. ros90IAV bL61604V 6b616041V €08901AV S97 SLI Joquinu UOISSDDNV VO “OD uueth VO “OD BLeIS VO “OD PINnUsaA, VO “OD SHON [9d VO “OD enng WO “0D ewoU0Ss VO “OD NOATASIS WO “0D oostourig ues WO “0D oostourly ues VO “OD uLeY VO “OD ule WO “09 AaraUOp] VO “09 AoraUOp] VO “0D odsiqgQ sinq ueg VO “0D odstgQ sin] urs VO “OD OostoueIy URS VO “0D oosrourly ues VO “OD eHON 19d VO “0D odstqgQ sin] ues VO “OD apisioaty WO “0= sajesuy soy WO “0—= AaraqUOpy VO “0D ewulou0s VO “0D ewoUuO0s VO “OD opesoplg” VO “OD opei0plg VO “OD une VO “OD zZNIp euRS VO— “OD ouls0pusjyq VO “OD oulsopusyq Ayypeso'T COTO AW8d LA e€lcO ANd LA 6810 dLAVAW 6SCO0 AW7d LA cecO AW2d LA 9910 AN2dLA pO80c F4otey “A uinjalogqiy suIqgANS yied [euoissay UspILyL yieg [euoissay Usp 1090 AW29d. LA ‘WS yieg [euoisay UspityL L090 AW2dLA ‘WS wunjalogiy suIgANS €090 AW2d LA ‘WS wuinjalogiy suIqAsS yieg [euoisay UspyLy, 0960 ANVdLA CETO ANd LA 8610 AN2d LA (u's) STTPM d 1900 ANd LA 6900 dLAV AWN 1L00 ON8dLA 0670 dLAV AW 89¢0 dLABAWN OrO00 dLA? AW 86¢0 dLABAW LeEOO0 ANd LA O0LO ANSdLA IOyONoA dUddID “gq Vinind ° uowwWwosy] vunisind ° ABlg “Y Sisuappaau °* DYYNMaU “ASS WRILIZJA VDYYNMAUL ° DJIUDZUDUL ‘dss Auledg ‘dD vilupzunu ° splempy 'S SIsuaysDUY]Y ° SIISM d lluaavs ‘dss u0g ‘DH Mayooy ° STISM d (Miseq) pupjuou ‘dss uog ‘D lMayooy ° luayooy ‘dss uog ‘D Mayooy ° SIISM d (Jooy 2 1dA00H) wnsolsivay “dss UOG ‘DH Mayooy ° zunyjy (‘M4Sse7q) puvosiouvsf ‘dss uogq ‘9 Mayooy ° []aMoH Vjnpidsiy ° Ad[pury vonvjs ° STOMA d SiSuajalqgvs * TSO 'f sSpuntupoa ° Jayeg vuopyisuap ° “MySeY lMayDq ° WPLLIDA, DUDUASSIU ° ARID “VW DiuDinuuunu * SITISM d Sisuaoulz0puau ° UOXPL — SS SS SSS =< < {oo SS @ Vv Vv Vv Vv Vv Vv Vv * xk Vv e * ** Vv e * *K Vv ** Kk Vv * Vv @ Vv Vv Vv Vv Vv ‘olory “1095 ‘OLOUV “OGNS e Vv “SSIN "1995S e@ Vv *K e Vv ‘OLD “1995 OAIN “ONS djdad S9c_ SII ‘NSAS UL poayIsodap aq |[[IM suouToadsg ‘soyreypy toes ‘IAS ‘AaseA [ORY ‘AJ ‘toyed sewmoyl ‘A ALA ‘SuOneIAdIgge IOJDI[OD ‘soleigr] YUeGUSyH 0} Jojo1 sIaquinU UOISsaddV 190 ‘PoyVOIpuUI OsTe St po[duies sjenprArIput oy} JO UOTISOd STWOUOXRL ‘(W) SASATVYNY d THe SLI AONV (@) JONANOAS S9Z (x) JONANOAS SL] YOA GANINVXY SNOLLOATIOD ‘7 ATEVL lee MARKOS ET AL.: PHYLOGENY OF THE ARCTOSTAPHYLOS HOOKERI COMPLEX 1998] 678980AV 8789804 V VO “OD epour[y uopiey otuej0og ‘DN ysing Msaizuaw sningly ® x dno1isjnO SI890IAV ZV “OD vung TETO dLABAW 1ajsuid ‘dss Aled ‘2D 1aj8uiid ‘y Vv 8 “OJOT JOIS pIS9OIAV VO “OD oostourly ues 6600 dLAZV AWN e €TS9OIAV VO “OD Aarauopy €rTO dLAVAWN Ag [pury (ysing) vpsojuawo] “Vy Vv xe VO “OD vieqieg ejueS 9S10 AN®d.LA ulyUueyD “Y Sisuaoisnfas “VY Vv VO “OD eregie | eyues LSIO AWNYdLA psojnpuv]s ‘dss “MISC Dsojnpuv]s Vv Vv VO “OD oulsopus|y 6600 ANdLA Jadig puviquinjoo “y v VO “OD vieqieg ejuRS SSIO ANd LA STIPM d apUuljpipo “Vy v €I890IHV VO “OD zip eyuRg 6L10 dLASAW “MISET SUPISIUDI “VY v e 9L61604V VO “OD ZNID eyes 6800 ANdLA e VO “OD ZMID ees 9800 ANdLA ARID “YW luossapun “Vy v ‘OO 1939S ZTIS9OIAV VO “OD seuin|d pIeO ANd LA Aled ‘D vplysia “y v e 0S61604AV {N “OD uesdQ9 Orr0 dLA xe TT89OLAV VO “OD oar] ues 6100 AW xe LI890ISV IZ8901HAV VO “OD Aarouopy 900 ANdLA Jasoidg (J) Isun-pan “Vy v e se OI890IAV VO “OD vulou0g 8970 dLAZ AWN e VO “oD eden 0660 dLAZAW Aled ‘D vuvipsofuv]s “Vv Vv 608901 V VO “OD eieqieg ejuRS LETO ANBd LA ® VO “OD odstqo sinq ues ISTO AN®d.LA ‘ISOIM 2 uosdar sipns “y Vv VO “OD oulpieuiog urs p610 AN x 80890ISV—s-: OT890IAV VO “OD o8aIq ues yieg [euoIsay Uep[lL yuny suasund -y v e x VO “0D Aarauoyy pS00 AN2dLA ‘NN Yj1wund “y Vv S97 SLI AV[eo0'T JayonoA, UOXeL didu soz SLI Joquinu UOISSIDNV “GHNNILNOZ) °C AIaV EL 192 MADRONO Cullings (1992). Amplified products were stored in 1X LTE. Asymmetric PCR using double-stranded tem- plate was performed with various combinations of the following primers: “ITS 2”, “ITS 3”, “ITS 4’, “ITS 5”° (White et al. 1990). The 50 wl reactions contained 5.8 pl water, 5.0 wl Tag enzyme buffer, 5.0 pl 25 mM MgCl, 2.0 wl 50% glycerol, 1.0 wl 100% DMSO, 0.950 wl 40 mM dNTP’s, 2.5 pl primer #1 (10 pM), 2.5 wl primer #2 (diluted to proper ratio), 0.25 wl Tag polymerase (5 units/wl), and 25 wl double-stranded template taken from 1X LTE solution. Amplification parameters followed Cullings (1992). Single-stranded DNA _ products were purified using spin columns. Columns were loaded with G-50 sephadex equilibrated with 1X STE (Sambrook et al. 1989). Both DNA strands were sequenced by the di- deoxy method using the Sequenase dGTP kit (U.S. Biochemical). **S-dATP was used for isotopic la- beling. The limiting primers in the single-stranded amplification were used as the sequencing primers. Samples were electrophoresed on an 8% polyacryl- amide gel. Gels were fixed with a solution of 10% methanol and 10% acetic acid for 30 min, trans- ferred to 3 MM Whatman paper, and vacuum dried at 80°C for 40 min. Gels were exposed to auto- radiographic film for at least 12 hours. 26S Region.—Double-stranded PCR _ products were amplified using 28KJ (Cullings 1992) and ei- ther the universal primer 28C (Hamby and Zimmer 1988) or 28B (Hamby et al. 1988). The 50yl re- actions contained 1X Promega PCR _ buffer (M190A), 0.75 mM dNTPs, 0.15—-0.5 wM each primer, 1.0—2.5 mM MgCl, 1.25 units Promega Taq, and 5 wl extracted DNA (dilutions ranged from 1:10 to 1:10,000). Cycling conditions con- sisted of a 3 min. denaturation at 94°C, followed by 40-44 cycles of a 20 sec denaturation at 95°C, a 45 sec annealing at 55°C and a 1.5—2.0 min ex- tension at 72°C, and a final extension at 72°C for 7 min. Amplified products were cleaned using the QIAGEN PCR product kit. Cleaned products were quantified by comparison to Hindlll digested lamda on a 1.5% standard agarose gel. Between 25 and 30 ng/pl of amplified product were subjected to 35 rounds of cycle sequencing using either the 28KJ primer or the 28B primer. Dye terminator chemistry was used according to manufacturers specifications with an annealing tem- perature of 51°C. Sequenced products were run on the ABI 377 automated sequencing system accord- ing to manufacturer specifications. Sequence analysis. ITS Region.—ITS sequences were aligned with those of Brassica napus L. (Oku- mura et al. 1992) and Daucus carota L. (Yokota et al. 1989) to determine the boundaries of the coding and spacer regions. DNA sequences were aligned manually using the DNA alignment program MacDNASIS (1994). Alignment of the ingroup and [Vol. 45 outgroup taxa required the introduction of gaps to accommodate five indels. All five indels occurred in ITS 1. The placement and length of the indels was unambiguous. The indels did not vary within the ingroup and were entered into the data matrix as gaps. 26S Region.—The 26S fragment corresponds to bp positions 334-616 of the 26S region determined by unambiguous alignment to Fragaria ananassa Ducheshe (GenBank accession # X58118) and Cit- rus limon (L.) Burm. f. (GenBank accession # X05910) Phylogenetic analysis. ITS Region.—ITS data were analyzed by Fitch parsimony as implemented in PAUP version 3.1.1 (Swofford 1993) using the branch-and-bound procedure to find all maximally parsimonious trees. All character-state changes were weighted equally. Bootstrap values were cal- culated from 100 replicate parsimony analyses (Fel- senstein 1985) using PAUP heuristic searches, sim- ple taxon addition sequence, TBR _ branch-swap- ping, and MULPARS. Due to the number of polymorphic taxa in the data set, it was not practical to exclude all taxa with ITS polymorphism from the analyses. We therefore chose to conduct two analyses that differed in their treatment of polymorphism in the ITS sequence data. In the first analysis polymorphic sites were coded using the [UPAC-IUB ambiguity codes and multiple states were recognized as polymorphic rather than uncertain in PAUP. In the second anal- ysis data were recoded from DNA data to multistate characters (Campbell et al. 1997). For example A = 1,G = 2 A/G = 3. 26S Region.—As with the ITS sequence data two analyses were conducted to account for 26S sequence polymorphism. In analysis one, polymor- phic sites were coded using the I[UPAC-IUB am- biguity codes and multiple states were recognized as polymorphic rather than uncertain in PAUP. In analysis two, data were recoded from DNA data to multistate characters (Campbell et al. 1997). In both 26S analyses data were analyzed by Fitch parsimony using PAUP version 3.1.1 (Swofford 1993). Heuristic searches were conducted with 10 replicates of random addition sequence, TBR branch-swapping, and MULPARS in effect to find maximally parsimonious trees. All character-state changes were weighted equally. Bootstrap values were calculated from 1000 replicate parsimony analyses (Felsenstein 1985) using PAUP heuristic searches, random starting trees, and TBR branch- swapping. Restriction digests of the ITS region. The ITS region was PCR amplified as described above. The approximately 700 bp fragment was subjected to digestion by two restriction endonucleases: Alu I and Hha |. These restriction endonucleases were chosen following a survey for phylogenetically in- 1998] formative restriction sites in the ITS sequences. Each reaction digest contained 4 yl double-stranded PCR products, 4 wl water, 1 wl 10X buffer, and | wl restriction endonuclease (diluted to ca. 7.5 units/ wl). Reactions were left at the manufacturer’s sug- gested incubation temperature for at least 16 h. Di- gested DNA was electrophoresed on 3% agarose gels using 1X TAE as the gel buffer. RESULTS Sequence analysis. ITS Region.—In the study taxa, ITS 1 was 253 bp long ITS 2 was 226 bp long, and the 5.8S was 164 bp long. These findings are similar to those found in other ITS studies of angiosperms (Baldwin et al. 1995). Of the aligned positions in ITS 1, ten sites (4.0%) were variable within the ingroup, seven were potentially infor- mative phylogenetically. In ITS 2, two sites (0.88%) were variable and both were potentially informative phylogenetically. There were no vari- able sites in the 5.8S subunit (Table 4). Polymor- phism within an individual, seen as two or more nucleotide states on the autoradiograph, was found to varying degrees throughout the taxa sequenced. Some taxa (e.g., A. hookeri ssp. hookeri) did not exhibit any polymorphic sites. Other taxa (e.g., A. pungens) were polymorphic at almost all variable sites (Table 4). At least two individuals per taxon were sequenced (excluding A. tomentosa, A. num- mularia, and the outgroup). Variation among indi- viduals of a single taxon was noted only in A. uva- ursi. All sequences obtained were included in the analysis (Table 2). 26S Region.—Two hundred and eighty-one bp of the 26S were sequenced and aligned. Of those 281 bp, 4 (1.4%) were phylogenetically informative (Table 4). Polymorphism was detected in three taxa, A. pungens and A. stanfordiana, and Arbutus men- ziesil. Phylogenetic analysis. IVS Region.—In analysis one (multiple states at a position recognized as polymorphism), nine equally parsimonious trees were generated using a branch-and-bound search (with furthest taxon addition sequence). Each of the nine trees required 100 evolutionary steps and has a CI (excluding uninformative characters) of 0.595. The nine maximally parsimonious trees differ to- pologically in their placement of one taxon, A. pun- gens. Arctostaphylos pungens is polymorphic at 7 of 9 phylogenetically informative sites, so it is not surprising that its phylogenetic position is unresol- ved. When A. pungens was removed and the anal- ysis was conducted as described above, the posi- tions of the remaining taxa relative to each other did not change. One maximally parsimonious tree was generated (Fig. 2A). This tree required 93 evo- lutionary steps and has a CI (excluding uninfor- mative characters) of 0.643. The analyses were also conducted with exclusion of the outgroup, Arbutus MARKOS ET AL.: PHYLOGENY OF THE ARCTOSTAPHYLOS HOOKERI COMPLEX 193 menziesii. The topology of the ingroup did not change with the omission of A. menziesii. Analysis two (multiple states at a site rescored as a novel multistate characteristic) generated one tree of 96 steps with a CI (excluding uninformative characters) of 0.905 (Fig. 2B). The relative topol- ogy is almost the same as in analysis one, the only difference being the placement of A. pungens. In analysis one, when A. pungens was included it fell out in several places in the tree but in analysis two it is nested within the clade containing A. hookeri ssp. franciscana, A. hookeri ssp. montana, and A. hookeri ssp. ravenii. In both analyses, the low amount of sequence divergence within the ingroup resulted in a lack of phylogenetic resolution and weak support of some clades. Despite these limitations, the results of the two analyses are nearly congruent (the only differ- ence in topology is accounted for by A. pungens) and some conclusions can be drawn. Two mono- phyletic clades were resolved in Arctostaphylos. One clade includes A. hookeri ssp. hookeri, A. hookeri ssp. hearstiorum, A. uva-ursi from New Jersey, and A. nummularia. The other clade con- tains A. hookeri ssp. raventi, A. hookeri ssp. mon- tana, A. hookeri ssp. franciscana, (A. pungens in analysis 2) A. uva-ursi from California, and A. to- mentosa. These data suggest that A. hookeri (sensu Wells 1993) is non-monophyletic, with subspecies found in two distinct clades that each include some representatives of other species in the genus. In ad- dition, the widespread and variable A. wva-ursi also appears to be non-monophyletic. 26S Region.—Analysis one of the 26S sequence data (multiple states at a position interpreted as polymorphism) resulted in eight equally parsimo- nious trees. Each of the eight trees required 34 evo- lutionary steps and has a CI (excluding uninform- ative characters) of 0.537. Analysis two (multiple states at a site rescored as a novel multistate char- acteristic) resulted in three most parsimonious trees which required 24 evolutionary steps. The consis- tency index of each tree (excluding uninformative characters) was 1.00. The topologies of the consen- sus trees from the two analyses were identical and discussions therefore will be restricted to this com- mon topology. One of the eight trees from analysis one was identical to the strict consensus trees (Fig. 3). Two distinct, moderately supported clades of Arctostaphylos taxa are present. Although there is no resolution among taxa within the two clades, some important results emerge: 1) A. hookeri is non-monophyletic, A. hookeri ssp. hookeri and A. hookeri ssp. montana fall into two distinct clades; 2) the results suggest that the subgeneric classifi- cation constructed by Wells (1992) does not rep- resent phylogenetic relationships of the taxa, sec- tion Arctostaphylos is paraphyletic; 3) the two clades reconstructed using the 26S sequence data 194 84 ot St bb bh bh bh) ee 2 1 73 MADRONO Arctostaphylos Arctostaphylos Arctostaphylos Arctostaphylos Arctostaphylos Arctostaphylos Arctostaphylos Arctostaphylos Arctostaphylos Arctostaphylos h. ssp. franciscana h. ssp. montana h. ssp. ravenii tomentosa uva-ursi; MNT Co. uva-ursi; SMT Co. h. ssp. hearstiorum h. ssp. hookeri nummularia uva-ursi; NJ Arbutus menziesii Arctostaphylos Arctostaphylos Arctostaphylos Arctostaphylos Arctostaphylos Arctostaphylos Arctostaphylos Arctostaphylos Arctostaphylos Arctostaphylos Arctostaphylos pungens h. ssp. franciscana h. ssp. montana h. ssp. ravenil tomentosa uva-ursi; MNT Co. uva-ursl1; SMT Co. h. ssp. hearstiorum h. ssp. hookeri nummularia uva-ursi; NJ Arbutus menziesii [Vol. 45 RRRLRBAAT Subgenus Micrococcus, Section Micrococcus Subgenus Arctostaphylos, Section Arctostaphylos Subgenus Arctostaphylos, Section Foliobracteata FiG. 2. The single most parsimonious trees produced from each of two analyses of data from the ITS region. Numbers above the lines represent the branch lengths and numbers below the lines represent bootstrap values. A) Analysis one, multiple states at a position recognized as polymorphic. Length = 93 steps. CI excluding uninformative characters = 0.643. RI = 0.905. RC = 0.885. Arctostaphylos pungens was not included in the analysis. B) Analysis two, multiple states at a position rescored into a novel multistate characteristic. Length = 96 steps. CI excluding uninformative characters = 0.905. RI = 0.920. RC = 0.901. 1998] MARKOS ET AL.: PHYLOGENY OF THE ARCTOSTAPHYLOS HOOKERI COMPLEX 195 Arctostaphylos stanfordiana Arctostaphylos rudis Arctostaphylos pungens Q Arctostaphylos patula a ie rom 7: 7 Arctostaphylos nummularia 3 POLE EAE ERY RET NELTESEEER Arctostaphylos mendocinoensis g Arctostaphylos h. ssp. hookeri Arctostaphylos viscida Arctostaphylos uva-ursi; MNT Co. by om Arctostaphylos tomentosa P Arctostaphylos pringlel & 59 Arctostaphylos nissenana ) Arctostaphylos h. ssp. montana i = = = = Arctostaphylos glauca Sooo Arctostaphylos canescens por Arctostaphylos andersonil Arbutus menziesii Subgenus Micrococcus, Section Micrococcus c= Subgenus Micrococcus, Section Nissenana mums «= SUbgenus Arctostaphylos, Section Arctostaphylos morn «=Subgenus Arctostaphylos, Section Foliobracteata mums «= SUDgenus Arctostaphylos, Section Pictobracteata Fic. 3. One of the 8 most parsimonious trees produced in analysis one (multiple states at a position recognized as polymorphic) using data from the 26S region. The topology is identical to the strict consensus trees of both analysis one and analysis two of the 26S data. Arbutus menziesii was used as the outgroup. Bootstrap values are given below the lines. Length = 34 steps. CI excluding uninformative characters = 0.537. RI = 1.00. RC = 1.00. correspond well to the two groups detected with the ITS-RFLP data (Fig. 3, Table 3). ITS-RFLP results. The ITS region from 34 taxa was analyzed with 2 restriction endonucleases. Combining the data from the endonucleases, each taxon had one of three primary pattern types: Group One, Group Two, and Mosaic (Table 3). Taxa were placed in Group One if the Alu I site was absent and the Hha I site was either present or polymorphic. With one exception, taxa were placed in Group Two if the Hha I site was absent and the Alu I site was either present or polymorphic. Taxa were placed in the Mosaic Group if the recognition site for both Alu I and Hha I was absent. Only one taxon, A. pungens, was polymorphic at both the Alu I and Hha I sites and it was placed in Group Two based on data from analysis of the 26S sequence data. DISCUSSION The results of this study suggest that A. hookeri sensu Wells (1993) combines taxa that do not con- stitute a natural group. The five subspecies of A. hookeri sensu Wells are members of two distinct ITS lineages (Fig. 2). One lineage includes the three subspecies which occur in the San Francisco Bay Area on serpentine soil: franciscana, montana, and ravenii. Also included in this lineage are A. tomentosa and A. uva-ursi (from California). The second lineage consists of two subspecies, hookeri and hearstiorum, which occur along the central coast of California and A. nummularia and A. uva- 196 TABLE 3. MADRONO [Vol. 45 NUCLEAR RIBOSOMAL DNA RESTRICTION SITE MUTATIONS OF 34 ARCTOSTAPHYLOS TAXA. Columns represent subgeneric taxa recognized by Wells (1992). The presence (1), absence (0), or polymorphic state (*) of restriction sites are shown. Subgenus Micrococcus Sect. Arctostaphylos A H A Group | . edmundsit . glauca . h. montana h. ravenii . klamathensis . manzanita . pumila . uva-ursi . viscida A. nissenana O l . bakeri . gabrielensis . h. franciscana . nevadensis te te et et et et ee ad SoC Coo oe ooo oc eg Group 2 _h. hearstiorum hh. hookeri A. mendocinoensis — * 0) A. nummularia = 0) . hispidula . patula . Stanfordiana | . densiflora = Kk > K . pungens . >Srdb SS Sb Mosaic . mewukka . parryanna . rudis >> > es Kem) ursi (from New Jersey). At this time, the phyloge- netic position of A. pungens is unresolved. The ITS-RFLP data set also provides support for the presence of two distinct lineages in Arctosta- phylos (Table 3). The two groups represented in the ITS-RFLP study correspond to the two clades seen in both trees based on the ITS sequence data (Fig. 2) and the 26S data (Fig. 3). These two groups were not represented in the subgeneric classification pro- posed by Wells (1992). Section Foliobracteata is the only section with members restricted to one group while members of section Arctostaphylos and subgenus Micrococcus are disassociated and occur in both primary groups. Within-individual polymorphism was detected in the ITS region for several of the taxa examined. Polymorphism was seen in both the ITS sequence data (Table 4A) and in the ITS-RFLP study (Table 3). Although widespread, the presence of polymor- phism does not seem to have dramatically altered the topology of the ITS tree (Fig. 2) or the inter- pretation of the ITS-RFLP data. With regard to the ITS sequence data, the position of only one taxon, A. pungens, changes with respect to the type of analysis done. All other taxa remain in the same relative phylogenetic position. In the ITS-RFLP Subgenus Arctostaphylos Sect. Foliobracteata Sect. Pictobracteata H A. i A H l A. andersonii 0) i A. pringlei O 1 l A. columbiana O 1 l A. refugioensis O ] ] A. tomentosa 0) l l | 1 l 1 me A. canescens 0) * a A. catalineae 0 ig ss A. glandulosa 0) k 0) 0) 0) 0) 0) 0) ok 0) O 0) study, with the exception of one taxon, the two pri- mary groups seen (Group One and Group Two) (Table 3) correspond to the two lineages detected with both ITS and 26S sequence data (Figs. 2—4). The 26S sequence data (Fig. 3) corroborate the finding that A. hookeri is not monophyletic. Al- though only two taxa of the A hookeri complex were sampled for 26S data, moderate levels of bootstrap support were found for the placement of A. hookeri ssp. hookeri and A. hookeri ssp. mon- tana in two distinct clades each including represen- tatives of other species of Arctostaphylos. Despite low levels of resolution, subgenus Arctostaphylos appears to be non-monophyletic, members are found in each of the two 26S clades (Fig. 3). The concordance between the ITS data and 26S data is not surprising given that the ITS region and 26S region are both part of the nrDNA 18S—26S repeat. The value of the 26S data to our study is not to provide an independent (un-linked) source of phy- logenetic evidence, but to augment the limited number of variable ITS characters most of which are complicated by the presence of within-individ- ual polymorphism—the 26S data provides more support for the two clades observed with the ITS data. Collectively, both lines of nrDNA data pro- 1o7 MARKOS ET AL.: PHYLOGENY OF THE ARCTOSTAPHYLOS HOOKERI COMPLEX 1998] LOLLYOOODDLWOWOVLLLDOD epTOsSTA'Y LOLLWOOODODLYNWOWVLLLDOD = «860XLNW ‘TSan eAN'Y LOLLYOOODDLYOVOWLELDOS eSOqUSsUIOR "WV LOLILOIDIOVLYLYOWLLLMD euetTprzozue As’ WV LOLLLOOOLY LW LYOWLLELOD stpni‘w LO LLMODDOMLWAWOWLLLYOS suebund’ wy LODLYOOOOOLV OV OWL TeTbutazd’ wy LOLLLODDOVLVLWOWLLLDOD * ep nqed-“y LOLLYWOOD LWOWOWLLLDOD BPUeUSSSTU WY LOLLLOODOWLVWLYOWLLLOOD +4 STSUSOUTOIOPUSU’ ¥ LO LLWNODDOLWOWOWLLLDOD eure qUOU YY LO LLLOOLIVLVLYWOWLLLOON Tisyooy UW LOLLYODDDIDLWOWOWLLLDOD eoneTb “wv LO LLWOODDOLWOWOWLLLDOD susoseueo "Vv LO LLWOODDDLWOWOWLLLDOD TTUOSITepue VY AYODALOADOOANODWUMAADAD TTSeTzuewu snqnqsy EBTTOVSOLCELTVIEB9SIELT LOOIOGPPETTOOSEEEETIC COCCCCCCOGCECTIELLILE * ¥ % ¥ “da LIVLIWWWLVLOLLILLLLLWOLLIELLLYOOLWLLOVOLLLWYOOLLOILOVOLOWOOOLWLLWYLLWLLLOWLOOOLLOLOAW ON n-n°¥ LIVLYWWLWLOLLLELLLWOLLLLIGLYOOOLVLLOWOLLLWWOOLLOLOWOLWWOLOLWLLVWLLWYLLLOVLLINOL-LOLOLY LWS*n-n' wy LIVLIWWWLVLOLLELLLLVOLLLLLLLWDOOLSLLODOLLLWYOO LLOLOVOLWWOLOLVILWYLLWILLOWLIVOLLOLOLY LNW n-n*W LIVLIVWWLVLOLLLLELLLWOLLLLLLLYODLILLOOOLLLWWOOLLOLS WOLVWOLNLY LIV LLY LLLOVLINOLLOLOOW eSOUSUIO} *¥ LIVLIWWWLWLOLLLLLLLWOLLLLLLLWOOLVLLOVOLLLWYOOLLOILOVOLOWOOOLVLLOVLLWILLOWLOOOLISLOOW =—- B FAP TNUUMU’ LIVLVWWLWLOLLILELLLWOLLLLILLWDOOLYLLOYOLLLWWOOLLILOWOLNWOOOLWLLUVLLWLLLOWLAAWLLOLOOW suebund‘y LIVLVWWLVLOLLILILLLWOLLLLLILWOOLILLOOOLLLWWOOLLOLOWOLWWOOOLWLLVWLLWLLLOWLLOVLLOLOOW TTUSAPI WY LIVLWWWLVLOLLIoLLLLWOLLILLLLWOOOLOLLDDOLLLWYOOD LLOLOWOLWWOOOLVLLVYLLWLILLOWLLOWLLOLOOW eue UOU’ LIVLWWWLWLOLLILILLLWOLLLLLLLWDOOLWLLOWOLLLWWOOLLOLOVOLOWOOOLVLLOVLLVLLLOWLOOOLLOLLOW TIEYZOOY 'W LIVLIWWWLVLOLLIEILLLWOLLLLLLLWDOOLWVLLOWOLLLWYOOLLOLOVOLIWOOOLWLLOVLLWILLOWLOIOOLLOLOOW "Asresy WV LIVLIVWWLVLO LLL ibLLLWOLLLLLILYOOOLYLLOYOLLLWYOOLLOLOVOLNWOOOLWLLVWLLWILLOWLLOWLLSLOOW «= PURBOSTOUPTZ Y 99D9DN999999NDIIIIOO/ I/. LIDIDIODO WO WOW LIODOO ND DD IDL WOO LIIOLIOIODOOWDOIDIIODDOPDLLIDOWIIOD ‘ZueU WY aa NNR Dae ek Na la al nen a a a ree et ee 68S TO68SPO668T8LESPELELZOOPZPPISTEPOLSPBEBEBIDUETB8ICLOEBSEIEDTOB8IES6ESLEBTONE9VCITSED ELELETTCCT T0006 668888LLLLOGVETTL0899SSSPPEELEGEB0DTPEEZOONDPTEECCZTIOLL999STVVVVECCT 99999999999999GS5SS5S55SSSSSSSSSSSPPPPPDPPPPVEECEEETCCCCCCCCLUITIIIIII * x x * *% XH *¥ x WY ee a a rr Ee Ee ‘(,) AQ PoyeoIpuUT ore SOUS BATFEULIOJUT A]JeonouasofAyd “D/O =S‘VV = M‘11/9 = MN‘ = A ‘DIV = UM ‘O/V = WD ‘LL ‘D ‘WV StoquiAg ‘NoIsay S97 SHL (aq GNV Noosa SLI FHL (VW NI SNOLLISOg AGILOATONN ATAVAVA ‘fP ATAVL 198 vide strong evidence for the existence of at least two, previously unrecognized lineages in Arcto- staphylos. Despite the apparent phylogenetic pattern recon- structed using nrDNA data, a history of lineage sorting or hybridization may complicate the inter- pretation of apparent phylogenetic signal (Avise 1989; Rieseberg 1991; Doyle 1992). The polymor- phism seen in the ITS data is suggestive of either hybridization or sorting of ancestral polymorphism. Some authors have interpreted patterns such as those seen in the ITS RFLP data as evidence of reticulate evolution (Sang et al. 1995; Rieseberg et al. 1990; Rieseberg 1991; Rieseberg and Brunsfeld 1992). Ellstrand et al. (1987) and Schierenbeck et al. (1992) have provided empirical evidence of in- trogression and hybridization between Arctostaphy- los taxa. Others have also postulated that first gen- eration hybrids, hybrid swarms, and taxa of hybrid origin exist in the genus (Dobzhansky 1953; Got- tlieb 1968; Keeley 1976; Kruckeberg 1977). How- ever, at this time we can not distinguish between patterns of reticulate evolution and lineage sorting in the taxa examined. This is the first study to address phylogenetic re- lationships among Arctostaphylos taxa using mo- lecular data. If the results presented here are indic- ative of the phylogenetic complexity of the genus as a whole, then uncovering phylogenetic patterns in the group may prove difficult. Furthermore, until phylogenetic analyses that include all of the rec- ognized taxa are complete, taxonomic decisions re- garding the circumscription of subgenera, sections, and species must be considered tenuous at best. A complete phylogenetic investigation is in progress that should help to build a better understanding of patterns of evolution in Arctostaphylos. ACKNOWLEDGMENTS A portion of this work was submitted by the first author to the faculty of San Francisco State University in partial fulfillment of the requirements for the degree of Master of Arts. The work was completed in the Conservation Ge- netics Laboratory, Department of Biology, San Francisco State University. SM would like to thank the members of her thesis committee: V. Thomas Parker, Robert W. Pat- terson, Cristian Orrego and Bruce G. Baldwin. Each pro- vided invaluable guidance and encouragement. Also, thanks to Ken Cullings for providing the primer c28KJ and many amplification protocols. Financial support was provided by the California Native Plant Society and the Hardman Foundation. The authors thank Robert W. Pat- terson and Bruce G. Baldwin for carefully reviewing the manuscript and providing many helpful comments. LITERATURE CITED ABRAMS, L. 1914. Genus Uva-Ursi. North American Flora 29(1):92-101. ADAMS, J. E. 1940. A systematic study of the genus Arc- tostaphylos. Journal of the Elisha Mitchell Society 56: 1-62. AVISE, J. C. 1989. Gene trees and organismal histories: A MADRONO [Vol. 45 phylogenetic approach to population biology. Evolu- tion 43:1192—1208. BALDWIN, B. G., M. J. SANDERSON, J. M. PORTER, M. FE WOICIECHOWSKI, C. S. CAMPBELL, AND M. J. DONo- GHUE. 1995. The ITS region of nuclear ribosomal DNA: A valuable source of evidence on angiosperm phylogeny. Annals of the Missouri Botanic Garden 82:247-277. Buscu. 1952. In Shishkin and Bobrov. Fl. U. R. S. S. 18: 79. CAMPBELL, C. S., M. J. WOJCIECHOWSKI, B. G. BALDWIN, L. A. ALICE, AND M. J. DONOGHUE. 1997. Persistent nuclear ribosomal DNA sequence polymorphism in the Amelanchier agamic complex (Rosaceae). Molec- ular Biology and Evolution 14(1):81—90. CULLINGS, K. W. 1992. Design and testing of a plant-spe- cific PCR primer for ecological and evolutionary studies. Molecular Ecology 1:233—240. . 1994. Molecular phylogeny of the Monotropoi- deae (Ericaceae) with a note on the placement of the Pyroloideae. Journal of Evolutionary Biology 7(4): 501-516. . 1996. Single phylogenetic origin of ericoid my- corrhizae within Ericaceae. Canadian Journal of Bot- any 74(12):1896—1909. DOBZHANSKY, T. 1953. Natural hybrids of two species of Arctostaphylos in the Yosemite region of California. Heredity. 7:73-79. Don, G. 1834. General Systematics 3:836. DoyYLE, J. J. 1992. Gene trees and species trees: Molecular systematics as one-character taxonomy. Systematic Botany 17:144—-163. Doy Le, J. J. AND J. L. DOYLE. 1987. A rapid DNA isola- tion procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19:11-—15. DRupDE, O. 1897. Ericaceae, in Engler and Prantl, Die na- turlichen Pflanzenfamilien, 4,1. Leipzig: Engelmann. Eastwoop, A. 1897. Proceedings of the California Acad- emy of Sciences Series 3 Bot. 1:83. . 1905. New species of Western Plants. Bulletin of the Torrey Botanical Club 32:193-—218. . 1934. A revision of Arctostaphylos with key de- scriptions. Leaflets of Western Botany 1:105—127. . 1937. Notes on Schizococcus with a key to the species. Leaflets of Western Botany II:49—64. ELLSTRAND, N. C., J. M. LEE, AND J. E. KEELEY. 1987. Ecological isolation and introgression: Biochemical confirmation of introgression in an Arctostaphylos (Ericaceae) population. Acta Oecologica 8:299—308. FELSENSTEIN, J. 1985. Confidence limits of phylogenies: An approach using the bootstrap. Evolution 39:783-— TON. GOTTLIEB, L. 1968. Hybridization between Arctostaphylos viscida and A. canescens in Oregon. Brittonia 20:83— O38: Hampy, R. K., L. Sims, L. ISSEL, AND E. A. ZIMMER. 1988. Direct ribosomal RNA sequencing: Optimization of extraction and sequencing methods for work with higher plants. Technical report, Plant Molecular Bi- ology Reporter 6:179—197. Hamsy, R. K. AND E. A. ZIMMER. 1988. Ribosomal RNA sequences for inferring phylogeny within the grass family (Poaceae). Plant Systematics and Evolution 160:29—37. HELLER. 1914. Catalog of North American Plants 3:276. HILEMAN, L. C., V. T. PARKER, M. C. VASEy. 1994. Pre- liminary generic relationships of the Arbuteae (Eri- 1998] caceae) based on molecular sequence data. American Journal of Botany 81(supplement, abstract):161. Hoover, R. E AND J. B. Roor. 1966. Two new shrubs from San Luis Obispo County, California. The Four Seasons 2(1):2. JEPSON, W. L. 1922. Revision of California Arctostaphyli. Madrono 1:78-—86. . 1939. A Flora of California, Vol. HII. Associated Students Store, Berkeley. CA. KEELEY, J. E. 1976. Morphological evidence of hybridiza- tion between Arctostaphylos glauca and A. pungens (Ericaceae). Madrono 23:427—434. Kron, K. 1996. Phylogenetic relationships of Empetra- ceae, Epacridaceae, Ericaceae, Monotropaceae, and Pyrolaceae: Evidence from nuclear ribosomal 18s se- quence data. Annals of Botany 77:293-—303. KRON, K. AND M. CHASE. 1993. Systematics of the Eri- caceae, Empetraceae, and Epacridaceae and related taxa based on rbcL sequence data. Annals of the Mis- souri Botanical Garden 80:735-—741. KRUCKERBERG, A. R. 1977. Manzanita (Arctostaphylos) hybrids in the Pacific Northwest: Effects of human and natural disturbance. Systematic Botany 2:233— 250. MaAcDnasis v. 3; copy. May 1994. Hitachi Software En- gineering Co., Ltd.; 1111 Bayhill Dr., Suite 395, San Bruno, CA 94066. Muwz, P. A. 1958. California miscellany. Aliso 4(1):95. . 1968. Arctostaphylos. Pp. 80—82 in Munz, P. A., Supplement to A California Flora. U.C. Press, Berke- ley, CA. OKUMURA, S. AND H. SHIMADA. 1992. Nucleotide sequenc- es of genes for a 5.8S and 25S rRNA from rapeseed (Brassica napuus). Nucleic Acids Research 20:3510. RAVEN, P. H. 1952. Plant notes from San Francisco, Cal- ifornia. Leaflets of Western Botany 6(11):208. AND D. I. AXELROD. 1978. Origin and relationships of the California Flora. U.C. Publications in Botany 72:1-134. RIESEBERG, L. H. 1991. Homoploid reticulate evolution in Helianthus (Asteraceae): Evidence from ribosomal genes. American Journal of Botany 78(9):1218—1237. AND S. J. BRUNSFELD. 1992. Molecular evidence and plant introgression,151—176. in P. S. Soltis, D. E. Soltis, and J. Doyle (eds.), Molecular Systematics of Plants. Chapman and Hall. New York, NY. , R. CARTER, AND S. ZONA. 1990. Molecular tests of the hypothesized hybrid origin of two diploid He- lianthus species (Asteraceae). Evolution 44:1498— Sal. Roor, J. B. 1976. A fresh approach to the genus Arctos- taphylos in California. The Four Seasons 5:2—24. . 1978. Studies in Arctostaphylos (Ericaceae). The Four Seasons 5:3-15. . 1980. California’s Arctostaphylos uva-ursi alli- ance. The Changing Seasons 1(2):19—26. SAMBROOK, J., E. E FRItsCcH, AND T. MANIATIS. 1989. in C. Nolan (ed.), Molecular cloning: A laboratory man- ual, second edition. Cold Spring Harbor Press, Plain- view, NY. SANG, T., D. J. CRAWFORD, AND T. FE Stugssy. 1995. Doc- umentation of reticulate evolution in peonies (Paeon- ia) using internal transcribed spacer sequences of nu- MARKOS ET AL.: PHYLOGENY OF THE ARCTOSTAPHYLOS HOOKERI COMPLEX 199 clear ribosomal DNA: Implications for biogeography and concerted evolution. Proceedings of The National Academy of Sciences 92:6813—-6817. SCHIERENBECK, K. A., G. L. STEBBINS, AND R. W. PATTER- SON. 1992. Morphological and cytological evidence for polyphyletic allopolyploidy in Arctostaphylos me- wukka (Ericaceae). Plant Systematics and Evolution 179:187—205. SHAPIN, S. 1966. Variation and hybridization in three Pa- cific Northwest taxa of Arctostaphylos. Senior thesis, Reed College, Portland, OR. SKINNER, M. W. AND B. M. PAvLIK. 1994. Inventory of Rare and Endangered Vascular Plants of California. California Native Plant Society Special Publication No. |, Fifth Edition. Sacramento, CA. STEBBINS, G. L. AND J. MAJor. 1965. Endemism and spe- ciation in the California flora. Ecological Monographs 35:1-35. . 1980. Polyploidy in plants: Unresolved problems and prospects. Pp. 495-520. in W. H. Lewis (ed.), Polyploidy: Biological relevance. Plenum Press New York. STEVENS, P. 1971. A classification of the Ericaceae: Sub- families and tribes. Botanical Journal of The Linnean Society. 64:1—53. SWOFFORD, D. L. 1993. PAUP - Phylogenetic Analysis Using Parsimony - version 3.1.1 Computer program distributed by the Illinois Natural History Survey, Champaign, IL. THORNE, R. 1992. Classification and geography of the flowering plants. Botanical Review. 58:225—348. WER, W. C., AND D. Q. HOPKINS. 1994. The Eocene or- chards and gardens of Republic, Washington. Wash- ington Geology 22:27—34. WELLS, P. V. 1968. New taxa, combinations, and chro- mosome numbers in Arctostaphylos (Ericaceae). Ma- drono 19:193-—210. . 1969. The relation between mode of reproduction and the extent of speciation in woody genera of the California chaparral. Evolution 23:264—267. . 1987. The leafy-bracted, crown sprouting man- zanitas, an ancestral group in Arctostaphylos (Erica- ceae). The Four Seasons 7:5—27. . 1988. New combinations in Arctostaphylos (Eri- caceae): annotated list of changes in status. Madrono 35:330-341. . 1991. The naming of the manzanitas. The Four Seasons 8:46—70. . 1992. Subgenera and sections of Arctostaphylos. The Four Seasons 9:64—69. . 1993. Arctostaphylos. Pp. 545-559 in J. C. Hick- man (ed.), The Jepson Manual: Higher plants of Cal- ifornia. University of California Press. Berkeley, CA. WHITE, T. J., S. L. BRuns, J. TAYLOR. 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. Pp. 315-322 in M. Innis, D. Gelfand, J. Sninsky, and T. White (eds.), PCR pro- tocols: a guide to methods and applications. Academ- ic Press, San Diego CA. YOKOTA, Y., T. KAwatTaA, Y. IIDA, A. KATO, AND S. TANI- FUJI. 1989. Nucleotide sequences of the 5.8S rRNA gene and internal transcribed spacer regions in carrot and broad bean ribosomal DNA. Journal of Molecular Evolution 29:294-301. MApRONO, Vol. 45, No. 3, pp. 200-209, 1998 MORPHOLOGICAL VARIATION IN CALIFORNIA ALPINE POLEMONIUM SPECIES DANIEL W. PRITCHETT! AND ROBERT PATTERSON Department of Biology, San Francisco State University San Francisco, CA 94132 ABSTRACT Relationships between the two California alpine species of Polemonium (P. chartaceum H. Mason and P. eximium E. Greene) as well as their species limits have been uncertain for many years. We sought to clarify these issues through multivariate analysis of floral and foliar characters. We made 16 measurements on each of 159 plants from populations throughout the ranges of P. chartaceum and P. eximium, and from populations of two putatively related species, P. elegans E. Greene from the Cascade Range and P. viscosum Nutt. from the Great Basin and Rocky Mountains. We used multidimensional scaling (MDS) to summarize in three dimensions patterns of variation with regard to the 16 measured variables. We used discriminant analysis to test the robustness of patterns identified in heuristic interpretation of the MDS. Our results suggest populations from the Klamath Range, White Mountains, northern Sierra Nevada (Sonora Pass), and southern Sierra Nevada each form distinct morphological-geographical entities. Pat- terns of similarity among these groups suggest that P. chartaceum and P. eximium warrant taxonomic revision, and that populations of P. chartaceum from throughout its range have affinities closer to P. elegans than to P. viscosum. Two alpine species of Polemonium are native to California: P. chartaceum H. Mason and P. exi- mium E. Greene. Both species, known commonly as sky pilots, are small herbaceous perennials with congested inflorescences and showy, blue-violet co- rollas. Both species are diploid (Pritchett 1993), and Grant (1959) reported that both are outcrossers, pollinated by bees and flies. Ranges of the two Cal- ifornia species as well as sampling locations of two related species (P. elegans E. Greene and P. vis- cosum Nutt, discussed below) appear in Fig. 1. Pol- emonium chartaceum is on List IB (plants rare, threatened, or endangered in California and else- where) of the California Native Plant Society (Skin- ner and Pavlik 1994). Unlike the Rocky Mountain P. viscosum, which has been the subject of considerable research (Ga- len 1990, 1985, 1983; Galen and Kevan 1980), the California species have received relatively little study. They have been examined only in the context of treatments of the entire genus or sections within it (Wherry 1942; Davidson 1950; Grant 1989; Pritchett 1993). Despite the taxonomic treatments provided by past researchers, several questions of taxonomy and biogeography remain. The first question regards the circumscription of P. chartaceum. When Mason (1925) described the species, he considered it en- demic to the White Mountains of eastern Califor- nia. Davidson (1950) later included Polemonium from the Klamath Range in northwestern California in P. chartaceum. Only three collections from the Klamath Range were available for examination at ' Present address: University of California White Moun- tain Research Station, 3000 East Line St., Bishop, CA 93514. that time (Pritchett 1994), and Davidson’s treatment resulted in P. chartaceum showing an unusual and disjunct distribution (Fig. 1). Monographers since Davidson (1950) have ac- cepted this circumscription, although occasionally collections from the Klamath Mountains (e.g., Den- ton 4239—WTU) are assigned to P. elegans rather than to P. chartaceum; and Whipple (1981) noted that the degree of leaflet dissection in Klamath Mountains plants differed from that described by Munz (1973). The California Natural Heritage Da- tabase (Horner 1976) has called specifically for a re-evaluation of the relationship between Klamath Mountains populations and those in the White and Sweetwater Mountains. A second series of questions exists regarding the taxonomic status of P. chartaceum and P. eximium, and relationships with regional congeners. Munz (1973) recognized P. chartaceum but wrote that it was ‘‘doubtfully specifically distinct from [P. exi- mium].”> He considered P. eximium ssp. charta- ceum as a likely alternative. Grant (1989) consid- ered it ‘“‘a matter of preference’’ as to whether to recognize P. chartaceum and P. eximium or to re- duce them both to subspecies of their Rocky Moun- tain relative P. viscosum, following Murray (1983). Both Davidson (1950) and Wilken (personal com- munication) suggest a third possibility: that further examination might lead to reduction of at least P. chartaceum to a subspecies of the P. elegans of the Cascade Range. Biogeographic patterns implicit in these pro- posed subspecific relationships differ considerably. Munz (1973) P. eximium ssp. chartaceum suggests that California taxa are most closely related to each other. Polemonium elegans ssp. chartaceum, how- ever, suggests north-south relationships along the 1998] g Cascade Mts - P. elegans Klamath Mts ’ - P. chartaceum Toiyabe & Toquima Mts B= P. viscosum Sierra Nevadal, - P. eximium - P. chartaceum \ | White Mts ( P. chartaceum Fic. 1. Distribution and Sampling Map. Polemonium chartaceum occurs in the White, Sweetwater, and Klamath Mountains and was sampled in all three ranges. Polemo- nium eximium occurs in the Sierra Nevada and was sam- pled in the four labeled areas. Polemonium elegans occurs in the Cascade Range and was sampled on Mt. Ranier. Polemonium viscosum occurs in alpine areas throughout the Rocky Mountains and Great Basin and was sampled in the Toiyabe and Toquima Mountains. See Table 1 for precise collection localities. TABLE 1. PRITCHETT AND PATTERSON: ALPINE POLEMONIUMS 201 Sierra Cascade axis, while P. viscosum ssp. char- taceum and eximium could suggest east-west rela- tionships across the Great Basin to the Rocky Mountains. As part of an examination of relationships among California alpine Polemonium species we assem- bled a data set of floral and leaflet measurements (Table 2). Measurements were made on plants col- lected from populations throughout the entire geo- graphic ranges of both CA species as well as pop- ulations of P. elegans and P. viscosum (Table 1, Fig. 1). In this paper we describe patterns of mor- phological variation revealed in these data through techniques of multivariate analysis. MATERIALS AND METHODS Sampling. Inflorescences (one per plant) were collected from throughout each population visit- ed—(Table 1, Fig. 1). After collections were pressed and dried, they were inspected for the pres- ence of intact, 5-merous flowers in which the stig- mas had spread apart. In large inflorescences sev- eral flowers typically met these criteria and the largest were chosen for dissection. This provided a simple means for consistent sampling among pop- ulations without lengthening the already-lengthy dissection and measurement protocol. We initially dissected four flowers per plant but were soon forced by constraints of inflorescence size and dis- section time to reduce the number to two, and, in two cases, one. Because we wished the individual plant to be the OTU rather than the individual flower we averaged SOURCES OF MATERIAL USED FOR MEASUREMENTS. Elevations are converted (1 ft = 0.3048 m) from those on U.S.G.S. 7.5’ and 15’ topographic maps. All collections made by DWP. # Plants Collection Mountain range Location measured number Klamath Trinity Co., Summit, 2658 m, 1.6 km. NW Mt. Eddy 6 101 Klamath Trinity Co., Mt. Eddy Summit, 2751 m. 6 102 Klamath Trinity Co., Summit 2707 m, 1.2 km. E Mt. Eddy 7 100 Sweetwater Mono Co., Summit South Sister Peak, 3456 m 19 111 White Mono Co., Marble Ck. Divide, 4084 m 6 115 White Mono Co., Saddle 0.5 km. SE Mt. Dubois, 4072 m 6 105 White Mono Co., SE slope White Mt. Peak, 4237 m Si 106 TOTAL P. chartaceum a7 Sierra Nevada (Sonora Pass) Mono Co., Summit 3307 m., 1.2 km. SE Leavitt Lake 10 118 Sierra Nevada (Sonora Pass) Mono Co., Summit Leavitt Peak, 3527 m 9 119 Sierra Nevada (Mt. Dana) Tuolumne Co., Summit Mt. Dana, 3979 m 16 103 Sierra Nevada (Mono Pass) Inyo Co., Ridge 0.4 km. SW Mono Pass, 3840 m 8 104 Sierra Nevada (Mono Pass) Fresno Co., Ridge 1.6 km. W Pine Creek Pass, 3658 m 8 112 Sierra Nevada (Mt. Gould) Fresno Co., SW slope Mt. Gould, 3810 m 8 107 Sierra Nevada (Mt. Gould) Fresno Co., SW slope Kearsarge Pass, 3658 m 8 12] TOTAL P. eximium 67 Cascade Pierce Co. (Wash.), Mt. Rainier, Panorama Pt., 2134 m 16 108 TOTAL P. elegans 16 Toiyabe Lander Co. (Nev.), Big Creek Summit, 3374 m 10 109 Toquima Nye Co. (Nev.), NW slope Mt. Jefferson, 3353 m =o 110 TOTAL P. viscosum Ne) [Vol. 45 MADRONO ** OCC OOSC—= FeCl TLE v9'0 VES 09°C pS'e v9? UOTHIOSXS JOYIUY *K CU Sais OWL eos c6'I OOl— 600— 16c— Co Rs UOTIISX9 BUSTS * 6 =U OT =U OT =U OT=U OT =U S6T="U4U OT=H=U 6GTH=U OT =U Dee No > Ls G1 oa v9'T CSV 10'V OC’ SOE 90°V V8'C 8c C col Sogo] JopBee] JO JOquUINN * 6S°9 LED Wego} vss ves LoS 6c'S 8cS O8'S aseq xATeD JOSM LS d Tata OL'8 08°6 cO'Ol L0O'6 C18 cL S O18 vC8 918 souarogumnoqto xATeD * 80'€ vor’ SEL She ad oc'¢ eVic LSC 8L°C yIsue] eqoy xATeD I> Os La MS laa Sv'9 LOL vol 6$'9 vO's c8'°S 00'S 89'S ves yisuey xATeD * cO'V 66°€ 86€ EDs cv’ Cs I8'e 90'7 8Se oseq oqn} B[[O1O) ** 8c Vl 9L01 ev il £O'Cl £9 Ol vS'6 Oo Vo'6 el Ol SOUSIOFUINIITS OQn} B[[OIOD oa SL Se 1S MS a S68 SIT cs Ol cSOl Sc 8 Ic9 cV9 cL9 69°9 yisue] aqni e{[OIOD * 8L'1 SOC STC OIC 89'1 061 19°T co" 6S'I yisue] BUSS ie te Mae Sl a cOTL 9G CLE VTE Cog ISL 1L'9 £9°6 ISL ysug] s]AIS I>QA>MLSOOAA WC 6L°C 66°C GEV c9'¢ OV'C 90'¢ 6C'€ C8°T 90URISIP UOTOSUT iM SoD ld > aS 68°8 6L'S ILS 90'7 SCTE SOL 96'S 98°9 Ly 9 yisue] use] L>MSOOGCITAA 06°C OCC COC Ic? VIC C81 col 861 88°1 9seq 9q0] PB[[O1O) Op SL dM SIs a LOS coo no CLS SO’ ICL Oe 18°¢ CIV YIPIM SqoT P[[OIOD be-o] Oat tea Ss I Loo cO'9 os Ivs cy v lev CCV v6v ITs YIsUe] IQO] BT[OIOD Lo Oda S Toa SITOC 6L'0C 06'6l co LI cc rl 66°Cl 6c tl 8L Vl CLS) Jajoutiead ago] e][OI0D JUIWIDINSvII 189} oS8uel adnpnyy L 3) 9) rat a | NM S Mo WY :apoo o1ydeis0ayH WNSOISIA WIN1WUIXA WNIIDIADYI SUuD8a]a :sataadg ‘asuey aqeAloyl = J ‘pinoy ‘yA = D ‘ssed ouopy = CD ‘vurqd JA = C isseg B10U0S = ‘J ‘SIW ONY = MA SSI JaweMIVeMS = C ‘aBueYy YUepY = Y ‘souey IAT = Y :sepood orydess0ayH ‘ssurdnoss ou paajosai s}so} asuvs apdnpnu yotyM JOJ Ing (SO'C > d) APuBoYIUSIS JayJIp 0) suBaUT POaMOYS WAONV YOIYM UL SJUDUTAINSKIL JIVSIPUT SYSTIAISY “"YISUS| BAIS sso] YISUST 9QN} LI[OIOD ST UOTIasxa BUSS ‘YISUS] 9Gn} LI[OIOD sso] ‘9OULISIP UOTJOSUT pue YSU] JUDUIeTY JO WINS dy} ST UOTIISxd JOyUY ‘UONLUIOJsUeN JOOI oIeNbs JaIJe UNOS & ST Saqo] JoYee] JO JoquINNY “WU Ul aie SJUDUIAINSeIJ, “SISATVNY LNVNIAIXOSIG YOA GANISAG SdNOUD OIHdVADOAD ANIN AO SLSAL JONVY AMdILTIAYY STASY-NVNMAN-LNAGALS JO SLTASAY GNV SNVEW “T ATAVL 1998] all 15 floral measurements (Table 2) from each flower from a given plant. These 15 averages were combined with the count of maximum leaflet lobes per rachis node to obtain a set of 16 measurements to represent each plant in the data set. To meet Pi- mentel’s and Smith’s (1986) requirements that sam- ple sizes (for multigroup discriminant analysis) equal or exceed the number of variables measured, we sampled at least 16 plants from every geograph- ic region of interest. Characters. We assembled a data set of floral and leaflet measurements (Table 2). We chose these characters because they have been used in various combinations by previous workers to distinguish al- pine Polemonium species (Wherry 1942; Davidson, 1950; Grant 1989). We treated two important char- acters, however, in novel ways. A two-state char- acter regarding anther exsertion (i.e., anthers ex- serted vs. anthers included) has been used to help distinguish P. chartaceum from P. eximium (Wher- ry 1942; Davidson, 1950; Grant 1989). Rather than record this two-state character we measured the flo- ral components that define anther exsertion (i.e. fil- ament length, filament insertion distance, and co- rolla tube length) (Table 2) and used the compo- nents—instead of the two-state character—in our analysis. Similarly, a two-state leaflet arrangement character, two-ranked vs. verticillate (Wherry 1942; Davidson 1950), or arranged in one plane vs. ar- ranged in whorls (Grant 1989), has been used. Grant (1989) used this character to help define sec- tions of the genus. Instead of recording this as a two-state character we sampled two well-developed leaves and counted the maximum number of leaflet lobes per rachis node. (The maximum—as opposed to the mean—number of lobes was used due to the extensive time and effort required to count leaflets at all nodes on a leaf to obtain a mean.) We made these changes in character treatments in search of finer resolution of morphological patterns and vari- ation among plants examined for the analysis. The measurements discussed above were made on plants sampled from populations throughout the entire geographic ranges of both California species, as well as populations of P. elegans and P. viscos- um (Table 1, Fig. 1). Dissection. After re-hydration, one longitudinal incision was made from the mouth of the corolla (between two corolla lobes) to the base of the co- rolla tube. The corolla tube was opened and flat- tened on a microscope slide. Corolla lobes were removed from the corolla tube (perpendicular to the long axis of the corolla tube) at their bases. The corolla lobes and tube were then mounted on the slide. The style was removed from the ovary and mounted on the slide with the three stigmas sepa- rated. Filaments were pulled up from the surface of the flattened corolla tube, bent back at the points of insertion and then flattened back on the surface of the corolla. This created at the point of insertion PRITCHETT AND PATTERSON: ALPINE POLEMONIUMS 203 a distinct angle that was used as a landmark for measurement. Calyx tubes were dissected with a longitudinal incision, opened, and mounted on slides in a procedure analogous to that used for dis- section of the corolla. Calyx lobes were not cut from the calyx tubes, however, as were corolla lobes. Measurement. Dissected flowers were placed un- der a dissecting microscope with a camera lucida attachment. The camera lucida was used to project on the magnified image of the dissected flowers the image of the cross hairs of a mouse on an adjacent digitizing tablet. This enabled us to measure both straight-line distances and the lengths of perimeters of curved features. Straight-line distances between pairs of morphological landmarks were measured by placing the image of the cross hairs on each landmark, then clicking the mouse button. Curved shapes were measured by holding a mouse button down and “‘dragging”’ the image of the cross hairs of the mouse along the image of the particular fea- ture to be measured. All floral measurements were made with the camera lucida and digitizing tablet using one of these two techniques, and were re- corded directly on the microcomputer to which the digitizing tablet was connected. Analytical techniques. We subjected character data to nonmetric multidimensional scaling [MDS]. This technique allowed us to summarize in three dimensions patterns of variation with regard to all 16 measurements. MDS has been shown to be ef- fective for numerical taxonomy (Rohlf 1972 as cit- ed in Pimentel 1979) and usually outperforms Prin- cipal Components Analysis (Pimentel 1979). To as- sess variation at both intra- and inter-specific scales, we examined measurements of plants assigned to P. chartaceum and plants assigned to P. eximium both separately and together. Before performing each MDS we standardized floral measurements and calculated average Euclid- ean distances between all pairs of OTU’s (plants). We used the resulting distance matrix first as the basis of a Principal Coordinates analysis. We then used the principal coordinates as an initial config- uration (following the procedure of Rohlf (1993)) for an MDS of the average Euclidean distance ma- trix. We also subjected the character data to multi- group discriminant analysis (MDA) (sensu Pimen- tel and Smith 1986) to test the robustness of pat- terns seen in results of MDS (Abbot et al. 1985). Herein, after square-root transformation of leaflet counts to correct for deviation from normality, we classified each OTU (plant) a priori to a group, and calculated linear combinations of measured vari- ables in which distances among the a priori groups were maximized (canonical variates analysis). We then calculated the scores of the OTU’s on the ca- nonical variates axes and compared them to scores of the centroids of the original groups. This allowed 204 computation of the probability of membership in each group for each OTU (Geisser classification). The extent to which predictions of group member- ship were successful was taken as evidence the par- ticular circumscription of groups could be support- ed based on the variables measured. Classification results in MDA thus helped provide an idea of ‘‘how different is different’? Pimentel (1979). We used several different circumscriptions of groups for different iterations of MDA. Results pre- sented below are based upon a geographic circum- scription: each plant was assigned to one of nine groups defined by the population and/or region where the plant was collected. The assignment of plants to groups is shown by the nine values in the ‘“Geographic Code”’ field in Table 2. Results of MDS suggested geographic patterning; these cir- cumscriptions represented geographic grouping at the finest scale (in terms of group sample sizes vs. numbers of variables) the data could support (Pi- mentel and Smith 1986). Other components of the MDA (in addition to the calculation of canonical variates and Geisser classifications mentioned above) included calcula- tion of ANOVA’s, Student-Newman-Keuls Multiple Range Test scores, and Generalized distances (Ma- hanobis Distance) among group centroids. These analyses aided interpretation of discriminant results by showing differences among the nine geographic groups with regard to means for each variable (ANOVA and Multiple Range Test) and by provid- ing a quantitative measure of the degree of resem- blance among the groups (Generalized distances) (Pimentel 1979). We used UPGMA clustering to summarize the Generalized distances among all pairs of group centroids into a single phenogram. To relate the floral measurements used in this analysis to anther and stigma exsertion characters used in previous treatments we calculated values of these characters for each of the nine groups defined for discriminant analysis. Anther exsertion was cal- culated by summing mean filament length and in- sertion distance and subtracting mean corolla tube length. Stigma exsertion was calculated by mean style length from mean corolla tube length. Calcu- lations were made from means calculated as part of the ANOVA mentioned above, and were not sub- ject to statistical analysis. We calculated similarity matrices, principal co- ordinates analyses and MDS using the NTSYS 1.8 software package (Rohlf 1993). We calculated dis- criminant analyses using the Bioxdtat II (Pimentel 1986) and CSS Statistica 3.1 packages. RESULTS Reductions of the 16-dimensional relationships in the original data set to three dimensions by means of MDS are shown in Figures 2—4. In Figure 2 plants from the Klamath Mountains and the White Mountains occupy different areas of the plot, MADRONO [Vol. 45 @ Klamath Mts. B White Mts A Sweetwater Mts Fic. 2. Multidimensional Scaling of P. chartaceum From Three Mountain Ranges. Symbols represent individual plants. Lines extending down from each symbol allow the position of the symbol to be interpreted with regard to all three axes. while plants from the Sweetwater Mountains are scattered in between. Plants from the Sonora Pass (Sierra Nevada) are separated from those of other Sierran populations in Figure 3. In Figure 4 Sonora Pass plants are separated from conspecific ones elsewhere in the Sierra Nevada (as in Fig. 3), al- though Figure 4 depicts the result of an MDS of P. chartaceum and P. eximium together. ANOVA’s show significant (P < 0.05) differ- ences among means of all 16 characters (Table 2) among the nine geographic groups defined for dis- criminant analysis. Patterns of differences among means were resolved by the Student-Newman- Keuls Multiple Range test (Table 2) for 11 of the 16 variables. The synthetic character “‘anther ex- sertion”’ (Table 2) separates populations currently assigned to P. chartaceum from those assigned to P. eximium with the exception of Sonora Pass pop- ulations (Table 2, Geographic Code ‘“‘L’). Geisser classification results (Table 3) show that the only misclassifications involve the Sweetwater Moun- tains—1 plant from the Sweetwater Mountains er- roneously assigned to the Klamath Mountains, 4 plants from the Sweetwater Mountains erroneously assigned to the Sonora Pass Sierra Nevada, and 1 plant from the Sonora Pass Sierra Nevada errone- ously assigned to the Sweetwater Mountains. The first three canonical variates axes (Fig. 5) account for 91% of the total variance among all eight axes. On the horizontal axis the Sonora Pass group as well as populations assigned to P. chartaceum are separated from populations assigned to P. eximium. On the vertical axis the Sonora Pass group is sep- arated from both P. chartaceum groups and south- ern Sierra P. eximium groups. The centroid repre- senting P. elegans is close to centroids assigned to P. chartaceum on both horizontal and vertical axes, 1998] A Sonora Pass @ Mt. Dana V_ Mono Pass @ Mt. Gould Fic. 3. PRITCHETT AND PATTERSON: ALPINE POLEMONIUMS 205 it | | Multidimensional Scaling of P. eximium from Four Regions in the Sierra Nevada. Symbols represent individual plants. Lines extending down from each symbol allow the position of the symbol to be interpreted with regard to all three axes. while the P. viscosum centroid is close only on the horizontal axis. Generalized distances among all nine group centroids on all eight canonical axes as summarized by UPGMA clustering (Fig. 6) are consistent with pattern of centroids on the first three canonical axes in Figure 5. DISCUSSION These results reveal a level of resolution of mor- phological variation finer than any previously at- Fic. 4. tained. Patterns are observed with regard to floral and foliar architecture as well as overall similarity among populations. Previous workers have described floral morphol- ogy of California alpine polemoniums in terms of discrete forms with regard to anther exsertion. Flowers of P. chartaceum are regarded as having exserted anthers, while those of P. eximium have included anthers (Davidson 1950; Munz 1973; Grant 1989; Wilken 1993). Results in Table 2 show Sonora Pass ene Sierra Nevada Mono Pass Mt. Gould | Klamath Mts Sweetwater Mts White Mts Multidimensional Scaling of P. chartaceum (from three mountain ranges) and P. eximium (from four regions in the Sierra Nevada). Symbols represent individual plants. Lines extending down from each symbol allow the position of the symbol to be interpreted with regard to all three axes. 206 TABLE 3. RESULTS OF GEISSER CLASSIFICATION. Numbers are numbers of plants. Rows are a priori group assign- ments and columns are predictions (6 misses/153 hits = 96% classification rate). Any numbers not on the diagonal represent misclassifications. See Table 2 for interpretation of geographic codes. Geographic code RoALK. (Ab D> “Gy ve aS" SW R lo-- (0 “0° “O04 0-02 “Or CO =) K 0 19 0 0 0 0 0 0 0 L Ov oO, 18.05 O-: 0 l OO D O° HO: 09 1G OO UOs~ “O SOe- 20 G O. “O) “Op ° 20 “te: "Os a0b.. "OF »O C 0. 0" 20:- (OF S08 Tee 0" 26h. S 0 l 4 0 0 0 14 O O T O° 2Ox BO! Os? OF OY ~ Oho TOT 10 W OF. Os Oe FO: -s0e 20 ae: 2On 6 9. that stigma exsertion also distinguishes these two floral forms, and more importantly, that there is a third floral form not previously recognized. This form is largely restricted to Sonora Pass plants, and is characterized by flowers with significantly (P < 0.05) greater insertion distances and significantly shorter filaments than those of flowers from any other areas. In terms of anther exsertion, Sonora Pass flowers (Table 2 Geographic code “‘L’) are intermediate between those of P. chartaceum and those of P. eximium from the central and southern Sierra Nevada. In terms of the components of an- ther exsertion, however, ““L’ flowers have extreme values, thus possess a unique floral form. Whether or not this form is recognized, future treatments will have to put in an exception clause for Sonora Pass plants if the customary two-state anther ex- sertion character is used. Variation in foliar architecture reveals a different pattern. Rather than being interpretable in terms of several discrete forms, there is only one relatively stable form of leaflet morphology in California. This form is characterized by a predominantly two- Sonora Pass White Mts Sweetwater Mts Cascade Mts Mono Pass Klamath Mts Mt. Gould Toiyabe and Toquima Mts Fic. 5. Canonical variates analysis of alpine Polemonium populations. Symbols indicate centroids along canonical variates axes of plants assigned to nine geographic groups defined for discriminant analysis. MADRONO [Vol. 45 part leaflet dissection and is restricted to plants in the Klamath Mountains. Leaflet dissection among plants of other regions is significantly greater than that of Klamath Moun- tain plants, and is best interpreted in terms of a cline rather than discrete forms. There is a signifi- cant increase in degree of dissection (Table 2) among California populations from north to south. This pattern is especially clear among P. eximium populations in the Sierra Nevada, but also holds for P. chartaceum. Given the importance of foliar mor- phology in regulation of CO, and H,O balances, this latitudinal increase in leaflet dissection might offer a fruitful subject of investigation for physio- logical ecologists. When leaflet and floral characters are considered together to assess overall similarity, geographic variation is apparent in both intra- and inter-specific MDS results (Figs. 2—4), and on a finer scale, in successful Geisser classification of almost every plant sampled from all nine geographic areas (Table 3). The six Geisser misclassifications are the only data that suggest any patterns of similarity that cross (rather than coincide with) geographic bound- aries. While populations in all nine geographic areas may be separable via discriminant analysis, not all areas have equally stable and distinct morphologi- cal forms. Data in Figures 2 and 4, and Geisser misclassifications (Table 3) suggest that populations in the Sweetwater Mountains may be more variable than populations in other areas. While we cannot rule out the possibility that this variation may be an artifact of unequal and generally small sample sizes, there are several biological hypotheses wor- thy of consideration. One hypothesis is that the variability may be evidence of sympatry between P. chartaceum and P. eximium. All Sweetwater Mountain plants have been assigned to P. charta- ceum by previous workers. The four Sweetwater Mountain plants misclassified as Sonora Pass plants (Table 3), however, have the short-filament floral morphology characteristic of Sonora Pass plants, which have been assigned to P. eximium. No field observations have been made of sym- patric populations of P. chartaceum and P. exi- mium. Rather, observations have been made in the Sweetwater Mountains that the same plant may have some flowers with the exserted (P. charta- ceum) design and other flowers with the short-fil- ament Sonora Pass design. This suggests that intro- gression and/or hybridization with plants from So- nora Pass may be a better interpretation than sym- patry. A final hypothesis is that this floral variation could be evidence that populations in the Sweet- water Mountains are composed of a form ancestral to the (florally) more uniform forms which now oc- cur elsewhere (Morefield personal communication). All hypotheses are consistent with the geography of the area. The terrain between the Sonora Pass populations and the Sweetwater Mountains (about 1998] 12.5 10.0 TS PRITCHETT AND PATTERSON: ALPINE POLEMONIUMS 25 Cascade Mts - P. elegans Klamath Mts - P. chartaceum Sweetwater Mts - P. chartaceum White Mts - P. chartaceum Sonora Pass - P. eximium Toiyabe/Toquima Mts - P. viscosum Mt. Dana - P. eximium Mono Pass - P. eximium Mt. Gould - P. eximium Fic. 6. Phenogram based on UPGMA clustering of generalized (Mahalanobis’) distances among centroids of geo- graphic groups defined for discriminant analysis. Numbers on the scale above the phenogram represent standard de- viations. The greater the number at which a group is joined to a cluster, the less the similarity. 30 km apart) is high enough so it would have had a habitat suitable for alpine Polemonium coloniza- tion as recently as the end of the last glacial period. Further investigation is needed to properly interpret the significance of morphological variation in Sweetwater Mountain plants. The southern Sierra Nevada is another area where populations which can be separated by means of discriminant analysis (Geographic Codes D, C, and G in Table 2; Mt. Dana, Mono Pass, and Mt. Gould in Fig. 5) should not necessarily be treat- ed as separate morphological entities. Patterns in Figures 3 and 4 show that, while there is consid- erable variation among plants in the Sierra Nevada (P. eximium), there is one principal morphological discontinuity that separates Sonora Pass popula- tions from those to the south. The intermediate po- sition of the Mt. Dana population, both morpholog- ically (Figs. 4 and 5) and geographically (Fig. 1), suggests a latitudinal cline may exist in the Sierra Nevada from populations near Sonora Pass south at least as far as to those around Mono Pass. We observed two additional characters (inflores- cence congestion and anther color) that display this pattern. While these characters were not quantified for use in multivariate analysis, repeated field ob- servations have shown Sonora Pass plants to have noticeably less congested inflorescences and a much greater abundance of yellow pollen (as op- posed to cream-colored pollen) than do plants from farther south in the Sierra Nevada. If variation in the Sierra Nevada is interpreted to represent two forms, and variation in the Sweet- water Mountains is hypothesized to result from in- trogression, four areas (of the seven areas in Cali- fornia defined for discriminant analysis) are left in which discrete morphological forms occur—the Klamath Range, the White Mountains, Sonora Pass, and the southern Sierra Nevada. The fact that the morphological forms identified above do not co-occur is consistent with theories of Polemonium evolution by previous workers. Grant (1989) hypothesized an allopatric mode of speciation. He wrote that the common ancestor of P. eximium and P. chartaceum ‘“‘had a semicontin- uous distribution in the far west at a cool stage of the Pleistocene. Character divergence among these taxa developed along with geographic isolation.” Davidson (1950) wrote “if cognizance were taken of the minute differences between popula- tions on different mountains one might eventually delimit as many subspecies as there are populations on isolated mountains.”’ In this statement Davidson implied allopatric speciation and anticipated the recognition of geographic-morphological entities such as those discussed above. He did not, however, consider the possibility that taxonomic treatment of these entities might require substantial revisions (as opposed to simply splitting species into subspe- cies). The phenogram in Figure 6 infers the taxonomic complexity of this group of populations. It suggests that Klamath Mountain populations have a greater affinity with Cascade Mountains populations of P. elegans and Sweetwater Mountain populations than with the White Mountain populations with which they are included in the current circumscription of P. chartaceum. The White Mountains group are sis- ter to the Cascade-Klamath-Sweetwater group, forming an elegans-chartaceum group. This treat- ment is consistent with Davidson’s and Wilken’s suggestions, but inconsistent with Murray’s reduc- tion of P. chartaceum to P. viscosum ssp. charta- ceum. Curiously, the Sonora Pass populations of P. ex- imium are sister to the elegans-chartaceum group rather than with other Sierra Nevada populations (Mt. Dana, Mt. Gould and Mono Pass) with which they are included in P. eximium. No previous treat- ments have placed any Sierra Nevada populations in the same taxon with those occurring in the Klam- ath, Sweetwater, or White Mountains. Populations of P. viscosum from the Toiyabe and Toquima Mountains are sister to the group consist- ing of P. elegans, P. chartaceum, and one popula- tion of P. eximium. This treatment is not consistent 208 with any published or proposed taxonomic treat- ment. The three remaining populations of P. exi- mium (groups Mt. Dana, Mt. Gould, and Mono Pass in Fig. 6) constitute a lineage clearly distinct from and sister to the rest of the alpine species of Pole- monium. Thus, with the use of explicit, quantitative analytical techniques, taxonomic grouping in this complex is even more difficult. One resolution of the question of rank would be the creation of a single western North American alpine Polemonium species, with all four California morphological en- tities, as well as P. elegans and P. viscosum, treated as subspecies (Stebbins personal communication). In this treatment the apparent arbitrariness of cur- rent circumscriptions and ranks regarding floral and foliar characters would be eliminated. This treat- ment, however, would not give sufficient attention to obvious morphological differences among pop- ulations from different regions. It may also obscure evolutionary relationships in its oversimplification. Nevertheless, current circumscriptions of P. chartaceum and P. eximium are inconsistent with patterns of floral and leaflet variation. These two taxa might just as easily be treated as four or five species, depending on the taxonomic interpretation of populations in the Sweetwater Mountains. Pat- ently the Sierran P. eximium is more taxonomically complex than has been thought. Three of these en- tities show greater similarity to P. elegans of the Cascade Range than to P. viscosum of the Rocky Mountains and Great Basin. This suggests that, in addition to revisions of circumscriptions of Cali- fornia species, relationships with regional conge- ners must be reconsidered as well. It is premature to make further inferences re- garding the relationships between California alpine Polemonium species and P. viscosum, due to the wide distribution and minimal sampling of P. vis- cosum in this examination. According to Grant (1989), however, there is an east-west clinal de- crease in floral size from Rocky Mountains popu- lations of P. viscosum to those in the Great Basin. The two populations sampled in this examination are the western-most in the Great Basin, and pre- sumably represent the small end of this cline. Since they are still larger in almost all measurements than most California populations (except southern Sier- ran), differences between P. viscosum and Califor- nia forms may be underestimated in this examina- tion. Further sampling of P. viscosum might rein- force, rather than change, the basic pattern of rel- ative similarities described above. One taxonomic conclusion supported by the cur- rent morphometric analysis is that the Klamath Mountain population is readily distinct from those from the White Mountains; based on these data, P. chartaceum should not refer to both entities. Ad- ditional support of the distinction between Klamath and White Mountain taxa is afforded by ITS se- quence data (de Geofroy et al. 1996), which show that White Mountain populations are more closely MADRONO [Vol. 45 related to P. eximium from the northern Sierra Ne- vada than they are to the Klamath Mountain pop- ulation on Mt. Eddy. Taxonomic recognition of the Klamath material as a distinct species is being un- dertaken (Pritchett and de Geofroy unpublished). Our primary goal in this study was to decipher the systematic complexity of the California alpine polemoniums and delineate the taxa that are rec- ognizable based on morphology. The next logical goal is to reconstruct the evolutionary history of the alpine polemoniums; however, the morpholog- ical characters that can be used for recognition of taxa are likely too few to allow a rigorous phylo- genetic reconstruction. Recently de Geofroy (1998) undertook a molecular phylogeny of the alpine pol- emoniums. Her results based on sequences of the ITS region of nuclear ribosomal DNA, are in gen- eral accord with our phenetic analysis of morpho- logical data. In particular her results support our conclusion that intra-specific variation within P. chartaceum and P. eximium require revision of these species. Final evaluation of all data, molec- ular and morphological, will generate a phyloge- netic model of the complex that can be used to test future hypotheses on the evolution of this genus. ACKNOWLEDGEMENTS We acknowledge funding to DWP from the California Native Plant Society, the University of California White Mountain Research Station, and the Hardman Foundation. We thank Randy Zebell, Eva Buxton, Isabelle de Geofroy, and Dieter Wilken for constructive criticism. We also thank Candace Galen and Joanna Schultz for their valu- able suggestions in their reviews of the manuscript. LITERATURE CITED ABBoTT, L. A., F A. BISBY, AND D. J. ROGERS. 1985. Tax- onomic analysis in biology. Columbia University Press, New York. DAVIDSON, J. F 1950. The genus Polemonium (Tournefort) L. University of California Publications in Botany 23: 209-282. DE GeEorrRoy, I., R. PATTERSON, C. ORREGO, AND R. ZE- BELL. 1996. Phylogeny and biogeography of the high- elevation species of Polemonium (Polemoniaceae). American Journal of Botany 83 (Supplement): 149. DE Georroy, I. 1998. Molecular phylogeny and biogeog- raphy of the alpine species of Polemonium (Pole- moniaceae). M.A. thesis. San Francisco State Univ., San Francisco, CA. GALEN, C. 1983. The effects of nectar thieving ants on seed set in floral and scent morphs of Polemonium viscosum. Oikos 41:245—249. . 1985. Regulation of seed set in Polemonium vis- cosum: Floral scents, pollination, and resources. Ecol- ogy 66:792-797. . 1990. Limits to the distributions of alpine tundra plants: Herbivores and the alpine skypilots. Oikos 59: 355-358. . AND P. G. KEvAN. 1980. Scent and color, floral polymorphisms and pollination ecology in Polemo- nium viscosum Nutt. American Midland Naturalist 104:28 1-289. 1998] GRANT, V. 1959. Natural History of the Phlox Family. Martinus Nijhof, The Hague. . 1989. Taxonomy of the tufted alpine and subal- pine polemoniums (Polemoniaceae). Botanical Ga- zette 150(2):158—-169. Horner, S. 1976. Polemonium chartaceum. Unpublished file document. California Natural Heritage Data Base, Sacramento, CA. Mason, H. L. 1925. Polemonium chartaceum p. 783 in W. L. Jepson, A manual of the flowering plants of California. California School Book Depository, San Francisco, CA. Munz, P. A. 1973. A California Flora and Supplement. University of California Press, Berkeley. CA. Murray, E. 1983. Notae Spermatophytae No. 3. Kalmia 13:23-24. PIMENTEL, R. 1979. Morphometrics. Kendall Hunt, Du- buque, IA. AND J. SMITH. 1986. BIOXTAT II. A multivariate toolbox. Sigma Soft, Placentia, CA. PRITCHETT AND PATTERSON: ALPINE POLEMONIUMS 209 PRITCHETT, D. W. 1993. A biosystematic examination of California alpine polemoniums. M.A. thesis. San Francisco State Univ., San Francisco, CA. . 1994. The habitat and distribution of Polemonium chartaceum in the Klamath Range: a clarification. Madrono 41:224—226. ROHLF, F J. 1993. NTSYS-PC. Numerical taxonomy and multivariate analysis system. v. 1.8. Exeter Software, Setauket, NY. SKINNER, M. W. AND B. M. PAVLIK (eds.). 1994. Inventory of rare and endangered vascular plants of California. California Native Plant Society, special publication No. 1, 5th ed., Sacramento, CA. WHERRY, E. T. 1942. The genus Polemonium in North America. American Midland Naturalist 27:741—760. WHIPPLE, J. J. 1981. A flora of Mt. Eddy, Klamath Moun- tains, California. M.A. thesis. Humboldt State Univ., Arcata, CA. WILKEN, D. H. 1993. Polemonium. p. 852 in J. Hickman (ed.), The Jepson manual: Higher plants of California. University of California Press, Berkeley, CA. MApRONO, Vol. 45, No. 3, pp. 210-214, 1998 ADULT SEX RATIO OF ARCEUTHOBIUM TSUGENSE IN SIX SEVERELY INFECTED TSUGA HETEROPHYLLA ROBERT L. MATHIASEN School of Forestry, Northern Arizona University, Flagstaff, AZ 86011 Davip C. SHAW Wind River Canopy Crane, University of Washington, 1262 Hemlock Road, Carson, WA 98610 ABSTRACT The adult sex ratio of Arceuthobium tsugense (C. Rosend.) G. N. Jones ssp. tsugense (western hemlock dwarf mistletoe) was 1:1 (n = 1608 plants) in the crowns of six large, A. tsugense-infected Tsuga heterophylla (Raf.) Sarg. (western hemlock) at the Wind River Canopy Crane Research Facility in south- central Washington. One tree, however, had a female-biased adult sex ratio and another had a male-biased adult sex ratio. Plants in the lower crowns (less than 20 m in height) exhibited a female-biased adult sex ratio. Our results suggest that an earlier study, which sampled plants near the ground and reported a female-biased sex ratio for A. tsugense, may have been biased because of the sampling method used. Dwarf mistletoes Arceuthobium spp. (dwarf mis- tletoes) are obligate, dioecious flowering plants that are parasitic on conifers. The ratio of male to fe- male adult plants is typically 1:1 (Hawksworth and Wiens 1972, 1996). However, many adult popula- tions of Arceuthobium tsugense (C. Rosend.) G. N. Jones ssp. tsugense (western hemlock dwarf mistle- toes) are reported to exhibit female-biased adult sex ratios of approximately 3:2, even though the em- bryonic sex ratio for this species is 1:1 (Wiens et al. 1996). Many of the adult sex ratio determina- tions for A. tsugense reported by Wiens et al. (1996) were based on samples of less than 100 plants (10 of 16 populations), but some of their adult sex ratio determinations (3 populations) used more than 400 plants. Each of the three populations that sampled >400 plants had significantly female- biased adult sex ratios, as did their entire sample population of 3057 plants (59 percent females). The Wind River Canopy Crane Research Facility is located in the Wind River Experimental Forest of the Gifford Pinchot National Forest in southern Washington. The facility uses a construction tower crane (Liebherr 550 HC) 75 m in height with a jib arm 85 m long that is capable of accessing 2.3 ha of an old-growth Tsuga heterophylla (Raf.) Sarg. (western hemlock) forest (Parker 1997, http:// depts.washington.edu/wrccrf. Accessed on Febru- ary 3, 1999). Approximately one third of the over- and under-story 7. heterophylla within the crane- accessible area are infected with A. tsugense. Ar- ceuthobium tsugense counted in the populations examined by Wiens et al. (1996) were growing within approximately 4 m of the ground. Because the Wind River Canopy Crane (WRCC) provides access to the upper and middle crowns of large, severely infected 7. heterophylla, the adult sex ratio of this population of A. tsugense could be deter- mined using several hundred plants examined at heights much greater than 4 m. The objectives of this study were to determine if the adult sex ratio of A. tsugense has a sex-bias when several hundred plants within the crowns of large, severely infected T. heterophylla are sampled and if sex ratio varies on individual trees or by height of infection. METHODS Six individuals of T. heterophylla severely in- fected with A. tsugense were sampled using the WRCC in 1996. Selected trees were greater than 40 m in height, infected in most, if not all, of their crown, and were accessible by the crane (most of their outer and upper crowns could be accessed). Trees had been assigned identification numbers when the crane site was stem mapped in 1995. Each | tree was sampled from the bottom of the crown on — the north side, gradually working up that side to the © tree top. Arceuthobium tsugense was sampled on | crane-accessible branches which were attached to | the north side of the trunk. Each crane-accessible infection on a sample branch was examined for in- fections by A. tsugense and adult plants (plants with | flowers). The approximate height of each branch sampled was recorded, based on the vertical dis- , tance between the crane gondola and the load jib. | Individual infections were detected by the spindle- | shaped swellings induced by A. tsugense and/or the | presence of A. tsugense on branches. When A. tsu- | gense shoots were absent on an obvious single in- | fection, the infection was recorded as nonreproduc- | tive. Where adult plants were observed a small sec- — tion of a plant was removed and examined to de- _ termine its sex (Hawksworth and Wiens 1996). | When necessary, shoots and their attached flowers | were examined with a 10x hand lens to aid sex | determination. Arceuthobium tsugense on witches’ | 1998] MATHIASEN AND SHAW: MISTLETOE SEX RATIO brooms that could not be distinguished as arising from separate infections were not sampled. In many cases separate infections that were close to each other on a branch or witches’ broom could be dis- tinguished because the adult plants they produced were of a different sex. When adjacent infections on the same branch were the same sex, the infection was treated as the same infection. After sampling the north side of a tree, the pro- cedure was repeated on its west, south and east side. The north and south sides of the six trees were sampled in late April. In mid September, the east and west sides of the trees were sampled using the above methods in order to increase the sample size of reproductive infections. The east side of one tree (# 1134) and the west side of one tree (# 2119) could not be accessed by the crane. Thus, four trees (#’s 1129, 2002, 2054 and 2131) were sampled on four sides and two on only three sides (#’s 1134 and 2119). Data were tallied by branch and tree for male and female plants and for nonreproductive infec- tions. Male and female plants also were tallied by height classes (<20, 20.0—24.9, 25.0—29.9, 30.0-— 34.9, 35.0—39.9, 40.0—45.0 and >45 m). These height classes were selected to provide a uniform representation of the number of branches sampled in each height class. Frequencies of male and female adult A. tsugense (sex ratios) were compared for individual trees and height classes using chi-square tests. Differences were judged to be significant at P = 0.05. Based on previous reports of sex ratios for A. tsugense (Wiens et al. 1996) we hypothesized that the sex ratio of A. tsugense in the trees we sampled would be female-biased at a ratio of approximately 3:2. Nonreproductive infections (%) Adult sex ratio (% females) Reproductive infections RESULTS The crane survey sampled 239 branches and 1980 dwarf mistletoe infections. Of these 1980 in- fections, 1608 had flowering plants, 805 male and 803 female, and 372 were nonreproductive. Thus, the adult sex ratio for this population of A. tsugense based on a sample of 1608 plants was essentially 1:1. There was variation in the sex ratio between the six trees sampled (Table 1). One tree (# 2119) had a significantly female-biased sex ratio (60 per- cent female) and one tree (# 2131) had a signifi- cantly male-biased sex ratio (44 percent female). Nonreproductive infections varied from 10 to 26 percent of the total number of infections sampled per tree (Table 1). More than 20 percent of the in- fections were recorded as nonreproductive on four of the trees. However, we did not sample every in- fection on each tree because many of the infections close to the main bole of the trees were inaccessible by the crane. In addition, we have no estimate of how many infections were missed on accessible branches because we could not distinguish the in- fections. Total infections sampled ADULT SEX RATIO OF REPRODUCTIVE INFECTIONS AND PERCENTAGE OF NONREPRODUCTIVE INFECTIONS OBSERVED ON SIX TSUGA HETEROPHYLLA AT THE WIND RIVER Branches Tree number 1129 CANOPY CRANE RESEARCH FacILity. 'Significant deviation in the adult sex ratio from 1:1. Chi-square statistics (P = 0.05). TABLE 1. 10 16 19 0.956 0.618 0.877 0.428 0.010 0.013 0.960 50 50 53 60! 44! 50 315 197 373 193 154 376 1608 348 254 44] 261 194 482 1980 Total/Mean 1134 2002 2054 2119 2131 2 TABLE 2. MADRONO [Vol. 45 ADULT SEX RATIO OF REPRODUCTIVE INFECTIONS AND PERCENTAGE OF NONREPRODUCTIVE INFECTIONS ON SIX TSUGA HETEROPHYLLA BY HEIGHT CLASS, WIND RIVER CANOPY CRANE RESEARCH FACILITY. !Significant deviation in the adult sex ratio from 1:1. Chi-square statistics (P = 0.05). Non- Height Number Number Adult reproductive class branches reproductive sex ratio infections (m) sampled infections (% females) (%) <20 26 82 61! 53 20.0—24.9 25 89 47 49 25.0—29.9 19 113 43 32 30.0—34.9 45 Syed! 53 1 35.0—39.9 4] 201 47 1S 40.0—45.0 44 398 50 7 >45 4] 375 50 3 Total/Mean 239 1608 50 19 The sex ratio and percentage of nonreproductive infections by height classes are presented in Table 2. The number of infections with plants increased as height class increased. There was little variation in the 1:1 sex ratio for height classes >30 m. The <20 m height class had a significantly female-bi- ased sex ratio of approximately 3:2 (P = 0.047). The majority of observable nonreproductive infec- tions (61 percent) were in the lower part of the crowns (<30 m). The sex ratio and percentage of nonreproductive infections for the tree with the largest sample of plants (# 2131) is summarized by height classes in Table 3. This tree had a significantly male-biased sex ratio (Table 1). The distribution of sex ratio by height class for this tree was female-biased in the lower crown (<20 m), but the sex ratio was con- sistently male-biased as height class increased. However, the sample size in the lower crown was small (only 14 plants), but plant counts increased dramatically as height class increased in all of the trees sampled (Table 3). The tree with the next greatest plant count (# 2002) had a similar distri- bution of plant numbers and non-reproductive in- fections were predominantly in the lower crown as well. It also had a female-biased sex ratio in the lower crown (<20 m), but had a consistent 1:1 sex ratio as height class increased, which resulted in an approximate 1:1 adult sex ratio for all A. tsugense sampled on that tree (Table 1). Other trees with a © 1:1 sex ratio (¥#’s 1129, 1134 and 2054) had similar distributions of male and female adult plants within | their crowns. Tree # 2119 had a female-biased sex — ratio because it had a predominance of female plants in its lower and middle crowns, but approx- imately equal numbers of male and female plants in its upper crown. However, only 154 plants were | sampled on tree # 2119, the lowest sample size for | an individual tree. | DISCUSSION Based on the 1608 A. tsugense we sampled, the © adult sex ratio for this population of A. tsugense is — approximately 1:1. This ratio differed from that (3: 2) reported for many other populations of A. tsu- | gense (Wiens et al. 1996). The differences in adult | sex ratios between these studies may be related to © sample sizes and sampling methods. For example, 10 of the 16 populations sampled by Wiens et al. TABLE 3. ADULT SEX RATIO OF REPRODUCTIVE INFECTIONS AND PERCENTAGE OF NONREPRODUCTIVE INFECTIONS ON TREE , # 2131 BY HEIGHT CLASS, WIND RIVER CANOPY CRANE RESEARCH FACILITY. 