roe Wet ct arate ei ua 4A” red in! sess ae nweatt es FA i yaa, muah VOLUME 53, NUMBER | JANUARY-MARCH 2006 MADRONO A WEST AMERICAN JOURNAL OF BOTANY é ya CONTENTS SHOOT MORPHOLOGY IN THE CLAYTONIA SIBIRICA COMPLEX (PORTULACACEAE) RODINFOPOUIATGNG TQTIY TAU) (Ol a eee seen l SYSTEMATICS OF SALVIA PACHYPHYLLA (LAMIACEAE) RODINGVI-LGVIOW ANG: TINGE AV OTS cote. sumone 0 aPibe ith te osannetensaceameruaree 1] THE EVOLUTION OF PINK: MORPHOLOGICAL AND GENETIC VARIATION AMONG THREE LITHOPHRAGMA (SAXIFRAGACEAE) SPECIES ReveCcd AMHUffluand JAMES EE AKICNATASON visaxcaschnccveaseass acnseveneilce ee 25 ADDITIONS TO THE VASCULAR FLORA OF WASHINGTON FROM A BIODIVERSITY STUDY ON THE HANFORD NUCLEAR RESERVATION Kathyrn A, Beck and Florence E.. (Caplow ids assestivctcivieess stant ieetanctacess 36 A RECONSIDERATION OF BRODIAEA MINOR (BENTH.) S. WATSON AND BRODIAEA PURDYI EASTWOOD (THEMIDACEAE), WITH THE RESURRECTION OF BRODIAEA NANA HOOVER Robert E. Preston cnc: Fes ceed, we PSN Esse evncess Sowceeeeees 46 FACTORS AFFECTING UNDERSTORY ESTABLISHMENT IN COASTAL SAGE SCRUB RESTORATION Matt V. Talluto, Katharine Nash Suding, and Peter A. Bowler................ 55 OBSERVATIONS OF FROND GROWTH AND DEVELOPMENT IN PENTAGRAMMA TRIANGULARIS SUBSP. TRIANGULARIS (PTERIDACEAE) OF SOUTHERN CALIFORNIA Gaya PEPIN, «INET io vce ens Ms oes NED eee cceeececeecetens 60 HABITAT, GEOLOGIC, AND SOIL CHARACTERISTICS OF SHASTA SNOW-WREATH (Neviusia cliftonit) POPULATIONS Len Lindstranddihand Julie: Kierstead NeISON 352 oo sion. cceeccceveccceeseceonse 65 ADDITIONAL TAXONOMIC STUDIES OF ARCEUTHOBIUM PENDENS (VISCACEAE): A RARE DwarRF MISTLETOE FROM CENTRAL MEXICO Robert L. Mathiasen and Carolyn M. Daugherty.......ccccccceeeeeeseceeeeeeeeeeees 69 A NEw SPECIES OF SILENE IN THE S/ILENE HOOKERI COMPLEX (CARYOPHYLLACEAE) FROM THE KLAMATH MOUNTAINS OF SHASTA- TRINITY NATIONAL FOREST, TRINITY COUNTY, CALIFORNIA Thomas W. Nelson, Jane P. Nelson, and Susan A. Erwith....cccccccccccceeee cess IZ LEPECHINIA ROSSI (LAMIACEAE), A NARROW ENDEMIC FROM THE WESTERN TRANSVERSE RANGES OF SOUTHERN CALIFORNIA SIEVE BOVG ONG OFIGNdO VISICUC ccs sre eat ttancnte ie see ated ee ra OTEWORTHY ATI ORINUAp ese tee rte cis sensei 86 Sisco eR ae tlc eta ee 85 OLLECTIONS BVVEAIS TIN GTi) NV een ote wwe Gh oa Meanie onan eens es oes cea aecdler e grenee 86 PUBLISHED QUARTERLY BY THE CALIFORNIA BOTANICAL SOCIETY MADRONO (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 mailing offices. Return requested. POSTMASTER: Send address changes to MADRONO, Roy Buck, Jepson Herbarium, University of California, Berkeley, CA 94720-2465. Editor—Dr. JOHN HUNTER c/o UC Davis Herbarium/Center for Plant Diversity Plant Biology One Shields Avenue University of California Davis Davis, CA 95616 johnhunterca@ sbcglobal.net Book Editor—JON E. KEELEY Noteworthy Collections Editors—DIETER WILKEN, MARGRIET WETHERWAX Board of Editors Class of: 2006—SUSAN J. MAZER, University of California, Santa Barbara, CA RICHARD OLMSTEAD, University of Washington, Seattle, WA 2007—SUSAN R. KEPHART, Willamette University, Salem, OR STEVE O'KANE, University of Northern Iowa, Cedar Falls, IA 2008—ELLEN DEAN, University of California, Davis, CA RoBERT E. PRESTON, Jones & Stokes, Sacramento, CA 2009—DONOVAN BAILEY, New Mexico State University, Las Cruces, NM MARK BorCHERT, USFS Ojai, CA CALIFORNIA BOTANICAL SOCIETY, INC. OFFICERS FOR 2005—2006 President: Michael Vasey, Department of Biology, San Francisco State University, 1600 Holloway Avenue, San Francisco, CA 94132, mvasey @sfsu.edu. First Vice President: Stefania Mambelli, Department of Integrative Biology, University of California, Berkeley, CA 94720, mambelli @ socrates.berkeley.edu. Second Vice President: Gretchen LeBuhn, Department of Biology, San Francisco State University, 1600 Holloway Avenue, San Francisco, CA 94132, lebuhn @sfsu.edu. Recording Secretary: Staci Markos, Friends of the Jepson Herbarium, University of California, Berkeley, CA 94720-2465, smarkos @ socrates.berkeley.edu. Corresponding Secretary: Susan Bainbridge, Jepson Herbarium, University of California, Berkeley, CA 94720- 2465, sjbainbridge @calmail.berkeley.edu. Treasurer: Roy Buck, Jepson Herbarium, University of California, Berkeley, CA 94720-2465, fax: 1.510.643.5390, roybuck @email.msn.com. The Council of the California Botanical Society comprises the officers listed above plus the immediate Past President, Bruce Baldwin, Department of Integrative Biology and Jepson Herbarium, University of California, Berkeley, CA 94720, bbaldwin @uclink4.berkeley.edu; the Editor of Madrono; three elected Council Members: James Shevock, National Park Service, Cooperative Ecosystems Studies Unit, 337 Mulford Hall, University of California, Berkeley 94720-3114, jshevock @nature.berkeley.edu; Jeffrey Corbin, Department of Integrative Biology, University of California, Berkeley, CA 94720, corbin @socrates.berkeley.edu; Dean Kelch, Jepson Herbarium, University of California, Berkeley, CA 94720, dkelch@sscl.berkeley.edu; Graduate Student Representatives: Elizabeth Zacharias, Department of Integrative Biology, University of California, Berkeley, CA 94720, ezachar@socrates.berkeley.edu; Vicente Garcia, Department of Integrative Biology, University of California, Berkeley, CA 94720, vicente @ socrates. berkeley.edu; Diana Benner, Department of Biology, San Francisco State University, 1600 Holloway Avenue, San Francisco, CA 94132, dbenner@sfsu.edu. Webmasters: Curtis Clark, Biological Sciences Department, California State Polytechnic University, Pomona, CA 91768-4032, jcclark @csupomona.edu; John La Duke, Department of Biology, University of North Dakota, Grand Forks, ND 58202, john_laduke @und.nodak.edu. @ This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). MADRONO, Vol. 53, No. 1, pp. 1-10, 2006 SHOOT MORPHOLOGY IN THE CLA YTONIA SIBIRICA COMPLEX (PORTULACACEAE) ROBIN O’QUINN'”? AND LARRY HUFFORD Washington State University, School of Biological Sciences Pullman, WA 99164-4236 ABSTRACT The Claytonia sibirica complex, including C. sibirica and C. palustris, exhibits considerable morphological variation that encompasses ecological diversity over a wide geographic range. Shoots are basically rhizomatous in the complex and least specialized in C. sibirica var. sibirica. Claytonia sibirica var. bulbillifera, a serpentine endemic of southern Oregon and northern California, forms succulent, storage scale leaves distal to its foliage leaves each growing season. These scale leaves, which consist primarily of leaf base, are generally lacking in other members of the sibirica complex and give the shoot systems of C. sibirica var. bulbillifera a bulb morphology. Claytonia palustris, like C. sibirica var. sibirica, forms an apically swollen rhizome, but differs in its habit by forming renewal shoots, born in the axils of the basal leaves, at the ends of plagiotropic, single long internodes. Key Words: Bulb, homology, leaf specializations, perennation, shoot architecture, serpentine. The Claytonia sibirica L. complex consists of understory herbs of coastal and mesic inland forests extending from northern Santa Cruz County, California, to coastal northeastern Si- beria (Miller et al. 1984; Chambers 1993; Miller 2003). The C. sibirica complex consists of annuals and perennials that exhibit considerable morpho- logical, ecological and cytological variation over its range. Shoot systems in the C. sibirica complex have been most often described as rhizomatous, although shoot system specializations associated with geographic and ploidy variation have been described. In the Klamath region (KR) of northwestern California and southwestern Oregon, C. sibirica have specialized underground structures involved in perennation (Gray 1877, 1887; Miller et al. 1884). Gray (1877) first described this KR form as C. bulbifera Gray and suggested it resembled C. sibirica but produced densely crowded peren- nating bulbs in a basal rosette. However, Gray’s (1887) revision of the North American Portula- caceae treated C. bulbifera as C. sibirica L. var. bulbillifera Gray and described it as “...only a form of C. sibirica with thickened bases of the radical leaves, which persist on the crown as bulblet-scales.” Miller et al. (1984, p. 266) suggested that the C. sibirica complex consists of three basic morpho- types that differ in “‘shape of basal leaves and the presence or absence of basal bulblets and rhizomes”. Gray’s (1877, 1887) C. sibirica var. bulbillifera represents one morphotype recognized by Miller et al. (1984), which they characterize as ‘Author for correspondence, e-mail: oquinn@pdx. edu *Present address: Portland State University, De- partment of Biology, Portland, OR 97208 bulbiferous, distinguished by elliptical basal leaves, and geographically localized to southern Oregon and northern California where it occurs frequently on serpentine substrates (Fig. 1). We will use this name to refer to the specialized KR morphotype. A second morphotype recognized by Miller et al. (1984) is found in shaded mesic habitats and has the deltoid basal leaf shape of Western C. sibirica var. sibirica GC. sibirica var. bulbillifera [ema] C. palustris 0 500 Miles ——----_— 0) 500 KM Fic. 1. Geographic distribution of the Claytonia sibirica complex. i) 90 r©. perfoliata C. rubra 100 — 0.005 substitutions per site FIG. 2. MADRONO C. palustris C. sibirica var. sibirica [Vol. 53 Perennial Claytonia, plus C. arenicola 1 C. exigua subsp. exigua M C. gypsophiloides C. saxosa C. parviflora subsp. grandiflora C. washingtoniana C. sibirica Complex Maximum likelihood cladogram from combined ITS and trnK/matK data showing phylogenetic relationships of the Claytonia sibirica complex from O’Quinn and Hufford (2005). the type specimen for C. sibirica var. sibirica. We apply the name C. sibirica var. sibirica to populations distributed in the Pacific Northwest along the Cascadian cordillera, the Columbia River Gorge, coast ranges northward from Santa Cruz county, California, to Alaska, and the Aleutian and Commander Islands (Fig. 1). It overlaps with C. sibirica var. bulbillifera in the KR (Fig. 1). Claytonia sibirica var. sibirica also has disjunct populations in the inland Northwest (northern Idaho, western Montana and _ sur- rounding portions of Oregon and British Colum- bia; Fig. 1). We distinguish the morphological variation in C. sibirica var. sibirica as western and eastern morphotypes. Miller et al. (1984) charac- terized C. sibirica var. sibirica as bulbiferous also, but less so than the endemic KR morphotype (=C.sibirica var. bulbillifera sensu Gray 1887). The disjunct eastern populations of C. sibirica var. sibirica are reported to lack swollen leaves. The third morphotype discussed by Miller et al. (1984) was later described as C. palustris by Swanson and Kelley (1987). Claytonia palustris 1s narrowly endemic to two small, mid-elevation regions at the northern and southern ends of the Sierra Nevada and in Siskyou County, Califor- nia, where it overlaps with C. sibirica var. bulbillifera at the eastern edge of its range (Fig. 1). This taxon is unique in the complex in preferring perennially wet, sunny habitats and was described by Swanson and Kelley (1987) as being strongly stoloniferous. O’Quinn and Hufford (2005) found robust support for the monophyly of the Muller et al. (1984) C. sibirica complex based on plastid and nuclear ribosomal DNA sequence data (Fig. 2). Notably, all members of the complex share a unique 10 base pair motif that includes a three- base pair insertion in the internal transcribed spacer region of the nuclear ribosomal DNA. Phylogenetic results recovered a sister taxon relationship between C. palustris and C. sibirica, but lineages within C. sibirica were not resolved. In this study we characterize shoot morphol- ogy of the C. sibirica complex, with a particular emphasis on specializations for