'Significant deviation in the adult sex ratio | from 1:1. Chi-square statistics (P = 0.05). Chi-square analysis was not performed on this data set for individual height , classes because of the small sample sizes. Non- Height Number Number Adult reproductive class branches reproductive sex ratio infections (m) sampled infections (% females) (%) <20 4 14 79 ie) 20.0—24.9 4 8 50 64 25.0—29.9 =) 29 35 41 30.0—34.9 8 45 44 24 35.0—39.9 8 71 44 15 40.0—45.0 10 118 39 14 >45 ° 91 45 1 Total/Mean 48 376 44! 22 1998] (1996) had less than 100 plants and they sampled within 4 m of the ground. Wiens et al. (1996) point- ed out that in their study, deviations from a female predominant 3:2 sex ratio were primarily related to small sample sizes or to host species. Their data also indicated that not all populations of A. tsugense display a female-biased sex ratio. One of the pop- ulations they sampled had a 1:1 sex ratio (Lake Cowichan, British Columbia) and was based on a relatively large sample size (257 plants). Therefore, if the populations of A. tsugense with female-biased adult sex ratios reported by Wiens et al. (1996) were resampled to obtain larger sample sizes from the entire crowns of infected trees, it is possible that adult sex ratios for those populations may show trends towards the 1:1 sex ratio reported here for the WRCC population of A. tsugense and reported for other species (Hawksworth and Wiens 1972, 1996). Wiens et al. (1996) inferred that the most likely reason for female-biased sex ratios in populations of A. tsugense is a greater longevity of female plants. Differential mortality of male and female plants also was hypothesized as a probable reason for the female-biased sex ratio of another obligately dioecious mistletoe, Phoradendron tomentosum (DC.) Gray, in central Texas (Nixon and Todzia 1985). Female plants of juniper mistletoe Phora- dendron juniperinum A. Gray (juniper mistletoe) are also reported to have greater longevity than males (Dawson et al. 1990). Smith (1971) reported that the life span of A. tsugense shoots was at least five years and although he did not discuss that fe- male plants live longer than males, his Tables 2 and 3 indicate that on young infections, female plants live longer than males. Therefore, the idea that greater longevity of female plants may account for the female-biased adult sex ratios reported for sev- eral populations of A. tsugense is plausible. Wheth- er or not female A. tsugense live longer than male plants throughout the crowns of severely infected trees remains unknown. But our data from the low- er crowns of 7. heteropyhlla we sampled at the WRCC site demonstrated a definite female-biased sex ratio (Table 1). In addition, we conducted a preliminary survey of infections by A. tsugense in an attempt to determine an adult sex ratio for the A. tsugense population growing within 4 m of the ground in the WRCC site (unpublished). Of the nearly 300 infections we observed near the ground, only three percent had plants that could be sexed and nearly all of these were females. These prelim- inary data and the crane survey data from the lower crowns of the large 7. heterophylla sampled (Tables 1 and 2) support the idea that female plants may be surviving longer than males in the lower crowns of T. heterophylla infected by A. tsugense. There- fore, sex ratio determinations based on samples from the lower crowns of T. heterophylla infected by A. tsugense will be consistently female-biased. Our data (Table 2) demonstrate that A. tsugense MATHIASEN AND SHAW: MISTLETOE SEX RATIO 213 is more abundant in the middle and tops of large, severely infected 7. heterophylla. Smith (1969) also reported that A. tsugense shoot production was more abundant in the middle and upper crowns of severely infected T. heterophylla. Less shoot pro- duction in the lower crowns of A. tsugense-infected trees may be related to the age of infections, as older infections often produce fewer shoots (Smith 1971). Because initial infection of trees occurs in the lower crown and moves upward in individual trees (Richardson and van der Kamp 1971; Par- meter 1978; Shaw 1982), infections in the lower crowns are generally older than infections in the upper crowns. However, many investigators have associated poor shoot production of A. tsugense with low light intensity in the lower crowns of in- fected trees or within dense forests (Weir 1916; Korstian and Long 1922; Gill 1935; Wagener 1961; Smith 1969; Richardson and van der Kamp 1971; Smith 1971) and Baranyay (1962) provided data to support this relationship. Whatever the reason for the large number of non- reproductive infections in the lower crown (age of infections, shading or other factors), sampling in- fected hemlock near the ground will mean that many infections of A. tsugense can not be used for adult sex ratio determinations. If the malate dehy- drogenase method of sex determination reported by Wiens et al. (1996) could be applied to A. tsugense tissue extracted from infected branches (endophytic system tissue), then the sex of nonreproductive in- fections could be determined. If the sex of all in- fections could thus be determined, then perhaps, the sex ratio for all infections sampled would be 1:1 since the embryonic sex ratio is 1:1 for A. tsugense (Wiens et al. 1996). Because of the variation in sex ratio by crown position and individual trees in the A. tsugense pop- ulation at the WRCC site, adult sex ratios should be determined for additional A. tsugense popula- tions using methods that minimize the bias we feel is associated with sampling a few trees from near the ground. Adult sex ratio determinations should be made by sampling several trees from throughout their crowns and plant counts should consist of at least 1000 individuals for each A. tsugense popu- lation sampled. Therefore, we recommend that A. tsugense adult sex ratio determinations be made by destructively sampling several trees per population. Destructive sampling allows access to branches in the middle and upper crowns where many more A. tsugense plants can be sexed. This will increase plant counts and should eliminate the bias that our data indicates is probably associated with sampling from the lower crowns of infected 7. heterophylla. The senior author has used this technique for Ar- ceuthobium laricis (Piper) St. John (arch dwarf mistletoe) in northeastern Washington (results re- ported in Table 3 of Wiens et al. 1996) and found destructively sampling several A. /aricis-infected western larch Larix occidentalis Nutt. (western 214 larch) provided a large sample (over 1500 plants) with a reasonable expenditure of time and effort. It also eliminated the potential bias of arbitrarily se- lecting mistletoe plants on infections near the ground and the possibility of sexing the same plants more than once. Destructive sampling also allows the ages of individual infections to be determined (Scharpf and Parmeter 1966; Smith 1971; Shaw 1982), thus making it possible to correlate adult sex ratio with age structure of the population. If female plants possess greater longevity than male plants, then older infections should be comprised of a higher proportion of females. ACKNOWLEDGMENTS The field assistance of Elizabeth Freeman, John Brown- ing and Andrew “Buz” Baker is greatly appreciated. Re- views of the original manuscript by Del Wiens, Terry Shaw and John Marshall also are appreciated. LITERATURE CITED BARANYAY, J. A. 1962. Phenological observations on Western hemlock dwarf mistletoe (Arceuthobium campylopodum Gill forma tsugensis). Canadian De- partment of Forestry, Entomology and Pathology Branch, Bi-monthly Progress Report 18(4):3—4. Dawson, T. E., J. R. EHLERINGER, AND J. D. MARSHALL. 1990. Sex-ratio and reproductive variation in the mis- tletoe Phoradendron juniperinum (Viscaceae). Amer- ican Journal of Botany 77:584—589. GILL, L. S. 1935. Arceuthobium in the United States. Con- necticut Academy of Arts and Science Transactions 32:111-245. HAWkKSworTH, E G. AND D. WIENS. 1972. Biology and classification of dwarf miustletoes (Arceuthobium). Agriculture Handbook 401, USDA Forest Service, Washington, D.C. HAWKSWORTH, E G. AND D. WIENS. 1996. Dwarf mistle- toes: biology, pathology, and systematics. Agriculture Handbook 709, USDA Forest Service, Washington, DC. KORSTIAN, C. E AND W. H. Lona. 1922. The western yel- MADRONO [Vol. 45 low pine mistletoe. Agriculture Bulletin 1112, USDA, Washington, DC. Nixon, K. C. AND C. A. TobzIA. 1985. Within-population, within-host species, and within-host tree sex ratios in mistletoe (Phoradendron tomentosum) in central Tex- as. American Midland Naturalist 114:304—310. PARKER, G. G. 1997. Canopy structure and light environ- ment of an old-growth Douglas-fir/western hemlock forest. Northwest Science 71:261—270. PARMETER, J. R., JR. 1978. Forest stand dynamics and eco- logical factors in relation to dwarf mistletoe spread, impact, and control. Pp. 16—30 in Proceedings of the symposium on dwarf mistletoe control through forest management. General Technical Report PSW-31, USDA Forest Service Pacific Southwest Forest and Range Experiment Station, Berkeley, CA. RICHARDSON, K. S. AND B. J. VAN DER Kamp. 1971. The rate of upward advance and intensification of dwarf mistletoe on immature western hemlock. Canadian Journal of Forest Research 2:313—316. SCHARPF, R. F AND J. R. PARMETER, JR. 1966. Determining the age of dwarf mistletoe infections in red fir. Re- search Note PSW-105, USDA Forest Service Pacific Southwest Forest and Range Experiment Station, Berkeley, CA. SHAW, C. G., II. 1982. Development of dwarf mistletoe in western hemlock regeneration in southeast Alaska. Canadian Journal of Forest Research 12:482—488. SMITH, R. B. 1969. Assessing dwarf mistletoe on western hemlock. Forest Science 15:277—285. SMITH, R. B. 1971. Development of dwarf mistletoe (Ar- ceuthobium) infections on western hemlock, shore pine, and western larch. Canadian Journal of Forest Research 1:35—42. WAGENER, W. W. 1961. The influence of light on estab- lishment and growth of dwarf mistletoe on ponderosa and Jeffrey pines. Research Note PSW-181, USDA Forest Service Pacific Southwest Forest and Range Experiment Station, Berkeley, CA. WIENS, D., D. L. NICKRENT, C. G. SHAW, E G. HAWK- SWORTH, P. E. HENNON, AND E. J. KING.1996. Embry- onic and host-associated skewed adult sex ratios in dwarf mistletoe. Heredity 77:55—63. WIER, J. R. 1916. Mistletoe injury to conifers in the North- west. Agriculture Bulletin 360, USDA Bureau of | Plant Industry, Washington, DC. | MapRONO, Vol. 45, No. 3, pp. 215-220, 1998 POPULATION ECOLOGY OF DUDLEYA MULTICAULIS (CRASSULACEAB); A RARE NARROW ENDEMIC T. ALEJANDRO MARCHANT, RUBEN ALARCON, JULIE A. SIMONSEN|, AND HAROLD KOoPpowITZ? Department of Ecology and Evolutionary Biology University of California, Irvine, CA 92697 ABSTRACT Dudleya multicaulis (Rose) Moran is a rare geophyte, endemic to the coastal plain of southern Cali- fornia. This species has a patchy distribution associated with the imperiled coastal sage scrub community. We investigate the population size, gene flow, and genetic structure at two different geographical scales; discrete colonies within the University of California, Irvine Ecological Reserve and from eight populations sampled across the range of this species. The objective of this protocol was first to determine what constituted a D. multicaulis population as interpreted from protein electrophoretic data, and second to estimate the population genetic structure and amount of gene flow among populations throughout the geographical range. This kind of information is important for rare species and should be used with any plan designed to protect them from further decline. Our data indicate that all individuals within the University of California, Irvine campus now, or in the past, functioned as one population and it is estimated to be in the order of 1200 individual adult plants distributed over approximately 63 acres. The low population genetic structure and high gene flow found at this scale may be explained by pollen transport between colonies. At the regional level we found there is little gene flow among populations across the range of the species, that there is a high level of intrapopulation genetic variation, but more importantly, that there is significant genetic differentiation among populations. We discuss the implications of our results for the conservation of genetic diversity in D. multicaulis. Dudleya (Crassulaceae, Rosales), a New World genus, consists mostly of perennial succulent herbs adapted to arid environments, many of which have restricted geographical distributions and_ specific habitat requirements (Moran 1951; Mulroy 1976). Dudleya species usually occur along the coastal plain of southern California and northern Baja Cal- ifornia with the highest number of species centered around San Diego (Mulroy 1976). Moran (1951) revised the genus and recognized fifty-five taxa grouped into the two subgenera: Dudleya and Has- seanthus (see also Uhl and Moran 1953). Most of the literature on this genus has dealt with the clas- sification and distribution of the different species (e.g., Bartel 1992; Bartel and Shevock 1983, 1990; Boyd et al. 1995; Moran 1951; McCabe 1997; Na- kai 1987; Uhl and Moran 1953). However, our knowledge of its ecology, genetics and habitat re- quirements remains poor (but see Clark 1989; Mul- roy 1979; Thomson 1993). This is particularly im- portant because of two factors that make many spe- cies within Dudleya prone to extinction: narrow en- demism with low population numbers and _ the increasing destruction and fragmentation of, and in- vasion by non-native species, into Southern Cali- fornia’s natural landscape (Keeley 1995; Painter 1995; Schierenbeck 1995). Our work focuses on the population ecology of | ' Present address: Department of Entomology, Univer- “sity of California, Riverside, CA 92697. * Author for correspondence. Dudleya multicaulis, (Rose) Moran a rare, narrow endemic of southern California. The California Na- tive Plant Society (CNPS) recognizes it as rare and endangered in California (List 1B, Elias 1986). The CNPS also notes that the major threats to this spe- cies are from development, road construction, graz- ing, and recreation (Elias 1986), all of which have increased dramatically in southern California. It is then reasonable to expect that populations of D. multicaulis will continue to be destroyed, and with them the genetic diversity of the species represent- ing perhaps adaptations to local environments. Dudleya multicaulis is a small, succulent geo- phyte endemic to the coastal plains of southern Cal- ifornia and is usually found growing on rocky out- crops (Dice 1990, personal observation). Its range extends from northern San Diego County to south- ern Los Angeles County and east to Riverside County. Dudleya multicaulis remains dormant dur- ing the dry months (usually June-—November) as an underground corm. Rainfall coupled with cold nights triggers the start of plant growth (personal observation). Depending on the timing and amount of rain, plants may emerge starting in mid-Novem- ber or as late as mid-January. The cymose inflores- cence usually appears in March and is fully devel- oped by April or early May. Plants flower for about forty days. A plant may have two to several inflo- rescences, each bearing three to many flowers. The small seeds are primarily gravity-dispersed, gener- ally traveling no more than 25 cm (Harker and DeViso unpublished). The mean number of seeds 216 TABLE |. MADRONO [Vol. 45 LOCATION, POPULATION REFERENCE CODES USED IN THE PAPER, NUMBER OF INDIVIDUAL PLANTS SAMPLED FROM EACH LOCATION AND SITE DESCRIPTION FOR THE LOCAL SCALE ANALYSIS. Population name Reference and location code UCI Ecological Preserve Twin Peaks TP Whoeler’s Folley WF Campus View CV Chancellor’s Hill CH UCI Environmental Hazard Site EH CC Sample Micro-habitat size description 40 Rocky Outcrop 40 Rocky Outcrop & Soil 40 Rocky Outcrop 28 Rocky Outcrop 38 Rocky Outcrop 24 Sandy Soil & Grasses Crystal Cove, Corona del Mar per fruit is 26.2 (Nau, = 200, SE = 11.4, Non. = 60, Orjuela and Marchant unpublished). It has been suggested that lichens from the genus Niebla (Ra- malinaceae) serve as a nutrient-rich seed trap for the propagation of Dudleya species on sheer rock (Reifner and Bowler 1995). A Dialictus (Apidae) species successfully pollinates plants of D. multi- caulis, but may not be the sole pollinator (Casares personal observation). Our preliminary data indi- cates that this species can self, but we have not investigated if this results in lower fitness of such progeny. Any attempt to design a management plan to pro- tect a particular species should be based on an un- derstanding of its genetic structure (Franklin 1980), as well as its demographic characteristics and nat- ural history. Therefore we investigated the level of genetic diversity and gene flow of D. multicaulis at two different scales: 1) within and among colonies at the University of California, Irvine. 2) Within and among populations throughout the range of this species. We provide estimated plant densities for each population and we further discuss the impli- cations of these findings for the conservation and protection of extant D. multicaulis populations. MATERIALS AND METHODS Plant sampling protocol: colony scale. We col- lected leaf material of individual plants from three areas; 1) the Environmental Hazard (EH) site lo- cated on the University of California, Irvine (UCI) main campus, which is about one km from the UCI TABLE 2. Ecological Reserve (this population has been iso- lated from contiguous populations of D. multicaulis by roads, parking lots, and construction develop- ments at the University), 2) the UCI Ecological Pre- serve (UCIEP) which is adjacent to the campus, and 3) Crystal Cove State Park about seven km distant (Table 1). We had previously surveyed the Ecological Preserve and recorded all known D. multicaulis populations on a vegetation map and used the natural topography of the area to define discrete colonies; Twin Peaks (TP), Chancellor’s Hill (CH), Woehler’s Folley (WF), and Campus View (CV). These colonies are mostly on rocky outcrops between twenty-five and 100 meters dis- tant from each other. We collected leaf samples from 24—40 individuals within these colonies dur- ing the winters of 1994 and 1995. At each plot we set up a transect and collected leaves from plants that were at least 50 cm apart. From each adult plant we removed one healthy leaf (about 5 cm long) which was transported on ice back to the lab- oratory to be homogenized using a _ phosphate grinding buffer (Soltis et al. 1983), and stored at —80°C for subsequent electrophoresis. This sam- pling protocol imposed no apparent damage to plants. Plant sampling protocol: regional scale. We col- lected leaf material from 8 distinct Orange County populations; Laguna Niguel, Orange, El Toro, Or- tega, San Clemente, Coastal, Laguna Beach, and UC, Irvine (Table 2, Fig. 1). The populations were separated from the UC Irvine population by 7 to LOCATION, REFERENCE CODES USED IN THE PAPER, NUMBER OF INDIVIDUAL PLANTS SAMPLED FROM EACH LOCATION, ESTIMATE OF POPULATION SIZE, SITE DESCRIPTION AND DISTANCE (km) FROM THE UCIEP POPULATIONS. Population name Reference Sample Population Micro-habitat Distance (km) and location code Size size estimate description from UCI Coastal CO 40 300 Sandy Soil and Grasses iz Laguna Niguel LN 40 300 Rocky Outcrop 15 El Toro ET 40 500 Sandy Soil and Grasses 14 Ortega Highway OH 40 200 Sandy Soil and Grasses 29 San Clemente SC 40 1000 Sandy Soil and Grasses a2 Laguna Beach LB 40 1200 Rocky Outcrop 10 Orange OC 40 300 Sandy Soil and Grasses 19 UCI Irvine UC 40 1200 Rocky Outcrop 0) 1998] Fic. 1. The shaded area represents the distribution of the Dudleya genus. The darker area shows the region of high- er species diversity. The inset shows the distribution of Orange County populations of Dudleya multicaulis sam- pled for this study. (Modified from Mulroy 1976). 32 km (Table 2). The main criteria for selecting the populations were their representative distribution of the species in Orange County and population size of about 500-1000 individuals. At each location leaf samples were collected as before. Electrophoretic analysis. Protein electrophoresis procedures followed Soltis et al. (1983). We pre- pared leaf material in 0.1 M phosphate grinding buffer (Ranker et al. 1989; Acquaah 1992) and placed the extract in microcentrifuge tubes. What- mann #3 filter paper strips were soaked in the su- pernatant and placed into wells in a gel of 11.7% starch concentration. Using the staining protocol of Ranker et al. (1983), and buffers and recipes from Soltis et al. (1983) and Acquaah (1992), we screened the following enzymes: Aldolase (ALD), Fructose-1,6-Diphosphate (FBP), and Menadinone Reductase (MNR), resolved in system 9; Malic En- Zyme (ME), Phosphoglucomutase (PGM), Phos- phoglucoisomerase (PGI), and Diaphorase (DIA) resolved on system 6; Acid Phosphatase (ACP), Es- terase (EST), Glucose-6-phosphate Dehydrogenase MARCHANT ET AL.: DUDLEYA POPULATION ECOLOGY 217 (G6P), and Glutamate Oxaloacetate Transaminase (GOT) resolved on system 8; Hexokinase (HXK), Shikimate Dehydrogenase (SKD) and Isocitrate De- hydrogenase (IDH) resolved on system TC-8; and Alcohol Dehydrogenase (ADH), Glyceraldehyde 3- phosphate Dehydrogenase (G3P), 6-phosphoglu- conate Dehydrogenase (6PG), and Malate Dehydro- genase (MDH) were resolved on a morpholine-ci- trate pH 6.1 from Wendel and Weeden (1989a). For the regional scale we utilized a digital camera (AlI- pha Imager 2000, Alpha Innotech Corp.) to obtain a computer image of each gel and photo imaging software (Adobe Photoshop 4.0) to assist in ana- lyzing the banding patterns produced by the allo- zymes. We inferred the genetic banding pattern based on the subunit structure and subcellular compartmen- talization of the enzymes (Gottlieb 1981). We de- noted allozymes alphabetically with the farthest oe 99 moving allozyme designated as “‘a. Statistical analysis. We estimated the following parameters using the computer software BIOSYS- 1 Release 1.7 (Swofford and Selander 1989): (1) the percentage of polymorphic loci using the 99% criterion; (2) the mean number of alleles; (3) the observed and expected heterozygosity; (4) the pop- ulation genetic structure using F' statistics (Nei 1977; Wright 1951); (5) Nei’s 1972 measures of genetic distance and similarity; (6) deviations from Hardy-Weinberg equilibrium using chi-squared goodness of fit test; and (7) generated a cluster analysis using the unweighted pair group method (UPGMA) based on Nei’s (1972) genetic similarity, and the modified Roger’s genetic distance. We es- timated Nm values for the interpopulation gene flow from the equation F;, = 1/(1 + 4Nm) following Wright (1951). Since we do not have estimates of the effective population number (NV) we cannot infer the migration rate. In this study we use Nm values only for comparisons of gene flow between the populations sampled at the local and regional scales. The use of F’,; to estimate gene flow is based in several assumptions, including neutrality of al- leles and genetic equilibrium, which may not hold in the case of recently fragmented populations such as for D. multicaulis. However, it can nevertheless be used as a parameter of comparison between the two scales of analysis used herein. Moreover, we have monitored individual plants for several years and we estimate a lifespan of about 15 years. Thus we feel that the recent human-induced fragmenta- tion has not yet produced significant effects on al- lele frequencies. RESULTS Genetic variation at the colony scale. For the detailed analysis of the UCI population we used more enzymes that for the geographical range. Of the original eighteen enzymes screened we were able to interpret fourteen putative loci from ten en- 21S TABLE 3. PARAMETERS OF GENETIC VARIATION AND Pop- ULATION GENETIC STRUCTURE AT THE LOCAL AND REGIONAL SCALE. P = proportion polymorphic loci at the 99% cri- teria; A = mean number of alleles per locus; H = mean expected heterozygosity; Fy; = the genetic correlation of individuals among sub-populations; F,, = the genetic cor- relation of individuals within each sub-population; and F’, = genes within individuals in the entire population. Scale P A H | ae F Fic DING Local 0.238 1.23 0.82 0.391 0.408 0.028 8.7 Regional 0.582 1.73 0.42 0.196 0.367 0.213 0.9 zymes systems: ALD, DIA-1, DIA-2, SKD, IDH, G6P, EST, PGI-1, PGI-2, G3P-1, G3P-2, PGM, 6PG-1, and 6PG-2. Of the fourteen loci examined, four were polymorphic in at least one of the colo- nies (28.6% at the 99% criterion). DIA-2 and EST were polymorphic in all populations, while PGI-1, PGI-2, G3P-1, G3P-2, PGM, 6PG-1, ALD, SKD, IDH, and G6P were monomorphic in all colonies. The mean number of alleles per locus was 1.3 (SE = (0.1) and the mean genetic diversity (unbiased H,) was 0.085 (SE = 0.046). The observed mean ge- netic diversity, when all the colonies were treated as one population, was not significantly different than the expected (H, = 0.062, SE = 0.036). Two loci deviated significantly from Hardy-Weinberg equilibrium (DIA-1 and EST), and they both showed positive fixation index values suggesting a lack of heterozygosity. The mean percentage of polymorphic loci for all colonies was 23.8% (SE = 3.7) and the mean num- ber of alleles per locus was 1.23 (SE = 0.1). All colonies had an observed level of heterozygosity less than the expected value and a population mean of 0.082 (SE = 0.046). We estimated three F-statistic parameters for the analysis of population structure at the UCIEP: F%;, which represents the genetic correlation of individ- uals among sub-populations, was 0.028; F),, which shows the genetic correlation of individuals within each sub-population, was 0.391; and F,;, which cor- MADRONO [Vol. 45 relates the genes within individuals in the entire population, which was 0.408 (Table 3). The esti- mated amount of gene flow (Nm) was ~8.7. Be- cause of the low number of polymorphic loci and overall low genetic variation, the coefficients for genetic distance and similarity showed little differ- entiation. Genetic variation at the regional scale. At the regional level we had fewer consistently resolvable loci for all populations and could only assay indi- viduals for the following enzymes: PGI, EST, ALD, 6PG, DIA, G3P, FBP, IDH, and MDH. Of all the loci examined 69.2% were polymorphic in at least one population (EST-1, EST-2, 6PG-2, DIA-2, G3P-1, IDH-1, IDH-2, FBP, and MDH). The pro- portion of polymorphic loci present in each popu- lation ranged from 0.4615—0.6154 with a mean of 0.5824, while the number of alleles and expected heterozygosity ranged from 1.69—1.85 and a mean of 1.73 and 0.130—0.208 with a mean of 0.424 re- spectively (Table 3). The calculated F statistic were as follow: F), = 0.196, Fi; = 0.367, Fs, = 0.213 (Table 3). The amount of gene flow calculated from the F,, value was 0.92. The clade showing the relationships among the populations revealed that the genetic distance between them ranged from 13% to 28% (Fig. 2). DISCUSSION Although one of the goals of many conservation programs is to maintain genetic diversity in species that are rare, threatened, and/or have small popu- lation sizes (Frankel and Soulé 1981; Simberloff 1988), researchers have usually neglected genetic considerations when generating plans for rare plant conservation (Barrett and Kohn 1991). We argue that information on the ecology, natural history, demographics, and also the genetic structure of rare and threatened plants is important and necessary for conservation plans. Information that includes all these parameters can increase the probability of | 228 .c4 220 116 2i2 08 204 .00 tp et ps a ps fas as fees ee penn a pa eet $s He fp os oe pee a pas os Gas ae es es eo RIFTASCAAEKAPHLAGAASRHOGAPCAACKRFRAPTAG Coastal eeRTOERTORKIX terteett HKLEKRAALERLDHAARATAKAAKKALLARACKLALLATAARE El Toro =z xz XORNRTRITKI EXEZOARCTORTKAITTATARTKITKIOTRTOCRIOKKTEKRAOTKRIEKIEAT UC Irvine t t RORKTKITTSE KAORTTROITKAETARTORIKAATTAOTKRARTKAITKAEOTRARTKRITKRATOKATOTKRITKIERT Ortega f e =z =z ERECT KTKRFEKRETC KT KXTFERLET KART RAFF KKTTERRTEKRFRKRETARATTKTFTAKFEKE Laguna Niguel 3 ekkerkteete t x SREERE TAREE AEF AHA ARKERERAREDRKTRKRERKRERKE? San Clemente =z SPRITOEKRTFLALIATTARETORKT t SRRETKETRETREPAREARE TRE RRETDARETKERRERKE 2 Laguna Beach ROKKRTKLLEAROAATHRELAAORATTKALKELAR TAK TAKATEAR TARA TAETARLARKTREKKELTRATAKILKATLCAT O range ha th rn tn rn than han Fa nn hn than Fann en ten tt 28 24 220 .16 2iZ 08 204 .00 Fic. 2. UPGMA cluster analysis of the Dudleya multicaulis populations using modified Rogers’ genetic distance. 1998] success for restoration and re-introduction, or even seed-banking strategies. Local scale analysis. The major component of F,, was F;, indicating that the genetic variation found by this data set was due mainly to differences among, individuals, not between colonies (Table 3). The high estimated gene flow (Nm = 8.7) which resulted from the limited genetic structure among colonies in the UCI population (F;; = 0.028) is indicative of a large panmictic population. Our re- sults indicate little genetic variation among colonies at UCI and accordingly the UPGMA cluster anal- ysis groups all individual plants at UCI as one pop- ulation. This includes the colonies found at the UCIEP and the now disjoined EH site on the main campus. Significant deviations from Hardy-Wein- berg equilibrium at each variable locus show that 87.5% of them were in heterozygote deficit. We interpret this deficit as a consequence of a high de- gree of relatedness among individuals possibly re- sulting from founder effects and subsequent mating among relatives. Regional scale analysis. One of the objectives of this study was to estimate historical or long term rates of gene flow (Nm) between populations sam- pled based on the genetic structure (F;;) among populations. We estimated Nm to be 0.92. Accord- ing to Wright’s island model (Wright 1951). Nm values much greater than | result in gene flow over- coming the effects of drift, and thus preventing lo- cal differentiation. On the other hand, values much less than 1 indicate that drift plays a dominant role. F, values in the range of 0.15 to 0.25 indicate great genetic differentiation (Wright 1978). Therefore, our data suggest that overall there is only a small amount of gene flow between populations across the range of the species and that there is a high level of intrapopulation genetic variation. More im- portantly, the data indicate that there is significant genetic differentiation among populations (Fig. 2). Since gene flow may determine the extent to which local populations function as an independent evolutionary unit, low gene flow among popula- tions may produce functionally unique populations which are evolving under different selection re- gimes (Slatkin 1994). Since D. multicaulis is char- acterized by geographically isolated populations, across its entire range, each population of D. mul- ticaulis may foster unique genotypic characteristics that could have evolved and adapted to microhab- itats. Moreover, when species are characterized by small and fragmented populations, genetic drift will dominate population genetic structure and presum- ably increase a population’s vulneribility to extinc- tion (Barrett and Kohn 1991). Concomitant reduc- tion in the genetic variation of a species may hinder its ability to adapt to changes in the environment and augment its susceptibility to disease (Fisher 1930; Hamilton 1982; Beardmore 1983). One of the primary goals of many conservation plans is to re- MARCHANT ET AL.: DUDLEYA POPULATION ECOLOGY 219 tain present genetic variation but, in species that are rare, and possibly endangered, present levels of ge- netic variation among extant populations may al- ready be lower than historical levels. Previous studies conducted by the authors indi- cate that the distribution of D. multicaulis appears to be constricting throughout its range. Approxi- mately 30% of the populations known from Orange County in 1981 were extripated by 1988 (Casares and Marchant personal observation) and possibly up to 50% may now be extinct (Webb and Mar- chant personal observation). An increasing number of species in California are becoming endangered as habitat loss and other threats continue affecting those species that are naturally rare (Skinner and Pavlik 1994; Skinner et al. 1995). Fiedler (1995) suggests that we should focus our conservation ef- forts on the protection of the rarest of species. Con- servation strategies for rare species with popula- tions threatened by anthropogenic activities should include estimations of genetic and demographic characteristics before population fragmentation and habitat conversion significantly affect them. ACKNOWLEDGMENTS This research has been in part supported by grants to the senior author from the California Native Plant Society, the Dean of Undergraduate Studies at UCI, and the UCI Eco- logical Preserve Committee. We thank Diane Campbell for her helpful input and review of the manuscript, Peter Bowler for his support and Alan Thornhill for his advice and encouragement. We also thank the two anonymous reviewers for helpful comments in the earlier manuscript. LITERATURE CITED ACQUAAH, G. 1992. Practical protein electrophoresis for genetic research. Dioscorides Press. Portland, OR. BARTEL, J. A. 1992. Dudleya. Pp. 525-530 in J. Hickman (ed.), The Jepson manual: higher plants of california. University of California Press, Berkeley, CA. BARTEL, J. A. AND J. R. SHEVOCK. 1983. Dudleya calcicola (Crassulaceae), a new species from the southern Si- erra Nevada. Madrono 30:210—216. BARTEL, J. A. AND J. R. SHEVOCK. 1990. Dudleya cymosa subsp. costafolia (Crassulaceae), a new subspecies from the southern Sierra Nevada, Tulare County, Cal- ifornia. Aliso 12:701—704. BARRETT, S. C. H. AND J. R. KOHN. 1991. Genetic and evolutionary consequences of small population size in plants: implications for conservation. Pp. 3-30 in D. A. Falk and K. E. Holsinger (eds.), Genetics and conservation of rare plants. Oxford University Press, New York, NY. BEARDMORE, J. A. 1983. Extinction, survival and genetic variation. Pp. 125-151 in C. M. Schoenwald-Cox, S. M. Chambers, B. MacBryde, and L. Thomas (eds.), Genetics and conservation. Benjamin-Cummings, Menlo Park, CA. Boyp, S., T. S. Ross, O. MISTRETTA, AND D. BRAMLET. 1995. Vascular flora of the San Mateo Canyon Wil- derness Area, Cleveland National Forest, CA. Aliso 14:109—139. Ciark, R. A. 1989. The ecological status and distribution of the endangered succulent Dudleya traskiae, on 220 Santa Barbara Island, California. Ph.D. dissertation. Univ. of California, Santa Barbara, CA. Dice, J. C. 1990. Rare plant monitoring: Many-stemmed Dudleya (Dudleya multicaulis). Crystal Cove State Park Resource Management Plan. State of California Department of Parks and Recreation. Extas, T. S. 1986. Conservation and management of rare and endangered plants. Proceedings from a confer- ence of the California Native Plant Society. FIEDLER, P. L. 1995. Rarity in the California flora: new thoughts and old ideas. Madrofio 42:127—141. FISHER, R. A. 1930. The genetical theory of natural selec- tion. Oxford University Press, Oxford. England. FRANKEL, O. H. AND M. E. SOULE. 1981. Conservation and evolution. Cambridge University Press, Cambridge. England. FRANKLIN, I. R. 1980. Evolutionary change in small pop- ulations. Pp. 135-150 in M. E. Soule and B. A. Wil- cox (eds.), Conservation biology: an evolutionary- ecological perspective. Sinauer, Sunderland, MA. GOTTLIEB, L. D. 1981. Electrophoretic evidence and plant populations. Progress in Phytochemistry 7:1—46. HAMILTON, W. D. 1982. Pathogens as causes of genetic diversity in their host populations. Pp. 269—296 in R. M. Anderson and R. M. May (eds.), Population bi- ology of infectious diseases. Springer-Verlag, New York, NY. KEELEY, J. E. 1995. Future of California floristics and sys- tematics: wildfire threats to the California flora. Ma- drono 42:175—179. McCaseE, S. W. 1997. Dudleya gnoma (Crassulaceae): a new, rare species from Santa Rosa Island. Madrono 44:48-58. Moran, R. V. 1951. A revision of Dudleya (Crassulaceae). Ph.D. dissertation. Univ. of California, Berkeley, CA. Mutroy, T. W. 1976. The adaptive significance of epi- cuticular waxes in Dudleya (Crassulaceae). Ph.D. dis- sertation. Univ. of California, Irvine, CA. Mutroy, T. W. 1979. Spectral properties of heavy glau- cous and nonglaucous leaves of a succulent rosette- plant. Oecologia 38:349—357. NAKAI, K. M. 1987. Some new and reconsidered Califor- nia Dudleya (Crassulaceae). Madrono 34:334—353. Nel, M. 1972. Genetic distance between populations. The American Naturalist 106:283—292. Ne!I, M. 1977. F-statistics and analysis of gene diversity in subdivided populations. Annals of Human Genetics 41:225-—233. PAINTER, E. L. 1995. Threat to the California flora: un- gulate grazers and browsers. Madrono 42:180-188. RANKER, T. C., H. HAUFER, AND P. S. SOLTIS. 1989. Genetic evidence for allopolyploidy in the neotropical fern Hemionitis pinnatifida (Adiantaceae) and the recon- struction of an ancestral genome. Systematic Botany 14:439—447. MADRONO [Vol. 45 REIFNER, R. E., JR. AND P. A. BOWLER. 1995. Cushion-like fruiticose lichens as Dudleya seed traps and nurseries in coastal communities. Madrono. 42(1):81—82. SCHIERENBECK, K. A. 1995. The threat to the California flora from invasive species, problems and solutions. Madrono 42:168—-174. SIMBERLOFF, D. 1988. The contribution of population and community biology to conservation science. Annuals Review of Ecological Systematics 19:473—512. SKINNER, M. W. AND B. M. PAVLIK (eds.). 1994. Inventory of rare and endangered vascular plants of California. California Native Plant Society Special Publication No. 1, 5th ed. California Native Plant Society, Sac- ramento. CA. SKINNER, M. W., D. P. Trpor, R. L. BITTMAN, B. ERTTER, T. S. Ross, S. Boyp, A. C. SANDERS, J. R. SHEVOCK, AND D. W. TAYLOR. 1995. Research needs for con- serving California’s rare plants. Madrono 42:211- 241. SLATKIN, M. 1994. Gene Flow and Population Structure. Pp. 3-15 in L. Real (ed.), Ecological genetics. Prince- ton University Press. Princeton, NJ. SOLTIS, D. E., C. H. HAUFLER, D. C. DARROW, AND G. J. GASTONY. 1983. Starch gel electrophoresis of ferns: a compilation of grinding buffers, gel and electrode buffers, and staining schedules. American Fern Jour- nal 73:9—27. SWOFORD, D. L. AND R. B. SELANDER. 1989. Biosys-1: A computer program for the analysis of allelic variation in population genetics and biochemical systematics. Version 1.7 THOMSON, P. H. 1993. Dudleya and Hasseanthus Hand- book. Bonsall Publications, CA. UHL, H. AND R. Moran. 1953. The cytotaxonomy of Dud- leya and Hasseanthus. American Journal of Botany 40:492-502. WEIR, B. S. AND C. C. COCKERMAN. 1984. Estimating F- Statistics for the analysis of population structure. Evo- lution 38:1358—1370. WENDEL, J. F AND N. EK WEEDEN. 1989a. Visualization and interpretation of plant isozymes. Pp. 5—45 in D. E. Soltis, and P. S. Soltis (eds.), Isozymes in plant bi- ology, Ist ed., Vol. 4. (Series editor: T. R. Dudley. Advances in plant sciences.) Dioscorides Press, Port- land, OR. WENDEL, J. F AND N. EF WEEDEN. 1989b. Genetics of plant isozymes. Pp. 46-72 in D. E. Soltis and P. S. Soltis (eds.), Isozymes in plant biology, Ist ed., Vol. 4. (Se- ries editor: T. R. Dudley. Advances in plant sciences.) Dioscorides Press, Portland, OR. WRIGHT, S. 1951. The genetic structure of populations. Annals of Eugenics 15:323-—354. . 1978. Evolution and the Genetics of Populations, Vol. 4. Variability within and among natural popula- tions. University of Chicago Press, Chicago. IL. MADRONO, Vol. 45, No. 3, pp. 221-230, 1998 SEQUOIADENDRON GIGANTEUM-MIXED CONIFER FOREST STRUCTURE IN 1900-1901 FROM THE SOUTHERN SIERRA NEVADA, CA ScoTT L. STEPHENS! Department of Environmental Science, Policy, and Management, University of California, Berkeley, CA DEBORAH L. ELLIOTT-FISk Department of Wildlife, Fish, and Conservation Biology, University of California, Davis, CA ABSTRACT Historical data collected from eight mixed conifer and four giant sequoia Sequoiadendron giganteum (Lindley) Buchholz (giant sequoia)-mixed conifer plots in the southern Sierra Nevada by George Sudworth in 1900-1901 were analyzed to determine historic forest structure. Although it is not possible to document details of the sampling methodology used by this early forest inventory, the plots were dominated by large trees of several species. Average diameter at breast height (DBH) was 110 cm (43 inches) in the mixed conifer plots and 145 cm (57 inches) in S. giganteum-mixed conifer plots for trees greater than 30.5 cm DBH. Results indicate that both shade intolerant and shade tolerant species were abundant. Average tree density was low at 278 trees/ha (111 trees/acre) in mixed conifer plots and 272 trees/ha (109 trees/acre) in S. giganteum-mixed conifer plots for trees greater than 30.5 cm DBH. The most common size classes are in the medium to large size classes for all S. giganteum-mixed conifer species. This is in contrast to published studies of current stands that have determined small size classes of shade tolerant species are occurring at higher frequencies. Early land uses such as logging and grazing at the turn of the 20th century impacted mixed conifer and S. giganteum-mixed conifer forests of the southern Sierra Nevada. Information from this study can assist in the characterization of the “natural range of variability”’ of these forests which could be used in their restoration and management. The United States Forest Service (USFS) has changed its philosophy of land management. Eco- system management has been selected and in Cal- ifornia it has been defined as the skillful, integrated use of ecological knowledge at various scales to produce desired resource values, products, services, and conditions in ways that also sustain the diver- sity and productivity of ecosystems (Manley et al. 1995). Determining which ecosystem structures are sus- tainable is a complex problem. The USFS has cho- sen pre-historical (the period before the influence of European settlement) ecosystem structure as the desired future condition but there is presently very little quantitative information in this area for the diverse ecosystems found in California. Historical and prehistoric information on the Structure of mixed conifer and S. giganteum-mixed conifer forests is also limited. Information of this type is useful in characterizing the ‘“‘natural range of variability”’ that the ecosystems historically op- erated in and can assist in specifying desirable fu- ture conditions in the restoration and management of these forests. Sources of this type of information include early photographs, personal journals, ' Current address: Natural Resources Management De- partment, California Polytechnic State University, San Luis Obispo, CA 93407. 805-756-2751 FAX 805-756- 1402 e-mail sstephen @calpoly.edu books, forest stand reconstruction from contempo- rary plot data, fire histories, and analysis of early forest inventories. Several investigators have examined past for- est structure in the southern Sierra Nevada, includ- ing sizes of forest aggregations based on tree di- ameter (Bonnicksen and Stone 1982) and forest structure as determined from tree age (Stephenson et al. 1991). Historic inventory data primarily from the northern Sierra Nevada and the Transverse Ranges of southern California have also been ana- lyzed (McKelvey and Johnston 1992). Results in- dicate that shade tolerant species such as Abies con- color (Gordon & Glend.) Lindl. and Calocedrus de- currens (Torrey) Florin have increased in abun- dance since fire suppression was initiated early in the 20th century (Parsons and DeBendeetti 1979). Each type of historic or prehistoric data has ad- vantages and disadvantages. Photographs can give excellent visual representations of past landscapes and can assist in the determination of species com- position, relative tree size, and density; but it is not possible to derive quantitative inventory data from them for analysis (Vankat and Major 1978). Books and early journals can give descriptions of the past landscapes, but in most cases, lack quantitative in- formation. Forest stand reconstruction based on sampling the diameter of current S. giganteum-mixed coni- fer forest trees (Bonnicksen and Stone 1982) can 222 give information on past and current forest struc- ture but also has limitations. This type of analysis attempts to recreate past landscapes based on tree aggregations and stand structure comparisons. In many cases, diameter at breast height (DBH) is used as a surrogate for tree age which can be in- accurate (Oliver and Larson, 1995). Problems with analysis and interpretation from forest aggrega- tions studies have been reviewed elsewhere (Ste- phenson 1987). Fire history investigations can give information on the past fire regime of an ecosystem if appro- priate trees are available for sampling (e.g., old, fire scarred trees that are resistant to decay). These his- tories can give accurate and precise information of the temporal and spatial distribution of the past fire regime, but use of this information to reconstruct past forest structure is difficult because of our lim- ited understanding of the effects of prehistoric fires. Prediction of the effects and behavior of past fires is difficult when the fuel complexes and forest structures they operated within are fundamentally different than the present. The spatial distribution of prehistoric fires has not been investigated thor- oughly making it impossible to estimate how ex- tensive prehistoric fires were. Limited information on the spatial extent of prehistoric fires is available in the southern Sierra Nevada (Kilgore and Taylor 1979; Swetnam et al. 1990; Swetnam et al. 1992; Caprio and Swetnam 1995). Early forest inventories can provide quantitative information on historic forest structure, however, the results from the analyses of these data can be biased. In most cases, the methods used in the in- ventory were not carefully recorded and it is not possible to determine how the samples were se- lected. Reconstruction from early forest inventory data are also limited because so few inventories were conducted. The objective of this paper is to analyze mixed conifer and S. giganteum-mixed conifer forest in- ventory data acquired in 1900-1901 from the southern Sierra Nevada to further our understand- ing of forest conditions and their management at the turn of the 20th century. STuDY SITE AND METHODS Forest survey area. The historic data were ob- tained from the area of the southern Sierra Nevada that is now Sequoia-Kings Canyon National Parks, the southern portion of Sierra National Forest and the northern portion of Sequoia National Forest (Fig. 1). The mixed conifer forest in this area is composed of S. giganteum, Pinus lambertiana Douglas, Pinus ponderosa Laws., A. concolor, and C. decurrens. The inventory also recorded Abies magnifica Andr. Murray and Pinus jeffreyi Grev. and Balf., but they were relatively rare. MADRONO [Vol. 45 Forest survey. Information on species composi- tion and diameter at breast height (DBH) of the mixed conifer and S. giganteum-mixed conifer for- ests was provided from an early forest inventory (Sudworth 1900a). George B. Sudworth, head of the dendrology project in Washington D.C., col- lected timber inventory data while employed by the United States Geological Survey (USGS). The pur- pose of this survey was to inventory the forest re- serves of the Sierra Nevada. The original unpub- lished field notebooks were the source of the in- ventory data analyzed in this paper. The field notebooks contain information from many different vegetation types in the southern Si- erra Nevada but only plots with S. giganteum- mixed conifer or mixed conifer data were used in this analysis. Exact plot locations are not given in the field notebooks but references to rivers, domi- nant mountains, and landmarks are included. Sud- worth may have carried an early USGS map with him during the inventory, but the location of this map is unknown. An incomplete set of photographs associated with Sudworth’s forest inventory are also available at the University of California, Berkeley, Bioscience and Natural Resources Li- brary. Eight mixed conifer plots and four S. giganteum- mixed conifer plots were recorded in the 1900-— 1901 field notebooks (Sudworth 1900a). Locations of plots that were recorded in mixed conifer forests include: 1) Westside of north fork of Kings river, one half way up slope. 2) Bubbs creek near Charlotte creek mouth (trib- utary Kings river). 3) Near sugar pine mill. 4) One mile west of sugar pine sawmill. 5) Sample area near fish camp. Headwaters of Big creek (tributary Merced river) and near head of Fresno river (Lewis fork). 6) Sample area near fish camp. Headwaters of Big creek (tributary Merced river) and near head of Fresno river (Lewis fork) (similar description used in plot 5). 7) Headwater of Chiquito creek; typical of this area down to the middle fork of the San Jouquin riv- et 8) Middle east slope, middle fork of San Jouquin river. Plot locations that were recorded in S. giganteum- mixed conifer forests include (only 3 plots had the locations recorded in the field notebooks): 1) North end of giant forest. 2) Near round meadow giant forest. 3) Round meadow giant forest. Sudworth recorded the species, DBH, and number of 4.9 m (16 ft) logs for each tree greater than 30.5 cm (12 inches) DBH. Plot size was typically 0.1 ha (0.25 acres) with one S. giganteum-mixed conifer | 1998] STEPHENS AND ELLIOTT-FISK: MIXED CONIFER FOREST STRUCTURE IN 1900-1901 223 California nIZzDo 2 124 1232 1222 2te 20° — 9° 118° 117° = 116° 115° 114° 113° 428 42° | | | xe ate | | 408 ao” 39° 39° | ve; ae Inventory Area | 38° Le KN 38° a Oe \\4 a | \ \ 36° 36 _ 359 ap | | 34° aia | | 332 = 050 100-150-200 250 300 3560 400 a | Kilometers 2 a ae, Se ee ener eee 2s es en eer ae a a _.|32° 124° VZs° 122° 121° 120° 119° 118° 117° 116° 115° 114° Fic. 1. George Sudworth’s Southern Sierra Nevada forest inventory area. plot of 0.2 ha (0.5 acres) recorded. Only plots with specified sample areas were used in this analysis. Many other much larger plots were recorded in Sudworth’s field notebooks in S. giganteum-mixed conifer forests, but the area sampled was roughly estimated and were therefore not conducive to quantification. The following plot values were calculated: av- erage basal area per hectare by species, average number of trees per hectare by species, average quadratic mean diameter by species, average per- cent plot basal area by species, and average percent plot stocking by species. Histograms of DBH for each species were also produced. Plot data are summarized and discussed, but a Statistical analysis was not performed. Selection of an appropriate analysis method requires informa- tion on sampling procedures which are unknown for this early forest inventory. RESULTS The smallest tree inventoried in most plots had a DBH of 30.5 cm. No comprehensive inventory data exists for trees below 30.5 cm DBH but the field notebooks had written descriptive observa- tions on regeneration and the impacts from early land uses which are summarized below. Mixed conifer plots. The eight mixed conifer plots were dominated by large trees of several spe- cies. The average quadratic mean diameter at breast height was 110 cm (43 in.) for all trees inventoried. Average tree density was 278 trees/ha or ranged 180—400 tree/ha (111 trees/acre, range 72-160 224 TABLE 1. Basal area Tree (m?/ha) Trees/ha A. concolor 75 113 (13.5) (20.7) C. decurrens 48 55 (11.9) (19.8) P. lambertiana 97 33 (25.3) (10.7) P. ponderosa 33 33 (16.7) (15.3) P. jeffreyi 14 18 (8.9) (10.3) A. magnifica 3 8 trees/acre) for trees greater than 30.5 cm DBH. Av- erage basal area was 271 m/*/ha (1166 ft? /acre). Table 1 summarizes all stand calculations for the mixed conifer plots. The largest trees in the mixed conifer plots were P. lambertiana with an average DBH of 152 cm (60 in.). The largest P. lambertiana recorded in the inventory had a DBH of 305 cm (120 in.). P. lam- bertiana made up only 19% of the trees/ha but con- tributed 36% of the basal area of the plots because of their large size. Abies concolor was the most common tree found in the plots contributing 41% of the individuals in- ventoried. Abies concolor accounted for 28% of the basal area of the plots, second to P. lambertiana. The average DBH of the A. concolor trees was the smallest of the species found in the mixed conifer forests. Pinus ponderosa and C. decurrens both have similar average DBH values. Calocedrus decurrens was more common contributing 20% of plot stock- ing compared to 12% for P. ponderosa. Pinus jef- freyi also had a similar DBH of 112 cm (44 in.) but was uncommon in the plots contributing 6% of plot stocking. Histograms of DBH by species are given in Figure 2. The following comments were written by Sud- worth in the original field notebooks and include information about regeneration and impacts from early European settlers in the mixed conifer plots (Sudworth 1900a). September, 1900. Westside of north fork of Kings river, one half way up slope. No reproduc- tion, sheep grazed till 2 years ago and burned over. September, 1900. Bubbs creek near Charlotte creek mouth (tributary of Kings river), an excep- tionally dense stand. No reproduction, complete shade, fire marks. September, 1900. Near sugar pine mill. Area cut, no reproduction, all timber sound but fire marked. September, 1900. 1 mile west of sugar pine saw- mill. In rich sandy loam, abundant reproduction, MADRONO [Vol. 45 SUMMARY OF AVERAGE STAND CALCULATIONS OF GEORGE SUDWORTH’S 8 MIXED CONIFER PLOTS OF THE SOUTHERN SIERRA NEVADA IN 1900-1901. (STANDARD ERROR) Percent DBH basal Percent (cm) area trees/ha 91 28 40 (3.5) 114 18 20 (157) f2 36 19 (10.5) 117 12 12 (21.4) 112 5 6 (20.7) 74 l 3 0.5—4 ft of all species. All timber severely fire marked at collar. October, 1900. Headwater of Chiquito creek; typical of this area down to the middle fork of the San Joaquin river. 60 concolor seedlings 3-6 ft high. No humans, sheep and cattle grazing of long standing. October, 1901. Heavy shade, no reproduction, humans, 8—10, steep, rocky loam soil, east slope. Sudworth’s notes indicate there were significant human settlement impacts to these ecosystems by the turn of the century. He noted recent evidence of fire in the majority of the plots, and he believed the fires were probably ignited by sheep herders in the area to increase forage production for livestock. In one plot, he noted regeneration of all species was present and in another that only white fir regener- ation was found, indicating regeneration was not uniform in the plots. Forests were relatively open during Sudworth’s inventory (Fig. 3). Repeat pho- tography has not been attempted because photo points were not permanently marked. Sequoiadendron giganteum-mixed conifer plots. The four S. giganteum-mixed conifer plots were dominated by large trees of several species and the average quadratic mean diameter at breast height was 145 cm (57 inches) for all trees inventoried. Omitting S. giganteum data, the average DBH of the remaining trees was 111 cm (44 inches) which is similar to the eight mixed conifer plots (110 cm). Sequoiadendron giganteum groves were also rela- tively open during the inventory (Fig. 4). Average tree density was 272 trees/ha (range 220-290 trees/ha) [109 trees/acre (range 88-116 trees/acre)] for trees greater that 30.5 cm DBH. Av- erage basal area was 2381 m/?/ha (2307 ft* /acre). Omitting S. giganteum data, average basal area of the remaining trees was 121 m*/ha (520 ft* /acre) which is less than 50% of the average basal area of the eight mixed conifer plots. Table 2 summa- 1998] STEPHENS AND ELLIOTT-FISK: MIXED CONIFER FOREST STRUCTURE IN 1900-1901 225 A. concolor 10 Be ae : | = 6 | ve 2 | 02M © OW re TR ANH KHRH ODMH Owor tT DOHaoanNnTtR | Ot NO CMODOewr nwt OD KF BHAN YM FT HO KR DO GD | SNS ee ON NEON NT, EE ONS UN ON DBH (cm) C. decurrens > o = o S o = Lu ee toto =. a oe (a Jat io ie em wa se ee on a ee | oO tr NON OC MO DOT ADA tT DO KF DOH GAANYM FT DO KR DOD D | - —- - b - - - - N N N N N N N N | DBH (cm) P. lambertiana 3] > 4 an) Es x 2 2 1 re 0 a. aos 929 © Oo ~r TRAN HMO ROH OaoOor tT BOON TR oO +t DO © MO DOT nt HD KF DOB DAN MYM FTF DOD KR DO D a b veal — 5 el — — — = N N N N N N N N DBH (cm) P. ponderosa 5 > 4 =8 2 =e 4 0 | P. jeffreyi | 2 > (s) o | g§ 1 | eee | LL. | 0 4 Jun) 92M” OO ere tT KR DAN HD ROH OHOnor tT DO TDA TR Oow~y+r DO © OD OH MO FTF DO KF OB HAN MYM FO KR DOD D - by rae = = — —_ —_ — —N N N N —N N N N DBH (cm) Fic. 2. Histograms of George Sudworth’s eight mixed conifer plots of the southern Sierra Nevada in 1900-1901. rizes all stand calculations for the S. giganteum- mixed conifer plots. Sequoiadendron giganteum were the largest trees in the twelve plots. The largest S. giganteum recorded in the inventory had a DBH of 536 cm (211 in.). S. giganteum made up only 32% of the trees/ha but contributed 77% of the basal area of the plots because of their large size. Compared to the mixed conifer plots, P. lambertiana was a much smaller component in the S. giganteum- mixed conifer plots. Abies magnifica and P. jeffre- yi were rare in the S. giganteum-mixed conifer Fic. 3. Tulare county. Interior of forest on bench of Peppermint Meadow, characteristic of east slope of the Kern river at the head of Dry Creek. Pinus ponderosa, P. jeffreyi, P. lambertiana, Calocedrus decurrens, Abies concolor, 1901. plots. Histograms of DBH by species are given in Figure 5. DISCUSSION Early land uses have impacted S. giganteum- mixed conifer forests of the Sierra Nevada (Ste- phenson 1996; Elliott-Fisk et al. 1997). Livestock grazing and logging were common in many areas of the Sierra Nevada in the late 1800’s (McKelvey and Johnston 1992). In 1900, few S. giganteum groves were in government ownership and logging was thought to be a major concern (Perkins 1900). A total of 470 ha (1173 acres) was privately held inside Sequoia and General Grant National Parks (later to become Grant Grove section of Kings Can- yon National Park) and the majority of the other groves were in private ownership by people who had every right, and in many cases every intention, to cut them into lumber (Perkins 1900). Sudworth’s field notes recorded that the majority of plots had no regeneration. Regeneration proba- bly occurred pre-historically in these forests with the creation of small canopy gaps. Sudworth veri- fied this by recording that very little reproduction 1998] STEPHENS AND ELLIOTT-FISK: MIXED CONIFER FOREST STRUCTURE IN 1900-1901 Fic. 4. Tulare county. Freeman Creek canyon with S. giganteum forest on north slope of basin. Sequoiadendron giganteum 1.75—2.5 meters in diameter and associated species of P. ponderosa, P. jeffreyi, P. lambertiana, A. concolor, C. decurrens, and occasional A. magnifica, 1901. occurred in mixed conifer forests except for occa- sional patches in open spaces (Sudworth 1900a). Patchy, high intensity fires may have created the _ Openings varying in size between 0.1—0.4 ha. in the S. giganteum-mixed conifer forests of the southern Sierra Nevada (Stephenson et al. 1991). Areas that had recently burned with a patchy high intensity fire could have abundant regeneration because duff and surface fuels would have been consumed pro- ducing a mineral soil seedbed, and resources such as light and water were available because of re- duced competition. Since Sudworth apparently did not sample areas that had recently experienced a localized high intensity fire, regeneration was sparse in the sampled plots. The plots, therefore, cannot be assumed to be an unbiased sample of the forest structure of mixed conifer and S. giganteum- mixed conifer forests in 1900-1901. The plots were dominated by large trees of sev- eral species. Both shade intolerant and shade tol- erant species were abundant in the plots. Age dis- tributions can vary dramatically in stands, often 228 TABLE 2. MADRONO OF THE SOUTHERN SIERRA NEVADA IN 1900-1901. (STANDARD ERROR) [Vol. 45 SUMMARY OF AVERAGE STAND CALCULATIONS OF GEORGE SUDWORTH’S 4 S. GIGANTEUM-MIXED CONIFER PLOTS Basal area DBH Percent Percent Tree (m?/ha) Trees/ha (cm) basal area trees/ha A. concolor 84 1S] $1 16 55 (29.9) (44.7) OS) P. lambertiana 32 29 127 6 11 (29.3) 71) 37.1) P. jeffreyi 3 2 114 0.3 l (O) (O) (O) S. giganteum 415 88 282 ae 32 (163.9) (40.3) (41.3) A. magnifica 3 2 122 0.7 1 (O) (O) (O) with no relation to diameter distributions (Oliver and Larson 1995) making it impossible to make conclusions on the age structure of the plots. The most common size classes are in the medium to large size classes for all mixed conifer and S. giganteum-mixed conifer species. This is in con- trast to published studies of current stands in Se- quoia National Park that have determined small size classes of shade tolerant species (A. concolor and C. decurrens) are occurring at higher frequencies relative to larger size classes (Parsons and De- Bendeetti 1979). If all trees less than 30.5 cm DBH are removed from the Parsons and DeBendeetti study, the remaining smaller shade tolerant size classes still have much higher frequencies than those Sudworth recorded. Sudworth’s notes indicate there were significant land use impacts on these forests by the turn of the 20th century. He noted recent evidence of fire in the majority of the plots, and believed most of the fires were ignited by sheep herders. Sheep herders burned to increase forage production and to remove S. giganteum 5 > 4 ) NS Tt N oO oe) i) - Ke) N oO ioe) oO sp] | 9 w » co oO N vt oO ad (o>) = oO + Oo @ oO N oO wo ty o>) [2) N vt i) = oO | — - _— - N WN N N is) Oo Om Oo vt > a + T+ +r WO WO DB\cm} A. concolor 5 4 | o e 31 | S os 2 2 u 1 0 | i 4 e 3 S o 2 | 2, LL 0 —— i) == oO + (ce) © oO N (sp) w yS (o>) oO N T+ oO ioe) oO = oO — 7 ea ae T~T NNN NSN Oo OO OHO FO FT +r wv t+ vt +r OO WO | DBH(cni} | | Fic. 5. Histograms of George Sudworth’s four S. giganteum-mixed conifer plots of the southern Sierra Nevada in 1900-1901. 1998] obstacles from the forest floor which impeded the movement of their livestock (Sudworth 1900b; McKelvey and Johnston 1992). Sudworth also believed livestock grazing in ri- parian areas was affecting the hydrology of some S. giganteum groves. In a previous inventory, Sud- worth believed springs and perennial streams were being effected by excessive sheep browsing which reduced S. giganteum regeneration at the “‘Calave- ras’’ Giant Sequoia grove in the central Sierra Ne- vada (now part of Calaveras State Park in Calaveras and Tuolumne counties) (Sudworth 1900b). Some of Sudworth’s inventory plots were re- cently harvested or in the process of being har- vested during the survey. He also witnessed the im- pacts of early logging in S. giganteum groves when he camped at the Enterprise Mill in 1901 (Sudworth 1900a). This mill operated two years and harvested many large S. giganteum within the present bound- aries of Mountain Home Demonstration State For- est, Tulare County. Most early logging operations in S. giganteum groves wasted a great deal of wood. When the trees were felled, the trunk and upper extremities fre- quently broke into almost useless fragments (Per- kins 1900). Additional waste that also occurred at the sawmill resulted in less than half of the standing volume of each harvested S. giganteum being con- verted into wood products (Perkins 1900). Slash produced by early logging operations in the S. giganteum-mixed conifer ecosystems was enormous. It was frequently 2 meters thick and was thought to be a certain source of future fires (Per- kins 1900). Early logging operations probably con- tributed to large, intense wildfires because of in- creases in surface fuel loads and increased ignitions from field crews. The absence of trees less than 30.5 cm DBH in Sudworth’s plots most likely occurred because they were relatively rare in the sampled plots. The ob- jective of the survey was to assess the timber re- sources in the Sierra Nevada, and therefore, areas with large dense stands were probably favored. Sudworth sampled areas dominated by large trees and regeneration in these areas would be low since the majority of site resources (light and water) were already being used by the existing mature trees. The plots sampled by Sudworth represent histor- ic conditions for areas dominated by very large trees in mixed conifer and S. giganteum-mixed co- nifer forests of the southern Sierra Nevada. How- ever, this analysis does not provide information on areas that were dominated by regeneration of trees of smaller size classes. Information from all forest mosaics is needed to completely describe the nat- ural range of variability that occurred in these eco- systems. This analysis gives information only on areas dominated by large trees, and therefore, is incomplete in describing the historic forest struc- ture. STEPHENS AND ELLIOTT-FISK: MIXED CONIFER FOREST STRUCTURE IN 1900-1901 2 NO Ne) CONCLUSION Although it is not possible to document the sam- pling methods used by this early forest inventory, the mixed conifer and S. giganteum-mixed conifer plots sampled by George Sudworth in the southern Sierra Nevada were dominated by large trees of several species. Shade intolerant and shade tolerant species were both abundant in the plots. This con- trasts to present forests where small shade tolerant species are more common and represents a struc- tural and compositional shift of forest condition. Mixed conifer forests were impacted by livestock grazing, fire, and logging at the turn of the 20th century. Some S. giganteum groves such as the Converse Basin Grove, now part of Sequoia Na- tional Forest in Fresno county, were almost com- pletely clear-cut at this time (Elliott-Fisk et al. 1997). Sheep grazing was intense and fires were frequently ignited by sheep herders to increase for- age production and to remove obstacles. Thus, even 100 years ago, these forests were subjected to sig- nificant European settlement alteration and do not reflect prehistoric conditions. Trees less than 30.5 cm DBH were probably rare in Sudworth’s plots. This analysis does not provide information on areas that were dominated by re- generation or by trees of smaller size classes. In- formation from all forest mosaics is needed to com- pletely describe the natural range of variability that occurred in these forests. Early land use decisions have impacted the pres- ent mixed conifer and S. giganteum-mixed conifer ecosystems of the southern Sierra Nevada. Knowl- edge of these practices and their ecological effects is useful in interpreting and understanding current forest structure. ACKNOWLEDGMENTS We are grateful to Craig Olsen for introducing us to George Sudworth’s field notebooks. We thank Qing Fu Xiao for help in figure production. We thank Bob Martin, Joe McBride, Carla D’ Antonio, Nate Stephenson, and two anonymous reviewers for their helpful comments on this manuscript. LITERATURE CITED BONNICKSEN, T. M. AND E. C. STONE. 1982. Reconstruction of a presettlement giant sequoia-mixed conifer forest community using the aggregation approach. Ecology 63(4):1134-1148. Caprio, A. C. AND T. W. SWETNAM. 1995. Historic fire regimes along an elevational gradient on the west slope of the Sierra Nevada, California. in J. K. Brown (ed.), Proceedings: Symposium on fire in wilderness and park management. USDA Forest Service General Technical Report INT-320. Intermountain Research Station, Ogden, UT. ELLIOTT-FiskK, D. L, S. L. STEPHENS, J. A.-AUBERT, D. Murpuy, AND J. SCHABER. 1997. Mediated Settlement Agreement for Sequoia National Forest, Section B. Giant Sequoia groves: an evaluation. Sierra Nevada Ecosystem project, Final Report to Congress, Adden- 230 dum (Davis: University of California, Centers for Wa- ter and Wildland Resources). KiLGoRE, B. M. AND D. TAYLOR. 1979. Fire history of a sequoia mixed conifer forest. Ecology 60:129—142. MANLEY, P. N, G. E. BROGAN, C. Cook, M. E. FLores, D. G. FULLMER, S. HUSARI, T. M. JIMERSON, L. M. Lux, M. E. McCain, J. A. ROSE, G. SCHMITT, J. C. SCHUy- LER, AND M. J. SKINNER. 1995. Sustaining ecosystems: a conceptual framework. USDA Forest Service report RS5-EM-TP-0O1. Pacific Southwest Region, San Fran- cisco, CA. McKELvey, K. S. AND J. D. JOHNSTON. 1992. Historical perspectives on forests of the Sierra Nevada and the Transverse Ranges of southern California: forest con- ditions at the turn of the century. USDA Forest Ser- vice General Techncal Report PSW-133. Pacific southwest research station, Albany, CA. OLIVER, C. D. AND B. C. LARSON. 1995. Forest stand dy- namics, updated ed. John Wiley and Sons, New York, NY. Parsons, D. J. AND S. H. DEBENDEETTI. 1979. Impact of fire suppression on a mixed-conifer forest. Forest Ecology and Management 2:21—33. PERKINS. 1900. Report on the big trees of California. USDA Division of Forestry. 56th Congress, Ist Ses- sion. Senate Document No. 393. Government printing office. STEPHENSON, N. L. 1987. Use of tree aggregations in forest ecology and management. Environmental Manage- ment 11:1-—5. STEPHENSON, N. L. 1996. Giant sequoia management is- sues: protection, restoration, and conservation. Sierra MADRONO [Vol. 45 Nevada Ecosystem Project, Final Report to Congress, Vol. II, Assessments and scientific basis for manage- ment options (Davis: University of California, Cen- ters for Water and Wildland Resources). STEPHENSON, N. L., D. J. PARSONS, AND T. W. SWETNAM. 1991. Restoring natural fire to the sequoia-mixed co- nifer forests: should intense fire play a role? in Pro- ceedings of the 17th Tall Timbers Fire Conference: high intensity fire in wildlands: management chal- lenges and options. Tallahassee, FL. SuDWoRTH, J. B. 1900a. Unpublished field note books of Sierra Nevada forest reserve inventory. (University of California, Berkeley, Bioscience and Natural Re- sources Library) SuDWoRTH, J. B. 1900b. Stanislaus and Lake Tahoe Forest Reserves, California, and adjacent territory. in An- nual reports of the Department of the Interior, 21st report of the U.S. Geological Survey, 56th Congress, 2nd session, senate document #3. Washington D. C, Government printing office. SWETNAM, T. W., C. H. BAISAN, P. M. BRown, A. C. CaA- PRIO, AND R. TOUCHAN. 1990. Late Holocene fire and climate variability in giant sequoia groves. Bulletin of the Ecological Society of America 71(2):342. SWETNAM, T. W., C. H. BAISAN, A. C. CApRIO, R. TOu- CHAN, P. M. BRown. 1992. Tree ring reconstruction of giant sequoia fire regimes. Final report to Sequoia- Kings Canyon and Yosemite National Parks. Coop- erative agreement DOI 8018-1-002, Laboratory of tree ring research, University of Arizona, Tucson, AZ. VANKAT, J. L. AND J. MAsor. 1978. Vegetation changes in Sequoia National Park, CA. Journal of Biogeography 5:377—402. MaAprono, Vol. 45, No. 3, pp. 231-238, 1998 DISTRIBUTION OF WINTER ANNUAL VEGETATION ACROSS ENVIRONMENTAL GRADIENTS WITHIN A MOJAVE DESERT PLAYA ROBERT W. LICHVAR!, WILLIAM E. SPENCER’, AND JONATHAN E. CAMPBELL? 'USAE Cold Regions Research and Engineering Laboratory, Remote Sensing Center, Hanover, NH 03755 -Department of Biological Sciences, Murray State University, Murray, KY 42071 sDepartment of Geography, University California Los Angeles, Los Angeles, CA 90095 ABSTRACT Discovery of distinct bands of winter annual vegetation on dry playas in the Mojave Desert caused us to investigate the relationships between edaphic factors and plant distribution. We established three tran- sects across the band of vegetation in Deadman Dry Lake and measured soil and plant characteristics. Monolepis nuttalliana, (Schultes) E. Greene Oligomeris linifolia, (M. Vahl) J. KE Macbr. and Schismus barbatus, (L.) Thell., all psuedohalophytic winter annuals, were encountered within the band. Soil texture and salinity appeared to be the primary determinants of the presence of vegetation. Distribution of winter annual vegetation within Deadman Playa appeared to be constrained by low soil moisture towards the outer edge, and high soil salinity towards the inner edge of the vegetation band. Dry lakes and playas have primarily been clas- sified according to soil chemistry, groundwater, sur- face features, and vegetation (Forshag 1926; Thompson 1929). Motts (1965) described six dif- ferent playa types according to surface and hydro- geomorphic conditions. Stone (1956) also devel- oped a classification scheme which divides playas into three categories: wet, dry, and mixed. Wet pla- yas, which have groundwater within 10 m of the surface throughout most of the year, are further grouped into either clay-encrusted, salt-encrusted, or crystal body. Dry playas are divided into two main groups: clay pan or lime pan. Mixed playas have an occurrence of both dry and wet features distributed across the playa. Seasonal ponded water is mostly associated with dry playas due to the hard and impenetrable surface while wet playas pond water less frequently because on gentle slopes and hummocky surfaces resulting from ground water influences. Phreatophytic species, those possessing extreme- ly long roots to reach the water table, are most often associated with wet playas (Hunt 1966). Likewise, bands of halophytic plants, associated with areas with high salt concentrations, and xerophytic shrubs are often found adjacent to dry clay pan pla- yas (Shreve 1925). Halophytic shrubs, which create centimeter high and higher soil mounds, have also been observed on one centimeter high and higher soil mounds located near the edges on dry, clay pan playas (Vasek 1983; Barbour et al. 1990). Similarly, Dahlgren et al. (1997) noted plant communities es- tablished on small dust dunes on the playas at Ow- ens Lake, California. Indeed, these dust dunes offer plants an area of decreased salinity and increased rooting zone. Nevertheless, both dry and wet playas within the Mojave Desert are usually devoid of any vegetation (Stone 1956). While delineating the boundaries of several playas during the winter of 1992, we observed the remnants of vegetation in a band along one particular dry clay pan playa (Deadman Dry Lake) (Lichvar and Pringle 1993). During February 1993, winter annual vegetation was again observed growing in a distinct band on Deadman Dry Lake. This playa lacked both soil mounds and phreatophytic species within the band of vegetation. While vegetation literature from oth- er parts of the country and the world described dif- ferent types of vegetation patterns ranging from playas that are entirely vegetated to dominated by shrub communities along the edges (Wondzell et al. 1990; Smettan et al. 1993; Hoagland and Collins 1997), this particular distribution pattern of vege- tation in relation to environmental gradients within playas of the Mojave Desert has not been well de- scribed. The objective of this research was to in- vestigate edaphic factors that may influence the dis- tribution of winter annual vegetation within Dead- man Dry Lake, San Bernardino County, CA. MATERIALS AND METHODS Study site. The study area (Deadman Dry Lake or Playa) is located within the Marine Corps Air- Ground Combat Center (MCAGCC) at 34°18'36” latitude and 116°08'06” longitude, 12 miles north of the town of Twentynine Palms, San Bernardino County, CA (Fig. 1). Deadman Playa results from local fault block activity and is located near the bottom of a 56,860 ha watershed (Londquist and Martin 1991). The playa itself is 6300 m long along its greatest axis, has an average width of 960 m, and is approximately 361 ha in size (Lichvar and Pringle 1993). Deadman Playa has an active wash that enters from the northwest and diffuses at the Transect Vegetated Band AN] Unvegetated Lake Bed Fic. 1. Location of vegetation band within Deadman Dry Lake, San Bernardino County, CA, and locations of Transects A, B, and C. Also shown is the location of sam- pling points in relation to the vegetation band of Transect A. Vegetation band is not drawn to scale (MCAGCC = Marine Corps Air—Ground Combat Center). northern part of the playa where the water remains until it evaporates. Deadman Playa is connected to Mesquite Dry Lake to the east by an inactive dry wash located at the southeast corner. Therefore, there is no evidence that surface water currently flows out of the southeast side of the playa. Thus, the association of plant occurrence with allocho- thonous soil material from the active wash is un- certain. The elevation at Deadman Playa is 554 m. Ad- jacent areas reach elevations of 1062 m and are gently sloping on all sides except to the north, where the playa abuts a bajada. This bajada is dom- inated by Creosote series (Larrea tridentata [DC.] Cov.), while the remaining areas are dominated by the halophytic saltbushes (Atriplex canescens [Pursh] Nutt. and A. polycarpa [Torrey] S. Watson). The playa itself is unvegetated except for the here- described annual plant community and is classified as a dry, hard, clay-type playa by Stone (1956). Deadman Playa is located in a region character- ized by a warm, hyperarid climate with hot sum- mers and mild winters (Minnich 1991). Freezing temperatures occasionally occur at Deadman Playa during December and January; summer tempera- tures often exceed 35°C. The temperature between January and March 1993 ranged from 25 to 33°C, while annual precipitation averages from 50 to 150 mm each year. A total of 89 mm of precipitation was received between January and March 1993. Plant and soil sampling. Three transects were es- tablished during March 1993 across the winter an- nual vegetation band within Deadman Playa, CA (Fig. 1). Transects were located by selecting rep- resentative locations within the vegetated band. Transects A and B contained five sampling loca- tions, while Transect C contained three sampling locations. Three replicate 1 m’* quadrats, spaced 1 m apart, were placed at each sampling location MADRONO [Vol. 45 along the three transects (Fig. 1). The length of each transect was 160 m, 280 m, and 200 m, re- spectively. Density, richness, and percent cover of all species were recorded and surface soil samples (O-10 cm deep) were taken within each quadrat. Plant densities were recorded for each species as the number of individuals per m’, while species richness was recorded as number of species per m/. Percent cover values were estimated for each spe- cies in each m’. Voucher specimens were verified by and deposited at UCR. Soil samples collected from each of the three replicate quadrats at all sam- pling locations were pooled and subsequently ana- lyzed for texture, electrical conductivity (EC), Wa- ter Content (WC), Ca*?, Mg*’, Na*!, and Sodium Absorption Ratio (SAR) by the U.S. Department of Agriculture, Natural Resources Conservation Ser- vice, National Soil Laboratory, Lincoln, NE (Soil Survey Laboratory Staff 1992). Water content was obtained by analyzing double bagged soil samples measured at 15 bars. In addition, SAR, a derived salinity measure using a sodium paste extract soil solution, was calculated following direct soil anal- ysis. Salinity and sodic status of the soils were clas- sified using criteria established by the Soil Science Society of America (Terminology Committee 1979). Data analysis. Pearson Product-Moment Corre- lation coefficients were calculated to assess the re- lationships among the ten environmental variables and the three vegetation variables (species richness, density, and percent cover). In addition, Canonical Correspondence Analysis (CCA: ter Braak 1988) was used to determine the relationship between the species density values, composition, and environ- mental variable values. Only identified species were input into this analysis. CCA 1s a direct gradient analysis technique that relies on the assumption of unimodal relationships between species and envi- ronmental variables. To determine the components that explained the greatest proportion of variance in the species data, stepwise forward selection of environmental variables was employed. Finally, Monte Carlo permutation analysis was performed on the first ordination axis to determine its’ signif- icance (Manly 1990). Subsequent to CCA, species density data for the 39 sampling locations were classified using TWINSPAN (Hill 1979). TWIN- SPAN is a polythetic, divisive classification tech- nique which bases groupings on percent similarity. To facilitate interpretation of the dendrogram, spe- cies data in the three replicate quadrats in each of | the 13 sampling locations were averaged. RESULTS Soils parameters. Particle size distribution of | soils within Deadman Playa varied within transects (Table 1). Clay and silt content was lower at the outer edge and increased toward the inner edge of the band within each transect, while sand content 1998] TABLE 1. LICHVAR ET AL.: DRY PLAYA PLANT DISTRIBUTION 255 MEANS (+1 SD) FoR CLAY, SILT, SAND, AND COARSE MATERIAL (>2m) CONTENT AND WATER CONTENT OF THE THREE POOLED REPLICATE QUADRATS FOR EACH SAMPLING LOCATION WITHIN DEADMAN PLAYA, CA. Sampling location % clay % silt Transect A 1 33,6 14.3 (0.44) (0.6) 2, 40.9 16.4 (2.0) (1.1) 3 39.2 J3.8 (0.7) (O) 4 41.7 17.1 (3.4) (1.2) 5 43.6 16.2 (2:7) (0.7) Transect B l 17.3 10.2 (3.50) (1.10) 2. 35.3 14.4 (0.7) Glick) 3 42.5 14.2 (2.8) (1.4) 4 43.6 14.7 (1.6) Gh.3% 3 50.4 20.9 (1.7) (1.0) Transect C l 12.6 VA (1.3) (h3) 2 28.3 32:0 (1.9) (522) 3 31.3 32.0 ely (3.4) was higher at the outer edge and decreased toward the inner edge. The percent coarse material, al- though generally low, was highest at the outer edge of the vegetation band, particularly in Transect B. EC and SAR varied widely within each of the tran- sects (0.7 to 10.2 mmhos cm™! and 8.3 to 67.5, respectively). Both EC and SAR were higher near the outer and inner edge of the vegetation band, and lower in the center of the vegetation band along Transect A (Table 2). In Transect B and C, both EC and SAR were lower along the outer edge and cen- ter of the band and increased toward the inner edge. The salinity and sodic status of soils within Deadman Playa varied between normal (non-saline, non-sodic) and saline-sodic (Table 2). Soils were generally sodic to normal at the center of the veg- etation band and saline-sodic at the inner and outer edges of the band. Soil moisture content (WC) in Deadman Playa ranged from 5.4% to 24.2%, and exhibited similar trends in each of the transects. Soil moisture was lowest at the outer edge of the vegetation band and increased towards the inner edge of the band. % sand % coarse WC (%) S271 2333 L373 (0.7) (0.58) (0.63) AD 0.83 19.40 (3.1) (0.29) (0.94) 45.3 0.50 19.00 (0.4) (0.9) (1.02) 41.2 0.50 20.07 (4.1) (O) (1.05) 40.2 0 20357, (2.8) (O) (0.25) IZ 23-1 7.90 (4.50) (3:78) (1.50) 50.2 a 16.10 (0.5) (2.08) (0.60) 43.0 0) 18.70 (4.2) (O) (3.4) 41.0 0) 20.60 (2.4) (O) Cl) 28.7 0) 2A 2 (2.7) (O) (1.1) 80.3 Tea 5.37 (1.9) (0.6) (0.31) 30.0 0.7 15:23 6.2) CIO) (1.82) 30:7 0.7 19.40 (5.7) (0.6) (0.56) Plant parameters. Six winter annual species were encountered within the vegetation band of Dead- man Playa. Monolepis nuttalliana (Schultes) E. Greene, native species considered as a facultative wetland species (FACW) (Reed 1988) that occurs primarily in alkaline areas with clay soils, was dominant and occurred in each of the three tran- sects. Oligomeris linifolia (M. Vahl) J. EF Machr. and Schismus barbatus (L.) Thell. also occurred, but less frequently. Oligomeris linifolia, a native facultative wetland species (FACW), is normally found in areas ranging from creosote scrub to al- kaline sinks, while S. barbatus, an upland species (UPL), is an alien grass occurring primarily in dry areas associated with sandy soils. Nama demissum A. Gray var. demissum was observed on the vege- tated band but was not present in any quadrat. Two additional species were observed within the quad- rats, but could not be identified because the material was immature. Subsequent to averaging the replicates within each sampling locations, species richness of winter annuals observed in Deadman Playa was greatest 234 TABLE 2. MADRONO [Vol. 45 MEANS (+1 SD) FOR ELECTRICAL CONDUCTIVITY (EC), SODIUM ABSORBANCE RATIO (SAR), CA: MG Ratio, AND SALINITY TYPE MEASURED AT 15 BARS FROM SOILS OF THE THREE POOLED REPLICATE QUADRATS FOR EACH SAMPLING LOCATION WITHIN DEADMAN PLAYA, CA. * Terminology Committee (1979). Sampling BC SAR location (mmhos/cm) (mol/mol) Transect A 1 10.15 54.33 (2.91) (8.99) 2, 2.40 24.33 (0.88) (6.85) 3 1.16 2133 (0.87) (1:25) 4 1.50 20.00 (0.14) 227) > 5.99 37.33 (4.69) (9.84) Transect B 1 0.74 1722 (0.092) (4.22) 2 1.10 16.00 (O) (2.65) 3 0.98 14.67 (0.33) (5:13) 4 1.18 18.00 (0.40) (9.64) 5 6.15 67.50 (0.08) (10.61) Transect C 1 0.84 12.00 (0.19) (4.58) 2 0.95 8.33 (0.13) (2.08) 3 2.59 18.93 Ci.13) (8.05) at the outer edge of the vegetation band in Transects A and B and lowest at the inner edge (Table 3). Richness along Transect C was greater than the richness observed along all sampling locations of Transects A and B. Average plant density was low- est at the outer and inner edges and greatest at the center of the vegetation band in Transects A and B. Transect C maintained disproportionately higher density values at all sampling locations. The aver- age percent area covered by vegetation was lowest at the outer and inner edges of the vegetation band in Transects A and B. While, average vegetation cover was low at the inner edge of Transect C along the vegetation band and increased toward the outer edge, percent cover was greater in all zones of Transect C than in Transects A and B. Correlations among soil and plant parameters. Pearson Product Moment Correlation coefficients suggested species density and percent cover exhib- ited a negative relationship with the salinity param- eters of the soil (Table 4). In addition, species rich- ness was correlated with soil texture and soil mois- ture parameters. Ca: Mg Salinity type? 8.18 saline-sodic Mo es sodic 3.30 sodic 4.35 sodic 7.68 saline-sodic 2.09 normal 3.33 sodic : 5.3) sodic 3.35 sodic 5.00 saline-sodic 4.00 normal 4.65 normal 6.76 sodic Canonical Correspondence Analysis (CCA) in- dicated the overall amount of variation in the spe- cies matrix accounted for by the environmental variables was 61.1% on the first axis and 10.5% on the second axis. However, initial CCA results sug- gested high multicollinearity existed between en- vironmental variables. Indeed, Variable Inflation Factors were greater than the threshold amount of 20 for most variables (ter Braak 1988). A stepwise forward selection of environmental variables re- ported that two variables (percent clay and EC) ac- counted for over 90% of the variance explained on the first ordination axis (Fig. 2). In addition, Monte Carlo permutation analysis indicated that the first axis of the ordination was significant when only percent clay and EC were used (P < 0.01). The first ordination axis was negatively correlated with both soil texture (percent clay) and salinity param- eters (EC) (Table 5). The second ordination axis, while similarly negatively correlated with percent clay was positively correlated with EC. TWINSPAN classified the 13 pooled sampling locations into three groups which were biologically 1998] TABLE 3. MEANS (+ | SD) FOR THE PLANT PARAMETERS OF THE THREE POOLED REPLICATE QUADRATS FOR EACH SAMPLING LOCATION ON DEADMAN PLAYA, CA. Transect distance Richness Density Cover (m) (No./m7?) (No./m7?) (%) Transect A 1 23 pis ey | eel (0.6) (18.5) (21) 2 ee 267.3 D57 (0.6) (76.8) (9:5) 5 1.0 233.3 50.0 (O) (28.9) (8.7) 4 1.0 58.3 10.0 (O) (16.1) (2.0) 5 0.0 0.0 0.0 (0.0) (0.0) (0.0) Transect B l 2.33 75.67 5.67 Cla) (29.0) (2.52) 2 1.00 203.09 18.33 (O) (14.43) (2.89) 3 1.67 142.67 43.33 (0.58) (62.53) (12.58) 4 1.00 141.67 24,607 (O) (52.04) (11.24) 5 0) 0) 0) (0) (O) (O) Transect C l 4.67 301.67 59.67 (0.58) (114.73) (9.29) 2 4.33 290.00 82.33 (1.16) (101.89) (4.51) SB 3.67 181.00 82.00 (1,16) (36.37) (3.57) meaningful (Fig. 3). Initially, the two innermost sampling locations (sampling location 5) on Tran- sects A and B were excluded from the analysis as they were unvegetated. TWINSPAN then calculat- ed the first major division in the remaining sam- LICHVAR ET AL.: DRY PLAYA PLANT DISTRIBUTION 239 pling locations based on spatial orientation. The three sampling locations at Transect C were sepa- rated from Transects A and B. Schismus barbatus, found primarily along Transect C, was an indicator of sandy conditions at those sampling locations. The second major division extracted by TWIN- SPAN indicates differences in soil texture and sa- linity. The outer sampling locations were recog- nized as having either lower WC or higher percent clay values than those seen in the center of the veg- etation band. Such conditions may allow M. nut- talliana to be more prevalent at the central sam- pling locations while the upland species, O. linifol- ia, was found exclusively in the outer sampling lo- cations. DISCUSSION Distinct horizontal gradients in EC, SAR, and other salinity factors were observed within tran- sects across the vegetation band within Deadman Playa. Vegetation was largely restricted to normal or sodic (non-saline) sites that generally exhibited higher soil moisture. Three saline-sodic sites were observed in this study. Of those three sites, one was sparsely vegetated and the other two were unveg- etated. Among the salinity parameters examined in this study, EC was described by CCA to be the factor related to salinity that explains the greatest amount of variation in the species richness, density and per- cent cover. EC was negatively correlated with total species density. Plant distribution seems to be re- stricted to sites with an EC of less than 5 mmhos/ cm. Normally, plant distribution decreases toward the center of the playa due to increasing salt con- centrations (Egbahl et al. 1989). Indeed, the aver- age EC value on the unvegetated inner sampling locations was 6.07 mmhos cm™!. While one vege- tated site did exceed an EC of 5 mmhos cm’, this site also had a relatively high percent sand content and WC. The occurrence of low EC at the outer edge of Transect B may be related to the extremely high sand content of this sampling location. It is TABLE 4. PEARSON PRODUCT MOMENT CORRELATION COEFFICIENTS AMONG SOIL AND PLANT PARAMETERS OF POOLED REPLICATE QUADRATS WITHIN DEADMAN PLAYA, CA. Levels of significance are indicated by * P = 0.05 and P = 0.001, all other correlations are insignificant (% coarse percent of coarse fragments, EC = Electrical Conductivity, SAR = Sodium Absorbance Ratio, WC = Water Content in Soil, Ca:Mg = ratio of Calcium to Magnesium). % clay % silt % sand % coarse EC SAR WC Ca:Mg_ Richness’ Density % cover % clay — O30" =U8l=* =O." 0.35 0.50 O.O8F= «0.22 =) 70" —0.41 =0.28 % silt — -—-0.78* -—0.42 0.07 0.02 0.44 0.35 0.24 0.09 0.53 % sand ae O72" =025.- =O! =O sa7? 0:33 0.34 0.18 —0.14 % coarse — =0.24 =0:27) =0.73* —O0A0 0:23 —0.06 =O.21 EC — 0.88** 0.38 0.80** —0.28 —0.67* —0.48 SAR — 0.48 0.57* —0.49 =0:68"- “=O1554 WC oe 0.27 =0i638* =0.36 =O18 Ca: Mg — O27 —0.46 =O] Richness — 0.54* Oey Density % cover = 0.78* 236 Olilin CCA Axis 2 Aa Aah Asad ; Pes p3_ 2B MADRONO [Vol. 45 CCA Axis 1 Fic. 2. CCA triplot of samples, species, and environmental variables (extracted via stepwise forward selection). Monnut = Monolepis nuttalliana, Olilin = Oligomeris linifolia and Schbar = Schismus barbatus. possible that percolation of surface water through the soil profile in the outer edge of Transect B is great enough to prevent the accumulation of salts at the surface. Additional hydrologic and soil po- rosity studies are needed to quantify the movement of water through the soil profile. The distribution of winter annual vegetation in Deadman Playa was also constrained by percent clay and soil moisture. Percent clay was extracted by CCA as the environmental variable which ex- plained the most variance in the species data. As one might expect, percent clay was highly corre- lated with WC (Table 4). Both percent clay and WC were negatively correlated with species richness. Monolepis nuttalliana was clearly the most com- mon species in the vegetated sites (>90% relative density), so variation in plant density primarily rep- resents variation in the distribution of M. nuttalli- TABLE 5. INTRA-SET CORRELATIONS OF ALL ENVIRONMEN- TAL VARIABLES WITH CCA ORDINATION AXES. Variable Axis | Axis 2 % clay (92 =U.15 % silt 0.04 —0.10 % sand 0.44 0.16 % coarse 0.38 0.39 Electrical Conductivity (EC) —0.42 0.57 Sodium Absorption Ratio (SAR) —0.59 0.41 Water Content (WC) =0 35 —0.16 Species Richness 0.9] 0.04 Density 0.53 =0.33 % cover 0.56 =022 ana. The relationship of M. nuttalliana cover to sa- | linity parameters suggests that this species is a | pseudohalophyte (a species present at moist. non- | saline sites within larger saline locations) rather © than a true halophyte (Waisel 1972). The xerophytic grass, Schismus barbatus, was | found primarily on Transect C. Transect C, had | lower average values for all salinity parameters © than Transects A and B. These relatively normal | soils may have allowed S. barbatus to outcompete the more halophytic species, M. nuttalliana and O. | linifolia in many sampling locations. Finally, O. lin- — ifolia was uncommon in most of the sampling lo- | cations. However, it was found in four of the six quadrats in the outer sampling locations. Although there are no clear indicators which explain O. lin- ifolia association with the outer quadrats, it does | appear to be found in areas with relatively high EC and percent sand. In summary, the winter annual taxa encountered | within Deadman Playa during the winter of 1993 — appear to be pseudohalophytes distributed largely | within a non-saline area within a larger saline lo- cation. The distribution of these annual taxa ap- pears to be limited by salinity near the inner edge of the band and low water availability at the outer edge of the band. The vegetation band represents a window of suitable conditions for growth in an oth- | erwise stressful location. It appears that, given suf- ficient precipitation during the winter, soil condi- tions are suitable for the germination and establish- — ment of winter annual vegetation within non-saline | locations of Deadman Playa. 1998] LICHVAR ET AL.: DRY PLAYA PLANT DISTRIBUTION 2357) (n=8) (n=3) No indicators Schbar (n=2) (n=6) Olilin No indicators C1 C2 C3 Al B1 A2 A3 A4 B2 B3 B4 Fic. 3. Results of TWINSPAN analysis of sampling locations expressed as pooled averages of replicate quadrats. Groups distinguished at each level are shown (n = number of sampling locations per group) with indicator species. Olilin = Oligomeris linifolia and Schbar = Schismus barbatus. ACKNOWLEDGMENTS We thank Mike Wilson at the Resources Conservation Service, National Soil Survey (NRCS) Laboratory for analyzing soil materials, Russell Pringle at NRCS, Wet- lands Institute for assisting with soil sampling, Andrew Sanders at the University of California, Riverside for ver- ifying plant specimens from this study, Dan Smith at USAE Waterways Experiment Station, Vicksburg, MS (WES) for assisting with CCA analysis, and Steve Sprech- er and James Wakeley at WES for reviewing of the manu- script. Comments from two anonymous reviewers were helpful. Finally, thanks to Robert Busch at WES for pre- paring the maps utilized in this study. Funding was pro- vided by Marine Corps Air-Ground Combat Center (MCAGCC), Twentynine Palms, CA. We thank Shelly Miller of MCAGCC for her assistance and MCAGCC for logistical support. LITERATURE CITED BARBOUR, M. AND J. Major. 1990. Terrestrial vegetation of California. California Native Plant Society, Special Publication Number 9. DAHLGREN, R. A., J. H. RICHARDS, AND Z. Yu. 1997. Soil and groundwater chemistry and vegetation distribu- tion in a desert playa, Owens Lake, California. Arid Soil Research and Rehabilitation 11:221—244. EGBAHL, M. K., R. J. SOUTHARD, AND L. D. WuitTIG. 1989. Dynamics of evaporite distribution in soils on a fan- playa transect in the Carrizo Plain, California. Soil Science Society of America Journal 53:898—903. FORSHAG, W. 1926. Saline lakes of Mojave Desert. Eco- nomic Geology 21:56—64. Hitt, M. 1979. TWINSPAN: A FORTRAN program for arranging multivariate data in an ordered two-way ta- ble by classification of the individuals and attributes. Department of Ecology and Systematics, Cornell Uni- versity, Ithaca, NY. HOAGLAND, B. W. AND S. L. CoLLins. 1997. Heterogeneity in shortgrass prairie vegetation: the role of playa lakes. Journal of Vegetation Science 8:277—286. Hunt, C. 1966. Plant ecology of Death Valley, California. USDI Geological Survey Professional Paper 509. U.S. Government Printing Office, Washington, DC. LICHVAR, R. AND R. PRINGLE. 1993. Delineation of the “Waters of the United States’? boundary at Deadman Dry Lake, Marine Corps Air-Ground Combat Center, Twentynine Palms, CA. Interagency Agreement M67399 92 MPG1003. LONDQUIST, C. AND P. MARTIN. 1991. Geohydrology and ground-water-flow simulation of Surprise Spring ba- sin aquifer system, San Bernardino County, CA. U.S. Geologic Survey Water-Resources Investigations Re- port 89-4099. MANLY, B. 1990. Randomization and Monte Carlo meth- ods in biology. Chapman and Hall, London, England. MINNICH, R. 1991. Natural resource management plan, MCAGCC, Twentynine Palms, CA. Draft Report, De- partment of Earth Sciences, University of California, Riverside, CA. Motts, W. 1965. Hydrologic types of playas and closed valleys and some relations of hydrology to playa ge- ology. Pp. 73-104 in J. Neal (ed.), Geology, Miner- alogy, and hydrology of US playas. US Air Force Cambridge Research Laboratories, Environmental Research Papers 96. REED, P. 1988. National list of plant species that occur in wetlands: California (Region 0). U.S. Fish and Wild- life Service Biological Report 88 (26.10). SHREVE, E 1925. Ecological aspects of the deserts of Cal- ifornia. Ecology 6:93—103. SMETTAN, U., M. JENNY, AND M. FACKLAM-MONIAK. 1993. Soil dynamics and plant distribution of a sand dune playa microchore catena after winter rain in the Wadi Arabia (Jordan). Catena 20:179—-189. SoIL SURVEY LABORATORY STAFF. 1992. Soil survey lab- oratory manual. Soil Survey Investigations Report No. 42, Version 2.0. USDA, Soil Conservation Ser- vice, National Soil Survey Center. Lincoln, NE. STONE, R. 1956. A geologic investigation of playa lakes. Ph.D. dissertation. University of Southern California, Los Angeles, CA. TER BRAAK, C. 1988. CANOCO—a FORTRAN program for canonical community ordination by [partial] [de- trended] [canonical] correspondence analysis and re- dundancy analysis (version 2.1). Technical Report: LWA-88-02, Agricultural Mathematics Group, AC Wageningen, The Netherlands. TERMINOLOGY COMMITTEE. 1979. Glossary of Soil Science Terms, Soil Science Society of America, Madison, WI. 238 THompson, D. 1929. The Mojave Desert region. U.S. Geo- logic Survey, Water Supply Paper, No. 578. VASEK, E 1983. Plant succession in the Mojave Desert. Crossosoma 9:1—23. WENT, E AND M. WESTERGAARD. 1949. Ecology of desert plants. HI. Development of plants in the Death Valley National Monument, California. Ecology 30:26—38. MADRONO [Vol. 45 WAISEL, Y. 1972. Biology of halophytes. Academic Press, New York, NY. WONDZELL, S. M., J. M. CORNELIUS, AND G. L. CUNNING- HAM. 1990. Vegetation patterns, microtopography, and soils on a Chihuahuan desert playa. Journal of Vegetation Science 1:403—410. MaprRONO, Vol. 45, No. 3, pp. 239-249, 1998 FLOWERING PHENOLOGY AND SEX EXPRESSION OF CROTON CALIFORNICUS (EUPHORBIACEAE) IN COASTAL SAGE SCRUB OF SOUTHERN CALIFORNIA BRADFORD D. MARTIN Department of Biology, La Sierra University, Riverside, CA ABSTRACT Croton californicus is reportedly a dioecious plant species. In the present study, monoecious morphs were observed in four populations of southern California. The prevalence of monoecious individuals was moderately low with relative abundance ranging from 2.5—18.0%. Female plants were more numerous than male plants in two of the four populations with male: female sex ratios ranging from 0.76—1.22. Individual plants were monitored monthly for flowering during the 1994 season. Though flowers were present throughout the year, their abundance varied widely. Flowering of male and female plants started to increase in January, with rapid increases in the production of flowers and fruits during April. Peak flowering and fruiting for female plants occurred in April and May respectively, while for male plants peak flowering occurred in May. A small, second peak flowering and fruiting appeared in October fol- lowed by a drop with almost no flowers produced during December Monoecious plants had a similar seasonal flowering pattern and were on average simultaneously bisexual for 59.6% of the flowering periods. Monoecious morphs also exhibited various degrees of sex expression from almost total maleness to total femaleness in a bimodal distribution pattern skewed towards maleness. This pattern indicates that C. californicus might be dimorphic with phase choices. Flower lifespan differed with gender. Staminate flowers lasted 4.5 d, whereas pistillate flowers lasted 13.2 d and fruits dehisced at 42.3 d. Pistillate flowers and fruits were significantly greater in biomass than staminate flowers. Male plants were significantly larger than female plants for crown diameter, while monoecious morphs were usually significantly larger than either female or male plants. Female plants exhibited higher floral biomass production and reproductive effort than male plants despite their smaller size. Monoecious plants produced more floral biomass than either female or male plants. Spatial segre- gation of male and female plants was also observed in one population. The idea that the sex of an individual of a di- oecious plant species is genetically determined and fixed throughout life was prevalent in past literature (Freeman et al. 1980). However, recent studies in- dicate a significant environmental influence on sex determination and gender modification in ‘‘dioe- cious plants’? (Freeman et al. 1980; Freeman et al. 1984; Lloyd and Bawa 1984; Sakai and Weller 1991; McArthur et al. 1992; Wheelwright and Bru- neau 1992). Many species described as dioecious, subdioecious, or sequential hermaphrodites contain individuals that are sexually labile or inconstant for sex expression (Freeman et al. 1980; Lloyd and Bawa 1984). Among these are several species listed as having well-differentiated sex chromosomes (Freeman et al. 1980). Many dioecious species are known to have pop- ulations with monoecious or hermaphrodite indi- viduals at low to moderate frequencies (Freeman et al. 1980; Willson 1983; Lloyd and Bawa 1984; Bul- lock 1985; Sakai and Weller 1991) and there exist many intergrades between the various sexual sys- tems of plants (Willson 1983; Lloyd and Bawa 1984; Bullock 1985; Thomson et al. 1989). The presence of bisexual individuals in diclinous spe- cies often indicates the occurrence of sex lability and sex change (Freeman and McArthur 1984; Freeman et al. 1984). Most botanical studies on gender modifications have focused on the proxi- mate causes of modification such as the mechanistic effect of hormones, photoperiod, temperature, and nitrogen levels (Lloyd and Bawa 1984). The adap- tive value of sex modifications received little atten- tion until the early 1980’s (Freeman et al. 1980; Freeman et al. 1984; Lloyd and Bawa 1984; McArthur et al. 1992; Ramadan et al. 1994) and to date is still not well understood. Euphorbiaceae exhibit diverse and unusual breeding systems with evidence of considerable gender modification. Bullock (1985) noted variable sex expression in some dioecious species of a trop- ical deciduous forest in Jalisco, Mexico, particular- ly in Jatropha and Bernardia of the Euphorbiaceae. For example, males vary in the tendency to produce single female flowers at the base of the inflores- cence. Croton is interesting for its diversity at this same locality: monoecy typifies most species, but the ratio of male to female flowers varies widely; two species consist of monoecious and female in- dividuals; and two species are dioecious (Bullock 1985; Tejada and Bullock 1988). Sex expression is determined by a genetic system in two species of the Euphorbiaceae (Shifriss 1956; Dellaporta and Calderon-Urrea 1993). However, some Euphorbi- aceae exhibit sex lability due to environmental in- fluence. Pruning, gibberellin, and potassium are known to induce femaleness in a few euphorbs (Heslop-Harrison 1957; Shifriss 1961). Although I-215 Fwy MADRONO [Vol. 45 San Bernardino | EH ote aOR Loma Linda Fic. 1. 1-215 Fwy Map of study area. Study populations of Croton californicus are indicated by triangles. OR = Orchard; PC = Plunge Creek; EH = East Highlands; and SB = San Bernardino. proximate causes have been studied for some eu- phorbs, little is known on the adaptive significance of gender modification. Croton californicus Muell. Arg. is reported to be a dioecious plant (Webster 1993). The presence of monoecious morphs in several populations of southern California (Martin 1995a, b) indicates that C. californicus is in a diclinous condition other than dioecy proper. Although some authors (Lloyd and Bawa 1984; Schlessman 1986) reject the common practice of separating bisexual individuals of di- morphic populations into a separate class such as monoecious, others have shown the importance of monitoring such morphs in order to understand the adaptive significance of sex lability and sex change (Freeman et al. 1984; McArthur et al. 1992; Ram- adan et al. 1994). However, it is critical that quan- titative data on gender expression and gender dy- namics of monoecious individuals be measured (Schlessman 1986). This study quantifies the monthly flowering and sex expression of male, fe- male, and monoecious plants of C. californicus for one season and documents the prevalence, charac- teristics, and phenotypic gender of monoecious in- dividuals. Data are presented that indicate plants of C. californicus change sex and that the species might be “‘dimorphic with phase choices’’ (Lloyd and Bawa 1984). Results contribute to a better un- derstanding of diclinous breeding systems and the adaptive significance of gender modification and sex lability. METHODS Species and study sites. Croton californicus is a subshrub inhabiting sandy soils, dunes, and washes | below 900 m in various plant communities of Cal- ifornia, Arizona, and Baja California. The small | (ca. 5 mm in diameter; Webster 1993), greenish- — white flowers are unisexual (or occasionally mor- — phologically bisexual; see Results), with five sepals _ and no petals. The pistillate flowers have twice | forked styles that resemble the numerous stamens of the staminate flowers. Flowers appear to be vis- ited by small bees and other small generalist insects | (Martin unpublished). Four populations of C. californicus were studied in the Santa Ana River floodplain of the Highland | and San Bernardino areas of southern California | (Fig. 1). The Orchard, Plunge Creek, and East | Highlands populations were components of River- — sidian coastal sage scrub association 2 (Kirkpatrick — and Hutchinson 1977). The San Bernardino popu- lation had a unique floodplain association and was 1998] previously studied (Martin 1995a). The East High- lands population burned approximately one year prior to the present study (Martin 1995b). Flowering phenology and sex expression. Monthly measurements of C. californicus were per- formed in the four study populations between Jan- uary and December 1994. Plot sampling, using 25 m?’ quadrats, was performed to record the flowering and sexual condition for C. californicus as well as to determine population density, cover, and fre- quency. Four to six quadrats were randomly sam- pled in approximately | ha of each study location and all plants were tagged and numbered for monthly monitoring. Plants were monitored at ap- proximately 4 wk intervals (mean time between censuses = 30.4 + 3.2 d) for the duration of the flowering season. Sexual condition was determined by counting all of the flowers and/or fruits on a plant. During the peak flowering months of April through July, counts were determined by utilizing a 0.25 m?* quadrat frame that was centered on the crown of each plant. The quadrat was subdivided with string into four 625 cm’ subquadrats. The northwest subquadrat or quadrant (if crown was larger than the 0.25 m?* quadrat frame) was counted and multiplied by four to determine an extrapolated flower and fruit count. It was necessary to make extrapolated counts due to some female plants pro- ducing more than 500 flowers and fruits and some male plants producing more than 700 flowers. Full floral counts were compared to extrapolated counts and found to be virtually identical. Cover was cal- culated from plant crown diameters which were measured during January 1994. Monoecious morphs. Monoecious morphs were determined to be plants that bore both staminate and pistillate flowers and/or fruits simultaneously or temporally during the 1994 season. Extrapolated counts could not usually be made for monoecious individuals due to an uneven spatial expression of sexuality. Plant structure of C. californicus was very complex so trying to count the number of male or female stems and branches was practically impossible to do. The number of bisexual racemes (male and female flowers adjacent to each other) and bisexual flowers on monoecious morphs was recorded in order to give some estimate of sexual uniformity or segregation within a plant. An estimated floral gender (EFG) was used to determine the maleness or femaleness of monoe- cious morphs (Thomson et al. 1989). This was quantified by the number of male and female flow- ers produced during the 1994 flowering season cal- culated from monthly mean floral count data ad- justed by the longevity of male and female flowers. The EFG was calculated as the number of pistillate flowers divided by all flowers. Thus, EFG describes phenotypic gender or morphological femaleness (Delesalle 1989) as a continuous variable poten- tially ranging from O (complete maleness) to 1.0 MARTIN: FLOWERING PHENOLOGY OF CROTON 241 (complete femaleness). The proportion of the year that monoecious morphs simultaneously bore sta- minate and pistillate flowers and/or fruits was also determined. In order to increase the sample size for monoecious plants, a stratified random sample was tagged outside the quadrats during December 1993 to include in the analysis with quadrat data. Floral morphology, development, and biomass. Floral morphology and development was measured during May and June 1995 in the East Highlands population. Two staminate flowers on ten different male plants (n = 20) and two pistillate flowers on ten different female plants (n = 20) were tagged in mature budding stages and monitored daily starting 1 May 1995 to determine floral longevity. Stages for male flowers were classified as follows: bud opening, anthesis, stamens withering (appressing), and flower abscission. Stages for female flowers and fruits were classified as follows: bud opening, anthesis, styles withering, small fruit, immature fruit, mature fruit, and abortion or dehiscence. The number of sepals and stamens on male flowers and the number of sepals, style branches, and carpels on female flowers were also counted. Flower and fruit biomass was measured on 17 April 1995 and 20 April 1995 in the East Highlands and Plunge Creek populations respectively. In each population, ten floral structures were measured on ten different plants (n = 100) each for staminate flowers, pistillate flowers, and fruits to determine a mean biomass for the different floral structures. Floral biomass was measured fresh due to the ex- treme fragility and small size of flowers. A total fresh and dry floral biomass was also measured to determine the amount of moisture in the fresh floral biomass for a more accurate estimate of floral bio- mass and reproductive effort. An estimate for year- ly mean floral biomass production for male, female, and monoecious plants was calculated from month- ly mean flower count, floral longevity (adjusting for sex differences in flower and fruit production and abortion rate), and mean dry floral biomass data. Climatic data and statistics. Monthly rainfall and mean high temperature data for 1994 was obtained for the San Bernardino area from the Western Re- gional Climate Center in Reno, Nevada. This in- formation was analyzed with monthly flowering data to observe any relationship between climatic and flowering patterns. One-way ANOVA followed by Tukey’s HSD test was utilized to statistically compare gender differ- ences in floral parts, flower and fruit biomass, and crown diameter. Linear correlation was used to ex- plore the relationship between EFG and _ plant crown diameter and the relationships of flowering with monthly rainfall and mean high temperature. Chi-square contingency analysis was used to test for spatial segregation and frequency of the C. cal- ifornicus sexual morphs. 242 MADRONO TABLE 1. CALIFORNIA DURING THE 1994 FLOWERING SEASON. [Vol. 45 RELATIVE ABUNDANCE (%) FOR SEXUAL MORPHS OF CROTON CALIFORINICUS IN FOUR POPULATIONS OF SOUTHERN Relative abundance Population Female n % n Orchard 15 38.5 17 Plunge Creek 38 47.5 35 East Highlands 45 55.6 34 San Bernardino 36 39.6 44 Total = 134 46.1 130 RESULTS Frequency of sexual morphs. Monoecious morphs of Croton californicus were present in all four of the study populations (Table 1). Relative abundance of monoecious morphs was moderately low with a total relative abundance of 9.3% and values ranging from 2.5—-18.0% for the different populations. Overall, females were more numerous than males. Male: female sex ratios ranged from 0.76—1.22 in the various populations with the total male: female ratio being 0.97. There were no sig- nificant differences in the frequency of sexual morphs or male: female sex ratios within or among populations. The Orchard and San Bernardino pop- ulations with male dominant sex ratios exhibited considerably higher numbers of monoecious morphs. Mean Number/Plant January February March April May Fic. 2. June Male Monoecious Male female % n % ratio 43.6 7 18.0 1.13 43.8 7 8.8 0.92 42.0 2 2 0.76 48.4 11 12.1 1:22 44.7 OM 9.3 0.97 Flowering phenology. Flowering of male and fe- male plants started to increase in January, with rap- id increases in the production of flowers and fruits during April (Fig. 2). Peak flowering for female plants occurred in April with peak fruiting in May, whereas for male plants peak flowering occurred during May. Both sexes displayed a sharp decline in flowering through August, with fruits exhibiting a similar pattern lagging until September. A small, second peak flowering and fruiting period appeared in October followed by a drop with almost no flow- ers produced during December. Monoecious plants had a similar seasonal flowering pattern (Fig. 3), but exhibited lower peak flowering periods with a larger discrepancy between the floral sexes at the peak and had more sustained flowering levels through the season. Also, the small, second peak —m— Male Flowers —@— Female Flowers —&— Female Fruits July August September October November December Month Mean number of staminate and pistillate flowers and fruits for male and female plants of Croton californicus ' in four populations of southern California during the 1994 flowering season. Plant sample sizes: January, male n = 130, female n = 134; December, male n = 121, female n {22. 1998] c & Oo = ®o a = =) Z Cc oO (eb) = > Pal = = > L— oO oO = s 3 < = © re] = ‘2 LL Fic. 3. MARTIN: FLOWERING PHENOLOGY OF CROTON —m— Male Flowers —@— Female Flowers —4&— Female Fruits = EI 3 2 2 8 2 > ie = £ 2 = £ 0.90 and only 4.8% of monoecious in- dividuals had EFGs ranging from 0.40—0.80. One monoecious morph from the stratified sample of the San Bernardino population that was monoecious during December 1993 became totally female dur- ing the 1994 season. Also, only one monoecious morph temporally separated its bisexuality during 244 TABLE 2. MADRONO [Vol. 45 PERCENT OF FLOWERING PERIOD (%) OBSERVED IN SIMULTANEOUS BISEXUALITY FOR MONOECIOUS PLANTS OF CROTON CALIFORNICUS IN FOUR POPULATIONS OF SOUTHERN CALIFORNIA DURING THE 1994 FLOWERING SEASON. Percent Length of of flow- simultaneous Length of ates bisexualit floweri iod ' isexuality Owering perio Re Range Mean + SD Range Mean + SD uality Population n (months) (months) (months) (months) (%) Orchard d 1-9 o.3 2 3.4 6-12 96 25 3.2 Plunge Creek 8 1-12 3.5 + 4.0 6-12 8.4 + 2.4 41.7 East Highlands 19 2-12 7.7 + 3.4 7-12 10.5 + 1.9 I San Bernardino 28 1-11 4.8 + 2.8 4-12 8.9 + 2.3 53.9 Total = 62 1-12 5.6 + 2.3 59.6 a5 4—12 9.4 the 1994 season and was also from the San Ber- nardino population. All other monoecious morphs produced male and female flowers and/or fruits si- multaneously. Monoecious plants were simulta- neously bisexual for 5.6 mo of a mean flowering period of 9.4 mo reflecting that 59.6% of the flow- ering period was spent in the bisexual condition (Table 2). Three main temporal floral progressions were ob- served in monoecious morphs for flowering peri- ods: 1) simultaneous bisexuality for the whole flow- ering period; 2) male expression initially followed by simultaneous bisexuality for the majority of the flowering period; and 3) male expression initially followed by simultaneous bisexuality followed by male expression towards the end of the flowering period. These three patterns occurred with a fre- quency of 25.8%, 17.7%, and 16.1% for the 62 monoecious plants respectively. The other 40.4% of the plants exhibited ten other floral progressions in- cluding sex change within the year. Seven plants (11.3%) alternately expressed one sex with simul- taneous bisexuality two or more times during the flowering period. Another seven plants (11.3%) ex- pressed one sex initially and eventually expressed the opposite sex exclusively during the middle or end of the flowering period. One of these seven plants changed from female to male without exhib- iting a period of simultaneous bisexuality. The spatial expression of sex on monoecious plants of C. californicus was quite variable. Many monoecious morphs predominantly expressed one sex with a minority of branches or portions of the plant expressing the opposite sex while some in- dividuals appeared to have a relatively even or equal dispersion and expression of male and female flowers. At a finer level of sexual segregation, bi- sexual racemes and bisexual flowers were also ob- served on 54.8% and 12.9% monoecious plants re- spectively (Table 3). Bisexual flowers were rela- tively uncommon and often malformed with fewer stamens and/or abnormally shaped pistils. One un- usual monoecious plant in the East Highlands pop- ulation produced 12 bisexual flowers and 40 bisex- ual racemes during April 1994. Floral morphology, development, and biomass. Flowers of C. californicus showed variation in number of floral parts. Male and female flowers usually had 5 sepals per flower but some flowers developed only 4; mean number of sepals per sta- minate or pistillate flower of 4.9 + 0.3 and 5.0 + 0.2 respectively. Male flowers averaged 12.2 + 1.5 stamens per flower with a range of 9-15. Female flowers averaged 3.3 + 0.5 carpels per flower with a range of 3—4, and averaged 13.4 + 2.4 style branches per flower with a range of 8-18. The mode for style branches was 12 which is equivalent to four style branches (twice bicleft) per carpel. There was no significant difference between male and female flowers for the number of sepals or for stamens versus style branches. Staminate flowers matured quickly relative to pistillate flowers (Table 4). Male flower buds reached anthesis in 3 d and abscised in 4—5 d. Fe- male flower buds reached anthesis in 5 d and fruits dehisced by 42 d. Spontaneous abortion of fruits occurred on average at 25 d with an abortion rate of 35%. Stamens of male flowers became appressed to each other after shedding pollen just before ab- Scission. Staminate and pistillate flowers were very small with fresh biomass of female flowers weighing twice as much as male flowers (Table 5). Female fruit biomass was almost ten times more than fe- male flowers and almost 20 times more than male flowers. Both female flowers and fruits were sig- nificantly greater (P < 0.001) in biomass than male flowers. Dry floral biomass was 28.0%, 30.3%, and 28.5% of the fresh floral biomass for male flowers, female flowers, and female fruits respectively. These differences are small enough that either fresh or dry floral biomass should approximate reproduc- tive effort equally well. Plant size and reproductive effort. Male plants of C. californicus were found to be larger when com- 1998] TABLE 3. NUMBER OF BISEXUAL RACEMES AND BISEXUAL FLOWERS OBSERVED ON MONOECIOUS PLANTS OF CROTON CALIFORNICUS IN FOUR POPULATIONS OF SOUTHERN CALIFORNIA DURING THE 1994 FLOWERING SEASON. Bisexual flowers Bisexual racemes Percent of Percent of plants with Number of plants with plants with Number of plants with flowers Number of racemes Number of MARTIN (%) flowers (%) flowers racemes racemes Population Orchard 0.0 [225 Sik 10 Plunge Creek 365 79.0 15 10 34 5 19 East Highlands San Bernardino : FLOWERING PHENOLOGY OF CROTON 245 0.0 12.9 Bond 54.8 139 Total pared to female plants (Table 6). Male crown di- ameter was larger in all populations and signifi- cantly larger (P < 0.001) when totals were analyzed for the four populations. Monoecious morphs were significantly larger (P < 0.001) than males or fe- males in two populations as well as for totals of the four populations, and then only with stratified random samples, which have much larger mean di- ameters than the monoecious plants in the quadrats. The mean crown diameters for monoecious plants with male dominant EFGs (<0.50) and female dominant EFGs (>0.50) was 56.6 + 26.0 cm and 47.7 + 10.9 cm respectively. These mean crown diameters were not significantly different and crown diameter was not correlated with EFG. Floral biomass estimates indicate that females al- located more energy into reproduction than males. Female plants produced a mean dry floral biomass of 7.60 g-plant”'-yr~' while males produced 7.11 g-plant''-yr-'. This difference becomes more pro- nounced when considering that females were small- er plants (60.1% of the mean crown area of male plants) and produced a proportionally higher amount of floral biomass than male plants. Utilizing mean crown area, females produced 90.0 g-m’-'-yr~! while males produced 50.6 g-m?"!-yr“!. Monoecious plants produced a mean dry floral bio- mass of 7.65 g-plant™'-yr~' for male flowers and 4.36 g-plant"'-yr~' for female flowers and fruits for a total mean dry floral biomass of 12.01 g-plant'-yr-'. These plants produced 49.5 g-m? ~!.yr~-! when utilizing mean crown area to calculate floral biomass production. Population spatial patterns. Plant densities for C. californicus were relatively high in three of the four populations (Table 7). Cover values were also high- est in these three populations. However, the cover in the recently burned East Highlands population was considerably lower. Frequency values were all 100% due to the large quadrat size used in sam- pling. Spatial segregation of the sexes was observed only in the East Highlands population. There was a significant difference (P < 0.001) in the segre- gation of male and female plants in the four quad- rats. Two quadrats were male dominant, with 72.2% of 18 plants and 77.8% of 9 plants being male, whereas the other two quadrats were female dominant, with 93.8% of 16 plants and 63.9% of 36 plants being female. This spatial segregation is even more notable when considering the large quadrat size (25 m7?) used. DISCUSSION Flowering phenology. Flowering of Croton cal- ifornicus occurs to some degree throughout the whole calendar year. Male and female plants exhib- ited similar seasonal patterns with synchronous pe- riods of maximal flowering, as found in other di- oecious species (Lloyd and Webb 1977; Bullock 246 MADRONO [Vol. 45 TABLE 4. NUMBER OF DAYS (MEAN + SD) FOR FLORAL DEVELOPMENT OF STAMINATE AND PISTILLATE FLOWERS ON MALE AND FEMALE PLANTS OF CROTON CALIFORNICUS. Values are from the East Highlands population during May 1995. Bud opening refers to the uncurling of stamens or styles in flower buds. Small fruits were <3 mm diameter; immature fruits 3—5 mm diameter; mature fruits >5 mm diameter. Floral development Male Female Floral stage n (days) n (days) Bud opening 20 1.0 + 0.0 20 2d ee 20 Anthesis 20 20 = 0.0 20 5.02 2:8 Stamen or style withering 20 3.0 = °O:8 20 Sele 2S Small fruit — — 20 13-227 16 Immature fruit — — 18 18:6, 171 Mature fruit — — 13 22.4 224 Abortion — — 7 24.6 + 4.3 Abscission or dehiscence 20 AS 2 09 13 AQ 23.1 and Bawa 1981; Armstrong and Irvine 1989; Carr 1991; Aronne et al. 1993). The correlation between rainfall and flowering indicates that flowering is primarily related to soil moisture. Flowering pat- terns for monoecious plants were very similar to male and female plants. The larger discrepancies between male and female flower production reflects the higher number of monoecious morphs with male dominant EFGs. The more sustained flower- ing patterns appeared related to the larger plant size of monoecious individuals compared to male and female plants. Larger plants possess very deep tap- roots (Martin unpublished) which allow monoe- cious plants to access moisture better and longer during the hot, dry summer months. Sex expression. Over 90% of C. californicus plants were constant in sex expression as male or female plants during the 1994 flowering season. The frequency of monoecious morphs observed in this study was considerable in some populations (18%) and many times higher than previously re- ported (Martin 1995a, b). This discrepancy is due to previous study sampling performed only once during a flowering season versus the 12 monthly samples measured in this study. Not all monoecious plants are simultaneously bisexual at any given time of the year. The one monoecious plant that changed sex between the 1993-1994 flowering sea- son and the 14 others that changed sex during the 1994 flowering season indicate that many monoe- cious plants and possibly some male and female plants are sexually labile and capable of sex change. Additional sex changes have been ob- served in C. californicus to support this hypothesis (Martin unpublished). Although monoecious plants reflect a complete array of sex expression with various degrees of maleness and femaleness, the bimodal distribution pattern for EFGs appears to indicate that C. cali- fornicus exhibits some stability to be either male or female. Bimodal gender distributions are indicative of plant species that are either diphasic or dimor- phic with phase choices (Lloyd and Bawa 1984). However, one could interpret this bimodality as di- morphism with male and female inconstancy ex- hibiting long tails of gender adjustment that merge into one another (Lloyd and Bawa 1984). The skew towards maleness in this bimodal distribution may reflect that male plants more often than female plants become monoecious or spend a longer period of time with a male dominant EFG if plants of C. californicus are changing sex. Similar bimodal dis- tribution patterns for EFGs or phenotypic genders have been observed in sex changing or sex labile plant species (Condon and Gilbert 1988; Delesalle 1989; Allison 1991). TABLE 5. FLORAL BIOMASS (MEAN + SD) FOR STAMINATE AND PISTILLATE FLOWERS ON MALE AND FEMALE PLANTS OF CROTON CALIFORNICUS. Values are from two populations during April 1995. All flowers and fruits were weighed fresh. Sample sizes are numbers in parentheses (n). All values differed significantly (P < 0.001) from ANOVA comparing different floral conditions. Floral biomass (g) Population Male flower Female flower Female fruit Plunge Creek 0.008 + 0.002 0.013 + 0.003 0.125 + 0.023 (100) (100) (100) East Highlands 0.007 + 0.001 0.014 + 0.004 0.147 + 0.027 (100) (100) (100) Total 0.007 = 0.002 0.014 + 0.004 0.136 =:0.027 (200) (200) (200) 1998] TABLE 6. CROWN DIAMETERS (MEAN + SD) FOR SEXUAL MORPHS OF CROTON CALIFORNICUS IN FOUR POPULATIONS OF SOUTHERN CALIFORNIA. Sample sizes are numbers in parentheses (n). K = number of 25 m? quadrats measured. a = monoecious morphs in quadrats. b = monoecious morphs in quadrats and from stratified random sample. F = female, M = male, Mon = monoecious’. Underlining between sexual morphs for Tukey’s pairwise comparisons designates nonsignificant differences. Tukey’s Crown diameter (cm) Male 30.0 + 19.9 pairwise comparisons Monoecious? Monoecious?® Female 25.0 2 117 Population Orchard F M Mon 0.16 40.6 + 21.0 (7) (17) a2 = 22:5 50 7-228 (15) 38.1 + 15.9 MARTIN: FLOWERING PHENOLOGY OF CROTON 247 F Mon M <0.05 45.1.-=°22.5 Plunge Creek (8) (35) (7) 45.5 + 29.0 57.4 + 15.9 (38) 2957 = 175 = ae <0.001 + 23.4 38.2 East Highlands (19) (34) 43.5 + 21.8 61.0 + 29.7 (45) 34.2 + 16.4 : “| LL, <0.001 37.3 + 14.8 San Bernardino (11) (44) 40.0 + 18.5 AZ. 22 2351 (36) 32.8 + 16.6 <0.001 55.6 + 24.9 18 Total 2) 2 (130) (134) TABLE 7. DENSITY, COVER, AND FREQUENCY OF CROTON CALIFORNICUS IN FOUR POPULATIONS OF SOUTHERN CALIFOR- NIA. Cover was calculated from plant crown diameters. K = number of 25 m? quadrats sampled. Density Cover Frequency Population K — (no./ha) (%) (%) Orchard 6 2600 13 100.0 Plunge Creek 4 8000 14.8 100.0 East Highlands 4 8100 99 100.0 San Bernardino 4 9900 14.5 100.0 Total = 18 6644 oD 100.0 When a plant is monoecious, its bisexual expres- sion is observed for a majority (59.6%) of the flow- ering period and 25.8% of the plants exhibited si- multaneous bisexuality for the whole period. This indicates the possibility of geitonogamous self-fer- tilization if C. californicus is self-compatible. Cro- ton californicus exhibited bisexuality at a local lev- el within an individual with a majority of monoe- cious individuals expressing bisexual racemes and very few individuals expressing bisexual flowers. Bisexual flowers were often malformed and may not be functional. Sex characters. Staminate flowers of C. califor- nicus were morphologically similar to the pistillate flowers. They were also short-lived and more nu- merous compared to the pistillate flowers. Similar floral characteristics have been observed in a mon- oecious species of the Euphorbiaceae (Bawa et al. 1982). Size appears to be a secondary sex character of C. californicus in which males are larger than fe- males, as has been previously reported (Martin 1995a, b). Most secondary sex characters of plants are subtle and usually are expressed in terms of growth, resource allocation, and timing or longevity (Lloyd and Webb 1977; Richards 1986). Energy ex- penditure into vegetative growth per plant structure may likely have been comparable for male and fe- male plants since there is no significant difference in leaf size (Martin 1995a) and overall leaf density appeared similar. Estimates of floral biomass production for the year may reflect greater reproductive effort for fe- males when compared to males. The smaller female plant size may be a result of this higher expenditure of energy for reproduction (Lloyd and Webb 1977; Shea et al. 1993; Allen and Antos 1993; Cipollini et al. 1994). Higher reproductive effort of females has been observed in many studies (Wallace and Rundel 1979; Allen and Antos 1993; Armstrong and Irvine 1989; Dawson et al. 1990; Vasiliauskas and Aarssen 1992; Shea et al. 1993) which often results in a shorter lifespan or higher mortality rate for females (Lloyd and Webb 1977; Allen and An- tos 1993; Cipollini et al. 1994). Although not sig- nificantly different, more females than males died in the present study. 248 Monoecious plants were larger than male and fe- male plants with 1.78 and 2.88 times more mean crown area respectively. Understandably, these plants have the potential of producing more floral biomass. Estimates of total mean dry floral biomass indicate that monoecious morphs produce 1.69 times and 1.58 times more floral biomass than male and female plants respectively. It is also interesting to note that monoecious individuals also produce 7.6% more male floral biomass than male plants and only 57% of the female floral biomass of fe- male plants. This 1.76 male: female floral sex ratio in part reflects the higher proportion of male dom- inant EFGs. It may also reflect that male plants be- come monoecious more often than female plants. Population spatial patterns. Density and cover data for populations of C. californicus were consid- erably higher than previously reported (Zembal and Kramer 1984; Martin 1995a, b) in all populations except Orchard. Total density and cover were also higher. Female biased sex ratios, as observed in this study and previously (Martin 1995a, b), are rarer in nature than male biased sex ratios (Willson 1983; Richards 1986). The sex ratio in plant populations may not be equal due to microhabitat segregation by the different sexes (Handel 1983; Waser 1984; Bierzychudek and Eckhart 1988; Dawson and Bliss 1989; Sakai and Weller 1991; Shea et al. 1993). Gross spatial segregation of the sexes was observed in the East Highlands population. However, envi- ronmental factors (e.g., soil moisture, soil fertility, etc.) were not measured in this study. It should be noted that this population was a young, fire-dis- turbed population. Although the overall sex ratio observed in this study was female dominant, the high number of male dominant monoecious plants would functionally create an overall male dominant sex ratio of 1.14. In most subdioecious species, one sex (male or female) is relatively constant in sex expression while the opposite sex is labile (Mc- Arthur and Freeman 1982; Freeman and McArthur 1984). The high proportion of male dominant mon- oecious individuals observed in this study may in- dicate that males of C. californicus are the more labile sex. Dimorphism with phase choices versus diphasy. Dimorphic species with phase choices contain two genetic morphs that are predisposed but not irrev- ocably committed toward male and female modes, respectively, and in certain conditions are induced to switch to the other mode (Lloyd and Bawa 1984). The two morphs switch at different thresh- olds of cueing factors so the sex ratio varies ac- cording to conditions. It is quite possible that C. californicus is a dimorphic species with phase choices. There are several observations that support this hypothesis. First, sex change has been observed in C. californicus within and between flowering seasons. Second, the frequency of monoecious morphs was relatively high (up to 18%), consider- MADRONO [Vol. 45 ing the data represents one season only. Studies conducted over several seasons typically reveal higher numbers of labile or monoecious individuals (McArthur and Freeman 1982). Third, sex ratios varied from 0.76—1.22 among populations suggest- ing that sex ratios are not stable and may be chang- ing with phase choices. Also, the San Bernardino population sex ratio in 1991 was 0.78 (female dom- inant; Martin 1995a, and in 1994 was 1.22 (male dominant). When monoecious morphs are added into ‘“‘male’’ or “‘female’’ categories the sex ratio for San Bernardino becomes 1.46. Fourth, monoe- cious EFGs exhibited bimodal gender distribution. The skew towards maleness might be explained by a lower threshold value for phase choice in males than for females (Lloyd and Bawa 1984). Diphasy seems unlikely in C. californicus for two reasons. First, it appears that there are two ge- netic morphs (male and female) which would in- dicate dimorphism with phase choices. Diphasic species contain individuals belonging to one genet- ic class but choose their sexual mode in a given season according to circumstances (Lloyd and Bawa 1984). Second, small, young sapling plants have been observed to be male or female in ap- proximately equal frequency and the mean plant size for females of all ages is smaller than for males. Most diphasic species choose male expres- sion early in life and become female as they in- crease in size (Lloyd and Bawa 1984). Further stud- ies are being conducted to follow C. californicus through several seasons to conclusively determine the breeding system. ACKNOWLEDGMENTS The author thanks Gary Bradley for the considerable time he spent in collecting data, discussion, and statistical advisement, and Russell Martin, Sherylle Martin, James Smith II, Jeffrey Tosk, Ryan VanDeventer, and James Wil- son for their assistance with the analysis of this study and manuscript. Special thanks go to Deborah Marr and Geof- frey Levin for their thorough reviews of the manuscript and James Ashby of the Western Regional Climate Center for providing climatic data. LITERATURE CITED ALLEN G. A. AND J. A. ANTOS. 1993. Sex ratio variation in the dioecious shrub Oemleria cerasiformis. Amer- ican Naturalist 141:537—553. | ALLISON, T. D. 1991. Variation in sex expression in Can- ada yew (Taxus canadensis). American Journal of Botany 78:569—578. ARMSTRONG, J. E. AND A. K. IRVINE. 1989. Flowering, sex | ratios, pollen-ovule ratios, fruit set, and reproductive | effort of a dioecious tree, Myristica insipida (Myris- ticaceae), in two different rain forest communities. American Journal of Botany 76:74-85. | ARONNE; G., C. C. WILCOCK, AND P. PIZZOLONGO. 1993. Pollination biology and sexual differentiation of Os- _ yris alba (Santalaceae) in the Mediterranean region. Plant Systematics and Evolution 188:1—16. Bawa, K. S., C. J. WEBB, AND A. F TuTTLE. 1982. The 1998] adaptive significance of monoecism in Cnidoscolus urens (L.) Arthur (Euphorbiaceae). Botanical Journal of the Linnean Society 85:213-—223. BIERZYCHUDEK, P. AND V. ECKHART. 1988. Spatial segre- gation of the sexes of dioecious plants. American Naturalist 132:34—43. BuLLock, S. H. 1985. Breeding systems in the flora of a tropical deciduous forest in Mexico. Biotropica 17: 287-310. AND K. S. BAWA. 1981. Sexual dimorphism and the annual flowering pattern in Jacaratia dolichaula (D. Smith) Woodson (Caricaceae) in a Costa Rican rain forest. Ecology 62:1494—1504. Carr; D. E. 1991. Sexual dimorphism and fruit produc- tion in a dioecious understory tree, /lex opaca Ait. Oecologia 85:381—388. CIPOLLINI, M. L., D. A. WALLACE-SENFT, AND D. FE WHIGHAM. 1994. A model of patch dynamics, seed dispersal, and sex ratio in the dioecious shrub Lindera benzoin (Lauraceae). Journal of Ecology 82:621-—633. ConpDoN, M. A. AND L. E. GILBERT. 1988. Sex expression of Gurania and Psiguria (Cucurbitaceae): Neotropi- cal vines that change sex. American Journal of Bot- any 75:875—884. Dawson, T. E. AND L. C. BLIss. 1989. Patterns of water use and the tissue water relations in the dioecious shrub, Salix arctica: The physiological basis for hab- itat partitioning between the sexes. Oecologia 79: 332-343. , J. R. EHLERINGER, AND J. D. MARSHALL. 1990. Sex-ratio and reproductive variation in the mistletoe Phoradendron juniperinum (Viscaceae). American Journal of Botany 77:584—589. DELESALLE, V. A. 1989. Year-to-year changes in pheno- typic gender in a monoecious cucurbit, Apodanthera undulata. American Journal of Botany 76:30—39. DELLAPORTA, S. L. AND A. CALDERON-URREA. 1993. Sex determination in flowering plants. Plant Cell 5:1241— 1251, FREEMAN, D. C. AND E. D. MCARTHUuR. 1984. The relative influences of mortality, nonflowering, and sex change on the sex ratios of six Atriplex species. Botanical Gazette 145:385—394. , AND K. T. HARPER. 1984. The adaptive Rienineance of sexual lability in plants using Atriplex canescens as a principle example. Annals of the Mis- souri Botanical Garden 71:265—277. , K. T. HARPER, AND E. L. CHARNOV. 1980. Sex change in plants: Old and new observations and new hypotheses. Oecologia 47:222—232. HANDEL, S. N. 1983. Pollination ecology, plant population structure, and gene flow. Pp. 163-211 in L. Real (ed.), Pollination biology. Academic Press, Inc., New York, NY. HESLOp-HARRISON, J. 1957. The experimental modification of sex expression in flowering plants. Biological Re- views of the Cambridge Philosophical Society 32:38— 90. KIRKPATRICK, J. B. AND C. EK HUTCHINSON. 1977. The com- munity composition of Californian coastal sage scrub. Vegetatio 35:21—33. LLoyp, D. G. AND K. S. Bawa. 1984. Modification of the gender of seed plants in varying conditions. Evolu- tionary Biology 17:255—338. AND C. J. WEBB. 1977. Secondary sex characters in plants. Botanical Review 43:177-—216. MARTIN: FLOWERING PHENOLOGY OF CROTON 249 MarTIN, B. D. 1995a. Monoecious morphs in Croton cal- ifornicus (Euphorbiaceae). Madrono 42:323-331. . 1995b. Postfire reproduction of Croton califor- nicus (Euphorbiaceae) and associated perennials in coastal sage scrub of southern California. Crossosoma 21:41—56. McArtTuHuR, E. D. AND D. C. FREEMAN. 1982. Sex expres- sion in Atriplex canescens: Genetics and environ- ment. Botanical Gazette 143:476—482. : , L. S. LUCKINBILL, S. C. SANDERSON, AND G. L. NOLLER. 1992. Are trioecy and sexual lability in Atriplex canescens genetically based?: Evidence from clonal studies. Evolution 46:1708—1721. RAMADAN, A., A. EL-KEBLAWyY, K. H. SHALTOUT, AND J. LovetTt-DousT. 1994. Sexual polymorphism, growth, and reproductive effort in Egyptian Thymelaea hir- suta (Thymelaeaceae). American Journal of Botany 81:847—-857. RICHARDS, A. J. 1986. Plant breeding systems. George Al- len & Unwin, London. SAKAI, A. K. AND S. G. WELLER. 1991. Ecological aspects of sex expression in subdioecious Schiedea globosa (Caryophyllaceae). American Journal of Botany 78: 1280-1288. SCHLESSMAN, M. A. 1986. Interpretation of evidence for gender choice in plants. American Naturalist 128: 416-420. SHEA, M., P. M. Dixon, AND R. R. SHARITZ. 1993. Size differences, sex ratio, and spatial distribution of male and female water tupelo, Nyssa aquatica (Nyssaceae). American Journal of Botany 80:26—30. SHIFRISS, O. 1956. Sex instability in Ricinus. Genetics 41: 265-280. 1961. Gibberellin as sex regulator in Ricinus communis. Science 133:2061—2062. TEJADA, C. A. D. P. AND S. H. BULLOCK. 1988. Seed fe- cundity of monoecious and female plants of a Croton species (Euphorbiaceae). Southwestern Naturalist 33: 233-234. THOMSON, J. D., K. R. SHIVANNA, J. KENRICK, AND R. B. Knox. 1989. Sex expression, breeding system, and pollen biology of Ricinocarpos pinifolius: A case study of Androdioecy? American Journal of Botany 76:1048—-1059. VASILIAUSKAS, S. A. AND L. W. AARSSEN. 1992. Sex ratio and neighbor effects in monospecific stands of Juni- perus virginiana. Ecology 73:622—632. WALLACE, C. S. AND P. W. RUNDEL. 1979. Sexual Dimor- phism and resource allocation in male and female shrubs of Simmondsia chinensis. Oecologia 44:34— 39. WASER, N. M. 1984. Sex ratio variation in populations of a dioecious desert perennial, Simmondsia chinensis. Oikos 42:343-348. WEBSTER, G. L. 1993. Euphorbiaceae. Pp. 567-577 in J. C. Hickman (ed.), The Jepson Manual. Higher plants of California. University of California Press, Berke- ley, CA. WHEELWRIGHT, N. T. AND A. BRUNEAU. 1992. Population sex ratios and spatial distribution of Ocotea tenera (Lauraceae) trees in a tropical forest. Journal of Ecol- ogy 80:425—4372. WILLSON, M. E 1983. Plant reproductive ecology. John Wiley & Sons, New York, NY. ZEMBAL, R. AND K. J. KRAMER. 1984. The known limited distribution and unknown future of Santa Ana River wooly-star (Eriastrum). Crossosoma 10:1—8. MApDRONO, Vol. 45, No. 3, pp. 250—254, 1998 LIMACINIASETA GEN. NOV. A CALIFORNIA SOOTY MOLD Don R. REYNOLDS Research and Collections, Natural History Museum of Los Angeles County, Los Angeles, CA 90007 dreynold @ almaak.usc.edu ABSTRACT Limaciniaseta californica is described as a sexually monomorphic species, developing on living plant surfaces. The distinguishing characters are dark cell-wall pigmentation, a stalked and setose ascocarp, a centrum with periphysoids, and an extenditunicate ascus that produces pigmented, traversely-septate as- cospores. Foliicolous ascomycete species adapted to the utilization of insect mediated plant exudates have been noted on the surface of living California plants by Millspaugh and Nuttal (1923), Miller and Bonar (1941), Farr et al. (1989) and Reynolds (1989c). These fungi have a dark, melanoid pigment in the cell walls of ascocarps and hyphae, and often form a mycelial mat of varying thickness on the living surfaces of leaves and nearby stems. A new species has been observed from among the foliicolous fun- gal assemblages associated with Baccharis pilularis D. C. in California. MATERIALS AND METHODS Observations were made from materials collect- ed in California since 1983. Representative collec- tions have been preserved as curated specimens in Herbarium LAM. Light microscopic observations were made with a Zeiss light microscope and a Ni- kon compound microscope utilizing squash mounts and hand-cut sections of whole ascocarps. Scanning electron micrographs were made with a Cambridge Electron Microscope at the University of Southern California Center for Electron Microscopy. TAXONOMY Limaciniaseta D. R. Reynolds, gen. nov. Mycelium in superficiebus plantarum vivarum. Hyphae ramosae, septatae, partietibus profunde pigmentiferis. Ascocarpus profunde pigmentifer, Stipitatus, globosus et ostiolatus. Setis in supero hemisphaero. Hyalnis periphysoidibus apice loculi ascocarpi. Ascus extenditunicatus. Ascosporae brunneae, transseptatae. The mycelium is superficial on living plant sur- faces. The hyphae are branched, septate and darkly pigmented. The ascocarp is pigmented, stipitate, globose and ostiolate. The setae are formed in the upper hemisphere of the ascocarp. Hyaline periphy- soids occur in the ascus locule. The ascus is an extendetunicate type, with brown transseptate as- cospores. Etymology. The name of the genus refers to a setose ascocarp similar to that of Limacinia Neger (Reynolds 1985). The mycelium is formed on living plant surfaces. The hyphae are branched, septate, with walls that are darkly pigmented. The ascocarp is darkly pig- mented, short-stalked, globose, and ostiolate; sev- eral setae form on its upper hemisphere. Multicel- lular, hyaline periphysoids are present in the asco- carp centrum originating from the inner surface at the apex of the ascal locule. The asci are arranged in a basal hymenium. The ascus wall is extenditun- icate. The eight, passively released ascospores are darkly pigmented and transversely septate at ma- turity. Limaciniaseta californica D. R. Reynolds, sp. nov. (Figs. 1-2) Ascocarpus usque ad 120 wm diametero, stipi- tatus, stipe ad 40 pm alto, subiculo, subtento hy- pharum profunde pigmentarum, plerumque oriun- darum a aliquot distinctis speciebus. Setae acumi- natae, Ostiolum setis profunde pigmentiferis, sep- tatis, 162 ad 180 pm _iongis, hyalinis. Periphysoidibus multiseptatis, apice loculi ascocar- pi, 2 X 25 wm. Hymenium maturione centrifugo intraliquidam hygroscopicam matricem. Maturus ascus ad 45 wm longus. Ascosporae octavae, fusi- forme, triseptatae, profunde brunneae, 13—22 x 6— 9 wm. Etymology. The species name indicates an oc- | currence in California. The ascocarp has a diameter | up to 120 pm. Its stipe measures up to 40 wm in height, formed from darkly pigmented subtending hyphae. The darkly pigmented setae are acuminate, 12-85 X 5—6 wm and surround an ostiole. The api- cal periphysoids are multiseptate and measure 2 X 25 wm. The mature hymenium forms centrifically | in a hygroscopic matrix. The mature ascus mea- sures up to 45 ym in length and the 8, fusiform ascospores are three-septate and brown, measuring 13-22 X 6-9 wm. Limaciniaseta californica is sexually monomor- 1998] REYNOLDS: LIMACINIASETA, GEN. NOV. Fic. 1. cation = 300. phic because it is known to exhibit only a sexual reproductive stage. The term “‘teleomorph”’ is un- derstood to apply only to the sexual stage in a pleo- morphic species, in the sense of Reynolds (1993), and should be used in opposition to an asexual or ‘“‘anamorph”’ reproduction stage occurring in the same life cycle. There is no evidence of pleomor- phy in L. californica. Thus, the species is termed monomorphic. Holotype. USA, California, Santa Barbara Coun- ty, Highway 101, south of EI Capitan State Park, on living surfaces of Baccharis pilularis DC, 17 April 1996, Don R. Reynolds and D. Minor, DRR136496 (holotype), on living surfaces of Bac- charis pilularis. Additional specimens examined. USA, Califor- nia, Contra Costa County, north end of Wildcat Canyon, 21 April 1931, L. Bonar, UC 966391, det. Morfea hendrickxii (Hansford) Batista & Ciferri. USA, California, Santa Barbara County, Highway 101, south of Santa Maria, on living surfaces of Baccharis pilularis, 17 April 1996, Don R. Reyn- olds and D. Minor. USA, California, Santa Barbara County, 17 April 1996, Don R. Reynolds and D. Minor, DRR136564. USA, California, Contra Costa State Park, Tilden State Park, on living surfaces of _ B. pilularis, 31 July 1983, Don R. Reynolds, DRR136024, USA, Oregon, Cannon Beach, on liv- Lamaciniaseta californica Apical view of setose ascocarps formed on a mycelial subiculum. SEM. Magnifi- ing surfaces of Umbellifera californica, 31 October 1935, J. R Thom (=Oregon State University 9280), BPI, det. Capnodium tuba Cooke & Harkness, (=IMUR 5247, det. Morfea tuba (Cooke & Hark- ness) Batista & Ciferri). Australia, Queensland, Tambourine Mountain, May 1934, A. Burge, BPI (det. Capnodium fuliginoides Rehm), type of Mor- fea helianthemi (Maire) Batista & Ciferri var. ma- jor. USA, Florida, Indian Town, 11 December 1919, C. V. Piper, IMUR 5520, det. Capnodium tuba (Cooke & Harkness) Ciferri & Batista. Costa Rica, San Jose, La Palma, P. C. Stanley 38094 (Plants of Costa Rica) BPI, type of Morfea miconia Batista & Ciferri—This is Trichomerium grandis- porum (Reynolds 1979) with hyaline ascospores rather than ‘“‘fuscoideae’”’ as characterized in the type description (Batista and Ciferri 1963). DRR138563, on living surfaces of Baccharis pilu- laris. The origin of hyphae giving rise to the ascocarps of L. californica and their possible connection to mitosporic fruit bodies is obscured in the collec- tions examined because of their formation in as- sociation with a plant surface mycelial layer of variable thickness to 3 mm. The layer is comprised of darkly pigmented hyphae similar to L. califor- nica and derives from several fungal species, in- cluding the mitospore and ascospore stages of the N N N Fic. 2. Lamaciniaseta californica Composite drawing representing longitudinal, median view of ascocarp. A short stalk subtends the fruitbody; two setae extend from their orgin near the ostiole; sterile hairs line the apical ostiole; periphysoids line the upper portion of the ascal chamber; the hymenium is depicted with two young asci in a Stage prior to ascospore and wall development and as a mature ascus with a cluster of 8 ascospores surrounded by a thinly stretched wall. Scale = approx. 10 pm. pleomorphic sooty mold, Capnodium salicinum. The hyphal strands immediately subtending the as- cocarp are comprised of darkly pigmented, rectan- gular shaped cells. The ascocarp forms on a colu- minate stalk that is comprised of firmly adhering, hyphal-like strands. The length of the stalk is vari- able; its width is 60—70% of the width of the as- cocarp, although, sometimes it is minimally pres- ent, especially where an underlying mycelial layer is substantially thickened. The acuminate setae form in the upper hemisphere surrounding the os- tiole (Fig. 1), which is lined with darkly pigmented, septate, pendulate but sometimes upwardly curving hairs. The multicellular, subulate, hyaline periphy- soids form a layer, originating from the cells of the interior, upper wall of the hymenial locule and ex- tending into the space between the ascocarp wall and the ascal layer (Fig. 2). The basal hymenium undergoes centrifugal maturation within a clear hy- groscopic matrix that fills the ascocarp cavity. The young, clavate ascus measures 16-18 wm before the ascospore initials are internally delimited by a refractive wall. The ascus apical wall is thickened and has a rudimentary nasse apicale at an early stage of ascospore formation similar to that of the fissitunicate type (Reynolds 1989a). The ascus in- creases approximately twice in volume during as- cospore formation, corresponding with a dimin- MADRONO [Vol. 45 ished apical wall thickness. The cylindrical ascus containing mature ascospores measures 36 wm in length. This fully developed ascus has the appear- ance of an extenditunicate ascus type (Reynolds 1989b) with the wall thinly stretched around the ascospores. The delineated ascospore wall is at first hyaline. After the formation of the first, centrally positioned cross septum, a brown pigmentation be- gins to intensify in the wall. A second cross septum subsequently divides each of the first two ascospore cells and the exterior spore surface ephemerially takes on an faintly echinulate appearance. The eight, fully formed ascospores are fusiform, three- septate, and golden brown in color in transmitted light. A passive release of the ascus and ascospores as a unit is suggested by their regular occurance on the surface of the ascocarp, particularly near the ostiole and in its immediate vicinity. The ascocarp of Limaciniaseta is similar to that of Scorias Fries (Reynolds 1979) but with less stalk. The setae are similar to those characteristic of species of Trichomerium Spegazzini (Reynolds 1982). The ascospores are comparable in shape and septation to those of Limacinia Neger (Reynolds 1985) and Trichomerium, and in pigmentation to those of Capnodium Montagne (Reynolds 1978) and Limacinia. The periphysoids are like those found in the centrum (Luttrell 1965) of the type species of Capnodium, Scorias, and Trichomerium. Two specimens cited under Morfea in Batista and Ciferri (1963) were found to be representative of Limaciniaseta californica. Batista and Ciferri (1963) utilized Limacinia Neger subgenus Morfea Arnaud as the basis of their genus Morfea. They | designated M. spongiosa Arnaud as the generic type species because it was the first taxon listed by Arnaud (1911). No specimens were cited as having been examined in support of this decision. The au- thors characterized their listed fungi with a ‘‘2-tu- nicate’’ ascus and ‘“‘brown, transversely plurisep- tate’? ascospores. The ascocarp was characterized | as globose to cylindric, pseudo-ostiolate, setose, | sessile, and formed from superficial, darkly pig- mented hyphae. Most of the nine species included | in this curious generic revision have been reas- | signed to other taxa (Barr 1955; Hughes 1976). One of the two L. californica specimens determined as a Morfea species (UC 966391) was listed as Morfea hendrickxii (Hansford) Batista & Ciferri. Specimen OSU 9280 was cited by Batista and © Ciferri (1963) as M. tuba Batista & Ciferri, a spe- | cies transferred to Capnodium without these au- thors having seen type material. A Thom collection | from Washington State (rather than California as stated) was indicated as the basis of an emendation | of M. tuba. The authors stated, “it is possible that | this material [the Thom specimen used as the basis _ | of the redescription and nov. comb. in Morfea] ... would be referable to Capnodium tuba, agreeing — with the description of Cooke and Harkness for per- — ithecia, since the ascospore were not described [by © 1998] them].”’ “In our taxonomy, this specimen ... does not belong to the genus Capnodium.”’ The Thom specimen (BPI) was found to be somewhat accurately depicted by Batista and Ci- ferri (1963) with the ascus and ascospores differing in size from that given in their emendation of the description. Using the Thom material, Batista and Ciferri (1963) interpreted the ascocarp as “‘harbor- ing at the top by short continuous setae ....”’ Fig- ure 53 (Batista and Ciferri 1963) shows three setae positioned near the ostiole of an upright, elongate fruit body. This perception of the ascocarp is made apparently in deference to the Cooke and Harkness (1884) reference to the position as “‘perithecium”’ hairs on C. tuba. The name Morfea tuba with the authorship (Cooke & Harkness) Batista and Ciferri (Batista and Ciferri 1963) is discounted for two reasons. First, the Thom specimen is inappropriate as an im- plied or undesignated lectotype. The Batista and Ci- ferri redescription is regarded as in serious conflict with the Cooke and Harkness protologue (Greuter et al. 1994: ICBN Article 9.13). Cooke and Hark- ness (1884) reported no ascus from the California material that was the basis of the name Capnodium tuba (‘‘ascis nondum visis’’). No specimen has been found that would properly serve as the type specimen for C. tuba. Further, the description of C. tuba by Cooke and Harkness (1884), repeated by Ellis and Everhart (1892), is likely that of a Lep- toxyphium species. A mitosporic fungus is clearly described by Cooke and Harkness (1884) as having an erect, branched, infundibuliform column that is “sursum ciliate.’”’ The mitosporic Leptoxyphium fruit body is accurately depicted by Hughes (1976) as “‘simple and variously proliferated.’ This same depiction also accounts for the position of cilia mentioned by Cooke and Harkness that occurred on a “‘perithecium.”’ ““Toward the apex [of the Leptox- yphium fruit body], component hyphae” form a funnel-shaped head comprised of hyphae that ‘‘ter- minate in a long ... hyaline cell and together they form a fringe of sterile hairs.’ The Cook and Hark- ness (1884) citation antedates the 1897 type de- scription for Leptoxyphium by Spegazzini (1918) and Hughes (1976). The second reason for ignoring the nomenclature of Batista and Ciferri (1963) is that their emenda- tion results in the name Morfea tuba Batista & Ci- ferri. This authorship constitutes a nomen nudum for lack of a Latin description or diagnosis (Greuter et al. 1994). CONCLUSIONS Several groups of sooty mold fungi species have been proposed as families (Hughes 1976). The tax- on concepts were based on the perception of a com- mon hyphal morphology type that provided a pu- tative morphological link of unique sexual and _ asexual reproductive structures with pleomorphic REYNOLDS: LIMACINIASETA, GEN. NOV. 259 implications at life cycle and phylogenetic levels. These delineated groups comprised of pleomorphic and mitosproric and ascosporic monomorphic spe- cies can be regarded as hypothetical monophyletic clades. Limaciniaseta californica has morphological characters derived from the hyphae and the asco- carp that suggest a sister relation to Capnodium, Trichomerium, and Scorias in one of the Hughes groups, the Capnodiaceae. An exception is the as- cus type; the capnodiaceous species have a fissitun- icate ascus. The extenditunicate ascus of L. califor- nica was first described for Meliolina sydowiana Stevens (Reynolds 1989b). Meliolina has been shown on the basis of 18s rDNA data to have a basal rather than a derived position on a filamen- tous ascomycete clade (Saenz 1997) and thus has no close relation to Meliola as predicted (Hughes 1993). ACKNOWLEDGMENTS Museum Associate David M. Minor assisted with the field work. Tod Stuessy contributed the Latin description. Keith Seifert provided valuable criticism. Financial sup- port was made available through the Los Angeles County Museum of Natural History Foundation. LITERATURE CITED ARNAUD, G. 1911. Contribution a l’étude des fumagines 2me Parite: Systematique et organisation des espeéces. Ann. Ecole. Natl. Agric. Montepellier, séries 2, 10: 211-330. BarR, M. E. 1955. Species of sooty molds from western North America. Canadian Journal of Botany. 33:497— 514. BATISTA, A. C. AND R. CIFERRI. 1963. Capnodiales. Sac- cardoa 2:1—195. ; CooKE, M. C. AND H. W. HARKNESS. 1884. New California fungi. Grevillea 12:83, 92. ELLs, J. B. AND B. M. EVERHART. 1892. The North Amer- ican Pyrenomycetes. Privately published. New Field, NJ. FARR, D. E, G. F BILLs, G. P. CHAMURIS, AND A. Y. ROSs- MAN. 1989. Fungi on plants and plant products in the United States. American Plant Pathology Society Press. St. Paul/Minneapolis, MN. GREUTER, W., EF R. BARRIE, H. M. BURDET, W. G. CHAL- ONER, V. DEMOULIN, D. L. HAWKSworTH, P. M. Jor- GENSEN, D. H. NICOLSON, P. C. SILVA, AND P. TREHANE. 1994. International code of botanical nomenclature (Tokyo code). Koeltz Scientific Books. K6nigstein, Germany. HUGHES, S. J. 1976. Sooty moulds. Mycologia 68:693— 820. . 1993. Meliolina and its excluded species. My- cological Papers 166:1—255. CAB International. Wal- lingford. LUTTRELL, E. S. 1965. Periphysoids, paraphyses and pseu- doparaphyses. Transactions British Mycological So- ciety 48:135—144. MILLER, V. AND L. BONAR. 1941. A study of the Perispo- riaceae, Capnodiaceae, and some other sooty molds from California. University of California Publications in Botany 19:405—417. 254 MADRONO [Vol. 45 MILLSPAUGH, C. E AND L. W. NutTTAL. 1923. Flora of Santa Catalina Island. Publication 212. Field Museum of Natural History, Botanical Series 5:1—413 + 5 plates and | figure. REYNOLDS, D. R. 1978. Foliicolous ascomycetes 2. Cap- nodium salicinum Montagne emend. Mycotaxon 7: 501-507. . 1979. The stalked capnodiaceous species. My- cotaxon 8:417—445. . 1982. Foliicolous ascomycetes 4: The capnodia- ceous genus Trichomerium Speg. emend. Mycotaxon 14:189—220. . 1985. Foliicolous ascomycetes 6: the capnodia- ceous genus Limacinia. Mycotaxon 23:153—-168. . 1989a. The bitunicate paradigm. Botanical Re- view 55:1-—52. . 1989b. An extenditunicate ascus in the ascostro- matic genus Meliolina. Cryptogamic Mycologie 10: 305-320. . 1989c. Foliicolous ascomycetes 8: Capnodium in California. Mycotaxon 34:197—216. . 1993. The fungal holomorph: an overview. in Reynolds and J. W. Taylor (eds.), The fungal holo- morph. Mitotic, meiotic and pleomorphic speciation in fungal systematics. D. R. CAB International. Wal- lingford, Great Britain. SAENZ, G. 1997. Phylogeny of the Erysiphales (powdery mildews), Meliola (black mildews), and Meliolina in- ferred from ribosomal DNA. Ph.D. dissertation. Univ. of California, Berkeley, CA. SPEGAZZINI, C. 1918. Notas micélogicas. (Buenos Aires) 4:281-295. Maprono, Vol. 45, No. 3, pp. 255—260, 1998 LIGNOTUBERS IN SEQUOIA SEMPERVIRENS: DEVELOPMENT AND ECOLOGICAL SIGNIFICANCE PETER DEL TREDICI The Arnold Arboretum of Harvard University, 125 Arborway, Jamaica Plain, MA 02130, USA ABSTRACT Seedlings of Sequoia sempervirens (D. Don) Endl. develop lignotubers as part of their normal ontogeny from detached meristems located in the axils of the two cotyledons. Within four months of germination, each axillary cotyledonary meristem gives rise to a large central bud with two or more collateral accessory buds. As seedlings age, bud and cortex proliferation phenomena associated with lignotuber formation spreads distally to include axillary meristems located immediately above the cotyledonary node. Ligno- tubers continue to expand throughout the life of a Sequoia, eventually forming massive, basal swellings that are covered with leafy shoots and/or suppressed shoot buds. The Sequoia lignotuber is a specialized organ of regeneration and carbohydrate storage that contributes to the long-term survival of the tree by producing buds that can develop into shoots following traumatic injury to the primary trunk and by generating new roots that increase the stability and the vigor of both young and old trees. In response to a variety of exogenous factors, Sequoia will also produce induced lignotubers, or burls, on the trunks of mature trees as well as on the “‘layered”’ lateral branches of young trees, where they come in contact with the soil. The California coast redwood, Sequoia semper- virens (D. Don) Endl. (Taxodiaceae (henceforth re- ferred to as Sequoia)), is famous not only for being one of the tallest trees in the world but also for its ability, unusual among conifers, to resprout vigor- ously after being cut down. While much has been written about the commercial and ecological im- portance of resprouting in redwood (Olson et al. 1990), little is known about the precise origin of these sprouts, which arise from a large “‘burl’’ lo- cated at the base of the tree (Becking 1968; Stone and Vasey 1968; Simmons 1973; Groff and Kaplan 1988). There are numerous reports in the literature of woody plants that can resprout from underground burls, technically known as lignotubers, following traumatic injury to the primary trunk. Anatomical studies on several arborescent taxa, including Eu- calyptus spp. (Carr et al. 1984), Arbutus unedo (Sealy 1949), Quercus suber (Molinas and Verda- guer 1993), and Ginkgo biloba (Del Tredici 1992, 1997), have shown that lignotubers are genetically _ determined structures that develop from buds lo- _ cated in the axils of both cotyledons and a few of _ the leaves immediately above them. Over time, lig- | hotubers can become quite large and contribute to the survival of plants in three ways: 1) they are a site for the production and storage of suppressed _ Shoot buds that can sprout following injury to the primary stem; 2) they are a site for the storage of _ carbohydrates and mineral nutrients, which may al- _ low for the rapid growth of these suppressed buds following stress or damage; and 3) for plants grow- _ ing on steep slopes, they can function as a kind of clasping organ that anchors the tree to the rocky substrate (Sealy 1949; James 1984; Del Tredici et al. 1992). Lignotuber-producing species are most common- ly found in Mediterranean-type ecosystems that are characterized by hot, dry summers and periodic fires (James 1984; Mesleard and Lepart 1989; Can- adell and Zedler 1994). The purpose of the present study is twofold: first, to determine whether or not the basal swellings produced by Sequoia fit the def- inition of an ontogenetic lignotuber (Carr et al. 1984; James 1984; Canadell and Zedler 1994), and second, to examine the ecological role that these Structures play in the tree’s native habitat in the coastal forests of northern California. MATERIALS AND METHODS Seeds of Sequoia sempervirens were extracted from green cones collected from the ground in Richardson Grove State Park, Humboldt County, California, on 28 October 1993. The cones had been dislodged from trees by squirrels feeding in the tree crowns. Cones were taken to the Arnold Arboretum in Boston, Massachusetts, where they were allowed to air dry until they shed their seeds. These were sown in a warm greenhouse (heated to a minimum temperature of 17°C) on 29 November 1993 and 13 April 1994. For both dates, germina- tion commensed about two weeks later. A minimum of ten undamaged seedlings were sampled at each of four time periods: 15, 37, 59, and 133 days after germination. At the time of sampling, two to three mm long segments of the primary axis, including tissue above and below the point of attachment of the cotyledons, were collected from the plants, fixed in FAA, dehydrated in a t-butyl alcohol series, 256 and embedded in paraplast. Serial transections, 10 microns thick, were cut on a rotary microtome and stained with Heidenhain’s hematoxylin and safranin (Johansen 1940). For the purpose of studying the later stages of lignotuber development, a dozen three-year-old redwood seedlings were purchased from a California nursery and cultivated in con- tainers in the greenhouse for three years. Observations on mature Seguoias, both logged and unlogged, were made during October 1993, at four sites in northern California: Redwood National Park and Humboldt Redwoods State Park in Hum- boldt County, Big Basin Redwoods State Park in Santa Cruz County, and Samuel P. Taylor State Park in Marin County. Seedlings of Sequoia, mostly one to four years old, were collected on 26 October 1993, along with a few small layered branches from older Sequoia saplings. Both the seedlings and the layers were brought back to the Arnold Arboretum for further study and documentation. At the same time, numerous “‘live burls’’ were purchased from tourist shops located near Redwood National Park. These were placed in shallow dishes of water in a warm greenhouse heated to a minimum temperature of 10°C. RESULTS Lignotuber development on seedlings: 15 to 180 days. Observations on greenhouse-grown seedlings indicate that lignotubers in S. sempervirens origi- nate from exogenous meristems located in the axils of the two cotyledons. On 15-day-old seedlings, the axillary cotyledonary meristems are poorly differ- entiated, being little more than a single, superficial layer of meristematic cells approximately 0.3 mm across, with no vascular connection to the stele (Fig. 1). By 37 days, structures identifiable as mer- istems develop in the axils of the cotyledons, but their vascular connection to the stele is only par- tially complete. By 59 days, the axillary cotyledon- ary meristems produce foliar primordia and estab- lish a complete vascular connection to the stele (Fig. 2). By 133 days, the axillary cotyledonary meristems develop into distinct buds that protrude from the stem by as much as 0.5 mm, and collateral accessory buds develop adjacent to the primary cot- yledonary bud (Fig. 3). By the time seedlings are six months old, clusters of buds are readily visible at both cotyledonary nodes, with some of them pro- ducing vegetative shoots. Lignotuber development on older seedlings: one to five years. On 3 to 5-year-old collected seedlings, the cotyledonary node is readily identifiable by the oppositely arranged pair of protruding bud-clusters at the base of the stem. Depending on the vigor of the seedling and the amount of damage it has sus- tained, one, both, or neither of the cotyledonary bud-clusters were producing leafy shoots, 0.5 to 4.0 cm long (Fig. 4). While sprouting is common in seedlings that have experienced damage to the pri- MADRONO [Vol. 45 mary stem, it also occurs in seedlings that showed no signs of injury. In addition to shoot production, the cotyledonary node region of the collected seed- lings often produce adventitious roots in response to partial burial. Typically, wild-grown Sequoia seedlings do not develop visible bud swellings at the cotyledonary node until they are between three and six-years-old (Becking 1968; Simmons 1973). In contrast, lig- notubers of greenhouse-grown seedlings produce abundant bud clutsters and/or sprouts by the time they reach one and a half years old. After four or five years of cultivation, the cortical swelling and bud proliferation associated with lignotuber for- mation spreads distally to engulf many of the nodes produced during the first growing season (Fig. 5). Lignotuber development on mature trees. Lig- notubers expand throughout the life of a Sequoia, eventually forming a massive, woody swelling at or just below ground level. The outer surface of this swelling is generally covered with shoot buds. On undamaged trees, lignotubers typically give rise to clusters of small leafy shoots encircling the base of the trunk. On trees damaged by logging or ero- sion, lignotubers give rise to large secondary trunks that equal or exceed the primary trunk. Mature trees that had been logged 90 to 100 years ago have now developed lignotuber sprouts well over a meter in diameter. When second-generation trees are found growing on a steep slope near a stream or a road cut, the woody lignotuber is readily observable as a massive “‘plate’’ of downward-growing tissue that follows the contours of the ground and extends two to three m out from the nearest trunk. As well as giving rise to new shoots, such exposed lignotubers are also the source of roots that help to anchor trees | to eroding slopes (Fig. 6). On rocky sites, the lig- notuber has a tendency to form a kind of clasping organ that envelops the adjacent substrate, further stabilizing the tree. Induced lignotuber development on layered branches. Induced lignotubers were observed to de- velop on the layered branches of 29-year-old Se- quoia saplings, the growth of which was limited by low light levels that prevail beneath a mature red- wood forest canopy. The stems of these plants are typically weak and spindly, and, when pinned down by a fallen limb or tree, they take root and turn © upwards to reestablish a vertical orientation. Typi- | cally a single, downward-growing lignotuber de- velops along the side of the stem in contact with the soil, although in a few cases more than one had | formed along a single stem. On such layered branches, the original connection to its parent trunk | typically withers away, leaving only the bowed shape of the stem and the off-center lignotuber as | evidence of it origin from a branch (Fig. 7). As is the case with lignotubers derived from ax- illary buds at the cotyledonary node, those formed by layered branches possess the ability to generate 1998] DEL TREDICI: SEQUOIA LIGNOTUBERS 257 aa m ® a. i ul Fics. 1-5. 