nutrient storage and perennation. Beyond Gray’s (1877) initial description of C. sibirica var. bulbillifera, the morphology of the so-called bulbiferous mor- photype of the KR populations has not been studied previously. This comparative study of the shoot systems in the C. sibirica complex addresses especially the morphological identity of struc- tures described as bulbs and bulbiferous and presents hypotheses for the origins of morpho- logical specializations. MATERIALS AND METHODS We sampled specimens of the western mor- photype of C. sibirica var. sibirica from the Williamette Valley, Columbia River Gorge and foothills of the Hood River valley, and of the eastern morphotype from the Lochsa and Clear- water River valleys (Table 1). Claytonia sibirica var. bulbillifera was collected in the Hlinois and Rogue River valleys of southern Oregon where this variety is the most common morphotype (Table 1). Samples of C. palustris were collected at the type locality at Jones Creek in Butte County, California, and seeds for greenhouse grown material were collected from a population at Stubbs Lake, Butte County, California (Table 1). Based on the cytogeographic results of Miller et al. (1984), we assume that our collections of the eastern morphotype of C. sibirica var. sibirica, C. sibirica var. bulbillifera, and C. palustris are diploid. Collections of the western morphotype of C. sibirica var. sibirica are potentially either diploid or tetraploid. Miller et al. (1984) suggested that diploids and polyploids have the same shoot morphologies. Comparisons of shoot system morphology for the four perennial forms of C. sibirica used scanning electron (SEM), and light (LM) micros- copy. Specimens from natural, greenhouse and common garden populations were sampled in May or June and August (Table 1) for fixation in formalin-acetic acid (FAA). Specimens for SEM were dehydrated in a graded ethanol series, critical-point dried, and mounted on aluminum stubs prior to gold coating. We examined 5-8 2006] TABLE |. O’QUINN AND HUFFORD: MORPHOLOGY IN THE CLA YTONIA SIBRIRICA COMPLEX — 3 COLLECTION DATA FOR SAMPLED POPULATIONS OF THE C. S/BIRICA COMPLEX. All vouchers are at WS. Taxon name Population location Collection voucher Collection date Claytonia sibirica var. Lochsa River sites: sibirica Eastern morphotype Glade Creek Eagle Summit Ashpile Creek Columbia River sites: Herman Creek Claytonia sibirica var. sibirica Western morphotype Bridal Veil Falls Hood River site: Pine Mountain Road Willamette River site: Corvallis, OR Southern Oregon sites: Davis Creek Brigg’s Valley Road Cave’s Camp Road Claytonia sibirica var. bulbillifera Eight Dollar Mtn. Road Claytonia palustris Northern California sites: Jonesville (type location) Stubb’s Lake (seed source for greenhouse grown collections) individuals per morphotype for SEM. Specimens were examined at an accelerating voltage of 15— 20 kV. Images were captured digitally using the program Quartz PCI (Quartz Imaging Corp. 1993-1998). Specimens for LM were dehydrated in a graded tertiary-butyl alcohol series (Johansen 1940), infiltrated and embedded in Paraplast™, sectioned at 16 Um, mounted on glass slides, stained with safranin-O and fast green, and examined with a Leitz light microscope. Micro- tomed sections were photographed or drawn using a drawing tube. To characterize shoot architecture and leaf base shape over ontogeny, we made cross and longitudinal sections through the basal rosettes of 3—5S individuals per examined population (Table 1) of western and eastern C. sibirica var, sibirica, C. sibirica var. bulbillifera, and greenhouse grown specimens of C. palustris. Seasonal growth along shoots was identified by discrete regions of leaf scars that differ from each other in length and circumference along a contin- uous shoot axis. RESULTS Shoot Architecture Claytonia sibirica var. sibirica. Perennials form an orthotropic to plagiotropic shoot with short internodes that bear helically arranged leaves, forming a rosette of photosynthetic leaves at the base of the newly elongating axis early in the R. O’Quinn 483 17 May 2002 R. O’Quinn 488 17 May 2002 R. O’Quinn s.n 25 Aug 2002 R. O’Quinn 492 20 May 2002 R. O’Quinn 528 19 March 2003 J. Schenk 774 28 June 2004 R. O’Quinn 490 R. O’Quinn s.n. R. O’Quinn 529 J. Schenk 774 20 May 2002 31 Aug 2002 19 Mar 2003 28 June 2004 J. Schenk 773 27 June 2004 . O’Quinn 494 . O'Quinn 504 . O'Quinn 365 . O’'Quinn 474 . O'Quinn 290 . O’Quinn 508 22 May 2002 23 May 2002 29 May 2001 OS Aug 2001 24 May 2000 24 May 2002 4 June 2000 Multiple collections between 2001 and 2004 . O’'Quinn 330 . Bjork 5704b QF FRRARARA growth season. Inflorescence branches and re- newal shoots form in the axils of the basal leaves (Fig. 3A). The main axis of the shoot enlarges in length to approximately 1—2 cm over the growth season and becomes globose/ovoid (0.5—1.0 cm in diameter) at its distal end (Figs. 3A, 4A); however, shoot size is variable and appears to depend on the age and growth conditions of the individual. Shoots older than one season have a distal globose/ovoid region and a proximal cylindrical region that consists of stem produced in the preceding one or two growth seasons. The main axis of the shoot rarely consists of more than three growing seasons of growth. Some shoots retain their taproot up to their third growth season (Fig. 4B); however, more com- monly the younger shoot axes disarticulate from older portions of rhizomes with taproots. The younger shoot axes form shoot-borne roots associated with nodes of the basal leaves. When a new growth cycle commences, several whorls of foliage leaves expand before the first inflorescences emerge. Each inflorescence has a pair of opposite, sessile leaves (Fig. 4C) and each flower is subtended by a small, oblanceolate bract. Inflorescences develop initially from the axils of the distal leaves in the basal rosette. Subsequently, inflorescences can form in the axils of more proximal leaf positions in the basal rosette, although branches developing in these leaf axils can also form renewal shoots (Fig. 3A). 4 MADRONO mg | Shoot axis Renewal shoots FIG. 3. T | Storage leaf oy Foliage leaf ) Leaf scar [Vol. 53 ae 7 Neu OH) b Shoot-borne root ye Inflorescence Taproot Diagram of shoot system architecture in the Claytonia sibirica complex. A. C. sibirica var. sibirica B. C. sibirica var. bulbillifera C. C. palustris. Arrows labeled ‘a’ show aerial rhizomes, arrows labeled *b’ show hypopodia. Renewal shoots have a basal rosette of helically arranged leaves. Elongation in the lower inter- nodes of these axillary branches (below their rosette of foliage leaves) can create aerial rhizomes that extend renewal shoots 1—5 cm away from the main axis (Fig. 3A). Axillary, aerial rhizomes have shoot-borne roots associat- ed with the nodes of the basal rosette leaves. In most shoots, all leaves are foliage leaves and have a leaf base, petiole and lamina. Foliage leaves have a range of forms, varying in size depending on growing conditions and probably ploidy level, but range from 3—30 cm in overall length and 5-8 cm in blade width (Fig. 4C). Leaf bases are crescentic in cross-section and the width to thickness ratio increases as they age (Fig. 4D— F). Petioles are terete in cross-section and roughly twice the length of the lamina. The laminas of basal leaves in C. sibirica var. sibirica are generally deltoid (Fig. 4C); however, Miller et al. (1984) illustrate a wide range of variation in lamina shape in tetraploid and hexaploid popula- tions. Foliage leaf color is consistently bright green for both eastern and western morphotypes. Some ramets produce late season scale leaves in addition to foliage leaves. The scale leaves consist largely of leaf base and have a rudimen- tary petiole and lamina (Fig. 4G). This hetero- blastic shift occurs uncommonly in populations of the western morphotype of C. sibirica var. sibirica but was not observed among any individuals from populations of the eastern morphotype. Claytonia sibirica var. bulbillifera. This variety has shoot morphology distinct from that of C. sibirica var. sibirica in stature, habit, perennation strategy and leaf specialization. Its shallow, subterranean shoot system is consistently smaller than that of C. sibirica var. sibirica, and its growth habit more lax (Fig. 5A). Claytonia sibirica var. bulbillifera shoot systems are gener- ally similar to those of var. sibirica in producing annually a globose/ovoid, orthotropic axis (Figs. 3B, 5B) that has a basal rosette of helically arranged leaves. Inflorescences form in the axils of leaves in the basal rosette as in C. sibirica var. sibirica. Renewal shoots that form in the axils of the earliest basal leaves can elongate as rhizomes. Claytonia sibirica var. bulbillifera produces specialized storage leaves that have a swollen, succulent leaf base and an unexpanded petiole and lamina (Figs. 3B, 5C—G) at nodes distal to the foliage leaves in the latter part of the growing season (Fig. 3B). At the beginning of the next growing season, these storage leaves can be either decaying or still turgid (Fig. 5D). With the resumption of shoot growth, the axis thickens and elongates distal to the storage leaf zone, new foliage leaves expand as a basal rosette, and inflorescences elongate from those rosette leaf axils. Foliage leaves have a distinct leaf base, petiole and narrowly to broadly elliptic lamina (Fig. 5A), and they are often gray green with a reddish hue, especially when associated with sunny, serpentine sites. Shoot-borne roots emerge in the region between the storage leaves and the newly expanding foliage leaves. By late spring, storage leaves are produced distal to the foliage leaf zone (Fig. 3B). During the summer, shoot systems of C. sibirica var. bulbillifera produce a range of leaf types from the typical foliage leaf described above to a modified form of foliage leaf, which has a succulent leaf base and expanded petiole (Fig. 5F) and lamina, as well as storage leaves (Fig. 5C—G). Inflorescences continue to expand from axillary buds of all leaf 2006] FIG. 4. O’QUINN AND HUFFORD: MORPHOLOGY IN THE CLA YTONIA SIBRIRICA COMPLEX — 5 Shoot system of Claytonia sibirica var. sibirica. A. Globose/ovoid shoot from active growth phase with foliage leaves and inflorescences removed (western morphotype). B. Shoot system of the eastern morphotype that shows three seasons of growth as discrete regions of leaf scars along a continuous shoot axis and retains a taproot. C. Habit. Arrows show opposite leaves subtending the inflorescence and deltoid lamina of the foliage leaf. D. Leaf base of foliage leaf (left) and scale leaf (right). E. Apex of a shoot system, showing the broad leaf base of a foliage leaf and two developing leaves. F. Cross section through the distal portion of a shoot showing the transectional shapes of leaf bases. G. Scale leaves that formed distally to the foliage leaves. agr = active growth rhizome, dl = deltoid lamina, fl = foliage leaf, flb = foliage leaf base, infl = inflorescence axis, lp = leaf primordium, ol = opposite leaves on inflorescence axis, psr = preceding season’s rhizome, sl = scale leaf, slb = scale leaf base, tr = taproot. Scale bar = 3.0 mm in A; 2.0 mm in D, G; 1.86 mm in E; 1.0 mm in F; and | cm in B, C. types throughout the growth season, which is extended for plants growing in more mesic sites. On drier sites, however, the above ground biomass withers and dies by late summer, leaving a shallowly subterranean shoot system that has prominent storage leaves (Fig. 5E). At the end of the growing season, C. sibirica var. bulbillifera preforms the foliage leaves and inflorescence buds that will expand during the next growing season. Claytonia palustris. Claytonia palustris is shal- lowly subterranean and often submerged. It differs from the rest of the sibirica complex in habit, degree