1-3. Transverse sections of the cotyledonary node region of 15- to 133-day-old seedlings of Sequoia sempervirens. (1) A 15-day-old seedling showing the relationship of the superficial meristems (arrows) to the cotyledons (c). Bar = 0.1 mm. (2) A 59-day-old seedling showing fully developed cotyledonary meristems (arrows), foliar pri- mordia, and the vascular connection to the stele. Bar = 0.1 mm. (3) A 133-day-old seedling showing fully developed cotyledonary bud (bottom) and two accessory collateral buds (top). Bar = 0.3 mm. Fic. 4. A three to four-year-old Sequoia seedling, collected from the wild, showing sprouting cotyledonay buds, accessory collateral buds, and the remnants of a cotyledon (c). Bar = 1.0 mm. Fic. 5. A five-year-old greenhouse grown seedling showing the proliferation of suppressed buds (arrow) at and above the cotyledonay node. Bar = 1.0 cm. MADRONO ms Fics. 6-9. 6. A large Sequoia growing along a stream bank in the Humboldt Redwoods State Park showing extensive root and trunk development from its exposed, downward-growing lignotuber. Photo by R. Becking. Fic. 7. A layered lateral branch of Sequoia. Note that the downward-growing induced lignotuber has produced both roots and a vegetative shoot. Bar = 1.0 cm. Fic. 8. An ancient Sequoia in Big Basin Redwoods State Park showing massive burl development. Fic. 9. A forest of Sequoias in Korbel, California, resprouting from their lignotubers three years after clear-cutting. both shoot buds and roots. How long it takes for a develop on the lower portions of the trunk of ma- branch to develop a visible lignotuber after it has ture redwood trees in response to traumatic injury been pinned to the ground is unknown, but is prob- from fire, wind, or floods. Typically lignotubers are ably at least two years. initiated above the point of injury, eventually grow- Induced lignotuber development on the trunk of | ing down over the wound to cover it. On very old mature trees. Large, lignotuber-like structures often trees, extensive growths of contorted callus tissue 1998] can project out from the trunk 50 cm or more (Fig. 8). If these burls come in contact with the ground, which they often do, they will develop both roots and shoots. When cut off and placed upside down in a dish of water in a warm greenhouse, they will produce numerous leafy shoots within two weeks, and roots after six months to one year. Observations suggest that burls on the trunks of old redwood trees originate as wound-induced cal- lus tissue which, as it proliferates, incorporates nearby shoot buds into its structure. There appear to be two distinct types of lignotubers on Sequoia, the contorted type, located mainly on the lower por- tions of the trunk, which is irregular in shape, downward-growing in orientation, and covered with sprouts and/or shoot buds. The second type, which occurs higher on the trunk, is nearly hemi- spherical in shape, lacks the downward orientation, and produces comparatively few sprouts or shoot buds. Trunk-burls are probably best interpreted as a case of uncontrolled bud and cortex proliferation induced by old age, traumatic injury, or environ- mental stress. They serve as sites for the production of new shoots and adventitious roots on trees that have been partially buried with silt from flooding or on leaning trees whose trunks have come into contact with the soil. DISCUSSION According to Strauss and Ledig (1985): “‘Archi- tectural patterns established during the first few months of life are indicative of development de- cades to centuries later, when the plant has in- creased a millionfold in size.’’ The lignotuber of S. sempervirens, which can be fully functional in four- month-old seedlings and remain functional on trees that are at least 1100 years old, clearly illustrates the truth of this observation. Regardless of the age or size of the tree, the redwood lignotubers often resprout within two to three weeks of logging. While most of these sprouts do not survive to ma- turity, enough of them do to effectively regenerate a new forest (Olson et al. 1990). One study done with an old-growth forest that had been clear-cut showed that the ability of redwoods to resprout (i.e., the number of sprouts per meter circumfer- ence) is greatest in trees that were between 200 and 400 years of age, and decreases rapidly thereafter. Trees more than one thousand years old are able to resprout at only 20 to 25% of the peak rate (Powers and Wiant 1970). The authors also reported that 92% of all surviving sprouts grow out from the lignotuber, 6% from the bole proper, and 2% from the cut surface of the stump. When a tree was growing on a slope greater than 20%, the sprouts are more numerous on the downhill side of the trunk. The remarkable ability of redwood trees to re- Sprout from its basal lignotuber, regardless of age, DEL TREDICI: SEQUOIA LIGNOTUBERS 259 is clearly the basis of the redwood’s persistence in the face of extensive clear-cutting (Fig. 9). Throughout its natural range, logging has served to transform S. sempervirens into a clonally reproduc- ing species that spreads by means of its under- ground lignotuber. Jepson (1910) described one col- ony of 45 large redwoods that formed a third-gen- eration “fairy ring,’ 17 m by 15 m across. As sig- nificant as lignotuber sprouting is for mature trees, however, the process is probably of greater impor- tance to seedlings and saplings that are struggling to survive in dense shade or on exposed slopes (Becking 1968; Canadell and Zedler 1994). Despite the abundant documentation on the im- portance of Sequoia lignotuber sprouting to forest- ry, there is very little information available on its significance in the absence of logging-related dis- turbance. One study on an uncut Sequoia forest in Humboldt County found that basal sprouting in red- wood was closely associated with the occurrence of fire (Stuart 1987). By correlating fire scars on the primary trunk of the tree with basal sprouts from its lignotuber, the author determined that during the ‘“‘pre-settlkement period”’ (between 1775 and 1875) fires occurred regularly in the forest, at an average interval of 24.6 + 2.8 years. Other studies, which analyzed fire scars on the cut stumps of old-growth redwoods, support the idea that fires were common in the redwood region prior to European settlement (Fritz 1931; Jacobs et al. 1985; Finney and Martin 1992). These findings are consistent with studies in other Mediterranean climates which report the oc- currence of lignotuber-producing angiosperms in habitats where fire, or other types of recurring dis- turbance, is common (James 1984; Mesleard and Lepart 1989). The trunk of a redwood tree, above the lignotu- ber, also shows a strong ability to resprout follow- ing traumatic injury. In the older forestry literature, there are numerous reports of large trees, entirely defoliated by fire, that sprout vigorously to form lush ‘“‘fire columns”? (Jepson 1910; Fritz 1931). Similarly, the author has observed recently blown down trees that sprouted new growth along the en- tire length of the horizontal trunk. Fink (1984) stud- ied the ability of Sequoia stems to resprout follow- ing injury and found that clusters of replacement buds developed exogenously in the needle axils of young branches over a one to two year period. Ex- cept for the length of time involved, the process he described for the development of preventitious shoot buds is very similar to that of the cotyledon- ary buds described in this paper. One must keep in mind, however, that the lignotuber formed at the cotyledonary node is under strict genetic control while those that develop elsewhere on the trunk are under environmental control. In this regard, Se- quoia is similar to Ginkgo biloba which produces positively geotropic lignotubers from axillary cot- yledonary buds (basal chichi), as well as induced lignotubers (aerial chichi) on its trunk and branches 260 (Del Tredici 1992, 1997). As is the case with Se- quoia, shoot and root regeneration by the Ginkgo biloba lignotuber play an important role in the per- sistence of the species in its native habitat in the temperate forests of eastern China (Del Tredici et al. 1992). From both the morphological and physiological perspectives, lignotuber-generated shoots can be considered ‘‘juvenile’’ relative to the rest of the tree (Greenwood 1995). This conclusion is supported by in vitro studies which found that tissue cultures started with lignotuber shoots from the base of a 90-year-old Sequoia were more vigorous and rooted more readily than those started with shoots from the crown of the same tree (Bon et al. 1994). The authors also identified numerous membrane-asso- ciated proteins that were synthesized in greater abundance in cultures derived from lignotuber shoots that those derived from the upper portions of the tree. Certainly it is not by chance that Se- quoia was the first conifer to be successfully cul- tured using in vitro techniques, and that the cultures were derived from lignotuber sprouts (Ball 1950). ACKNOWLEDGMENTS I would like to express my thanks to: Dr. Rudolf Beck- ing of Arcata, California, who freely shared his extensive knowledge of the redwood forest with the author; Sheila Morris of Vancouver, British Columbia, who provided technical support with the staining and sectioning of the seedling material; and Dr. Paul Groff, whose work first stimulated my ideas on the subject of Sequoia lignotubers. I would also like to thank the Highsted Foundation of Redding, Connecticut, for financial support; the green- house staff of the Arnold Arboretum, for growing the plants used in this study; and the Organismal and Evolu- tionary Biology Department of Harvard University for al- lowing me to use their microscopes. LITERATURE CITED BALL, E. 1950. Differentiation in a callus culture of Se- quoia sempervirens. Growth 14:295—325. BECKING, R. 1968. The ecology of the coastal redwood forest and impact of the 1964 floods upon redwood vegetation. Final report, National Science Foundation Grant NSF GB #4690. Bon, M.-C., EF RICCARDI, AND O. MONTEUUIS. 1994. Influ- ence of phase change within a 90-year-old Sequoia sempervirens on its in vitro organogenic capacity and protein patterns. Trees 8:283—287. CANADELL, J. AND P. H. ZEDLER. 1994. Underground struc- tures of woody plants in Mediterranean ecosystems of Australia, California, and Chile. Pp. 177-210 in M. T. Kalin Arroya, P. H. Zedler, and M. D. Fox (eds.), Ecology and biogeography of Mediterranean ecosystems in Chile, California, and Australia. Springer-Verlag, New York, NY. Carr, D. J., R. JAHNKE, AND S. G. M. Carr. 1984. Initi- ation, development, and anatomy of lignotubers in MADRONO [Vol. 45 some species of Eucalyptus. Australian Journal of Botany 32:415—437. DeEL TREDIcI, P. 1992. Natural regeneration of Ginkgo bi- loba from downward growing cotyledonary buds (ba- sal chichi). American Journal of Botany 79:522—530. . 1997. Lignotuber development in Ginkgo biloba. Pp. 119-126 in T. Hori et al. (eds.), The world of Ginkgo. Springer-Verlag, Tokyo. , H. LING, AND G. YANG. 1992. The Ginkgos of Tian Mu Shan. Conservation Biology 6:202—209. FINK, S. 1984. Some cases of delayed or induced devel- opment of axillary buds from persisting detached meristems in conifers. American Journal of Botany 71:44-51. FINNEY, M. A. AND R. E. MARTIN. 1992. Short fire inter- vals recorded by redwoods at Annadel State Park, California. Madrono 39:25 1-262. Fritz, E. 1931. The role of fire in the redwood region. Journal of Forestry 29:939—950. GREENWOOD, M. S. 1995. Juvenility and maturation in co- nifers: current concepts. Tree Physiology 15:433- 438. GRoFF, P. AND KAPLAN, D. R. 1988. The relation of root systems to shoot systems in vascular plants. Botanical Reviews 54:387—422. Jacoss, D. E, D. W. COLE, AND J. R. MCBRIDE. 1985. Fire history and perpetuation of natural coast redwood ecosystems. Journal of Forestry 83:494—497. JAMES, S. 1984. Lignotubers and burls—their structure, function and ecological significance in Mediterranean ecosystems. Botanical Reviews 50:225-—266. JEPSON, W. L. 1910. The silva of California. The Univer- sity Press, Berkeley. CA. JOHANSEN, D. A. 1940. Plant microtechnique. McGraw- Hill, New York, NY. MESLEARD, F AND J. LEPART. 1989. Continuous basal sprouting from a lignotuber: Arbutus unedo L. and Erica arborea L., as woody Mediterranean examples. Oecologia 80:127—131. Mo.inas, M. L. AND D. VERDAGUER. 1993. Lignotuber ontogeny in the cork-oak (Quercus suber; Fagaceae) II. Germination and young seedling. American Jour- nal of Botany 80:182-191. OLSON, D. F, D. E Roy, AND G. A. WALTERS. 1990. Se- quoia sempervirens (D. Don) Endl. Pp. 541-551 in R. M. Burns and B. H. Honkala (eds.), Silvics of North America, Vol. 1, conifers. U. S. D. A. Forest Service Handbook 654. Powers, R. FE AND H. V. WIANT. 1970. Sprouting of old- growth coastal redwood stumps on slopes. Forest Sci- ence 16:339-341. SEALY, J. R. 1949. The swollen stem-base in Arbutus une- do. Kew Bulletin 4:241-251. Simmons, N. 1973. Description of bud collars on redwood seedlings. Masters Thesis, Humboldt State Universi- ty, Arcata, CA. STONE, E. C. AND R. B. VASEY. 1968. Preservation of coast redwood on alluvial flats. Science 159:157—161. Strauss, S. H. AND E T. Lepic. 1985. Seedling architec- ture and life history evolution in pines. American Naturalist 125:702—715. STuaRT, J. D. 1987. Fire history of an old-growth forest of Sequoia sempervirens (Taxodiaceae) in Humboldt Redwoods State Park, California. Madrono 34:128— 141. MADRONO, Vol. 45, No. 3, pp. 261-270, 1998 CAREX SERPENTICOLA (CYPERACEAE), A NEW SPECIES FROM THE KLAMATH MOUNTAINS OF OREGON AND CALIFORNIA PETER F ZIKA, KELI KUYKENDALL, AND BARBARA WILSON Carex Working Group, Herbarium, Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR 97331 Carex serpenticola P. F Zika (Carex section Ac- rocystis) is described from ultramafic deposits in the Klamath Mountains of southwestern Oregon and northwestern California. It appears most similar to C. globosa Boott, from which it is distinguished by frequently unisexual culms and dark purple pis- tillate scales. C. serpenticola perigynia are shorter, and differ in the length of stipe and nerving on the faces. Bisexual culms closely resemble C. concin- noides Mackenzie, from which it is separable by its three subplumose styles. Lilla Leach was an early botanical explorer in the remote regions of the Klamath Mountains. On 1 May 1931, while collecting on serpentine bedrock deposits in the mountains east of Game Lake, Curry Co., OR (Leach and Leach 1938; Love 1991, and personal communication), she gathered the earliest known specimens of an unknown Carex. For de- cades Leach’s serpentine sedge was overlooked, un- til the Carex Working Group began work on an atlas of Oregon Carex in 1993. Reviewing the col- lections of Leach and subsequent botanists, we were puzzled by plants labeled *‘C. globosa?”’ and disjunct from its known range in California (Fig. 1). Further field and herbarium investigations dis- closed that this sedge was a distinctive undescribed species. As most or all collections of this new Car- ex were from ultramafic bedrock zones, we chose a name to reflect its preference for serpentine soils. Carex serpenticola P. F Zika, sp. nov. (section Ac- rocystis). (Fig. 2)—-TYPE: USA, California, Del Norte Co., sunny moist slope, SW aspect, ser- pentine soil, with Libocedrus, Umbellularia, Rhamnus californica Eschsch., Quercus vaccin- ifolia Kellogg, Carex mendocinensis Olney, Mimulus guttatus DC., Danthonia californica Bolander, SE of Azalea Lane, N of Middle Fork Smith River, ca. 0.8 km SE of bridge in the town of Gasquet, 41°51'N, 123°57.5'W (TI17N R2E sect. 21 SW%4; Humboldt Meridian), elev. 120 m, 29 April 1997, P. F. Zika and K. Kuykendall 13062 (holotype, OSC; isotypes, HSC, MICH, NY, UC, US). Differt a Carex globosa Boott culmis staminatis aut pistillatis interdum moneociis; perigyniis 3.1— 3.6 mm longis, enervibus, stipitibus 0.4-0.8 mm longis; squamis pistillatis atropurpureis sine nervis secundariis ad nervum medium parallelis. Perennial, rhizomatous between small tufts of shoots; capable of forming mats 2 m in diameter; rhizomes 2.5—7.5 (—10) cm long, 1.2—2.2 (—3.4) mm thick; young rhizomes clad in purple-black scales 5—7 mm long, the largest with fibrillose sheaths. Base of erect shoot scaly with bladeless sheaths, dark red-purple; ventral face of bladeless sheaths deteriorating with age and leaving a pin- nate-reticulate network of persistent veins. Leaves mostly basal, margins scabrous; leaf blades variable depending upon exposure, long, straight and nearly flat when in favorable sites, folded, short, and fal- cate in harsh sites; all leaves somewhat V-shaped in cross section, widest leaves (1.5—) 2.2—3.5 (—5.0) mm wide when spread flat, longest basal leaves 8—28 (—35) cm long, semi-evergreen, never glaucous; sheath mouth and /igu/e minutely sca- brous (at 40x). Ligule wider than tall, obtuse, membranous, discolored, 0.2 mm thick. Only shoots of the previous season producing flowering culms. Cu/ms 8—38 cm tall, bearing 2—5 highly re- duced green or purple-margined leaves, culms erect in flower, pistillate and bisexual culms arching to the ground in fruit (Fig. 2F). Culms usually either pistillate (Fig. 2D, E, F) or staminate, thus the spe- cies superficially appearing dioecious. Bisexual culms rarely with terminal spike gynecandrous (Fig. 2C). Most bisexual culms with a solitary ter- minal staminate spike and sessile 4—9-flowered pis- tillate spikes (Fig. 2B). Staminate culms with 1-3 lateral aborted spikes marked by short inflorescence bracts (Fig. 2A). Some rhizomes bearing shoots with both bisexual and unisexual culms. Bracteal sheaths essentially absent, O—2 mm long, dark pur- ple-margined when present. Inflorescence bract minute, purple, shorter than lateral spike, to green and leafy, equaling or exceeding inflorescence, 0.1— 9.2 cm long; bracts minutely scabrous-ciliate at tip; blades of larger inflorescence bracts dark purple- margined at base, with green midrib varying in width, sometimes with purple stripes flanking green midrib, sometimes with slender white-hyaline mar- gin (this, if present, 0.1—0.2 mm wide). Staminate inflorescence 13—24 mm long, (1.4—) 2.0—4.1 mm diameter; stamens 3 per scale, dried anthers 2.1— 4.0 mm long, staminate scales oblanceolate, 4.6— 5.6 mm long, 1.1—1.6 mm wide, purple-black, fad- ing to reddish, minutely erose-margined. Pistillate inflorescence with terminal spike up to 65 flowered and 8—47 mm long, 6-9 mm wide. Bisexual inflo- rescence 15-28 mm long, 4—9 mm wide; subter- 262 * = C. serpenticola @ = C. globosa Fic. 1. Distributions of Carex serpenticola and C. glo- bosa. Mexican populations of C. globosa are not mapped. minal pistillate spikes 5—8 mm long, sessile or rare- ly on erect peduncles 5—10 mm long, lateral spikes 4—9 flowered. Axillary basal or near-basal pistillate spikes rarely present (Zika and Kuykendall 13062 UC) on some pistillate or bisexual culms, on pe- duncles 1—11 cm long. Pistillate scales lanceolate, glabrous or scabrous-tipped, longer and narrower than perigynia, acuminate, 3.8—4.6 mm long, 1.3-— 1.8 mm wide, widely spreading as perigynia ma- ture, dark purple. Pistillate scale midrib green or yellow-brown, flat or slightly keeled, glabrous or Slightly scabrous; rarely 1-2 pistillate scales on plant with 1-2 faint lateral nerves (Rolle 327 ID). Pistillate scale margins with hyaline-border 0.1—0.4 mm wide. Stigmas ca. 3 mm long above orifice of perigynium, stigmas 3, subplumose. Perigynium obovate, gradually tapered to a cuneate base, 3.1— 3.6 X 1.5-1.8 mm, short hairy, green, some rip- ening to a dark purple, plump, trigonous, wingless, with marginal ribs, the faces nerveless (rarely 5 nerves up to mid-length), facial nerves hard to see on fresh material. Perigynia abruptly short beaked, beak 0.5—1.0 mm long, dark purple except for hy- aline margin of shallowly toothed orifice, teeth 0.2 mm long. Dried stipe shriveled and discolored, 0.4— 0.8 mm long below achene, plump and pale when fresh. Achene_ globose-trigonous, faces convex, MADRONO [Vol. 45 tightly filling the central perigynium body, 1.9—2.2 x 1.4—-1.8 mm, medium to light brown, finely pa- pillate (at 40) in longitudinal rows; style base 0.1 mm long, cleanly articulated. Paratypes. USA, Oregon, Curry Co.: Berry Cab- in, between Game Lake and Horse Sign Butte, ca. 13 km S of Agness, 1 May 1931, Leach 3325 (ORE); Pistol River Hill Rd., E of Ismert Ln., 8 May 1988, Stansell s.n., 3114 (WS), 3 May 1995, Stansell 3041 (WS); 8 km S of Gold Beach, 23 Jul 1945, Peck 23951 (WILLU); Lemmingsworth Gulch, 17 Apr 1984, Stansell s.n. (OSC); 2.4 km W of Doe Gap, 26 Jun 1993, Rolle 608 (OSC); ca. 1.6 km SW of Signal Butte, May 1990, Stansell s.n. (OSC, RSA, UC, WTU); ca. 3.8 air km NE of Sig- nal Butte, 1 May 1989, Stansell s.n. (OSC), 6 May 1995, Wilson et al. 7630 (WS); Dry Butte Trail, Kalmiopsis Wilderness, 4 Jun 1994, Stansell 3023 (IDS); near Hunter Cr. Bog, 19 Jun 1997, Zika et al. 13167 (OSC); Josephine Co.: S of Mikes Gulch, 4.3 air km ENE of Fiddler Mt., 20 Apr 1993, Zika 11972 (WTU); road to Day’s Gulch, Josephine Cr., 15 Apr 1984, Greenleaf 1473 (ORE); ca. 0.4 air km NW of Eight Dollar Mt., 30 May 1988, Zika 10485 (OSC); 2.1 air km SSW of Eight Dollar Mt., 21 Apr 1993, Zika 11975 (MICH, OSC, WTU); 2.4 air km SSW of Eight Dollar Mt., 21 Apr 1993, Zika 11976 (WS); ca. 1.2 road km SW of Illinois River bridge, FS Rd. 4201, 28 May 1994, Wilson et al. 6790 and 6791 (OSC), 23 Mar 1996, Clery et al. 57 (OSC); near small roadside pool, 1.3 road km E of Illinois River bridge on FS Rd. 4201, 22 April 1995, Zika 12297 (UBC, UC), 14 May 1994 Wilson et al. 6793 (GH, ID, MICH, OSC, RSA, WTU); slope above and N of Illinois River bridge on FS Rd. 4201, 24 May 1996, Zika 12865 (WTU); BLM fen, S of Eight Dollar Mt., 18 April 1984, Stansell s.n. (OSC), 23 March 1996, Clery 55 et al. (OSC, WS, WTU); Star Flat Rd., NNW of Eight Dollar Mt., 17 May 1990, Rolle 329 (WTU); NE of Wood- cock Mt., 23 May 1995, Perkins 950523 (IDS), 24 May 1995, Newhouse 95001b (WTU); near N rid- geline of Woodcock Mt., 15 Jun 1995, Perkins 950615 (UBC, WS), 23 May 1995, Newhouse 95002c (OSC); Oak Flat Rd., 21 km W of Selma, 18 Apr 1969, White and Lillico 148 (OSC), Pearsoll Peak Rd., ca. 32 road km W of Selma, 13 May 1962, Addor 1425 (ORE); Chrome Ridge, 29 May 1995, Rolle 893 (US). California, Del Norte Co.: N side of Middle Fork of Smith River, on Old Gasquet Toll Rd., near Gasquet, 1 Jun 1935, Parks and Tra- cy 11200 (HSC), 14 May 1994, Wilson 6803 (RSA); Gasquet Mountain, close to town of Gas- quet, 18 Jun 1979, Clifton and Griswold 5379 (HSC); Humboldt Flat, 5.6 km S on French Hill Rd. from Rte. 199, 14 May 1983, Janeway 292 and 293 (HSC); Old Gasquet Toll Rd., 0.2 km N of southern entrance, 12 Apr 1975, Barker 247 (HSC); French Hill Rd., 0.6 km above Rte. 199, Gasquet, 3 May 1995, Zika 12322 (OSC); Old Gasquet Toll 1998] B =| VV IE Fic. 2. ZIKA ET AL.: C. SERPENTICOLA, NEW SPECIES C 263 I cm Carex serpenticola P. FE Zika. A) Carex serpenticola 3 spikes, note bracts marking aborted lateral spikes (Rolle 608 OSC; Curry Co., Oreg.); B—E) infructescences (Zika 12322 OSC; Del Norte Co., Cal.); B) terminal spike 3, lateral spike(s) 2; C) terminal spike 2/d, lateral spikes 2; D, E) all spikes 2; F) lax habit when fruiting (Newhouse 95001B WTU; Josephine Co., Oreg.). Rd., 1.1 km above river bridge, 3 May 1995, Zika — 12323 (IDS, MICH, OSC, WTU); Old Gasquet Toll Rd., 2.3 km N of southern entrance, 7 Mar 1979, _ Barker 211 (HSC); Old Gasquet Toll Rd., 4.3 km N of southern entrance, 15 May 1974, Barker 261 - (HSC); N entrance to Old Gasquet Toll Rd., con- _ fluence of E and W Forks of Patrick Cr., 19 May 1997, Zika 13082 (OSC); near Stony Cr., ca. 2 km _ W of Cold Spring Mt., 1 Jun 1980, Baker 1667 _ (HSC); Stony Cr. Bog, 13 May 1973, Smith 6712 _ (HSC); near Browns Mine, | Jun 1980, York 899 _ (ASC); near Cold Spring Mt., 1 June 1980, Baker _ 1650 (HSC); near Diamond Cr., North Fork, 2 Jun _ 1980, York 944 (HSC); along Elk Camp Ridge NE of Gasquet, 7 Mar 1975 and 12 Apr 1975, Klipfel 358 and 359 (HSC); near Wimer Rd., close to town _ of Smith River, 21 May 1979, Clifton and Overton _ 2706, 2976, and 3372 (HSC), 6 May 1994, Stansell 3117 (WTU); Low Divide Rd., 2300 feet, 6 May _ 1994, Stansell 3115 (WS); near junction of Wimer _ and Low Divide Rds., 29 Apr 1997, Zika and Kuy- _ kendall 13068 (OSC); Low Divide, Wimer Rd., 29 Apr 1997, Zika and Kuykendall 13075, 13076 (OSC); ridge S of Black Butte, 18 Jun 1994, Stan- sell 3060 (WS). Distribution. Carex serpenticola appears to be restricted to the Klamath Mountains of California and Oregon, west of the Cascade Mountain high- lands. We documented populations in three coun- ties: Del Norte Co., CA, and Curry and Josephine Cos., OR. Carex serpenticola ranges as far south- west as the xeric serpentine plant communities in the Smith River drainage, in the Gasquet region of Del Norte Co. Carex globosa ranges north to within 8 km of C. serpenticola, in the wet maritime red- wood forest zone on the lower reach of the Smith River in Del Norte Co. (Zika 13087 WTU). From there C. globosa ranges south and east on the coast and in the coastal mountains. We mapped the dis- tribution of the two species based on our collections and herbarium materials (Fig. 1). Affinities. Based on morphology, we considered placement of Carex serpenticola in the three (rather loosely defined) subgeneric sections discussed be- low. 264 Sectional placement. Although similar to Carex section Scirpinae in gross morphology, C. serpen- ticola fails to fit there because its pistillate and bi- sexual culms are multispicate. Culms in C. sect. Scirpinae are generally unispicate, but when mul- tispicate and bisexual (e.g., C. scabriuscula Mack.), the perigynial architecture is quite different. A bet- ter fit is in the circumboreal C. sect. Acrocystis. Members of the section are generally plants of up- land habitats characterized by plump, pubescent, tristigmatic perigynia in short, subterminal spikes, subtending a staminate spike, with essentially sheathless inflorescence bracts and basal foliage that is glabrous in the region described in this study Carex serpenticola shares the growth habit, upland ecology and general morphology of species in sec- tion Acrocystis. We note that recent unpublished molecular investigations (Roalson personal com- munication.) support our morphological data, and place C. serpenticola among the western members of C. sect. Acrocystis. The weak boundary between Carex section Ac- rocystis and C. sect. Digitatae led Koyama (1962) to synonymize them. North American authors still segregate them, although traditional key characters such as achene shape (Mackenzie 1931-1935; Her- mann 1970) fail to separate the sections. Carex ser- penticola does not share characters now used for defining C. sect. Digitatae: well-developed bracteal sheaths, insignificant bracteal blades, and poorly developed perigynial beaks that are essentially toothless and <0.5 mm long. Carex concinna R. Br. and C. concinnoides Mackenzie are anomalies in C. sect. Digitatae, with unusual warty and non- plumose styles. Carex concinnoides is also excep- tional with its highly reduced bracteal sheath, and relatively elongate beak, and closely resembles C. serpenticola. These species could be viewed as transitional between the sections Acrocystis and Digitatae, and molecular studies are recommended for a fresh definition of the sectional boundaries. Similar species. Carex section Acrocystis 1s found in the Old and New World, however no Eur- asian taxa occur in North America (Hultén 1968; Fernald 1950). None of the Eurasian species are a close match to the unusual key characters of C. serpenticola, which combines long rhizome inter- nodes, unisexual culms and a dark (almost black) scale color. Eastern North American taxa in the sec- tion are not similar to C. serpenticola. Although C. umbellata Schk. ex Willd. does show a tendency to produce unisexual culms, it differs substantially in other characters. Based on morphology, geography, and ecology, Carex serpenticola is most similar to C. globosa among western members of C. sect. Ac- rocystis. Collections of the two species have con- fused more than one respected caricologist. Both plants share features such as stem height, leaf length, leaf width, loosely spreading habit, wide ob- ovoid perigynia with pronounced swollen stipes MADRONO [Vol. 45 and distinct, abruptly tapered beaks, and sharp- tipped pistillate scales that are widely spreading in the mature infructescence. Unisexual culms are common on C. serpenticola, but rare among similar taxa in western North America. Carex globosa culms can be unisexual (Howell 1949), further sug- gesting it is related to C. serpenticola. Geographi- cally the two taxa are adjacent to each other along the Pacific coast (Fig. 1). And ecologically, both species occur on ultramafics, an unusual edaphic tolerance in the section. Important morphological distinctions between Carex globosa and C. serpenticola follow. Ordi- narily C. globosa does not produce unisexual stems. The perigynia of C. globosa are longer, 3.9— 5.1 mm, including prominent discolored stipes. Measured from the base of the swollen achene to the base of the perigynium, stipes are 1.2—2.3 mm long. The perigynial faces of C. globosa have 2-5 thick longitudinal nerves (Fig. 3G). Pistillate scales are predominantly green with 2—4 nerves parallel- ing the midvein (Fig. 3G). By comparison, most C. serpenticola culms are unisexual. This species has perigynia 3.1—3.6 mm long, with stipes 0.4—0.8 mm long. Carex serpenticola perigynia have thick mar- ginal ribs, but usually lack nerves on the faces. If facial nerves are present, they are generally thin and short. The perigynia of the two taxa look dif- ferent because of the relatively long stipes charac- teristic of C. globosa (Fig. 3G, S). Finally, pistillate scales in C. serpenticola are predominantly dark purple-black with a single midvein, and lack par- allel nerves. On rare occasions a few scales or per- igynia on a culm have some nerves suggesting C. globosa, but the majority of scales and perigynia are not nerved. In summary, the two taxa can be separated reliably by perigynium dimensions and nerving, scale color and nerving, and presence or absence of unisexual culms within populations. In addition, the two have different habits in nature. Carex globosa is a loosely clumped species, with more vertical than horizontal rhizome development. Carex serpenticola tends to spread, forming diffuse mats or turfs, or producing small clumps of vertical shoots on elongate horizontal rhizomes. Habit dif- ferences are not easily assessed on most herbarium sheets. Perigynium length is a reliable discriminator be- tween Carex globosa and C. serpenticola, with a . clear hiatus. Carex serpenticola perigynia are 3.1— | 3.6 mm long. Munz and Keck (1965) reported C. | globosa perigynia were 4—5 mm long and Jepson © (1925) reported they were 5 mm long. Although | Mastrogiuseppe (1993) reported C. globosa peri- gynia at 3.3—-5.0 mm long, we found a range of 3.9— 5.1 mm for C. globosa perigynia from Marin Co., | CA, and north and could not find mature specimens — in the lower range reported by her. The C. globosa © holotype has perigynia 4.0—4.7 mm long. In the herbarium, Carex serpenticola collections | commonly have been mistaken for C. concinnoides. 1998] ZIKA ET AL.: C. SERPENTICOLA, NEW SPECIES 265 / Lhe 1 mm 1 mm | Fic. 3. Perigynia, scale and stigmas. All drawings prepared from dried herbarium materials. B) Carex brainerdii perigynium (Zika 12372 OSC; Klamath Co., Oreg.); C) C. concinnoides stigma and perigynium (Miller 56 OSC; Grant Co., Oreg.); G) C. globosa scale and perigynia (Zika 12316 OSC; Marin Co., Cal.); R) C. rossii perigynium (Zika 12300 OSC; Josephine Co., Oreg.); S) C. serpenticola stigma (Zika 11972 WTU) and perigynium (Zika 12297 UBC; _ both from Josephine Co., Oreg.). _ The latter has a much wider range, across much of _ western North America. Where sympatric, the two _ have a similar habit and habitat, although C. con- _cinnoides is less tolerant of wetland soils. Mixed _ collections and our observations show they some- times grow intermingled. Their perigynia (Fig. 3C, S), dark scales, and arching bisexual culms (Fig. 2F) are strikingly similar, suggesting a close rela- tionship. However, the subtle but profound differ- ences between the species suggests to us conver- 266 gent evolution (presumably for seed dispersal), rather than recent divergence from common ances- try. Only C. serpenticola generates unisexual culms. Furthermore, they are quickly and reliably separated by the thick warty stigmas typical of C. concinnoides (Fig. 3C) which contrast with the slender, subplumose stigmas in C. serpenticola (Fig. 3S). Carex concinnoides also has 4 styles and quadrangular-based globose achenes. Carex serpen- ticola has 3 styles and trigonous-based globose achenes. Although Peck (1961), Hitchcock et al. (1969), and others have suggested Carex concinnoides can have three stigmas, our study revealed only 4-stig- matic flowers, echoing the results of St. John and Parker (1925). We speculate that Hermann (1970) may have perpetuated reports of three stigmas in C. concinnoides based on his examination of un- recognized C. serpenticola from northern Califor- nia. As recently as 1977, Hermann annotated 3- styled C. serpenticola as C. concinnoides (e.g., Parks and Tracy 11200 HSC, Barker 261 HSC). Synonymy. We found no available synonyms of California and Oregon species of section Acrocystis that might apply to C. serpenticola, nor has A. A. Reznicek (personal communication). The holotype of Carex globosa (Nuttall s.n., K!) is merely la- beled “California”? and lacks a specific locality or date (Boott 1845). It was most likely collected along the coastline, at a point accessible to ships of the day, no further north than Monterey, in Cali- fornia Alta, in spring of the year 1836, according to historical summaries of Nuttall’s travels in the American west (e.g., McKelvey 1955). This places the C. globosa holotype well south of the known range of C. serpenticola. We have also examined MADRONO [Vol. 45 the holotype of C. concinnoides (Williams s.n., NY!) from ‘‘Columbia Falls, MT,”’ far to the north- east of the known range of C. serpenticola (Mac- kenzie 1906). Taxonomic rank. We rejected placement of the new taxon as a subspecies of Carex globosa. More than two consistent morphological features differ- entiate them. They are entirely allopatric. Both can occur on low-elevation serpentine meadows along the coast (C. globosa in California, C. serpenticola in Oregon), but in these similar ecological condi- tions they maintain their morphological differences. We found no intermediate plants to suggest wide- spread introgression or hybridization in populations we Studied in the field. We believe these differences are best summarized by assigning the rank of full species to C. serpenticola. Dichotomous identification key. Our key is for identification of all pubescent-fruited Carex from the Klamath Mountains, west of the Cascade Mountains. The key is actually composed of 3 keys, one for staminate culms, one for pistillate culms, and one for bisexual culms. We excluded species such as C. mendocinensis and C. scopulorum Holm, as their perigynia have only sparse apical bristles and are not densely or uniformly pubescent. Our key includes all six C. sect. Acrocystis taxa known from in or near the Klamath region (C. brainerdii Mackenzie, C. brevicaulis Mackenzie, C. globosa, C. inops L. Bailey ssp. inops, C. rossii Boott and C. serpenticola). Like Hitchcock et al. (1969), we elected to treat C. rossii in the broad sense, includ- ing C. brevipes W. Boott and C. diversistylis Roach. We followed Dunlop (1997) and submerged C. gi- . gas (Holm) Mackenzie in C. scabriuscula Mack. A KEY TO PUBESCENT-FRUITED CAREX in the Klamath Mountains 1. Culms all unisexual 2. Culms staminate 2' Culms pistillate 3. Base of flowering culms scaly, lacking fibers; spikes 2—4 3’ Base of flowering culms lacking scales, fibrous; spikes solitary 4. Ventral surface of sheath glabrous; ligules wider than tall 4’ Ventral surface of sheath pubescent; ligules taller than wide C. scirpoidea Michaux var. pseudoscirpoidea (Rydb.) Cronq. | Bsn eee ot Seen ee Stee aes cca Ss) |i 5. Perigynia flat except over relatively small achenes, sessile, bases truncate 5' Perigynia plump, filled by achenes, with distinct stipes or tapered to cuneate bases C. scabriuscula ie ARGS age ie OL em 4 6. Spikes 2—4, lowest well-separated or obvious, elliptic-ovate to subglobose (Fig. 2D, E); perigynia beaks bidentate 6’ Spikes usually solitary, rarely with 2 spikes, these overlapping; oblong-cylindrical; perigynia beak tips essentially entire, toothless 1’ Some or all culms bisexual 7. Stigmas 4, (Fig. 3C); stigmatic surface warty, non-plumose at 15x (Fig. 3C); achenes quadrangular-based 7’ Stigmas 3 (Fig. 3B, G, R, S); stigmatic surface finely plumose at 15x (Fig. 3S); achenes trigonous-based 8. Foliage pubescent; perigynia with 3 flat faces, obviously trigonous 8' Foliage glabrous; perigynia with 2 flat faces or plump-globose, never obviously trigonous C. serpenticola C. scirpoidea var. pseudoscirpoidea | C. concinnoides | C. gynodynama Olney © 1998] ZIKA ET AL.: C. SERPENTICOLA, NEW SPECIES 9. Perigynia body flattened except over the relatively small achene; perigynia shiny, purplish-black C. scabriuscula 9’ Perigynia body plump, filled by the relatively large achene; perigynia dull, green to brown or purple 10. Numerous short-peduncled basal spikes hidden among the rosettes 11. Pistillate scales with 1 prominent midvein, perigynia lacking nerves 12. Perigynia > 1.7 mm wide, body subspherical; staminate scales scabrous near tip of midab; coastal dunes and headlands... 64.22.46 «444 ske ees ee aces C. brevicaulis 12’ Perigynia < 1.6 mm wide, body elliptic; staminate scales not scabrous; Coast Range and “nlanG & 4.4 ale 2 e-5 sa. chess ee ee eee ee ee eae eee C. rossii (Fig. 3R) 11’ Pistillate scales with 3—5 prominent nerves; perigynia strongly nerved............. 13 13. Foliage pale or glaucous; perigynia with beaks ca. equal stipes; perigynia bodies barrel-shaped, broadest near middle; leaves generally papillate on underside at 40; all basal spikes erect on short peduncles ................. C. brainerdii (Fig. 3B) 13’ Foliage green, not glaucous; perigynia with stipes ca. 1.5—2 times as long as beaks, perigynia bodies obovoid, broadest near beaks; leaves papillate only on veins or epapillate at 40; some basal spikes arched on long peduncles ...... C. globosa (Fig. 3G) 10’ Lacking short-peduncled basal spikes 14. Staminate spikes 2—3, the terminal spike > 30 mm long; culms 60—100 cm tall; emergent in shallow marshes ........... C. pellita Muhl. 14’ Staminate spikes solitary and < 30 mm long; culms < 50 cm tall; upland mesic or xeric sites 15. Scales and base of inflorescence bract green to red (like apples); pistillate scales and lower staminate scales with conspicuous white margins 0.4—0.8 mm wide; popula- tions lacking unisexual culms; approaching our area on volcanic substrates in the Cascade Mountains C. inops ssp. inops 15’ Scales and base of inflorescence bract dark purple or black (like beets); pistillate scales and lower staminate scales with thin white margin 0.1—0.2 mm wide; popu- lations primarily unisexual culms; ultramafic substrates in the Klamath Mountains Sexual expression. The facultative production of unisexual culms is unusual in Carex section Acro- cystis. Asian and eastern North American taxa with sporadic unisexual culms are otherwise not very similar to Carex serpenticola. Howell (1949) noted that some Carex globosa, when on serpentine, can produce a “‘terminal inflorescence . . . either entire- ly staminate or ... [with] 1 (or rarely 2) pistillate flowers.”” His voucher (Howell 19368A RSA!) ap- proaches C. serpenticola, with some culms that ap- parently aborted most or rarely all lateral pistillate flowers. These plants, however, maintained the di- agnostic heavily nerved perigynia and 3—5-nerved _ pistillate scales (Fig. 3G) of C. globosa. Fewer than 1% of the C. globosa stems we examined in the field and herbarium exhibited unisexual culms (by reduction). Culms of C. globosa (commonly) and _C. serpenticola (rarely) produce essentially basal _long-peduncled pistillate spikes, which might be confused with unisexual culms. Carex serpenticola produces two culm types, bi- _ sexual and unisexual. We sampled 179 stems from 15 populations, and found 92% of the culms were unisexual (Fig. 2A, D). Unisexual culms were sta- _ Minate (74 stems, 41% of sample) or pistillate (90 stems, 50%). One rhizome can form both truly sta- minate culms and functionally staminate culms. _ The latter exhibit a terminal staminate spike and one (or more) aborted lateral pistillate spikes. Aborted spikes are marked by a bract (Fig. 2A). _Some culms in most populations develop 1-2 lat- C. serpenticola (Fig. 2) eral pistillate spikes near the terminal staminate spike (Fig. 2B). One rhizome can bear both unisex- ual and bisexual culms. Thus C. serpenticola, like most members of the genus, is not dioecious. Stan- dley (1985) found dioecy in fewer than 10 of the ca. 1500 Carex species worldwide. Bisexual culms of C. serpenticola have staminate (Fig. 2B) or rare- ly gynecandrous (Fig. 2C) terminal spikes. Our field observations raise questions about the influ- ence of age, site conditions, and vagaries of climate on the sex of inflorescences. More research is need- ed to explain sexual expression in C. serpenticola. Ecology. Carex serpenticola is associated with the largest ultramafic exposures in North America. The soils derived from ultramafic bedrock are gen- erally referred to as serpentine by biologists. In a region otherwise heavily forested, serpentine soils support sparse, low-productivity plant communities like Darlingtonia fens, Umbellularia-Rhododen- dron occidentale (Torrey & A. Gray) A. Gray ri- parian strips, Pinus jeffreyi Grev. & Balf. savannas, and transitional or successional scrublands domi- nated by Quercus vaccinifolia, Rhamnus californi- ca, and Lithocarpus densiflorus (Hook. & Arn.) Rehder var. echinoides (R. Br. Campst.) Abrams (Jimmerson et al. 1995). We located populations of C. serpenticola in or at the margins of each of these serpentine vegetation assemblages. Habitats sup- porting the most robust individuals of C. serpenti- cola were vernally moist to mesic meadows and 268 TABLE 1. ULTRAMAFIC ENDEMICS (SENSU KRUCKEBERG 1984; SMITH AND SAWYER 1988) FOUND AT CAREX SERPEN- TICOLA SITES. Arnica cernua Howell Calochortus howellii S. Wats. Cardamine nuttallii E. Greene var. gemmata (E. Greene) Rollins Erythronium citrinum S. Watson Hastingsia bracteosa S. Wats. Lomatium howellii (S. Watson) Jepson Poa piperi A. Hitchc. Salix delnortensis C. Schneider Sanicula peckiana J. F Macbr. Senecio hesperius Greene Viola cuneata S. Watson riparian strips, at or near the upland-wetland bound- ary. Populations were observed between 60 and 1200 m elevation on slopes of all aspects. Slopes were flat or gentle at those sites with the largest plants. Compact plants with short culms and falcate leaves were found on steeper, well-drained slopes. The forests were usually serpentine savannas or ri- parian strips, dominated by Pinus jeffreyi, Caloced- rus decurrens (Torrey) Florin, Arctostaphylos nev- adensis A. Gray, A. patula E. Greene, A. viscida C. Parry, Ceanothus cuneatus (Hook.) Nutt., C. pum- ilis E. Greene, and Juniperus communis L. Some associated non-endemic taxa, typical of serpentine soils, were Aspidotis densa (Brackenr.) Lellinger, Carex mendocinensis Olney, Cerastium arvense L., Deschampsia cespitosa (L.) Beauv., Erythronium citrinum S. Watson, Festuca idahoensis Elmer, Hastingsia alba (Durand) S. Watson, Microseris howellii Gray, Poa secunda J. S. Presl, Ranunculus calijornicus Benth., R. occidentalis Nutt., Scirpus criniger Gray, Trillium rivale S. Watson, and Viola lobata Benth. Carex serpenticola appears to be a narrowly dis- tributed edaphic endemic. Many of its associates are also endemic to the ultramafic exposures in the area (Table 1). Unlike C. serpenticola, C. globosa is a bodenvag species, abundantly documented from both ultramafic and other bedrock exposures, at least from Marin Co., California and south. Cu- riously, we have not seen evidence that C. globosa grows on serpentine soils north of Marin County. Individuals of C. serpenticola from wetland mar- gins are often large and easy to locate, at least prior to the summer drought. Small individuals with a compact growth form, widely scattered across drier serpentine slopes, are harder to find. We have noted that C. globosa occupies different habitat than C. serpenticola where their ranges con- verge in northern California. The former is found in partial to full shade, under or at the edge of dense forests with sparse understory, close to the coast and at low elevations (<300 m). Carex globosa is on well-drained non-serpentine soils in this wet cli- matic region. Dominant woody plants are often Se- MADRONO [Vol. 45 quota, sempervirens, (D. Don) Endl. Quercus chry- solepis Leibm., Lithocarpus densiflorus, or Pseu- dotsuga menziesii (Mirbel) Franco, over Toxicod- endron diversilobum (Torrey & A. Gray) E. Greene and Vaccinium ovatum. From Marin Co., Califor- nia, and south, however, Carex globosa is well doc- umented in a variety of different plant communities and on serpentine or non-serpentine soils (e.g., Har- dham 1962; Howell 1949). Carex serpenticola individuals flower and fruit earlier than most Carex in the region, which may help explain the paucity of specimens found in her- baria. Flowering occurs from the first week of March to early May. Mature fruiting plants were collected between late April and June, depending on elevation and exposure. Pollination ecology might explain why the stig- matic surfaces differ in C. serpenticola and C. con- cinnoides. Carex serpenticola and C. sect. Acro- cystis display inconspicuously colored, slender sub- plumose stigmas (Fig. 3S). They are typical wind- pollinated sedges, and we never observed insects visiting their flowers. Leppik (1955) reported in- sect-pollinated Carex spp. with bright white floral displays composed of showy bracts, and bright white anthers and stigmas. Carex concinnoides may be evolving towards entomophily. We observed Hymenoptera occasionally visiting the striking white clustered presentation of stigmas on C. con- cinnoides, topped with showy pale yellow to whit- ish anthers. The broad, conspicuous warty stigmas of C. concinnoides (Fig. 3C) might attract pollina- tors and enhance pollen capture from non-aerial sources. We are not aware of seed dispersal investiga- tions among the representatives of Carex sect. Ac- rocystis in the Klamath region. Some taxa (e.g., C. brainerdii, C. brevicaulis, and C. rossii) present their mature fruits close to ground level on abbre- viated culms. Other members of the section posi- tion some of their fruit output on the ground with long, weak, basal peduncles (e.g., C. globosa). Carex concinncides (sect. Digitatae) and C. ser- penticola have erect flowering culms that arc to place infructescences on the soil surface (Fig. 2F). All these growth habits suggest adaptations facil- itating diaspore collection by ants in upland hab- itats (Handel 1976, 1978). Seed dispersal by ants is well documented in a variety of plants and hab- itats (e.g., Beattie 1985; Bossard 1991; Boyd 1996). The plump, pale fresh stipes in C. globosa and C. serpenticola resemble structures described or illustrated as ant lures in sedges such as C. com- munis Bailey, C. jamesii Schwein., C. laxiculmis Schwein., C. pedunculata Muhl., C. pilulifera L. and C. umbellata Schkuhr (Kjellsson 1985a, b; Beattie and Culver 1981). Dried herbarium mate- rial (Fig. 3) does not reveal how expanded the stipe is in nature. We observed the velvety tree ant (Liometopum occidentale, Formicidae, Dolichod- erinae) carrying off mature perigynia we scattered 1998] on the ground near fruiting C. serpenticola. We suspect myrmecochory in C. serpenticola, which differs from most Carex in its upland habitat, its ‘‘weeping’’ stems, and peculiar perigynial bases. Conservation status. Carex serpenticola appears to be a narrow endemic (Fig. 1), but its status, dis- tribution and abundance are still poorly document- ed. We discovered populations inland on extensive serpentine areas, as well as coastal colonies on small, isolated serpentine knobs. C. serpenticola may be a widespread and characteristic member of the regional ultramafic flora, on what geologists re- fer to as the Josephine ophiolite. ACKNOWLEDGMENTS We are grateful for the efforts of collectors such as Veva Stansell and Wayne Rolle. Members of the Carex Working Group contributed field time, herbarium analysis, com- puter expertise, and partial funding for publication costs. The line drawings are by John Megahan. Anthony Rez- nicek and anonymous reviewers generously improved the manuscript. We are pleased to acknowledge Rhoda Love’s assistance with historical details, an upgrading of the Latin by Kenton Chambers, and ant identification by Lynn A. Royce. Key ant observations were made by Elizabeth Gould. Partial funding for travel came from the Native Plant Society of Oregon, (including Corvallis, Emerald, and Siskiyou chapters), Carex Working Group, the Cali- fornia Native Plant Society (North Coast, Lassen, and Shasta chapters), Oregon Natural Heritage Program, Rogue River National Forest, Siskiyou National Forest, and the Coos Bay District of the Bureau of Land Man- agement. Curators at the following institutions provided loans or access to collections: A, BM, CAS, GH, HSC, K, MICH, NY, OSC, ORE, RSA, SOSC, UC, US, VT, WILLU, WS, and WTU. For allowing us to review spec- imens, we thank the staff at Oregon herbaria not listed in Holmgren et al. (1990), including Roseburg and Coos Bay Districts of the Bureau of Land Management, Crater Lake National Park, Oregon Caves National Monument, and the Douglas Co. Museum. LITERATURE CITED BEATTIE, A. J. 1985. The evolutionary ecology of ant- plant mutualisms. Cambridge University Press, New York. NY. BEATTIE, A. J. AND D. C. CuLver. 1981. The guild of myrmecochores in the herbaceous flora of West Vir- ginia forests. Ecology 62:107—-115. Boott, F 1845. Caricis species novae vel minus cognitae. Proceedings of the Linnaean Society 1:254—261. Bossarb, C. C. 1991. The role of habitat disturbance, seed predation and ant dispersal on establishment of the exotic shrub Cyfissus scoparius in California. Amer- ican Midland Naturalist 126:1—13. Boyp, R. S. 1996. Ant-mediated seed dispersal of the rare chaparral shrub Fremontodendron decumbens (Ster- culiaceae). Madrofio 43:299-315. DuNnLop, D. A. 1997. Taxonomic changes in Carex (sec- tion Scirpinae, Cyperaceae). Novon 7:355—356. FERNALD, M. L. 1950. Gray’s Manual of Botany, 8th ed. D. Van Nostrand Co., New York. NY. HANDEL, S. N. 1976. Dispersal ecology of Carex pedun- culata (Cyperaceae), a new North American myr- ZIKA ET AL.: C. SERPENTICOLA, NEW SPECIES 269 mecochore. American Journal of Botany 63:1071— 1079. HANDEL, S. N. 1978. The competitive relationship of three woodland sedges and its bearing on the evolution of ant-dispersal of Carex pedunculata. Evolution 32: 151-163. HARDHAM, C. B. 1962. The Santa Lucia Cupressus sar- gentii groves and their associated northern hydroph- ilous and endemic species. Madrono 16:173—178. HERMANN, FE J. 1970. Manual of the Carices of the Rocky Mountains and Colorado Basin. Agriculture Hand- book No. 347. U.S. Forest Service, Department of Agriculture, Washington. DC. HitcHcock, C. L., A. CRONQUIST, AND M. OwnBey. 1969. Vascular plants of the Pacific Northwest. Part |: Vas- cular cryptogams, gymnosperms, and monocotyle- dons. University of Washington Press, Seattle. WA. HOLMGREN, P., N. H. HOLMGREN, AND L. C. BARNETT. 1990. Index herbariorum. Part |: The herbaria of the World. 8th ed. New York Botanical Garden, Bronx. NY. Howe LL, J. T. 1949. Marin flora: Manual of the flowering plants and ferns of Marin County, California. Uni- versity of California Press, Berkeley. CA. HULTEN, E. 1968. Flora of Alaska and neighboring terri- tories. Stanford University Press, Stanford. CA. JEPSON, W. L. 1925. A manual of the flowering plants of California. University of California Press, Berkeley. CA. JIMERSON, T. M., L. D. Hoover, E. A. MCGEE, G. DE- NitTo, AND R. M. CreASy. 1995. A field guide to serpentine plant associations and sensitive plants in northwestern California. USDA Forest Service, Pa- cific Southwest Region. Publication R5-ECOL-TP- 006. KJELLSSON, G. 1985a. Seed fate in a population of Carex pilulifera L. 1. Seed dispersal and ant-seed mutualism. Oecologia 67:416—423. KJELLSSON, G. 1985b. Seed fate in a population of Carex pilulifera L. Il. Seed predation and its consequences for dispersal and seed bank. Oecologia 67:424—429. Koyama, T. 1962. Classification of the family Cyperaceae (2). Journal of the Faculty of Science, University of Tokyo 8:149—278. KRUCKEBERG, A. R. 1984. California serpentines: flora, vegetation, geology, soils, and management prob- lems. University of California Press, Berkeley. CA. LEACH, J. AND L. LEACH. 1938. Botanizing in Oregon’s hinterland. Mazama 20:59-—62. Leppik, E. E. 1955. Dichromena colorata, a noteworthy entomophilous plant among Cyperaceae. American Journal of Botany 42:455—458. Love, R. M. 1991. The discovery and naming of Kal- miopsis leachiana and the establishment of the Kal- miopsis Wilderness. Kalmiopsis 1|:3—8. MACKENZIE, K. K. 1906. Notes on Carex—I. Bulletin of the Torrey Botanical Club 33:439. MACKENZIE, K. K. 1931-1935. North American Flora. (Poales) Cyperaceae—Cariceae. The New York Bo- tanical Garden, Bronx. NY. MASTROGIUSEPPE, J. 1993. Carex, sedge. in J. C. Hickman (ed.), The Jepson manual. Higher Plants of California. University of California Press, Berkeley. CA. McKe vey, S. D. 1955. Botanical exploration of the Trans-Mississipp1 West 1790-1850. The Arnold Ar- boretum of Harvard University, Jamaica Plain. MA. Munz, P. A. AND D. D. KEcK. 1965. A California flora. University of California Press, Berkeley. CA. 270 MADRONO [Vol. 45 Peck, M. E. 1961. A manual of the higher plants of Or- cies, section, and subgenus of Carex. American Jour- egon, 2nd ed. Binfords and Mort, Portland. OR. nal of Botany 12:63—69. SMITH, J. P. AND J. O. SAWYER. 1988. Endemic vascular STANDLEY, L. A. 1985. Paradioecy and gender ratios in plants of northwestern California and southwestern Carex macrocephala (Cyperaceae). The American Oregon. Madrono 35:54—69. Midland Naturalist 113:283—286. ST. JOHN, H. AND C. S. PARKER. 1925. A tetramerous spe- Mapbrono, Vol. 45, No. 3, pp. 271-272, 1998 NOTES ANTENNARIA DIOICA (ASTERACEAE: INULEAE): ADDI- TION TO THE VASCULAR FLORA OF CALIFORNIA.—Jer- ry G. Chmielewski, Department of Biology, Slip- pery Rock University, Slippery Rock, PA 16057, USA. The California flora includes about one-third of the North American Antennaria species (Bayer 1984; Bayer and Stebbins 1993; Chmielewski 1997; Chmielewski and Chinnappa 1988; Chmie- lewski et al. 1990). Specifically, Jepson (1925) in- cluded eight species, Munz (1959) included 10 spe- cies, Ferris (1960) included 13 species, and Steb- bins and Bayer (1993) included 14 species. Only four species, A. argentea Benth., A. dimorpha (Nutt.) Torrey & A. Gray, A. geyeri A. Gray, and A. luzuloides Torrey & A. Gray were consistently included among the species by these respective au- thors. Here I report the occurrence of Antennaria dioi- ca (L.) Gaertner in the California flora. Although Jepson (1925) included A. dioica, the varieties con- gesta DC., corymbosa Jepson, kernensis Jepson, marginata Jepson, and rosea Eat. were not consis- tently considered to be part of the species, but rath- er distinct species or synonyms of others, in sub- sequent treatments of either the flora (Munz 1959; Ferris 1960; Stebbins and Bayer 1993) or genus (Bayer 1984; Bayer and Stebbins 1993; Chmie- lewski 1997; 1998; Chmielewski and Chinnappa 1988; Chmielewski et al. 1990). Antennaria dioica, the mountain or dioecious cat’s-paw, is a diploid (Fedorov 1969, Bayer 1984), dioecious, mat-forming species that was described from Europe. The species is characterized by green, glabrous, occasionally tomentose adaxial leaf sur- faces and white or pink, broadly obovate (staminate plants) or oblong-obovate (pistillate plants) invo- lucral bracts. It previously was reported to be dis- tributed across northeastern Europe from Scandi- navia and the British Isles, eastward into Asia as far as Japan, northward to the Kamchatka Penin- sula, eastward to Bering Island and western Aleu- tian Islands (Hultén 1949; 1968; Anderson 1959; Polunin 1959; Welsh 1974). The species typically inhabits heaths, dry grasslands, sandy or stoney slopes, and gravelly soil in the alpine zone (Polunin 1959; Tutin et al. 1976). The single collection of Antennaria dioica that is the basis of this report consists of three staminate shoots bearing immature, pink-bracted involucres. These specimens are morphologically similar to hundreds of collections of A. dioica that I have ob- served from the Aleutian Islands and Europe. Hul- tén (1949) indicated that Aleutian Island material did not appear to be closely related to any Ameri- can species of Antennaria known to him. Circum- scription of the species by Jepson (1925), however, indicated similar features to A. marginata E. L. Greene [syn. = A. dioica var. marginata (E. L. Greene) Jepson] from southwestern North America (Bayer and Stebbins 1987; Chmielewski et al. 1990). The two taxa differ in that the flowering stems and leaves of A. marginata bear distinctive, purple, glandular hairs, whereas those of A. dioica do not. Further, the stolon surface in A. marginata is densely woolly, but only pubescent in A. dioica (Bayer and Stebbins 1993). The several thousand kilometer range extension (Aleutian Islands to California) is unusual, although not without precedent in the genus. It was not until the morphological limits were clarified for species of Antennaria from the arctic and Cordilleran regions that several species, including A. alborosea A. E. Porsild, A. alpina (L.) Gaertner, A. aromatic Evert, and A. densifolia A. E. Porsild, were shown to have large range extensions (Malte 1934; Porsild 1950; Porsild and Cody 1980; Evert 1984; Bayer 1989; Chmielewski and Chinnappa 1988; Chmie- lewski et al. 1990; Chmielewski 1993, 1996, 1998). The occurrence of A. dioica in the California moun- tains is suggestive of a more continuous, north- ward, preglacial distribution. The lack of previous reports from continental North America may be the result of factors working independently or in con- cert. First, the species may be restricted to the Si- erra Nevada and represents either a disjunct popu- lation due to recent colonization, or alternatively is a remnant of a once more extensive distribution. Second, inaccessibility to much of the subalpine and alpine habitat in western North America at a reasonable cost has concurrently limited botanical exploration and collection. Because this part of the subalpine and alpine California flora is remote, it is among some of the least collected and explored. Third, to the uninitiated, Antennaria species are of- tentimes difficult to determine, especially when the quality of the specimen is poor, that is, when not in good flowering or fruiting stages. Since neither the first nor second scenarios may be reasonably tested, the third may be at least partially dismissed based on my past experience with the genus. Ad- mittedly, I have not seen all western North Amer- ican collections however. It is my belief that in its entirety the second scenario is the most likely. Antennaria dioica (L.) Gaertner, De Fruct. Sem. PI. 2: 410. 1791. Basionym: Gnaphalium dioicum L. Sp. Pl. 850. 1753. Type locality: Habitat in Eu- ropae apricis aridis. Range extension: California, Inyo Co., Sierra Nevada, near Margaret Lakes, east of Bishop Pass, among granite fell fields and O72 MADRONO scattered Pinus albicaulis, elev. 10,800 ft, 19 Jul, 1962, Betty H. Johnson 692, CAS 898202. ACKNOWLEDGMENTS I thank the curators at CAS for the loan of material on which this report is based. LITERATURE CITED ANDERSON, J. P. 1959. Flora of Alaska and adjacent parts of Canada. The Iowa State University Press, Ames, IA. BAYER, R. J. 1984. Chromosome numbers and taxonomic notes for North American species of Antennaria (As- teraceae: Inuleae). Systematic Botany 9:74-83. BAYER, R. J. 1989. A systematic and phytogeographic study of Antennaria aromatica and A. densifolia (As- teraceae: Inuleae) in the western North American cor- dillera. Madrono 36:248—259. BAYER, R. J. AND G. L. STEBBINS. 1987. Chromosome numbers, patterns of distribution, and apomixis in An- tennaria (Asteraceae: Inuleae). Systematic Botany 123305—319, BAYER, R. J. AND G. L. STEBBINS. 1993. A synopsis with keys for the genus Antennaria (Asteraceae: Inuleae: Gnaphaliinae) of North America. Canadian Journal of Botany 71:1589—1604. CHMIELEWSKI, J. G. 1993. Antennaria pulvinata Greene: The legitimate name for A. aromatica Evert (Aster- aceae: Inuleae) Rhodora 95:261—276. CHMIELEWSKI, J. G. 1996. The dense-leaved pussy’s toes, Antennaria densifolia (Asteraceae: Inuleae): an ad- dition to the vascular flora of British Columbia. Ca- nadian Field-Naturalist 110:314—317. CHMIELEWSKI, J. G. 1997. A taxonomic revision of the Antennaria media (Asteraceae: Inuleae) polyploid species complex in western North America. Brittonia 49:309-327. CHMIELEWSKI, J. G. 1998. Antennaria alpina (Asteraceae: Inuleae): revision of a misunderstood arctic-alpine polyploid species complex. Rhodora 100:39—68. CHMIELEWSKI, J. G. AND C. C. CHINNAPPA. 1988. Range extension of Antennaria aromatica Evert (Asteraceae: Inuleae). SIDA 13:256—257. CHMIELEWSKI, J. G., C. C. CHINNAPPA, AND J. C. SEMPLE. 1990. The genus Antennaria (Asteraceae: Inuleae) in [Vol. 45 western North America: morphometric analysis of Antennaria alborosea, A. corymbosa, A. marginata, A. microphylla, A. parvifolia, A. rosea, and A. um- brinella. Plant Systematics and Evolution 169:151— ie: EverRT, E. EF 1984. A new species of Antennaria (Astera- ceae) from Montana and Wyoming. Madrofo 31: 109-112. Feporov, A. A. 1969. Chromosome numbers of flowering plants. Leningrad, Academy of Science, USSR, V. L. Komarov Botanical Institute. FerRRIS, R. S. 1960. Vol, [V. Bignoniaceae to Compositae. Bignonias to sunflowers. 474—485 in L. Abrams and R. S. Ferris, Illustrated flora of the Pacific states. Washington, Oregon, and California. Stanford Uni- versity Press, Stanford, CA. HuLTEN, E. 1949. Flora of Alaska and Yukon. Lunds Universitets Arsskrift. N.E Avd. 2, 46:1511—1534. HUuLTEN, E. 1968. Flora of Alaska and neighboring terri- tories: a manual of the vascular plants. Stanford Uni- versity Press, Stanford, CA. JEPSON, W. L. 1925. A manual of flowering plants of Cal- ifornia. Associated Students Store, University of Cal- ifornia, Berkeley, CA. MALTE, M. O. 1934. Antennaria of arctic America. Rho- dora 36:101—117. Munz, P. A. 1959. A California flora. University of Cali- fornia Press, Berkeley, CA. PoLuNIN, N. 1959. Circumpolar arctic flora. Oxford Uni- versity Press, Amen House, London. PorsILp, A. E. 1950. The genus Antennaria in northwest- ern Canada. Canadian Field-Naturalist 64:1—25. PorsILb, A. E. AND W. J. Copy. 1980. Vascular plants of continental Northwest Territories, Canada. National Museums of Natural Sciences, National Museums of Canada, Ottawa. STEBBINS, G. L. AND R. J. BAYER. Antennaria. 1993. 196— 198 in The Jepson Manual: Higher plants of Califor- nia, J. C. Hickman (ed.), University of California Press, Berkeley, CA. TuTIN, T. G., V. H. HEywoop, N. A. BurGgs, D. M. Moore, D. H. VALENTINE, S. M. WALTERS, AND D. A. WEBB. 1976. Flora Europaea, Vol. 4. Plantaginaceae to Compositae (and Rubiaceae). Cambridge Univer- sity Press, Cambridge. WELSH, S. L. 1974. Anderson’s flora of Alaska and adja- cent parts of Canada. Brigham Young University Press, Provo, UT. Maprono, Vol. 45, No. 3, p. 273, 1998 REVIEW THE ONCE AND FUTURE FOREST: A GUIDE TO FOREST RESTORATION STRATEGIES. By L. J. Sauer. 1998. Is- land Press, Covelo, CA. 350 p. Hardcover, $50.00. ISBN 155963555255. This book focuses on the restoration of urban forests and parks in the northeastern United States, although concepts covered can be transported to other areas, particularly those related to the social aspects of restoration. The book begins by review- ing important ecological concepts such as succes- sion, fragmentation, and disturbance. Other issues discussed include how wildlife, the hydrologic cy- cle, and landforms/soils have been affected by ur- banization, the section on changes in hydrology is particularly interesting. Issues surrounding the restoration of areas dom- inated by non-native and opportunistic native spe- cies are extensively covered. Restoration is pre- sented as a combination of art and science, and community involvement is paramount. One central idea present throughout the book is the necessity of local community involvement in restoration proj- ects. I have personally witnessed many restoration projects that have failed because the people in- volved never sought local involvement, particularly in urban areas. The book advocates the need for monitoring to accompany all restoration projects and tells us that this is often neglected. I also be- lieve monitoring is critically needed if the science of restoration is to move forward. Watershed or large spatial scales is advocated in restoration projects. It is certainly not possible to include all such lands in all projects but people should consider how lands beyond their local sites could effect the project. The book tells us that there is no “‘step-by-step’’ method in restoration, rather a long-term effort is necessary requiring a high de- gree of expertise and commitment. The book also challenges the idea of ‘‘wilderness’’ and I believe correctly states that many landscapes have been af- fected by the practices of indigenous peoples for long periods. The book states that the most important condi- tions for a successful restoration project are for it to be community- and science-based. To be com- munity-based, it must represent a consensus, which in turn requires that it be participatory. Science- based restoration projects must be documented and monitored. The book advocates the best way to convey real information to a community is to have them gather the information themselves, through monitoring both before and during restoration. The later chapters provide specific techniques that can be used in restoration projects. This section of the book can be thought of as a practical resto- ration guide, the sections on building soil systems and restoring natural water systems are very good. Specific field examples are primarily given from the northeastern United States, particularly Central Park in New York. This section could have been improved by including more examples from other areas in the United States (southeast, midwest, west) but compromises must be made to keep a book manageable, the author also writes from her strengths in the northeastern United States. This book could be improved by including additional references; in some cases, only one view is pre- sented on topics that currently have diverse inter- pretations. I recommend this book to people interested in developing community-based restoration projects. As populations grow the ecosystem effects of ur- banization will increase and this book gives infor- mation on how to manage and restore these areas. —Scott Stephens, Natural Resources Manage- ment Department, Cal Poly State University, San Luis Obispo, CA 93407. Volume 45, Number 3, pages 187—274, published 22 June 1999 SUBSCRIPTIONS—-MEMBERSHIP Membership in the California Botanical Society is open to individuals ($27 per year; family $30 per year; emeritus $17 per year; students $17 per year for a maximum of 7 years). Late fees may be assessed. Members of the Society receive Maprono free. Institutional subscriptions to MADRONO are available ($60). Membership is based on a calen- dar year only. Life memberships are $540. Applications for membership (including dues), orders for subscriptions, and renewal payments should be sent to the Treasurer. Requests and rates for back issues, changes of address, and undelivered copies of MapRONO should be sent to the Corresponding Secretary. INFORMATION FOR CONTRIBUTORS Manuscripts submitted for publication in MADRONO should be sent to the editor. It is preferred that all authors be members of the California Botanical Society. Manuscripts by authors having outstanding page charges will not be sent for review. Manuscripts may be submitted in English or Spanish. English-language manuscripts dealing with taxa or topics of Latin America and Spanish-language manuscripts must have a Spanish RESUMEN and an English ABSTRACT. Manuscripts and review copies of illustrations must be submitted in triplicate for all articles and short items (NOTES, NOTEWORTHY COLLECTIONS, POINTS OF VIEW, etc.). Follow the format used in recent issues for the type of item submitted. Allow ample margins all around. Manuscripts MUST BE DOUBLE-SPACED THROUGHOUT. For articles this includes title (all caps, centered), author names (all caps, centered), addresses (caps and lower case, centered), abstract and resumen, 5 key words or phrases, text, acknowledgments, literature cited, tables (caption on same page), and figure captions (grouped as consecutive paragraphs on one page). Order parts in the sequence listed, ending with figures. Each page should have a running header that includes the name(s) of the author(s), a shortened title, and the page number. Do not use a separate cover page or ‘erasable’ paper. Avoid footnotes except to indicate address.changes. Abbreviations should be used sparingly and only standard abbrevia- tions will be accepted. Table and figure captions should contain all information relevant to information presented. All measurements and elevations should be in metric units, except specimen citations, which may include English or metric measurements. Authors of accepted papers will be asked to submit an electronic version of the manuscript. Microsoft Word 6.0 or WordPerfect 6.0 for Windows is the preferred software. Line copy illustrations should be clean and legible, proportioned to the MApDRONO page. Scales should be in- cluded in figures, as should explanation of symbols, including graph coordinates. Symbols smaller than | mm after reduction are not acceptable. Maps must include a scale and latitude and longitude or UTM references. In no case should original illustrations be sent prior to the acceptance of a manuscript. [Illustrations should be sent flat. No illustrations larger than 27 X 43 cm will be accepted. Presentation of nomenclatural matter (accepted names, synonyms, typification) should follow the format used by Sivinski, Robert C., in Maprono 41(4), 1994. Institutional abbreviations in specimen citations should follow Holmgren, Keuken, and Schofield, Index Herbariorum, 8th ed. Names of authors of scientific names should be abbreviated according to Brummitt and Powell, Authors of Plant Names (1992) and, if not included in this index, spelled out in full. Titles of all periodicals, serials, and books should be given in full. Books should include the place and date of publication, publisher, and edition, if other than the first. All members of the California Botanical Society are allotted 5 free pages per volume in MAprono. Joint authors may split the full page number. Beyond that number of pages a required editorial fee of $40 per page will be assessed. The purpose of this fee is not to pay directly for the costs of publishing any particular paper, but rather to allow the Society to continue publishing MApRONo on a reasonable schedule, with equity among all members for access to its pages. Printer’s fees for illustrations and typographically difficult material @ $35 per page (if their sum exceeds 30 percent of the paper) and for author’s changes after typesetting @ $4.50 per line will be charged to authors. At the time of submission, authors must provide information describing the extent to which data in the manu- script have been used in other papers that are published, in press, submitted, or soon to be submitted elsewhere. r Sie oy : arn my AN q ‘CONTENTS NOTEWORTHY COLLECTIONS VOLUME 45, NUMBER 4 OCTOBER-—DECEMBER 1998 MADRONO A WEST AMERICAN JOURNAL OF BOTANY QUERCUS GARRYANA HOoK. (FAGACEAE) STAND STRUCTURE IN AREAS WITH DIFFERENT GRAZING HISTORIES Randall D. Jackson, Kenneth O. Fulgham, and Barbara Allen-Diaz...... 215 RELATIVE CONTRIBUTION OF BREEDING SYSTEM AND ENDEMISM TO GENOTYPIC DIVERSITY: THE OUTCROSSING ENDEMIC TARAXACUM CALIFORNICUM VS. THE WIDESPREAD APOMICT 7: OFFICINALE (SENSU LATO) Jennifer C. Lyman and Norman C. ELIStrand .......ccccssssccccccccccceeeeeeeeeeeeee 283 EFFECTS OF SIMULATED OIL FIELD DISTURBANCE AND TOPSOIL SALVAGE ON ERIASTRUM HOOVERI (POLEMONIACEAE) Jay M. Hinshaw, Gary L. Holmstead, Brian L. Cypher, and David C. ATG OTS ON Sukisececteti suas ethea tad ventas te A 290 REESTABLISHMENT OF ERIASTRUM HOOVERI (POLEMONIACEAE) FOLLOWING OIL FIELD DISTURBANCE ACTIVITIES Gary L. Holmstead and David C. AndePrson ..........ccccccceeeeeeeeestttntneeeeeeeeees 295 Coast LIVE Oak REVEGETATION ON THE CENTRAL COAST OF CALIFORNIA Anuja Parikh and Nathan Gale occ. 0ticcsceccoavestetceelectecsssseesequsssnsnnnnescsecesecs 301 ALPINE VASCULAR FLORA OF HASLEY BASIN, ELK MounrtTaINs, COLORADO, USA Randy V. Seagrist and Kevin J. TQylOY .....cccccccccseessessssssccccccccceeeeeeeeeeeeeeeees 310 ALPINE VASCULAR FLORA OF BUFFALO PEAKS, Mosquito RANGE, COLORADO, USA Randy. Seagrist and: Kevitt J: AGylor ic.0 och yieate cactveiecccecctenctcnssseaecoess 319 CATTIFOR NIA ove Ree Av oe EEGs, Ss Dass Mul sooo ence cacao wag eeeseteroeeeeee 326 IMD ON TA NAG eee A ss Sed 4 es tice Ga pS sea ON Nc ensune ete dcue en 328 BRIFISHEC ORUMBIA issn icccsc ce Abasuilooses eacesex cs eaceus bah ose voess Mulgan usein Sete dt an we adtacuspueneaueees 330 PRESIDENT’S REPORT FOR VOLUME 45S .0.0..........cccccssssssscescossssecescoeessrccenees 331 EDITOR’S REPORT FOR VOLUMEAS: 32 seiiiisccccesecccee SEU i vesvecceccdatesssaseeoes SEV REVIEWERS OF MANUSCRIPTS oe si se races cvecetesdhccsnastenciactensieeeees 333 INDEX TO VOI UIVUE ooh nection ced SEI. ae ave oad eens nd ons an vei ewer eee 334 DATES OF PUBLICATION «xc.ccetnniniau Weta 336 PDE DG AMON ec eaee haea esate ccc reins cer sc occowe Sesten soso Saat neuen eeee finer aeons ene il TABLE OF CONTENTS FOR VOLUME 45 200.0... ccccccccccsssseceeceeeseececeeaneeees 1V PUBLISHED QUARTERLY BY THE CALIFORNIA BOTANICAL SOCIETY Maprono (ISSN 0024-9637) is published quarterly by the California Botanical Society, Inc., and is issued from the office of the Society, Herbaria, Life Sciences Building, University of California, Berkeley, CA 94720. Subscription information on inside back cover. Established 1916. Periodicals postage paid at Berkeley, CA, and additional mail- ing offices. Return requested. PosTMASTER: Send address changes to MADRONO, ‘/ Mary Butterwick, Botany De- partment, California Academy of Sciences, Golden Gate Park, San Francisco, CA 94118. Editor—KnrisTINA A. SCHIERENBECK California State University, Chico Department of Biology Chico, CA 95427-0515 kschierenbeck @csuchico.edu Editorial Assistant—Davw T. PARKS Book Editor—Jon E. KEELEY Noteworthy Collections Editors—DIkETER WILKEN, MARGRIET WETHERWAX Board of Editors Class of: 1998—FREDERICK ZECHMAN, California State University, Fresno, CA Jon E. KeELey, U.S. Geological Service, Biological Resources Division, Three Rivers, CA 1999—Timotny K. Lowrey, University of New Mexico, Albuquerque, NM J. Mark Porter, Rancho Santa Ana Botanic Garden, Claremont, CA 2000—Pame a S. Sottis, Washington State University, Pullman, WA JOHN CALLAway, San Diego State University, San Diego, CA 2001—Robsert PATTERSON, San Francisco State University, San Francisco, CA PaAuLa M. ScuIFFMAN, California State University, Northridge, CA 2002—NorMAN ELLSTRAND, University of California, Riverside, CA Cara M. D’ Antonio, University of California, Berkeley, CA CALIFORNIA BOTANICAL SOCIETY, INC. OFFICERS FOR 1998-1999 President: R. JoHN LitTLe, Sycamore Environmental Consultants, 6355 Riverside Blvd., Suite C, Sacramento, CA 95831 First Vice President: Susan D’ ALcAmo, Jepson Herbarium, University of California, Berkeley, CA 94720 Second Vice President: David Ket, California Polytechnic State University, Biological Sciences Department, San Luis Obispo, CA 93407 Recording Secretary: ROXANNE BITTMAN, California Department of Fish and Game, Sacramento, CA 95814 Corresponding Secretary: | SUSAN BAINBRIDGE, Jepson Herbarium, University of California, Berkeley, CA 94720. suebain @SSCL.berkeley.edu Treasurer: Mary Butterwick, Botany Department, California Academy of Science, Golden Gate Park, San Fran- cisco, CA 94118. butterwick.mary @epa.gov The Council of the California Botanical Society comprises the officers listed above plus the immediate Past President, WAYNE R. FERREN, JR., Herbarium, University of California, Santa Barbara, CA 93106; the Editor of Maprono; three elected Council Members: MARGRIET WETHERWAX, Jepson Herbarium, University of California, Berkeley, CA 94720; JAMEs SHEvock, USDI National Park Service, Pacific West Region, 600 Harrison Street, Suite 600, San Francisco, CA 94107; DIANE Exa, U.S. Fish and Wildlife Service, 3310 El Camino Avenue, Sacramento, CA 95825; Graduate Student Representative: DENNIS P. WALL, Jepson Herbarium, University of California, Berke- ley, CA 94720. This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Maprono, Vol. 45, No. 4, pp. 275—282, 1998 QUERCUS GARRYANA HOOK. (FAGACEAE) STAND STRUCTURE IN AREAS WITH DIFFERENT GRAZING HISTORIES RANDALL D. JACKSON Ecosystem Sciences Division, 151 Hilgard Hall #3110, Universi Berkeley, CA 94720 rjackson @ nature.berkeley.edu KENNETH O. FULGHAM Department of Rangeland Resources & Wildland Soi Humboldt State University, Arcata, CA 95521 fulghamk @ axe.humboldt.edu BARBARA ALLEN-DIAZ Ecosystem Sciences Division, 151 Hilgard Hall #3110, University of California Berkeley, CA 94720 ballen @nature.berkeley.edu ABSTRACT We estimated seedling, sapling, and mature tree densities for two historically defined grazed-classes (Low and HI) of the Mad River Ranger District, Six Rivers National Forest, CA. Using Quercus garryana Hook. vegetation coverages on a geographic information system and a variable probability sampling scheme, 9 oak stands were randomly selected from each grazed-class. We employed a second sampling stage by selecting 3 simple random samples of 100-m? quadrants from each of the 18 oak stands selected in the first sampling stage. We found greater seedling and mature tree densities for the HI grazed-class and greater sapling densities for the Low grazed-class. Sapling densities were roughly double and seedling densities about 5 times mature tree densities regardless of grazed-class. We suggest that increased grazing intensity creates favorable environments for seedling survival, but may ultimately reduce the number of seedlings transitioning to the sapling size-class. Our results showing roughly 2:1 sapling to mature tree ratios indicate that Q. garryana regeneration is occurring on these rangelands. The historical clearing of oak woodlands to max- imize forage yield for livestock grazing or to con- vert to agricultural production has been well doc- umented (Allen-Diaz and Holzman 1991; Bartolo- me et al. 1986; Bartolome and Standiford 1992; Muick and Bartolome 1987). Along with suburban encroachment, these land-use practices have been implicated as factors in the postulated decline of Quercus douglasii Hook. & Arn (blue oak), Q. lob- ata Nee (valley oak), and Q. engelmannii E. Greene Engelmann oak regeneration (Griffin et al. 1987; Muick and Bartolome 1985; Pavlik et al. 1991). Regeneration problems manifested as bimodal age structure biased towards seedlings and mature trees with a lack of sapling age-classes have been cited for each of the above species (Bartolome et al. 1986; Griffin 1971; Griffin 1976; Lathrop et al. 1990). Quercus garryana Hook. has the longest north- south distribution among western Quercus spp. The northern range of Q. garryana extends onto Van- couver Island in Canada at 50°N latitude and the southern range runs into Los Angeles County at 34°N latitude (Stein 1990). The coastal range of Q. garryana terminates in Marin County, but it is found further south along the western slope of the Sierra Nevada in a limited and disjunct distribution. The relationships between Northcoast Q. garry- ana populations and livestock grazing, land-use practices, and suburban encroachment have re- ceived limited attention (Griffin et al. 1987; Muick and Bartolome 1985; Reed and Sugihara 1986). Su- gihara et al. (1987) defined plant community types for Northcoast Q. garryana-dominated landscapes and Saenz and Sawyer (1986) studied grazing ef- fects on Q. garryana understory composition. Lit- erature treating potential domestic livestock grazing effects on Q. garryana is limited to a statewide sur- vey conducted by Muick and Bartolome (1987) and modeling efforts by Anderson and Pasquinelli (1984). Muick and Bartolome (1987) found no sig- nificant Q. garryana distribution patterns related to livestock grazing on the Northcoast. They assessed the adequacy of the existing Q. garryana sapling populations to replace trees lost through mortality by examining sapling to mature tree ratios on 13 Northcoast plots. Sapling to adult tree ratios were less than 1:1 for Q. garryana in their study, yet 75% of all individuals sampled were seedlings in- dicating no lack of seedling establishment, but a failure of seedlings to survive the transition to sap- ling size-classes. Bolsinger (1988) assessed the age structure of Q. garryana by drawing inferences from seedling and 276 Six Rivers National Forest | | Eureka Grazing allotments (not to scale) Los Angeles YY Mad River Ranger District Fic. 1. Study sites located within Mad River Ranger Dis- trict, Six Rivers National Forest, Humboldt County, CA. MADR 45 (4) Art#Ol 46 7% JACKSON Figt @1 sapling counts from systematic plots established across California. He found 45% of Q. garryana- type plots contained no seedlings and 38% con- tained no saplings suggesting the potential for in- sufficient replacement populations or that present spatial distributions may be shifting. Anderson and Pasquinelli (1984) predicted declining Q. garryana distribution under “‘present trend”’ conditions. Their model parameterized livestock grazing, wildlife pressure, fire frequency, and weather conditions specific to the Northcoast’s Sonoma and Mendoci- no Counties. We designed this study to describe Q. garryana stand structure and to examine livestock grazing in- tensity effects on this structure. Specifically, we sought to determine whether differentially grazed stands contained significantly different densities of seedlings, saplings and/or mature trees. This infor- mation should provide insight about the regenera- tion status of Q. garryana on California’s North- coast. METHODS Study site. We conducted this study June and July 1995 on the Mad River Ranger District (MRRD), Six Rivers National Forest (SRNF), and Humboldt County, CA (Fig. 1). SRNF les within sections of Humboldt, Del Norte, Siskiyou, and Trinity Coun- ties in the northern California coast range. MRRD is typified by a Mediterranean climate of cool, moist winters and warm, dry summers. Average January temperatures range from —2 to 4°C (28 to 40°F) and average July temperatures range from 17 to 27°C (64 to 80°F) (Oakeshott 1978). Annual pre- cipitation ranges from 127 to 152 cm (50 to 60 in). MADRONO [Vol. 45 TABLE 1. SUMMARY OF HISTORICAL GRAZING ALLOTMENT DATA USED FOR ASSIGNING SIX GRAZING ALLOTMENTS TO ONE OF TWO GRAZED-CLASSES. 