of internode elongation, vegetative reproduction, production of modified leaves and size. Shoot systems are weakly orthotropic to plagiotropic, consisting of a swollen ovoid stem with alternately arranged, sheathing leaves with dorsiventrally flattened bases. Leaves form in an open basal rosette that has longer internodes and fewer foliage leaves than the rosettes of the sibirica varieties (Figs. 3C, 6A—E). No leaf specializations for storage or perennation were observed in greenhouse grown or field-collected material. Greenhouse grown material grew only vegetatively. Renewal branches form in the axils of the lowermost leaves of the basal rosette and 6 MADRONO [Vol. 53 FIG. 5. Shoot system of C. sibirica var. bulbillifera. A. Habit. Arrow shows elliptical lamina of the foliage leaf. B. Dissected shoot system showing two seasons of growth. The stem is thicker in the region of active growth than for the preceding season. Foliage leaves, inflorescences and shoot-borne roots have been removed. C. Swollen scale leaves attached to distal portion of a rhizome. D. Shoot system showing overwintered scale leaves proximal to newly expanding foliage leaves and inflorescences. E. Subterranean bulb. F. Shoot system showing characteristics of transition from foliage to scale leaf zones. G. Cross section through distal portion of shoot showing transectional shape of scale leaves. agr = active growth rhizome, fl = foliage leaf, infl = inflorescence axis, lp = leaf primordium, p = petiole, psr = preceding season’s rhizome, rl = rudimentary lamina, sl = storage leaf, slb = storage leaf base, tr = taproot. Scale bar = 1.0 cm in A; 2.0 mm in B, C, E; 5.0 mm in D; 1.0 mm in F, G. inflorescences in the axils of the uppermost (Fig. 3C). Inflorescence axes have a subequal pair of oblanceolate to broadly elliptic leaves and flowers are subtended by small oblanceolate bracts. Under natural and greenhouse growth conditions, C. palustris has a size comparable to C. sibirica var. bulbillifera. The axillary buds that become renewal shoots extend plagiotropically from the axils of rosette leaves and become highly elongated (5—15 cm) (Fig. 6D, E). Most of this elongation 1s in a single, basal internode (1.e., a hypopodium sensu Bell [1991]) that initially has a slightly swollen apical zone with unexpanded leaf primordia (Fig. 6E). The apical zone, (Fig. 6A) which consists of few nodes, becomes orthotropic, undergoes radial thickening in the axis, and foliage leaves expand. Shoot-borne roots are formed at nodes of these swollen, orthotropic renewal shoots, which then replicate the architecture of primary shoots over the course of the growing season. Modified Leaves A. heteroblastic shift from foliage leaves to scale leaves was observed in all examined ramets of C. sibirica var. bulbillifera (Figs. 3B, 5D, 7), but was uncommon among ramets of C. sibirica 2006] O’QUINN AND HUFFORD: MORPHOLOGY IN THE CLA YTONIA SIBRIRICA COMPLEX 7 FIG.6 Shoot system of Claytonia palustris A. Renewal shoot apex showing two leaf primordia at apical meristem. B. Renewal shoot showing prominently swollen axis apex. C. Renewal shoots. D. Stoloniferous habit. E. Renewal shoot with hypopodia. am = apical meristem, fl = foliage leaf, h = hypopodium, Ip = leaf primordium, rs = renewal shoot, sl = scale leaf. Scale bars = 100 um in A; 1.2 mm in B; 1.0 cm in D; 5 mm in C, E. var. sibirica. The scale leaves of both varieties have rudimentary laminas that have a primordial shape and size and are frequently dislodged from the leaf base at maturity (Figs. 4D, G, 5C—F, 7A). All scale leaves are supplied by a single vascular strand, which broadens to form one medial and two lateral bundles that are embedded in a ground tissue of large, starch-filled, isodiametric cells. The epidermis is a single cell layer thick. Scale leaf form, however, differs between the two varieties. Scale leaves of C. sihirica var. sibirica are similar in size and shape to the leaf bases of foliage leaves (Fig. 4D, F, G). In contrast, the scale leaves of C. sibirica var. bulbillifera are radially thicker than the bases of aan lf A B C ayy D FIG. 7 Heteroblastic leaf series from one individual ramet of Claytonia sibirica var. bulbillifera (collected July 2004). A = Swollen scale leaves proximal to foliage leaves of the active growing season. B = Transition leaves with swollen bases, short petioles and small laminas. C = Foliage leaves. D = Swollen scale leaves at distal end of shoot. most foliage leaves, although transitional leaf forms that had a thickened base, short petiole, and small lamina were found among early season foliage leaves directly preceding the formation of foliage leaves (Figs. 5F, 7D). The thickening of scale leaves of C. sibirica var. bulbillifera is centered primarily in cells adaxial to the primary vascular strand, producing a flattened adaxial surface (Fig. SC, G). In constrast, scale leaves of C. sibirica var. sibirica had limited adaxial thickening and retained the adaxial concavity of foliage leaf bases (Fig. 4D—F). Modified leaves in the western morph of C. sibirica var. sibirica were found only in late season collections and always in the distal portion of the shoot. This contrasts with our observations of C. sibirica var. bulbilli- fera, in which the late-forming scale leaves (1.e., storage leaves) persisted through the winter attached to the stem axis and were subjacent to the expanding foliage leaves and inflorescences of the next growing season (Fig 5D). DISCUSSION Being a Bulb Perennial members of the C. sibirica complex have similar globose to ovoid primary shoot axes that bear annually a basal rosette of leaves, from which axillary inflorescences and renewal shoots are formed (Fig. 3A—C). Although these shoot systems are fundamentally rhizomatous (sensu Bell 1991), some variants in the C_ sibirica complex have been described as having bulbs, 8 MADRONO bulblets, or bulbils, and being bulbiferous (Gray 1877, 1887; Miller et al. 1984). Thus, it is important to clarify the bulb aspects of shoot systems in the C. sibirica complex to understand how they represent modifications of the basic rhizomatous shoot system. Bulbs and bulblets are usually described as orthotropic shoot systems that bear fleshy (especially enlarged) scale leaves along very short internodes (Arber 1925; Rees 1972; Dahlgren and Clifford 1982; Bell 1991). Shoot systems of C. sibirica var. bulbillifera meet the criteria for bulb morphology. The production of relatively large, fleshy scale leaves between the cyclic intervals of reproductive growth seen in C. sibirica var. bulbillifera results in a bulb mor- phology that presumably serves as a perennation specialization of the basic rhizomatous form shared with other members of the complex. Gray’s characterization of the KR form as having bulbs in a basal rosette (Gray 1877) and a crown of bulblet-scales (Gray 1887) calls attention to architectural variation: renewal shoots that formed in the axils of foliage leaves can have the form of bulbs when distal scale leaves swell late in the growing season and on the primary axis new succulent scale leaves of the current growing season would be formed as a crown distal to the foliage leaves. Bulbs of C. sibirica var. bulbillifera differ from those found common- ly among various geophytic monocots. For example, geophytic monocots often have a thin, dry scale leaf or leaves (the tunic) that surrounds the entire bulb (Mann 1952; Rees 1972; McNeal and Ownbey 1973). Because they have very short internodes and leaves that lack petioles, it can appear that foliage leaves of geophytic monocots emerge from the rosette of fleshy scale leaves (Arber 1925; Dahlgren and Clifford 1982). Both of these distinctive aspects of monocotyledonous bulbs are lacking in C. sibirica var. bulbillifera, which has neither thin, dry scale leaves nor foliage leaves without petioles. Despite Dahlgren and Clifford’s (1982) assertion that bulbs are a specialization found only in monocotyledons, we and others (Rees 1972; Cronquist 1981; Bell 1991), have recognized that a few clades of dicotyledons have also converged on bulb mor- phology. Claytonia sibirica var. bulbillifera appears to be a serpentine endemic, and we hypothesize that the serpentine environment provided the selection for its bulb morphology. Kruckeberg (1984) discussed the general infertility of serpentine soils and the low turnover of nitrogen and phosphorus in communities associated with these soils. He emphasized that these unique nutritional and chemical characteristics have not only ecological but also evolutionary consequences, namely the origin of endemic species and subspecific eco- types of plants adapted strictly to the serpentine environment. In the KR region, the growing [Vol. 53 season is limited largely to the late winter and spring and the above-ground foliage of herba- ceous perennials has generally senesced by later summer. This relatively short growing season for herbaceous perennials in the KR is reminiscent of that faced by spring ephemerals of eastern deciduous forests. Lapointe (2001) emphasized that subterranean perennating structures, includ- ing bulbs, corms, thick rhizomes, and tubers, were evolutionary responses to the strong selec- tion that spring ephemerals face for the rapid allocation of high levels of nutrients for shoot growth during the early spring when cool temperatures may limit enzymatic activity for photosynthesis. Herbaceous perennials of the KR region would face similar selection; moreover, this selection would be enhanced by the nutrient limitation of the serpentine environment. Thus, selection for a bulb morphology in this complex, in which ancestral heteroblastic variation would have included the formation of thick scale leaves as exemplified by C. sibirica var. bulbillifera, would help to circumvent the early season need for the rapid uptake of nutrients and augment the general nutrient-limited environment imposed by serpentine substrates by making nutrients avail- able largely from scale leaves that are specialized for nutrient storage (and were provisioned over the course of the preceding growing season). Morphological Transitions and Homology Claytonia sibirica. Arber (1925) emphasized the morphological continuity between bulbs and rhizomes, and we observe this transition in C. sibirica. The shoot architecture of both varieties of C. sibirica is largely the same, but in var. bulbillifera we find specialization in the consistent formation of swollen scale leaves distal to the foliage leaves. At the end of the growing season, the bulb of var. bulbillifera consists of a tight aggregation of swollen storage leaves clustered around the preformed, but unexpanded, foliage leaves and inflorescences of the next growing season. Not all ramets of C. sibirica var. sibirica form scale leaves at the end of the growing season, but when scale leaves develop they have largely the size and shape of foliage leaf bases and are arranged in a relatively loose rosette at the tip of the shoot and are fewer in number than the swollen scale leaves of var. bulbillifera (ctf. Fellows 1971). Given the positional and morpho- logical similarity of scale leaves in both varieties, we hypothesize that they are homologous. Miller et al. (1984) suggested that attributes of diploids, such as the morphotypes described here for varieties sibirica and bulbillifera, could have been combined in hybrid populations and this could account for the presence of scale leaves in some ramets of var. sibirica. Alternatively, the formation of scale leaves by some _ perennial 2006] ramets of var. sibirica may simply represent variation in populations irrespective of hybrid- ization or polyploidy. Instead these bulb-like modifications may be similar to the precursors of the distinctly bulbous var. bu/billifera. Additional populations of var. sibirica over its geographic range and habitat conditions need to be sampled for morphological variation, ploidy level, and ancestry to ascertain the phylogenetic homology of shoot system variants. Claytonia palustris. In contrast to Miller’s (1984) description of C. palustris as having ** ..branched rhizomes that are bulbiferous,” we did not observe shoot systems in our sampling of this species that had the morphology of bulbs (cf. also Swanson and Kelley 1987). Primary and renewal axes of C. palustris become swollen and have short internodes that bear scale leaves at the end of the growth season, but these scale leaves do not enlarge as storage structures, a critical feature of bulbs. The initial elongation of axillary renewal shoots is centered in a single internode, a hypopodium (sensu Bell 1991), that functions in a manner similar to the droppers of various monocotyledonous geophytes, (e.g., Erythro- nium), in positioning the orthotropic portion of the renewal axis at a distance from the parent shoot (Arber 1925; McLean and Ivimey-Cook 1951). The hypopodia of C. palustris are homol- ogous to the aerial rhizomes of other members of the C. sibirica complex but differ in the distance they remove renewal shoots from the parent plant. Aside from the formation of hypopodia during the initial elongation of renewal shoots, shoot architecture is very similar in C. palustris and C. sibirica. However, C. palustris is further distinguished from C. sibirica by the formation of leaf bases that completely ensheath the shoot axis, and these leaf bases lack the radial thickening that is common in C. sibirica. Taxonomic Implications Miller et al. (1984) did not recognize the morphotypes in the C. sibirica complex as different taxonomic entities, although they clearly describe morphological variation attributable to genetic differences. Chambers (personal commu- nication) contends that C. sibirica exhibits a high degree of genetic variation over its wide latitu- dinal range but does not find clear delineations between types to warrant taxonomic recognition. However, plants cultivated from seed and grown Over successive years under uniform greenhouse conditions show that plants from the KR maintain a strongly bulbiferous phenotype (O’Quinn unpublished data), from which we infer that shoot system plasticity in the formation of enlarged, fleshy scale leaves is limited. Because of their distinctive bulb morphology, discrete O’QUINN AND HUFFORD: MORPHOLOGY IN THE CLA YTONIA SIBRIRICA COMPLEX 9 geographic distribution and preference for ser- pentine soils, we have followed Gray’s (1887) treatment in recognizing KR populations as C. sibirica var. bulbillifera. ACKNOWLEDGMENTS We thank OSC for loaning herbarium specimens; Lynn Kinter, John Schenk, and Curtis Byj6rk for contributing field collections; Chris Davitt and Valerie Lynch-Holm for assistance with microscopy; Marc Toso for photographic assistance; Ken Chambers for insightful conversations; and editor John Hunter, Ken Chambers, and Pam Diggle for helpful reviews of our manuscript. This work was funded in part by a Betty W. Higinbotham Award and a Noe Higinbotham Award to Robin O’Quinn. LITERATURE CITED ARBER, A. 1925. Monocotyledons: a morphological study. Cambridge University Press, Cambridge. BELL, A. D. 1991. Plant form: an illustrated guide to flowering plant morphology. Oxford University Press, New York, NY. CHAMBERS, K. L. 1993. Claytonia. Pp. 898-900 in J.C. Hickman (ed.), The Jepson manual: higher plants of California. University of California Press, Berkeley, CA. CRONQUIST, A. 1981. An integrated system of classifi- cation of flowering plants. Columbia University Press, New York, NY. DAHLGREN, R. M. T. AND H. T. CLIFFORD. 1982. The monocotyledons: a comparative study. Academic Press, London. FELLowsS, C. E. 1971. A cytotaxonomic study of the origin of Claytonia washingtoniana. M.S. thesis. Humboldt State University, Arcata, CA. GRAY, A. 1877. Characters of new species etc. Proceedings of the American Academy of Arts and Sciences 12:54. . 1887. Contributions to American Botany. XV. 1. Revision of some polypetalous genera and orders precursory to the Flora of North America. Proceedings of the American Academy of Arts and Sciences 22:270-314. JOHANSEN, D. A. 1940. Plant microtechnique. McGraw- Hill Book Company, Inc., New York, NY. KRUCKEBERG, A. 1984. California serpentines: flora, vegetation, geology, soils and management prob- lems. University of California Press, Berkeley, CA. LAPOINTE, L. 2001. How phenology influences physi- ology in deciduous forest spring ephemerals. Physiologia Plantarum 113:151—157. MANN, L. K. 1952. Anatomy of the garlic bulb and factors affecting bulb development. Hilgardia 21:195-251. McLEAN, R. C. AND W. R. IVIMEY-Cook. 1951. Textbook of theoretical botany, Vol. 1. Longmans, Green and Company, London. McNEAL, D. W. AND M. OwNBEY. 1973. Bulb morphology in some western North American species of A//ium. Madrono 22:10—24. MILLER, J. M. 2003. Claytonia L. Pp. 485-488 in Flora of North America Committee, Flora of North America north of Mexico, Vol. 4 Magnoliophyta: Caryophillidae, pt 1. Oxford University Press, New York, NY. 10 MADRONO , K. L. CHAMBERS AND C. E. FELLOWS. 1984. Cytogeographic patterns and relationships in the Claytonia sibirica complex (Portulacaceae). Sys- tematic Botany 9:266—271. O’QUINN, R. L. AND L. HUFFORD. 2005. Molecular systematics of Montieae (Portulacaceae): implica- tions for taxonomy, biogeography and ecology. Systematic Botany 30:314—-331. [Vol. 53 QUARTZ PCI, 1993-1998. Scientific Image Manage- ment Systems. Quartz Imaging Corporation. Ver- sion 4.20. REES, A. R. 1972. The growth of bulbs. Academic Press, London. SWANSON, J. R. AND W. A. KELLEY. 1987. Claytonia palustris (Portulacaceae), a new species from California. Madrono 34:155—161. MADRONO, Vol. 53, No. 1, pp. 11—24, 2006 SYSTEMATICS OF SALVIA PACHYPHYLLA (LAMIACEAE) ROBIN M. TAYLOR! AND TINA J. AYERS Department of Biological Sciences, Northern Arizona University, Flagstaff, AZ 86011-5640 ABSTRACT Most populations of Salvia pachyphylla occur along mountain ranges adjacent to the Mojave Desert of southern California, southwestern Nevada, and northern Baja California, Mexico. A smaller disjunct group occurs in eastern Arizona near the southern edge of Navajo and Hopi reservation lands near Winslow, AZ. This study was undertaken to determine whether there are morphologically and genetically distinct geographical groups within S. pachyphylla and whether these groups form a cohesive unit easily separated from S. dorrii. Specimens of S. pachyphylla and broadly sympatric taxa in S. dorrii were examined in a morphometric analysis of twelve characters. A preliminary molecular analysis using the nuclear ribosomal DNA internal transcribed spacers (ITS-1 and ITS-2) and the embedded 5.8 S subunit was also performed on the same taxa. Morphometric analysis supports the continued recognition of S. pachyphylla and S. dorrii as distinct species and the recognition of three subspecies within S. pachyphylla, requiring two new subspecies, eremopictus and meridionalis, described here. The molecular data support the recognition of the S. dorrii species complex as a whole, but do not support the separation of S. dorrii and S. pachyphylla as distinct species, although the Mexican populations of S. pachyphylla appear genetically distinct. Key Words: Salvia pachyphylla, disjunct distribution, morphometrics, sequence data. Most Salvia pachyphylla Munz_ populations occur in the Transverse Ranges of the California Floristic Province and the mountain ranges of the Mojave Desert of southern California, south- western Nevada, and northern Baja California Norte, Mexico. A smaller, disjunct group occurs in eastern Arizona near the southern edge of Navajo and Hopi reservation lands near the city of Winslow, AZ. This interesting disjunct distri- bution raises questions as to whether the geo- graphical groups are morphologically distinct and whether these groups form a cohesive unit easily separated from S. dorrii. Salvia pachyphylla was first collected by the Parish brothers in the San Bernardino Mountains and described as Audibertia incana var. pachys- tachya by Gray (1878). Samuel B. Parish (1898) elevated this taxon to Audibertia pachystachya. Amos A. Heller (1900) transferred Audibertia pachystachya to the genus Ramona. Harvey M. Hall (1902) transferred Ramona pachystachya to Salvia and recognized it as a variety of S. carnosa [now known as S. dorrii], giving it the name var. compacta. Philip A. Munz elevated it to species level (Salvia compacta), creating a homonym of S. compacta Kuntze (Munz 1927). Finally, Munz (1935) renamed S. compacta as S. pachyphylla Munz. Salvia pachyphylla is a member of the Salvia dorrii (Kellogg) Abrams complex (Strachan 1982), which is comprised of only these two ‘Present address: United States Department of the Interior, BLM Medford District Office, 3040 Biddle Road, Medford, OR 97504. Email: Robin_Taylor@ or.blm.gov. species. Epling (1938) and Strachan (1982) recognized the close relationship between 5S. pachyphylla and S. dorrii based upon morpho- logical characters. They are both woody shrubs with peeling bark, opposite leaves in fascicles, and crowded verticils containing pink to magen- ta-colored bracts. Strachan (1982) used quantita- tive characters of the leaves and flowers to separate the two species. The leaves of S. pachyphylla are usually much larger than those of S. dorrii (20-50 mm vs. 4-30 mm). The inflorescence bracts the corollas are also much longer in S. pachyphylla (bracts 10-20 mm, corollas 17-28 mm vs. bracts 5-14 mm, corollas 9-18 mm in S. dorrii). One of the qualitative differences between the two species is the position of hairs on the corolla. Salvia pachyphylla has a ring of hairs within the lower portion of the corolla tube, whereas S. dorrii flowers possess hairs on the lower lip of the corolla that extend slightly into the throat. The distance between the base of the tube and the hairs is nearly the same in both species, which suggests that the length of the tube determines the final placement of the hairs (Strachan 1982). Salvia pachyphylla has hairs on the adaxial side of the bracts, whereas S- dorrii does not. The two species are hypothesized to be reproductively isolated because they are geographically isolated and flower at different times (Strachan 1982). Salvia pachyphylla flowers from July to October whereas S. dorrii flowers from March to July (September), which Strachan thought would exclude any gene flow. Salvia pachyphylla is found on north facing slopes at elevations of 1500-3050 m. The three subspecific taxa of S. dorrii found in the southwestern U.S. 12 occur at much lower elevations (850—1900 m), even though Strachan (1982) noted higher eleva- tions for the northern part of the range of var. dorrii. Strachan (1982) noted that there may be three morphologically distinct groups within S. pachyphylla. The California plants have very large, obovate leaves (23-63 mm) and _ large bracts (11-20 mm). The Arizona plants have short, spatulate leaves (20-44 mm) and smaller bracts (8-14 mm). The Baja plants have small linear to narrowly spatulate leaves (26-45 mm) and intermediate bracts (10-19 mm). This mor- phological variation suggests either that the three groups have diverged or that the delimita- tion of S. pachyphylla from S. dorrii may have created artificial groups. Two hypotheses are plausible: 1) Salvia pachy- phylla once had a more continuous distribution across the Southwest until climate change, and subsequent contraction of pinyon-juniper wood- land, split the species into the three groups seen today. If this was the case, one would expect to find some degree of morphological and genetic differentiation among populations across the range of the species; 2) The morphological variation seen in the Arizona populations of S. pachyphylla may be a reflection of a closer relationship between the Arizona populations and sympatric subspecies within S. dorrii. If this is true, one would expect to see more continuous morphological variation and genetic similiarity between the two species. To resolve these issues the following ques- tions were asked: 1) Are Salvia pachyphylla and S. dorrii distinct species? 