'AU = Animal unit = 1 individual. Mean annual Oak Grazed- Allotment ‘AU Oak ha ha-AU™! class Long Ridge 88 1859 | LOW Norris-Green 108 1839 17 LOW Barry Creek 109 1484 14 LOW Soldier Creek 66 616 9 HI Buck Mountain 9] 766 8 HI Van Duzen 166 1166 e HI Eighty percent of this precipitation occurs between November and April. Miles (1993) described MRRD soils as derived from the Franciscan Me- lange Complex with Quercus garryana stands oc- curring predominantly on the Oxalis-Hecker-Doty association. These are well-drained, pale brown loams with moderate to strong blocky structure and a slightly acidic nature. Grazed-classes. As an initial grazing allotment screening, we asked MRRD range personnel to identify 3 MRRD grazing allotments with high his- torical grazing intensity and 3 with relatively low historical grazing intensity. These allotments were to represent each of two grazed-classes (LOW or HI; Table 1). MRRD range personnel stated that each of the 3 Low grazed-class allotments were histori- cally grazed very lightly, and had no permitted grazing activity for 5 to 10 years prior to this study. MRRD range personnel also confirmed that histor- ical permitted use (1950 to 1981) continued on the 3 HI grazed-class allotments after 1981. We derived our grazed-classes by defining graz- ing intensities that were based on allotment-wise Q. garryana hectares (estimated using 1990’s GIS technology) because we believed that the majority of available forage, hence grazing pressure, oc- curred on the oak woodland/savanna vegetation type. We used annual animal units records (1950 to 1981) and total Q. garryana hectare estimates for each allotment to quantify historical grazing inten- sities. Actual allotment-wise stocking rates were set by MRRD range managers using their assessments of suitable grazing hectares. Suitable hectares were comprised of oak savanna; coniferous forest; and small, open grassland areas known as “‘glades’’. However, Sawyer et al. (1977) described the un- derstory layer of the Northcoast Douglas-fir—tan- oak type (Pseudotsuga menziesii (Mirbel) Franco— Lithocarpus densiflorus (Hook. & Arn.) Rehder) as sparse, less developed, and dominated by shrubby L. densiflorus while others have confirmed this type of coniferous forest provides much less herbaceous productivity and available forage than oak wood- lands (Sharrow and Leininger 1982). While glades provide excellent forage, the fraction of the allot- 1998] ments in this type is very small. Therefore, we as- sumed that oak woodlands provide the bulk of for- age for grazing animals and only used Q. garryana hectares in grazing intensity quantification. Sampling. Assuming that the total number of seedlings, saplings, and mature trees were a func- tion of the area of any oak stand, we employed a 2-stage sampling design consisting of a variable probability selection scheme as the 1*-stage and simple random sampling (SRS) as the 2"-stage. Nine Q. garryana stands were selected from each of 2 grazed-class subsets with probability propor- tional to the area of the remaining Q. garryana stands within that particular grazed-class. Hence, larger Q. garryana stands had higher selection probabilities. We incorporated these probabilities, known as 1|*-order’ inclusion probabilities (Overton and Stehman 1995), into Horvitz-Thompson (HT) estimation of the size-class densities for each grazed-class (Horvitz and Thompson 1952). We calculated sample-based sampling variance esti- mates for the two-stage HT estimator using the Sen- Yates-Grundy (SYG) method (Yates and Grun- dy 1953). This method requires that the 2"¢-order inclusion probabilities for the selected 1*'-stage units (Q. garryana stands) be known. Second-order inclusion probabilities are the probabilities that any 2 particular Q. garryana stands were both chosen during the selection process. We estimated 2"¢-order inclusion probabilities using the list sequential method outlined by Sunter (1977). The 2" sam- pling stage entailed SRS of three 100-m? quadrats from each unit (18 oak polygons) selected in the 1“ stage. We then scaled seedling, sapling, and mature tree estimates from each grazed-class, deriving mean values (per 100 m/’) for standardization and comparison ease. Ninety-five percent confidence in- tervals were constructed around each size-class density estimate for statistical comparisons. We obtained Universal Transverse Mercator (UTM) coordinates for each point selected during the 2"'-stage of sampling from a geographic infor- mation system database. We arbitrarily decided that these UTM coordinates would serve as the south- eastern corner of each 100-m’ sampling quadrat. Quadrats were located in the field with a handheld geographic positioning system. We delineated boundaries of each quadrat with a handheld com- pass and cloth tape. Seedlings (<1 cm basal di- ameter), saplings (<10 cm and >1 cm basal di- ameter), and mature trees (>10 cm basal diameter) were enumerated for each quadrat. Definition of size-classes followed Muick and Bartolome (1987). We did not age individual seedlings, saplings, or mature trees. We estimated several site variables at each quad- rat to determine potential environmental disparities among the two grazed-classes. These variables in- cluded 1) slope, 2) azimuth, 3) canopy cover, and, 4) dominant herbaceous understory type. The sine JACKSON ET AL.: QUERCUS GARRYANA STAND STRUCTURE Pie of slope in degrees and the cosine of azimuth in degrees from N were multiplied to create a solar insolation indicator variable northness (sensu Borchert et al. 1989) which potentially varied from +100 to —100. Greater positive values of this vari- able indicate a N-facing steep slope while higher negative values indicate a S-facing steep slope. We ocularly estimated canopy cover at each site. We characterized the herbaceous understory type at each quadrat by making 6 randomly located her- baceous species cover estimates within the quadrat. Each estimate was made with a 10-point frame us- ing the first foliar intercept criterion for each pin placement (Heady and Rader 1958). We then clas- sified species into one of the following functional groups: 1) annual grass, 2) annual forb, 3) perennial grass, 4) perennial forb, or 5) woody perennial. Bare ground and dry organic matter 10-point frame hits were also recorded and categorized. We estimated edaphic characteristics at a ran- domly chosen | of 3 quadrats at each of the 18 Q. garryana stands. Soil pits were excavated for esti- mation of the following variables: 1) solum depth, 2) rooting depth, 3) clay percent (by feel), 4) rock fragment percent (2-mm mesh sieve), 5) pH (LaMotte colorimetric method), 6) moist Munsell color value, 7) textural-class (by feel), and 8) avail- able water-holding capacity (USDA 1993). Para- metric tests were not performed to assess signifi- cant differences for these variables between grazed- classes, nor to correlate them to Q. garryana size- class estimates because they were collected with the variable probability sampling scheme described above. By definition, unequal probability samples do not meet the parametric test requirement of in- dependent, identical distributions between groups (Sarndal et al. 1992). Therefore, we present contin- uous data in tabular form (Table 2) and discrete data in graphical (Fig. 2) form for grazed-class comparisons. Where tabular and graphical exami- nation suggested differences between the grazed- classes in a given variable, we tested joint bivariate distributions with a two-dimensional Kolmogorov- Smirnov (2DKS) nonparametric procedure (Garvey et al. 1998). Two-DKS uses a bootstrapping tech- nique (500 Monte Carlo simulations) to iteratively compare randomly derived expectation matrices with an observed joint-distribution matrix. This technique is effective at assessing both linear and non-linear relationships among bivariate distribu- tions and makes no assumptions about functional responses. RESULTS Comparison of 95% confidence intervals showed that seedling and mature tree densities were greater in the HI grazed-class while sapling densities were greater in the LOW grazed-class (Fig. 3). Mean es- timates are reported + the 95% confidence interval and statistical significance determined by overlap or 218 MADRONO [Vol. 45 TABLE 2. SAMPLE SIZE (N), MEANS, AND STANDARD ERRORS (SE) OF CONTINUOUS ENVIRONMENTAL VARIABLES FOR EACH GRAZED-CLASS. ! Units defined in text. Variable n Canopy cover (%) 27 Slope (degrees) P| 'Northness 2] 2.6 AWC 2 4.0 pH 9 6.2 Solum depth (cm) 9 Rooting depth (cm) 9 Clay (%) 9 Rock fragment (%) 9 Moist Munsell color value* 9 not. The seedling density estimate for the HI grazed- class was 33.5 + 3.3 seedlings per 100 m? com- pared to 19.1 + 1.2 seedlings per 100 m/? for the LOW grazed-class. The mature size-class followed the same trend in that the HI mature tree estimate was greater than the LOW mature tree estimate (HI = 5.0 + 0.5, Low = 3.9 + 0.4). Sapling size-class density estimates were the opposite. There were on average more saplings per 100 m* in the Low grazed-class (10.8 + 0.3) than in the HI grazed-class (9.3 = 1.1; Pig. 3). The greatest disparity in environmental variables between the two grazed-classes was average slope (HI = 17.5 + 1.4, Low = 26.9 + 1.4; Table 2). Other continuous variables appeared quite similar SO were not analyzed further than reporting means and standard errors. Examination of the dominant herbaceous understory type standard errors (Fig. 2) indicated similar distributions of this categorical variable between grazed-classes. Soil textural-class differences were evident; all 9 Low grazed-class sites were classified as loams while 5 of 9 HI grazed-class soils were determined to be loams with 50.0 45.0 40.0 35.0 30.0 25.0 20.0 15.0 10.0 5.0 0.0 Mean % cover Annual forb Annual grass Dry organic matter Perennial HI LOW SE Mean SE 4.5 63.7 4.4 1.4 26.9 1.4 4.9 aA 4.6 0.5 a3 0.3 0.1 6.2 0.1 4.1 118.4 3.1 2.0 TOD 0.7 1.1 17.4 0.5 3.6 42.7 1.1 0.1 3.5 0.1 1 clay loam and 3 sandy loams. Only 1 soil pit per site was dug, therefore no dispersion statistics were estimable from these data. Two-DKS results indicated that the joint distri- butions between slope and seedling density, slope and sapling density, and slope and mature tree den- sity (Fig. 4) were not significantly different from random (P = 0.60, 0.53, and 0.50; respectively). DISCUSSION Greater seedling densities coupled with lower sapling densities in areas with greater grazing in- tensities suggest that herbivory of surrounding veg- etation may positively influence Q. garryana seed- ling survival and recruitment, but that incidental trampling and herbivory of the seedlings may dis- courage transition to the sapling stage. While this needs experimentation, there are experimental re- sults for Q. douglasii which corroborate our infer- ences and may help to explain our observations. Quercus douglasii and Q. garryana are taxonomi- cally and ecologically similar species. Both belong Grazed-class LOW HI Bare ground Perennial forb Woody grass perennial Functional group Fic. 2. frame estimates. Functional group cover means and standard errors (n = 9 for each grazed-class) as determined by 10-point 1998] 40.0 35.0 + a 30.0 | S| S = 25.0 o Q. = 20.0 | £ 5 10.0 | 5.0 4 0.0 - Seedling LOW-grazed 19.1 @ HI-grazed 33.5 Fic. 3. to the white oak subgenus Lepidobalanus and oc- cupy apparently similar ecological sites—shallow, rocky hill slopes—in distinctive Californian cli- matic regions (Rundel 1979; Rundel 1986; Stein 1990). Additionally, these species freely hybridize along climatic transition zones (Tucker 1979). In greenhouse experiments, Gordon et al. (1989) found greater Q. douglasii seedling emergence and growth responses in annual forb (Erodium spp.) seeded plots compared to annual grass (Bromus diandrus L.) seeded plots. They attributed this to increased rates of water stress in the annual grass plots. This illustrates that competition between Q. douglasii seedlings and neighboring herbaceous vegetation does occur, however, Gordon et al. (1989) obtained their results absent defoliation. Welker and Menke (1990) reported that defoliation of annual vegetation surrounding growing Q. doug- lasit seedlings was beneficial in reducing evapo- transpiration of competing vegetation, thereby re- ducing the rate at which the Q. douglasii seedlings ~ > 5 ° n c re) v ) se ae (=) e Bn As & ae 4 se) Q, g g 2 n up a [es 10 20 30 40 10 Slope (%) Fic. 4. JACKSON ET AL.: QUERCUS GARRYANA STAND STRUCTURE Slope (3%) 219 Sapling Mature tree 10.8 3.9 9.3 5.0 Seedling, sapling, and mature tree density means and 95% confidence intervals for each grazed-class. encountered water stress. These results offer poten- tial explanations for our findings, whereby herbiv- ory is releasing Q. douglasii seedlings from re- source competition with surrounding vegetation. Resources being competed for are probably water, nutrients, and space. However, this phenomenon is likely overwhelmed at higher grazing intensities where grazing selectivity lessens and seedlings are depredated more or less incidentally. Hall et al. (1990) showed that at very high stock- ing rates, where animals are actually competing for space as well as forage, increased seedling depre- dation rates occurred. However, with reduced stocking rates, Hall et al. (1990) found little to no livestock preference for Q. douglasii seedlings. Given their similarities, these results likely hold for Q. garryana. Seedling depredation is probably a function of forage production and availability, ro- dent populations (Davis et al. 1990), and stocking rate. Experimental evidence for Q. garryana is needed. als) 20 Mature tree density 10 20 30 40 a0) 20 30 40 Slope (%) Joint bivariate distributions of slope and seedling density, sapling density, and mature tree density. 280 We did not measure microtopographic relief or estimate soil disturbance, but these factors may be important reasons for higher seedling densities. Be- cause livestock grazing creates soil disturbance (Kauffman and Krueger 1984), areas of higher grazing intensity tend to have greater micro-relief and therefore a higher probability of radicle inter- ception and penetration with the soil (Watt 1919). During autumn months, Q. garryana dormancy decreases interspecific competition for light and water between annual grasses and Q. garryana seedlings (Hibbs and Yoder 1993). This reduction in photosynthesis and growth allows seedling un- derground biomass persistence and promotes car- bohydrate storage in root systems (Hibbs and Yoder 1993) while the annual grasses dominate available resources. Domestic stock grazing in the spring re- sults in annual grass defoliation as well as the in- cidental Q. garryana seedling defoliation and tram- pling. However, Hibbs, and Yoder (1993) have shown that Q. garryana is a prolific and adventi- tious sprouter. They found 20-year-old Q. garryana seedling roots with aboveground shoots less than 3 years in age. Hence, should the Q. garryana seed- ling survive defoliation and water stress until an- nual grasses have been grazed or completed their life cycle, the Q. garryana seedling taproot and fine root system can eventually access resources at low- er soil depths where annual grasses are not com- peting. Indeed, we usually found a dense network of fine Q. garryana roots dispersed throughout the solum to a depth of 75+ cm (Table 2), while annual grass roots were not observed below the surface 10- cm. Seedlings able to establish these deeper roots should increase their survival chances via resource partitioning (sensu Brown 1998). However, re- source partitioning is likely not achieved until Quercus seedlings have successfully negotiated in- terspecific resource competition mediated by cattle dietary preferences for herbaceous grasses and forbs. Coupled with an ability to survive many sea- sons of foliage removal via sprouting, resource par- titioning may account for greater Quercus seedling densities in areas with higher grazing intensities. Litter decomposition rate is another potentially important factor that may be differentially treated across grazing regimes. Desiccation is a major cause of germination failure in California’s pre- dominantly dry oak woodlands (Borchert et al. 1989). However, the consistently moist, but cool Northcoast environs are more likely to result in low decomposition rates leading to higher litter accu- mulation with a concomitant increase in acorn rot. Opening of the herbaceous layer to increased solar insolation via phytomass removal by livestock may reduce the potential for acorn rot, thereby increas- ing germination and seedling success. These hy- potheses all need in situ experimental testing in Q. garryana stands. Significantly greater mature tree densities in the HI grazed-class may be the best explanation for both MADRONO [Vol. 45 greater seedling densities and lesser sapling densi- ties in this group. It follows that more trees lead to more acorns leading to more seedlings. Transition from seedling to sapling is probably inhibited by lack of sufficient light and other resources due to crowding by mature trees combined with seedling defoliation and depredation by livestock. Allen- Diaz and Bartolome (1990) concluded that while recruitment of Q. douglasii seedlings was frequent, none of their seedlings made the transition into sap- ling size-classes. Why was mature tree density greater in the HI grazed-class? The HI-grazed allotment group was generally found on gentler slopes that may have had some indirect effects on sapling to mature tree transition. For example, fires are known to burn less intensively on gentler slopes (Albini 1976) which might have provided a more favorable environment for tree development. Alternatively, deeper soils on sites with gentler slopes may have provided a mi- croclimate more conducive to sapling to tree tran- sition. However, our soil data revealed no evidence for this. Examination of maps and field observation revealed no obvious geomorphological or topo- graphical patterns that might account for these dif- ferences. Aside from the grazed-class comparisons of size- class densities, our results indicate that natural Q. garryana regeneration on these MRRD sites is not at risk. Representing roughly double the mature tree stock, densities of about 10 saplings per 100 m/? seems a surplus. Should one tree from a well- stocked stand die at any given time, several sap- lings should be available for replacement. To truly assess regeneration status, mortality and survival rates between each size-class must be known. Our single season study did not permit estimation of mortality rates even for seedlings. Our point-in- time estimates of age structure do provide compel- ling evidence that regeneration of this species is occurring. Annual assessment of seedling mortality rates coupled with creative techniques for estimat- ing sapling to tree transition probabilities and tree mortality rates are needed to verify our results and quantify the regeneration status of this resource. CONCLUSIONS Higher seedling but lower sapling densities with increased grazing intensity indicates that grazing at higher stocking rates does not affect Q. garryana seedling recruitment. There seems to be plenty of seedlings beneath and around tree canopies, but seedlings in the higher grazed areas are making the transition to the sapling size-class in lesser numbers than those in the more lightly grazed areas. How- ever, our results indicate that Q. garryana regen- eration is potentially occurring regardless of graz- ing pressures as evidenced by the roughly 2:1 sap- ling to tree ratio estimates found for both grazed classes. 1998] ACKNOWLEDGMENTS Comments by Ayn Shlisky, James Bartolome, and three anonymous reviewers greatly improved this manuscript. Data acquisition assistance from Six Rivers National For- est personnel is greatly appreciated. Thanks to the UC Riverside Baseball Team and Bob and Carmen Reilly for vehicular support. LITERATURE CITED ALBINI, E A. 1976. Estimating wildfire behavior and ef- fects. USDA Forest Service, Intermountain Forest and Range Experiment Station. GTR-INT-30. ALLEN-DIAZ, B. AND J. W. BARTOLOME. 1990. Survival of Quercus douglasii (Fagaceae) seedlings under the in- fluence of fire and grazing. Madrono 39:37-53. ALLEN-DIAZ, B. H. AND B. HOLZMAN. 1991. Blue oak com- munities in California. Madrono 38:80—95. ANDERSON, M. V. AND R. L. PASQUINELLI. 1984. Ecology and management of the northern oak woodland com- munity, Sonoma County, California. MA thesis. Son- oma State University. BARTOLOME, J. W., P. C. Muick, AND M. P. MCCLARAN. 1986. Natural regeneration of California hardwoods. In T. R. Plumb and N. H. Pillsbury (eds.), Proceed- ings of the symposium on multiple-use management of California’s hardwood resources, Nov. 12-14, 1986; San Luis Obispo, CA. Pacific Southwest Forest and Range Research Station. BARTOLOME, J. W. AND R. B. STANDIFORD. 1992. Ecology and management of Californian oak woodlands. Jn P. FE Ffolliott et al. (eds.), Ecology and management of oak and associated woodlands: perspectives in the southwestern United States and northern Mexico, April 27—30, 1992; Sierra Vista, AZ. USDA Forest Service GTR RM-218. BOLSINGER, C. 1988. The hardwoods of California’s tim- berlands, woodlands, and savannas. USDA Forest Service Resource Bulletin. PNW-RB-148. BORCHERT, M. I., F W. Davis, J. MICHAELSEN, AND L. D. OYLER. 1989. Interactions of factors affecting seed- ling recruitment of blue oak (Quercus douglasii) in California. Ecology 70:389—404. Brown, C. S. 1998. Restoration of California Central Val- ley grasslands: applied and theoretical approaches to understanding interactions among prairie species. Ph.D. dissertation. University of California, Davis, CA. Davis, E W., M. BORCHERT, L. E. HARVEY, AND J. C. MI- CHAELSEN. 1990. Factors affecting seedling survivor- ship of blue oak (Quercus douglasii H. & A.) in Cen- tral California. Jn R. Standiford (ed), Proceedings of the Symposium on oak woodlands and hardwood rangeland management, Oct. 31—Nov. 2, 1990, Davis, CA. Pacific Southwest Forest and Range Experiment Station, Berkeley, CA. Garvey, J. E., E. A. MARSCHALL, AND R. A. WRIGHT. 1998. From star charts to stoneflies: Detecting rela- tionships in continuous bivariate data. Ecology 79: 442-447. Gorpon, D., J. WELKER, J. MENKE, AND K. RICE. 1989. Competition for soil water between annual plants and blue oak (Quercus douglasii) seedlings. Oecologia 79:533—41. GRIFFIN, J., P, MCDONALD, AND P. Mulck. 1987. California oaks: a bibliography. Pacific Southwest Forest and Range Research Station. GTR-PSW-96. JACKSON ET AL.: QUERCUS GARRYANA STAND STRUCTURE 281 GRIFFIN, J. R. 1971. Oak regeneration in the upper Carmel Valley, California. Ecology 52:862—68. GRIFFIN, J. R. 1976. Regeneration in Quercus lobata sa- vanna, Santa Lucia Mountains, California. American Midland Naturalist 95:442-85. HALL, L., M. GeorGE, T. ADAMS, P. SANDS, AND D. McCreary. 1990. The effect of season and stock den- sity on blue oak establishment. /n R. Standiford (ed), Proceedings of the symposium on oak woodlands and hardwood rangeland management, Oct. 31—Nov. 2, 1990; Davis, CA. Pacific Southwest Forest and Range Experiment Station, Berkeley, CA. HEADY, H. E AND L. RADER. 1958. Modifications of the point frame. Journal of Range Management 11:95— 96. Hipss, D. E. AND B. J. YODER. 1993. Development of Oregon white oak seedlings. Northwest Science 67: 30-36. Horvitz, D. AND D. THompson. 1952. A generalization of sampling without replacement from a finite universe. Journal of the American Statistical Association 47: 663-685. KAUFFMAN, J. B. AND W. C. KRUEGER. 1984. Livestock impacts on riparian ecosystems and streamside man- agement implications: a review. Journal of Range Management 37:430—438. LATHROP, E., C. OSBORNE, A. ROCHESTER, K. YEUNG, S. SORET, AND R. Hopper. 1990. Size class distribution of Q. engelmannii on the Santa Rosa Plateau, Riv- erside, California. /n R. Standiford (ed), Proceedings of the symposium on oak woodlands and hardwood rangeland management, Oct. 31—Nov. 2, 1990; Davis, CA. Pacific Southwest Forest and Range Experiment Station, Berkeley, CA. MILES, S. 1993. Soil survey of Six Rivers National Forest area: USDA Soil Conservation Service, Pacific Southwest Region. Muick, P. AND J. BARTOLOME. 1987. An assessment of natural regeneration of oaks in California. California Division of Forestry. Contract# 8CA42136. MulIck, P. AND J. W. BARTOLOME. 1985. Research Studies in California related to oaks. Wildland Resources Center, University of California, Berkeley. Report No. 7. OAKESHOTT, G. 1978. California’s changing landscapes: a guide to the geology of the state. New York: Mc- Graw-Hill Book Company. OVERTON, W. S. AND S. V. STEHMAN. 1995. The Horvitz- Thompson theorem as a unifying perspective for probability sampling: with examples from natural re- source sampling. The American Statistician 49:261— 268. PAVLIK, B., P. Muick, S. JOHNSON, AND M. Popper. 1991. Oaks of California. Los Olivos, CA: Cachuma Press, Inc. REED, L. AND N. SUGIHARA. 1986. Prescribed fire for res- toration and maintenance of Bald Hills oak wood- lands. In T. R. Plumb and N. H. Pillsbury (eds.), Pro- ceedings of the symposium on multiple-use manage- ment of California’s hardwood resources, Nov. 12— 14, 1986; San Luis Obispo, CA. Pacific Southwest Forest and Range Research Station. RUNDEL, P. W. 1979. Adaptations of mediterranean-climate oaks to environmental stress. /n T. R. Plumb (ed.), Proceedings of the symposium on the ecology, man- agement, and utilization of California oaks, June 26— 28, 1979; Claremont, CA. Pacific Southwest Forest and Range Research Station. 282 RUNDEL, P. W. 1986. Origins and adaptations of California hardwoods. Jn T. R. Plumb and N. H. Pillsbury (eds.), Proceedings of the symposium on multiple-use man- agement of California’s hardwood resources, Nov. 12-14; San Luis Obispo, CA. Pacific Southwest For- est and Range Research Station. SAENZ, L. AND J. O. SAWYER. 1986. Grasslands as com- pared to adjacent Quercus garryana woodland un- derstories exposed to different grazing regimes. Ma- drono 33:40—46. SARNDAL, C., B. SWENSSON, AND J. WRETMAN. 1992. Mod- el Assisted Survey Sampling. Springer-Verlag, New York, NY. SAWYER, J. O., D. A. THORNBURGH, AND J. R. GRIFFIN. 1977. Mixed evergreen forest. Jn M. J. Barbour and J. Major (eds.), Terrestrial vegetation of California. John Wiley and Sons, New York, NY. SHARROW, S. H. AND W. C. LEININGER. 1982. Forage pref- erences of herded sheep as related to brush control and seasonal browsing damage to Douglas-fir regen- eration. Oregon Agricultural Experiment Station. Project #0066. STEIN, W. I. 1990. Quercus garryana Dougl. ex Hook. MADRONO [Vol. 45 Oregon white oak. Jn R. M. Burns and B. H. Honkala (eds.), Silvics of North America: hardwoods. USDA Forest Service, Washington, DC. SUGIHARA, N. G., L. J. REED, AND J. M. LENIHAN. 1987. Vegetation of the Bald Hills oak woodlands Redwood National Park, California. Madrofio 34:193—208. SUNTER, A. 1977. List sequential sampling with equal or unequal probabilities without replacement. Journal of Royal Statistical Society; Series C 26:261—268. USDA. 1993. Soil Survey Manual: Agricultural Hand- book No. 18. U.S. Dept. of Agriculture: For sale by the Supt. of Docs., U.S. G.P.O., Washington, DC. WatTT, A. S. 1919. On the causes of failure of natural regeneration in British oakwoods. Journal of Ecology 7:173-—203. WELKER, J. AND J. MENKE. 1990. The influence of simu- lated browsing on tissue water relations, growth and survival of Quercus douglasii (Hook and Arn.) seed- lings under slow and rapid rates of soil drought. Functional Ecology 4:807-17. YATES, E AND P. GRUNDY. 1953. Selection without replace- ment from within strata with probability proportional to size. Journal of the Royal Statistical Society; Series B 15:235-61. MADRONO, Vol. 45, No. 4, pp. 283-289, 1998 RELATIVE CONTRIBUTION OF BREEDING SYSTEM AND ENDEMISM TO GENOTYPIC DIVERSITY: THE OUTCROSSING ENDEMIC TARAXACUM CALIFORNICUM VS. THE WIDESPREAD APOMICT T. OFFICINALE (SENSU LATO) JENNIFER C. LYMAN!” AND NORMAN C. ELLSTRAND! ‘Department of Botany and Plant Sciences, University of California, Riverside, CA 92521-0124 and *Department of Biology, Rocky Mountain College, 1511 Poly Drive, Billings, MT 59102-1796 ABSTRACT We used allozymes to compare the population genetic structure of Taraxacum californicum Munz & I. M. Johnston, an outcrossing endemic of the San Bernardino Mountains of California, to that of its widespread weedy relative, 7. officinale Wigg. (sensu lato), an obligate apomict. The average number of allozyme phenotypes per population of the endemic was six times that of the widespread species. The endemic had about twice as much genotypic diversity (estimated by D) per population than the widespread species, and that diversity was distributed much more evenly per population than in the widespread species. Both species had about the same level of average interpopulation differentiation. For this pair of related species, breeding system apparently plays a more important role than endemism in determining population genetic structure. Restricted geographical distribution and small population size are characteristics of endemic plant species. Population genetic theory predicts that these conditions should influence population genet- ic structure, and, as a consequence, endemic species are expected to lack or to have reduced genetic polymorphism (reviewed by Ellstrand and Elam 1993). Empirical evidence that supports those the- oretical expectations is accumulating; and genetic diversity tends to increase with species range size (e.g., Hamrick and Godt 1990; Karron 1991). Breeding systems are also known to influence population genetic structure. Theory predicts that predominantly selfing and asexual species should exhibit lower variation within populations and greater interpopulation differentiation than out- crossers (Jain 1976; Baker 1959; Levin and Kerster 1971). Again, empirical data support these expected trends (e.g., Ellstrand and Roose 1987; Hamrick and Godt 1990). It is not clear which factor, breeding system or endemism, should have a greater influence on pop- ulation genetic structure. The answer to this ques- tion will be an important one for plant conservation managers who are rarely able to assess the genetic diversity of every species put in their charge. Such information will help managers judge under what circumstances loss of genetic variation might be most severe. For example, all other things being equal, if breeding system is more important than endemism, then outcrossing species will have more genetic variation to lose as their populations be- come increasingly fragmented under disturbance whereas clonal or selfing populations will tend to start with relatively low variation. Population genetic comparisons of predominant- ly outcrossing endemic species and related selfing or asexual widespread species would reveal valu- able information about the relative contributions of breeding system and endemism to population struc- ture. If endemism is the more important factor, then genetic analyses should reveal a paucity of genetic polymorphism within populations of the endemic species compared to a widespread selfing or asex- ual congener. Alternatively, if breeding system in- fluences population structure more than endemism, then within-population genotypic diversity of an outcrossing species with a restricted distribution should be greater than that of a selfing or asexual widespread species. And because interpopulation gene flow is typically higher in outcrossing species than a selfing or asexual species (Hamrick and Godt 1990), then we would expect that an outcrossing endemic should exhibit less interpopulation differ- entiation than a widespread selfing or asexual spe- cles. A genus well suited for comparative study of the effects of endemism and breeding system on pop- ulation genetic structure is Taraxacum. The great majority of species in this large genus are asexual, producing seeds that are genetically identical to the maternal parent through apomixis (agamospermy); and most of the remainder are predominantly or obligately outcrossing species (Grant 1981). The Species we chose for study are the T. californicum Munz & I. M. Johnston (California dandelion) and 284 T. officinale Wigg. (common dandelion). Although T. californicum is in the section Ceratophora and T. officinale is in Ruderalia, the two species are so closely related that they occasionally spontaneously hybridize when they come in contact (Skinner and Pavlik 1994). Taraxacum californicum is a predominantly out- crossing perennial herb endemic to moist, subalpine (1950—2400 m) meadows of the eastern San Ber- nardino Mountains of southern California (Munz 1974; Krantz 1980; Hickman 1993). The species was recently listed as endangered by the U.S. Fish and Wildlife Service (Federal Register 1998). Taraxacum officinale, also a perennial herb na- tive to Europe, is a pantemperate weed of lawns, meadows, and disturbed places (in California, from O-—3300 m); and, it occurs throughout North Amer- ica (Munz 1974; Hickman 1993). The taxon is ob- ligately apomictic. While treated as a single species by most North American researchers, T. officinale probably represents what many European experts judge to be a complex of agamospecies (Grant 1981; Richards personal communication). Like oth- er North American researchers (Taylor 1987; King 1993), we were unable to identify any character that allowed for the easy assignment of separate taxa. Therefore, we consider the species to be “‘sen- su lato’’. The purpose of this study was to survey the ge- notypic diverity within and among populations of T. califiornicum, an outcrossing endemic, and T. of- ficinale, its widespread apomictic congener. We compared their patterns of genetic diversity to dis- tinguish the relative contributions of breeding sys- tem and endemism on their population genetic structure. Additionally, we add some baseline in- formation on the genetic biology of T. californicum with bagging experiments and chromosome counts. METHODS AND MATERIALS Collection of material. We used unopened flower buds for our genetic analysis. In the case of T. cal- ifornicum, in June 1982 we collected one unopened flower bud from each of 30 randomly selected plants in each of 5 representative populations from its range in the eastern San Bernardino Mountains of California (Fig. 1). The buds were stored in plas- tic bags and kept cool until they could be extracted for allozyme study in Riverside. We also collected achenes to be germinated for chromosomal analysis from 3 individuals located in the population near Bluff Lake. In the case of T. officinale, flower buds were ob- tained from plants grown from seed. During June and July 1980 we collected a mature infructescence (head) from each of 30 randomly selected plants in each of 22 T. officinale populations across the Unit- ed States, detailed in Lyman and Ellstrand (1984). Heads were considered to be mature when they were fully opened, with achenes exposed to the MADRONO [Vol. 45 wind. The heads were transported to Riverside for germination as described below. In collecting both species, plants sampled were separated spatially (<>1 m) to insure that rosettes were not attached to the same taproot (Naylor 1941). All collection sites were isolated from each other by at least 4 kilometers. Germination. Achenes were germinated in 4-inch styrofoam cups in a greenhouse at the University of California at Riverside and later transferred to clay pots in the lathhouse. Germination occurred readily within 3—7 days. One seedling per maternal parent was grown to maturity for electrophoretic analysis. Additional seedlings were used for chro- mosome counts. Chromosome counts. Chromosome numbers were counted for the offspring of five individuals from the Bluff Lake population of 7. californicum and from two individuals from each population of T. officinale. We germinated seeds from these pop- ulations on moist filter paper in petri dishes under lab conditions. Four- to six-day-old root tips were placed in vials with 0.2% colchicine solution for 2 h and then fixed in 1:3 aceto-alcohol (Richards 1972a). The root tips were hydrolyzed in 0.1 N HCl for 11 minutes at 60°C and then stained in Feulgen solution for at least 1 h before the squashes were prepared (L6ve and Love 1975). We squashed 2- mm lengths of root tips on microscope slides in 1% aceto-propionic acid before viewing root tip mei- otic cells under the microscope. A minimum of three metaphase plates per plant were examined and counted. Testing for apomixis and autogamy in California dandelion. We bagged flowers of T. californicum to test whether it can set seed without the assistance of animal vectors (i.e., by apomixis or autogamy). In June 1982 at Wildhorse Meadows, the least dis- turbed site, small-mesh net bags were placed over one unopened capitulum on each of five plants of T. californicum. The bags were secured tightly to the ground with metal spikes to prevent insect movement in and out of the bags. In July the bags were removed, and each head was placed in a sep- arate plastic bag so that seed set, as judged by whether achenes were fully developed and filled, could be determined in the laboratory. Electrophoretic analysis. Unopened flower buds from up to 30 plants per population of 7. califor- nicum collected from the field were subjected to isozyme analysis by starch gel electrophoresis. For T. officinale, we analyzed unopened buds from up to 20 plants per population of plants grown from field-collected seed. Individuals buds were homog- enized in two drops of extraction buffer (0.01 M DTT buffered with 0.1 M Tris-HCl, pH 7.0). Ho- mogenates were adsorbed to paper wicks which were inserted vertically into 12% electrostarch gels. We used the Tris-EDTA-borate continuous gel and = 1998] LYMAN AND ELLSTRAND: GENETIC DIVERSITY IN TARAXACUM Belleville Meadow AAr Lower Holcomb = Valley Wr Barton FiAts igor Sta. Collection sites of 5 populations of Taraxacum californicum, San Bernardino Mountains, CA (inset area BiG. 1. enlarged). electrode buffer system of Heywood (1980). Elec- trophoresis was conducted for 4 h at 50 milliamps. Plastic containers of ice were placed on the gels during the run to prevent overheating. Internal stan- dards were run on each gel to determine the elec- trophoretic equivalence of bands from different populations. We assayed for three enzymes—alco- hol dehydrogenase (ADH), phosphoglucoisomerase (PGI), and phosphoglucomutase (PGM)—using the staining procedures described by Heywood (1980). Genetic analysis of isozyme patterns in related spe- cies (e.g., Roose and Gottlieb 1976) suggest that electrophoresis resolves five loci (one for ADH, two for PGI and PGM). Despite the polyploidy of these species (see Results), banding patterns were simple, and alleles were easily assigned. RESULTS Chromosome studies revealed a tetraploid chro- mosome complement in the nucleus of all individ- uals of T. californicum surveyed. In all counts, the ~ v Wildhorse Meadows INo ah UP Heartbar Horse Camp number of chromosomes was 31 (2n = 31). Similar levels of unusual aneuploid tetraploidy have been reported in some sexual European species of Tar- axacum (Richards 1972a, b, 1973). Taraxacum of- ficinale, however, has been shown to be a triploid, X = 8, 2n = 24 (Munz 1974). Our analysis of two individuals from each of 22 populations of this spe- cies showed the same chromosome number and ploidy level. Chromosome size was distinctly dif- ferent in the two species. Taraxacum californicum chromosomes were observed to be more than twice the size of those of T. officinale. Three of five bagged capitula produced no seed at all. The number of unfertilized ovules per head was 53, 73, and 76. Two remaining capitula pro- duced a few filled achenes, 8 of 78 in one, and 2 of 67 in another. Overall, 2.9% of the ovules pro- duced seed when foreign pollen was excluded. Un- bagged inflorescences on the same plants had full seed set (100%). These data support the hypothesis that 7. californicum is a self-incompatible outcross- ing species. 286 TABLE 1. ALLELES IN 7. CALIFORNICUM AND T. OFFICINALE. Taxon fae Locus Allelle californicum _ T. officinale PGI-1 a x x b x x re x X d X X e x f x PGM-1 a X x b X Cc x X d X x e x f x ADH a x x b x x Cc x 8 d x e ps Genotypic variation was present within and among the populations of 7. californicum and T. officinale investigated. We report genotypic rather than allele frequency data throughout this study be- cause dosage effects due to polyploidy in both spe- cies were not clearly discernible on the starch gels, making it impossible to determine allele frequen- cies. Levels of variation were different for the two species. Of the five loci considered, two were monomorphic for both T. officinale and T. califor- nicum. The polymorphic loci were shared by both species. Some alleles were common to both species, whereas others were species-specific (Table 1). We found 56 unique genotypes (allozyme phenotypes) among 147 individuals of T. californicum surveyed. The range of genotypes per population was 17-19 (x = 18.2; Fig. 2). No population contained more than six individuals of the same genotype. In con- trast, only 21 genotypes were found among the 518 individuals surveyed, ranging from 1—9 (x = 3.2) genotypes per population (Fig. 2). Pielou’s correction version of the Gini Index (D) describes the degree of diversity within a popula- tion (cf. Ellstrand and Roose 1987). This index em- phasizes changes in common rather than rare class- es. D can be as little as O (a uniform sample) or as large as 1 (every individual different). The correct- ed values for the five populations of T. californicum ranged from 0.94—-0.97 (x = 0.96). Values for T. officinale ranged from 0.00—0.89 (x = 0.50). Even- ness values (£), which reflect how evenly the ge- notypes within a population are distributed among individuals, were also calculated (Fig. 3) (cf. Ells- trand and Roose 1987). This value also ranges from zero (extreme skewness) to 1.0 (complete equita- bility). The striking bimodal distribution of E for T. officinale showed that in some populations one or a few dominant clones predominated but that in the MADRONO [Vol. 45 BT. officinale | BT. californicum POPULATIONS Serie aa 13] Li H tH re ao i fi k | le ii ia ! 1 w fl fn LI aos i r T -! J 7 T 1 2 3 4 5 6 7 8 9 10 11 #12 #13 #14 #«15 «#«16=«6«170«18~=«619 GENOTYPES Fic. 2. Number of allozyme phenotypes per population for T. californicum and T. officinale. others, the clones were equitably distributed among the individuals of the population. All five T. cali- fornicum populations, however, had evenness val- ues close to 1.0 (Fig. 3). Both evenness and diver- sity differed significantly between these species (P < 0.01; Mann-Whitney U test). Interpopulational differentiation within each spe- cies was measured using Hedrick’s (1971) formula for genotypic similarity. This index is as follows: 2 Bi Eiy PI a 2\i=1 j=l The values for this formula range from O (no sim- ilarity among populations) to 1 (complete genotypic identity among populations). The values of the five populations of JT. californicum were uniformly high, from 0.70—0.96 (x = 0.83). The values for the 22 populations of T. officinale ranged from 0.35-1.0 (x = 0.81). QT. officinale @T. califomicum NUMBER OF POPULATIONS 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 EVENNESS Fic. 3. Evenness values per populations for T. officinale and 7. californicum. 1998] DISCUSSION In our comparison of two congeners having dif- ferent recombination systems and range sizes, we found that the predominantly outcrossing endemic species T. californicum had substantially more ge- notypic diversity than the apomictic and wide- spread species T. officinale (sensu lato). The aver- age number of allozyme phenotypes per population of the endemic was six times that of the widespread species. Likewise, the endemic had about twice as much genotypic diversity per population as the widespread species, and that diversity was distrib- uted much more evenly per population than in the widespread species. Both species had about the same level of average interpopulation differentiation as estimated by Hed- rick’s J. The similarity is curious, given that the two species were sampled over such different scales (T. californicum over a scale of tens of kilometers; T. officinale over hundreds of kilometers). We would expect much more interpopulation differentiation in T. officinale for two reasons. First, we would expect more isolation (and consequently more differentia- tion) among distantly sampled populations com- pared to those sampled at a finer scale. Second, as noted above, all other things being equal, we would expect an apomictic species to have reduced gene flow relative to a sexual species (and consequently more differentiation). Gene flow by seed is the only means available to those species that reproduce without fertilization, whereas sexual species can disperse their genes by both seed and pollen. The best interpretation we can make is that the level of isolation among the meadows that make the home of the California dandelion are roughly the same as the level of isolation among our widely sampled common dandelion populations. Results from our bagging experiment are com- patible with an outcrossing breeding system for T. californicum based on self-incompatibility. Bagged capitula set few or no seed; unbagged capitula on the same plants had full seed set. Although this Species may be pseudogamous or semigamous, re- quiring pollination to stimulate apomictic seed pro- duction (Richards 1986), such syndromes are un- known for Taraxacum (Grant 1981). Also, net bags might have raised the temperature of the capitula to the point that apomictic seed production was dis- rupted (Stebbins personal communication). How- ever, as noted above, self-incompatible, endemic montane Taraxacum are known in Europe (Rich- ards 1973), and apomixis requiring pollination has never been reported for the genus (Grant 1981). Finally, the high genotypic diversity we discovered in T. californicum argues against high levels of ap- omictic seed production. Prior genetic surveys of agamoOspermous species typically show much lower genotypic diversity (e.g., Ellstrand and Roose 1987; Diggle et al. 1998). We acknowledge that demon- Strating sexuality conclusively requires further ex- LYMAN AND ELLSTRAND: GENETIC DIVERSITY IN TARAXACUM 287 perimentation, but, presently, all of the evidence supports outcrossing as T. californicum’s breeding system. We recognize that using a small number of mark- er loci can underestimate genotypic diversity. Add- ing more characters may identify more genotypes (Ellstrand and Roose 1987). Nonetheless, it is clear that the two species have different population ge- netic structures for the same set of genetically based markers. Genotypic diversity in T. officinale is low, but virtually every individual of T. califor- nicum was found to be genotypically distinct. In- deed, we are confident that if we were able to add a few more markers, we would be able to distin- guish among those few individuals of 7. californi- cum that shared a genotype in this study. For this pair of species, the difference in breed- ing system is apparently more important in deter- mining population genetic structure than the differ- ence in range size. The relatively high levels of genotypic diversity persist in the apparently out- crossing T. californicum despite its limited geo- graphic range. We are aware of only one other study that compares an endemic, outcrossing spe- cies with a widespread species that has an alterna- tive breeding system. The self-incompatible endem- ic of New Mexico’s Organ Mountains, Oenothera organensis Munz, is nearly monomorphic at several allozyme loci despite high polymorphism at self- incompatibility locus (Levin et al. 1979). Most of its congeners that have been studied have more ge- netic diversity (Levin et al. 1979), and almost all of these have an essentially clonal reproductive sys- tem (permanent translocation heterozygosity; Grant 1981). Thus, the trend in Oenothera is opposite that found here for Taraxacum. Taraxacum californicum is believed to be a relict of the section Ceratophora (Jepson 1925) that is thought to have spread through the northern hemi- sphere during an interglacial period (Richards 1973). Fossil records of the section date to 1 X 10° years B.P (Chearney and Mason 1936). Subse- quently T. californicum is presumed to have be- come isolated from the many other species of the section, which occur in a circumpolar distribution (Richard 1973). Given its long isolation, the fact that its total population size may be less than 6 x 10° plants, and its restricted habitat, the polymor- phism of 7. californicum is surprisingly high. The remarkable genetic diversity that T. califor- nicum exhibits despite its endemism suggests that breeding system plays a greater role in maintenance of genetic variation than the constraints of endem- ism do to limit it. But there is another possibility to account for the origin and maintenance of ge- netic variation. The ploidy difference between T. californicum and T. officinale may explain at least some of the greater genotypic variation displayed in T. californicum. If this species does, in fact, be- long to the section Ceratophora, then, like the tet- raploid European members of this section, it is 288 probably an allotetraploid derived from two diploid sexual species (Richard 1973). The resulting sexual amphiploid will initially breed true for a highly het- erozygous genotype but will release that variation slowly over generations through rare tetrasomic re- combination (Grant 1981; Roose and Gottlheb 1976). Another possibility is that T. californicum is not yet in evolutionary equilibrium. If fragmentation and endemism are relatively recent in terms of the mean generation time of the species, the patterns we observed better represent historic inertia than the factors that are currently molding patterns of diversity (Ellstrand and Elam 1993). Because the species was only described in this century, we know little of its history. If 7. californicum popu- lation genetic structure is not in equilibrium, we might expect to see the effects of isolation and en- demism eventually work to erode the current levels of genetic diversity, but such changes could take decades. A puzzling feature of the chromosome studies is the aneuploid condition found in the individuals of T. californicum at Bluff Lake. Investigators have noted the same phenomenon in a number of Euro- pean Taraxacum species (Malecka 1962, 1967a, b, 1969; Sorensen and Gudjonsson 1946; Richards 1970, 1972a, b, 1973), including many sexual spe- cies (e.g., Sorensen and Gudjonsson 1946). The ex- tent of the aneuploid condition and its role in re- productive events in 7. californicum are presently unknown. This research has uncovered relatively high lev- els of genotypic variation in an endemic species in comparison with its widespread apomictic conge- ner. We conclude that endemics need not necessar- ily be genotypically depauperate species relative to widespread congeners. Instead, it is clear that the organization of genetic variation may be subject to other constraints such as breeding system and his- tory. Further descriptive and experimental popula- tion genetics studies are needed for this and other Species pairs to determine the nature of these con- straints. ACKNOWLEDGMENTS We wish to thank Janet Clegg, Adrienne Edwards, Betty Lord, Maile Neel, Andy Schnabel, Nancy Smith-Huerta, Bob Soost, and Frank Vasek for their helpful comments on an earlier draft of this paper. Special thanks to Laura Heraty for her expert editing and work on the figures. LITERATURE CITED BAKER, H. G. 1959. Reproductive methods as factors in speciation in flowering plants. Cold Springs Harbor Symposium. Quantitative Biology 24:177-191. CHEARNEY, R. W. AND H. L. MAson. 1936. A Pleistocene flora from Fairbanks, Alaska. American Museum Novitates 887:17. DIGGLE, P. K., S. LOWER AND T. A. RANKER. 1998. Clonal diversity in Alpine populations of Polygonium vivi- MADRONO [Vol. 45 parum. International Journal of Plant Sciences 159: 606-615. ELLSTRAND, N. C. AND D. R. ELAM. 1993. Population ge- netic consequences of small population size: impli- cations for plant conservation. Annual Review of Ecology and Systematics 24:217—242. ELLSTRAND, N. C. AND M. L. Roose. 1987. Patterns of genotypic diversity in clonal plant species. American Journal of Botany 74:123-131. FEDERAL REGISTER. 1998. Endangered and threatened wildlife and plants; Final rule to determine endan- gered or threatened status for six plants from the mountains of southern California. Federal Register 63:49006—49022. GRANT, V. 1981. Plant Speciation, 2nd ed. Columbia Uni- versity Press, New York, NY. HAMRICK, J. L. AND M. J. W. Gopt. 1990. Allozyme di- versity in plant species. Pp. 43-63. in A. H. D. Brown, M. T. Clegg, A. L. Kahler, and B. S. Weir (eds.) Plant Population Genetics, Breeding, and Ge- netic Resources. Sinauer, Sunderland, MA. HEpbrRIck, P. W. 1971. A new approach to measuring ge- netic similarity. Evolution 25:276—280. HeEywoop, J. S. 1980. Genetic correlates of edaphic dif- ferentiation and endemism in Gaillardia. Ph.D. dis- sertation. University of Texas, Austin, TX. HICKMAN, J. C. 1993. The Jepson Manual: Higher Plants of California, University of California Press, Berke- ley, CA. JAIN, S. K. 1976. The evolution of inbreeding in plants. Annual Review of Ecology and Systematics 7:469— 495. JEPSON, W. L. 1925. A manual of flowering plants of Cal- ifornia. University of California Press, Berkeley, CA. KARRON, J. D. 1991. Patterns of genetic variation and breeding systems in rare plant species. Pp. 87—98. in D. A. Falk and K. E. Holsinger (eds.) Genetics and Conservation of Rare Plants. Oxford Press, New York, NY. KING, L. M. 1993. Origins of genotypic variation in North American dandelions inferred from ribosomal DNA and chloroplast DNA restriction enzyme analysis. Evolution 47:136—151. KRANTZ, T. P. 1980. Rare plant status report. California Native Plant Society, Sacramento, CA. Levin, D. A. AND H. W. KERSTER. 1971. Neighborhood structure in plants under diverse reproductive meth- ods. American Naturalist 105:345—354. , K. RITTER, AND N. C. ELLSTRAND. 1979. Protein polymorphism in the endemic Oenothera organensis. Evolution 33:534—542. Love, A. AND D. LOvE. 1975. Plant chromosomes. Strauss and Cramer, Vaduz. Liechtenstein. LYMAN, J. C. AND N. C. ELLSTRAND. 1984. Clonal diversity in Taraxacum officinale (Compositae), an apomict. Heredity 53:1—10. MALECKA, J. 1962. Cytological studies in the genus Tar- axacum. Acta Biologica Cracovienska 4:117—136. . 1967a. Chromosome numbers in five Taraxacum species from Mongolia. Acta Biologica Cracovienska 10:73-83. . 1967b. Cytoembryological studies in Taraxacum scanicum. Acta Biologica Cracovienska 12:195—206. . 1969. Further cyto-taxonomic studies in the ge- nus Taraxacum section Erythrospermum. Acta Biol- ogica Cracovienska 12:57—70. Munz, P. A. 1974. A flora of southern California. Uni- versity of California Press, Berkeley, CA. 1998] Nayor, E. 1941. The proliferation of dandelions from the roots. Bulletin of the Torrey Botanical Club 68: 351-358. RicHarps, A. J. 1970. Hybridization in Taraxacum. New Phytologist 69:1103—1121. . 1972a. The karyology of some Taraxacum spe- cies from alpine regions of Europe. Botanical Journal of the Linnaeus Society 65:47—57. . 1972b. In A. Love, 1. 0. P. B. Chromosome num- ber reports. Taxon 21:165—166. . 1973. The origin of Taraxacum agamospecies. Botanical Journal of the Linnean Society 66:189-— 211, . 1986. Plant breeding systems. Allen & Unwin, London, England. LYMAN AND ELLSTRAND: GENETIC DIVERSITY IN TARAXACUM 289 Roose, M. L. AND L. D. GOTTLIEB. 1976. Genetic and biochemical consequences of polyploidy in Trago- pogon. Evolution 30:818—830. SKINNER, M. W. AND B. M. PAvLik. 1994. California Na- tive Plant Society’s inventory of rare and endangered vascular plants of California. CNPS Special Publ- cation No. 1, 5th ed., Sacramento, CA. SORENSEN, T. AND C. GUDJONSSON. 1946. Spontaneous chromosome aberrants in apomictic Taraxaca. Kon- geligo danske Videnskabernes Selskab, Skrifter Biol- ogiske 4:1—48. TAYLOR, R. J. 1987. Population variation and biosyste- matic interpretations in weedy dandelions. Bulletin of the Torrey Botanical Club 114:109—120. MApRONO, Vol. 45, No. 4, pp. 290-294, 1998 EFFECTS OF SIMULATED OIL FIELD DISTURBANCE AND TOPSOIL SALVAGE ON ERIASTRUM HOOVERI (POLEMONIACEAB) JAY M. HINSHAwW,! GARY L. HOLMSTEAD,? BRIAN L. CYPHER,! AND DAvID C. ANDERSON? EG&G Energy Measurements, Inc., RO. Box 127, Tupman, CA 93276 ABSTRACT The effects of simulated oil field disturbance and topsoil G.e. E. hooveri seed bank) salvage on E. hooveri reestablishment were evaluated to develop effective strategies for conserving Eriastrum hooveri (Jeps.) Mason, a federally threatened plant. The study was conducted at two experimental sites at the former Naval Petroleum Reserve No. 1 (NPR-1), Kern County, CA. This species was initially present at Site 1 and nearly absent at Site 2. Six replications of five treatments were established simulating salvage and non-salvage of E. hooveri seed-laden soil before and after seed maturation and dispersion. Eriastrum hooveri densities were estimated in 1993 (pre-disturbance) and 1995 (post-disturbance). In this study we found that 1) surface disturbance negatively affected E. hooveri density for at least two growing seasons, 2) E. hooveri recolonized disturbed plots in two growing seasons from seed naturally dispersed from adjacent habitat, 3) topsoil salvage and respreading did not significantly affect the recolonization of E. hooveri on disturbed plots, 4) the timing of topsoil salvage had no effect, 5) E. hooveri was established at very low densities on several plots with no previous E. hooveri using topsoil from occupied habitat as a seed source, and 6) E. hooveri cover was inversely related to total vegetation cover but not to exotic grass COver. Eriastrum hooveri (Jepson) H. Mason, is a small annual herb endemic to the southern San Joaquin Valley and southern Coast Range regions of Cali- fornia (Munz 1973; Patterson 1993; Moe 1995). Plants exhibit wiry stems, alternate thread-like leaves, and small white flowers arranged in dense bracteate heads (Patterson 1993; Moe 1995). The species occurs in annual grassland and chenopod scrub habitats in portions of seven California coun- ties at elevations ranging from 50 to 910 m (Steb- bins et al. 1992; Lewis 1992; CDFG 1993; Patter- son 1993; Danielsen et al. 1994; Skinner and Pavlik 1994). Eriastrum hooveri often occurs in sandy loam soils derived from alluvial and colluvial par- ent material and underlying sedimentary rocks. Habitats occupied by E. hooveri commonly over- lie extensive hydrocarbon deposits; thus, oil and gas development and production activities have his- torically resulted in impacts to habitat suitable for this species. Such impacts primarily comprise soil disturbance from grading and facility and infra- structure construction activities. Although effects of oil and gas field related disturbances on E. hooveri were the focus of this study, the U.S. Fish and Wildlife Service (USFWS) cited impacts from ag- ricultural development, urbanization, and water ' Present address: NPRC Endangered Species and Cul- tural Resources Program, Critique, Inc., 1601 New Stine Road, Suite 240, Bakersfield, CA 93309. * Present address: Idaho Power Company, Environmen- tal Affairs Department, 1221 West Idaho Street, Boise, ID 83702. 3 Present address: Bechtel Nevada, PO. Box 98521, M.S. RSL-25, Las Vegas, NV 89193-8521. projects as the primary threats to the species’ ex- istence (USFWS 1990). Eriastrum hooveri was listed as threatened by the USFWS in 1990 (USFWS 1990), largely in re- sponse to Taylor and Davilla’s (1986) findings and the paucity of field observations during the three- year period of drought preceding federal listing. However, the results of more recent botanical sur- veys conducted during non-drought years showed that this species was more common and widespread than originally believed (Lyman et al. 1991; Steb- bins et al. 1992; Lewis 1992, 1994). The need for its continued listing as threatened has been ques- tioned (Lewis 1992, 1994; Willoughby 1995). Lew- is (1994) suggested that the protection of large tracts of E. hooveri habitat on federally managed lands would ensure survival of the species. The Bu- reau of Land Management has submitted a proposal to the USFWS recommending the species be de- listed and the USFWS has indicated that it may follow that recommendation. Currently, federal agencies continue to manage EF. hooveri popula- tions on federally administered lands in accordance with Section 2(c)(1) of the Endangered Species Act of 1973. The conservation of E. hooveri within its range in petroleum producing areas such as the former NPR-1I (now referred to as the Elk Hills Oil Field), necessitates the understanding of the effects of oil and gas developmental activities on the species. Primary strategies recommended by the USFWS and used by the U.S. Department of Energy (DOE) to mitigate impacts to E. hooveri populations at the Elk Hills Oil Field included population avoidance, or, if unavoidable, salvage and replacement of E. 1998] Kilometers 5 0 iG: 1. 18G of the Reserve. Cartography by Mark R. M. Otten. hooveri seed-laden topsoil. Subsequent to comple- tion of oil field projects seed collection and reseed- ing was often not possible due to project timing and annual variation in E. hooveri seed production. Therefore, topsoil salvage and respreading follow- ing completion of the project or on nearby areas in need of habitat restoration typifies impact mitiga- tions to this species. However, the effects of dis- turbance on E. hooveri, the effectiveness of topsoil salvage, and the effects of topsoil salvage timing were unknown. This study investigated the effects of simulated oil field disturbance on, and the effi- cacy of topsoil salvage for, E. hooveri prior to and following seed maturation and dispersion. Entire journals are devoted to the topic of natural lands restoration and management (e.g., Restora- tion Ecology, Restoration & Management Notes) and the scientific literature has a profusion of books and articles describing the effects of various kinds of habitat manipulation on unwanted alien and de- sirable native and naturalized plants. Methods and results of transplanting, reseeding, and introduction of sensitive or endangered plants have been studied (Hiatt et al. 1995), especially those susceptible to poaching such as rare cacti and orchids (Lyons 1987; Allen 1994). However, except for research conducted by Holmstead and Anderson (1998) and reported in this issue, we are unaware of field trials involving experimental use of topsoil as a seed source at study sites occupied by threatened or en- dangered annual plants. METHODS The DOE conducted a manipulative field study (with USFWS approval) from April 1993 to July 1995 at the former NPR-1, 40 km southwest of Ba- kersfield, Kern County, CA (Fig. 1). Two E. hoo- vert study sites were located in Section 18G (Sec- Naval Petroleum Reserve No. 1 Section 18G Study Site Locations HINSHAW ET AL.: ERIASTRUM HOOVERI CONSERVATION Zo 1 Sacramento ¢ Los Angeles Map of Naval Petroleum Reserve No. 1. Two Eriastrum hooveri study site locations are shown within Section tion 18, Township 31 South, Range 24 East, Mount Diablo Base & Meridian). Site 1 was about 170 m above sea level; and Site 2, located 850 m north of Site 1, was about 190 m above sea level (Fig. 2). Vegetation at both sites is characteristic of the Val- ley Saltbush Scrub community as described by Hol- land (1986). Prior to the study, E. hooveri was known to occur in relatively high densities at Site 1 and was believed absent at Site 2. The regional climate is hot and dry in summer, and is cool and wet in winter with periodic fog. Annual ambient air temperatures generally range from 0—38°C (National Weather Service, no date). Annual precipitation averaged 156 mm between 1975 and 1994, occurring mostly as rain from No- vember—April (National Climatic Data Center 1975-1995). Precipitation contributing to the grow- ing season for annual plants (October—March pre- cipitation) was 225 mm in 1993, 113 mm in 1994, and 227 mm in 1995 (National Climatic Data Cen- ter 1992-1995). In the spring of 1993, thirty 6 AN = April, no topsoil imported; AT = April, with topsoil; JN = July, no topsoil imported; JT = July, with topsoil; C = control. 4 Pre-disturbance measurements. > Means within a column with different letters are significantly different at a = 0.05. Sites |! Site 2? Treatment? 1993+ 1995 1993+ 1995 AN 182 A> O.85A 0) OA (0.7436) (0.5051) AT 2.32 A 0.52 A 0) 0.02 A (1.5372) (0.1956) (0.0167) JN 11.8 A LISA 0) OA CHATI9): (03265) JT 247A 0.85 A 0) 0.18 A (1.2785) (0.2377) (0.1641) C 2.30 A 437 B 0) 0.02 A (1.0139) (0.8758) (0.0167) from 2.30 to 4.37 plants per 0.5 m? on Site | control plot transects, but again, this increase was not sta- tistically significant. DISCUSSION The effects of surface disturbance on E. hooveri are poorly understood. A common perception held by the authors of this paper and other botanists who have studied FE. hooveri is that colonies of this spe- cies appear to be tolerant of some undetermined level of disturbance and that the species is adapted to generally open microhabitats (e.g., Lewis 1992, 1994; Holmstead and Anderson 1998). Eriastrum hooveri plants are often present on previously dis- turbed areas (Taylor et al. 1988; Lyman et al. 1991; Lewis 1992; Holmstead and Anderson 1998), sometimes with the disturbance apparently defining E. hooveri colony boundaries (Lewis 1994). Lewis (1994) found that 49 of 53 E. hooveri sites threat- ened by off-highway vehicle usage were situated on previously disturbed sites. Cypher (1994) ob- served higher E. hooveri survival rates on grazed than ungrazed areas, and no difference in E. hoo- veri fecundity between grazed and ungrazed areas. Holmstead and Anderson (1998) suggested that some level of habitat disturbance is compatible with FE. hooveri conservation. In our study, E. hooveri density was negatively correlated with total vege- tation cover, although the relationship was admit- tedly weak. This is consistent with our general field observations. Many E. hooveri locations on and ad- jacent to NPR-1 are 1) naturally or artificially dis- _ turbed sites supporting early successional species, and 2) relatively open microhabitats at sites domi- nated by later successional species. We found no HINSHAW ET AL.: ERIASTRUM HOOVERI CONSERVATION 295 correlation between B. madritensis ssp. rubens cov- er and E. hooveri cover, so the amount of overall vegetation cover, rather than exotic grass cover, seems to limit EF. hooveri growth. Our results support the hypothesis that this spe- cies readily recolonizes relatively small sites sub- jected to simulated oil field disturbance. During the study, E. hooveri recolonized disturbed Site | plots two growing seasons after disturbance. If precipi- tation prior to the 1994 growing season had not been below average (113 mm versus 143 mm nor- mal), E. hooveri recolonization might conceivably have occurred by the first growing season, as ob- served by Holmstead and Anderson (1998). In our study, respreading of seed-laden topsoil led to the growth of E. hooveri at very low densities on several previously unoccupied Site 2 plots; how- ever, E. hooveri densities were lower than on Site 1, probably due to the lack of seed dispersal from adjacent occupied habitat. Because of the extremely low densities that resulted, it appears that topsoil importation for the purpose of establishing E. hoo- veri on unoccupied habitat may not be an effective conservation measure. Although E. hooveri reestablishment was achieved on Site 1, FE. hooveri density was signifi- cantly lower on disturbed plots than control plots. This lower density is probably temporary because E. hooveri density on disturbed plots studied by Holmstead and Anderson (1998) was similar to or higher than on control plots after five growing sea- sons (Hinshaw unpublished). In our study, further monitoring will be needed to determine the recov- ery period for E. hooveri at Sites 1 and 2. Eriastrum hooveri densities on Site 2 plots that received topsoil collected in July were higher than on plots receiving topsoil collected in April, but the difference was not statistically significant. This slight difference may have resulted from initially higher E. hooveri densities on Site 1 JN plots from which the topsoil was collected (Table 1). These data support the conclusion that timing of topsoil salvage did not affect post-disturbance FE. hooveri densities. Apparently, seed dispersal from adjacent habitat and seeds contained in the soil seed bank contributed more to recovery than did the 1993 seed crop. Therefore, a mitigation requirement to delay oil field activities until after E. hooveri seed set would appear to be both ineffective and unnec- essary for E. hooveri conservation. Eriastrum hooveri densities on Site 1 were sim- ilar to or lower on disturbed plots receiving topsoil than disturbed plots with no topsoil. This result was unexpected because topsoil removal was equivalent to soil seed bank removal. Eriastrum hooveri plants on the plots with no topsoil probably resulted from seeds naturally dispersed from adjacent occupied habitat. On these plots, topsoil salvage did not ap- pear to be an effective strategy for enhancing the recolonization of this species on relatively small disturbances. Seeds from adjacent habitat apparent- 294 ly dispersed onto disturbed sites, producing plants after 1-2 growing seasons. Therefore, topsoil sal- vage and respreading on relatively small distur- bances within areas occupied by E. hooveri would seem unnecessary for purposes of species conser- vation. Funding for future studies of E. hooveri is un- certain because this species apparently is slated for delisting (Warren personal communication). Should further research occur, however, we recommend that germination studies be conducted under con- trolled conditions to learn more about seed bank dynamics of this species. Habitat manipulation studies of the effects of flooding, fire, herbivory, and anthropogenic surface disturbance on E. hoo- veri would certainly add further insights useful in developing management strategies for conserving this species. In addition, we strongly support Lew- is’ (1992, 1994) contention that further field inven- tories are needed for this cryptic herb. In conclusion, E. hooveri density was negatively affected by simulated oil field disturbance for at least two growing seasons, simulated topsoil sal- vage did not enhance E. hooveri reestablishment on disturbed plots, the timing of topsoil salvage did not affect the density of subsequent E. hooveri plants, and E. hooveri cover was not related to ex- otic grass (B. madritensis ssp. rubens) cover, but was inversely related to total vegetation cover. ACKNOWLEDGMENTS We express our gratitude to the U.S. Department of En- ergy and Chevron USA, Inc. for funding this project. We thank the hard-working field crew members who gathered the raw data from 1993-1995. Finally, we thank Dr. Ellen Cypher, Dr. Kent Ostler, and two anonymous reviewers for their constructive review of the draft manuscript. LITERATURE CITED ALLEN, W. H. 1994. Reintroduction of endangered plants. BioScience 44:65—68. BonHaM, C. D. 1989. Measurements for terrestrial vege- tation. John Wiley and Sons, Inc., New York. CALIFORNIA DEPARTMENT OF FISH AND GAME (CDFG). 1993. Rarefind: Eriastrum hooveri. California Natural Diversity Data Base. California Department of Fish and Game, Sacramento, CA. CYPHER, E. A. 1994. Demography of Caulanthus califor- nicus, Lembertia congdonii, and Eriastrum hooveri, and vegetation characteristics of endangered species populations in the southern San Joaquin Valley and the Carrizo Plain Natural Area in 1993. Unpublished report to the California Department of Fish and Game, Sacramento, CA. DANIELSEN, K. T., T. M. AUSTIN, AND C. WONG-LEE. 1994. Field inventory of Caulanthus californicus (Califor- nia jewelflower) in Los Padres National Forest. Un- published report to U.S. Forest Service, Goleta, CA. HIATT, H. D., T. E. OLSON, AND J. C. FISHER, JR. 1995. Reseeding four sensitive plant species in California and Nevada. Pp. 94-101 in Roundy, B. A. et al. (compilers), Proceediigs: wildland shrub and arid land restoration symposium. General Technical Re- MADRONO [Vol. 45 port INT-GTR-315. USDA Forest Service Intermoun- tain Research Station, Ogden, UT. HOLLAND, R. F. 1986. Preliminary descriptions of the ter- restrial natural communities of California. California Department of Fish and Game, Sacramento, CA. HOLMSTEAD, G. L. AND D. C. ANDERSON. 1998. Reesta- blishment of Hoover’s wooly-star (Eriastrum hooveri) following disturbance. Madrofo 45(4):(in press). Lewis, R. 1994. Eriastrum hooveri field inventory: 1994. Unpublished report to U.S. Bureau of Land Manage- ment, Caliente Resource Area, Bakersfield, CA. . 1992. Eriastrum hooveri field inventory: 1992. Unpublished report to U.S. Bureau of Land Manage- ment, Caliente Resource Area, Bakersfield, CA. LYMAN, J. C., K. WILSON, S. BIANCO, B. WUNNER, AND G. CLIFTON. 1991. Botanical survey for Eriastrum hoo- veri on Naval Petroleum Reserve No. 1, Kern Coun- ty, California. BioSystems Analysis, Inc. Report No. J-598 to EG&G Energy Measurements, Inc., Tupman, CA. Lyons, G. 1987. Landscaping with extinction. Pp. 226— 244 in D. Fuller and S. Fitzgerald (eds.), Conserva- tion and commerce of cacti and other succulents. World Wildlife Fund, Washington, D.C. Mog, L. M. 1995. A key to vascular plant species of Kern County, California. California Native Plant Society, Sacramento, CA. Munz, P. A. AND D. D. Keck. 1973. A California flora and supplement. University of California Press, Berkeley, CA. NATIONAL CLIMATIC DATA CENTER (NCDC). Bakersfield, CA. 1975-1995. National Climatic Data Center. Asheville, NC. NATIONAL WEATHER SERVICE. No date. The climate of Kern County. Kern County Board of Trade, Bakers- field, CA. PATTERSON, R. W. 1993. Eriastrum. Pp. 826 and 828 in J. C. Hickman (ed.), Jepson manual: higher plants of California. University of California Press, Berkeley, CA. Sas INSTITUTE INC. 1990. SAS/STAT user’s guide, version 6. Cary, NC. SKINNER, M. W. AND B. M. PAVLIK (eds.). 1994. Inventory of rare and endangered vascular plants of California. Special Publication No. 1, California Native Plant So- ciety, Sacramento, CA. STEBBINS, J. C., T. C. MALLORY, W. O. TRAYLER, AND G. W. Molise. 1992. Botanical resources report: Califor- nia Aqueduct. Unpublished report prepared for the San Joaquin Field Division, California Department of Water Resources. TAYLOR, D. W. AND W. B. DAVILLA. 1986. Status survey for three plants endemic to the San Joaquin Valley and adjacent areas, California. BioSystems Analysis, Inc. Report No. J-235, U.S. Fish and Wildlife Service, Sacramento, CA. TayYLor, D. W., R. E. PALMER, G. L. CLIFTON, AND K. A. TEARE. 1988. Reconnaissance survey for federal can- didate rare plants, Elk Hills Naval Petroleum Reserve | No. 1. BioSystems Analysis, Inc. Report No. J-361 to EG&G Energy Measurements, Inc., Tupman, CA. U.S. FisH AND WILDLIFE SERVICE (USFWS). 1990. Endan- gered and threatened wildlife and plants: determina- i tion of endangered and threatened status for five | plants from the southern San Joaquin Valley. Federal _ Register 55:29361—29370. WILLOUGHBY, J. 1995. The future of California’s floristic | heritage on public lands. Madrofio 42(2):242—268. MaprRONO, Vol. 45, No. 4, pp. 295-300, 1998 REESTABLISHMENT OF ERIASTRUM HOOVERI (POLEMONIACEAE) FOLLOWING OIL FIELD DISTURBANCE ACTIVITIES GARY L. HOLMSTEAD Idaho Power Company, Environmental Affairs Department, 1221 West Idaho Street, Boise, ID 83702 DAvID C. ANDERSON EG&G Energy Measurements, Inc., 101 Convention Center Drive, Las Vegas, NV 89109 ABSTRACT Little is known about the ecology of Eriastrum hooveri (Jepson) H. Mason or about its tolerance to oil field related habitat disturbance. Taylor and Davilla (1986) suggested the species was closely associated with dense cryptogamic soil crust, characteristic of undisturbed sites. We monitored reestablishment of E. hooveri on two sites disturbed by construction activities (a pipeline and a well pad) at the U‘S. Department of Energy’s Naval Petroleum Reserve No. 1, Kern County, CA. Before construction, topsoil from the sites was stockpiled. After construction, the topsoil was replaced, and the pipeline site and a portion of the well pad site were seeded with a mix of native shrub, grass, and forb species. Part of the well pad site was left unseeded so that reestablishment of E. hooveri could be compared between seeded and unseeded plots. Sites were monitored during the first two growing seasons following disturbance (1991 and 1992). Vegetation characteristics of the disturbed sites were compared with adjacent undisturbed habitat. Eriastrum hooveri recolonized both disturbed sites in the first growing season. Generally, the density and frequency of occurrence of E. hooveri on our transects increased from the first to the second growing season. Cryptogamic crust cover was low (=4.6%) on both the disturbed and undisturbed sites in both years. Our observations suggest that 1) E. hooveri is able to quickly recolonize heavily disturbed sites, at least if topsoil is conserved and weather conditions are favorable; and that 2) cryptogamic crust cover may not be as important a correlate with the occurrence of FE. hooveri as previously thought. Eriastrum hooveri (Jepson) H. Mason (Hoover’s woolly-star) is a small annual herb of the Polemon- iaceae and endemic to the San Joaquin Valley, Cal- ifornia. It was federally listed as threatened in July 1990, due to threats of agricultural land conversion, urbanization, reservoir construction, and oil and gas development (U.S. Fish and Wildlife Service 1990). Little is known about the ecology of E. hooveri. Taylor and Davilla (1986) provided some general observations on germination, soil seed reserves, growth phenology, and reproduction of the species. Early observations by Taylor and Davilla (1986) typically associated E. hooveri populations on sites without dense annual plant cover and with dense cryptogamic soil crusts (eucaryotic algae, lichens, bryophytes, cyanobacteria, and fungi), which nor- mally take several years to develop (Anderson et al. 1982a). A 1988 reconnaissance survey of NPR- 1 (EG&G Energy Measurements 1988) and on-go- ing surveys of NPR-1 (EG&G Energy Measure- ment 1992) have identified several E. hooveri pop- ulations in formerly disturbed sites. Most of these sites are on or near abandoned or infrequently-used roadways, suggesting that the species can respond favorably to disturbance. Eriastrum hooveri re- sponse to disturbance has not been experimentally investigated. In this paper, we evaluate the reestablishment of _ E. hooveri on two oil field construction projects, an underground pipeline and a well pad, during the first two growing seasons following disturbance. We describe the chronology of each disturbance and subsequent restoration activities; and we com- pare the density and frequency of occurrence of E. hooveri, and other characteristics of associated veg- etation, between the disturbed sites and adjacent undisturbed habitat. On the well pad site, experimental plots were used to compare reestablishment of EF. hooveri be- tween plots of seeded topsoil (1.e., replaced topsoil seeded with a mix of shrub, grass, and forb species) and unseeded topsoil. We hypothesized that by not seeding the conserved topsoil, plant competition would be reduced and E. hooveri reestablishment would be enhanced. Study area. The majority of the San Joaquin Val- ley is cultivated and very few remnants of native plant communities remain (Preston 1981). Al- though NPR-1 is an active oil and gas producing facility, it encompasses large tracts of native and naturalized vegetation in the San Joaquin Valley. It is located approximately 40 km southwest of Ba- kersfield, Kern County, CA, and consists of 19,120 ha. NPR-1 is located on the Elk Hills formation, which is in the most arid portion of cismontane California (Major 1977). Geomorphologically, the Elk Hills constitute a subsidiary upland of the Inner South Coast Ranges. The main ridge, oriented in a 296 TABLE 1. MADRONO [Vol. 45 LIST OF SPECIES AND SEEDING RATES USED TO SEED REDISTRIBUTED TOPSOIL ON THE PIPELINE STUDY SITE (DECEMBER 1990) AND THE WELL PAD STUDY SITE (JANUARY 1991), NAVAL PETROLEUM RESERVE No. 1, KERN COUNTY, CA. * PLS = Pure Live Seed, which equals (purity < germination)/100. PLS? Study site Life-form Scientific binomial kg PLS*/ha /m? a. Pipeline Shrub Atriplex polycarpa (Torrey) S. Watson leak 16 Atriplex lentiformis (Torrey) S. Watson 0.6 5 Eriogonum fasiculatum (Benth.) Torrey & A. Gray 4.5 344 Isomeris arborea Nutt. 3.4 5 Grass Vulpia myuros (L.) C. Gmelin lee 198 Forb Lupinus densiflorus Benth. 0.5 3 Phacelia tanacetifolia Benth. 0.5 107 b. Well Pad Shrub Atriplex polycarpa (Torrey) S. Watson 23 By Eriogonum fasiculatum (Benth.) Torrey & A. Gray 4.5 344 Isomeris arborea Nutt. 3.4 5) Grass Vulpia myuros (L.) C. Gmelin Tal 198 Forb Trifolium hirtum All. Vt 34 Lupinus densiflorus Benth. 0.5 3} Phacelia tanacetifolia Benth. 0.5 107 northwest-southeast direction, is flanked by deeply incised canyons and subsidiary ridges. The ridges and drainages extend into gently sloping, alluvial plains along the outer boundaries. Elevations range from 93 to 473 m above sea level. NPR-1 lies within the Valley Grassland vegeta- tion type (Heady 1977). Dominant shrubs include Atriplex polycarpa (Torrey) S. Watson, Hymenoclea salsola A. Gray, and Isomeris arborea Nutt. Her- baceous cover is dominated by Bromus madritensis L. and Erodium cicutarium (L.) L Her. The climate in this region is hot and dry in sum- mer, and cool and wet in winter with frequent fog. Temperatures in summer often exceed 38°C, and seldom go below 0°C in winter. Precipitation occurs primarily as rain falling between November and April (O’Farrell et al. 1987). Since 1981, when weather data collection began on NPR-1, annual precipitation has averaged 124 mm and ranged be- tween 51 and 226 mm. METHODS Pipeline study site. In 1990, a gas company com- pleted construction along two underground natural gas pipelines that crossed NPR-1. A large diameter pipeline was installed to replace a small pipeline constructed in 1930. A small pipeline, located ap- proximately 1 km away, was also removed. The construction corridor along the pipelines ranged be- tween 15-20 m wide, and was approximately 38 km long. Approximately 7 ha of habitat containing E. hooveri populations were disturbed within the pipeline corridors. Construction activities were de- layed until August, several months after the typical flowering season for E. hooveri (April—May), to al- low existing FE. hooveri plants to set seed. Prior to pipeline trenching operations, a road grader was used to scrape 7—8 cm of topsoil to the edge of the construction corridor in all FE. hooveri habitat. A road grader was then used to replace the topsoil following construction. Due to the deep, powdery nature of the disturbed soil, straw mulch was applied at a rate of about 9,000 kg/ha to im- prove soil structure. Straw was crimped into the soil using a sheep’s foot-type roller-crimper. All E. hooveri habitat was drill seeded with a mix of shrubs, forbs, and grasses at a rate of 11.2 kg of pure live seed (PLS) per hectare (Table 1a). Seed- ing was completed in December 1990. In spring 1991, the length of the pipeline corridor in E. hooveri habit was divided into 0.1 km seg- ments and ten segments were randomly selected. In each segment a random starting point was selected. At each starting point, two parallel 25-m line tran- ! sects were established to monitor vegetation. One | transect was located outside of the pipeline corridor | in undisturbed habitat, and the other transect was | located down the centerline of the reclaimed pipe- line corridor. Vegetation was sampled along each transect in the spring of 1991 and 1992. An ocular point projection device (ESCO Associates Inc., Boulder, CO) was used to estimate ground cover. | A total of 100 points or “‘hits’’ (a dimensionless | plot such as a point frame-type sample) were sam- pled along each transect; 10 points at 2.5 m inter- vals. Points were recorded as bare ground, litter, cryptogamic crust, or live vascular plant material, | by species. The elements classified as litter includ- ed both dead standing and detached biomass. Cryp- | togamic crust included just those elements that are | identifiable in the field, without magnification. The © density of E. hooveri was determined by recording | the number of plants that occurred within a 2 X 25 — m belt transect. The densities of grasses and forbs were determined by counting the number of indi- vidual plants within five 1 < 1 m quadrats placed | at 5 m intervals along the transect. Eriastrum hoo- veri frequency was estimated by using five 1 x 1 m quadrats placed at 5 m intervals along each tran- sect, and counting the number of quadrats contain- \ 1998] ing E. hooveri. Frequency of occurrence was ex- pressed as the percentage of quadrats containing E. hooveri. Well pad study site. A second study site was es- tablished near a new water well. In July 1990, con- struction began on the well before a biological sur- vey was conducted. About half of the area proposed for the well pad (0.4 ha) was scraped and 8—10 cm of topsoil was stockpiled. A survey of the site iden- tified several small stands (5—50 plants each) of E. hooveri in the surrounding undisturbed areas of the well pad site. To avoid further disturbance to this E. hooveri population, the new well was relocated to a nearby existing well pad. After consultation with the U.S. Fish and Wildlife Service, the US. Department of Energy established vegetation mon- itoring transects at the original well site to docu- ment reestablishment of E. hooveri. In August 1990, the stockpiled topsoil was spread back over the disturbed area, and the site was divided into nine study plots, about 10 x 50 m each. Three disturbed plots were drill seeded in January 1991 with a mix of shrubs, forbs, and grasses at a rate of 13 kg of PLS/ha (Table 1b), straw mulched at a rate of about 3,400 kg/ha, and crimped. Three disturbed plots were not seeded, and three plots were selected in adjacent undis- turbed E. hooveri habitat. Due to the pattern of dis- turbance created by construction equipment, plots were arranged side by side starting with a seeded topsoil plot adjacent to unseeded topsoil plot, which was then adjacent to undisturbed habitat. This order of treatments was replicated three times. A permanent 25-m line transect was established down the centerline of each plot to monitor vege- tation. A random starting location was selected for the first transect and all additional transects were aligned parallel to the first. Vegetation was moni- tored in the spring of 1991 and 1992 using the methods previously described for the pipeline study Site. Precipitation data. Monthly precipitation was re- corded with an All Weather Rain Gauge (Produc- tive Alternatives, Inc., Fergus Falls, MN) at eight stations on NPR-1. Total annual precipitation is ex- pressed on a water-year (WY) basis (e.g., WY91 = 1 July 1990 to 30 June 1991). Data analysis. On the pipeline site, the density and frequency of E. hooveri, and other vegetative characteristics (e.g., total plant cover, cryptogamic cover, density of grasses and forbs) were compared between undisturbed habitat and seeded topsoil plots. On the well pad site, the density and fre- quency of E. hooveri, and vegetative characteristics were compared between undisturbed habitat, seed- ed topsoil, and unseeded topsoil plots. On each study site, and for each plot type, the effects of year (1991 and 1992) and treatment were evaluated us- ing repeated measures analysis of variance. On the _ well pad site, linear contrasts between treatments HOLMSTEAD AND ANDERSON: ERIASTRUM HOOVERI REESTABLISHMENT 299, were used to separate mean values. Means were considered significantly different at alpha <0.05. SAS/STAT v.6 software (SAS Institute Inc. 1990) was used to perform statistical computations. RESULTS Annual precipitation was 137 mm in WY9I1 (10% above average) and 155 mm in WY92 (25% above average). Pipeline study site. Eriastrum hooveri was pres- ent on all transects on the seeded topsoil plots and undisturbed habitat in 1991 and 1992. The density and frequency of FE. hooveri plants increased sig- nificantly from 1991 to 1992 and were significantly higher on the undisturbed habitat (Table 2a). From 1991 to 1992, density of E. hooveri on the undis- turbed habitat increased from 2.1 to 5.3 plants/m’, and on the seeded topsoil plots it increased from 0.1 to 0.9 plants/m? (Fig. la). Between 1991 and 1992, the frequency of E. hooveri increased from 38 to 68% on the undisturbed habitat, and from 18 to 26% on the seeded topsoil plots (Fig. la). Total plant cover and density of grasses and forbs increased significantly from 1991 to 1992 and were higher on the undisturbed habitat (Table 2a). From 1991 to 1992, total plant cover, and the density of grasses and forbs generally increased significantly in both treatments (Table 3a). In 1991, cryptogam cover was 4.6% on the un- disturbed habitat, and absent on the seeded topsoil plots. In 1992, cryptogam cover was 1.3% on the undisturbed habitat and 0.5% on the seeded topsoil plots (Table 3a). Well pad study site. As on the pipeline site, E. hooveri was present on all transects in both dis- turbed and undisturbed habitat in 1991 and 1992. The density of E. hooveri increased significantly from 1991 to 1992, but frequency of FE. hooveri was not significantly different between years (Table 2b). The density and frequency of EF. hooveri plants were not significantly different between treatments (Table 2b). From 1991 to 1992, density of E. hoo- veri increased from 0.8 to 1.9 plants/m? on the un- disturbed habitat, increased from 0.4 to 2.3 plants/ m’ on the seeded topsoil plots, and increased from 0.4 to 4.0 plants/m’ on the unseeded topsoil plots (Fig. 1b). From 1991 to 1992, the frequency of E. hooveri increased from 46.7 to 53.3% on the un- disturbed habitat, decreased from 46.7 to 6.7% on the seeded topsoil plots, and increased from 26.7 to 33.3% on the unseeded topsoil plots (Fig. Ib). Total plant cover and density of grasses and forbs increased significantly from 1991 to 1992 and were highest on the undisturbed habitat (Table 2b). From 1991 to 1992, total plant cover and density of grass- es and forbs increased significantly in all treatments (Table 3b). Total plant cover and density of grasses and forbs were higher on the seeded topsoil plots compared to the unseeded topsoil plots, but only total plant 298 MADRONO [Vol. 45 TABLE 2. SUMMARY OF REPEATED MEASURES ANALYSIS OF VARIANCE ON THE EFFECTS OF YEAR AND TREATMENT ON ERIASTRUM HOOVERI DENSITY AND FREQUENCY, AND KEY VEGETATION CHARACTERISTICS ON THE PIPELINE AND WELL PAD STUDY SITES, NAVAL PETROLEUM RESERVE No. 1, KERN County, CA. * Factors are listed in decreasing order of mean = undisturbed habitat, ST = seeded topsoil, and UST = unseeded topsoil. Factors differing significantly (P < 0.05, linear contrasts) are indicated by “‘>’’. values where UND Study site a. Pipeline b. Well Pad Characteristic E. hooveri Density E. hooveri Frequency Total Plant Cover Grass Density Forb Density E. hooveri Density E. hooveri Frequency Total Plant Cover Grass Density Forb Density Factor P value Mean difference? Year <0.001 1992 > 199] Treatment 0.006 UND > ST Year X Treatment 0.012 Year 0.007 1992 > 1991 Treatment 0.003 UND > ST Year X Treatment 0.094 Year <0.001 1992 > 1991 Treatment 0.346 UND = ST Year X Treatment 0.053 Year 0.013 1992 > 1991 Treatment O21 UND = ST Year X Treatment 0.011 Year <0.001 1992 > 1991 Treatment 0.001 UND > ST Year X Treatment 0.933 Year 0.008 1992 > 1991 Treatment 0.538 UST = UND = ST Year X Treatment 0.249 Year 0.436 1992 = 1991 Treatment 0.374 UND = UST = ST Year X Treatment 0.199 Year <0.001 1992 > 1991 Treatment 0.002 UND > ST > UST Year X Treatment 0.003 Year <0.001 1992 > 1991 Treatment 0.006 UND > ST = UST Year X Treatment 0.024 Year 0.008 1992 > 1991 Treatment 0.538 UND = ST = UST Year X Treatment 0.249 cover was significantly higher (Table 2b). Density and frequency of FE. hooveri were higher on the unseeded topsoil plots compared to the seeded top- soil plots, but these differences were not significant (Table 2b). No cryptogamic soil crust was observed on any of the plots in 1991. In 1992, cover of cryptogams was 2.0% on the seeded topsoil plots, 1.0% on the unseeded topsoil plots, and 0.7% on the undis- turbed habitat (Table 3b). DISCUSSION Although these investigations are opportunistic in nature in that they were conducted without the opportunity to set up ideal experimental conditions, some observations can be made. These observa- tions should be substantiated with appropriate ex- perimental studies. The results of this investigation demonstrate that E. hooveri can quickly colonize disturbed sites, at least when topsoil is conserved and returned, and adequate rainfall is received. At both study sites, E. hooveri occupied all disturbed plots after one grow- ing season, and its density increased on the dis- turbed plots, at both study sites, from the first to the second growing season (Fig. 1). The frequency of E. hooveri increased on all disturbed plots except the seeded topsoil plots on the well pad study site, from the first to the second growing season (Fig. 1). During the first few growing seasons following disturbance, we expected that the density and fre- quency of E. hooveri on the disturbed plots would not be as high as on undisturbed habitat. This ex- pectation was confirmed at the pipeline study site (Table 2a). However, on the well pad site, neither the density nor the frequency of E. hooveri were significantly different between disturbed plots and undisturbed habitat (Table 2b). The small sample sizes (n = 3) for each treatment on the well pad site probably reduced the ability to detect signifi- cant differences. We hypothesized that by not seeding the con- served topsoil at the well pad site, plant competi- tion would be reduced and E. hooveri reestablish- ment would be enhanced. Unseeded plots had lower plant cover and lower densities of grasses and forbs than seeded topsoil plots in both 1991 and 1992 (Table 3b). However, E. hooveri density and fre- quency were not significantly higher on the un- seeded plots than on the seeded plots (Table 2b). 