2) Are there morpho- logically or genetically distinct geographical subgroups within S. pachyphylla? 3) If so, how much variation is found within the Arizona populations? 4) Can morphology or genetic markers be used to suggest an origin for the Arizona plants? To answer these questions, morphometric analyses were performed on morphological characters and molecular sequence data were gathered from populations of S. pachyphylla throughout the geographical range of the species and populations representing the three subspe- cies of S. dorrii found in the Southwest. Morphometric analyses are commonly used to study variation among populations and species (Dodd and Helenurm 2000; Battaglia and Patterson 2001). Recent studies of the internal transcribed spacer region (ITS) (Baldwin et al. 1995; Ballard et al. 1999; Meerow et al. 2000; Urbatsch et al. 2000) have shown that this 1s a valuable region for phylogenetic studies at the species level and that the ITS region has enough nucleotide sequence variability for resolution of lower-level phylogenetic questions (Baldwin 1995: Soltis et al. 1998). 2 MADRONO [Vol. 53 METHODS Morphometric Analysis One hundred and eighty-two herbarium speci- mens were used in a morphometric analysis to determine morphological variation within the Salvia dorrii species complex. A complete list of the specimens used is included in Taylor (2002) and most are included in the exsiccatae listed in the taxonomic treatment below. Sample sizes are included in the figures below. Two specimens of S. pachyphylla from two different populations from Nevada were put into the California group due to similar morphological characters. Both S. dorrii subsp. mearnsii and S. dorrii subsp. dorrii were used in the morphometric analysis because populations of each exist in Arizona in close proximity to the populations of S. pachyphylla. Other varieties of S. dorrii were not used due to their physical distance from the Arizona popula- tions of S. pachyphylla. Calipers and a LEICA S6E (0.6—4) dissecting scope were used to measure the 12 characters discussed below. Four categorical and eight continuous characters were used in the analysis. Measurement of bract length and width were determined by taking the average of three measurements using the lowest bract on the first full flowering verticil. Measurement of the hairs on the abaxial side of the bracts was made by averaging the majority of hair lengths. Hairs along the margin of the bract were calculated by averaging the five longest hairs along the margin. Corolla length was measured on rehydrated flowers at full anthesis. Rehydrated corollas were cut longitudinally to discern whether a ring of hairs was present within the corolla and whether there were hairs on the lip. Leaf length and width were averages based upon the three largest leaves on each specimen. The average of two internode lengths was taken starting at the base of the lowest bract and measuring to the base of the next to the last verticil. The adaxial side of the bracts was observed to determine the presence of hairs. The abaxial side of the bracts was observed to determine whether the glands were sunken into the leaf tissue or raised. Nine of the twelve characters (bract width, bract length, abaxial bract hair length, marginal bract hair length, corolla length, presence of hairs within corolla, leaf length, leaf width, internode length) were analyzed using the computer program SYSTAT version 8.0 (SPSS 1998) to perform a Discrimi- nant Function Analysis (DFA) and Principle Components Analysis (PCA). Six of the 12 characters (bract length, bract width, leaf length, leaf width, internode length, corolla length) were analyzed through Analysis of Variance (AN- OVA) using JMP version 4.0.4 (SAS Institute 2001). 2006] TABLE l. Herbarium (ASC). Taxon Salvia pachyphylla AZ Salvia pachyphylla CA Salvia pachyphylla MX Salvia dorriu subsp. Mearnsii Salvia dorrii var. dorrii Salvia dorrii var. pilosa Salvia mohovensis Salvia davidsonii TAYLOR AND AYERS: SALVIA PACHYPHYLLA Locality Meteor Crater Meteor Crater N. Winslow N. Winslow Dilkon Dilkon Petrified Forest Santa Rosa Santa Rosa San Bernardino Mtns San Bernardino Mtns Sierra San Pedro Martir Sierra San Pedro Martir Cottonwood Perkinsville Sedona N. Cameron Cameron Shadow Mtn San Bernardino Mtns San Bernardino Mtns Kingston Mtns San Bernardino Mtns San Bernardino Mtns San Bernardino Mtns Grand Canyon County/State Collector/Coll. # GenBank Accession # Coconino Co., AZ R. Taylor 03B AF538906 Coconino Co., AZ Re. Taylor 03C AF538907 Navajo Co., AZ R. Taylor 04C AF538908 Navajo Co., AZ R. Taylor 04E AF538909 Navajo Co., AZ R. Taylor 26A AF538911 Navajo Co., AZ R. Taylor 27 AF538912 Coconino Co., AZ R. Taylor 19 AF538910 Riverside Co., CA S. Rhodes 9924 AF538913 Riverside Co., CA S. Rhodes 9925 AF538914 San Bernardino S. Rhodes 9926 AF538915 Co., CA San Bernardino S. Rhodes 9928 AF538916 CouCA Baja, Mexico S. Rhodes 00124 AF538917 Baja, Mexico S. Rhodes 00127 AF538918 Yavapai Co., AZ R. Taylor 14A AF538900 Yavapai Co., AZ R. Taylor 15A AF538901 Yavapai Co., AZ R. Taylor 16D AF538902 Coconino Co., AZ R. Taylor 05C AF543682 Coconino Co., AZ R. Taylor 06A AF538898 Coconino Co., AZ R. Taylor 08 AF538899 San Bernardino S. Rhodes 9927 AF538903 Co. CA San Bernardino S. Rhodes 9929 AF538904 Co., CA San Bernardino S. Rhodes 9931 AF538905 Co., CA San Bernardino R. Taylor 13C AF538921 Coz, CA San Bernardino R. Taylor 13E AF53892 Co., CA San Bernardino R. Taylor 11A AF538920 Co. CA Coconino Co., AZ R. Scott 882 AF538919 COLLECTION USED IN THE MOLECULAR ANALYSIS. All specimens deposited in the Deaver Molecular Analysis Twenty-six samples were used in the ITS analysis to represent the three geographic groups of S. pachyphylla and the three subspecific taxa of S. dorrii that are broadly sympatric (Table 1). Salvia davidsonii Greenm. and S. mohavensis Greene were used as outgroups for this analysis because they are southwestern representatives within the genus Sa/via but have never been recognized as part of the Salvia dorrii species complex. Salvia mohavensis is placed within the same section Audibertia and sub-section Jepsonia as the ingroup (Epling 1938). Samples were taken from populations throughout the Southwest over a four-year period. The two Baja California, Mexico samples were obtained from herbarium Sheets in the Deaver Herbarium (ASC). Genomic DNA was extracted from silica-dried leaf tissue and fresh leaf tissue using a modified CTAB protocol of Doyle and Doyle (1987). Quality and quantity were assessed with gel electrophoresis on a 1% agarose gel. The entire ITS region (ITS 1/5.8 s/ITS 2) was then amplified using primers created with the Oligo program version 6.56; primer sequences are as follows: for- ward primer: ITSAL22F 5’ GTTTCCGTAGGT- GAACCTGC 3’; internal forward primer: IT- SAL291F 5’ CTCGGCAACGGATATCTCG 3’; and reverse primer ITSAL693R 5’ TTAAACT- CAGCGGGTGATCC 3’. DMSO was added to aid in the reduction of secondary structure (Soltis et al. 1998). Amplification procedures were as follows: four minutes of denaturing at 95°, thirty seconds at 95°, thirty seconds of annealing at 55°, One minute of extension at 72°, thirty two cycles, ten minute extension at 72°, and a_ holding temperature at 4. To assess possible parology, Polymerase Chain Reaction (PCR) products were cloned, due to noise within the sequence and problems with amplification. Cloning was accomplished using an Invitrogen TOPO TA cloning kit (Electro- poration protocol) following the manufacturer’s recommended procedures. Amplification of the clones were as follows: Five minutes of de- naturing at 95°, thirty seconds at 95°, thirty seconds of annealing at 56°, one minute of 14 MADRONO mo oO Lf. oO FACTOR(2) FACTOR(1) Fic. 1. [Vol. 53 GROUPS 3 S. pachyphylla AZ (n=22) © S. pachyphylla MX (n=23) S. pachyphylla CA (n=101) S. dorrii subsp. mearnsii (n=9) S. dorrii subsp. dorrii (n=27) PCA scatter plot of morphological characters. Factor | is the first principal component and Factor 2 is the second principle component. Sample numbers (n) for each taxon in parentheses. See text for discussion of factors. extension at 72°, thirty five cycles, three minute extension at 72°, and a holding temperature at 4°. PCR concentrations and amplification were assessed electrophoretically on a 1% agarose gel using a low DNA mass ladder. No length mutations were seen in any of the clones. PCR products from three clones were purified using Qiagen’s QIAquick PCR _ purification columns and protocols. Double-stranded PCR products were sequenced on polyacrylamide gels at the Arizona Research Lab in Tucson using a big dye terminator chemistry kit version 2 and an ABI 377 machine. Forward and reverse sequences and ABI electropherograms were edited in DNA STAR- Seqman II version 5.01 (1989;—2001) after reverse complementation to resolve any ambiguities. Sequences were placed in DNA STAR-Megalign version 5.01 (1993:—-2001) and aligned using Clustal W and then aligned visually. All but one of each set of redundant sequences were excluded from the alignments. Sequence align- ments were saved as PAUP files and analyzed using PAUP 4.0b10 (Swofford 2002). Heuristic searches were performed with tree bisection and reconnection (TBR) branch swapping and ran- dom taxon addition. All searches were run using only informative characters. Gaps were coded as missing data and as a 5'" element. One hundred bootstrap replicates were performed using the heuristic search and TBR branch swapping. RESULTS Morphometric Analysis Principal component analysis (PCA) of all specimens produced the scatter plot shown in Fig. 1. The factors (1.e., principal components) are linear combinations of characters that best account for variation in the data. The first factor was comprised mostly of bract length, corolla length, ring of hairs, leaf length, and leaf width. Factor two was comprised mainly of one character, bract hair length. The loadings of variables for each factor can be found in Taylor (2002). The principle components analysis shows clear separation between S. dorrii and S. pachyphylla. Clear separation is also seen between the Arizona populations and the California populations of S. pachyphylla although the Mexico specimens over- lap with both the Arizona and California clusters. The PCA results were identical when examining scatter plots containing specimens of S. pachy- phylla and S. dorrii together or just examining scatterplots containing only S. pachyphylla speci- mens (PCA not shown), the trends still remain 2006] TAYLOR AND AYERS: SALVIA PACHYPHYLLA Ibs A Bract Length ~B_Bract Width 1: - 2 : ' | p<00001 : a P<0.0001 & ' 1.84 : 4 i & 4 i 31.2 H 5 1.64 = S | ———— == anaes t O - —== sf a ane as gies Ee cies : a OC) PM Ge 0.84 L i 1.24 § 4 3 : ) "< O o4, & 0. T T AZ CA MEXICO 95% CA MEXICO 95% Confidence Confidence Locality Interval Locality Interval C Leaf Length D Leaf Width jens eam a 7 65 2.55 @ P <0.0001 ‘ P <0.0001 . 4 a = 4 5 2- : a a a = : Z 5 44 3 H % { - eS ——————— == = at = a) = Ss - i 4 ; ee ® a H = {—-—> a 3 {= 2 2 os ‘ — AZ CA MEXICO 95% AZ CA MEXICO 95% Confidence Confidence Locality Interval Locality Interval E Internode Length F Corolla Length P< 0.0001 ‘ : £ ie : s : ! : : 6 4 s {o) SS ee s @ a eee = Fo — Oo ees ae = & ee. 0 GE aS —— oe 3 > ea [e) s = = . oO ea s i : T T - ~ ~ =] a } a Sra a AZ CA MEXICO = 95% CA MEXICO =: 99% Confidence Confidence Locality Interval Locality Interval Fic. 2. Distribution and 95% confidence intervals for six morphological characters in S. pachyphylla separated by locality. Horizontal line represents overall mean for all groups. Diamonds show means (center line) and the 95% confidence limits (top and bottom lines) for each group. Circles also represent the 95% confidence intervals for each group. the same between the S. pachyphylla geographic shown in Figure 2. All six characters showed groups (Taylor 2002). significant differences among the three geograph- The distributions and 95% confidence intervals ical groups of S. pachyphylla (ANOVA, P < for six of the nine morphological characters are 0.05). For five of the six morphological char- 16 MADRONO [Vol. 53 TABLE 2. CLASSIFICATION MATRIX OF ALL SPECIMENS USED IN THE MORPHOMETRIC ANALYSIS. A B c D E Jo CORRECT A S. pachyphylla AZ 21 0 1 0 0 95 B S. pachyphylla MX l 21 ] 0 0 91 C S. pachyphylla CA l 12 88 0 0 87 D S. dorrii ssp. mearnsii 0 0 0 9 0 100 E S. dorrii ssp. dorrii 0 0 0 0 oF 100 TOTAL 23 33 90 9 ay 91 acters, P values were less than 0.0001. All pairs of the three geographical groups had significantly different means (P < 0.05, Each Pair Student’s t- test). The Mexican populations appear interme- diate between the California and Arizona popu- lations with respect to bract shape (length and width), and internode length. The California populations appear intermediate in corolla length, while the Arizona populations appear intermediate for leaf width and hair length on the floral bracts. The classification matrix, derived from a Dis- criminant Function Analysis (DFA) was used to test group membership based on_ pre-defined geographic distributions of S. pachyphylla. The classification matrix presented in Table 2 had percentages of correct grouping ranging from 87— 100%. This matrix indicates complete taxonomic separation between populations of S. pachyphylla and S. dorrii. No misidentifications were seen. The matrix also provides adequate support for the recognition of all three geographical groups as subspecific taxa within S. pachyphylla. Based upon the matrix produced, the Arizona and Mexican specimens will rarely be misidentified (91-95% correct placement), while the morpho- logical variation found in the Californian popu- lations may result in correct placement only 87% of the time. Only a few of the Californian specimens showed extreme variability. Most of the Californian specimens contributed to the limited variability seen in the results discussed above. A jackknife matrix derived from an analysis of all specimens using the same group- ings ranged from 85—100% correct grouping (Taylor 2002). Molecular Analysis Sequence characteristics of the ITS region within the S. dorrii complex and outgroups are summarized in Table 3. Sequences from cloned DNA showed no variation that would suggest that paralogues were present. Sequences obtained for the ITS 1 region are within the ranges reported by Baldwin et al. (1995), but the size in base pairs (268-277) is larger than all taxa reported, except for the Brassicaceae and Mal- vaceae. The ITS 2 region is within the reported base pair range. The ITS 1 region also contains 27% higher sequence divergence compared to the ITS 2 region. Levels of divergence within the S. dorrii species complex, including both ITS regions and the 5.8 S, ranged between 0.1—7.7%. The levels of divergence were much higher when the outgroups were included with ranges between 9.3-28.8%. Rates of divergence are consistent with other studies that compare lower level taxa (Schilling et al. 1998). The G+C content within the ITS | and ITS 2 are similar to each other and are much higher than G+C content found in all other families except Rosaceae (Baldwin et al. 1995): Of the 695 nucleotide positions sequenced during this molecular analysis, 71 were parsimo- ny informative. One hundred equally most parsimonious trees (length 85) were found using 10 random sequence additions with all characters equally weighted and gaps coded as missing data (consistency index (CI) = 0.90, retention index (RI) = 0.94, rescaled consistency (RC) = 0.84) (Fig. 3). One hundred equally parsimonious trees (length 95) were also found using gaps coded as TABLE 3. SEQUENCE CHARACTERISTICS FOR THE SALVIA DORRIT COMPLEX AND OUTGROUPS. ITS 1 ITS 2 5.88 Aligned Length 268 252 162 Un-aligned Length 268-277 243-252 162 Sequence Divergence within S. dorrii complex (%) 0.1-7.7 0.1—2.5 0.1-1.2 Sequence Divergence including outgroups (%) 13.1—28.8 11.0-21.4 9.3—10.6 G+C Content (%) 65-71 61-71 49-58 Informative Sites within S. dorrii complex 5 6 | Informative Sites including Outgroups 26 29 16 Informative Sites within S. dorrii complex gaps as 5th state 5 8 n/a Informative Sites including Outgroups gaps as 5th state 03 34 n/a 2006] TAYLOR AND AYERS: SALVIA PACHYPHYLLA 17 6 S dorrii subsp mearnsii rt14A 2° S dorri subsp mearnsii rt16D 1 S pachyphylla CA sr9928 S dorrii subsp mearnsii rt15A S dorrii subsp dorrii rtO05C S dorrii subsp dorril rtO6A S dorrii subsp dorrii rt08 S dorrii subsp pilosa sr9927 L S dorrii subsp pilosa sr9931 15 , S pachyphylla AZ rt03B 1 | 1 Ss pachyphylla AZ rt03C , ! S pachyphylla AZ rt04C S pachyphylla AZ rt04E S pachyphylla AZ rt19 : 1 S pachyphylla AZ rt26A : S pachyphylla AZ rt27 99 S pachyphylla CA sr9924 S pachyphylla CA sr9925 1 § dorrii subsp pilosa sr9929 S pachyphylla CA sr9926 15 : S pachyphylla MX sr0124 63 S pachyphylla MX sr0127 S MOHAVENSIS rt13C 48 S MOHAVENSIS rt11A 100 S MOHAVENSIS rt13E S DAVIDSONII rs882 — 1 change Fic. 3. Phylogram of ITS sequence data; one of one hundred equally most parsimonious trees showing branch lengths above the nodes and bootstrap support for selected nodes; initials after taxon names refer to collector/ collection numbers: rt = R. Taylor collections; sr = S. Rhodes collections; rs = R. Scott collections. 18 MADRONO a 5'* state (CI = 0.90; RI = 0.94). No major changes in topology were evident by coding gaps as missing or as a 5" state. There is little resolution between S. pachyphylla and the taxa sampled from the S. dorrii complex with one notable exception, the S. pachyphylla specimens from Mexico. The S. pachyphylla populations from Mexico appear genetically distinct from all of the rest of the S. dorrii complex. The S. pachyphylla populations from the Arizona group form a weak clade. There is no support for a group representing the S. pachyphylla popula- tions that occur in California. The S. dorrii complex as a whole formed a strong clade (bootstrap = 99%). DISCUSSION The morphometric data not only suggest continued recognition of S. dorrii and S. pachy- phylla as separate species, but the data also support recognition of three distinct groups within S. pachyphylla. The molecular data do not fully reflect the amount of morphological divergence seen in the S. dorrii species complex. The molecular data show strong support for the S. dorrii complex as a whole but provide little else, except that the Mexican populations of S. pachyphylla appear to be genetically distinct from the rest of the 3S. dorrii complex. Lack -of molecular divergence has been documented in the Asteraceae where morphology evolves faster than DNA sequences even in rapidly evolving gene regions (Baldwin et al. 1998; Baldwin 2000). ITS sequences in woody plants evolve much more slowly than ITS sequences in herbaceous annuals of recent origin (Baldwin et al. 1995). Some populations of S. pachyphylla and S. dorrii subsp. mearnsii have large, gnarled trunks with peeling bark, which suggests that they may be very long- lived shrubs. A small pilot study looking at the 3’ end of matK and its adjacent spacer region of the chloroplast DNA of four taxa within the S. dorrii complex corroborates the ITS results and sug- gests that S. pachyphylla may be paraphyletic (Taylor 2002). Many different Native American tribes have used both species within the Sa/via dorrii complex for medicinal or ceremonial purposes in the past and some continue to use it today (Zigmond 1981; Huisinga 2001). Salvia pachyphylla current- ly is wild-harvested by the Navajo and Hopi as ceremonial tobacco and medicine (Phyllis Hogan personal communication). Trade can function as a mode of transportation to a new locality, which can help organisms overcome geographical bar- riers enabling genetic and morphological differ- entiation (Brooks and Johannes 1990). Most populations of S. pachyphylla occurring within Arizona are found in close proximity to ruins, suggesting that they might be products of pre- [Vol. 53 historic introduction. If the Arizona populations had shown little or no morphological or genetic variation and had nested within one of the other geographical groups, this mode of introduction, although untestable, might have been more plausible. The results of the PCA and ANOVAs show that there has been significant morphological differentiation of all S. pachyphylla populations in Arizona when compared to the California and Mexico populations. These results thus refute the possibility that the Arizona populations were remnants of trade between groups of indigenous peoples. In addition, there is evidence that the presence of the species predated human migration into northern Arizona. Plant macrofossils from packrat middens are an ideal method for reconstructing past plant species distributions and ages (Cole 1990). The Cricetid rodent, Neotoma stephensi (the packrat), has been known to collect plant material within a 30-100 meter radius from the nest for nest building materials (Van Devender et al. 1987). Packrat urine contains high amounts of calcium oxalate, which aids in solidifying the midden into rock-like deposits (Wells 1976). The crystallized urine envelops and protects the plant macrofossil from decay, preserving it for years to come (Cole 1990). It has been shown that midden deposits are preserved for over 50,000 years in optimal conditions (Wells 1976; Cole 1990). Material from packrat middens show species within the S. dorrii complex to be at least 39,900 years old based on carbon dating of leaves (K. Cole unpublished data). Leaves of S. pachyphylla found in Nevada have been dated to between 10,060—11,940 years old (K. Cole unpublished data), which predates Archaic Cul- tural Groups of PaleoIndians (Fish and Fish 1984). Samples of S. dorrii subsp. dorrii from packrat middens found in Arizona have been dated between 12,015—17,400 years old (K. Cole unpublished data). There are no known speci- | mens of S. pachyphylla from middens in Arizona — to date, but, based upon the midden data — presented above, it is plausible that S. pachy- | phylla occurred in Arizona well before settlement _ by PaleoIndian groups. The present day distri- bution of the three extremely small populations | of S. pachyphylla in Arizona adjacent to ruins | may be the result of past wild-harvesting or natural remnant populations that are soon to > become extinct. The fourth Arizona population | of S. pachyphylia is extensive and has very large | individuals, seedlings, and many different in- | termediate age classes indicating that the occur- | rence of S. pachyphylla in Arizona is probably | not a remnant of past use by Paleoindians. | The research presented here is not definitive | but suggests that the distribution of S. pachy- — phylla was once more continuous, with vicariant | 2006] events causing the contraction of the popula- tions to the geographically distinct groups seen today. This vicariance might have been caused by climate changes such as rising temperatures and decreasing moisture during the formation of the Mojave Desert (Raven and Axelrod 1978; Stott 1981; Axelrod 1983: Schaffer 1993). The contraction to the current disjunct distribution in Arizona might also have occurred as late as the Holocene, when pinyon-juniper woodland again dominated the landscape until warmer, drier climate