1998] A. Pipeline Study Site 10 9 YA 1991 8 “27 Hl 1992 Zs | a 5 B, 2 3 2 |g Y : OF er UND ST UND B. Well Pad Study Site UND ST UST UND ST UST Fic. 1. Summary of Eriastrum hooveri density and fre- quency on the pipeline and well pad study sites where ST = seeded topsoil, UST = unseeded topsoil, and UND = undisturbed habitat, U.S. Naval Petroleum Reserve No. 1, Kern County, CA. Vertical bars indicate each standard er- ror of the mean. The capacity for E. hooveri to quickly invade disturbed sites is supported by other observations on NPR-1. In spring 1992 we observed seven pop- ulations of E. hooveri within a firebreak corridor that is maintained around the perimeter of NPR-1 (Holmstead and Anderson unpublished). Various sections of the firebreak have been annually disked for many years depending on the amount of vege- tative cover present on the firebreak. Since 1990, all known E. hooveri populations have been avoid- ed by disking operations. However, two of the pop- HOLMSTEAD AND ANDERSON: ERIASTRUM HOOVERI REESTABLISHMENT 200 ulations we observed in spring 1992 were in sec- tions of the firebreak that had been disked in 1991 and five populations were in areas disked in 1989. The degree of reestablishment of E. hooveri in this study may partially be attributable to favorable growing conditions during WY91 and WY92. An- nual plant population sizes vary widely from year to year (Holland 1987), and this variation is tradi- tionally attributed to the vagaries of annual weather. A low rainfall year may result in very low numbers of a species, or even years when no plants are ob- served, while a higher rainfall year may result in large numbers of a species. Taylor and Davilla (1986) observed that FE. hooveri germinated rela- tively late (January—February) as opposed to after the first rainfall (October-November). In WY91, rainfall was 10% above average, and 91% (124 of 137 mm) occurred from January—March. In WY92, rainfall was 25% above average, and 80% (124 of 155 mm) occurred from January—March. Abundant rainfall in both WY91 and WY92, and concentra- tion of this rain between January—March may have promoted high germination and establishment of E. hooveri. None of the disturbed or undisturbed plots at ei- ther study site had ‘“‘dense patches of abundant soil cryptogams”’ that Taylor and Davilla (1986) re- ported were a principal correlate with the presence of E. hooveri. Cryptogamic cover was absent on most of the plots in 1991 (Table 3). The highest percent cover of cryptogams was 4.6%, which oc- curred on the undisturbed plots at the pipeline study site in 1991. In 1992, average cryptogam cover was =2.0% on all plots. These amounts of cryptogamic cover are well below what would be considered dense cover. Mean cover of cryptogams in non- grazed areas in Utah deserts was estimated at 53.6% (Brotherson et al. 1983). Average cover of TABLE 3. WEGETATION CHARACTERISTICS ON THE PIPELINE AND WELL PAD STUDY SITES DURING THE FIRST AND SECOND GROWING SEASONS (1991 AND 1992) FOLLOWING CONSTRUCTION ACTIVITIES, NAVAL PETROLEUM RESERVE No. 1, KERN County, CA. Standard errors of the mean are in parentheses. 1991 1992 Seeded Unseeded Undisturbed Seeded Unseeded Undisturbed Study site Factor topsoil topsoil habitat topsoil topsoil habitat a. Pipeline Sample Size 10 --— 10 10 — 10 Cover (%) Total Plant 31.3 (4.6) — 41.8 (1.9) 67.6 (3.1) —- 65.8 (4.9) Cryptogams 0.0 — 4.6 (1.4) 0:5-(0:3) —- 13:(0;3) Density (no./m?) Grass 24.2 (6.8) oo 95.0 (9:6) 124.4 (31.1) a 94.4 (8.6) Forb 13.8 (2.5) — 61.8 (9.4) 49.8 (9.3) —- 98.9 (14.9) b. Well Pad Sample Size 3 3 3 3 3 3 Cover (%) Total Plant 26.7 (4.5) 12.0 (1.7) 55.0 (4.5) 62.7 (3.6) 5o75' Glss) 6939) Cryptogams 0.0 0.0 0.0 20° (1:0) 1.0 (1.0) OFF (0.3) Density (no./m7?) Grass 21.3 (4.3) 1.7 (0.5) 34.5 (8.8) 130.5 (8.5) 85.4 (10.3) 216.7 (31.6) Forb S22 oe 7.00.6) 36.7 9) 36.7 (19.8). > 31.7 (8) G3 412.9) 300 cryptogams on grazed sites in Utah deserts was 6.3% on light developed crusts and 20.9% on mod- erate-heavy developed crusts (Anderson et al. 1982b). Our observations suggest that EF. hooveri is not restricted to dense cryptogamic soil crusts. Quantification of site characteristics associated with E. hooveri populations are needed. Some explanation for the low cryptogamic cover observed in this study compared to observations by Taylor and Davilla (1986), and for cover estimates reported for other investigations of cryptogamic crusts may be explained by 1) observer variability in cover estimates, 2) differences in antecedent pre- cipitation that might have made the crust more ap- parent in one sample year than another, or 3) the components of cryptogamic crusts that are included in the cover estimates. Algal cover estimates can be quite subjective since algae are less obvious than lichens and mosses and their cover estimates are dependant upon experienced observers (Anderson et al. 1982a). Cryptogamic crust is known to con- sist of eucaryotic algae, lichens, bryophytes, cy- anobacteria, and fungi that live on and just below the soil surface (Belnap 1994). Most general field inventory investigations focus on just those crust components that are visible without magnification, while more specific research on cryptogamic crust utilize laboratory detection procedures and estimate total cover of all components. The total cover es- timates referenced in this paper used similar field methods. The occurrence of E. hooveri on disturbed areas, and its apparent capacity to quickly occupy dis- turbed sites indicate that it may not be dependent on pristine habitats or dense cryptogamic cover. Some level of disturbance may be compatible with E. hooveri conservation. However, long term mon- itoring studies of disturbed FE. hooveri populations and recently colonized disturbed sites are warrant- ed. Such studies should investigate the persistence and vigor (size, flowering, and fruiting) of the plants over time. ACKNOWLEDGMENTS We thank the staff members at EG&G/EM who helped collect data on the study plots. B. Cypher, J. Hayes, T. O’Farrell, and G. Warrick reviewed the manuscript and provided valued comments. This investigation was con- ducted for the U.S. Department of Energy/Naval Petro- leum Reserves in California, and Chevron U.S.A. through the Nevada Field Office under Contract No. DE-AC08- 88NV 10617. MADRONO [Vol. 45 LITERATURE CITED ANDERSON, D. C., K. T. HARPER, AND S. R. RUSHFORTH. 1982a. Recovery of cryptogamic soil crusts from grazing on Utah winter ranges. Journal of Range Management 35:355—359. . 1982b. Factors influencing development of cryp- togamic soil crust in Utah deserts. Journal of Range Management 35:180-—185. BELNAP, J. 1994. Potential role of cryptobiotic soil crusts in semiarid rangelands. Pp. 179-185 in S. B. Monsen and S. G. Kitchen (compilers), Proceedings—Ecolo- gy and Management of Annual Rangelands. General Technical Report INT-GTR-313. USDA Forest Ser- vice Intermountain Research Station, Ogden, UT. BROTHERSON, J. D., S. R. RUSHFORTH, AND J. R. JOHANSEN. 1983. Effects of long-term grazing on cryptogam crust cover in Navajo National Monument, Arizona. Journal of Range Management 36:579-581. EG&G ENERGY MEASUREMENTS. 1988. The occurrence and status of candidate species listed by the U.S. Fish and Wildlife Service on Naval Petroleum Reserve #1, Kern County, California. Report No. EGG 10617- 2015, Santa Barbara Operations, Goleta, CA. . 1992. Endangered species program, Naval Petro- leum Reserves in California—Annual report FY91. Report No. EGG 10617-2131, Santa Barbara Opera- tions, Goleta, CA. Heapy, H. F 1977. Valley grassland. Pp. 491-514 in M. G. Barbour and J. Major (eds.), Terrestrial vegetation of California. J. Wiley and Sons, New York, NY. HOLLAND, R. F 1987. What constitutes a good year for an annual plant?—Two examples from the Orcuttieae. Pp. 329-333 in T. S. Elias (ed.), Conservation and management of rare and endangered plants. Proceed- ings from a conference of the California Native Plant Society. Sacramento, CA. Mayor, J. 1977. California climate in relation to vegeta- tion. Pp. 11—74 in M. G. Barbour and J. Major (eds.), Terrestrial vegetation of California. J. Wiley and Sons, New York, NY. O’FARRELL, T. P., G. D. WARRICK, N. E. MATHEWS, AND T. T. KATo. 1987. Report of endangered species stud- ies on Naval Petroleum Reserve #2, Kern County, © California. EG&G Energy Measurements Report No. EGG 10282-2189, Santa Barbara Operations, Goleta, CA. PRESTON, W. L. 1981. Vanishing landscapes: Land and life _ in the Tulare Lake Basin. University of California | Press, Berkeley, CA. SAS INSTITUTE INc. 1990. SAS/STAT User’s Guide, Ver- | sion 6. Cary, NC. TAYLOR, D. W. AND W. B. DAVILLA. 1986. Status survey | for three plants endemic to the San Joaquin Valley, | California. Report No. J-235, U.S. Fish and Wildlife Service, Sacramento, CA. U.S. FISH AND WILDLIFE SERVICE. 1990. Endangered and | threatened wildlife and plants; determination of en- | dangered or threatened status for five plants from the — Southern San Joaquin Valley. Federal Register 55: _ 29361-29370. MapRONO, Vol. 45, No. 4, pp. 301-309, 1998 COAST LIVE OAK REVEGETATION ON THE CENTRAL COAST OF CALIFORNIA ANUJA PARIKH AND NATHAN GALE FLx, 1215 Bajada, Santa Barbara, CA 93109 ABSTRACT As part of a revegetation program for Quercus agrifolia Nee, we examined the reported effect of nurse plant facilitation on seedling establishment and growth. To investigate this location effect, acorns were planted directly in the ground in 100 positions under shrubs, and in 100 positions in the open. In addition, we tested for the effect of protection by covering half the planting positions with cages. In the first year, acorns planted in the open had higher germination but lower survival than acorns planted under shrubs, resulting in no significant location effect on the success of seedling establishment. The protection effect was significant, with the success of caged seedlings almost double that of uncaged seedlings. After two years, no location effect on seedling survival or growth was found. Cages continued to have a significant positive effect on seedling survival, but they tended to retard their growth. Over five years of monitoring, no significant effect of nurse plant facilitation on seedling survival or growth was found, although the number of seedlings that survived under shrubs was greater than that in the open. We also explored potential relationships of associated vegetation type and crowding on acorn seedling development. Seed- ling establishment, survival, and growth were associated with differences in vegetation type, and were higher in planting sites with more mesic vegetation. Crowding due to multiple seedlings growing in one planting position versus single seedlings did not negatively affect the growth of seedlings. Of 100 nursery- grown seedlings transplanted in the field in the first year, 99 were surviving at the end of the five-year monitoring period, and on average, they were much larger than the direct acorn plantings. As with the acorn seedlings, no significant nurse plant location effect was found for seedlings transplanted in the open or under shrubs. Interest in the ecology, recruitment, and revege- tation success of Quercus agrifolia Nee var. agri- folia (coast live oak) in California has heightened in recent years. A primary issue of ecological and conservation concern is that the extent of oak woodland and savanna habitats has decreased due to causes that are not well understood, but may in- clude herbivory, competition, drought stress, cattle grazing, and development. The effects of these fac- tors have been examined for California oaks, par- ticularly Quercus douglasii Hook. & Arn. (blue oak) (Griffin 1971, 1976; Muick and Bartolome 1987; Borchert et al. 1989; Davis et al. 1991), but considerably fewer studies have concentrated on Q. agrifolia. Since there appears to be a net loss of oak populations statewide, the replacement of oak seedlings has been required as mitigation for de- velopment projects that result in the removal of adult oak trees. In many of these projects, the re- vegetation success with oaks has been poor. To un- derstand fully the reasons for the lack of natural oak recruitment and the limited success of revege- tation efforts, long-term studies of 10 or more years would be required; realistically, such long-term Studies are seldom possible and systematic data be- yond two years are rare. In a two-year study on coast live oak establish- ment in Central California, Callaway and _ D’Antonio (1991) addressed the question of “‘nurse plant” interactions with Q. agrifolia seedling sur- _ Vivorship and found that seedlings grown under shrubs had higher survival rates than seedlings growing in open areas. They suggested that shrubs may reduce environmental stresses on young oak seedlings and provide protection from herbivory. Other researchers found that Q. agrifolia seedling survival was facilitated by shade and protection by caging (Muick 1991; Plumb and Hannah 1991). As part of an environmental mitigation program, the U.S. Air Force was required to compensate for the loss of mature oaks that occurred during the construction of new facilities, roads, and railroads on San Antonio Terrace at Vandenberg Air Force Base (AFB). The project goal was to establish a total of 70 viable Q. agrifolia seedlings by the end of the five-year mitigation and monitoring program in 1995. Revegetation was carried out by growing oak seedlings from acorns planted directly in the ground (acorn seedlings), and by transplanting nursery-grown seedlings at the field site (transplant seedlings). We collected survival and growth data consistently over five years for both acorn and transplant seedlings, allowing us to quantify losses that took place during seedling germination and es- tablishment phases. In this paper, we examine the location effect of nurse plant treatments on the sur- vival and growth of the acorn and transplant seed- lings. In addition, we investigate the protection ef- fect of caging to explore the role of seed predation/ herbivory in the early phases of acorn seedling es- tablishment. Finally, we summarize and compare the survival and growth patterns of all seedlings over five years. 302 Study site. Vandenberg AFB occupies almost 40,000 ha north of Point Conception on the Central Coast of California. The San Antonio Terrace (34°49'N, 120°35'W) is located in the northern part of the base, between San Antonio Creek to the south and Shuman Creek to the north. The Terrace is ecologically important since much of its area comprises a unique ecosystem of stabilized sand dunes. Dune slopes and ridges are covered by coastal dune scrub vegetation dominated by Eri- cameria ericoides (Less.) Jepson (goldenbush) and Artemisia californica Less. (California sagebrush). Many interdune swales contain wetlands that sup- port a number of different plant communities, in- cluding a variety of marsh and woodland vegetation types. The woodlands are dominated primarily by Salix lasiolepis Benth. (arroyo willow), but in sev- eral of the dune swales, Q. agrifolia is the dominant canopy species. The area chosen for the oak revegetation project was located west of Live Oak Springs, a wetland of approximately 2 ha that was created on San An- tonio Terrace as part of the environmental mitiga- tion program. The site is within a swale near estab- lished oaks. Soils supporting the oaks are sandy, well drained, and have an ample supply of subsur- face water. The presence of mature Q. agrifolia in- dicated environmental conditions favorable for the growth of this species; additionally, the shrubby habitat in the area provided some protection for seedlings from browsing animals. No cattle grazing occurs at the site, but Odocoileus hemionus (mule deer) and Sylvilagus bachmani (brush rabbits) are common in the area, and Sus scrofa (feral pigs) are known to be found nearby. Small mammals that prey on acorns, such as Neotoma fuscipes (wood rats), Thomomys bottae (pocket gophers) and Sper- mophilus beecheyi (ground squirrels), are present at the site, although not in large numbers. METHODS Planting and field methods. Oak acorns were col- lected from Q. agrifolia on San Antonio Terrace in November and December 1990. To ensure genetic diversity among acorns, several different collection locations were selected. Since live oak acorns re- quire a 30—90 day cold stratification period before they germinate (Schopmeyer 1974), the acorns were stored in a refrigerator until direct planting was carried out in February 1991, during the winter rainy season. Although only 70 seedlings were re- quired for mitigation, many more acorns and seed- lings were planted to compensate for: 1) potentially low viability of collected acorns; 2) mortality of seedlings resulting from drought; and 3) potential losses of acorns and seedlings due to herbivory. Locations for planting were chosen within the elevational zone of occurrence of neighboring ma- ture oaks. For the acorn plantings, 100 locations were chosen under separate shrubs, and 100 loca- MADRONO [Vol. 45 tions were chosen about 1 m away from each shrub in open areas. The shrub species used as nurse plants were E. ericoides, A. californica, Baccharis pilularis DC. (coyote brush), and Toxicodendron diversilobum (Torrey & A. Gray) E. Greene (west- ern poison oak). A total of 800 acorns were planted, four in a hole at each location, placed sideways just below the surface of the soil. Fifty randomly se- lected positions from each treatment (shrub/open) were covered with protective cages constructed of — hardware cloth (mesh size about 1 cm?) and shaded with 50 percent shadecloth. Cages were approxi- mately 30 cm in diameter and 45 cm in height. To assure germination and to maintain emerging seed- lings, all positions were hand-watered once a week | in the initial months after planting, but less fre- quently thereafter. Six hundred acorns were sent to a nursery to grow into seedlings. The nursery used a soil mix that included 40 percent sand, 3 percent redwood bark, and peat moss. A slow release fertilizer (19- 6-12 nitrogen-potassium-phosphorus) was added, along with small amounts of other nutrients. In ad- | dition, a pH buffer of dolomitic lime maintained | the mix at a pH of approximately 6 to 6.2 (sandy soils on the Central Coast are generally acidic). The acorns initially were planted in flats. When tap roots of seedlings reached the bottom of the flats, _ the seedlings were transplanted into 1-gallon con- © tainers, where lateral root systems could develop. | Transplanting to the 1-gallon containers occurred in — April 1991, approximately 60 to 90 days after ger- mination. In November 1991, 100 of 401 available | nursery-grown seedlings were transplanted in the field (hereafter referred to as transplant seedlings) at the same site as the acorn seedlings. Approxi- | mately half the transplants, chosen randomly, were located under shrubs, and half in open areas (an | exact 50-50 split was not possible due to site con- | straints). Half were protected by cages similar to | the acorn plantings, but these transplants were not | selected completely at random because some seed- ; lings already were too large to be caged. Due to | this size bias, the protection treatment for the trans- plant seedlings was not analyzed statistically. As seedlings outgrew the cages, they were re- | moved or replaced by larger cages starting in Sum- | mer 1992, and continuing each year. All cages were removed in Fall 1995 at the end of the mitigation | program. A drip irrigation system was installed in | Fall 1991, with water emitters located at sites where ° seedlings survived from acorn plantings, and at all. transplant locations. The seedlings were watered | once per month in drier seasons (June—November). | Irrigation ended in fall/winter each year, as soon as normal precipitation began, and was discontinued | altogether in Fall 1995 at the end of the mitigation program. Neither the acorn nor the transplant seed- | lings received any artificial fertilization treatments | after planting on San Antonio Terrace. The lack of | long-term supplemental water and nutrients was de- | 1998] signed to encourage seedlings to adapt to prevailing rainfall conditions and nutrient sources in their nat- ural environment. Data collection and analysis. Seedling survival was monitored approximately every two weeks in the first four months after acorn seedlings germi- nated in April 1991, then monthly through August 1992, and then annually through 1995. Transplants were monitored after November 1991 at the same time as the acorn seedlings. Survival data included counting seedlings and documenting their condi- tion. Measurements of height, stem diameter, and leaf number were taken in the summers of 1991, 1992, and 1993 as indicators of growth of surviving acorn seedlings. Only height and stem diameter measurements were taken for transplant seedlings, since measuring leaf number would not have been time-effective. In 1994 and 1995, only height and stem diameter measurements were taken for all seedlings. Survival and growth data were analyzed using contingency table analysis and two-way analysis of variance (ANOVA) respectively, testing for the ef- fects of location (plantings under nurse shrubs or in the open) and protection (cages present or ab- sent) on the acorn seedlings. Additional factors ex- amined included associated plant species or vege- tation type, and the potential effects of crowding due to multiple seedlings growing in a single po- sition. For the transplant seedlings, only the effect of location was examined using ANOVA. Statisti- cal analyses were carried out using CSS:STATIS- TICA (StatSoft, Inc. 1991). RESULTS Acorn seedling germination and survival—Year 1. Using data collected to 18 October 1991, percent germination, seedling survival, and seedling suc- cess were assessed for the location and protection treatments. Percent germination refers to the max- imum number of acorns (out of the four planted) that germinated at any time. Percent survival is the proportion surviving after germination to a partic- _ular time. Percent success is the proportion of plant- ed acorns that germinated and survived (germina- tion multiplied by survival). The germination, survival, and success data were examined using contingency table analysis. For germination, the effect of location was not statisti- cally significant (P > 0.05); the overall mean was 41.8 percent for locations in the open and 33.3 per- cent for locations under shrubs (Fig. 1a). However, a significant effect of protection was found (x? = 16.85, P < 0.01), with a mean germination for caged acorns of 44.3 percent, and a mean of 30.8 percent for acorns without cages (Fig. 1b). Acorn _ germination stabilized at about the 60th day after _ first germination; the highest germination occurred for open and caged acorns, the lowest for unpro- tected acorns planted under shrubs (Fig. 2a). PARIKH AND GALE: COAST LIVE OAK REVEGETATION 303 The survival of acorn seedlings through this pe- riod was higher for seedlings that germinated under shrubs than those in the open, but this difference was not significant, with 77.5 percent surviving un- der shrubs, and 60.9 percent in open areas (Fig. 1a). Caged acorn seedlings survived at a rate almost double that of uncaged acorns; a significant differ- ence, with means of 84.6 percent versus 46.5 per- cent (x? = 26.61, P < 0.001; Fig. 1b). With respect to the success of acorn plantings, there was no significant effect of location (Fig. 1a), but again, a significant effect for protection was found (x? = 39.95, P < 0.001). Mean seedling suc- cess was 37.0 percent for caged seedlings and 19.0 percent for uncaged seedlings (Fig. 1b). The higher germination rate for acorns planted in the open was offset by a lower survival rate; and conversely, the lower germination for acorns under shrubs was compensated for by a higher survival rate over time (Fig. 2b). Therefore, overall success was not sig- nificantly different between acorn seedlings under shrubs and those in the open. This difference is most apparent when comparing the success of caged seedlings under shrubs and those in the open (Fig. 2c). In addition to the effects of location and protec- tion, we also examined the potential effect of spe- cies of nurse plant on percent germination, survival, and success. The species factor could not be con- trolled equally due to physical constraints of the planting area, however, the dominant shrub species at each planting location was recorded. Four major shrubs were present: A. californica, E. ericoides, T. diversilobum, and B. pilularis associated with an understory of Carex praegracilis W. Boott (clus- tered field sedge). Although beneath-shrub sample sizes were not appropriate for robust statistical analysis, a consistent seedling germination-surviv- al-success pattern was observed, with Baccharis/ Carex sites being followed by planting sites dom- inated by Toxicodendron, Ericameria, and Artemis- la (Fig. Ic). To summarize Year 1 results, germination was maximum for caged acorns, particularly in the open, suggesting that these acorns were protected from herbivory and may have received extra soil moisture due to the shadecloth covering the cages; moreover, increased sunlight in open conditions may have facilitated seedling germination. After germination rates leveled off, survival was better for protected seedlings and those under shrubs; pro- tection from above-ground herbivory and good soil moisture retention apparently was provided both by cages and by shrubs. Acorns germinated and seed- lings survived better in the Baccharis/Carex sites than in sites dominated by other shrub species. The higher success of seedlings in Baccharis/Carex sites may reflect higher soil moisture conditions, since these species occupy mesic environments such as wetland swales on the San Antonio Terrace. 304 a. Location Effect 7 | See eaeN i ee Saar | ( ier ee ee my a | |e a — ot im I\ eae Y i Md N YG, SY Seedling Location [| Germination [SS Survival Success c. Associated Species Effect | | s| \\— xf $—_$- WV i S\N s7T NN -N | N@ > oft —N-—_Nt _4N— NZ LN ON] |Nev_| INA 30 f Nt =N NA ING | ANL__fFINL_| IN@L_J IN@L wala Va WT NE ot NA NZI ING INF NAF NZ NE een er cenes,Bactherstcare [_] Germination Survival Success HG... of location, protection, and associated species. Acorn seedling survival and growth—Year 2. At the end of Summer 1992, the effects of location and protection on the number of surviving acorn seed- lings and their growth were examined. After this time, the removal of cages from seedlings that out- grew them precluded further experimental testing and analysis of the cage protection effect. The acorn seedlings had been sown in February and germinated by April 1991; therefore, they were ap- proximately 6 months old during first measure- ments in 1991, and 15 months old during the next period of measurements in 1992, having progressed by then through two growing seasons in the field. As in Year 1, the effect of the location treatment on seedling success in Year 2 was not found to be significant. With respect to the survival of acorn plantings, 45 positions in the open had live seed- lings in 1991, decreasing to 40 in 1992. Under shrubs, 49 positions with live seedlings in 1991 de- creased to 43 in 1992. A greater difference in in- MADRONO [Vol. 45 b. Protection Effect MMMM SV 1 | : LLL miei Yl WY ZA ?) fe} (o} oO Q: co) je) 3 WW esas Sai) \/ an) 7 Seedling Protection [__] Germination Survival Success Percent germination, survival, and success of acorn seedlings from data collected 18 October 1991: the effects dividual seedling mortality was observed, with seedling losses between 1991 and 1992 in the open being three times that under shrubs: 24 of 113 seed- lings observed in 1991 were lost in the open, but only 8 of 106 seedlings under shrubs. The effect of the protection treatment in Year 2 continued to be significant (x? = 14.66, P < 0.01). High germination and survival rates for caged po- sitions through 1991 resulted in 146 acorn seed- lings, while only 73 survived in uncaged positions. By 1992, seedling survival remained better in caged positions compared to uncaged positions, with 121 seedlings in 54 caged positions versus 66 seedlings in 29 uncaged positions. ANOVA was used to examine growth measure- ments taken of acorn seedlings in the first two years. Height and stem diameter of seedlings plant- ed in the open or under shrubs did not show a sig- nificant location effect. Likewise, the effect of pro- tection by caging on the growth of acorn seedlings a. Percent Germination of Seedlings Over Time Open/Cage Shrub/Cage ~< Open/No Cage =s Shrub/No Cage Percent Germination ey ° T — 1 ul 1 = 1 al al 1 1 100 120 140 160 180 200 220 240 260 280 300 4/91 10/91 b. Percent Survival of Germinated Seedlings Over Time Shrub/Cage + Open/Cage Percent Survival Shrub/No Cage Open/No Cage 4 lg 1 1 = 1 1 1 L 1 100 120 140 +160 180 200 220 240 260 280 300 4/91 10/91 c. Percent Success of Seedlings Over Time — Open/Cage Shrub/Cage Open/No Cage oS Shrub/No Cage Percent Success iq 1 | 1 N us be t. \ L at 100 120 140 160 180 200 220 240 260 280 300 4/91 10/91 Day Number in 1991 (January 1 = Day 1) Fic. 2. Percent germination, survival, and success of acorn seedlings with respect to location and protection, April to October 1991. was not apparent after the first growing season (Figs. 3a and 3b). Subsequently, absolute changes in height and stem diameter were substantially greater for the non-caged seedlings, which doubled in size by 1992. In contrast, the caged seedlings a. Acorn Seedling Height by Protection, 1991-1992 22 Wefan he : ae a | ar |e | aaa ane eT ae = ar || eae! 12 eee IN) Hire NG ee ff (Ve INE NE INET IN INE se IN— IW INE [__] Not Caged Caged al Fic. 3. Growth of acorn seedlings for the protection effect, PARIKH AND GALE: COAST LIVE OAK REVEGETATION 305 lagged in growth, resulting in a significant differ- ence in both growth parameters for the cage pro- tection effect in 1992 (height: F, .., = 17.93, P < 0.0001; stem diameter: F, ,,; = 23.55, P < 0.0001). The cages apparently tended to retard the growth of seedlings. For the live acorn seedlings in 1991 and 1992, data were collected on leaf number of seedlings at the same time that the other two growth parameters were measured. In 1991, mean leaf number was 7.34, and by 1992 it was 17.24—an increase of about 135 percent. In both years, the effects of lo- cation and protection were significant: leaf number was higher for seedlings located in the open and not caged. For seedlings growing under shrubs or caged, leaf number was lower in both years, likely due to the effect of shading by shrubs or the shade- cloth around the cages. Acorn seedling survival and growth over five years. In early 1991, the mean germination for acorns planted in the field was about 38 percent, yielding 249 seedlings in June, a seedling estab- lishment success of 31 percent. The initial mortality of acorn seedlings was 12.0 percent between June and October 1991. Mortality rates were 14.6 per- cent between 1991 and 1992; 12.3 percent between 1992 and 1993; 9.8 percent between 1993 and 1994; and 6.8 percent between 1994 and 1995. There were 32, 23, 16, and 10 seedlings lost in these four periods, respectively. By July 1995, 138 live acorn seedlings were recorded (Fig. 4a), a five- year seedling success of 17 percent. Of the 200 planting positions, 72 had at least one live seedling, a success by position of 36 percent. At various times over the course of the project, the oak seedlings were subject to browsing, mold attacks, woolly oak aphids, leaf mining, and drought stress. These factors apparently resulted in the mortality of acorn seedlings. We also noted in b. Acorn Seedling Stem Diameter by Protection, 1991-1992 Mean Stem Diameter in cm LL [__] Not Caged Caged 1991 to 1992. 306 a. Seedling Survival, 1991-1995 260 240 220 » 200 fe.) £ 180 ae} F ® 4160 Acorn Seedlings 77) g 140 =e i20 e Transplant Seedlings @ 100 Be | E 80 Zz 60 40 20 0 days 0 200 | 400 | 600 | 800 | 1000 | 1200 ! 1400 | 1600 month/year 5/91 12/91 6/92 1/93 7/93 2/94 8/94 3/95 Time After Planting Acorns on 5 February 1991 Fic. 4. the field that heavy understory cover in the planting positions, both of native and non-native species, tended to have a negative effect on the survival and growth of seedlings in initial years. The understory species probably grew rapidly in response to irri- gation of the planting positions in dry months; we observed that clearing the planting positions did ap- pear to facilitate seedling growth, although this fac- tor was not tested experimentally. With respect to the effect of location on acorn seedling survival, the number of planting positions under shrubs with live seedlings only slightly ex- ceeded the number in the open during the five years of the project. Initially, higher germination led to a greater number of individual seedlings in the open in 1991, but these seedlings suffered higher mor- tality in later years (Fig. 4b). By 1995, therefore, the absolute number of seedlings surviving under shrubs was greater than that in the open (79 versus 59). This difference, however, was not found to be statistically significant (x? = 5.23, P > 0.10). As discussed previously for the first two years, patterns of acorn seedling growth, analyzed sepa- rately each year for years 3 through 5, showed no significant effects of location, with mean height and stem diameter measures being comparable for seed- lings growing in the open and under shrubs. In contrast to the lack of an effect of nurse plant location on the growth of oak seedlings, we ob- served in the field that acorn seedlings growing in the Baccharis/Carex sites were considerably larger than those elsewhere. We therefore combined growth and survival data for the planting sites as- sociated with the shrub species Artemisia, Erica- meria, and Toxicodendron, and compared them to data from the Baccharis/Carex sites. We found that from 1992 onwards, both height and stem diameter were significantly greater for acorn seedlings in the Baccharis/Carex sites (Figs. 5a and 5b). By 1995, means for height were 33.8 cm versus 55.5 cm for MADRONO [Vol. 45 b. Acorn Seedling Survival by Location, 1991-1995 Under Shrub Number of Live Seedlings Number of Positions With Live Seedlings 1992 [_] Open KX Under Shrub Seedling survival, and acorn seedling survival by location, 1991 to 1995. the two vegetation types; and stem diameter means were 0.6 cm versus 1.1 cm (height: F, j3, = 13.42, P < 0.001; stem diameter: F,,;, = 21.80, P < 0.0001). A further examination of survival of acorn seed- lings in the Baccharis/Carex vegetation type showed that although the number of planting po- sitions and absolute number of live seedlings were smaller in the first two years, these seedlings sur- vived considerably better over the longer term than those planted in the other sites (Fig. 5c). By 1995, the number of positions in each type was similar, but 82 live seedlings survived in the Baccharis/ Carex sites, compared to 56 in the Artemisia/Eri- cameria/Toxicodendron sites. This finding indicat- ed that more multiple seedlings were present by 1995 in the Baccharis/Carex sites, while more self- thinning occurred at the other sites. Acorn seedling self-thinning. The issue of thinning the oak acorn seedlings in reference to positions with multiple seedlings arose while evaluating the reve- getation program in later years. To assess for pos- sible crowding effects, survival and growth data were analyzed for the years 1991 to 1995. The num- ber of positions with multiple seedlings decreased over the years, indicating that self-thinning had oc- curred, particularly where 3 or 4 seedlings were present (Fig. 5d). Analysis of the growth data re- vealed no significant crowding effects on height or stem diameter over each of the five years, and mul- tiple seedlings (2 to 4) growing in a single position were not significantly different in size than single seedlings. Moreover, no observable differences in health of the single versus multiple seedlings were noted in the field. Therefore, we decided not to thin the seedlings at the end of the monitoring period, and to allow natural mortality to continue. Transplant seedling survival and growth over five years. Oak seedlings grown in the nursery ger- a. Acorn Seedling Height by Vegetation Type, 1991-1995 | Mean Stem Diameter in cm 5 40 | \ 3 30 Ky N VN NF > 20 = N N V) Ni | IN NAN IN »T =e TINT IN | IN | ISL! INU INU (0) N Aw BS iA fe NI A 1991 1992 1993 1994 1995 [_] Artemisia/Ericameria/ Baccharis/Carex Toxicodendron c. Acorn Seedling Survival by Vegetation Type, 1991-1995 ae) o @ Baccharis/Carex 3 xe) 8 Artemisia/Ericameria/ E Toxicodendron Zz Number of Positions With Live Seedlings 1991 1992 1993 1994 1995 [__] Artemisia/Ericameria/ KXQ Baccharis/Carex Toxicodendron Fic. 5. minated by April 1991, and thus were approxi- mately the same age as the acorn seedlings. One hundred of these seedlings were transplanted in the field in November 1991, after having progressed through one growing season in the nursery. Of the 100 transplants, only one seedling (planted in the open and uncaged) died (in March 1992) through- out the five-year monitoring period (Fig. 4a). Since seedling mortality was negligible, there clearly was no evident effect of location or protection on the survival of transplant seedlings. By the time the nursery-grown seedlings were transplanted in the field in 1991, they already were much larger than the acorn seedlings. In fact, some of the transplants were too large to be caged, and the protection treatment could not be applied at ran- dom, therefore, the effect of caging on growth was not examined. Height and stem diameter measure- ments were taken at the time of transplanting to provide baseline growth data. In 1991, mean height of transplant seedlings was 45.6 cm and mean stem diameter was 0.81 cm, compared to a mean height of 6.7 cm and a mean stem diameter of 0.14 cm PARIKH AND GALE: COAST LIVE OAK REVEGETATION 307 b. Acorn Seedling Stem Diameter by Vegetation Type, 1991-1995 LLM 1 VM ooo 8S 8g Ny ow FAN AHN ®@B © So 8 LS F SS [__] Artemisia/Ericameria/ Toxicodendron Baccharis/Carex d. Frequency of Positions by Number of Live Acorn Seedlings Per Position NANNAAANARARARRRRARARARAAANY INN] NAAN ZLLLA Number of Positions pe) fo) 7 YY Pa’ 7 cx 4 VILLLLLLLLLLLLLAL v7 ras WUMMMMUIM111/17) NANAAAANAAAANAAAANNNNANANNARAARARARNY VZLLLLLLLLLLLLLLLLL LA) | VZZLLILILILELLLELELEL LA RT v, Bx A 94 ra ra ra CX © LLLALLALLLLLLLLLLLL = © © =k = 3 m respectively). We used position of logs and their state of decay | to determine the species present in the canopy be- ° fore the gap was created (the gap-forming species). | Ten of the sampled gaps had been formed by more | than one canopy species. In all cases, the removal | of one species created most of the gap’s area, and | this species was considered the gap-forming spe- | cies. We considered the tree species present in a gap’s » uppermost vegetation layer to be the species filling | that gap. (For 78 of the 80 gaps, the sapling or shrub layer was uppermost.) Twenty gaps had more | than one tree species in the uppermost layer. In | these gaps, the species with the greatest cover in the upper layer, we identified as the species filling - the gap. Where an angiosperm and a conifer species | were both in the uppermost layer, the conifer could | have been considered the gap-filling species, even if at a lower cover, because of its greater maximum | size and potentially faster growth (McArdle et al. 1949; Porter 1965). However, in no gap was Se- quoia both in the uppermost layer and at a lower. cover than an angiosperm in that layer, and in only | three gaps was Pseudotsuga both in the uppermost layer and at a lower cover than an angiosperm spe- cies. If Pseudotsuga had been considered the spe- cies filling those three gaps, the results would have. 6 ed 35 e - Below Canopy ° = 30 ° Below Gap ° ‘ ie) 2 % 25 2 é Pune : OD 20 e a e e e . Ke) ° : : 5 15 Fe ° ‘ . : ® ae ° E O e SO Mae e © : © 10 S60 ; : oe jae .° r g e e . 7 SF ° ‘ ° een tae ee 0 e 0 e t 0 30 60 90 120 150 180 Aspect (Degrees from North) Fic. 1. Aspect and direct light reaching the understory for 80 locations in the watershed of Maddock Creek, Santa Cruz Co., CA. Percent direct light is percent of light in- cident upon canopy that reaches understory. remained similar, and their interpretation would not have been altered. RESULTS Understory light environments. For the year and the watershed as a whole, total light reaching the understory was about 12% of that incident upon the canopy. Direct light averaged 12.5 + 7.3% and ranged from 0.1 to 35%. Indirect light had a com- parable average (11.8%), but less variation (1 SD = 3.6, range 4.8—21.0%). More light reached points beneath canopy gaps (n = 10, direct light mean 16.7 + 9.2%, indirect light mean 13.7 + 3.6) than beneath the canopy (n = 70, direct light mean 11.9 + 6.9%, indirect light mean 11.6 + 3.6%; Mann-Whitney U, P = 0.03 and 0.02, respectively). However, light levels at gap locations were not disjunctly higher than light lev- els at canopied locations. The range of light levels within gaps fell within the range of light levels at _canopied locations (8.1—33% and 0.1—35% direct light, respectively). | TABLE 1. HUNTER ET AL.: UNDERSTORY LIGHT AND GAP DYNAMICS 3 Understory light environments also differed be- tween aspects (Fig. 1). Twice as much direct light reached the understory of south-facing slopes (90— 270°, n = 38, mean = 16.9 + 7.1%) than of north- facing slopes (270—0—90°, n = 42, mean = 8.4 + 5.3%; Mann-Whitney U, P < 0.001). Twenty-two percent more indirect light reached the understory of south-facing slopes (13.8 + 3.4%) than of north- facing slopes (10.8 + 2.9%; Mann-Whitney U, P < 0.001). The higher indirect light levels on south- facing slopes indicate a more open canopy. Interestingly, species composition of the canopy also influenced understory light levels. On south- facing slopes, canopy species composition was patchier and separated into two distinct canopy types: (1) a Pseudotsuga-hardwood type with emer- gent Pseudotsuga above a layer of angiosperms dominated by Arbutus and the Quercus species, and (2) an upland redwood type with emergent Sequoia above a layer of Lithocarpus. Light levels beneath the Pseudotsuga-hardwood canopy (22.4 + 8.4%, n = 7) were substantially higher than beneath the Sequoia-Lithocarpus canopy (12.6 + 6.4%, n = 7; Mann-Whitney U, P = 0.05). This difference be- tween canopy types was comparable to that be- tween canopied and gap locations, and indicates that local variation in canopy composition can af- fect understory light environments significantly. These two canopy types covered 37% of south-fac- ing locations, and the remaining locations were in- termediate in canopy structure and light environ- ment. There was a correspondence between light levels and distribution of tree species in the understory. For Arbutus, Pseudotsuga, and Quercus, understory saplings in the shrub layer occurred at significantly high light levels (Table 1). Sequoia saplings in the shrub layer were not found at high light levels, but established seedlings in the herbaceous layer were found at significantly high light levels (mean = 20.5 + 7.0% direct light). This result is consistent with the species’ biology: saplings of Sequoia are MEAN LIGHT ENVIRONMENTS FOR UNDERSTORY INDIVIDUALS OF CANOPY TREE SPECIES. N is the number of locations (out of 80) at which species were present in a given vegetation layer. Means (+1 SD) are of percent of light incident upon the canopy that reaches the understory. Asterisks denote an average significantly higher than locations _ without the species; * P < 0.05, ** P < 0.01 (Mann-Whitney U test). Species N Shrub layer Arbutus menziesii 2 Quercus chrysolepis 4 Pseudotsuga menziesii 6 Lithocarpus densiflorus 69 Sequoia sempervirens 22 Herbaceous layer Sequoia sempervirens 5 Quercus chrysolepis ° Pseudotsuga menziesii > Lithocarpus densiflorus 24 Direct light ave. Indirect light ave. 25:6 =: 8.6" 7A S12 22.5 =. O1* 16.0 = 2.9* 19 GO 227/075 145° 3167 1 ey mses es 1 EY case ns LO, 9° 7.0 10.0 + 3.4 20)5. 227 OF 17.0 = 14.7 se a apes OG as 14.2 + 4.5 | De ena fo) 126° 24 12.0: 84 } A aaa Fs: 4 MADRONO North-facing South-facing z \N ; a N Arbutus Litho. Pseudo. Quercus Sequoia Fic. 2. Frequency of tree species beneath canopy gaps on north-facing (n = 36) and south-facing (n = 44) as- pects of Maddock Creek’s watershed, Santa Cruz Co., CA. Arbutus = Arbutus menziesii, Litho. = Lithocarpus den- siflorus, Pseudo. = Pseudotsuga menziesii, Quercus = Quercus chrysolepis and Sequoia = Sequoia sempervi- rens. shade-tolerant and persist in the understory for de- cades, but establishment of seedlings is strongly af- fected by light levels (Jacobs 1987). Only Litho- carpus, by far the most abundant and widespread tree in the understory, had no correspondence be- tween light levels and seedling or sapling distri- bution. Gap dynamics. Most tree species were absent from a large portion of gaps (Fig. 2). Only Litho- carpus was present in almost all gaps (95%), and at a high cover (mean = 36.4 + 27.4%). Pseudo- tsuga and Quercus were in a large portion of gaps (49 and 45% respectively) but when present were at a low cover (5.5 + 9.4 and 6.1 + 11.6%), while Arbutus and Sequoia were rarely present (15 and 16% respectively) and were at a low cover (mean = 7.5 + 11.2 and 12.3 + 19.5%). Lithocarpus also was filling the most gaps (Fig. 3). Of 80 gaps, Lithocarpus was filling 68%, Pseu- dotsuga 15%, Quercus chrysolepis 9%, Sequoia 5%, and Arbutus 4%. Because Lithocarpus had formed significantly less of the gaps (38%) than it was filling, it was increasing in importance within the canopy (x? test, df = 1, P = 0.0001). Pseudo- tsuga and Arbutus had formed significantly more of the gaps (35 and 14% of gaps, respectively) than they were filling, and thus were declining in im- portance within the canopy (x’ test, df = 1, P = 0.004 and 0.025 respectively). Sequoia and Quer- cus did not have significant differences between the number of gaps formed and the number filled, though an increase in Quercus is suggested (P = 0.086). There was also a relationship between gap-form- ing and gap-filling species. Of gaps formed by Lith- ocarpus (n = 30), Lithocarpus filled 90%, signifi- cantly more than expected (x’ test, df = 1, P = 0.001). This result may be due in part to a signifi- cantly lower presence of other species in gaps [Vol. 46 A. Gap-Forming Species 80 [S55] North-facing Q South-facing © 60 (ren ro) c 3 40 ® oO . Quercus Sequoia Arbutus B. Gap-Filling Species Percent of Gaps Rew RY Arbutus Pseudo. Quercus Sequoia Litho. Fic. 3. Species forming and filling gaps on north-facing (n = 36) and south-facing (n = 44) aspects of Maddock Creek’s watershed, Santa Cruz Co., CA. Species abbre- viations as in Figure 2. formed by Lithocarpus (x? test, df = 1, P = 0.002). Gaps formed by Lithocarpus (n = 30) averaged 1.8 + 0.8 tree species versus 2.4 + 0.9 present below gaps formed by other species (Mann-Whitney U, P = 0.003). Also, Lithocarpus was the only tree spe- cies present below 43% of gaps formed by Litho- carpus, while just 10% of gaps formed by other species had only Lithocarpus present beneath them (x? test, df = 1, P < 0.001). The data also suggest self-replacement by Sequoia. Three of four gaps formed by Sequoia were being filled by basal sprouts of the same tree(s) that had formed the gap. Although there was no relationship between gap area and species filling the gap, there was a rela- tionship between slope aspect and gap-filling spe- cies (Fig. 3). Species less tolerant of shade (Arbu- tus, Quercus chrysolepis, and Pseudotsuga) filled significantly more south-facing gaps (39%) than north-facing gaps (14%; x? test, df = 1, P = 0.014). As a consequence, succession differed between north and south-facing slopes. On north-facing slopes, Pseudotsuga formed significantly more gaps than it filled (y? test, df = 1, P = 0.002) and there- fore declined in importance, while on south-facing slopes there was no significant difference in the number of gaps formed and filled by Pseudotsuga (P = 0.23). The data also suggested differences be- 1999] tween aspects in the dynamics of Lithocarpus and Quercus. DISCUSSION In upland redwood and Pseudotsuga-hardwood forests, tree species differ in shade-tolerance. For example, Quercus species, Arbutus and Pseudotsu- ga are clearly less tolerant than Lithocarpus, Se- quoia and Umbellularia californica (Hook. & Arn.) Nutt. (Waring and Major 1964; Unsicker 1974; Tappiener et al. 1986; Burns and Honkala 1990; Sugihara 1992; Hunter 1997a and 1997b). Together with the influence of fire, these interspecific differ- ences in shade-tolerance probably determine most patterns of sapling recruitment into the canopy. Prior to fire suppression, surface fires would have removed most understory regeneration, including Lithocarpus saplings (Kauffman 1986). This effect would have limited the successful recruitment of understory Lithocarpus into the canopy, while cre- ating seedbed and understory conditions favorable for the other canopy species (Jacobs 1987; Her- mann and Lavender 1990; Hunter 1994). Currently, however, Lithocarpus saplings accumulate in forest understories (Tappeiner and McDonald 1984; Hun- ter 1997a). By affecting sapling establishment prior to gap formation, the transmission of light through cano- pies can influence substantially the species com- position of regeneration (Canham et al. 1994). This was the case in this watershed’s forests. Here, most gaps were filled by saplings that had established prior to the gap’s formation. As a consequence, the most abundant species throughout the understory, the shade-tolerant Lithocarpus, was also the species filling the majority of canopy gaps. Because forest canopies vary on a fine scale in species composition, leaf area and height, light reaching the understory is also variable (Baldocchi and Collineau 1994). In this study, canopied loca- tions received from 0.1 to 35% of the direct light incident upon the canopy. This range of light levels had ecological significance because four of the five canopy species had understory regeneration asso- ciated with higher understory light levels, and be- cause the less shade tolerant species were filling more gaps on south-aspects, where light levels were higher. Interestingly, much of the variation in understory light seems to be due to interspecific differences in Shade cast. In this study, on south-facing slopes, light levels beneath a Pseudotsuga-mixed hard- wood canopy were nearly twice those beneath a Sequoia-Lithocarpus canopy (ave. 20.7% vs. 12.6%), a difference comparable to that between gaps and canopied locations (16.7% vs. 11.9%). In other studies, there is also evidence that the more shade-tolerant species can develop a denser crown and therefore allow less light to pass through to the understory (Waring & Major 1964; Unsicker 1974; Minore 1986; Harrington et al. 1984). HUNTER ET AL.: UNDERSTORY LIGHT AND GAP DYNAMICS ) If much of the variation in understory light is due to interspecific differences in transmission of light through the crown, then this attribute could be an important cause of observed patterns in the stand dynamics in California’s coastal forests. For ex- ample, the self-replacement of Lithocarpus (ob- served in this study) may be promoted by low lev- els of light passing through the crowns of canopy- sized Lithocarpus, allowing advance regeneration of the shade-tolerant Lithocarpus but not of less tolerant species. Similarly, relatively high levels of light passing through the crowns of Quercus spe- cies and Arbutus may contribute to their replace- ment by Pseudotsuga and Umbellularia, which oc- curs in several types of forest and woodland veg- etation (McBride 1974; McDonald and Littrell 1976; Hunter 1995; Safford 1995; Barnhart et al. 1996). For the dominant trees of California’s coastal for- ests, the magnitude of interspecific differences in shade cast, and the influence of these differences upon succession both deserve further investigation. ACKNOWLEDGEMENTS For their assistance, we thank D. Hunter, G. Hunter, R. Pearcy, R. Rousseau, and S. Ustin. For their comments on drafts of this manuscript and related manuscripts, we thank T: Niesen, R. Patterson, R. Rejymanek, J. Sawyer, K. Schierenbeck, R. J. Zazoski, and three anonymous review- ers. LITERATURE CITED AGEE, J. K. 1991. Fire history along an elevational gra- dient in the Siskiyou Mountains, Oregon. Northwest Science 65:188—199. ANDERSON, M. C. 1964. Studies of the woodland light climate. I. The photographic computation of light conditions. Journal of Ecology 52:27-41. BALDOCHI, D. AND S. COLLINEAU. 1994. The physical na- ture of solar radiation in heterogeneous canopies: spa- tial and temporal attributes. Pp. 21—71 in M. M. Cald- well and R. W. Pearcy (eds.), Exploitation of envi- ronmental heterogeneity by plants. Academic Press, New York, NY. BARBOUR, M. G. AND J. Mayor. 1988. Introduction. Pp. 3-11 in M. G. Barbour and J. Major (eds.), Terrestrial vegetation of California, Expanded edition. California Native Plant Society, Sacramento, CA. BARNHART, S. J., J. R. MCBRIDE, AND P. WARNER. 1996. Invasion of northern oak woodlands by Pseudotsuga menziesii (Mirb.) Franco in the Sonoma Mountains of California. Madrono 43:28—45. BECKER, P., D. W. ERHART, AND A. P. SMITH. 1989. Anal- ysis of forest light environments part I. Computerized estimation of solar radiation from hemispherical can- opy photographs. Agricultural and Forest Meteorol- ogy 44:217—232. Burns, R. M. AND B. H. HONKALA. 1990. Silvics of North America. Handbook 654, Forest Service, U.S. De- partment of Agriculture, Washington, D.C. CANHAM, C. D., A. C. FINZI, S. W. PACALA, AND D. H. BURBANK. 1994. Causes and consequences of re- source heterogeneity in forests: interspecific variation 6 MADRONO in light transmission by canopy trees. Canadian Jour- nal of Forestry Research 24:337-349. CHAZDON, R. L. AND C. B. FIELD. 1987. Photographic es- timation of photosynthetically active radiation: eval- uation of a computerized technique. Oecologia 73: 525-532. FINNEY, M. A. AND R. E. MARTIN. 1992. Short fire inter- vals recorded by redwoods at Annadel State Park, California. Madrono 39:251-—262. GREENLEE, J. M. 1983. Vegetation, fire history and fire potential of Big Basin Redwoods State Park. Disser- tation, University of California, Santa Cruz, CA. GRS. 1992. Vegetation data collection using the GRS can- opy-densitometer. GRS Inc., Arcata, CA. HARRINGTON, T. B., J. C. TAPPEINER, AND J. D. WALSTAD. 1984. Predicting leaf area and biomass of 1- to 6- year-old tanoak (Lithocarpus densiflorus) and Pacific madrone (Arbutus menziesii) sprout clumps in south- western Oregon. Canadian Journal of Forest Research 14:209-213. HERMANN, R. K. AND D. P. LAVENDER. 1990. Pseudotsuga menziesii (Mirb.) Franco Douglas-Fir. Pp. 527—540 in R. M. Burns and B. H. Honkala (eds.), Silvics of North America, Vol. 1, conifers. Handbook 654, For- est Service, U.S. Department of Agriculture, Wash- ington, D.C. HICKMAN, J. C. (ed.). 1993. The Jepson manual: higher plants of California. University of California Press, Berkeley, CA. Hunter, J. C. 1989. The pattern of canopy gaps in an old- growth forested landscape. M. A. thesis, San Francis- co State University, San Francisco, CA. . 1994. The litter layer as a barrier to seedling emergence in California’s mixed evergreen forests. Bulletin of the Ecological Society of America 75: $102. 1995. Architecture, understory light environ- ments, and stand dynamics in northern California’s mixed evergreen forests. Ph.D. dissertation, Univer- sity of California, Davis, CA. 1997a. Fourteen years mortality in two old- growth Pseudotsuga-Lithocarpus forests in northern California. Journal of the Torrey Botanical Society 124:273-279. . 1997b. Correspondence of environmental toler- ances with leaf and branch attributes for six co-oc- curring species of broadleaf evergreen trees in north- ern California. Trees 11:169—175. AND V. T. PARKER. 1993. The disturbance regime of an old-growth forest in coastal California. Journal of Vegetation Science 4:19-—24. JAcoss, D. F 1987. The ecology of redwood seedling es- tablishment. Ph.D. dissertation, University of Califor- nia, Berkeley, CA. , D. W. COLE, AND J. R. MCBRIDE. 1985. Fire his- tory and perpetuation of coast redwood ecosystems. Journal of Forestry 83:494—497. KAUFFMAN, J. B. 1986. The ecological response of the shrub component to prescribed burning in mixed con- ifer ecosystems. Ph.D. dissertation, University of Cal- ifornia, Berkeley, CA. MCARDLE, R. E., W. H. MEYER, AND D. BRucE. 1949. The yield of Douglas fir in the Pacific Northwest. Tech- [Vol. 46 nical Bulletin 201, U.S. Department of Agriculture, Washington, D.C. MCBRIDE, J. R. 1974. Plant succession in the Berkeley Hills, California. Madrono 2:317—380. McDOoNnaALb, P. M. AND E. E. LITTRELL. 1976. The bigcone Douglas-fir/canyon live oak community in southern California. Madrofio 23:310-—320. MInorE, D. 1986. Effects of madrone, chinkapin, and tan- oak sprouts on light intensity, soil moisture, and soil temperature. Canadian Journal of Forestry Research 16:654-658. PEARCY, R. W. 1989. Radiation and light measurements. Pp. 97-116 in R. W. Pearcy, J. Ehleringer, H. A. Moo- ney, and P. W. Rundel (eds.), Plant physiological ecol- ogy: field methods and instrumentation. Chapman and Hall, London. PORTER, D. R. 1965. Site index curves for three hardwood species in the redwood region of Humboldt County, California. M.S. thesis, Humboldt State College, Ar- cata, CA. Rice, C. L. 1985. Fire history and ecology of the North Coast Range Preserve. Pp. 367-372 in J. E. Lotan, B. M. Kilgore, W. C. Fischer and R. W. Mutch (eds.), Proceedings of the symposium on wilderness fire. General Technical Report INT-182, U.S. Forest Ser- vice, Department of Agriculture, Washington, D.C. Ricu, P. M. 1989. A manual for analysis of hemispherical canopy photography. Los Alamos National Labora- tory, Los Alamos, NM. SAFFORD, H. D. 1995. Woody vegetation and succession in the Garin Woods, Hayward Hills, Alameda County, California. Madrono 42:470—489. SAWYER, J. O., D. A. THORNBURGH, AND J. R. GRIFFIN. 1988. Mixed evergreen forest. Pp. 359-382 in M. G. Barbour and J. Major (eds.), Terrestrial vegetation of California, Expanded edition. California Native Plant Society, Sacramento, CA. SUGIHARA, N. G. 1992. The role of treefall gaps and fallen trees in the dynamics of old growth coast redwood forests. Ph.D. dissertation, University of California, Berkeley, CA. TAPPEINER, J. C. AND P. M. MCDONALD. 1984. Develop- ment of tanoak understories in conifer stands. Cana- dian Journal of Forest Research 14:271-—277. TAPPEINER, J. C., P. M. MCDONALD, AND T. E HUGHES. 1986. Survival of tanoak (Lithocarpus densiflorus) and Pacific madrone (Arbutus menziesii) seedlings in forests of southwestern Oregon. New Forests 1:43- 55. UNSICKER, J. E. 1974. Synecology of the California bay tree, Umbellularia californica, (H. & A.) Nutt. in the Santa Cruz Mountains. Ph.D. dissertation, University of California, Santa Cruz, CA. WARING, R. H. AND J. MAJsor. 1964. Some vegetation of the California coastal redwood region in relation to gradients of moisture, nutrients, light, and tempera- ture. Ecological Monographs 34:167—215. WHITTAKER, R. H. 1960. Vegetation of the Siskiyou Mountains, Oregon and California. Ecological Mono- graphs 30:279-338. ZINKE, P. J. 1988. The redwood forest and associated north coast forests. Pp. 679-698 in M. G. Barbour and J. Major (eds.). Terrestrial vegetation of California, Ex- panded edition. California Native Plant Society, Sac- ramento, CA. Maprono, Vol. 46, No. 1, pp. 7-12, 1999 THE REPRODUCTIVE BIOLOGY AND HOST SPECIFICITY OF OROBANCHE PINORUM GEYER (OROBANCHACEAE) MARK W. ELLIS* USDA Forest Service, Leavenworth Station, 600 Sherbourne, Leavenworth, WA 98826 RONALD J. TAYLOR Department of Biology, Western Washington University, 516 High Street, Bellingham, WA 98225 RICHY J. HARROD USDA Forest Service, Leavenworth Station, 600 Sherbourne, Leavenworth, WA 98826 ABSTRACT This study is the first examination of the reproductive biology of Orobanche pinorum Hook., a rare, parasitic flowering plant of western North America. Plants from six separate populations were treated to test reproductive strategies, including xenogamy, autogamy, and agamospermy. Seeds that developed from the flowers of treated plants were counted to determine the reproductive success of each strategy. The results showed O. pinorum to be predominantly autogamous, producing a mean of nearly 700 seeds per capsule and over 70,000 seeds per plant. There was some evidence for xenogamy, but potential pollinators were seen only on six occasions in three years. Two were collected—solitary bees Osmia exigua and Ashmeadiella cactorum. No evidence was found for agamospermy. Orobanche pinorum appeared to parasitize only one host—the shrub Holodiscus discolor (Pursh) Maxim. INTRODUCTION Orobanche pinorum Hook. (pine broomrape, Fig. la—d) is in the Orobanchaceae, a family of achlo- rophyllous, obligate root parasites (Kut 1969). An endemic of western North America, this uncommon species has been found from northern California through Oregon, and north to central Washington and Idaho (Munz 1930; Hitchcock et al. 1959; Abrams and Ferris 1960; Heckard 1993). In Wash- ington state, O. pinorum is listed as a ‘‘monitor”’ species due to its relative rarity (Washington Nat- ural Heritage Program 1997). A literature search provided no published studies on this species other than descriptive reports and a single chromosomal count in one specimen (Heckard and Chuang 1975). The flowers of Orobanchaceae are apparently adapted for cross-pollination, displaying traits typ- ically associated with bee pollination—tubular co- rolla complete with landing platform, nectar guides, contrasting colors around the floral entry, and nec- tar rewards (Kuijt 1969). The campanulate corolla of the zygomorphic flowers is yellowish with pur- ple exterior veins which probably serve as nectar guides. The two broad, deltoid lobes of the upper lip form a hood over the stamens and stigma while the three lanceolate lobes of the lower lip provide a potential insect landing platform (Fig. la—c). Reproductive strategies besides xenogamy have * Present address: 858 Buena Vista Street, Moss Beach, CA 94038. also evolved within Orobanchaceae. While the flowers of most species are chasmogamous, cleis- togamy has been reported in Epifagus, Cistanche, and Boschniakia (Thieret 1971; Musselman 1980; Olsen and Olsen 1980). Although Kuijt (1969) ar- gues that insects play the most important role in pollination, the work of Musselman (1980) and Musselman et al. (1981) show autogamy to be com- monly employed in varying degrees. Jenson (1951) found that populations of Orobanche uniflora L. in New England were obligately agamospermous, and Reuter (1986) found that populations of Orobanche fasciculata Nutt. in Wisconsin produced similar numbers of seeds via agamospermy, autogamy, and xenogamy. Species within Orobanchaceae parasitize a vari- ety of herbaceous and woody host species (Mus- selman 1980). The majority of broomrapes appear to have broad host ranges, with more narrow host- specificity found occasionally among physiological races (Musselman 1980). Although most references suggest O. pinorum is hosted by coniferous species (Geyer 1851; Hitchcock et al. 1959; Munz and Keck 1968; Peck 1961), the most recent accounts (Heckard 1993; Smith-Kuebel and Lillybridge 1993) suggest instead the shrub Holodiscus discol- or (Pursh) Maxim. The purpose of this study was to examine the reproductive strategy of O. pinorum, collect and identify any potential pollinators, and examine host specificity. g MADRONO Fic. 1. a. upper % of Orobanche pinorum stalk with flowers and terminal buds. b. frontal view of O. pinorum flower. c. lateral view of O. pinorum flower. d. O. pinorum flower with ventral half of corolla tube excised. e. O. uniflora flower with ventral half of corolla excised. Drawings by Eve Ponder. METHODS Study area. All field research sites were located in Washington state, within the Leavenworth Rang- er District of the Wenatchee National Forest. The TABLE 1. [Vol. 46 sites were on moderate to steep slopes, in lithosolic substrates with gravitational instability and poor horizontal development. The forest communities were generally mixed-coniferous, dominated by Pseudotsuga menziesii (Mirbel) Franco, Pinus pon- derosa Laws., and Abies grandis (Douglas) Lindley with less than 50% canopy cover. The area is un- derlain by the Chumstick Formation, a bedrock composed of middle Eocene materials including micaceous feldspathic sandstone, interbedded peb- bly sandstone, conglomerate, and shale (Tabor et al. 1987). The great soil groups in this region include cryorthods, haplorthods, and xerochrepts (Franklin and Dyrness 1973). The climate is characterized by warm to hot summers (maximum temperatures ap- proaching 38°C) and cold winters (minimum tem- peratures below —18°C) (Donaldson and Ruscha 1975). Mean annual precipitation at the study sites ranges from 64—89 cm with two thirds falling be- tween October and March. The summers are char- acterized by drought, with less than 2.5 cm of rain falling from July through August. Table 1 provides summary descriptions of the study sites. Individuals of O. pinorum from six populations were selected and treated in two reproductive ex- periments. Data from two additional populations were used to estimate the mean number of flowers per plant. In this study, populations of O. pinorum were considered distinct if they were separated by at least a half mile with no individuals found be- tween populations. Data were collected during the growing seasons of 1993, 1994, and 1995. First reproduction experiment. Six treatments were applied to test reproductive strategies of O. pinorum (Table 2). The criteria for selecting plants were health and phenology. In each population, plants were selected if they appeared healthy and showed no visible signs of insect infestation (i.e., caterpillar silk, feces, or herbivory). To ensure that all reproductive structures would be well devel- oped, buds just prior to anthesis were treated. Treat- ments were assigned randomly to each suitable plant, using a calculator’s random function. For each treatment, eight plants from five different pop- ulations (Sites 1-5) were selected with two flower buds treated per plant. In Treatment | (T1), autogamous seed production APPROXIMATE LOCATION, ASPECT, POPULATION SIZE, SLOPE AND ELEVATION OF O. PINORUM Study Sites. All sites were located in Washington state, within the Leavenworth Ranger District of the Wenatchee National Forest. Site # Elev. (m) Slope (°) Aspect Pop. size Location l 720-1100 30-45 >100 Negro Creek 2 880 20-30 16—34 North Ruby Creek 3 830 30 2-5 South Ruby Creek 4 610 10 SW 3-9 Lower Camas 5 990 <10 9-22 Upper Camas 6 1170 30 se 2-3 King Creek 7 520-670 40 >100 W. Spromberg Canyon 8 500-575 40 41 E. Spromberg Canyon 1999] ELLIS ET AL.: OROBANCHE PINORUM POLLINATION 9 TABLE 2. TREATMENT DESCRIPTIONS FOR THE FIRST REPRODUCTION EXPERIMENT. Treatment # Treatment test Tl Autogamy T2 Control T3 Potential Xenogamy T4 Vectoral Xenogamy TS Agamospermy- | T6 Agamospermy-2 was tested by bagging the selected buds in a tight- weave cotton cloth to exclude pollinators and wind- borne pollen. In T2 (Control), selected plants and buds were left unmodified. In T3 (Potential Xenog- amy), buds selected were emasculated and hand- crossed with pollen from other plants to test the potential for seed production via vectoral cross-pol- lination (insects, wind, etc.). In T4 (Vectoral Xe- nogamy), buds selected were emasculated and left open to pollination to test directly for xenogamous seed production. In T5 (Agamospermy-1), buds se- lected were emasculated and then bagged to ex- clude pollinators and wind-borne pollen to test for agamospermy. In T6 (Agamospermy-2), both an- thers and stigmas were excised from the selected buds, and the buds were then bagged to exclude pollinators and wind-borne pollen. This treatment provided a more thorough test of agamospermy, since no stigmatic surface was available for the ger- mination of pollen that the experimental design might fail to exclude from the flower. Once the capsules of the treated buds matured, they were collected prior to dehiscence and stored to dry. Once dry, the capsules were opened and the seeds removed and spread evenly on a grid and counted under 15X magnification. Second reproduction experiment. Four treat- ments were applied to test for the relative success of four potential pollination strategies in O. pino- rum (Table 3). Again, healthy and phenologically suitable plants were selected for treatment, and the treatments were applied randomly for each plant, using a calculator’s random function. Nine plants were selected for each treatment from sites 1—6. As in the first pollination experiment, flower buds that appeared ready to open were selected for treatment. For each treatment, the entire plant was bagged to exclude potential pollinators and wind-borne pol- Treatment procedure Flowers bagged to exclude pollinators Flowers not bagged or modified Flowers emasculated and hand-crossed Flowers emasculated, open to pollination Flowers emasculated, bagged to exclude pollinators Flowers emasculated, stigmas excised, bagged to exclude pollinators len. Capsules were collected and seeds removed and counted as in the first reproductive experiment. In Tl (Xenogamy/Vicinism), selected buds were emasculated and hand-crossed with pollen from flowers on nearby plants in the same population to test the potential reproductive success of vicinism. In T2 (Xenogamy/Non-Vicinism), buds selected were emasculated and hand-crossed with pollen from different populations to test the potential re- productive success of xenogamy between plants less likely to be related. All cross-pollinations in T2 were made between plants from populations separated by at least one mile. In T3 (Potential Gei- tonogamy), buds selected were emasculated and hand-pollinated with pollen from flowers on the same plant to test the potential reproductive success of geitonogamy. In T4 (Geitonogamy), buds from the middle or lower portions of the plant were emasculated and left available for pollination via pollen falling from flowers above them. Although the bag would help prevent wind from blowing pollen from one flower to another, pollen that did exit a flower might fall onto flowers below. For that reason, buds tested in treatments T1—T3 were all located at the tops of plants so that no flowers were directly above them. Unlike the first reproduction experiment, hand- crossing was employed as the pollination method in the treatments T1—T3. In this way, potential vari- ations in results due to the different success rates between hand-pollination and natural self-pollina- tion were averted. Statistics. The research was designed for the two reproductive experiments so that all data could be analyzed by Analysis of Variance (ANOVA) using a completely randomized design with subsampling. The alpha level of significance was preset at 0.05. The mean number of seeds per capsule for O. TABLE 3. TREATMENT DESCRIPTIONS FOR THE SECOND REPRODUCTION EXPERIMENT. All plants bagged to exclude polli- nators. All flowers emasculated. Treat- ment # Treatment test Treatment procedure 4 tI Xenogamy/Vicinism Flowers hand-crossed with pollen from plants in the same population Tr? Xenogamy/Non-Vicinism T3 Potential Geitonogamy T4 Geitonogamy Flowers hand-crossed with pollen from plants in different populations Flowers hand-crossed with pollen from other flowers on the same plant Flowers selected from the middle or lower region of the inflorescence 10 MADRONO [Vol. 46 TABLE 4. MEAN SEED COUNTS PER CAPSULE FOR THE FIRST REPRODUCTION EXPERIMENT. * Treatments significantly dif- ferent at the 0.05 level. Treatment Mean # Tl 12 T3 T4 T5 T6 689 TZ = a sa ee * 248 T3 a a = = = 13 T4 - * - 0 T6 * = pinorum was calculated from the data in T2 (Con- trol) of the first reproduction experiment. The mean number of flowers per plant was calculated from 33 plants over a two-year period. An estimated mean number of seeds per plant was then derived by mul- tiplying the mean number of flowers per plant by the mean number of seeds per capsule. Potential pollinators. Over 100 hours were spent observing plants for insect pollination activities. These observations were made during visits to the sites for data collection, applications of experimen- tal treatments, and fruit collections. Since prelimi- nary observations suggested that there were few if any insect pollinators, no systematic method of ob- servation was employed. The importance of flower visitors was inferred by a combination of factors: the number of observed visitations, pollen load ra- tios, insect behavior, and the results of treatments that tested for xenogamy. Pollen load ratios of in- sect specimens were determined using a dissecting microscope and pollen was identified by comparing the grains with those taken from O. pinorum an- thers. Host specificity. Four specimens of O. pinorum were excavated (including one in Oregon) to iden- tify the host species. Host roots were located and traced back to their sources. Since O. pinorum 1s rare in Washington, Oregon, and California (Jepson 1970; Heckard 1973; Heckard and Chuang 1975; Heckard 1993), the number of excavations were in- tentionally limited in number. RESULTS First reproduction experiment. The mean num- bers of seeds per capsule for each of the six treat- TABLE 5. MEAN SEED COUNTS FOR THE SECOND REPRO- DUCTION EXPERIMENT. * Treatments significantly different at the 0.05 level. Treat- ment Mean # Tl T2 T3 T4 Treatment test 319 Tl * Xenogamy/Vicinism 264 12 * Xenogamy/Non-Vicinism 194 T3 * Potential Geitonogamy Geitonogamy Treatment test Autogamy Control (unmodified) Potential Xenogamy Vectoral Pollination Agamospermy-1 (emasculated only) Agamospermy-2 (emasculated and stigma excised) ments are given in Table 4. The treatment testing for autogamy showed the highest production of seeds, with the control a close second. The hand- crossed treatment testing for potential Xenogamy produced less than half as many seeds as the con- trol and autogamy treatments, but nearly 20 times more than the test for naturally occurring vectoral pollination. Neither treatment for agamospermy produced any seeds at all. Second reproduction experiment. The mean numbers of seeds per capsule for the four treat- ments are given in Table 5. The capsules in Treat- ments Tl (Xenogamy/Vicinism), T2 (Xenogamy/ Non-Vicinism), and T3 (Potential Geitonogamy) did not produce significantly different numbers of seeds from each other, but produced significantly more seeds than capsules in T4 (Geitonogamy). Seed production. The mean number of flowers per plant was 96 (95% Confidence Interval {CI} + 14) and the mean number of seeds per capsule was 689 (95% CI + 126). The estimated mean number of seeds per plant is 71,656. Potential pollinators. During the summers of 1993, 1994, and 1995, insects were seen entering O. pinorum flowers on only six different occasions. All were small hymenopterans. Two were collected and identified as solitary leafcutting bees in the su- perfamily Apoidea, family Megachilidae (Dr. Terry Griswold, bee taxonomist, Utah State University, Logan, UT, personal communication). Ashmeadiel- la cactorum is a small, long-tongued, dark brown bee with light stripes derived from densely packed white hairs at the sulci separating the abdominal tergites. The second species was Osmia exigua, a slightly larger, long-tongued green bee. Just before capture, both bees had been actively and system- atically entering flowers, crawling completely into the corolla tube, then backing out, flying to another | flower, and repeating the process. Osmia exigua could be heard apparently buzz-pollinating the | flowers it entered, suggesting that it was actively collecting pollen. Both bees had a ventrally located abdominal sco- pa where the majority of pollen grains were col- » lected. Pollen examined on the body of A. cactorum © was exclusively that of O. pinorum. The body of O. exigua held two types of pollen, approximately | 1999] 75% of which was from O. pinorum (the other pol- len type was not identified). Host specificity. In all four excavations of O. pi- norum, the host root was traced back to H. discolor. Although species composition varied at O. pinorum sites, all O. pinorum individuals seen in this study were found within five meters of one or more H. discolor plants. DISCUSSION Reproductive strategy. The results suggest pre- dominant autogamy is the reproductive method em- ployed by O. pinorum (although pseudogamy has not been ruled out). Facultative autogamy is a com- mon reproductive strategy for annuals and para- sites, and for species within Orobanchaceae in par- ticular (Holzner 1982; Musselman et al. 1981). Open habitats select for colonizing species, and successful colonizers tend to have restricted recom- bination systems (Grant 1981; Begon and Mortimer 1986). Such colonizers can rapidly multiply and oc- cupy a Site. Although the flowers of O. pinorum show most of the hallmarks of bee pollination, the positioning of their stamens promotes selfing. In O. uniflora, the predictable arrangement of the androecium rel- ative to the stigma would tend to discourage self- pollination and facilitate crossing (Fig. le). The short filaments prevent the anthers from reaching the stigma, and the anthers are located beneath the more or less horizontal style. In this way the stigma is positioned to receive pollen carried from other flowers by the intruding insect pollinator. As the vector continues further, the anthers are positioned to dust it with pollen. In contrast, the maturing an- thers of O. pinorum flowers examined in this study grew toward the stigma, with one or more contact- ing it directly (Fig. 1b, d). Hairs on the pollen sacs became entangled in the sticky fluid of the stig- matic surface, so that when the anthers dehisced, self-pollination was assured. As a result, the ma- jority of ovules were self-fertilized, producing hun- dreds of seeds per capsule without relying on the vagaries of insect vectors. In this study, insect pol- lination accounted for an average of less than 20 mature seeds per capsule while self-pollination pro- duced an average of nearly 700 (Table 4). The parasitic habit of this species is highly spe- cialized, and O. pinorum appears to occupy a niche with few if any direct competitors. One may predict that inbreeding would tend to maintain genetic sta- bility in this parasite, and that predominant autog- amy would in turn be reinforced by the stabilizing selection of its parasitic niche (Grant 1981). Thus predominant autogamy provides a stabilizing mech- anism for this parasite as well as a means of de- pendably high seed production necessary to assure continued contacts with its host. No evidence was found for pollen selection other than first-come-first-served, since pollen from the ELLIS ET AL.: OROBANCHE PINORUM POLLINATION 1a same plants, neighboring plants, and plants in dis- tant populations all produced seeds with equal suc- cess. The apparently homogamous flowers of O. pi- norum ensure selfing, which occurs approximately at (and on rare occasion before) the onset of anthe- sis. With no physiological barrier to xenogamy, O. pinorum retains the potential for cross-pollination and the subsequent infusion of new genetic varia- tion. Apparent pollinators. The recovery of two flow- er-visiting hymenopterans lends support to the po- tential for occasional out-crossing. That the pollen loads of both bees were composed primarily of O. pinorum pollen (entirely on A. cactorum) suggests a specificity that could be the result of a long-term relationship between pollinator and plant. Although the average vectoral seed production shown in this study was very low (13/flower, Table 4), its occur- rence establishes out-crossing as a clear possibility. Still, the evolutionary direction may be toward increasingly restricted recombination. The findings here document that insect visitation is infrequent and likely to be preceded by self-pollination when it occurs. Furthermore, many if not most visits by hymenopteran vectors may result in geitonogamous pollination, as they systematically probe successive flowers on the same plant. The occasional dehiscence of O. pinorum anthers before anthesis noted in this study suggests the pos- sibility of a trend toward cleistogamy. Olsen and Olsen (1980) found that a population of Boschni- akia hookeri Walp., a related species within Oro- banchaceae, had achieved obligate autogamy via cleistogamy. Host specificity. The type specimen of O. pino- rum was collected by Andreas Geyer nearly 150 years ago (Geyer 1851). In his notes he wrote that it was “‘growing on the roots of Abies balsaminea.”’ It seems evident that Geyer coined the epithet pi- norum based on what he assumed was the host fam- ily. Most of the regional floras and identification books that followed mention coniferous hosts for O. pinorum (Munz 1930; Hitchcock et al. 1959; Abrams and Ferris 1960; Munz and Keck 1968; Peck 1961), but most fail to mention the host spe- cies, and the reliability of these accounts is doubt- ful. Musselman (1980) questioned the biological veracity of many reported hosts of species in Oro- banchaceae. There appears to be no mention in the literature of gymnosperm hosts for any of the Oro- banchaceae other than O. pinorum. Confirmation of such a finding would be an important contribution. Abrams and Ferris (1960) mention anecdotal re- ports of O. pinorum parasitizing H. discolor, which may be the earliest published suggestion of this as- sociation. Major and Taylor (1977) list ““Holodiscus microphyllus (with Orobanche pinorum)” among alpine vegetation habitats in California’s Cascades. Two recent works also name Holodiscus as the host. Heckard (1993) mentions “‘Holodiscus spp.” 2 MADRONO which suggests the possibility that H. discolor, H. microphyllus, and H. boursiéri might all be hosts. He goes on to state emphatically that O. pinorum is “not known on conifers.”’ Finally, Smith-Kuebel and Lillybridge (1993) specifically mention H. dis- color as the host, and include the comment that O. pinorum was “‘originally thought to be parasitic on conifers.’’ The findings of this study on host spec- ificity, although limited, support these last two re- ports. A species limited to a single host appears unusu- al in Orobanchaceae. The range of host families is very diverse, including herbs, shrubs, and trees. Monocot hosts are rare, and gymnosperm hosts may not exist. Musselman (1980) argues that most species in Orobanchaceae are very promiscuous, with broad host ranges, and that only physiological races are restricted to narrow host ranges. Further- more, he claims that O. crenata may have the most narrow host range, restricted to legumes. Thus O. pinorum may be unusual in having a narrow host range—possibly the most narrow within Oroban- chaceae—if it is truly restricted to three or fewer Holodiscus species. ACKNOWLEDGMENTS This study was supported financially by the Leaven- worth Ranger District, Leavenworth WA, and Western Washington University, Bellingham, WA. Assistance was provided by the following Forest Service personnel—EIl- len Kuhlmann, Dottie Knecht, Valeri Campanyitsev, and Laurie Malmquist. The detailed illustrations were drawn by Eve Ponder. Terry Griswold (Utah State University) determined the two megachilid bees. LITERATURE CITED ABRAMS, L. AND R. S. FERRIS. 1960. Illustrated flora of the Pacific states: Washington, Oregon, and Califor- nia. Stanford University Press, Stanford, CA. BEGON, M. AND M. Mortimer. 1986. Population ecology: a unified study of animals and plants, 2nd ed. Black- well Scientific Publications, Oxford, U.K. DONALDSON, W. R. AND C. RUSCHA. 1975. Washington cli- mate for Chelan, Douglas and Okanogan counties. Washington State University, Cooperative Extension Service, Pullman, WA. FRANKLIN, J. E AND C. T. DyRNEss. 1973. Natural vege- tation of Oregon and Washington. USDA Forest Ser- vice General Technical Report PNW-8. Pacific North- west Forest and Range Experiment Station, USDA Forest Service, Portland, OR. GEYER, A. 1851. Catalogue of Mr. Geyer’s collection of plants gathered in the Upper Missouri, the Oregon Territory, and the intervening portion of the Rocky Mountains. Hooker’s Journal of Botany and Kew Garden Miscellany 3:287—300. GRANT, V. 1981. Plant speciation, 2nd ed. Columbia Uni- versity Press, New York, NY. HECKARD, L. R. 1973. A taxonomic reinterpretation of the Orobanche californica complex. Madrono 22:41—70. 1993. Orobanchaceae: broom-rape family. Pp. 804-808 in J. C. Hickman (ed.), The Jepson manual: [Vol. 46 higher plants of California. University of California Press, Berkeley, CA. AND T. I. CHUANG. 1975. Chromosome numbers and polyploidy in Orobanche (Orobanchaceae). Brit- tonia 27:179-186. HiTcHcock, C. L., A. CRONQUIST, M. OWNBEY, AND J. W. THOMPSON. 1959. Vascular plants of the Pacific Northwest. University of Washington Press, Seattle, WA. HOLZNER, W. 1982. Concepts, categories and characteris- tics of weeds. Pp. 3—20 in W. Holzner and M. Numata (eds.), Biology and ecology of weeds. Dr. W. Junk Publishers, Boston, MA. JENSON, H. W. 1951. The normal and parthenogenetic forms of Orobanche uniflora in the eastern United States. Cellule 54:135-142. JEPSON, W. L. 1970. A manual of the flowering plants of California. University of California Press, Berkeley, CA. KuuT, J. 1969. The biology of parasitic flowering plants. University of California Press, Berkeley, CA. Mayor, J. AND D. W. TAYLOR. 1977. Alpine. Pp. 601—675 in M. G. Barbour and J. Major (eds.), Terrestrial veg- etation of California. John Wiley and Sons, New York, NY. Munz, P. A. 1930. The North American species of Oro- banche, section Myzorrhiza. Bulletin of the Torrey Botanical Club 57:61 1-624. AND D. D. KECK. 1968. A California flora. Uni- versity of California Press, Berkeley, CA. MUSSELMAN, L. J. 1980. The biology of Striga, Oroban- che, and other root-parasitic weeds. Annual Review of Phytopathology 18:463—489. , C. PARKER, AND N. DIXON. 1981. Notes on au- togamy and flower structure in agronomically impor- tant species of Striga (Scrophulariaceae) and Oroban- che (Orobanchaceae). Beitraege zur Biologie der Pflanzen 56:329-343. OLSEN, S. AND I. D. OLSEN. 1980. The seed of Boschniakia hookeri (Orobanchaceae). Botanisk Tidsskrift 75: 159-172. Peck, M. E. 1961. A manual of the higher plants of Or- egon, 2nd ed. Binfords and Mort, Publishers, Port- land, OR. REUTER, B. C. 1986. The habitat, reproductive ecology and host relations of Orobanche fasciculata Nutt. (Orobanchaceae) in Wisconsin. Bulletin of the Torrey Botanical Club 113:110—117. SMITH-KUEBEL, C. AND T. R. LILLYBRIDGE. 1993. Sensitive plants and noxious weeds of the Wenatchee National | Forest. USDA Forest Service, Pacific Northwest Re- | gion, R6-WEN-93-014. STEBBINS, G. L. 1957. Self fertilization and population | variability in the higher plants. The American Natu- | ralist 91:337-—354. TABOR, R. W., V. A. FRIZZELL, J. T. WHETTEN, R. B. WAITT, D. A. SWANSON, G. R. BYERLY, D. B. BOOTH, M. J. HETHERINGTON, AND R. E. ZARTMAN. 1987. Geo- | logic map of the Chelan 30-minute by 60-minute | quadrangle, Washington. Miscellaneous Investiga- | tions Map-I-1661. USDI Geological Survey, Reston, | VA. THIERET, J. W. 1971. The genera of Orobanchaceae in the | southeastern United States. Journal of the Arnold Ar- | boretum 52:404—434. WASHINGTON NATURAL HERITAGE PROGRAM. 1997. Endan- gered, threatened and sensitive vascular plants of | Washington. Department of Natural Resources, | Olympia, WA. MADRONO, Vol. 46, No. 1, pp. 13-19, 1999 ALIEN ANNUAL GRASSES AND FIRE IN THE MOJAVE DESERT MATTHEW L. BROOKS Department of Biology, University of California, Riverside, CA 92521 ABSTRACT Fires have become more frequent in the Mojave Desert since the 1970’s, threatening native plants and animals. This study describes how different annual plants facilitate the spread of fire during summer when weather conditions are optimal for fire. Eight annual plant taxa were evaluated: the alien grasses Bromus (B. madritensis L. subsp. rubens (L.) Husnot, B. tectorum L., B. trinii Desv.) and Schismus (S. arabicus Nees, S. barbatus (L.) Thell.), the alien forb Erodium cicutarium (L.) LU Hér, the native grass Vulpia (V. microstachys (Nutt.) Munro, V. octoflora (Walter) Rydb.), the native forbs Amsinckia tessellata A. Gray, Descurainia pinnata (Walter) Britton, and Phacelia tanacetifolia Benth., and other native forbs (119 species combined). Frequency and cover of dead annual plants were measured to describe the composition of fine fuels in summer (7 to 14 July), and compared to measurements of live annual plants in spring (12 April to 11 May) to determine their persistence as fuels after they senesced (summer: spring ratio). These data were collected during 1995 at 34 sites in the central, southern, and western Mojave Desert. Also described were the effects of each annual plant group in facilitating the spread of three 2.25 ha experi- mental fires conducted in August, 1995. Absolute frequency and cover and summer: spring ratios were highest for Bromus and Schismus, and lowest for native forbs. Alien annual grasses contributed most to the continuity and amount of dead annual plants and to the spread of summer fires. Fire spread rapidly (12 m/min) and continuously across interspaces with Bromus and slowly (1 m/min) and discontinuously with Schismus. No other annual plant group produced sufficient continuous biomass to carry fire across interspaces. Fire management must include the control of alien annual grasses in the Mojave Desert. The frequency of fire, the number of fires caused by humans (Fish and Wildlife Service 1994; Brooks 1998a), and the dominance of alien annual grasses (Hunter 1991; Brooks 1998b; Kemp and Brooks 1998) all increased between the 1970’s and 1990’s in the Mojave Desert. Fires caused by humans are most common near urban developments, major roads, and where off-highway vehicle use is unlim- ited (United States Department of the Interior re- cords), whereas fires caused by lightening are typ- ical of more remote wilderness areas. Alien annual grasses dominate all of these areas, and biomass of one species in particular, Bromus madritensis L. subsp. rubens (L.) Husnot, is strongly correlated with size and frequency of fire (Brooks 1998b). Abundance of alien annual grasses is also positive- ly correlated with the frequency of fire in the So- noran Desert (Brown and Minnich 1986; Schmid and Rogers 1988). Thus, dominance of alien annual grasses appears to be a primary environmental cor- relate of fire in the Mojave Desert. Areas dominated by alien annual grasses often have lower biomass and diversity of native forbs (Brooks 1998b; Brooks and Berry 1999), but it is unclear why landscapes dominated by them are more flammable than those dominated by native forbs in the Mojave Desert. Possible reasons in- Present address: United States Department of the Inte- rior, United States Geological Survey, Biological Re- _ sources Division, Western Ecological Research Center, 41734 South Fork Dr. Three Rivers, CA 93271, phone/ fax: 559-561-6511, matt_brooks @usgs.gov clude the higher surface to volume ratio of grasses compared to forbs that makes them easier to ignite (Kauffman and Uhl 1990), the more continuous cover of fuel that annual grasses often create on the landscape (Pyne et al. 1996), and the apparent abil- ity of alien annual grasses to remain rooted and upright longer than native forbs allowing them to persist as flammable fuels into the summer when the threat of fire is highest (Brooks and Berry 1999). Widely spaced shrubs and bunchgrasses with rel- atively bare interspaces between them characterize native Mojave Desert plant communities (Rundel and Gibson 1996). Frequent breaks in the continu- ity of fine fuels hinder the spread of fire, which is a primary reason fire is considered to be historically uncommon in this region (Humphrey 1974; O’Leary and Minnich 1981; Brown and Minnich 1986). The ability of alien annual grasses to pro- duce high amounts of persistent flammable fuels in perennial plant interspaces seems to promote Mo- jave Desert fires (Brooks 1998b). The purpose of this study was to compare the roles of alien annual grasses and other annual plants in facilitating the spread of fire in the Mojave Des- ert. This was accomplished by measuring the fre- quency and cover of fine fuels produced by differ- ent annual plant species and describing how flames spread through these fuels during experimental summer fires. Frequency was measured to evaluate the continuity and cover was measured to evaluate the amount of annual plant fuels. The summer: spring ratios of frequency and cover were calculat- 14 MADRONO ed to determine the amount that each decreased be- tween spring and summer. Absolute frequency and cover during summer were measured to compare the characteristics of annual plant fuels during the time of year when high temperatures, low relative humidity, and low fuel moisture levels create con- ditions that are ideal for fire. METHODS Eight annual plant taxa were analyzed: Bromus [B. madritensis subsp. rubens, B. tectorum L., B. trinii Desv.], Schismus (S. arabicus Ness, S. bar- batus {L.] Thell.), Erodium cicutarium (L.) L Hér, Vulpia (V. microstachys [Nutt.] Munro, V. octoflora [Walter] Rydb.), Amsinckia tessellata A. Gray, Des- curainia pinnata (Walter) Britton, Phacelia tana- cetifolia Benth., and other natives (119 forb spe- cies, Brooks 1998b Appendix 3.1). The first two are alien grasses, the third is an alien forb, the fourth is a native grass, and the remaining species are native forbs. These taxa are among the most widespread and abundant annual plants in the cen- tral, southern, and western Mojave Desert (Brooks 1998b). The one exception is Vulpia, which was included because of its possible ecological similar- ities with the alien annual grasses. Some species were grouped and analyzed as genera because they could not be reliably distinguished during the sum- mer. Plant nomenclature followed Hickman (1993). Frequency and cover of annual plants. Frequen- cy and cover of annual plants were measured at each of 34 sites located in the central (n = 16), southern (n = 8), and western (n = 10) Mojave Desert (Fig. 1). Sites were chosen by randomly se- lecting half the townships located within each of the three regions and randomly selecting one of the 1 mi’ sections within each township from those that did not contain playas, mountaintops, or private lands. Final sites were located within each section adjacent to but greater than 50 m from dirt roads and greater than 2 km from paved roads and human habitations. A single 360 m transect with twenty-five sam- pling points placed 15 m apart was established at each of the study sites. The transect was oriented parallel to the elevational contour to sample alter- nating run-off and run-on microtopographic posi- tions. At each site twenty-five replicate measure- ments were made within two microhabitats, beneath the north side of perennial plant canopies (>50 cm dia.) and in the interspace between them (>1 m from perennial plant canopies). Within each micro- habitat cover and frequency were estimated using a 22 cm long point-frame of ten equally spaced 1.5 mm diameter pins (Greig-Smith 1964). The frame was oriented perpendicular to the ground, each pin was lowered to the surface of the soil, and the num- ber of times each pin touched above-ground parts of annual plant species was recorded. Frequency was estimated as the proportion of the pins in each [Vol. 46 frame that touched at least one plant part and cover was estimated as the total number of pin touches per 10-pin frame. More detailed site and sampling design descriptions were presented by Brooks (1998b). Data were obtained during spring and summer 1995 following a winter with 200% of average rain- fall (National Oceanographic and Atmospheric Ad- ministration 1995). Measurements were first made from 12 April through 11 May when most species had reached peak biomass and just before they be- gan to senesce. Only living plant material was mea- sured. The second measurements were made from 7 through 14 July, when all germinated annual plants were dead. Because spring measurements were followed by biomass clipping (Brooks 1998b), summer measurements were recorded 20 cm from the spring measurements. Dead annual plant material included some senescent biomass from previous years in addition to plants that grew during the spring. Only material that was alive in spring 1995 was used to calculate the summer: spring ratios. Annuals that were alive in spring could be identified in summer because they were golden brown and often contained inflorescences and leaves. Annuals still standing after one year typically acquire a gray hue and lose their inflores- cences and leaves. Average frequency and cover of the eight annual plant taxa from each site were used as replicates to calculate average values across the entire study re- gion (n = 34 replicates). Sites that did not possess at least one point-frame hit for a given plant taxa during the spring were not used in the estimate of i summer : spring ratios. Some of the less common | species (e.g., P. tanacetifolia) were not detected at — some sites, reducing the number of replicates for — this group. Hence, sample sizes varied among | groups and differences among them were tested us- _ ing Tukey’s studentized range test that is robust for — unequal sample sizes (P = 0.05) (Sokal and Rohlf 1 1995). Arcsine transformations were performed on ratios and square root transformations were per- formed on absolute counts prior to testing. Experimental fires. To evaluate the role of dif- | ferent annual plants in facilitating the spread of fire, detailed observations were made during three ex- | perimental fires in the western (35°14'30’N, | 117°51'15”W), central (35°07'30"N, 117°07'45"W), | and southern (34°41'30’N, 117°57'30"W) Mojave Desert (Fig. 1). These fires were conducted on 16, 22, and 24 August 1995 respectively. Each site was | 150 X 150 m (2.25 ha) and dominated by Bromus | spp. and/or Schismus spp. Dead annual plants were | ignited using a continuous flame line applied with | a drip-torch (diesel/unleaded gas mix) along the up- © wind border of each site. The total amount of annual plant fuels (nearest | dry 25 kg/ha) and the dominant annual plant spe- cies in the beneath-canopy and interspace micro- | vO ve LL ee a an 1999] St. George Mojave Desert Mojave Los Angeles / 4 ] : N / rea | 100 km ' x =Annual Plant Frequency and Cover Measurements w =Fire Observations Mexico 7 ~~ Ariz ssn <%Na Fic. 1. Locations of the 34 study sites in the central, southern, and western regions of the Mojave Desert. habitats were determined using visual estimation. These estimates were based on the author’s expe- rience physically sampling annual plant biomass and visually estimating cover in the Mojave Desert. Rates of fire spread and flame lengths were record- ed at 15 random interspace points during each fire. The total amount of burned area (nearest 25%) and continuity of burning were visually estimated after each fire. Air temperature, relative humidity, cloud cover, wind speed and direction, at the beginning and end of each fire were recorded, because weath- er conditions can affect fire behavior. RESULTS Frequency and cover of annual plants. The pro- portion of point frame hits for Bromus spp. was 90% B. madritensis subsp. rubens and 10% B. tec- torum and B. trinii combined during spring 1995. The proportion of Schismus spp. species were not estimated because S. arabicus and S. barbatus could not be reliably distinguished in the field. The proportion of Vulpia species was 60% V. octoflora and 40% V. microstachys. Estimates of total annual plant cover were highly correlated with concurrent measurements of above-ground live biomass during spring 1995 (Brooks 1998b, r = 0.64). Summer: spring frequency was highest for Bro- mus spp. and Schismus spp. and lowest for the other natives category (Fig. 2). Vulpia spp., A. tessellata, D. pinnata, and P. tanacetifolia had intermediate frequency ratios, but very low absolute frequencies (Fig. 3). Erodium cicutarium had an intermediate BROOKS: ALIEN ANNUAL GRASSES AND FIRE 15 0.8 . 0.6 summer:spring frequency ro) AK 0.2 Fic. 2. Summer: spring frequency of annual plants av- eraged over 34 sites in 1995 (+1 SE). Dissimilar letters indicate significant differences using Tukey’s studentized range test (P < 0.05). * alien species. frequency ratio and absolute frequency. The com- bination of high summer: spring frequencies and high absolute frequencies during summer indicate that Bromus spp. and Schismus spp. contributed most to the frequency of dead annual plants in the summer. Summer: spring cover was highest for Bromus spp. and lowest for the other natives category (Fig. 4). Schismus spp., A. tessellata, D. pinnata, and P. tanacetifolia had intermediate ratios, but absolute cover of Schismus was significantly higher than all groups except Bromus spp. (P < 0.05, Fig. 5). Sum- mer: spring and absolute cover were relatively low for E. cicutarium. Similar to the frequency results, the combination of high summer: spring cover ra- tios and high absolute cover during summer indi- cate that Bromus spp. and Schismus spp. contrib- uted most to the cover of dead annual plants in the summer. Experimental fires. The central Mojave site had the lowest amount of fine fuels (Table 1) and the lowest wind speed and highest relative humidity during the fire (Table 2). As a result, fire did not spread beyond ignition points. MADRONO 50 40 ~ 30 S > S) © © > o 20 10 ° > 5 @ & & & i) S 4) SS ~~ ~~ SS @ oe SS FF SE § SF Fe Ke) ey Rey @ ‘. CS Cr Fic. 3. Absolute frequency of annual plants averaged over 34 sites in 1995 (+1 SE). Dissimilar letters indicate significant differences using Tukey’s studentized range test (P < 0.05). * alien species. The southern Mojave site had relatively high amounts of fine fuels in the beneath-canopy and interspace microhabitats (Table 1). Relative humid- ity was relatively low and wind speeds were mod- erate, so fire spread relatively fast across interspac- es and 50% of the site burned over large continuous areas (Table 2). The western Mojave site had high amounts of fine fuels in the beneath-canopy microhabitat, but only moderate amounts in the interspaces (Table 1). Relative humidity was relatively low and wind speeds were high (Table 2), but fire spread rela- tively slow and 50% of the total site burned in many small patches. Low humidity, moderate to high wind speeds, and substantial interspace biomass of fine fuels comprised mostly of alien annual grasses were as- sociated with high rates and continuities of fire spread. In contrast, relatively high humidity, low wind speeds, and virtually no fine fuels between shrubs were associated with no fire spread. Differ- ences in fire behavior were not attributed exclu- sively to weather or fuel composition, because these variables were confounded. However, fires [Vol. 46 0.8 summer:spring cover Fic. 4. Summer: spring cover of annual plants averaged over 34 sites in 1995 (+1 SE). Dissimilar letters indicate significant differences using Tukey’s studentized range test (P < 0.05). * alien species. were always fueled by the dead stems of alien an- nual grasses. Fire spread was extensive and rapid where Bro- mus spp. was codominant with Schismus spp. in interspaces (southern Mojave site; Table 1). Fire spread was patchy and slow where only Schismus spp. was dominant in interspaces (western Mojave site). Average wind speed during the fire at the western Mojave site was twice that at the southern Mojave site (Table 2), yet fire spread faster and more continuously across the latter site (Table 1). Thus, high interspace biomass of Bromus spp. and Schismus spp. resulted in greater fire danger at the southern Mojave site, even though wind speeds were much higher at the western Mojave site. Where Bromus spp. was abundant in interspaces fire spread approximately 12 m/min with flame lengths up to 30 cm. The heat generated by Bromus spp. was sufficient to ignite and consume dead stems of native forbs and plant litter. Schismus spp. was also effective at carrying fire across interspac- es, but only at 1 m/min with flame lengths 5—10 cm. Flame lengths from Schismus spp. could not easily be seen and often only burned the top 25% of Schismus spp. stems, indicating that tempera- tures were relatively low. Most dead native forb stems and litter material were unburned. Although | Bromus spp. and Schismus spp. both facilitated the — spread of fire, only Bromus spp. produced long | ) 1999] 20 15 10 absolute cover (point hits / 10-pin frame) Fic. 5. Absolute cover of annual plants averaged over 34 sites in 1995 (+1 SE). Dissimilar letters indicate sig- nificant differences using Tukey’s studentized range test (P < 0.05). * alien species. flame lengths that consumed considerable amounts of annual plant biomass. Erodium cicutarium and native annuals did not contribute significantly to the spread of fire, because of low frequency and cover. Flames fueled by Bromus spp. were sufficient to consume small shrubs such as Ambrosia dumosa (A. Gray) Payne, Krascheninnikovia lanata (Pursh) A. D. J. Meeuse & Smit, Hymenoclea salsola A. Gray, and Lycium andersonii A. Gray, whereas flames fueled by Schismus spp. were rarely hot enough to ignite these shrubs. Fire intensity in Bro- mus spp. was usually insufficient to ignite large Shrubs such as Larrea tridentata (DC.) Cov. How- BROOKS: ALIEN ANNUAL GRASSES AND FIRE 17 ever, Larrea tridentata containing large accumula- tions of Bromus spp. stems and dead shrub stems in the sub-canopy were highly susceptible to burn- ing. In these cases fire carried from Bromus spp. stems, to dead shrub stems, to live shrub stems, and typically resulted in the entire shrub being con- sumed by flames. DISCUSSION The alien annual grasses Bromus spp. and Schis- mus Spp. appear to be necessary for fire to spread across the Mojave Desert landscape. These were the only annual plant taxa that produced abundant and continuous cover of fine fuels that persisted into the summer fire season. Intermediate fuels produced by large forbs can add to the continuity and amount of flammable bio- mass, although fine fuels produced by alien annual grasses are generally required to sustain a fire (Brooks personal observation). These large forbs include the alien mustards Brassica tournefortii Gouan, Hirschfeldia incana (L.) Lagr.-Fossat, Sis- ymbrium altissimum L., Sisymbrium irio L., and Descurainia sophia (L.) Webb, and the weedy na- tive A. tessellata. They are especially common along roads where fires frequently start and some of the alien species are rapidly expanding their range and becoming established away from roads in the Mojave and Colorado deserts (Kemp and Brooks 1998; Brooks and Berry 1999). Hence, large weedy forbs may present a more widespread fire hazard in the future. Thick layers of annual plant litter often develop where alien annual grasses are abundant (Brooks and Berry 1999). Accumulations of litter led to par- ticularly hot temperatures, long flame residency times, and continuous burn patterns in experimental fires conducted during summer 1995 and 1996 in the Mojave Desert (Brooks unpublished). Plant lit- ter decomposes slowly in desert regions and grasses can be among the slowest (Facelli and Pickett 1991). Thus, litter accumulation may be another mechanism by which alien annual grasses facilitate the spread of Mojave Desert fires. The current study suggests that Bromus spp. fuel fast moving hot fires whereas Schismus spp. fuel slower moving cooler fires. This pattern is gener- TABLE 1. SITE DATA FOR EXPERIMENTAL FIRES CONDUCTED IN AUGUST 1995. Bie ticle Interspace (kg/ha) Dominant annuals Be (spp. >25% relative cover) Sus Beneath- Inter- (m/min) Area burned Site canopy space Beneath-canopy Interspace CSE). -€% of 2.25 ha) Central Mojave 300 25 Bromus/Schismus Schismus 0 (0) O Southern Mojave 700 200 Bromus Bromus/Schismus 12 (8) 50 (continuous) Western Mojave 800 100 Bromus/Schismus Schismus 1 (1) 50 (patchy) 18 MADRONO TABLE 2. WEATHER DATA FOR EXPERIMENTAL FIRES CONDUCTED IN AUGUST 1995. Air Time temperature Site (PST) CC) Central Mojave begin 1030 35 end 1115 41 Southern Mojave begin 1020 oD end 1130 35 Western Mojave begin 1130 37 end 1215 38 ally consistent among seasons and years (Brooks personal observation). The immediate ecological effects of Bromus spp. fires are probably more sig- nificant, because they are more intense and often consume perennial shrubs. However, Schismus spp. can facilitate the spread of fire between patches of Bromus spp., and promote fires at arid low eleva- tion sites where Bromus spp. are less abundant (Brooks 1998b). During the 1990’s in the Mojave Desert some fires fueled mostly by Schismus spp. exceeded 40 ha (100 acres) before they were extin- guished by fire crews (United States Department of the Interior records). High postfire dominance of alien annual grasses can promote subsequent fires in the Great Basin desert (Whisenant 1990; Peters and Bunting 1994; Billings 1994) and other ecosystems (D’ Antonio and Vitousek 1992). Post-fire plant communities in the Mojave and Sonoran deserts are also typically dominated by alien annual grasses (O’Leary and Minnich 1981; Brown and Minnich 1986; Brooks unpublished), so previously burned areas appear to be more susceptible to fire than unburned areas. This grass/fire cycle is a significant ecological threat because most native plant species are poorly adapted to survive fire in the deserts of southwest- ern North America (Tratz 1978; O’Leary and Min- nich 1981; Wright and Bailey 1982; Brown and Minnich 1986; Billings 1994; Lovich and Bain- bridge 1999). Drought years may reduce the dominance of Bromus spp. in both recently burned and unburned areas decreasing the chance of fire (Minnich per- sonal communication), but these effects vary among sites. For example, the winter of 1998-1999 was very dry in the Mojave Desert and most Bro- mus spp. seedlings did not survive to maturity on low elevation bajadas, whereas many survived and reproduced on high elevation hills and mountains (Brooks personal observation). High elevation sites may provide more mesic conditions that allow Bro- mus spp. to survive drought better than at lower elevations. Recurrent fire is most prevalent at these high elevation sites (Brooks unpublished) where high biomass of alien annual grasses and the phys- ical effects of steep slopes promote fire. Establish- ment of the grass/fire cycle appears to be more like- [Vol. 46 Rel. Cloud Wind speed humidity cover Wind (km/h) (%) (%) direction (gusts) as 0) SSW O (8) 25 0) SSW O (8) 17 0) SSE 54 O) 15 0) SSE 8 (13) 19 0) NNE 16 (18-32) 10 0) NNE 16 (18-32) ly on high elevation slopes than on low elevation bajadas. Management of fire in the Mojave Desert should focus on minimizing the dominance of alien annual grasses and preventing the establishment of new plant species that can increase landscape flamma- bility (Brooks and Berry 1999). Such species in- clude large forbs and the perennial bufflegrass (Pennisetum ciliare, Brooks et al. 1999). Sources of ignition from human activities should also be minimized, especially where alien annual grasses are abundant and topography is conducive to the spread of fire. ACKNOWLEDGEMENTS I thank Richard Franklin, Don Orsburn, Phil Gill, Chuck Robbins, and the fire crews of the United States Department of the Interior, Bureau of Land Management, California Desert District (BLM-CDD) for their assistance in conducting the experimental fires. I also thank Larry Forman, Bob Parker, Tom Egan, and the other BLM-CDD resource managers for their assistance in acquiring the permits needed to conduct the fires. Mary Price, Kristin Berry, John Rotenberry, Richard Minnich, Edith Allen, Bradford Martin, and an anonymous reviewer provided helpful reviews of this manuscript. The Interagency Fire Coordination Committee of the United States Department of the Interior provided financial support. LITERATURE CITED BILLINGS, W. D. 1994. Ecological impacts of cheatgrass and resultant fire on ecosystems in the western Great Basin. Pp. 22-31 in S. B. Monsen and S. G. Kitchen (eds.). Proceedings—Ecology and Management of Annual Rangelands, 18-22 May 1992, Boise ID. General Technical Report INT-GTR-313, Department of Agriculture, Forest Service, Intermountain Re- search Station. Brooks, M. L. 1998a. Effects of fire on the desert tortoise, Gopherus agassizii. International Conference on Tur- tles and Tortoises, 30 July—2 August 1998, North- ridge, CA. . 1998b. Ecology of a biological invasion: Alien annual plants in the Mojave Desert. Ph.D. disserta- tion, University of California, Riverside, CA. Brooks, M. L. AND K. H. BERRY. 1999. Ecology and Man- agement of Alien Annual Plants in the California De- serts. California Exotic Pest Plant Council Newsletter (in press). 1999] Brooks, M. L., T. C. ESQUE, AND C. R. SCHWALBE. 1999. Effects of exotic grasses via wildfire on desert tor- toises and their habitat. Desert Tortoise Council Sym- posium, 5—8 March 1999, St. George, UT. Brown, D. E. AND R. A. MINNICH. 1986. Fire and creosote bush scrub of the western Sonoran Desert, California. American Midland Naturalist 116:411—422. D’ ANTONIO, C. M. AND P. M. VITOUSEK. 1992. Biological invasions by exotic grasses, the grass/fire cycle, and global change. Annual Review of Ecology and Sys- tematics 3:63—87. FACELLI, J. AND S. T. A. PickeTT. 1991. Plant litter: its dynamics and effects on plant community structure. Botanical Review 57:1-31. FISH AND WILDLIFE SERVICE. 1994. Desert Tortoise (Mo- jave Population) Recovery Plan. United States Fish and Wildlife Service, Portland, OR. GREIG-SMITH, P. 1964. Quantitative Plant Ecology, 2nd ed. Butterworth, London. HICKMAN, J. C. (ED.) 1993. The Jepson Manual: Higher Plants of California. University of California Press, Berkeley, CA. HuMPHREY, R. R. 1974. Fire in deserts and desert grass- land of North America. Pp. 365-401 in T. T. Ko- zlowski and C. E. Ahlgren (eds.). Fire and Ecosys- tems. Academic Press, New York. HunrtTER, R. 1991. Bromus invasions on the Nevada Test Site: present status of B. rubens and B. tectorum with notes on their relationship to disturbance and altitude. Great Basin Naturalist 51:176—182. KAUFMAN, J. B. AND C. UHL. 1990. Interactions of anthro- pogenic activities, fire, and rain forests in the Amazon basin. Pp. 117—134 in J. G. Goldammer (ed.). Fire in the Tropical Biota: Ecosystem Processes and Global Challenges. Springer-Verlag, Berlin. Kemp, P. R. AND M. L. Brooks. 1998. Exotic Species of California Deserts. Fremontia 26:30—34. Lovicu, J. E. AND D. BAINBRIDGE. 1999. Anthropogenic degradation of the southern California desert ecosys- tem and prospects for natural recovery and restora- tion. Environmental Management (in press). NATIONAL OCEANOGRAPHIC AND ATMOSPHERIC ADMINIS- BROOKS: ALIEN ANNUAL GRASSES AND FIRE Ibe) TRATION. 1995. Climatological data annual summary, National Oceanographic and Atmospheric Associa- tion: 99. O’ LEARY, J. F AND R. A. MINNICH. 1981. Postfire recovery of creosote bush scrub vegetation in the Western Col- orado Desert. Madrono 28:61—66. PETERS, E. F AND S. C. BUNTING. 1994. Fire conditions pre- and postoccurence of annual grasses on the snake river plain. Pp. 31—36 in S. B. Monsen and S. G. Kitchen (eds.). Proceedings—Ecology and Manage- ment of Annual Rangelands, 18—22 May 1992, Boise ID. General Technical Report INT-GTR-313, Depart- ment of Agriculture, Forest Service, Intermountain Research Station. PYNE, S. J., PR. L. ANDREWS, AND R. D. LAVEN. 1996. In- troduction to wildland fire. Second Edition. John Wi- ley and Sons, New York. RUNDEL, P. W. AND A. C. GIBSON. Ecological communities and processes in a Mojave Desert ecosystem: rock valley Nevada. Cambridge University Press, Cam- bridge. SCHMID, M. K. AND G. F RoGers. 1988. Trends in fire occurrence in the Arizona upland subdivision of the Sonoran Desert, 1955 to 1983. The Southwestern Naturalist 33:437—444. SOKAL, R. R. AND FE J. ROHLF. 1995. Biometry. W. H. Freeman and Company, New York. TRATZ, W. M. 1978. Postfire vegetational recovery, pro- ductivity, and herbivore utilization of a chaparral-des- ert ecotone. Master’s thesis, California State Univer- sity, Los Angeles. CA. WHISENANT, S. G. 1990. Changing fire frequencies on Ida- ho’s snake river plains: ecological and management implications. Pp. 4—7 in E. D. McArthur, E. D. Rom- ney, E. M. Smith, and S. D. Tueller (eds.). Proceed- ings—Symposium on Cheatgrass Invasion, Shrub Die-off, and Other Aspects of Shrub Biology and Management, 5—7 April 1989, Las Vegas, NV. Gen- eral Technical Report INT-276, Department of Agri- culture, Forest Service, Intermountain Research Sta- tion. WRIGHT, H. E. AND A. W. BAILEY. 1982. Fire Ecology, United States and Canada. Wiley, New York. MAbRONO, Vol. 46, No. 1, pp. 20-24, 1999 SYNCHRONY AND ASYNCHRONY OF ACORN PRODUCTION AT TWO COASTAL CALIFORNIA SITES WALTER D. KOENIG Hastings Reservation, University of California, 38601 E. Carmel Valley Road, Carmel Valley, CA 93924 DALE R. MCCULLOUGH Department of ESPM, University of California, Berkeley, CA 94720 CHARLES E. VAUGHN Hopland Research and Extension Center, University of California, 4070 University Road, Hopland, CA 95449 JOHANNES M. H. KNops School of Biological Sciences, University of Nebraska, 348 Manter Hall, Lincoln, NE 68588-0118 WILLIAM J. CARMEN 145 Eldridge, Mill Valley, CA 9494] ABSTRACT We measured annual acorn production of oaks Quercus spp. at Hastings Reservation and at Hopland Research and Extension Center, located 320 km apart in the outer coast ranges of California, for 16 years between 1982 and 1997. Of the three species measured at both sites, acorn production by Quercus lobata Nee (valley oak) and Quercus douglasii Hook. & Arn. (blue oak) was significantly correlated between sites, whereas acorn production by Quercus kelloggii Newb. (California black oak) was not. Both Q. lobata and Q. douglasii acorn production was significantly correlated with mean April temperatures and rainfall at their respective localities, but more closely with April temperatures at Hastings and with rainfall at Hopland. Synchrony in acorn production between Quercus spp. requiring one year to mature acorns was significantly greater than among those requiring two years to mature acorns. The geographic extent of the populations producing acorn crops synchronously differs between species, but in some cases may extend over distances of at least several hundred kilometers. Mast-fruiting, or masting, is a population phe- nomenon (Kelly 1994). That is, a single tree may produce highly variable numbers of seeds from year to year, but it is only when a population of trees produce seeds synchronously from one year to the next that masting can be considered to occur. Recent studies in New Zealand (Norton and Kelly 1988), the midwestern United States (Sork et al. 1993) and California (Koenig et al. 1994a, 1996) have begun to elucidate the patterns and causes of masting behavior in forest trees. However, work has only recently begun to address the question: what is the geographic extent of the ‘population’ pro- ducing seeds synchronously? Here we analyze data relevant to this question using data on three species of Quercus. Specifical- ly, we independently collected data on acorn pro- duction by Q. lobata (valley oak), Q. douglasii (blue oak), and Q. kelloggii (California black oak) oaks at two sites in coastal California 320 km apart: Hopland Research and Extension Center in Men- docino County (38°58.5'N, 123°07'W; hereafter ‘*‘Hopland’’) and Hastings Natural History Reser- vation in Monterey County (36°23'N, 121°33’W; hereafter ‘“‘Hastings’”’), over a 16 year period from 1982 to 1997 (Fig. 1). We ask: 1) does the popu- lation of masting oaks extend over this distance in California? and 2) is acorn production correlated with similar environmental factors at the two sites? METHODS At Hopland, acorn production by each species was censused using 10 traps consisting of plastic garbage bags 0.46 m in diameter, each placed at a random location under a tree of the appropriate spe- cies. Traps were checked in December at the end of the season and the total number of sound acorns trapped log-transformed (/n[N+1]). Species cen- sused included Q. lobata, Q. douglasii, Q. kellog- gii, and interior live oak Q. wislizenii A. DC. (in- terior live oak). At Hastings, acorn censuses were done visually between mid-September and early October just pri- or to acorn fall (Koenig et al. 1994b). For each tree two observers scanned different areas of the tree’s canopy and counted as many acorns as possible in 15 sec. Counts were added to yield the number of acorns counted in 30 sec (N30). Values for each 1999] Fic. 1. A map of central coastal California showing the locations of Hopland Research and Extension Center and Hastings Natural History Reservation relative to the San Francisco Bay area (center). Lines are county boundaries; Hopland is in Mendocino County while Hastings is in Monterey County. tree were log-transformed (/n[N30+1]) and aver- aged to yield the mean log-transformed number of acorns counted per tree of each species. The rela- tive merits of visual surveys versus traps for cen- susing acorns in California oak woodland habitat are discussed in Koenig et al. (1994b), who also provide data demonstrating that values derived from visual counts are significantly correlated with numbers of acorns obtained by trapping for Q. lo- bata at Hastings. Species included and the number of trees cen- sused per species were Q. lobata (87), Q. douglasii (57), Q. kelloggii (21), Q. agrifolia Nee (coast live oak; 63), and (Q. chrysolepis Liebm. canyon live oak; 21). Thus, three species (Q. lobata, Q. doug- lasii, and Q. kelloggii) were surveyed at both lo- calities. These species differ in that both Q. lobata and Q. douglasii are ‘‘1-year species’ requiring a single year to mature acorns, whereas Q. kelloggii is a “‘2-year species’’ requiring two years to mature acorns. Of the three live oak species censused at one or the other site, Q. agrifolia (at Hastings) is a one-year species while Q. wislizenii (at Hopland) and Q. chrysolepis (at Hastings) are two-year spe- CIES. Weather data came from stations located near headquarters at both sites. Variables analyzed in- cluded seasonal rainfall (1 Sept. of year x-1 to 31 August of year x) and mean April temperature KOENIG ET AL.: SYNCHRONY OF ACORN PRODUCTION 21 6 2) 4 s) 2 1 Q. douglasii le 0.63** @ S SS = CQ. ss QO. kelloggii r=0.14 S Hastings Fic. 2. Correlations between the log-transformed mean acorn crops of Q. lobata, Q. douglasii and Q. kelloggii at Hopland and Hastings between 1982 and 1997 (N = 16 years). Spearman rank correlations and their significance values (** = P < 0.01; *** = P < 0.001) are listed. (mean of the daily averages of the maxima and minima). Statistical analyses were made using non- parametric Spearman rank correlations and Mann- Whitney U tests. P-values are two-tailed; values listed are means + SD. RESULTS Annual acorn production at the two sites was highly correlated for Q. lobata and Q. douglasii, but not for Q. kelloggii (Fig. 2). Prior studies of Q. lobata and Q. douglasii at Hastings have demon- strated significant correlations between acorn pro- duction by these two species and mean April tem- peratures during the peak of flowering and polli- nation (Koenig et al. 1996), whereas no environ- mental variable has as yet been identified to correlate with acorn production by @Q. kelloggii. Correlations between acorn production of these 22 MADRONO TABLE 1. [Vol. 46 SPEARMAN RANK CORRELATION COEFFICIENTS BETWEEN ACORN PRODUCTION AT HASTINGS AND HOPLAND AND ENVIRONMENTAL VARIABLES MEASURED AT THE SAME SITES BETWEEN 1982 AND 1997 (N = 16 YEARS). * = P S 0.05; ee P< 0:01; 7** = P= 0.001. Mean April temp. Seasonal rainfall Mean April temp. (year x — 1) Seasonal rainfall (year x — 1) Hastings Q. lobata O.825"* = O10 =—0657* 0.04 Q. douglasii i sese =0,25 —0.48 0.26 Q. kelloggii 0.04 O.59* 0:09 0.01 Hopland Q. lobata 0.637** =0:32 SO ae 0.35 Q. douglasii 0.50* =035 =Q. 72% 0.57* Q. kelloggii 0.12 0.42 ZON9 —0.28 three species at both Hastings and Hopland over the 16 years analyzed here and both mean April tem- perature and seasonal rainfall are summarized in Table 1. Correlations between the environmental §vari- ables at the two sites over the 1982—1997 period were high (rainfall: r, = 0.79, n = 16, P < 0.001; mean April temperature: r, = 0.94, n = 16, P < 0.001). However, the relationships between these environmental factors and acorn production were not identical at the sites. Acorn production by Q. lobata and Q. douglasii were positively correlated with mean April temperature at both sites; however, the correlations were much higher at Hastings than at Hopland. Interestingly, the reverse pattern holds for the relationship between acorn production by these two species and seasonal rainfall: all corre- lations were negative, but they were considerably stronger at Hopland than at Hastings. Multiple re- gression analyses with these two variables yielded identical results. Acorn production by Q. douglasii at Hopland was also significantly positively corre- lated with seasonal rainfall lagged one year while Q. kelloggii at Hastings was positively correlated with mean April temperature lagged one year; both these correlations were low and not statistically sig- nificant at the other site. We also compared interspecific correlations be- tween one-year (Table 2) and between two-year (Tabie 3) species of oaks, both within and between sites. For the one-year species, 9 of 10 interspecific correlations were significant, including all six of the cross-site comparisons. Overall, the mean interspe- TABLE 2. cific correlation coefficient was 0.64 + 0.16 and did not differ significantly between cross-site (mean = 0.62 + 0.10) and within-site comparisons (mean = 0.68 + 0.24; Mann-Whitney U-test, z = 0.6, P = 0.52). For the two-year species, none of the six pairwise correlation coefficients was significantly different from zero and the mean interspecific cor- relation coefficient was 0.11 + 0.21, significantly less than the interspecific correlations between one- year species (Mann-Whitney U-test, z = 3.2, P < 0.002). Among the two-year species, only the cor- relation between Q. kelloggii and Q. chrysolepis at Hastings came close to being significant (P = 0.06). DISCUSSION These results demonstrate that for Q. lobata and Q. douglasii annual acorn crops are highly syn- chronous between two sites in coastal California 320 km apart. Furthermore, acorn crops of these species at the two sites are both correlated with mean April temperatures. This is consistent with the hypothesis that masting occurs over large geo- graphic areas in these two species, and furthermore suggests that the proximate cues used by trees to synchronize reproductive effort are also similar over large distances (Koenig et al. 1996). However, patterns were not identical at the sites. At Hastings, both species were positively and strongly correlated with mean April temperature and negatively, but less strongly, correlated with seasonal rainfall, whereas at Hopland the pattern was reversed. This suggests that the precise mix of environmental fac- SPEARMAN RANK CORRELATION COEFFICIENTS BETWEEN ACORN PRODUCTION OF ONE-YEAR QUERCUS SPECIES, BOTH WITHIN AND BETWEEN HASTINGS AND HOPLAND, BETWEEN 1982 AND 1997 (N = 16 YEARS). * = P S 0.05; ** = P= 0.01, *** => P= 0,001. Q. agrifolia (Hastings) Q. lobata (Hastings) 0.38 Q. douglasii (Hastings) 0.59* Q. lobata (Hopland) 0:53* Q. douglasii (Hopland) 0.53* Q. lobata Q. douglasii Q. lobata (Hastings) (Hastings) (Hopland) OZ se Oe — OS7* 0637 O289%** a 1999] TABLE 3. SPEARMAN RANK CORRELATION COEFFICIENTS BETWEEN ACORN PRODUCTION OF TWO-YEAR QUERCUS SPE- CIES, BOTH WITHIN AND BETWEEN HASTINGS AND HOPLAND, BETWEEN 1982 AND 1997 (N = 16 YEARS). All P > 0.05. d. d. Q. kelloggii chrysolepis wislizenti (Hastings) (Hastings) (Hopland) Q. chrysolepis (Hastings) 0.48 — — Q. wislizenit (Hopland) 0.05 0.16 — QO. kelloggii (Hopland) 0.14 —0.05 —0.10 tors influencing acorn production may differ at dif- ferent localities, potentially explaining the lower correlation between acorn production at the sites (r, = 0.82 for QO. lobata and 0.73 for Q. douglasii) compared to that for the critical environmental fac- tor (r, = 0.94 for mean April temperatures). In contrast, we found no evidence of synchrony between acorn production by Q. kelloggii surveyed at the two sites. Because Q. kelloggii requires two years to mature acorns, it is possible that the en- vironmental factors synchronizing reproduction are more complicated, and thus less likely to be geo- graphically synchronous, than those used by the one-year species. Consistent with this hypothesis, analyses at Hastings Reservation using the 16 years of data between 1980 and 1995 found no relation- ship between any plausible environmental variable and acorn production in Q. kelloggii (Koenig et al. 1996). However, with the 1982 to 1997 data used here, there is a Statistically significant correlation between mean April temperature lagged one year and acorn production by @Q. kelloggii at Hastings. This correlation is not significant in the Hopland data over the same time period. More data will be needed before we will be able to understand what environmental factors influence acorn production in this species. We also performed pairwise interspecific com- parisons of the acorn crops of the one-year and the two-year species both within and between sites. All interspecific comparisons between one-year species were positive and 9 of 10 were statistically signif- icant (Table 2). Mean correlations were high and there was no difference between the within-site compared to the between-site cross-correlations. In contrast, only 4 of 6 comparisons between two-year Species were positive and none was statistically sig- nificant (Table 3). These results support the hy- pothesis that acorn production of two-year Quercus Species is geographically less synchronous than that of one-year Quercus species. Analyses of acorn production from data reported in the literature similarly indicate that the geograph- ic extent of synchrony in acorn production by two- year Quercus species is less than that of one-year Species (Koenig and Knops 1997). At the proximate KOENIG ET AL.: SYNCHRONY OF ACORN PRODUCTION 23 level, this could be because the two-year species are sensitive to more complicated environmental factors or because they are sensitive to different sets of environmental factors in different sites, as suggested by our results for Q. lobata and Q. doug- lasii. At a more ultimate level, it could be because the ecological factors selecting for mast-fruiting in two-year species differ from those important to one-year species, due for example to differences in the habitats they inhabit. Hypotheses suggested to favor masting include predator satiation (Silver- town 1980), wind pollination (Smith et al. 1990), and several other lesser ways by which efficiency may be increased by devoting more resources to reproduction in some years than others (Norton and Kelly 1988; Kelly 1994). These results add to the small but growing amount of data available concerning the geographic scale of, and the proximate factors involved in, syn- chronizing acorn production by Quercus spp. Com- bined with prior analyses demonstrating significant synchrony between acorn production of Q. lobata and Q. douglasii at Hastings and at Jasper Ridge in San Mateo County 130 km away, and between acorn production of Q. agrifolia not only at Jasper Ridge and at Hastings but also at Pozo, 290 km south of Jasper Ridge (Koenig et al. 1996), these data extend synchrony in acorn production by Quercus spp. to a distance of over 300 km. The proximate cue used to synchronize acorn produc- tion by Q. lobata and Q. douglasii throughout this range appears to be either spring temperature or seasonal rainfall, which are themselves correlated over large distances. Whether these patterns extend throughout the state is currently under investiga- tion. In contrast, we found no statistical synchrony in acorn production by Q. kelloggii between Hopland and Hastings. This negative finding is consistent with prior analyses indicating that the environmen- tal factors affecting synchrony in acorn production in oaks requiring two years to mature acorns are more difficult to discern than those used by one- year species of Quercus. Also supporting the contention that synchrony is lower in oaks requiring two years to mature acorns are the results of interspecific correlations, both within and between sites, between one-year and be- tween two-year Quercus species. Correlations be- tween annual acorn production of one-year species were all positive and, with a single exception, sta- tistically significant. In contrast, correlations be- tween annual acorn production of two-year species were not consistently positive or negative and none was significantly different from zero. Geographic synchrony in acorn production appears to be greater both within and between one-year species than within or between two-year species of Quercus. These results suggest a complex pattern of spa- tial autocorrelation in acorn production by Quercus spp. Within and even between one-year species, the 24 MADRONO extent of geographic synchrony in acorn production appears to be large, possibly encompassing the en- tire state. However, for two-year Quercus species, geographic synchrony appears neither to be as ex- tensive nor to cross species boundaries. What this means for a particular locality depends largely on the geographic scale being considered. On a local scale of a few square kilometers, many California sites contain only one-year Quercus spe- cies and thus may be subject to relatively frequent community-wide acorn crop failures due to the syn- chrony in acorn production across one-year Quer- cus species. Such synchrony is likely to extend over large geographic areas thousands or even tens of thousands of square kilometers in size. However, once the geographic scale over which one is con- cerned starts to encompass such larger areas, the topographic heterogeneity and complexity of the California landscape will generally ensure that sites containing both one-year and two-year Quercus species will be present somewhere within the area. Thus, despite large-scale geographic synchrony in at least several of the most widespread species of California oaks, the diversity of habitats occurring over moderately large geographic areas makes it unlikely that the acorn crop of all species will fail in any particular year (Koenig and Haydock 1999). Masting by oaks has been shown to have cas- cading effects on communities in the eastern United States via its affect on mouse populations (Jones et al. 1998). Consequently, large-scale geographic synchrony in acorn production such as is suggested here could plausibly have major effects on com- munities over similarly large geographic areas, es- pecially to the extent that the species involved are specialized on the acorns of one-year Quercus. No vertebrate acorn predator of which we are aware is specialized in this way. However, at least some of the many invertebrate species that depend on acorns are restricted to the acorns of particular Quercus subgenera, and sometimes usually a single Quercus species (Russo 1979; Cornell 1985). The effects of geographic synchrony in acorn produc- tion on populations of such taxa remain to be doc- umented. ACKNOWLEDGMENTS Support for these projects came from the University of California’s Integrated Hardwood Range Management Program, the University of California Agricultural Field Stations, Barry Garrison and the California Department of Fish and Game, and California Agricultural Experiment Projects 4031-MS and 5873-MS. We thank the many peo- ple who helped survey acorn production including Paul [Vol. 46 Beier, Chris Byrne, Tom Kucera, Ron Mumme, Pam Mul- ligan, Mary O’Bryan, Mark Stanback, and Floyd Weck- erly. Thanks are also due to Al Murphy and Robert Timm, the directors of Hopland, Mark Stromberg, the manager of Hastings, and an anonymous reviewer for comments on the manuscript. LITERATURE CITED CORNELL, H. V. 1985. Local and regional richness of cy- nipine gall wasps on California oaks. Ecology 66: 1247-1260. GRIFFIN, J. R. AND W. B. CRITCHFIELD. 1972. The distri- bution of forest trees in California. U.S. Department of Agriculture Forest Service Research Paper PSW- 82. JONES, C. G., R. S. OSTFELD, M. P. RICHARD, E. M. SCHAUBER, AND J. O. WOLFF. 1998. Chain reactions linking acorns to gypsy moth outbreaks and lyme dis- ease risk. Science 279:1023—1026. KELLY, D. 1994. The evolutionary ecology of mast seed- ing. Trends in Ecology and Evolution 9:465—470. KOENIG, W. D. AND J. HAyYDOCK. 1999. Oaks, acorns, and the geographical ecology of acorn woodpeckers. Jour- nal of Biogeography 26:159—165. , R. L. Mumme, W. J. CARMEN, AND M. T. STAN- BECK. 1994a. Acorn production by oaks in central coastal California: variation within and among years. Ecology 75:99-109. , J. M. H. KNops, W. J. CARMEN, M. T. STANBECK, AND R. L. MumMeE. 1994b. Estimating acorn crops using visual surveys. Canadian Journal of Forest Re- search 24:2105—2112. , > : , AND 1996. Acorn production by oaks in central coastal Califor- nia: influence of weather at three levels. Canadian Journal of Forest Research 26:1677—1683. AND . 1997. Patterns of geographic syn- chrony in growth and reproduction of oaks within California and beyond. Pp. 101-108 in N. H. Pills- bury, J. Verner, and W. D. Tietje (tech. coords.), Pro- ceedings of a symposium on oak woodlands: ecology, management, and urban interface issues. U.S. De- partment of Agriculture Forest Service General Tech- nical Report PSW-GTR- 160. Norton, D. A. AND D. KELLY. 1988. Mast seeding over 33 years by Dacrydium cupressinum (Lamb. (rimu)(Podocarpaceae) in New Zealand: the impor- tance of economies of scale. Functional Ecology 2: 399-408. Russo, R. 1979. Plant galls of the California region. Box- wood Press, Pacific Grove, CA. SILVERTOWN, N, J. W. 1980. The evolutionary ecology of mast seeding in trees. Biological Journal of the Lin- nean Society 14:235-—250. SmTIH, C. C., J. L. MANRICK, AND C. L. KRAMER. 1990. The advantage of mast years for wind pollination. American Naturalist 136:154—166. Sork, V. L., J. BRAMBLE, AND O. SEXTON. 1993. Ecology of mast fruiting in three species of North American deciduous oaks. Ecology 74:528-541. MaprONo, Vol. 46, No. 1, pp. 25-37, 1999 FIRE SEASON AND MULCH REDUCTION IN A CALIFORNIA GRASSLAND: A COMPARISON OF RESTORATION STRATEGIES Marc D. MEYER! AND PAULA M. SCHIFFMAN Department of Biology and Center for the Study of Biodiversity, California State University, Northridge, CA 91330-8303 ABSTRACT Prescribed burning and mulch reduction via grazing are two restoration strategies employed for the enhancement of native flora in California grasslands. However, the effectiveness of these methods to restore native species and suppress alien species is poorly understood. In particular, the effectiveness of different seasons of burning to restore native vegetation has been attributed to several factors, including plant phenology patterns (phenology hypothesis), fire intensities (intensity hypothesis), and accumulated mulch biomass (mulch hypothesis). In order to test these hypotheses and compare the efficacy of burning and grazing as restoration tools, the short-term effects of fire season and mulch reduction on grassland vegetation were evaluated in the Carrizo Plain Natural Area (San Luis Obispo Co., CA). Warm-season (late-spring and fall) burning significantly increased the cover and diversity of native vegetation and decreased the cover and seed viability of alien grasses relative to control treatments. Winter burning and mulch reduction did not increase the cover or diversity of native plants and were only moderately effective at reducing alien plant cover. Seed germination data showed that the seeds of one common native plant species, Phacelia ciliata Benth, responded positively to fire. These results indicated that fire season is a significant factor in grassland restoration, and that the success of different fire seasons for restoration is determined by plant phenology patterns, season-specific fire intensities, and potentially the removal of all mulch biomass. Warm-season prescribed burning and not grazing or cool-season burning is the most effective strategy for restoring native annual vegetation to California grasslands. Fire is an important restoration tool for ecolo- gists, yet very few studies have examined the sig- nificance of fire season for the regeneration of plant communities. Controlled burning programs allow resource managers to simultaneously reduce unde- sirable species and enhance desirable species (Towne and Owensby 1984; Parsons and Stolhgren 1989; Whisenant and Ursek 1990; Howe 1995). Such a shift in community composition may be achieved by burning strategies that incorporate the effects of fire season. These management strategies have proven to be particularly effective in the res- toration of a variety of native grassland communi- ties, including California bunchgrass (Menke 1992), Texas wintergrass (Whisenant et al. 1984), North American fescue (Grilz and Romo 1994), Kansas tallgrass prairie (Towne and Owensby 1984), Wisconsin tallgrass prairie (Howe 1995), and California valley grassland (Parsons and Stohl- gren 1989). California’s valley grassland is an example of an annual dominated community in which fire is a nat- ural and regular phenomenon (Parsons 1981). Al- though historic accounts of species composition are Sparse, it is believed that native winter annuals once dominated this community in the relatively arid central and southern part of the state (Wester 1981; Keeley 1990). Today, however, these grasslands are ' Present address: Graduate Group in Ecology, Depart- ment of Wildlife, Fish, and Conservation Biology, Uni- versity of California, Davis, CA 95616-8751. dominated by a variety of introduced annual grass- es primarily of Mediterranean origin (Heady 1977; Heady et al. 1992; Keeley 1990). Most of these alien annuals are tolerant of grazing and drought, and are highly competitive (Heady 1977; Heady et al. 1992; Keeley 1990). Furthermore, they produce large amounts of mulch that suppress the growth of native plants (Heady 1956; Bartolome 1979). Most of these alien grass species in California are sus- ceptible to fire and lack any adaptations to frequent burning (Evans and Young 1970; Smith 1970; Men- ke 1989). Burning reduces the aboveground bio- mass of alien grasses, and promotes growth of na- tive and alien forbs (Hervey 1949; Larson and Dun- can 1982; Menke 1989). However, most of these post-fire changes are transitory: Two years after a fire, alien grasses regain dominance, mulch biomass returns to pre-burn levels, and native forbs revert to a state of rarity (Hervey 1949; Bentley and Fen- ner 1958; Parsons and Stohlgren 1989). Several controlled burning strategies have emerged within recent years for the enhancement of native plants in grasslands. One strategy focuses on plant phenology patterns to maximize the effects of burning. When undesirable species are more vul- nerable and desirable species are less susceptible to the negative effects of fire, a controlled burn may enhance native plants or reduce exotics. Such a burning strategy has been effective in enhancing the biomass of native species in California (Parsons and Stohlgren 1989) and under-represented ‘“‘phe- nological guilds”’ elsewhere (Howe 1994, 1995). In 6 MADRONO California grasslands, carefully-timed burns have been moderately successful in the restoration of na- tive perennial bunchgrass communities (Ahmed 1983; Keeley 1990; Menke 1992), and may be a beneficial strategy for the enhancement of native annual communities. For instance, during the late spring, alien grass seeds in California grasslands are still contained in the inflorescence and have un- developed seed coats, making them extremely vul- nerable to fire (McKell et al. 1962; Menke 1992). Likewise, in the winter, alien grasses are in the seedling stage when they are highly susceptible to disturbance. A fire during either the late-spring or winter should be detrimental to alien grass survi- vorship. Unlike alien grasses, native forbs maintain a persistent soil seed bank that is relatively pro- tected from severe environmental stresses through- out the year (Young et al. 1981; Rice 1989a). Con- sequently, a carefully-timed burn in the late-spring or winter could be used to facilitate the concurrent reduction of alien grasses that are in a vulnerable stage of development and enhancement of native forbs that persist as seeds in the protected soil en- vironment. A second burning strategy for the enhancement of native diversity focuses on the relationship be- tween fire season and fire intensity. By employing a controlled burn of relatively high intensity, re- source managers can achieve elevated levels of alien species mortality. Such high-intensity fires have been successful in reducing seed survivorship of alien grasses, such as Taeniatherum caput-me- dusal (McKell et al. 1962). Native grassland spe- cies that are adapted to intense fires, however, would likely benefit from the effects of a high tem- perature burn. Such is the case with California grassland forbs, which occur in high densities fol- lowing intense, warm-season fires (Hervey 1949; Larson and Duncan 1982; Parsons and Stohlgren 1989). Consequently, the application of a high in- tensity fire may be suitable for California annual grassland restoration. Unfortunately, our under- standing of the relationship between fire season and fire intensity does not extend beyond the conven- tional wisdom that fires during the moist winter season are of low intensity and those during the dry summer season are of high intensity. A third burning strategy for the enhancement of native plants focuses on the removal of mulch. This approach considers the amount of mulch to be the principal factor determining changes in the floristic composition in California grasslands (Heady 1977; Heady et al. 1992). Mulch refers to all dead above- ground herbage, both that lying on the soil surface and that remaining in the canopy (Heady 1956; Heady et al. 1992). Mulch acts as an insulating lay- er that promotes the establishment of alien grasses and suppresses the growth of native seedlings (Evans and Young 1970; Smith 1970). In the ab- sence of mulch, grassland composition changes to one dominated by native and alien forbs (Heady [Vol. 46 1956, 1965). Based on these observations, previous authors have assumed that burning and grazing cause similar changes in grassland composition (Heady 1977; Heady et al. 1992). Furthermore, a burn that is timed to occur when it has the greatest capacity to reduce mulch should be the most effec- tive for the restoration of native forbs (although such a burn would also be expected to enhance alien forbs; Hervey 1949; Heady 1972, 1977; Heady et al. 1992). However, to date, no study has directly compared the effects of fire and mulch re- duction on a California annual plant community. Consequently, it is not clear to what extent the di- rect effects of fire (e.g., high temperatures during burning) and the indirect effects of fire (e.g., higher soil moisture depletion and soil surface temperature fluctuations following mulch reduction; Evans and Young 1970) influence composition changes in Cal- ifornia grasslands. The purpose of this study was to directly com- pare the effects of fire and grazing, and evaluate the effectiveness of different burning regimes for simultaneously reducing alien plant cover and en- hancing native plant cover and diversity in a Cali- fornia annual grassland. Three hypotheses, each ad- dressing a specific burning strategy, were tested. These hypotheses state that the optimal burn treat- ment for the restoration of native plants is (1) burn- ing when alien grass seeds or seedlings are most susceptible to disturbance (the phenology hypoth- esis), (2) burning when fire temperatures attain their greatest annual intensity (the intensity hypothesis), or (3) burning when fire has the greatest capacity to remove mulch (the mulch hypothesis). The spe- cific burn season conforming to the intensity and mulch hypotheses were determined from fire tem- perature and mulch biomass data, respectively. The most effective burn seasons according to the phe- nology hypothesis were determined from phenolo- gy patterns of California annual plants. It is impor- tant to note that these hypotheses are not mutually exclusive, and that the optimal burn season may encompass the expected predictions of two or even all three hypotheses. METHODS Experimental site. The study site was located at the Carrizo Plain Natural Area in San Luis Obispo County, CA, a 81,000 ha preserve in the southwest corner of the San Joaquin Valley region. It is the largest remaining fragment of valley grassland in the state. The site was on a flat part of the plain at an elevation of 595 m. The soil is clayey and sea- sonally dry (from April-May to October—Decem- ber) and derived from Miocene marine sandstones, | siltstones, and shales of the surrounding Caliente and Temblor Ranges (Reid et al. 1993). Maximum temperatures average 9°C in January and 29°C in July. Precipitation, which occurs almost exclusively as winter rain, averages about 14.5 cm annually and ee Cen a eer ae a ne 1999] varied during this study from 31.1 cm/year (1994— 1995) to 14.8 cm/year (1995-1996; Buttonwillow, CA; National Oceanic and Atmospheric Adminis- tration 1996). Vegetation at the site was dominated by fast-growing winter alien annuals. 97.2% and 1.5% of the total annual vegetation cover was com- prised of alien and native plants, respectively. Grass cover at the site was 90.6%, forb cover was 8.1%, and bare ground was 1.3%. The site was dominated by four alien grass species: Bromus madritensis L., Hordeum murinum L., Avena barbata Link, and Av- ena fatua L. (in order of abundance; nomenclature follows Hickman 1993). Other less common plant species include the alien species Erodium cicutar- ium (L.) L. Hér, Bromus diandrus Roth, B. hordea- ceous L., B. tectorum L., Lactuca serriola L., and the native species Amsinckia tessellata A. Gray, Monolopia lanceolata Nutt., and Lotus humistratus E. Greene. Mulch, which is characteristic of grass- dominated late-successional California grasslands, constituted an additional 99.3% cover in the exper- imental plots. Burrowing rodents, such as Thomo- mys bottae (gophers) and Dipodomys spp. (Kan- garoo rats) were absent from the study plots but present in adjacent areas. Like grasslands through- out the region, the site had a long history of cattle and sheep grazing and periodic dryland farming (i.e., site has been plowed; Cronise 1868; Burcham 1957; Preston 1981; Stromberg and Griffin 1996). During but not prior to the period of this study, livestock were excluded from the research site by a barbed-wire fence. Experimental design. Thirty 6 X 6 m plots were arranged in a rectangular grid and surrounded by a barbed-wire fence. Each individual experimental plot was surrounded by a 2 m buffer zone that was mowed for the duration of this study (June 1995 to May 1996). Six replicate plots were randomly sub- _ jected to each of the following experimental treat- ments: (1) unmanipulated (control), (2) mulch re- duction, (3) late-spring burn (17 June 1995), (4) fall burn (24 Sept. 1995), or (5) winter burn (9 Feb. 1996). The amount of time it took to burn each plot increased from late-spring (15—30 seconds) to fall (30—60 seconds) to winter (60—120 seconds). Fires were ignited by starting a backing fire followed by a flanking fire and finally a head fire. This fire ig- _ nition sequence was used to ensure fire safety rather _ than for any specific experimental reason. Wind conditions during each burn were <5 mph, and _ each fire produced flames that were estimated to be 0.5 to 1.5 m in height. Soil surface burn tempera- _ tures were measured by placing a temperature sen- sitive indicator profile (Omega Engineering Inc.) on _ the soil surface in the center of each plot. Fall- burned plots also contained a temperature indicator profile at the plot periphery in order to evaluate the variability in fire temperature within plots. Each temperature profile contained sixteen waxes with a _ Tange of melting points (ranging from 107°C to MEYER AND SCHIFFMAN: FIRE SEASON AND MULCH REDUCTION Zi 427°C) that were pasted on a 20 X 20 cm ceramic tile. Mulch reduction plots were mowed with an Echo weedeater and the mulch canopy was re- moved with a hand rake and by hand on 30 Sep- tember 1995. This treatment was done to simulate the effects of grazing on mulch biomass. Care was taken not to rake the soil surface and the bottom several centimeters of the mulch canopy was left in place in order to avoid removing seeds from the soil surface seed bank. In addition, the basal 6 cm of the mulch canopy was not removed in order to minimize disturbance to this seed bank. Vegetation Sampling. Vegetation in the experi- mental plots was sampled once in April 1995 for pre-treatment data and once in April 1996 for post- treatment data. A single post-treatment sampling period appeared sufficient to capture the short-term effects of each treatment on above ground vegeta- tion, since: (1) previous authors have shown the effects of fire and mulch reduction to occur pri- marily in the first post-burn or post-mulch reduction year in California grassland (Heady 1956; Parsons and Stohlgren 1989), and (2) our visual survey of the treatment plots in the second post-treatment growing season (using a Daubenmire [1959] cover class system and pooling spieces according to the following groups: alien grasses, aliens forbs, native forbs), revealed that grasses had regained domi- nance and native forbs had become rare (<5% cov- er). The point-intercept method was used for both censuses, and estimates of percent cover for each plant species and species diversity of native plants were determined (Barbour et al. 1987; Schiffman 1994; Sawyer and Keeler-Wolf 1995). For the pre- treatment census, three 6 m long parallel transects were randomly positioned along an east-west axis within each plot. Pins were placed at 20 cm inter- vals along each transect (Schiffman 1994) and the plant species (or bare ground) encountered at each pin point were recorded (for a total of 90 sample points per plot). Avena barbata and A. fatua were lumped because of their delayed phenology and the difficulty in distinguishing them from one another given only vegetative characters. In the post-treat- ment census, eleven 6 m parallel transects were systematically placed along an east-west axis with- in each plot. Pins were placed at 50 cm intervals along this transect and data were recorded in the same manner as in the pre-treatment census (121 total sample points per plot). Analyses of variance of the pre-treatment data revealed that there were no significant differences among plots in terms of the percent cover of any plant species encountered (F,, > 7.71, P > 0.05 for each species), with the exception of Amsinckia tessellata (F,, = 14.26, P = (0.002), which was rarely encountered in either census. In order to estimate the relative biomass of mulch among treatments, the height of the mulch canopy was sampled in April 1996 in ten randomly selected locations within each plot. 28 MADRONO Seed bank and seed viability tests. Ten soil sam- ples were collected for seed bank assessment from randomly selected points within 5 control plots (collected on 30 September 1995), 6 late-spring burn plots (collected on 24 September 1995), and 6 fall burn plots (collected on 9 December 1995). Burned plot seed samples were collected following burn treatments, and all seed samples were collect- ed prior to winter rainfall. Cores of dry soil (6 cm in diameter and 5 cm in depth) were extracted using a cylindrical bulb planter and transported to the lab- oratory where they were stored in plastic bags for 3-12 weeks at room temperature. The soil samples were then thoroughly mixed with vermiculite (1:1 soil-to-vermiculite ratio), placed in 5 cm diameter plastic pots in a greenhouse, and watered for a pe- riod of 4—6 weeks. All germinating forb seedlings were identified to species when possible. All ger- minating grasses were lumped into a single grass category. Soil surface seed samples were collected from locations adjacent to each soil coring point within the control, late-spring, and fall burned plots. These samples were used to estimate seed viability of the soil surface seed bank following each treatment; samples were not used to estimate seed rain. Seed samples were gathered by hand at the same time that soil seed bank samples were collected, placed in plastic bags, and stored in the laboratory at room temperature for a period of 3—12 weeks. Seeds of Bromus madritensis, Hordeum murinum, Avena barbata and Avena fatua. (the four most abundant plant taxa in the Spring of 1995) were separated from the rest of the mulch material and placed on moist filter paper in petri dishes. Avena spp. were combined since identification to species was not possible. Each petri dish contained 10—20 seeds, and all dishes were kept moist for the duration of the seed viability study. The total number of tested seeds varied among the three grass species due to variability in sample seed densities. Bromus mad- ritensis seed viability trials consisted of twenty seeds per plot, times six replicates. Hordeum mu- rinum seed viability trials consisted of 100 seeds per plot, times six replicates. Avena seed viability trials consisted of 20 seeds per plot times three to five replicate plots. The petri dishes containing these seeds were placed in a growth chamber set at a constant 20°C and a 12-hour light/dark cycle. In addition, since Avena spp. are known to exhibit ex- tended seed dormancy (Richardson 1979), petri dishes containing Avena seeds were treated with 5 ml of 10°? M gibberellic acid (GA,) solution to ensure radicle emergence of viable seeds (Richard- son 1979). Seeds of all three taxa were scored as viable if they exhibited radicle emergence within a 20-day period. Seed samples of the native forbs Phacelia ciliata, Monolopia lanceolata, and Amsinckia_ tessellata were also collected from experimental plots (June 1996), placed in plastic bags, and stored in the lab- [Vol. 46 oratory at room temperature for five months. One- hundred seeds of each native species collected were subjected either to an open flame from an alcohol lamp for approximately 0.5 seconds (flame treat- ment) or were unmanipulated (control treatment). Seeds were then put in petri dishes (10 seeds per dish) between two moist pieces of filter paper, placed in a growth chamber set at 20°C and a 12- hour light/dark cycle, and observed for radicle emergence for a 20 day period. Data analysis. Cover percentages were calculat- ed for all major species (frequency of each species in 121 sampling points) encountered in the 1996 vegetation census. In addition, percent cover data for grasses versus forbs and native versus alien plants were determined by pooling species belong- ing to each of these categories. Species diversity estimates were calculated using a modified Shan- non index of diversity (H’ = —{Pn(P, + 0.0001); Magurran 1988). This index yielded nearly identi- cal results as the standard Shannon index and was necessary to compensate for zero values of diver- sity encountered in samples lacking species in the native or alien plant guilds. Soil surface tempera- tures during burning were estimated using the me- dian melting point temperature between the highest | melted and lowest unmelted indicator values from | each temperature profile. Percent cover, species di- versity, litter height, seed viability, seed bank, and | soil surface burn temperature data were analyzed | with Model I one-way ANOVA. All multiple pair- wise comparisons were analyzed using a Tukey’s HSD test with a Bonferroni adjusted a (Rice 1989b). ANOVA assumptions of randomness, in- dependence, and normality were generally shown . to be met. The ANOVA assumption of equality of | variances was assessed using the F,,,, test for ho- | moscedasticity. In cases where this last assumption | was not met, the nonparametric analog of ANOVA, | the Kruskal-Wallis test, was used to compare ex- perimental treatments. In the native seed flame- treatment experiments, a x’ test was used to test if | seed germination was significantly enhanced by fire | (for each species, samples were pooled among pre- | treatment plots). Differences were considered to be . significant at a = 0.05. All statistical tests were — conducted using SYSTAT version 5.2. RESULTS Fire intensities. Fire intensities differed signifi- — cantly among the three burn seasons (F,, = 69.97, P < 0.001) (Fig. 1). In particular, the winter burn | was significantly lower in intensity than the late-_ spring or fall burns. Fall burning yielded the high- | est soil surface burn temperatures, but these tem- peratures were not significantly different than late- | spring burned plots. In addition, coefficients of variation for soil surface burn temperatures within — a burned plot and among burned plots were similar (CV = 9.9 and 12.2, respectively). 1999] 400 ANOVA: F=69.97 df=2 P<0.001 300 200 100 Soil Surface Burn Termperature ( °C) Winter burn Fall burn Late-Spring burn Fic. 1. Mean fire temperature (+SE) for winter, fall, and late-spring burns. Different lowercase letters indicate sig- nificant differences between treatments in this and follow- ing figures (n = 6 per treatment). General patterns of post-fire vegetation cover. The percent cover of alien annual plants was sig- nificantly greater in the control plots than in the burned plots (Fig. 2a). Mulch reduction plots were intermediate with respect to alien plant cover. Among the three burn treatments, alien plant cover was highest in the winter burned plots and lowest in the late-spring and fall burned plots. In contrast to the alien cover data, the percent cover and di- versity of native plants were significantly lower in the control, mulch reduction, and winter burn treat- ments (H,, = 21.47, P < 0.001 for native cover; H,, = 23.05, P < 0.001 for native diversity). Fall and late-spring burn treatments had the greatest cover and diversity of native plants, one order of magnitude greater than the control, mulch reduc- tion, or winter burned plots (Fig. 2b and 3, for cov- er and diversity, respectively). Both the grass and forb guilds responded strong- ly to the fire and mulch reduction treatments. Alien grass cover (Fig. 4a) was similar to the cover of all alien plants (Fig. 2a) in treatment plots. It was high- est in the control plots, intermediate in the mulch reduction plots, and lowest in the burned plots. Moreover, among burn treatments, grass cover was lowest in the late-spring burn treatment, interme- diate in the fall burn treatment, and greatest in the winter burned plots. Interestingly, patterns of forb cover (Fig. 4b) were not similar to the patterns of native plant cover (Fig. 2b) in these experimental plots. Instead, the cover of forbs was greatest in late-spring and control plots, and lowest in the mulch reduction, winter burned, and fall burned plots. Species cover. The native annual forbs Monolo- _ pia lanceolata and Phacelia ciliata had significant- _ ly higher cover following the fall and late-spring burn treatments (H,, = 21.36; P < 0.001 for M. MEYER AND SCHIFFMAN: FIRE SEASON AND MULCH REDUCTION 20 (a) 200 ANOVA: F=66.76 df=4 P<0.001 O > O O c & oO 100 cS Oo < 3S Control Mulch Winter Fall L-Spr (D) Kruskal-Wallis test: H=21.47 df=4 P<0.001 © > 20 O O c & Ou @ = 10 (40) Zz oS Control Mulch Winter Fall L-Spr Fic. 2. Mean % alien (a) and native (b) plant cover (+SE) for the five treatments. Note that y-axis scales in these paired figures and the following paired figures are different. Kruskal-Wallis test: H=23.05 df=4 P<0.001 = 0.50 7) _ @ fe «0.40 7p) @ —. ty 0:30 Io o 7) @ 0.20 = — ig) £& 0.10 ge ae 0.00 SAREE A cae FORE TSNE SS SRS Control Mulch Winter Fall L-Spr Fic. 3. Mean native plant species diversity (+SE) as in- dicated by a modified Shannon index of diversity (H’). 30 MADRONO ANOVA: F=41.82 df=4 P<0.001 50 % Alien Grass Cover SS SC Mulch Winter Control Fall L-Spr (b) Kruskal-Wallis test! H=17.46 df=4 P=0.002 d % Forb Cover Control Mulch Winter Fall Fic. 4. Mean % alien grass (a) and forb (b) cover (+SE) for the five treatments. L-Spr lanceolata; H,, = 25.25, P < 0.001 for P. ciliata) (Fig. 5). These species were completely absent from the control and mulch reduction treatment plots and had <1% cover in the winter burned plots. Most of the native cover found in fall and late-spring burned plots (95.5% and 88.8%, respec- tively; Fig. 2b) was due to the combined presence of these two native forbs. Other native forbs, in- cluding Amsinckia tessellata and Lotus humistratus, had very little cover in any of the treatments. No significant treatment effect was detected for either of these uncommon native species (H,, = 2.20, P = 0.699 for A. tessellata; H,, = 5.78, P = 0.216 for L. humistratus). Patterns of abundance of individual alien forb species were different from the trends detected for the native forb species. Erodium cicutarium, an alien grassland forb, had significantly greater cover in late-spring burned plots (Fig. 6a). All other treat- ments yielded relatively low amounts of cover for this alien forb. In contrast, the alien forb Lactuca serriola had the greatest cover in control plots, fol- [Vol. 46 (a) Monolopia lanceolata Kruskal-Wallis test: H=21.36 df=4 P<0.001 b 4 ae O O m4 . 