resembling present day conditions occurred at the end of the Pleistocene (Martin and Mehringer 1965). Examples of disjuct distributions from northern Baja California, Mexico and northeastern Arizona are known from other taxa such as Errazurizia Phil. and phylogenetic studies of these groups might lead to a better understanding of southwestern biogeog- raphy. Results of the morphological data support the recognition of subspecific taxa within S. pachy- phylla, as presented below. Due to the size of most of the Arizona populations and the importance of this plant to tribes in the Southwest, monitoring of populations and of annual harvesting 1s recommended. Future re- search within the S. dorrii complex should include larger sample sizes, additional molecular mar- kers, and ecological observations to understand whether the recognition of one or two distinct species within the Salvia dorrii species complex is warranted. TAXONOMIC TREATMENT Salvia pachyphylla Munz Rose Sage Aromatic, branching perennial shrub with gray peeling bark 35-100 cm tall 40-150 cm wide, generally much wider than tall. LEAVES oppo- site, fascicled, glandular, 2.0-6.3 cm long and 0.5-2.5 cm in wide, fleshy, obovate to rhombic, attenuate at the base of the leaf, tip acute to obtuse, abaxial and adaxial sides covered in appressed white hairs. INFLORESCENCE of l-many verticils subtended by many bracts: bracts green to magenta in color, glandular, ciliate, scarious, rotund to orbicular, pubescent. FLOWERS several per verticil; calyx connate, lobed at the top, green to purple, pubescent: corolla connate, blue to violet, limb comprised of an upper lip and two bilateral lobes, tube containing a ring of hairs; stamens 4, exerted from corolla. FRUIT 1-4 nutlets. SEEDS tan to black. Found on north facing slopes in conifer forests, 1219-2830 m (4000-9284 ft). Three sub- species distributed in southwestern North Amer- ica (Fig. 4). Uses: smoke or tea to calm the mind, and used in the treatment of epilepsy (Whiting | 1939; Zigmond 1981). TAYLOR AND AYERS: SALVIA PACHYPHYLLA Ae KEY TO THE SUBSPECIES OF S. PACHYPHYLLA 1. Internode length between verticils (6.5-) 8.5— 10 (-15.5) mm, bract width (3.5-) 5.5—5.9 (9.2) mm, endemic to northeastern Arizona S. pachyphylla subsp. eremopictus 1. Internode length between verticils (14.5-) 18.5—25 (-45) mm, bract width (6.5-) 9-13 (-15.7) mm. 2. Shrubs with a well-defined woody trunk, leaf length (2.6-) 3—3.6 (-4.4) mm, leaf width (5-) 7-11 (-14) mm, endemic to Baja California, Mexico Pas nlie dk Geen Wee Sate S. pachyphylla subsp. meridionalis Subshrubs mostly branching below _ the ground, leaf length (2.3-) 3.7-4.2 (-6.3) mm, leaf width (10-) 13—17 (-25) mm, southern California and Nevada N Salvia pachyphylla P. A. Munz subsp. pachy- phylla Rose Sage Ulustration: Brittonia 34(2): 167. 1982. Audibertia incana var. pachystachya A. Gray, Syn. Fl. N. Amer., ed. 2. 2(1): 461. 1886. A. pachystachya Parish, Erythea 6:91. 1898; not Salvia pachystachya Trautv. Ramona pachysta- chya A. A. Heller, Muhlenbergia 1:4. 1900. Salvia carnosa var. compacta H. M. Hall, Univ. Calif. Publ. Bot. 1:111. 1902, nom. superfl. S: compacta Munz, Bull. S. Calif. Acad. Sci. 26:22. 1927; not S. compacta Kuntze, 1891. S. pachy- phyulla Epling ex Munz, Man. S. Calif. Bot. 445. 1935. Type: UNITED STATES. Califor- nia: San Bernardino Co.: Bear Valley, San Bernardino Mts., Aug 1882, Parish & Parish 330 (lectotype, GH; isolectotypes A, DS, F). Subshrubs extensively branching below the ground, 30—-45.5 cm tall, 61-120 cm wide. IN- TERNODES 1.2-4.5 cm. LEAVES obovate, 2.3-6.3 cm long, 1.0—2.5 cm wide. FLOWERS 1.2-2.5 cm long. BRACTS many; 1.1—2.03 cm long; 0.8—-1.57 cm wide; hairs on bracts 0.01- 0.02 mm long; hairs on bract margins 0.01— 0.055 mm. SEEDS 3 mm long, 2 mm wide. Paratypes. U.S.A., California. Inyo Co., Pana- mint Mtns, Death Valley National Monuemnt, Aguereberry Point, 1961 m, 02 July 1983, RF. Thorne 56130.1 (RSA); Jail Canyon, 2438 m, 11 July 1977, A. P. Romspert 13 (RSA): Rogers Peak, 2591 m, 10 July 1974, L. DeBuhr 44818 (RSA): Wildrose Canyon, 1920 m, 3 July 1974, L. DeBuhr 44793 (RSA); T19S R45E sect. 27, 1951 m, 15 August 1968, J.L. Reveal 1786 (RSA); Dolomitic Rocky Ridge, UTM 497130E, 4023510N, 1951 m, 2 July 1983, P.M. Peterson 1183 (RSA); Narrow Canyon above Townes Pass, 17 June 1937, C. Epling (RSA); Thorndike’s Ranch, 2286 m, 7 July 1937, C. Epling (RSA);Wild Rose Canyon to Telescope Peak, 2682 m, 8 July 1937, P.A, Munz 14793 (RSA); 2133 m, 16 May 1931, R. Hoffmann 459 (RSA); Kern Co., Scodie Mtns, Walker Pass Trailhead on Pacific Crest Trail, T26S R37E sect. 19, 1707 m, 6 August 1988, B. 20 MADRONO 120°0'0"W [Vol. 53 115°0'0"W 110°0'0"W Legend * Salvia pachyphylla subsp. eremopictus Salvia pachyphylla subsp.meridionalis + Salvia pachyphylla subsp. pachyphylla 120°0'O"W FIG. 4. Ertter 7891 (NY, RSA); Riverside Co., San Jacinto Mtns, 1402 m, June 1901, H.M. Hall 2160 (NY, RSA); north slope, 1829 m, 20 August 1922, E.C. Jaeger 1010 (RSA); Tahquitz Peak, 22 August 1933, L. Crutcher 3 (RSA); Tahquitz Valley, 2652 m, 31 August 1930, J. Ewan 2149 (RSA); Santa Rosa Mtns, 1981 m, 21 August 1952, P.A. Munz 17995 (NY, RSA); 2 September 1970, C.W. Tilforth 297 (RSA); dry _ slopes, 2134 m, 29 June 1922, P.A. Munz S&88 (RSA); near Santa Rosa Springs campground, 2088 m, 15 115°0'0"W 110°0'0"W Distribution of the three subspecies of Salvia pachyphylla. June 1978, C. Davidson 7382 (RSA); 2138 m, 21 August 1971, N.R. Zabriskie 438 (RSA); near | Vendeventer Flat, 1676 m, | July 1933, V. Duran 3490 (NY,RSA); stony slopes near creek, 1524 m, | 25 June 1922, P.A. Munz 5813 (RSA); Toro Peak, | 14 August 1938, P.A. Munz 15363 (RSA); Vandeventer Flat, 2073 m, 21 August | 1952, P.A. Munz 17981 (NY, RSA); Cultivated, 2 | 2438 m, September 1954, E.K. Balls 19778 (NY, RSA); virgin spring, 2652 m, 14 August 1938, P.A. Munz 15346 (RSA): Toro Mtn, July 1901, ELE. Schel- 2006] lenger (RSA); San Bernardino Co., Blk. Hawk Mine, near Victorville, 4 July 1926, M.E. Jones (RSA); Clark Mtns 4.2 miles NNW of Mountain Pass, 2134 m, 6 July 1972, B. Prigge 2 (RSA); 4 miles NW of mountain pass, 2250 m, 6 July 1973, B. Prigge 1197 (NY, RSA); Pachalka Spring, 1981 m, 6 October 1935, CB. Wolf (RSA); Hanna Flats, near Fawnskin, 1829 m, 23 July 1941, G.7. Hastings (NY); Keystone Mine, 45(air)miles E of Baker, 5.5 (air)miles Sof Ivanpah at Jct. Roads to Keystone Mine and Keyston Springs, 1680 m, 29 August 1973, J. Henrickson 12620 (RSA); Kingston Mtns, 2 miles from peak, 1676 m, 23 October 1977, J. Henrick- son 16297 (RSA); Kingston VABM, 2136 m, 18 September 1980, S. Castagnoli 228 (RSA); Mid- Shut-Up canyon, 1524-1676 m, 23 October 1977, J. Henrickson 16305 (NY); on a rocky ridge, 2134 m, 27 July 1949, J. Roos 4507 (RSA); Porcupine Canyon, 2134 m, 21 September 1980, R.F. Thorne 54790 (RSA); New York Mtns, Keystone Canyon, 1580 m, 7 July 1973, J. Henrickson 11056 (RSA); 1737 m, 29 October 1976, R.F. Thorne 47965 (RSA); 29 July 1952, P.C. Everett 17299 (RSA): Keystone Spring, 1707 m, 13 October 1935, P.A. Munz 13874 (RSA); San Bernardino Mtns, 8.7 miles SE of Lucerne Valley, 1398 m, 18 June 1978, C. Davidson 7300 (RSA); Barton Flats, 1981 m, 30 October 1955, L. Benson 15607 (RSA); Bear Lake, 2000 m, 7 July 1931, E.W. Clokey 5292 (NY); Bear Valley, | August 1901, L.R. Abrams 2077 (RSA); 1895, A. Davidson (RSA); Big Bear Lake, 22 August 1935, C.L. Hitchock 2825 (RSA); 6 July 1924, J. M.J. (RSA); 8 August 1964, BC. Templeton 10191 (RSA); Big Meadows, 1990 m, 27 July 1925, J.B. Feudge 1242 (RSA); 2134 m, N.C. Cooper 2886 (RSA); Cactus Flat, 1829 m, 25 June 1926, P.A. Munz 10501 (RSA); Cactus Flat, August 1915, F. Grinnell (RSA); Camp Osceola, on the upper Santa Ana River, 1829 m, 21 July 1936, E.R. Johnson (NY ); Coon Creek, Heart Bar State Park, 34°10’N, 116°45'W, 2286 m, 9 August 1992 , S.D. White 599 (RSA); Cushenbury canyon, 1450 m, 5 May 1978, R.F. Thorne 51874 (NY); 23 September 1927, M. Jones (RSA); Cushenbury Grade, 1219m 9 July 1927, J.T. Howell 318 (RSA); Fish Creek, 2743 m, 14 July 1924, P.A. Munz 8497 (RSA): Santa Ana River, 1966 m, 22 October 1931, CB. Wolf (NY); Foxesee Creek, 2438 m, 22 August 1920, FW. Peirson 1060 (NY); from Bear Lake to Holcomb Valley, 2134 m, 5 July 1930, FW. Peirson 9011 (RSA); Heart Bar campground, 34°10'N, 116°43’W, 2438 m, 17 July 1989, B. Wagner (RSA); Holcomb Valley, 3N10, 3N16, 7390 m, 12 & 13 July 1979, RF. Thorne 53493 (NY); Green Lead Mine road, 2195 m, 7 August 1931, J/. Ewan 4867 (NY); , T3N R1IW sect. 34/26, 2134, 27 June 1979, J. Strachan 2994 (NY); Johnson Grade, 1981 m, 5 July 1935, M.B. Dunkle 4015 TAYLOR AND AYERS: SALVIA PACHYPHYLLA a (NY); Baldwin Lake, 1951 m, 22 August 1932, POA, Mung 12707. (SSA): Ubucerme:+-V alley, OMYA’s Crystal Creek Haul road, 34°20.5'N, 116°56.5 W, 1770 m, 27 August 1998, S.D. White 7092 (NY); Marble Canyon, 34°20.5'N, 116°52.5'W, 1585 m, 26 August 1998, S.D. White 7094, 7097 (RSA); near Bear Valley, September 1893, T. Minthorn (RSA); Nelson Ridge, south of Smarts Ranch, 34°16'.1"N, 116°45'37.1"W, 1951 m, 29 July 1998 , S. Boyd 10259 (NY); Old Rose Mine, 2134 m, 9 October 1937, P.A. Munz 14955 (NY); Onyx summit, 2450 m 30 August 1975, C. Davidson 3234 (RSA); 34°11'13’N, 116°42'53”"W, 2551 m, 27 March 1999, S. Rhodes 9928 (RSA); Rose Mine, 2134 m, 21 October 1945, H. Crooks 93 (NY); Santa Ana Canyon, 24 July 1906, H.M. Hall 7549 (NY, RSA); Santa Ana River, 1920 m, 21 August 1922, P.A. Munz 6147 (RSA); SE Terrace Springs and W of Arrastre Creek, 34°19'44"N, 116°45’58”"W, 1463 m, 25 June 1998, V. Soza 308 (RSA): Seven Oaks, July 1902, C. Wilder 395 (NY); South Fork Public campground, 1981 m, 24 July 1947, P.A. Munz 12053 (NY, RSA); South Fork, 1890 m, 26 July 1906, J. Grimmell 307 NY; Sugarloaf Mountain, 2591 m, 22 July 1926, P.A. Munz 10779 (RSA); 3 August 1932, F.R. Fosberg 8617 (RSA);Warrens Well, 1280 m, 30 June 1938, F.C. Jaeger (NY); San Diego Co., Tantillas Mtns, 1875, E. Palmer 304 (NY, RSA); Tulare Co., Chimney Creek Campground, opposite Lamont bench mark, 3 miles S of the BLM Chimney Creek Camp- ground, 29 June 1985, D.W. Mc Neil 3110 (NY); Kern Plateau, east of Long Valley, 1676—1859 m, 8 August 1967, J.T. Howell (NY); Kernville, head of Sand Canyon, T24S R36E sec29Se, 1981 m, 10 June 1986, B. Ertter 6382 (NY). Distribution. Inyo, Kern, Riverside, San Ber- nardino, San Diego, Cos., California and Clark and Tulare Cos., Nevada. 1219-2682 m (4000— 8800 ft). Flowering from June—October. Habitat. North facing slopes. Loose sand, limestone, or granitic soil. Found among pines and junipers. Epithet etymology. The epithet refers to the thick leaves. Salvia pachyphylla subsp. eremopictus R. Taylor subsp. nov. (Fig. 5) Arizona Rose Sage— TYPE: USA, Arizona: Navajo Co., 16.5 mi N of Interstate 40 on Hwy 87, just past mile marker 362; 2.5 mi N of Little Painted Desert State Park. UIM Zone, 12-5, 35503375, 3893128N, 1676 m elev., 28 October 1999, R. Taylor 04 (holotype, RSA; isotypes, ARIZ, ASC, ASU, NY). Similis subsp. pachyphylla, sed differt floris internodis brevis, 6.5—-15.5 mm longi, bracteae 3.5—9.2 mm lata. pe: MADRONO Fic. 5. IHustration of Salvia pachyphylla subsp. eremopictus: A—flowering branch; B—longitudinal view of flower; scale bars = | cm. Shrubs with well defined woody trunk, 35-— 50 cm tall, 40-150 cm wide. LEAVES spatulate, 2.0-4.4 cm long, 0.6—1.7 cm wide. INFLORES- CENCE internodes 0.65—1.55 cm. FLOWERS 1.3—2.45 cm long. BRACTS many; 0.83—1.4 cm long; 0.35—0.92 cm wide; hairs on bracts 0.01 mm long; hairs on bract margins 0.01—0.03 mm. SEEDS 2.5—4 mm long, 1.5—2.5 mm wide. Paratypes. U.S.A., Arizona. Apache Co., Petrified Forest National Monument, Chinde Mesa, 12S 603424E 3892175N, 1804 m elev., 17 July 1998, M. Hansen s.n. (ASC); 2 September 2000, R. Taylor 19 (ASC); Coconino Co., Meteor Crater, N-facing slope of crater, 35°01.923’'N, 111°01.520'W, 1646 m elev., 28 October 1999, R. Taylor 03 (ASC); 18 October 1998, M. Hansen s.n. (ASC); 26 May 1994, J. Beasley s.n. (ASC); 21 September 1998, S. Hill 31025 (ASC); Navajo Co., Navajo Reservation, NW of Dilkon, 35°28'N, 110°19’W, 1937 m, 7 October 2001, R. Taylor 26 (ASC, NY, RSA). Distribution. Salvia pachyphylla subsp. eremo- pictus is known only from the southern Colorado Plateau in Apache, Coconino and Navajo Coun- ties on Navajo reservation, National Park Ser- vice, State land, and private land at 1539-1937 m (5500-6356 ft) (Fig. 4). Flowering from (May) July—October usually after summer monsoon rains begin. [Vol. 53 Habitat. Salvia pachyphylla subsp. eremopictus is found on barren north-facing slopes and washes on basalt and painted desert soils derived from Chinle shale. Found among juniper and salt bush. Epithet etymology. The subspecific epithet refers to the “‘Painted-Desert” substrates derived from Chinle shale where historical populations of this taxon had been found. During this study three new populations were found on volcanic substrates adjacent to “‘Painted-Desert” forma- tions. Salvia pachyphylla subsp. meridionalis R. Taylor subsp. nov. Baja Rose Sage Illustration: Brittonia 34(2): 167. 