2 1 Seeks Ss] [eeaeee 0 SEs Control Mulch Winter Fal L-Spr (b) Phacelia ciliata Kruskal-Wallis test: H=25.25 df=4 P<0.001 b a ® > O O os Control Mulch Winter Fall L-Spr Fic. 5. Mean % cover (+SE) of the native forbs Mon- olopia lanceolata (a) and Phacelia ciliata (b). lowed by the late-spring burned plots (Fig. 6b). Mulch reduction, winter burned, and fall burned plots contained relatively low cover of L. serriola. Four abundant alien grass species, Bromus mad- ritensis, Hordeum murinum, Avena barbata, and Avena fatua, collectively accounted for most of the grass cover in all treatment plots (Fig. 4a). These species tended to have their greatest cover in un- burned plots, with the exception of the Avena spp. The cover of B. madritensis was lowest following winter, fall, and late-spring burn treatments (Fig. 7a). Late-spring and fall burning also significantly | reduced the cover of H. murinum (Fig. 7b). Insig- nificant differences were found between experi- | mental treatments for Avena spp. (F,4 = 2.57, P = 0.062). Similarly, Bonferroni analyses revealed no | significant differences (P > 0.05) between all pair- | wise treatment comparisons of Avena spp. cover. Mulch reduction and late-spring burn treatments | had the lowest cover of Avena spp. (Fig. 7c). Patterns of post-fire mulch biomass. Burning and — mulch reduction both significantly reduced the © i} 1999] (a) Erodium cicutarium Kruskal-Wallis test: H=8.27 df=4 P=0.082 20 _ oO > O O 3S 10 0 SS “ RSS SS Ss Ss Ss Control Mulch Winter Fall L-Spr (b) Lactuca serriola Kruskal-Wallis test: H=24.88 df=4 P<0.001 a 30 _ S O 20 O oS 10 Control Mulch Winter Fall L-Spr Fic. 6. Mean % cover (+SE) of the alien forbs Erodium cicutarium (a) and Lactuca serriola (b). height of the mulch canopy (H,, = 23.09, P < 0.001; Fig. 8a). As would be expected, the mulch canopy was highest in the control plots. It was sig- nificantly shorter in the winter burned and mulch reduction plots, and virtually non-existent in the fall and late-spring burned plots. Bare ground was sub- stantially greater in fall and late-spring burned plots and lower in control, mulch reduction, and winter burned plots (H,, = 23.14, P < 0.001; Fig. 8b). Seed viability and seed bank estimates. Seed bank data complemented the vegetation data. The _ densities of germinating forb seeds were signifi- _ cantly greater in plots burned in the late-spring than those burned in the fall or left unburned (control; | Fig. 9a). This high density of forb seeds in the late- spring can be attributed to the high density of Er- odium cicutarium seeds that remained viable after - the late-spring burn treatment (Fig. 9a). Germinat- ing grass seed bank densities, on the other hand, were greatest in the control plots and lowest in the _ fall and late-spring burned plots (Fig. 9b). MEYER AND SCHIFFMAN: FIRE SEASON AND MULCH REDUCTION JI (a) Bromus madritensis 80 ANOVA: F=44.08 df=4 P<0.001 60 40 20 0 Control Mulch Winter Fall (b) Hordeum murinum Kruskal-Wallis test: H=21.53 df=4 40} _a i 0) se) > O O 20 oS 10 0 Saas Control Mulch Winter Fall L-Spr (c) Avena spp. 30 ANOVA: F=2.57 df=4 P=0.062 40 30 20 10 fe) 2 ; <= Control Mulch Winter Fall L-Spr Fic. 7. Mean % cover (+SE) of the alien grasses Bromus madritensis (a), Hordeum murinum (b), and Avena spp. (c). For Avena spp., there are no significant pairwise dif- ferences between treatments. Seed viabilities for the alien grass taxa (B. mad- ritensis, H. murinum, and Avena spp.) were highest in control plots and lowest in late-spring burned plots (Fig. 10). Although, the differences in viabil- ity between late-spring and fall burn treatments were significant for Avena spp. seeds, these differ- ences were not significant for the seeds of the other two alien grass species. In the native seed viability tests, 32% of flame- treated Phacelia ciliata seeds and 1% of control seeds demonstrated radicle emergence, indicating 32 MADRONO Height of Standing Mulch (cm) Mulch Winter Fall L-Spr Control (b) Kruskal-Wallis test: H=23.14 df=4 P<0.001 xo) c 20 = 2 (o) ced) , a 6 10 o~ Control Mulch Winter Fall L-Spr Fic. 8. Mean (+SE) height of the mulch canopy (a) and % bare ground (b) for the five treatments. that P. ciliata seed germination was significantly enhanced by fire (x? = 34.87, df = 1, P < 0.001). No flame-treated or control seeds of Monolopia lanceolata and Amsinckia tessellata germinated. DISCUSSION Burn season restoration strategies. Fire season has a significant influence upon the regeneration patterns of a variety of terrestrial plant communities (Martin 1983; Towne and Owensby 1984; Whis- enant et al. 1984; Menke 1992; Glitzenstein et al. 1995; Howe 1995). In this study, fire season influ- enced short-term vegetation cover and diversity, seed viability, and fire intensity in a California an- nual grassland. Burn treatments also varied in their effect on native annual vegetation: warm season burns (fall and late-spring) were far more effective at increasing native species cover and diversity and at reducing the cover of alien species than cool sea- son burns (winter). In particular, late-spring burning was most effective at enhancing native cover and [Vol. 46 ANOVA: F=8.95 df=2 P<0.001 (Forbs), F=8.52 df=2 p=0.004 (Erodium cicutarium) b 14 Forbs Erodium cicutarium # of Seeds per Sample rS oOo NM fF OD Control Fall burn L-Spring burn (b) 200 ANOVA: F=69.97 df=4 P<0.001 100 # of Grass Seeds per Sample 0 S Fall burn L-Spring burn Control Fic. 9. Mean seed bank densities (+SE) of forbs and Erodium cicutarium (a) and alien grasses (b) for control, fall burn, and late-spring burn treatments. ANOVA: Bromus: F=36.7 df=2 P<0.001 Hordeum: F=124.6 df=2 P<0.001 Avena: F=48.2 df=2 P<0.001 100 [J Control Fall burn Late-Spring burn © oO 60 40 % Seed Viability 20 0 Avena Bromus Hordeum Fic. 10. Mean alien grass seed viability (+SE) for con- | trol, fall burn, and late-spring burn treatments. | 1999] reducing alien grass cover and seed viability (al- though it did increase the cover of E. cicutarium). These results indicate that fire season is an impor- tant factor for the restoration of native annuals in California. Yet, the underlying mechanism of the success of warm season burning (particularly late- spring burning) is unclear. Consequently, it is not known whether the success of warm season burning is a result of fire season and plant phenology pat- terns (phenology hypothesis), a fire season-intensity relationship (intensity hypothesis), or increased bare ground (mulch hypothesis). Each of these hy- potheses are evaluated below. The phenology hypothesis predicts that the most effective burn season for the restoration of native plants occurs when alien plants are most vulnerable to fire. During the winter in California grasslands, alien plants are in a vulnerable seedling stage (Chiariello 1989; Heady et al. 1992). Unlike alien grasses which lack inter-annual seed dormancy, na- tive plants maintain a persistent seed bank in the protected soil environment (Young et al. 1981; Rice 1989a). Given these phenology patterns (particular- ly the greater vulnerability of alien species) it was predicted that the winter season would be an ideal time in which to burn grassland vegetation; this is assuming that substantial seed germination of alien grasses has already occurred and that fuel loads are of sufficient biomass to facilitate a fire. Native and alien plant cover data, however, did not support this prediction: winter burning was only partially effec- tive at reducing alien cover. Moreover, it failed to increase native cover and diversity at least during the winter immediately following burning. There are three possible explanations for these results. The failure of winter burning to immediately re- store native vegetation may be due to a delay in the response of native species to fire. For instance, in a previous experiment, native species such as Monolopia lanceolata did not show increased cover following a winter burn until the second post-fire year (P. M. Schiffman unpublished). However, based on visual surveys of all of the experimental plots (see Methods), we found no delayed effect of fire in the second year of this study. Alternatively, winter fires may not burn hot enough to induce na- tive seed germination and cause sufficient alien plant mortality. Lastly, fire may kill germinating na- tive seedlings along with alien seedlings, since the phenologies of these two groups are similar. Ad- ditional study of the effects of fire season will be needed to determine the plausibility these latter two possibilities. The phenology hypothesis predicts that late- Spring is also an optimal burn season for the res- toration of native vegetation. In the late-spring, alien grass seeds have a high moisture content and incompletely developed seed coats and, hence, are more vulnerable to fire than later in the season when their seed coats are dry and hardened (McKell et al. 1962). Thus, based purely on plant MEYER AND SCHIFFMAN: FIRE SEASON AND MULCH REDUCTION 33 phenology patterns, a fire in the late-spring would be expected to have a greater negative impact upon alien grass seeds than a fire in the fall. This pre- diction was supported by the data: alien grass cover and seed viability were lower in late-spring burned plots. In particular, the cover and seed viability of Bromus madritensis and Avena spp. decreased sig- nificantly after the late-spring burn. These results contrasted with the previous findings of Parsons and Stohlgren (1989) who found that fall burning was more effective at reducing Avena and Bromus biomass than late-spring burning. However, this in- congruence may have been due to differences in the seasonal moisture content of grass seeds be- tween the two studies. Parsons and Stohlgren (1989) observed that alien grass seeds had a higher moisture content and were more vulnerable to fire in the fall rather than in the late-spring. If we as- sume that the levels of seed moisture in this study were higher in the late-spring, then this result would explain the discrepancy between the results of Parsons and Stohlgren (1989) and this study. Moreover, it would lend support to both the phe- nology hypothesis and McKell et al.’s (1962) seed phenology post-fire succession hypothesis. Evidence for the intensity hypothesis was con- flicting. High temperature, warm-season fires gen- erated the greatest native species cover and diver- sity and lowest alien species cover. This positive association between fire intensity and native cover and negative association between fire intensity and alien cover was consistent with the intensity hy- pothesis. However, given that burn temperatures were generally higher or the same on fall burned plots, fire intensity alone does not explain why late- spring burning was more effective than fall burning at reducing alien grass cover and seed viability. This finding may be the result of a threshold re- sponse of alien grass seeds to elevated burn tem- peratures. That is, as fire intensity increases beyond some critical temperature range, it becomes an in- significant factor influencing alien grass seed mor- tality. For instance, seed viability in dry seeds of the California alien grasses, Bromus hordeaceus L. and Taeniatherum caput-medusae, drops dramati- cally at 180 to 200°C, but above and below this critical temperature range, seed viability decreases at an extremely low rate (McKell et al. 1962). The same is true of moist seeds of these two species but at a range closer to 160 to 180°C (McKell et al. 1962). Evidence for the mulch hypothesis was also con- flicting. Heady (1956) and Heady et al.’s (1992) mulch hypothesis makes two predictions: (1) the complete removal of mulch (with only bare ground remaining) should mimic the effects of fire, and (2) there is a negative linear relationship between mulch biomass and the proportion of native and exotic forb species present. The results of this study lend some support the first prediction but do not support the second. Even though 89% of the mulch 34 MADRONO TABLE 1. [Vol. 46 PREDICTIONS AND EVALUATION OF THE THREE BURN SEASON RESTORATION HYPOTHESES. The season to the left of the greater-than sign (>) indicates the more effective burn season for native species restoration. The equality sign (~) is used to signify that both treatments are similar in effectiveness. Mulch reduction refers to a significant drop in the height of the mulch canopy, whereas mulch removal refers to the removal of all mulch biomass (both in the canopy and on the soil surface). Predicted optimal burn Hypothesis a) winter b) late-spring 1. Phenology hypothesis 2. Intensity hypothesis 3. Mulch hypothesis canopy was removed in the mulch reduction treat- ment, this treatment did not enhance native cover or diversity and was significantly less effective at reducing alien cover than the fall and late-spring burn treatments. These results may have been due to the fact that compositional changes, particularly the increase in native and alien forbs, do not occur until all the mulch canopy is removed and the ground is exposed. Thus, the greater cover of native forbs in the burned plots may be a result of the increased amount of bare ground following burning and not changes in total mulch biomass. In partic- ular, the positive association between bare ground and native cover that was observed in this study supports this conclusion. Furthermore, bare ground has been associated with specific environmental cues that enhance germination or seedling survi- vorship in several annual forb species (Rice 1989a). For instance, Rice (1985) found that Erodium bo- trys (Cav.) Bertol and Erodium brachycarpum (Godron) Thell. had significantly higher germina- tion on bare ground than under litter. Similarly, in the absence of dead grass blades of Poa annua L., emergence and survivorship of Capsella bursa-pas- toris (L.) Medikus and Senecio vulgaris L. seed- lings increase significantly (Bergelson 1990). Therefore, the amount of bare ground, rather than the biomass of mulch, was the more likely factor causing changes in native species cover and diver- sity. Of the three burn treatments evaluated in this study, the late-spring burn prediction of the phe- nology hypothesis received the most support from the experimental results (Table 1). The evidence also supported the prediction that warm-season burns were more effective than cool-season burns at restoring native vegetation (intensity hypothesis). The mulch hypothesis was partially rejected based on differences in the effectiveness of mulch reduc- tion and fire treatments for enhancing native cover and reducing alien cover. Responses of species and guilds. Individual spe- cies responded differently to fire and mulch reduc- tion treatments. The native forbs Phacelia ciliata and Monolopia lanceolata were abundant only after season or result fall > late-spring > winter a) mulch reduction ~ fall burn b) mulch removal =~ fall burn Prediction supported? a) no b) yes yes (fall & late-spring > winter) no (late-spring > fall) a) no b) prediction not tested a warm-season fire. Species such as these that re- spond positively to fire are often classified as “‘fire annuals”’ (Keeley et al. 1985). However, only P. ciliata appeared to be dependent on fire for signif- icant germination: the seeds of P. ciliata showed a significant increase in radicle emergence following flame/heat/smoke treatment, a trait found in other species of the family Hydrophyllaceae (Keeley and Fotheringham 1998) and genus Phacelia (Keeley and Keeley 1982). The alien forb Erodium cicutar- ium was classified as an ‘“‘opportunistic annual” (Keeley et al. 1985). Opportunistic annuals are abundant after fire but also occur in higher densities in areas where bare ground is common (i.e. canopy gaps). Moreover, unlike P. ciliata and M. lanceo- lata, E. cicutarium was common only on late-spring burn plots. The germination of FE. cicutarium seeds was also greatest for seed bank samples collected from late-spring burned plots. Parsons and Stohl- gren (1989) noted a similar increase in Erodium botrys density following late-spring burning, but not fall burning. Similarly, York (1997) found a de- cline in Erodium brachycarpum following a late September fire. Such a pattern may be the result of enhanced germination following exposure to the extreme temperature fluctuations of bare ground during the summer, a seed characteristic found in several species of Erodium (Rice 1985). Erodium seeds probably did not experience these wide fluc- tuations on fall burn plots since these plots were insulated with mulch throughout the summer. By the time of the fall burn, cooler, more moderate temperatures of the fall season predominated, and the conditions necessary for Erodium seed germi- nation were lacking. Species were also classified as mulch-dependent or fire-intolerant. The only mulch-dependent spe- cies was Lactuca serriola. Lactuca serriola cover was highest on control plots where mulch biomass was an order of magnitude or two greater than any of the treatment plots. In addition, Bromus madri- tensis and Hordeum murinum were considered fire- intolerant species. Bromus madritensis cover was significantly reduced in all burned plots, and H. mu- rinum cover decreased significantly in warm-season 4 : 1999] burned plots. Plants that exhibited none of the above patterns included Avena spp. and two native herbs that were uncommon in this study (Lotus humistratus and Amsinckia tessellata). Plant guilds also varied in their response to fire and mulch reduction. Grasses were negatively af- fected by the removal of mulch but most signifi- cantly reduced in burned plots. Forbs, however, had greater cover in control and late-spring burned plots and lower cover in fall and winter burn plots. These post-fire patterns in forb cover contrasted with the results of Hervey (1949) and Parsons and Stohlgren (1989), who found forb cover to be consistently greater in burned plots, regardless of the season of burning. This inconsistency in forb cover can be attributed to the abundance of the alien forb, Lac- tuca serriola, in the control plots of this study. Lac- tuca serriola is common to areas of high mulch biomass in central California grasslands (M. Meyer, personal observation). However, this species was absent from the study plots of Hervey (1949) and Parsons and Stohlgren (1989). More importantly, the observation of decreased L. serriola cover fol- lowing a fire demonstrated that not all species with- in the forb guild respond similarly to fire. Implications for grassland management. The ef- fects of fire on grassland composition have long been considered to be consistent with the effects of mulch reduction (Heady 1956, 1972; Heady et al. 1992). As a result, the influence of fire and live- stock grazing on grassland vegetation have also been viewed as equivalent, since both methods fa- cilitate a reduction in mulch biomass (Hervey 1949; Heady 1972). It is surprising, however, that virtu- ally no studies directly comparing the effects of fire and mulch reduction or fire and grazing on Cali- fornia grassland composition have ever been con- ducted. Nevertheless, the direct comparison of burned and mulch reduction plots in this study strongly suggest that fire and grazing do not facil- itate equivalent changes in community composi- tion, particularly with respect to changes in native cover and diversity. This result is consistent with Stromberg and Griffin (1996) who found no in- crease in native species cover or diversity in grass- lands recently grazed by cattle. The dissimilarity between burned and mulch reduction plots in this present study may have been the result of differ- ences in the amount of bare ground, since warm- season burned plots had significantly more post- treatment bare ground than mulch reduction plots. Exposed bare ground following fire can cause changes in soil moisture and soil surface albedo, relative humidity, and temperature (Evans and Young 1970). Alternatively, enhanced native cover and diversity may be due to direct effects of fire. While neither of these two alternatives can be re- jected with the results of this study, the latter pos- sibility cannot be discounted for two reasons. First, alien grass seed viability was significantly reduced MEYER AND SCHIFFMAN: FIRE SEASON AND MULCH REDUCTION os) following fire (a result that cannot be attributed to mulch reduction since the fall-burned seed samples were covered by a layer of mulch until they were burned). Second, radicle emergence of Phacelia cil- lata, a species endemic to California grasslands, in- creased significantly following flame treatment. These results not only indicate that the direct ef- fects of fire do have a very significant impact upon the germination and survivorship of California grassland species, but also infer that fire may have been historically important within this plant com- munity. These results indicate that the direct effects of fire do have a significant impact upon the ger- mination and survivorship of California grassland species. Consequently, this evidence suggests that fire is a more effective method than grazing for native species restoration. Successful fire management in California grass- lands will also require an understanding of the vari- able nature of plant phenology patterns within these communities. High variability in the annual rainfall in California grasslands generates high variability in the phenology patterns of grassland plants. For instance, in a dry early-rainfall year, grassland plants set seed much earlier in the season (e.g., April) than in a year of abundant, late rainfall (seeds set in June). This change in the timing of seed set with rainfall should, in turn, influence the effectiveness of a particular fire season to kill alien grass seeds and increase native plant cover. Con- sequently, the success of the mid-June burn in this study should not be considered to be the only date for the restoration of native biological diversity to a California annual grassland. Instead, a successful burn should be timed just prior to when alien grass- es set seed, whether this occurs later or earlier in the spring season. Fire and grazing are the primary tools for grass- land management. Yet, our understanding of the ef- ficacy of each of these strategies to restore native vegetation still remains largely unexplored. Cali- fornia grassland restoration and management, in particular, have long rested upon the conclusions of only a few experiments and studies that have fo- cused almost entirely on the effects of mulch re- duction (e.g., Heady 1956; Heady et al. 1965). Moreover, virtually all previous studies of Califor- nia grassland vegetation have focused on the res- toration of a single perennial bunchgrass species, Nassella pulchra (A. Hitche.) Barkworth (e.g., Ah- med 1983; Fossom 1990; Menke 1992; Dyer and Rice 1997). This study provides a new perspective on the use of fire and grazing in California grass- land management and emphasizes the influence of these management strategies on the diverse annual component of grassland communities. The conflict- ing results of this study with previous ones under- scores the need for more comparative studies of fire, grazing, and other management practices used to restore native annual vegetation to California grasslands. 36 MADRONO ACKNOWLEDGMENTS We thank Tom Valone, Paul Wilson, Jennifer Matos, Jane Bock, and Carla D’ Antonio for valuable comments and useful advice on earlier drafts of this paper. We are grateful to Justin Albert, Roy van de Hoek, Jamie Kneitel, Patrick Byrne, Steve Mason, and Dominique Joyner for assistance in the field. We are also grateful to The Nature Conservancy, U.S. Bureau of Land Management, and Cal- ifornia Department of Fish and Game for providing fence materials and the opportunity to do research at the Carrizo Plain Natural Area. Thanks go to Debbie Santiago and the BLM fire crew for managing the prescribed burns. Fund- ing for this research was provided by the Southern Cali- fornia Botanists, the California State University Pre-Doc- toral Program, and the University Corporation and Grad- uate Studies, Research, and International Programs Office at the California State University, Northridge. Additional funding and support were generously provided by the NIH-IMSO graduate training grant program (granted to M. D. Meyer). LITERATURE CITED AHMED, E. O. 1983. Fire Ecology of Stipa pulchra in the California Annual Grassland. Ph.D. dissertation, Uni- versity of California, Davis, Davis CA. BARBOUR, M. G., J. H. BURK, AND W. D. Pitts. 1987. Terrestrial Plant Ecology. Benjamin/Cummings Pub- lishing, Inc., Menlo Park, CA. BARTOLOME, J. W. 1979. Germination and seedling estab- lishment in California annual grassland. Journal of Ecology 67:273-281. BENTLEY, J. R. AND R. L. FENNER. 1958. Soil temperatures during burning related to postfire seedbeds on wood- land range. Journal of Forestry 56:737—740. BERGELSON, J. 1990. Life after death: site pre-emption by the remains of Poa annua. Ecology 71:2157—2165. BIONDINI, M. E., A. A. STEUTER, AND C. E. GRYGIEL. 1989. Seasonal fire effects on the diversity patterns, spatial distribution and community structure of forbs in the northern mixed prairie, USA. Vegetatio 85:21-31. BurcHAM, L. T. 1957. California Range Land. California Department of Natural Resources, Sacramento. CA. CHIARIELLO, N. R. 1989. Phenology of California Grass- lands. Pp. 47-58 in L. EF Huenneke and H. Mooney (eds.). Grassland structure and function: California annual grassland. Kluwer Academic Publishers, Dor- drecht, Netherlands. CRONISE, T. F 1868. The Natural Wealth of California. H. H. Bancroft & Co., New York. DAUBENMIRE, R. 1959. A canopy-coverage method of veg- etational analysis. Northwestern Science 33:43—64. Dyer, A. R. AND K. J. Rick. 1997. Intraspecific and diffuse competition: the response of Nassella pulchra in Cal- ifornia grassland. Ecological Applications 7:484— 492. EvANS, R. A., AND J. A. YouNG. 1970. Plant litter and establishment of alien annual weed species in range- land communities. Weed Science 18:697—703. Fossom, H. C. 1990. Effects of prescribed burning and grazing on Stipa pulchra (Hichc.) seedling emergence and survival. M.S. thesis. University of California, Davis, Davis, CA. GLITZENSTEIN, J. S., W. J. PLATT, AND D. R. STRENG. 1995. Effects of fire regime and habitat on tree dynamics in North Florida longleaf pine savannas. Ecological Monographs 65:441-—476. [Vol. 46 Gritz, P. L., AND J. T. Romo. 1994. Water relations and growth of Bromus inermis Leyss (smooth brome) fol- lowing spring or autumn burning in a fescue prairie. American Midland Naturalist 132:340-348. HeEApy, H. F 1956. Changes in a California annual plant community induced by manipulation of natural mulch. Ecology 37:798-812. . 1965. The influence of mulch on herbage pro- duction in an annual grassland. Proceedings of the 9th International Grassland Congress 1:391—394. . 1972. Burning and the grasslands in California. Proceedings of the Annual Tall Timbers Fire Ecology Conference 12:97—107. . 1977. Valley grassland. Pp. 491-514 in M. G. Barbour and J. Major (eds.). Terrestrial Vegetation of California, Wiley Press, New York. , J. W. BARTOLOME, M. D. Pitt, G. D. SAVELLE, AND M. C. StrRoub. 1992. California prairie. Pp. 313- 335 in R. T. Stroud (ed.). Natural Grasslands. Elsevier Science. Amsterdam, Netherlands. HERVEY, D. FE 1949. Reaction of a California annual-plant community to fire. Journal of Range Management 2: 116-121. HICKMAN. J. C. (ED.). 1993. The Jepson manual: higher plants of California. University of California Press, Berkeley, CA. Howe, H. FE 1994. Response of early- and late-flowering plants to fire season in experimental prairies. Ecolog- ical Applications 4:121—133. . 1995. Succession and fire season in experimental prairie plantings. Ecology 76:1917—1925. KEELEY, S. C. AND J. E. KEELEY. 1982. The role of alle- lopathy, heat, and charred wood on the germination of chaparral herbs. Proceedings of symposium on dy- namics and management of Mediterranean-type eco- systems. USDA Forest Service General Technical Re- port 58:1-7. KEELEY, J. E., B. A. MORTON, A. PEDROSA, AND P. TROT- TER. 1985. Role of allelopathy, heat, and charred wood in the germination of chaparral herbs and suf- frutescents. Journal of Ecology 73:445—458. . 1990. The California valley grassland. Pp. 2—23 in A. A. Schoenherr (ed.), Southern California bota- nists special publication No. 3, Claremont, CA. AND C. J. FOTHERINGHAM. 1998. Smoke-induced germination in California chaparral. Ecology 79: 2320-2336. Larson, J. R. AND D. A. DUNCAN. 1982. Annual grassland response to fire retardant and wildfire. Journal of Range Management 35:700-—703. MAGurRRAN, A. E. 1988. Ecological Diversity and its Mea- surement. Princeton University Press, Princeton, NJ. Martin, S. C. 1983. Responses of semidesert grasses and shrubs to fall burning. Journal of Range Management 36:604-—610. McKELL, C. M., A. M. WILSON, AND B. L. Kay. 1962. Effective burning of rangelands infested with medu- sahead. Weeds 10:125-131. MENKE, J. W. 1989. Management controls on productivity. Pp. 173-199 in L. EF Huenneke and H. Mooney (eds.), Grassland Structure and Function: California Annual Grassland. Kluwer Academic Publishers, Dordrecht, Netherlands. . 1992. Grazing and fire management for native perennial grass restoration in California grasslands. Fremontia 20:22-—25. NATIONAL OCEANIC AND ATMOSPHERIC ASSOCIATION. 1996. 1999] Climatological data annual summary: California. Washington, DC. Parsons, D. J. 1981. The historical role of fire in the foothill communities of Sequoia National Park. Ma- drofio 28:111—120. AND T. J. STOHLGREN. 1989. Effects of varying fire regimes on annual grasslands in the Southern Sierra Nevada of California. Madrono 36:154—168. PRESTON, W. L. 1981. Vanishing Landscapes: Land and Life in the Tulare Lake Basin. University of Califor- nia Press, Berkeley, CA. Reip, D. A., R. C. GRAHAM, R. J. SOUTHARD, AND C. AM- RHEIN. 1993. Slickspot soil genesis in the Carrizo Plain, CA. Soil Science Society of America Journal 57:162-168. Rick, K. J. 1985. Responses of Erodium to varying mi- crosites: the role of germination cuing. Ecology 66: 1651-1657. . 1989a. Impacts of seed banks on grassland com- munity structure and population dynamics. Pp. 211- 230. in M. A. Leck, V. T. Parker, and R. L. Simpson (eds.). Ecology of soil seed banks. Academic Press, New York. RicE, W. P. 1989b. Analyzing tables of statistical tests. Evolution 43:223-—225. RICHARDSON, S. G. 1979. Factors influencing the devel- opment of primary dormancy in wild oat seeds. Ca- nadian Journal of Plant Science 59:777-—784. SAWYER, J. O. AND T. KELLER-WOLF. 1995. A manual of MEYER AND SCHIFFMAN: FIRE SEASON AND MULCH REDUCTION a California vegetation. California Native Plant Soci- ety, Sacramento, CA. SCHIFFMAN, P. M. 1994. Promotion of exotic weed estab- lishment by endangered giant kangaroo rats (Dipo- domys ingens) in a California grassland. Biodiversity and Conservation 3:524—537. SMITH, T. A. 1970. Effects of disturbance on seed germi- nation in some annual plants. Ecology 51:1106—1108. STROMBERG, M. R., AND J. R. GRIFFIN. 1996. Long-term patterns in coastal California grasslands in relation to cultivation, gophers, and grazing. Ecological Appli- cations 6:1189-1211. TOWNE, G. AND C. OWENSBY. 1984. Long-term effects of annual burning at different dates in ungrazed Kansas tallgrass prairie. Journal of Range Management. 37: 392-397. WESTER, L. 1981. Composition of native grasslands in the San Joaquin Valley, California. Madrono 28:231-241. WHISENANT, S. G., D. N. UECKERT, AND C. J. SCIFRES. 1984. Effects of fire on Texas wintergrass communi- ties. Journal of Range Management 37:387-391. AND W. URSEK. 1990. Spring burning Japanese brome in a western wheatgrass community. Journal of Range Management 43:205-—208. YorK, D. 1997. A fire ecology study of a Sierra Nevada foothill basaltic mesa grassland. Madrofo 44:374— 383. Younca, J. A., R. A. EvANs, C. A. RAGUSE, AND J. R. LARSON. 1981. Germinable seeds and periodicity of germination in annual grasslands. Hilgardia 49:1—37. MADRONO, Vol. 46, No. 1, pp. 38-45, 1999 FLORA OF A VERNAL POOL COMPLEX IN THE MAYACMAS MOUNTAINS OF SOUTHEASTERN MENDOCINO COUNTY, CALIFORNIA KERRY L. HEISE University of California, Hopland Research and Extension Center, Hopland, CA 95449 ADINA M. MERENLENDER Environmental Science, Policy, and Management, University of California, Berkeley, CA 94720-3110 ABSTRACT Vernal pools in the Mayacmas Mountains of southeastern Mendocino County, CA typically occupy topographic depressions related to landslide dams and fissures. A group of pools within a 1290 ha study area range in size from 180 to 3069 m’, and are located at elevations between 329 and 902 m on slopes of oak woodland and chaparral. Eryngium aristulatum Jepson var. aristulatum and Isoetes howellii En- gelm. codominate shallow, wide-margined pools and are associated with vernal pool specialist taxa such as Gratiola, Navarretia, Plagiobothrys, and Pogogyne. Deep, narrow-margined pools are characterized by cosmopolitan wetland taxa such as Callitriche, Carex, Eleocharis, Juncus, and Ranunculus. Plant surveys conducted in 1996 and 1999 indicate no significant change in species abundance or composition between the two years. The characteristic flora and fauna of California’s vernal pools and their distribution is well docu- mented (Jain 1976; Holland and Jain 1987; Zedler 1987; Keeler-Wolf et al. 1995; King et al. 1996; Bauder and McMillan 1998; Holland 1998). Of the16 vernal pool regions in the state described by Keeler-Wolf et al. (1995), the Mendocino Region, which lies entirely in Mendocino County, is one of the least known. The purpose of this study was to provide a description of the flora of a vernal pool complex in the arid, mountainous portion of south- eastern Mendocino County based on observations from 2 years. These interior cismontane vernal pools occur on relatively unstable soils that are de- rived from marine sedimentary rocks. A number of floristic studies have been completed in the area and the occurrence of vernal pools has been men- tioned (Neilson and McQuaid 1981; Murphy and Heady 1983; De Nevers 1985; Smith and Wheeler 1991, 1992; Baad 1998), but our knowledge of ver- nal pool floristics remains poorly understood. Study area and methods. The 1290 ha study area is located in the interior north coast ranges of southeastern Mendocino County. Lying just east of the Russian River and 6 km northeast of U.S. Hwy TABLE 1. HOPLAND RESEARCH AND EXTENSION CENTER, MENDOCINO COUNTY, CA. Elevation = 947 mm. Year Jul Aug Sep Oct Nov Dec 1995-1996 5 350 1997-1998 Ze 17 4] 21S 92 1998-1999 24 184 74 paper enereeecen 101 near Hopland, CA, the area includes portions | of the University of California, Hopland Research and Extension Center (HREC) and adjacent public and private lands (Fig. 1). The area consists of moderately steep, predominately southwest-facing | slopes in the Mayacmas Mountains, with elevations ranging from 183 to 914 m. Many high gradient ephemeral creeks bisect this terrain creating a series | of parallel ridges and deep gullies. Lying approxi- | mately 65 km inland from the Pacific Ocean, the study area experiences a typical Mediterranean cli- mate of hot-dry summers and cool-wet winters. Rainfall for the winters of 1995-1996 and 1998- 1999 totaled 1074 mm and 880 mm respectively; | the 35-year average is 947 mm (Table 1). The 12 vernal pools in this study range in ele- — vation from 329 m to 902 m and are situated on ; benches originating from old landslides and soil slips. The soils are predominately fractured sand- . stones and shales (Sutherlin Series) or glaucophane ~ schist and related metamorphic rocks (Yorkville Se- | ries) of the Franciscan Formation (Gowans 1958), ° which are especially prone to landslipping. The | pool basins are underlain with a moderately com- | pact clay hardpan. Mixed oak woodland and savan- © MONTHLY RAINFALL (mm) FOR WINTERS OF 1995-1996 AND 1998-1999 AT THE UNIVERSITY OF CALIFORNIA 244 m. 35 year average = | Jan Feb Mar Apr May — Jun Tot 214 244 88 88 64 2A 1074 525, 502 150 71 11i 1546 101 287 149 a7 + 880 1999] Location of Hopland Research ae : ao. eemeters A and Extension Center e Vernal Pools N in California Contours (50m intervals) Hopland Research and Extension Center Boundary ore oy < . Sian a = ae ‘ ges LO oe Se a x \ Oo , ) ‘| CRE wy ONG =. Fic. 1. Location of vernal pools at the Hopland Research and Extension Center, Mendocino County, CA. Twining pool is located 5 km northwest of Hog pool. Township 13 North, Range 11 West. na dominate the lower elevations, whereas chapar- ral and patches of closed cone forest are common above 675 m (Murphy and Heady 1983). Some pools are shaded by mature Quercus spp. (Quercus douglasii Hook. & Am., Q. lobata Nee, Q. garra- yana Hook., Q. agrifolia Nee), but otherwise are surrounded by annual grassland or chamise chap- arral. A group of sag ponds just outside of the study area support perennial wetland taxa such as Scir- pus, Typha, and Potamogeton, which were absent from the larger vernal pools. Cattle were excluded from the Twining pool in 1996; all other pools, ex- cept Bluebird and Riley, are grazed by sheep. We visited each pool in mid April, May and June of 1996 and 1999 to inventory the vascular plants. A species list was compiled for each pool, noting all species from pool center to upper strand line or interface where there was an obvious change in species composition from the adjacent upland veg- etation. The DAFOR scale (Kent and Coker 1992): dominant (5), abundant (4), frequent (3), occasional (2), and rare (1) was used to obtain a subjective measure of species abundance for each pool. Tree canopy cover was determined from an ocular esti- mate of noon shade over the pools in June. Since all pools were round in shape the mean of 4 radii were used to determine pool area. Representative _ species new to the study site were collected and _ deposited into the HREC herbarium. RESULTS AND DISCUSSION Of the 90 vascular plants and one bryophyte _ found at the 12 vernal pools during the April—June, 1995-1996 and 1998-1999 sampling period, 63 TABLE 2. SUMMARY DATA FOR VERNAL POOLS AT THE HOPLAND RESEARCH AND EXTENSION CENTER AND ADJACENT PRIVATE LAND, MAYACMAS MOUNTAINS, MENDOCINO County, CA. Note: No grazing in Bluebird since 1989, Riley since 1956, and Twining since 1996. Rifle Coon Weather Weather Coon Junction Junction Bluebird West East Riley North South Small Large Hog Range Twining Huntley Elevation (m) Grazing (Yes, No) Pool area (sq m) HEISE AND MERENLENDER: MAYACMAS VERNAL POOLS 60 45 45 a2 105 65 100 32 40 115 68 13 20 30 33 42 Maximum water depth (cm) 110 100 90 40 70 38 30 10 100 Water depth (cm) 30 May 1996 Water depth (cm) 18 Jun 1996 38 12 80 55 53 100 Water depth (cm) 26 May 1999 Water depth (cm) 21 Jun 1999 % tree canopy cover qT 40 12 35 20 29 25 23 45 18 Be) 33 16 24 33 27 Number of plant species 40 MADRONO [Vol. 46 TABLE 3. ABUNDANCE DATA FOR VERNAL POOLS AT THE HOPLAND RESEARCH AND EXTENSION CENTER AND ADJACENT PRIVATE LAND, MENDOCINO County, CA. Data collected April-June, 1996 and 1999. 5 = Dominant, 4 = Abundant, 3 = Frequent, 2 = Occasional, 1 = Rare. Nomenclature follows Hickman (1993). Junc- Junc- tion tion Hunt- Hunt- Blue- Blue- West- West- Species ley-96 ley-99 bird-96 bird-99 96 99 Achyrachaena mollis Schauer Agrostis exarata Trin. Aira caryophyllea L. 2 ps 1 Allium unifolium Kellogg Anagallis arvensis L. 2 1 1 1 Aristida oligantha Michaux Avena barbata Link Z 1 Briza maxima L. Briza minor L. 2 2 3 2 Brodiaea elegans Hoover 1 2 1 1 Brodiaea stellaris S. Watson Bromus diandrus Roth 1 1 Bromus hordeaceus L. 2 2 1 1 Callitriche heterophylla Pursh var. bolanderi (Hegelm.) Fassett 2 2 Carex athrostachya Olney Carex feta L. Bailey Carex subbracteata Mackenzie Castilleja attenuata (A. Gray) Chuang & Heckard Centunculus minimus L. Cerastium glomeratum Thuill. 3 2 Ceratophyllum demersum L. Cicendia quadrangularis (Lam.) Griseb. Crassula aquatica L. Schonl. 2 2 Cynodon dactylon (L.) Pers. 4 3 Cynosurus echinatus L. Dactylis glomerata L. Danthonia californica Bolander Deschampsia danthonioides (Trin.) Munro 2 pi 4 4 4 3 Deschampsia elongata (Hook.) Munro Downingia cuspidata (E. Greene) E. Greene Elatine californica A. Gray Eleocharis acicularis (L.) Roemer & Schultes 4 4 3 4 Eleocharis macrostachya Britton 5 Elymus glaucus Buckley Eremocarpus setigerus (Hook.) Benth. 3 3 Eryngium aristulatum Jepson var. aristulatum 2 5) 5 5 ) 5 Festuca rubra L. Galium aparine L. 2 2 Geranium bicknellii Britton Geranium dissectum L. 1 Glyceria leptostachya Buckley Gratiola ebracteata Benth. e) + 4 4 Heterocodon rariflorum Nutt. Hordeum marinum Hudson subsp. gussoneanum (Parl.) Thell. 3 3 2 2 Z + Hypochaeris glabra L. Isoetes howellii Engelm. 5 5 2 2 Juncus bufonius L. 2 2 Juncus patens E. Meyer Juncus tenuis Willd. Juncus xiphioides E. Meyer S Lasthenia californica Lindley Lasthenia glaberrima A.DC. Leptodictyon riparium (Hedw.) Warnst. Lilaea scilloides (Poiret) Hauman Lolium multiflorum Lam. 2 3 5 4 Lolium temulentum L. Lythrum hyssopifolium L. 2 2 3 2 Madia gracilis (Smith) Keck Mentha pulegium L. ye Microseris douglasii (DC.) Schultz-Bip. 1999] HEISE AND MERENLENDER: MAYACMAS VERNAL POOLS 41 TABLE 3. EXTENDED Junc- Junc- tion tion Coon Coon Coon Coon Rifle Rifle Twin- Twin- East- East- Riley- Riley- North- North- South-South- WS- WS- WL- WL- Hog- Hog- Range-Range- ing- ing- 99 99 96 99 96 99 96 99 96 99 96 99 96 99 96 99 96 99 3 1 1 1 4 2 2 3 2 2 1 1 ] 2 2 a pe 2 1 Z ] 2 1 ] 1 l 1 ] 5 1 ] 3 3 1 2 3 2 2 2 2 2 2 2 2 1 1 1 ] 1 1 2 1 3 ] 2 ] ] ] 1 1 1 3 | 4 3 4 2 | 3 1 2 2 2 2 2 4 S, 5 1 2 4 1 ] ] ] 2 5 1 2 1 2 2 2 2 2 1 I 1 3 2 p. 3 2 4 3 2 Z ps 3 2 1 2 2 ] 2 ] ] 2 1 2 1 1 2 1 2 3 3 2 3 2 2 5 4 Z 4 2 2 1 2 1 3 1 2 1 5 5 3 2 y 2 4 4 3 4 5 5 > 4 5 5 4 5 5 3 p) 5 5 2 2 S ) 2 5 5 5 5 5 i) 5 5 4 4 2 2 1 2 1 3 1 2 l 3 3 3 3 3 3 3 | 4 2 4 4 3 2 3 3 ] 2 2 4 2 2 3 2 3 4 3 4 2 2, 3 3 Zz 1 5 5 5 5 5 4 =) 5 5 i) 2 1 4 2 2 2 2 l 2 2 4 1 ] Zz 1 2 ] 3 1 2 S 2 > 2 3 2 p2 5 - 4 4 4 4 2 2 5 a 2 4 4 2 2 3 Zz £) g) 1 2 Zz 2 2 Z ] 1 ] 2 2 2 2 l 2 5 5 40 MADRONO [Vol. 46 TABLE 3. ABUNDANCE DATA FOR VERNAL POOLS AT THE HOPLAND RESEARCH AND Ex’ SION CENTER AND ADJAc ENT PRIVATE LAND, MENDOCINO County, CA. Data collected April—June, 1996 and 1999. 5 = Dominant, 4 = Abundant, 3 = Frequent, 2 = Occasional, 1 = Rare. Nomenclature follows Hickman (1993). Hunt- ley-96 Species Hunt- ley-99 bird-96 bird-99 Blue- Blue- Junc- tion West- 96 June- tion West- 99 Achyrachaena mollis Schauer Agrostis exarata Trin. Aira caryophyllea L. Allium unifolium Kellogg Anagallis arvensis L. Aristida oligantha Michaux Avena barbata Link Briza maxima L. Briza minor L. Brodiaea elegans Hoover 1 Brodiaea stellaris S. Watson Bromus diandrus Roth Bromus hordeaceus L. Callitriche heterophylla Pursh var. bolanderi (Hegelm.) Fassett Carex athrostachya Olney Carex feta L. Bailey Carex subbracteata Mackenzie Castilleja attenuata (A. Gray) Chuang & Heckard Centunculus minimus L. Cerastium glomeratum Thuill. 3 Ceratophyllum demersum L. Cicendia quadrangularis (Lam.) Griseb. Crassula aquatica L. Schénl. Cynodon dactylon (L.) Pers. Cynosurus echinatus L. Dactylis glomerata L. Danthonia californica Bolander Deschampsia danthonioides (Trin.) Munro 2) Deschampsia elongata (Hook.) Munro Downingia cuspidata (E. Greene) E. Greene Elatine californica A. Gray Eleocharis acicularis (L.) Roemer & Schultes 4 Eleocharis macrostachya Britton S) Elymus glaucus Buckley Eremocarpus setigerus (Hook.) Benth. 3 Eryngium aristulatum Jepson var. aristulatum 2 Festuca rubra L. Galium aparine L. 2 Geranium bicknellii Britton Geranium dissectum L. Glyceria leptostachya Buckley Gratiola ebracteata Benth. 3 Heterocodon rariflorum Nutt. Hordeum marinum Hudson subsp. gussoneanum (Parl.) Thell. 3 Hypochaeris glabra L. Isoetes howellii Engelm. 5 Juncus bufonius L. 2 Juncus patens E. Meyer Juncus tenuis Willd. Juncus xiphioides E. Meyer Lasthenia californica Lindley Lasthenia glaberrima A.DC. Leptodictyon riparium (Hedw.) Warnst. Lilaea scilloides (Poiret) Hauman Lolium multiflorum Lam. Lolium temulentum L. Nw is) Lythrum hyssopifolium L. 2 Madia gracilis (Smith) Keck Mentha pulegium L. Microseris douglasii (DC.) Schultz-Bip. nN Nn i) od nw tv i) i) is) 1999] HEISE AND MERENLENDER: MAYACMAS VERNAL POOLS 41 TABLE 3. EXTENDED Junc- Junc- tion tion Coon Coon Coon Coon Rifle Rifle Twin- Twin- East- East- Riley- Riley- North- North- South-South- WS- WS- WL- WL- Hog- Hog- Range-Range- ing- _ ing- 99 99 96 99 96 99 96 99 96 99 96 99 96 99 96 99 96 99 3 1 1 1 4 2 2 2 3 2 2 1 1 1 2 2 2 2 2 1 2 1 2 1 1 1 1 1 1 1 5 4 1 1 3 3 1 2 3 2 2 2 2 yd 2 2 2 1 1 1 1 1 1 2 1 3 1 2 1 1 1 1 1 1 3 1 4 3 4 2 1 3 1 2 2 2 2 2 4 5 5 1 2 4 1 1 1 1 2 3 1 2 1 2 2 2 2 2 1 1 1 3 2 2D Joe et a a aD SB 2 1 2 2 1 2 1 1 2 1 2 2 1 1 2 1 2 3 3 2 3) 2 2 S} 4 2 4 2 2 1 2 1 3 1 2 1 5 5 3 2 2 2 4 4 3 4 5 5 5) 4 5) 3) 4 S) 5 3 2 5 5 2 2 3 3 2 5 5) 5 5 5 5 5 5 4 4 2 2 1 2 1 3 1 2 1 3 3 3 3 3 3 3 1 4 2 4 4 3 2 3 3 1 2 ee 4 2 2 Bin pat Weg! adie ee aia ea eee ie 2 1 5 5) 5} 5 5 4 5 5) 5 5 2 1 4 2 2 2 2 1 2 2 2 4 1 1 2 1 2 1 3 1 3 ai ne B 2 3 rae 5 5 4 4 4 4 2 2 5 5 2 4 Dias DE desi ul baid Blt 1 2 2 2 2 2 1 1 1 2 2 2 2 1 2 3) 5 2 42 MADRONO TABLE 3. CONTINUED Species Mimulus guttatus DC. Mimulus pilosus (Benth.) S. Watson Montia fontana L. Navarretia intertexta (Benth.) Hook. Phalaris aquatica L. Plagiobothrys bracteatus (T. J. Howell) I. M. Johnston Poa annua L. Poa secunda J. S. Presl Pogogyne zizyphoroides Benth. Polypogon interruptus Kunth Polypogon monspeliensis (L.) Desf. Potamogeton foliosus Raf. Psilocarphus tenellus Nutt. Ranunculus aquatilus L. Ranunculus lobbii (Hiern) A. Gray Rumex crispus L. Silene gallica L. Sisyrinchium bellum S. Watson Spiranthes pornifolia Lindley Spirodela polyrrhiza (L.) Schleiden Stellaria media (L.) Villars Taeniatherum caput-medusae (L.) Nevski Trichostema laxum A. Gray Trifolium dubium Sibth. Trifolium variegatum Nutt. Trisetum canescens Buckley Triteleia hyacinthina (Lindley) E. Greene Veronica peregrina L. subsp. xalapensis (Kunth) Pennell Vulpia bromoides (L.) S. FE Gray Vulpia myuros (L.) C. Gmelin Zigandenus micranthus Eastw. Total number of species (70%) were native species (Appendix 1). Species richness ranged from 12 at Coon North pool to 45 at Hog pool (Table 2). Centunculus minimus L., Crassula aquatica (L.) Schénl, Deschampsia dan- thonioides (Trin.) Munro, Downingia cuspidata (E. Greene) E. Greene, Elatine californica A. Gray, Eryngium aristulatum Jepson var. aristulatum, Gra- tiola ebracteata Benth., [soetes howellii Engelm., Plagiobothrys bracteatus (T. J. Howell) I. M. John- ston, Pogogyne zizyphoroides Benth., and Ranun- culus lobbii (Hiern) A. Gray, were restricted to ver- nal pools in the study area, but absent from other nearby wetland types. Of these species, Hog pool had the highest number (7) compared to Riley and Coon North pools, which had none. Eleocharis aci- cularis (L.) Roemer & Schultes, E. macrostachya Britton, Eryngium aristulatum var. aristulatum, Gratiola ebracteata, Hordeum marinum Hudson subsp. gussoneanum (Parl.) Thell., and Plagioboth- rys bracteatus had the highest constancies, occur- ring in 8 out of 12 pools. Species richness was highest in a band approximately 1 meter below the high strand line where a mix of introduced annuals and native wetland species occurred. Below this [Vol. 46 Junc- Junc- tion tion Hunt- Hunt- Blue- Blue- West- West- ley-96 ley-99 bird-96 bird-99 96 99 2 pi 2 3 3 3 2 1 3 3 5 2 1 2 1 2 1 1 1 3 3 1 Z 1 2 2 1 2 1 3 2 | 1 3 3 2 3 2 2 1 2 3 1 l 2 3 1 29 2) 24 24 15 15 band and extending toward the pool centers, Er- yngium aristulatum var. aristulatum, Eleocharis macrostachya, and Isoetes howellii were common and usually the dominant species (Table 3). Of 27 exotic species, Briza minor L., Hordeum marinum ssp. gussoneanum, Lolium multiflorum Lam., and Polypogon monspeliensis (L.) Desf. were the most abundant, commonly encroaching into the pools from the outer margin. Exotic species were rare to- ward the pool centers. There were no significant changes observed in species composition and abundance between the two sampling periods of 1995-1996 and 1998—- 1999 (Table 3). Although the two winters fell close to the 35 year rainfall average of 947 mm, a wet late spring and early summer in 1996 lengthened the period of inundation by several weeks over those in 1999. Surprisingly, the pool phase was shorter after the El Nifio winter of 1997-1998, which produced 1546 mm of rainfall (Table 1). Re- sults from cover estimates taken in September of 1997 are shown in Table 4 and reflect the summer pool vegetation. The variables that seem to influence differences eT io “at —— = — 1999] HEISE AND MERENLENDER: MAYACMAS VERNAL POOLS 43 TABLE 3. EXTENDED CONTINUED Junc- Junc- tion tion Coon Coon Coon Coon Rifle Rifle Twin- Twin- East- East- Riley- Riley- North- North- South-South- WS- WS- WL- WL- Hog- Hog- Range-Range- ing- ing- 99 99 96 99 96 99 96 99 96 99 96 99 96 99 96 99 96 99 1 2 1 | ] 3 1 1 2 1 1 1 ] 4 4 1 Zz 4 2 2 2 l 3 3 3 3 4 2 2 4 4 | 2 3 3 1 2 l | ps | l 1 1 4 4 3 2 l 2 2 1 2 2 2 | 2 2 2 2 1 | 3 3 2 1 2 3 5 3 4 2 ps 3 5 4 3 2 Z 2 2 2 ps 1 4 1 1 1 | 1 1 2 1 1 2 1 1 1 z 1 3 1 1 ps 2 l 2 4 4 3 ps zZ 2 2 3 2 l 1 2 1 | l 1 Z 2 l 2 1 Z 1 Zz 2 l 2 3 3 l 2 2 l 19 24 29 28 11 12 ZY 30 22 24 15 20 43 40 15 18 35 32 between pools in this study are pool depth and pro- file, length of inundation phase, degree of shade, and management. Deep pools such as Riley, Coon South, Coon North, and Hog experience longer pe- riods of inundation and often support taxa more typical of perennial wetlands such as Juncus, Eleo- charis, and Carex. Riley and Coon South pools are _ also steep-profiled with narrow margins, a topog- raphy that did not support vernal pool specialist plants common on shallow-profiled pools with wide margins. Coon South was the only vernal pool un- derlain with a dense mat of the aquatic moss Lep- _todictyon riparium. Although Coon North and Riley pools are both deep, densely shaded, steep- _ profile pools, they had little in common floristically. Coon North, which receives heavy sheep use, was essentially devoid of herbaceous vegetation, while Riley, protected from livestock use since 1956, had a dense band of Carex subbracteata Mackenzie, _ Eleocharis macrostachya, and Agrostis exarata i} | it Trin around its upper perimeter. Both pools lack Eryngium aristulatum var. aristulatum and Isoetes howellii which were characteristic of many of the other vernal pools. Hog is the largest (3069 m7) and deepest (115 cm) vernal pool but has a very shal- low profile, thus supporting a diverse mix of pe- rennial wetland and vernal pool specialist taxa. Huntley, Bluebird, Junction West, Junction East, Weather Small, Weather Large, Rifle Range, and Twining pools are relatively shallow with wide margins and little or no tree canopy. Eryngium aris- tulatum var. arisulatum was a dominant species in both Junction West and Junction East, which lie adjacent to each other. Junction East was the deeper of the two with I. howellii as a codominant, where- as Lolium multiflorum codominated in Junction West. Weather Large and Weather Small, paired pools similar in many respects, were both codom- inated by E. aristulatum var. aristulatum and I. howellii. Huntley and Bluebird pools lie within 500 m of each other at similar elevations and are both codominated by E. macrostachya and E. aristula- tum var. aristulatum. Rifle Range pool was origi- nally a shallow profile vernal pool with an area of ca. 1200 m’. Excavation of the fragmented shales that overlay the surface resulted in the creation of several small, deep pools, which are dominated by E. aristulatum var. aristulatum in the basins and 42 MADRONO [Vol. 46 TABLE 3. CONTINUED Species Junc- June- tion tion Hunt- Hunt- Blue- Blue- West- West- ley-96 ley-99 bird-96 bird-99 96 99 Mimulus guttatus DC. Mimulus pilosus (Benth.) S. Watson Montia fontana L. Navarretia intertexta (Benth.) Hook. Phalaris aquatica L. Plagiobothrys bracteatus (T. J. Howell) I. M. Johnston Poa annua L. Poa secunda J. S. Presl Pogogyne zizyphoroides Benth. Polypogon interruptus Kunth Polypogon monspeliensis (L.) Desf. Potamogeton foliosus Raf. Psilocarphus tenellus Nutt. Ranunculus aquatilus L. Ranunculus lobbii (Hiern) A. Gray Rumex crispus L. Silene gallica L. Sisyrinchium bellum S. Watson Spiranthes pornifolia Lindley Spirodela polyrrhiza (L.) Schleiden Stellaria media (L.) Villars Taeniatherum caput-medusae (L.) Nevski Trichostema laxum A. Gray Trifolium dubium Sibth. Trifolium variegatum Nutt. Trisetum canescens Buckley Triteleia hyacinthina (Lindley) E. Greene Veronica peregrina L. subsp. xalapensis (Kunth) Pennell Vulpia bromoides (L.) S. F. Gray Vulpia myuros (L.) C. Gmelin Zigandenus micranthus Eastw. Total number of species (70%) were native species (Appendix 1). Species richness ranged from 12 at Coon North pool to 45 at Hog pool (Table 2). Centunculus minimus L., Crassula aquatica (L.) Schénl, Deschampsia dan- thonioides (Trin.) Munro, Downingia cuspidata (E. Greene) E. Greene, Elatine californica A. Gray, Eryngium aristulatum Jepson var. aristulatum, Gra- tiola ebracteata Benth., Isoetes howellii Engelm., Plagiobothrys bracteatus (T. J. Howell) I. M. John- ston, Pogogyne zizyphoroides Benth., and Ranun- culus lobbii (Hiern) A. Gray, were restricted to ver- nal pools in the study area, but absent from other nearby wetland types. Of these species, Hog pool had the highest number (7) compared to Riley and Coon North pools, which had none. Eleocharis aci- cularis (L.) Roemer & Schultes, E. macrostachya Britton, Eryngium aristulatum var. aristulatum, Gratiola ebracteata, Hordeum marinum Hudson subsp. gussoneanum (Parl.) Thell., and Plagioboth- rys bracteatus had the highest constancies, occur- ring in 8 out of 12 pools. Species richness was highest in a band approximately 1 meter below the high strand line where a mix of introduced annuals and native wetland species occurred. Below this 2 2 2 3 w w i) Nw rs) is) Y= is) i) i) i) is) Nn is) 1999] HEISE AND MERENLENDER: MAYACMAS VERNAL POOLS 43 TABLE 3. EXTENDED CONTINUED Junc- Junc- tion tion Coon Coon Coon Coon Rifle Rifle Twin- Twin- Fast- East- Riley- Riley- North- North- South-South- WS- 9 #99 % %9 9% 99 996 99 96 WS- WL- WL- Hog- Hog- Range-Range- ing- ing- 99 96 99 96 99 96 99 96 99 1 2 1 1 1 is) 4 4 1 BO jim vie 2 1 1 1 l ha 4 i 2 2 1 2 aA 9) Pe 2 3 5 3 5 4 3 2 1 4 1 1 2 1 l 2 1 3 1 1 3 1 2 A @ ¥ 2 1 1 Dh es ee 2 QR PD 2 1 2 eee 290 28) 1 12) 27) 30) 22 l 1 3 ae fh 3 2 2 1 3 fk A l 2 3 3 1 2 2 1 l 3 2 1 2 2 2 2 I 3 Al 2 2 2 2 2 2 2 1 1 1 1 1 1 2 1 2 2 2 2 2 Ne 1 1 2 1 2 1 2 1 2 ch A's} i) band and extending toward the pool centers, Er- yngium aristulatum var. aristulatum, Eleocharis macrostachya, and Isoetes howellii were common and usually the dominant species (Table 3). Of 27 exotic species, Briza minor L., Hordeum marinum ssp. gussoneanum, Lolium multiflorum Lam., and Polypogon monspeliensis (L.) Desf. were the most abundant, commonly encroaching into the pools from the outer margin. Exotic species were rare to- ward the pool centers. There were no significant changes observed in species composition and abundance between the two sampling periods of 1995-1996 and 1998- 1999 (Table 3). Although the two winters fell close to the 35 year rainfall average of 947 mm, a wet late spring and early summer in 1996 lengthened the period of inundation by several weeks over those in 1999. Surprisingly, the pool phase was shorter after the El Nifio winter of 1997-1998, which produced 1546 mm of rainfall (Table 1). Re- sults from cover estimates taken in September of 1997 are shown in Table 4 and reflect the summet pool vegetation. The variables that seem to influence differences between pools in this study are pool depth and pro- file, length of inundation phase, degree of shade, and management. Deep pools such as Riley, Coon South, Coon North, and Hog experience longer pe- tiods of inundation and often support taxa more typical of perennial wetlands such as Juncus, Eleo- charis, and Carex. Riley and Coon South pools are also steep-profiled with narrow margins, a topog- taphy that did not support vernal pool specialist plants common on shallow-profiled pools with wide margins. Coon South was the only vernal pool un- derlain with a dense mat of the aquatic moss Lep- todictyon riparium. Although Coon North and Riley pools are both deep, densely shaded, steep- Profile pools, they had little in common floristically. Coon North, which receives heavy sheep use, was essentially devoid of herbaceous vegetation, while Riley, Protected from livestock use since 1956, had 4 dense band of Carex subbracteata Mackenzie, Eleocharis macrostachya, and Agrostis exarata around its upper perimeter. Both pools lack Eryngium aristulatum var. aristulatum and Isoetes owellii which were characteristic of many of the other vernal pools. Hog is the largest (3069 m?) and deepest (115 cm) vernal pool but has a very shal- low profile, thus supporting a diverse mix of pe- rennial wetland and vernal pool specialist taxa. Huntley, Bluebird, Junction West, Junction East, Weather Small, Weather Large, Rifle Range, and Twining pools are relatively shallow with wide margins and little or no tree canopy. Eryngium aris- tulatum var. arisulatum was a dominant species in both Junction West and Junction East, which lie adjacent to each other. Junction East was the deeper of the two with /. howellii as a codominant, where- as Lolium multiflorum codominated in Junction West. Weather Large and Weather Small, paired pools similar in many respects, were both codom- inated by E&. aristulatum var. aristulatum and I. howellii. Huntley and Bluebird pools lie within 500 m of each other at similar elevations and are both codominated by E. macrostachya and E. aristula- tum var. aristulatum. Rifle Range pool was origi- nally a shallow profile vernal pool with an area of ca, 1200 m*. Excavation of the fragmented shales that overlay the surface resulted in the creation of several small, deep pools, which are dominated by E. aristulatum var. aristulatum in the basins and & & TABLE 4. PERCENT COVER OF DOMINANT SPECIES BELOW LOWER EDGE OF OUTSIDE VEGETATION BAND, HOPLAND RESEARCH AND EXTENSION CENTER AND ADJACENT PRIVATE LAND, MENDOCINO CounTy, CA. Data collected on 23 Sept. 1997. The no grazing status of Junction West, Junction East, and Hog pools are for observations made inside 16 m? grazing exclosures placed inside pools. Grazing Rifle Range Weather Weather Coon Coon Junction Junction Junction Junction Huntley Bluebird West Twining East Riley North South Small Large Hog Hog 80 East West Species Leptodictyon riparium Aristida oligantha 25 2D Cynodon dactylon Deschampsia danthonioides Eleocharis acicularis 25 15 5 10 20 Eleocharis macrostachya Eremocarpus setigerus 70 65 50 10 45 45 70 Eryngium aristulatum Hordeum marinum Tsoetes howellii 25 Lolium multiflorum 10 50 Lythrum hyssopifolium Mentha pulegium MADRONO 10 45 10 45 Trichostema laxum Barren (Litter/Soil) 10 40 60 25 30 25 100 80 85 25 90 [Vol. 46 Deschampsia danthonioides around the rims. The Twining pool is located ca. 5 km to the northwest of HREC on private land; cattle were fenced out in 1996. Dominants include [. howellii, E. macro- stachya, and Mentha pulegium L. Narrow endemic or rare plants, so characteristic of geologically older vernal pool complexes in oth- er parts of the state (Thorne 1984), are absent from the pools in the study area. Instead, many of the species restricted to vernal pool habitat here and nearby vernal pools (de Nevers 1985; Baad 1998) are widely distributed outside California, except Eryngium aristulatum var. aristulatum, which is found only in the San Francisco Bay Area and the North Coast Ranges of California. These cismontane pools occur in a relatively wet part of the state, resulting in prolonged periods of inundation and the establishment of perennial wet- land taxa. Keeley and Zedler (1998) note that such a habitat might be more aptly described as a vernal marsh rather than a vernal pool. In addition, their relatively recent origin and limited lifespan differ- entiate them from pools of older geomorphological origin such as those associated with southern Cal- ifornia coastal terraces, Central Valley alluvial ter- races, or more widespread lava flow landforms (Holland 1978; Keeler-Wolf et al. 1995). ACKNOWLEDGMENTS We are grateful to Chuck Williams for his knowledge of the vernal pools in the region, Elizabeth Nance for her assistance in the field, Colin Brooks for GIS and GPS data processing, Bob Keiffer, Chuck Vaughn, Greg de Nevers, and two anonymous reviewers for their comments on the manuscript, and Greg Giusti for his enthusiasm and support. LITERATURE CITED BAAD, M. FE 1998. Biological survey of Lost Valley, South Cow Mountain Recreation area, Bureau of Land Management, Department of the Interior. Unpub- lished report. B950-RFP 7-0001. BAUuDER, E. T. AND S. MCMILLAN. 1998. Current distri- bution and historical extent of vernal pools in south- ern California and Baja California, Mexico. Pp. 56— 70 in C. W. Witham, E. Bauder, D. Belk, W. Ferren, and R. Ornduff (eds.), Ecology, conservation, and management of vernal pool ecosystems—proceedings from a 1996 conference. California Native Plant So- ciety, Sacramento, CA. DE NEverRS, G. 1985. Pepperwood flora. Pepperwood Ranch Natural Preserve. Sonoma County, CA. Cali- fornia Academy of Sciences, San Francisco, CA. Gowans, K. D. 1958. Soil Survey of the Hopland Field Station. California Agricultural Experiment Station, Hopland, CA. HICKMAN, J. C. (ed.). 1993. The Jepson manual: higher plants of California. University of California Press, Berkeley, CA. HOLLAND, R. F 1978. The geographic and edaphic distri- bution of vernal pools in the great Central Valley, California. California Native Piant Society, Special Publication #4, Sacramento, CA. . 1998. Great Valley Vernal Pool Distribution, Pho- 1999] torevised 1996. Pp. 71-75 in C. W. Witham, E. Bau- der, D. Belk, W. Ferren, and R. Ornduff (eds.), Ecol- ogy, conservation, and management of vernal pool ecosystems—proceedings from a 1996 conference. California Native Plant Society, Sacramento, CA. AND S. JAIN. 1987. Vernal Pools. Pp. 515-533 in M. Barbour and J. Major (eds.), Terrestrial vegetation of California. California Native Plant Society, Special Publication No. 9. JAIN, S. (ed.). 1976. Vernal pools: their ecology and con- servation. Institute of Ecology, University of Califor- nia Special Publication 9. KEELER-WOLF, T., D. R. ELAM, AND S. A. FLINT. 1995. California vernal pool assessment preliminary report. State of California, The Resources Agency, Depart- ment of Fish and Game, Sacramento, CA. KEELEY, J. E. AND P. H. ZEDLER. 1998. Characterization and global distribution of vernal pools. Pp. 1-14 in C. W. Witham, E. Bauder, D. Belk, W. Ferren, and R. Ornduff (eds.), Ecology, conservation, and manage- ment of vernal pool ecosystems—proceedings from a 1996 conference. California Native Plant Society, Sacramento, CA. KENT, M. AND P. COKER. 1992. Vegetation description and HEISE AND MERENLENDER: MAYACMAS VERNAL POOLS 45 analysis, a practical approach. John Wiley and Sons, New York. KING, J. L., M. A. SIMOvICH, AND R. C. BRusca. 1996. Species richness, endemism and ecology of crusta- cean assemblages in northern California vernal pools. Hydrobiologia 328(2):85-116. Murpuy, A. H. AND H. EF HEADY. 1983. Vascular plants of the Hopland Field Station, Mendocino County, California. The Wasmann Journal of Biology 41(1- 2):53—96. NEILSON, J. A. AND D. McQuAID. 1981. Flora of the Ma- yacmas Mountains. California Energy Commission. Unpublished report. P700-81-16. SMITH G. L. AND C. R. WHEELER. 1990-1991. A flora of the vascular plants of Mendocino County, California. University of San Francisco, San Francisco, CA. THORNE, R. F 1984. Are California’s vernal pools unique? In S. Jain and P. Moyle (eds.), Vernal pools and in- termittent streams. U.C. Davis Institute of Ecology Publication 28:1-8. ZEDLER, P. H. 1987. The ecology of southern California vernal pools: A community profile. Biological Report 85, National Wetlands Research Center, U. S. Fish and Wildlife Service. MADRONO, Vol. 46, No. 1, pp. 46—48, 1999 NOTES IMPACT OF A NON-NATIVE PLANT ON SEED DISPERSAL OF A NATIVE.—Salsola tragus L. (Mosyakin 1996), or tumbleweed, was introduced to the United States from Eurasia in the late 1800’s (Young 1991). It has subsequently invaded large portions of the arid western United States and Can- ada. The impact that Salsola spp. have upon neigh- boring native plant survival during succession has been the subject of a number of studies (e.g., Lodhi 1979; Allen and Allen 1988; Johnson 1998). How- ever, the impact of the invader upon the seed dis- persal of natives has not been quantified. Plants such as Salsola spp. that have a shrubby growth form act as barriers that slow wind currents and trap wind-dispersed seeds (Day and Wright 1989). Sal- sola tragus often reaches densities higher than na- tive species on disturbed sites. As a result, S. tragus may impact the seed dispersal of other species to a greater extent than natives with a similar architec- ture. Thus, the presence of the non-native S. tragus has the potential to alter the ability of natives to colonize successfully by altering seed dispersion patterns and plant survival. Castle Mountain Mine near Searchlight, NV, contains a large (ca. 50 ha) overburden area where unusable excavated material from deep within the mountain is piled. The surface consists of heavily compacted rock fragments with nutrient-poor (total Kjeldahl nitrogen <1.5 mg/g), alkaline (pH = 8.0) soils that have little organic matter (<1.59%) and clay (<1.0%) (Walker unpublished data). The most common colonizers include the non-native annual S. tragus and the more diffusely constructed native annual Eriogonum deflexum Torrey (Hickman 1993), or flat-topped buckwheat. Although both species have wind-dispersed seeds, the two plants differ in their specific mechanisms of seed dispers- al. Salsola tragus disperses its seeds after the main stem abscises at the base. The wind subsequently pushes the elliptical plants, jarring seeds loose from leaf axils as the plants bounce on the ground over a considerable distance. Stallings et al. (1995) found that Salsola tragus seeds were evenly dis- tributed across the landscape if they were dropped during the tumbling phase. Seeds were concentrat- TABLE 1. ed in areas below the plant before the stem abscised and where the plant came to rest for a time (i.e., obstructions). By contrast, E. deflexum have winged seeds carried from the maternal plant by wind cur- rents. The dispersal of these seeds depends upon the speed and direction of the wind, which may be altered by plants in the area. Salsola tragus in southwestern Wyoming had wind speeds of 13 m/ sec one meter above the plant, 11 m/sec 15 cm above the plant, 9 m/sec 10 cm in front (..e., up- wind) of the plant, and 2.5 m/sec 10 cm behind (i.e., downwind from) the plant (Allen & Allen 1988). We hypothesized that S. tragus plants pro- vide a barrier to E. deflexum seed dispersal through slowed wind currents. Eriogonum deflexum plants may in turn affect S. tragus seed dispersal by fur- nishing a physical obstruction to movement during the tumbling phase. On the overburden pile at Castle Mountain Mine, we selected a flat, 600 m?’ plot dominated by S. tragus and, to a lesser extent, E. deflexum. We hap- hazardly chose sixty plants representative of the en- tire 600 m? plot: fifteen each of live S. tragus, dead (but attached to the ground by the main stem) S. tragus, live E. deflexum, and dead E. deflexum. Most plants chosen were at least 50 cm from their | nearest neighbor (dead E. deflexum were occasion- ally as close as 20 cm to their nearest neighbor). In September 1996 we collected seed and plant litter within a 10 cm X 15 cm quadrat under each plant and from the nearest plant-free clearing within 25 cm to 100 cm east of each focal plant’s eastern | canopy edge. We used the difference between each canopy sample and its adjacent open area for anal- | ysis. This approach adjusted for differences in spe- | cies substrate preferences on a microsite scale (e.g., living E. deflexum were more common in seed- catching rocky areas while living S. tragus were | more common on sandy substrates). Litter was sep- arated from the inorganic soil by sieving each sam- ple and floating the organic matter. Each litter sam- ple was dried and weighed. Due to the time-con- suming nature of seed counting, five random sam- ples from each treatment were selected for seed counts. All seeds were identified for each sample, DESCRIPTIVE STATISTICS FOR ALL STATUS AND SPECIES COMBINATIONS ADJUSTED FOR MICROSITE (UNDER CAN- OPY—ADJACENT OPEN AREA) [MEAN (SE)]. Seed densities expressed as number of seeds per cm’. Litter mass expressed as grams dry weight per cm’. Dead E. deflexum Seed densities E. deflexum S. tragus 4.30 0.09 Ie3D: + (1.20) 0.23, ~(0:19) Litter mass 0.0010 (0.0005) Live E. deflexum (1:93) (0.09) 0.0139 (0.0024) Dead S. tragus Live S. tragus 9.62 (4:31) 4.05 (2.78) 0.6720 (0.2390) 232. (124) 3.85 ~Gl30) 0.0315 (0.0053) 1999] NOTES 47 TABLE 2. SUMMARY FROM Two-Way ANOVA wITH STA- TUS OF PLANT (DEAD OR ALIVE) AND SPECIES (E£. DEFLEXUM or S. TRAGUS). Seed densities and litter adjusted for mi- crosite (under plant—adjoining open area) and natural log T transformed prior to analysis. F ratios are reported, with —e— Dead Plants asterisks denoting significance level (* = P < 0.050; —s# - Living Plants ** = P < 0.010). = E. iD deflexum S. tragus Lv 5 seed seed a Ne Source of variation df densit densit df Litter rs y Mf 53 Status 1 001 0.14 1 16.19%* See Species 1 0.00 LOQ7** > 1 edOO5e* ge Status X Species 1 0.40 2.61 1 18.81** WW Error 16 56 Total 19 59 Eriogonum deflexum Salsola tragus Plant Species but seeds other than S. tragus and E. deflexum were very rare and not included in the analysis. The data were analyzed with a two-way ANOVA (Minitab 1991) which included main effects of species and status (dead or alive) and the interaction between them. —e— Dead Plants — - Living Plants RESULTS Samples from open areas had less litter (mean + SE; 0.003 = 0.001 vs. 0.032 + 0.007 g. dry wt./ cm’), fewer E. deflexum seeds (1.30 + 0.54 vs. 5.69 + 1.39 seeds/cm’), and fewer S. tragus seeds (0.19 + 0.08 vs. 2.25 + 0.84 seeds/cm?’) than areas under plant canopies, as expected. After adjusting the un- der plant sample for microtopography using the open sample, the presence of S. tragus plants re- sulted in E. deflexum seed densities statistically equal to those under living E. deflexum plants, re- gardless of plant status (Tables 1 and 2; Fig. 1A). Eriogonum deflexum Salsola tragus These results suggest that S. tragus plants served Plant Species as an effective trap for E. deflexum seeds. However, S. tragus seed density was more than 16 fold great- er under S. tragus than under E. deflexum (Tables 1 and 2; Fig. 1B), indicating that the native E. de- flexum does not provide a major barrier to tumbling S. tragus. Plant status, species, and the interaction —e— Dead Plants between them were all significant for litter data (Ta- — =~ Living Plants bles 1 and 2; Fig. 1C). Thus S. tragus plants were also more effective than E. deflexum in catching and/or retaining litter under the canopy. The data suggest that E. deflexum seeds under E. deflexum fall from the plant and are blown away, while E. deflexum seeds under S. tragus have been captured by the slowed wind currents (Fig. 1A) and ~~ aN > Do L™ ~~ 3 2 BD Oo ® DO Oo YH Litter Mass (g. dry wt. / cm 2) <_ Fic. 1. Interactions between plant status and species for a) Eriogonum deflexum seeds, b) Salsola tragus seeds, and c) litter [mean + SE]. Means adjusted for microsite by subtracting adjacent open sample from canopy sample. Numbers shown in Table 1. Eriogonum deflexum Salsola tragus Plant Species 48 MADRONO perhaps held by litter under dead S. tragus (Fig. 1C). Figure 1B indicates S. tragus seeds under S. tragus plants probably came from the plant above them, since E. deflexum had very few S. tragus seeds under their canopies (Table 1). We conclude that S. tragus has a significant effect upon the dis- persion of E. deflexum seeds, but E. deflexum plants do not affect the dispersion of S. tragus seeds. However, the time period for which this pattern is obtainable may be strongly influenced by tumble- weed abscission later in the season, a possibility not addressed in this data set. Further work is needed to identify the overall impact that the seed-catching function of the alien S. tragus has upon native seedling establishment and plant success. We showed that a native plant’s seed distribution can be affected by Salsola tragus. Other work has indicated that Salsola spp. can pos- itively or negatively interact with nearby natives throughout the plants’ lives. Allen and Allen (1988) propose Salsola kali may facilitate native grass seed colonization and establishment but may later com- [Vol. 46 pete with adults for water and nutrients. Salsola kali has an extensive root system and efficient uptake of phosphorus (Itoh and Barber 1983). In addition, S. kali can significantly alter soil nutrient concen- trations in mixed culture, perhaps due to its rapid growth rate (Allen 1982). Salsola tragus can even make phosphorus more available to neighboring plants through a high oxalate concentration in can- opy leachates (Cannon et al. 1995; Hageman et al. 1988). Clearly, Salsola spp. affect neighboring na- tives, but the net effect of this interference is un- known. This study has shown that Salsola spp. can concentrate native plant seeds under the Salsola spp. canopies, where germination and growth may then be either facilitated or inhibited. Further un- derstanding of the effects of Salsola spp. on native colonizers will enhance efforts to reintroduce native species to damaged ecosystems. —CHERYL H. VANIER AND LAWRENCE R. WALKER, De- partment of Biological Sciences, University of Nevada, Las Vegas, NV 89154. MADRONO, Vol. 46, No. 1, pp. 49-50, 1999 NASCENT INFLORESCENCES IN ARCTOSTAPHYLOS PRINGLEI: RESPONSE TO KEELEY PHILIP V. WELLS Department of Botany, University of Kansas, Lawrence, KS 66045 In an account of the absence of nascent inflores- cences in Arctostaphylos pringlei, C. Parry Keeley (1997) pays lip service to the diversity of these de- velopmental structures in the genus but illustrates only the expanded paniculate type of his own spe- cies, ““A. rainbowensis’’. Unfortunately, he seems unaware that there is wide variation in stages of development of the nascent inflorescence in differ- ent species of Arctostaphylos. Whatever the failing may be, I cut to the point by illustrating the nascent inflorescence on one of my collections of A. prin- glei (Fig. 1), dated November 6, 1986, from the San Bernardino Mountains, CA (typical of five or more on aS many specimens). Fic. 1. Life-size scan of branchlet of Arctostaphylos pringlei, as it was collected on November 6, 1986. The bracteose nascent inflorescence is at the upper left, where it terminates a 1986 shoot of the year. Note the distal position above the mature leaves of the year. 50 MADRONO All species of Arctostaphylos have the inflores- cences terminal at the ends of branchlets. Terminal meristems shift from a vegetative mode producing leaves to a flowering mode producing a dormant, embryonic (nascent) inflorescence at the tip of the branchlet. The nascent inflorescence terminates growth on that axis; it forms as the new leaves ma- ture below it on the same shoot. Both nascents and new leaves are produced on shoots of the current year, following completion of flowering on separate shoots of the preceding year. Thus, in Figure 1, the new (1986) shoots have fully mature leaves, but the bracteose, distal nascent inflorescence is immature and dormant on November 6, 1986; the separate shoots bearing ripe fruits (the 1985 shoots) were collected but are not shown. The period of dormancy of the nascent inflores- [Vol. 46 cence varies from 5—10 months; in most species this prevents flowering in the summer and fall, when the drought of the Mediterranean isoclime is most intense. Most of the coastal manzanitas bloom in the dead of winter, peaking in January. Among the later species is A. glandulosa, which often flow- ers in March (the tetraploid crown-sprouters are late-bloomers). Latest of all to bloom are the mon- tane species, notably A. pringlei, which flowers in April and May. Since the formation of nascent in- florescences on the new shoots may be much later in A. pringlei than in most other species, they may well be overlooked. LITERATURE CITED KEELEY, J. E. 1997. Absence of nascent inflorescences in Arctostaphylos pringlei. Madrono 44:109-111. MapbRONO, Vol. 46, No. 1, pp. 51-54, 1999 NASCENT INFLORESCENCES IN ARCTOSTAPHYLOS PRINGLEI: RESPONSE TO KEELEY AND WELLS MICHAEL C. VASEY AND V. THOMAS PARKER Department of Biology, San Francisco State University Character states involving nascent inflorescences in Arctostaphylos (Ericaceae) are of great taxonom- ic value. Accordingly, as Keeley (1997) observes, the general absence of a nascent inflorescence in the genus is worthy of notice. Arctostaphylos prin- glei C. Parry is a species of arid montane environ- ments in southern California, Arizona, and northern Baja California that consists of two subspecies, subsp. pringlei and subsp. drupacea (C. Parry) P. Wells that differ in the fusion of nutlets in the fruit. The former occurs in Arizona, the latter in southern Califonrnia, and both subspecies have been found in northern Baja. Like other montane species of Arctostaphylos, flowering occurs between mid- _ spring through early summer. Typically, Arctosta- phylos spp. develop a dormant (nascent) inflores- cence at the tips of their new stem growth during the time fruits mature and disperse in late spring and summer. Nascents can be observed from the end of stem growth until flowering the next year. In contrast, Keeley (1997) has observed that A. pringlei does not produce nascents after stem growth, but produces inflorescences as flowering begins. Hence, the controversy raised by Wells (Wells, 1999) concerns the developmental timing of the formation of an inflorescence with floral buds, 1.00 0.90 0.80 rather than an all or nothing type of character state, as would be implied by “‘nascents present versus absent.”’ Wells implies that Keeley’s (1997) general ob- servation is incorrect. He cites five specimens that he collected in November 1986 to demonstrate that A. pringlei does indeed produce nascent inflores- cences. One of us (MCV) has observed A. pringlei in the field in northern Baja California and in Ari- zona. Jon Keeley had mentioned the lack of inflo- rescences in A. pringlei before a trip to the Sierra San Pedro Martir Mountains in the fall of 1995, which made it a character of interest. On November 25, 1995, one individual (and only one) was found with a nascent inflorescence; other shrubs in the area appeared to lack this structure. Given this con- troversy, we decided to distinguish between the separate interpretations of Keeley and Wells by posing a pair of simple alternative hypotheses: 1) the development of nascents should occur just be- fore and during flowering (flowering phase); versus 2) the development of nascents should occur during and following fruiting (fruiting phase). Confirming hypothesis | and rejecting hypothesis 2 would sup- port Keeley (1997), while confirming hypothesis 2 and rejecting 1 would support Wells (this issue). 0.70 0.60 El A pungens GA. canescens A. pringlei 0.50 0.40 0.30 0.20 Flowering | Fic. 1. Fruiting Percentage of specimens examined that contained presumed nascent inflorescences for A. pungens, A. canes- cens, and A. pringlei. Data are presented for two phenological stages, if the specimen was in flower, and if the specimen was maturing fruit. 52 MADRONO [Vol. 46 Oo NM FF OD WO CE oOo +r MO Oo KR O DD CO ~m Fic. 2. Sequence of flowering, fruiting and nascent inflorescence production found in specimens housed at California Academy of Sciences. A) sequence for A. canescens; B) sequence for A. pungens. TABLE 1. CHI-SQUARE ANALYSES OF PRESENCE OR ABSENCE OF NASCENT INFLORESCENCES AGAINST FLOWERING OR FRUIT- ING AMONG COMBINATIONS OF A. PRINGLEI, A. CANESCENS, AND A. PUNGENS. Values in the right column include both Chi- square values and significance levels. Nascents Flowering Fruiting Total Chi-Square A. pringlei with 11 4 LS without 42 44 86 2.169 NS A. pungens with 5 29 34 without 12 3 15 22.502 p<<0.001 A. canescens with 6 30 36 without fal 3 14 20.083 p<<0.001 A. pungens+ with 1] 59 70 A. canescens without 23 6 29 39.653 p <<0.001 Species W/nascents W/o nascents Chi-Square Flowering A. pringlei 11 42 53 A. pungens+ 11 23 34 22152 A. canescens 22 65 87 NS Fruiting A. pringlei 4 44 48 A. pungens+ 59 6 65 79.438 A. canescens 63 50 £13 p<<0.001 i) 1999] VASEY AND PARKER: RESPONSE TO KEELEY AND WELLS 53 poe eae i | —@—pdFlower ~~ Od Fruit '—~f&— pdNascent Fic. 3. —@ ppFlower = ppFruit a 3 ppNascent Sequence of flowering, fruiting and nascent inflorescence production found in specimens housed at California Academy of Sciences. A) sequence for A. pringlei subsp. drupacea; B) sequence for A. pringlei subsp. pringlei. METHODS Initially, 101 sheets of specimens from the Cal- ifornia Academy of Sciences (CAS) of both sub- species of A. pringlei collected throughout its range were examined for phenological stage and presence or absence of apparent nascent inflorescences (note: shortly prior to flowering, immature inflorescences appear similar to ‘“‘nascents’’). Collection numbers for each sheet were recorded as well as dates of collection, county, phenological stage (early flow- ering, flowering, early fruiting, fruiting, and past fruiting), and presence or absence of nascent inflo- rescences. For comparative purposes, two other mid-montane species, 49 sheets of A. pungens and 50 sheets of A. canescens, were examined in a sim- ilar way (a total of 99 sheets for the two species combined). Chi-square 2 X 2 contingency analysis was used to test for differences in nascents between flowering and fruiting stages (Zar 1984). RESULTS Arctostaphylos pungens Kunth and A. canescens Eastw. are typical of other species in the genus. The development of nascent inflorescences occurs at the time fruits are maturing on the tips of newly elon- gating stems (Fig. 1, 2). Approximately 90% of the specimens in the fruiting phase are developing nas- cent inflorescences. Note that this pattern holds true for both of these species and there is no significant difference between them (Table 1). Arctostaphylos pringlei is not substantially different from the other two species during the flowering phase (Fig. 1, 3), however, during the fruiting phase, less than 10% of the A. pringlei specimens possessed nascent in- florescence structures (Fig. 1, 3). This pattern held true equally for both subspecies suggesting that this unusual developmental character is a shared feature in the A. pringlei lineage. The stark contrast be- tween A. pringlei and the other species during the fruiting phase (Fig. 1, Table 1) are consistent with our hypothesis 1, which supports the conclusions of Keeley (1997). Granted, four out of 48 specimens of A. pringlei were found to present nascent inflorescences ap- parently established during the fruiting phase. We underscore ‘‘apparently’’ given the possibility that these individuals represent shrubs that may be flow- 54 MADRONO ering out of season. One CAS specimen (#563576), for example, was flowering in September, clearly an ‘“‘exception from the rule”’ that may be associ- ated with phenological opportunism on the part of this individual; such opportunism is common in the genus. In contrast, 59 out of 65 sheets of A. pun- gens and A. canescens had nascents during the fruiting phase (Fig. 1, 2). This difference in the pat- tern of nascent inflorescence establishment between A. pringlei and other species of Arctostaphylos dur- ing the fruiting phase is significant. These findings are very consistent with the observations of Keeley (1997). DISCUSSION Given its near universal occurrence in Arctosta- phylos, the nascent inflorescence developmental character is logically ancestral in this genus. In that case, the general lack of nascent inflorescences in A. pringlei during the fruiting phase is likely a de- rived condition. With this feature combined with its unusual pink deciduous bract characters, characters that Wells (1992) interpreted as warranting subsec- tional status for this species, A. pringlei appears to be a distinctive lineage within the genus. Keeley (1997) introduces an important observation con- cerning the general lack of nascent inflorescences [Vol. 46 in A. pringlei. Our analysis of specimens from CAS confirms his observations with nascents rarely oc- curing during the fruiting phase in A. pringlei in decided contrast to other species in the genus in which a large majority of individuals establish nas- cents during this phenological stage (Fig. 1, 2, 3, Table 1). Having observed numerous populations of Arc- tostaphylos in the field, we have come to the con- clusion that few single characters are completely consistent in this genus. Instead, within Arctosta- phylos, consistency is revealed in a suite of char- acters that distinguish species reliably. That rare ex- ceptions occur to the “lack of nascent inflores- cence” status of A. pringlei is hardly surprising. Taking the position that ‘‘exceptions must make the rule” in this instance and that Keeley’s two years of population observations are somehow uninfor- med or incorrect seems unlikely to advance our un- derstanding of this complex group. LITERATURE CITED KEELEY, J. E. 1997. Absence of nascent inflorescences in Arctostaphylos pringlei. Madrono 44:109-111. WELLS, P. V. 1999. Nascent inflorescences in Arctostaphy- los pringlei. Madrono. 46:49—50. ZAR, J. H. 1984. Biostatistical analysis, 2nd ed. Prentice- Hall, Inc., Englewood Cliffs, NJ. a MaApRONO, Vol. 46, No. 1, pp. 55-57, 1999 DEINANDRA BACIGALUPII (COMPOSITAE—MADINAB), A NEW TARWEED FROM EASTERN ALAMEDA COUNTY, CALIFORNIA BRUCE G. BALDWIN Jepson Herbarium and Department of Integrative Biology, University of California, Berkeley, CA 94720-2465 ABSTRACT Deinandra bacigalupii is a new tarweed known from alkaline meadows in the vicinity of Livermore, Alameda County, California. The taxon has been treated as conspecific with Deinandra [Hemizonia] increscens, but represents a separate lineage that is morphologically, ecologically, and geographically distinct. Unlike other members of Deinandra, D. bacigalupii combines the following morphological char- acteristics: proximal primary-stem leaves mostly entire or irregularly lobed, distal cauline leaves mostly narrowly linear or lance-linear, ray florets mostly 8 per head, ray corolla limbs 2—4 mm long, anthers yellow, and disc pappi of highly irregular, erose scales <1 mm long or reduced to crowns of minute bristles. Results of molecular phylogenetic and morpho- logical studies of Deinandra Greene sensu Baldwin (1999) [=Hemizonia DC. sect. Madiomeris Nutt. sensu Tanowitz (1982) plus ‘‘Fruticosae”’ or “*Zon- amra’’ (see Clausen 1951; Keck 1959)] lead me to conclude that plants from eastern Alameda County, California, treated by Tanowitz (1982) as geograph- ically disjunct members of Hemizonia increscens (H. M. Hall ex D. D. Keck) Tanowitz [= Deinandra increscens (H. M. Hall ex D. D. Keck) B. G. Bald- win] constitute a distinct lineage. Although mor- phologically similar to D. increscens, the Alameda County plants possess mostly entire or irregularly lobed (rather than pinnatifid) proximal primary- stem leaves, yellow (not dark-purple) anthers, and a shorter, more irregular pappus. Ecologically, the Alameda County plants are somewhat unusual in Deinandra for occurring in poorly drained, alkaline habitats more typical of the closely-related genus Centromadia Greene [=Hemizonia DC. sect. Cen- tromadia (Greene) D. D. Keck]. Results of molec- ular phylogenetic analyses of nuclear rDNA spacer sequences place the Alameda County plants closer to D. corymbosa (DC.) B. G. Baldwin than to a lineage comprising representatives of D. increscens subsp. increscens and D. increscens subsp. villosa (Tanowitz) B. G. Baldwin (Baldwin unpublished). The chromosome number shared by D. increscens and the Alameda County plants (2n=12 II), but not shared with D. corymbosa (2n=10 ID), is modal and putatively basal in Deinandra. Based on the fore- going morphological, ecological, and phylogenetic considerations, I propose a new species to accom- odate the distinctive Deinandra populations from eastern Alameda County. Deinandra bacigalupii B. G. Baldwin, sp. nov. (Fig. 1)—TYPE: USA, California, Alameda County, north of Livermore, junction of Ames Street and Raymond Road, in sandy alkaline soil, 31 Aug 1966, R. F. Hoover 9954 (holotype, UC; isotypes, CAS, OBI, UC). A ceteris speciebus Deinandrae differt charac- teribus conjuncte foliis proximalibus plerumque in- tegris vel irregulariter lobatis; foliis caulinis distal- ibus plerumque anguste linearibus vel lanceolatis- linearibus; floribus radiorum (6—)8(—9), limbis cor- ollarum 2—4 mm longis; antheris flavis; squamis papporum irregulariter erosis <1 mm longis vel pappis coroniformibus setis minutis. Annual herbs, strongly odorous. Stems erect, branched in distal half or to near base (the branches ascending-virgate), tawny or whitish (or purplish), shiny near base, to 4 dm high, sparsely to densely hirsute, minutely stipitate-glandular distally, the glands yellowish or clear. Leaves sessile, mostly cauline, evenly distributed, alternate (except in ba- sal rosette), ascending to appressed along stems, usually much longer than internodes; blades of pri- mary-stem leaves narrowly oblanceolate (near base of stem) to linear or lance-linear, =1 dm long, grad- ually reduced distally, mostly entire or irregularly lobed, slightly revolute, sparsely hirsute and mi- nutely stipitate-glandular, the glands yellow or clear; blades of branch-stem leaves linear to lance- linear, <1 cm long on distal branches, slightly rev- olute, uniformly hirsute and _ stipitate-glandular. Capitulescences loosely corymbiform, the side branches overtopping central branches. Peduncles inconspicuous (