1982.—_TYPE: MEXICO, Baja California Norte, Sierra San Pedro Martir, La Encantada, rock hillsides about margin of meadow, 2100 m elev., 18 Septem- ber 1930, 1. L. Wiggins and D. Demaree 48872 (holotype, RSA; isotypes, NY). Similis subsp. pachyphylla, sed differt folia anguste, 5-14 mm lata. Shrubs with well defined woody trunk, 30— 45.5 cm tall, 61-120 cm wide. INTERNODES 1.4-3.35 cm. LEAVES linear to spatulate, 2.6— 4.5 cm long, 0.5—1.4 cm wide. FLOWERS 1.5— 2.1cm long. BRACTS many; 0.97-1.9 cm long; 0.63—1.4 cm wide; hairs on bracts 0.01— 0.015 mm long; hairs on bract margins 0.03— 0.06 mm. SEEDS 3.0—4.0 mm long, 2.0—3.0 mm wide. Paratypes. MEXICO, Baja California Norte. Observatoria UNAM, San Pedro Martir, 2830 m, 13 October 1985, A. Gonzalez (ASC); La Grulla, San Pedro Martir, 2286 m, 1926, C.G. Abbott (NY); UNAM Observatory, San Pedro Martir Mtns National Park, 1372 m, 5 August 1995, H.D. Hammond 10844 (NY); UNAM Observatory, San Pedro Martir Mtns National Park, Vallecitos, 2590-2743 m, 1 September 1985, J. Donahue 96055 (NY); Cerro Verado Blanco, San Pedro Martir, north-north west of the observatory, 31°3’N,115°9'W, 2345 m, 16 September 1998, J. Rebman 5610 (RSA); Laguna Hanson, 9 July 1938, M.B. Dunkle 5416 (RSA); Yerba Buena, Sierra San Pedro Martir, 31°00'N, 115°27'W, 16 August 1967, Moran! Thorne 14200 (RSA); La Corona, Sierra San Pedro Martir, 30°58’'N, 115°35'W, 2000 m, 30 August 1963, R. Moran 11272 (RSA); Rancho Mezquite, Sierra Juares, 32°18'’N, 116°00’W, 1450 m, 3 September 1966, R. Moran 13446 (RSA); Arroyo Copal, Sierra San Pedro Martir, 31°04’N, 115°28'W, 2550 m, 24 August 1968 R. Moran 15436 (RSA); Laguna Hanson, Sierra de Juares, Con- stitucion National Park, 32°02.5’N, 115°55’W, 1610 m, 15 September 1983, R.F. Thorne 57116 (RSA); Sierra Juarez, Estado de Baja California, 2006] 3 miles east of Laguna Hanson, 32°02'N, 115°52'W, 26 July 1994, J. Rebman 2839 (NY);14 miles south of La Rumerosa, 32°21'N, 116°00’W, 29 June 1962, R. Moran 9805 (RSA); La Corona de Abajo, Parque Nacional Sierra San Pedro Martir, 2080 m, 27 August 1988, R. Noyes 638 (RSA); Cerro Botella Azul, Sierra San Pedro Martir, 19 July 1988, S. Boyd 2643 (RSA); Vallecitos, Sierra San Pedro Martir, 1 mile S of La Tasajera, 20 July 1988 , S. Boyd 2725 (RSA); Vallecitos, Sierra San Pedro Martir, 2456 m, 21 September 1930 , Wiggins & Demaree 4971(NY, RSA); La Encantada, Sierra San Pedro Martir, 30°55’N, 115°24’W, 2200 m, 19 August 1967, Moran & Thorne 14370 (RSA); La Encantada, Wiggins & Demaree 4887 (NY, RSA); La Encantada, Sierra San Pedro Martir, 2200, 18 September 1930, Wiggins & Demaree 4892 (NY, RSA); Yellow pine belt, between Ojos Negros and Neji Rancho, 16 September 1929, Wiggins & Gillespie 4137 (NY, RSA). Distribution. Salvia pachyphylla subsp. meridio- nalis is distributed in the Sierra San Pedro Martir and Sierra Juarez from 32°00’N, 116°00’W to 30°25'’N, 115°30’W at 1372-2830 m (4500— 9284 ft) (Strachan 1982, Fig. 2). Flowering from June-August. Habitat. Salvia pachyphylla subsp. meridionalis is found on north-facing rocky slopes derived from coarse sand and granitic soil. Found among pines. Epithet etymology. The subspecific epithet refers to the southern-most distribution of this taxon. ACKNOWLEDGMENTS We gratefully acknowledge loans from the curators of NY and RSA. Special thanks to H. David Hammond, Monica Hansen, Phyllis Hogan, Kristin Huisinga, Suzanne Rhodes, Daniela Roth, and Susie Vogel for field collections. The Navajo Nation and Petrified Forest National Park graciously provided plant collecting permits. Stephen Shuster, Randy Scott, and Mar-Elise Hill provided technical assistance with the statistical analyses. H. David Hammond, Nancy Morin, and two anonymous reviewers provided con- structive reviews of the manuscript. Field work was supported by a Northern Arizona University (NAV), Henry Hooper Research grant. This paper is a portion of a thesis submitted to NAU, by the senior author, in partial fulfillment of a Masters degree. LITERATURE CITED AXELROD, D. I. 1983. Paleobotanical history of the western deserts. Pp. 113-130 in S. G. Wells and D. R. Haragan (eds.), Origin and evolution of deserts. University of New Mexico Press, Albu- querque, NM. BALDWIN, B. G. 1995. A new prospect for California botany: integrating biosystematics and phyloge- netics. Madrono 42:154—167. TAYLOR AND AYERS: SALVIA PACHYPHYLLA ZO —. 2000. Roles for modern plant systematics in discovery and conservation of fine-scale biodiver- sity. Madrono 47:219—229. , M. J. SANDERSON, J. M. PORTER, M. F. WOJCIECHOWSKI, C. S. CAMPBELL, AND M. J. DONOGHUE. 1995. The ITS region of nuclear ribosomal DNA: a valuable source of evidence on angiosperm phylogeny. Annals of the Missouri Botanic Garden 82:27—277. , D. J. CRAWFORD, J. FRANCISCO-ORTEGA, S.C. KIM, T. SANG, AND T. F. Stussy. 1998. Molecular phylogenetic insights on the origin and evolution of oceanic island plants. Pp. 410-441 in P. S. Soltis, D. E. Soltis, and J. J. Doyle (eds.), Molecular systematics of plants. Il: DNA sequencing. Kluwer Academic Publishers, Boston, MA. BALLARD, H. E., K. J. SYTSMA, AND R. R. KOWAL. 1999. Shrinking the violets: phylogenetic relation- ships of infrageneric groups in Viola (Violaceae) based on internal transcribed spacer DNA se- quences. Systematic Botany 23:439-458. BATTAGLIA, R. E. AND R. PATTERSON. 2001. A morphometric analysis of the Leptosiphon androsa- ceus complex (Polemoniaceae) in the central and south coast ranges. Madrono 48:62—78. Brooks, R. R. AND D. JOHANNES. 1990. Phytoarch- aeology. Historical, ethno- and economic botany series, Vol. 3. Dioscorides Press, Portland, OR. CoLeE, K. L. 1990. Reconstruction of past desert vegetation along the Colorado River using packrat middens. Palaeogeography, Palaeoclimatology, Pa- laeoecology 76:349—366. Dopp, S. C. AND K. HELENURM. 2000. Floral variation in Delphinium variegatum (Ranuncula- ceae). Madrono 47:116—126. DOYLE, J. AND J. DOYLE. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19:11—15. EPLING, C. 1938. The California Salvias: A review of Salvia, section Audibertia. Annals of the Missouri Botanic Garden 25:95—189. FisH, S. K. AND P. R. FISH. 1984. Prehistoric agricultural strategies in the southwest. Anthropo- logical Research Papers No. 33. Arizona State University, Tempe, AZ. GRAY, A. 1878. Synoptical flora of North America. Ed. 1, Vol. 2, pt. 1. Iverson, Blakeman, Taylor and Co., New York, NY. HALL, H. M. 1902. A botanical survey of San Jacinto Mountain. University of California Publications in Botany 1:1—140. HELLER, A. A. 1900. Some changes in nomenclature. Muhlenbergia 1:1-8. HUISINGA, K. D. 2001. Cultural influence as a factor in determining the distribution of a rare sage, Salvia dorrii subspecies mearnsii. Pp. 228-237 in J. Maschinski and L. Holter (eds.), Southwestern rare and endangered plants: proceedings of the Third Conference... Proceedings RMRS-P23. U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fort Collins, CO. MARTIN, P. S. AND MEHRINGER, P. J. 1965. Pleistocene pollen analysis and biogeography of the Southwest. Pp. 443-451 in H. E. Wright and D. G. Frey (eds.), The Quaternary of the United States. Princeton University Press, Princeton, NJ. MEEROW, A. W., C. L. Guy, Q. LI, AND S. YANG. 2000. Phylogeny of the American Amaryllidaceae 74 MADRONO based on nrDNA ITS sequences. Botany 25:708—726. Munz, P. A. 1927. The southern California species of Salvia (including Ramona). Bulletin of the Southern California Academy of Science 26:17—29. . 1935. A manual of southern California botany. Scripps Publishing Fund, Claremont, CA. PARISH, S. B. 1898. Little or little-known plants of southern California: I. Erythea 6:85—92. RAVEN, P. H. AND D. I. AXELROD. 1978. Origin and relationships of the California flora. University of California Publications in Botany 72:1—134. SAS INSTITUTE, INC. 2001. JMP version 4.0.42. SAS Institute, Inc., Cary, NC. SCHAFFER, J. P. 1993. California’s geological history and changing landscapes. Pp. 49-54 in J. C. Hickman (ed.), The Jepson manual: higher plants of Califor- nia. University of California Press, Berkeley, CA. SCHILLING, E. E., L. C. RANDAL, AND R. D. NOYES. 1998. Phylogenetic relationships in Helianthus (Asteraceae) based on nuclear ribosomal DNA internal transcribed spacer region sequence data. Systematic Botany 23:177—187. SOLTIS, D. E., P. 8S. SOLTIS, AND J. J. DOYLE. 1998. Molecular systematic of plants Hl: DNA sequenc- ing. Kluwer Academic Publishers, Boston, MA. SPSS INc. 1998. SYSTAT version 8.0° graduate pack., SPSS Inc., Chicago, IL. STOTT, P. R. 1981. Historical plant geography: an introduction. Allen & Unwin, London, England. STRACHAN, J. L. 1982. A revision of the Salvia dorrii complex (Lamiaceae). Brittonia 34:151—169. Systematic [Vol. 53 SWOFFORD, D. L. 2002. PAUP* (Phylogenetic analysis using parsimony (*and other methods) Version 4.010b. Sinauer Associates, Sunderland, MA. TAYLOR, R. M. 2002. The origin and adaptive radiation of Salvia pachyphylla (Lamiaceae). M.S. thesis. Northern Arizona University, Flagstaff, AZ. URBATSCH, L. E., B. G. BALDWIN, AND M. J. DONOGHUE. 2000. Phylogeny of the coneflowers and relatives (Heliantheae: Asteraceae) based on nuclear rDNA internal transcribed spacer (ITS) sequences and chloroplast DNA restriction site data. Systematic Botany 25:539—565. VAN DEVENDER, T. R., R. S. THOMPSON, AND J. L. BETANCOURT. 1987. Vegetation history of the desert southwestern North America: the nature and timing of the Late Wisconsin-Holocene tran- sition. Pp. 323—352 in W. F. Ruddiman and H. E. Wright (eds.), The geology of North America: North America and adjacent oceans during the last glaciation, Vol K-3. Geological Society of Ameri- ca, Boulder, CO. WELLS, P. V. 1976. Macrofossil analysis of wood rat (Neotoma) middens as a key to the Quaternary vegetational history of arid America. Quaternary Research 6:223—248. WHITING, A. F. 1939. The Ethnobotany of the Hopi. Museum of Northern Arizona Bulletin, Num- ber 15. Museum of Northern Arizona, Flagstaff, AZ. ZIGMOND, M. L. 1981. Kawaiisu ethnobotany. Uni- versity of Utah Press, Salt Lake City, UT. MADRONO, Vol. 53, No. 1, pp. 2535, 2006 THE EVOLUTION OF PINK: MORPHOLOGICAL AND GENETIC VARIATION AMONG THREE LITHOPHRAGMA (SAXIFRAGACEAE) SPECIES REBECCA A. HUFFT AND JAMES E. RICHARDSON! Department of Ecology and Evolutionary Biology, University of California Santa Cruz, Santa Cruz, CA 95064 hufft@darwin.ucsc.edu ABSTRACT Prior molecular work using ITS and chloroplast sequence data revealed that the endemic Lithophragma trifoliatum and the broad-ranging L. parviflorum form a tight clade with a third species, L. affine. We used AFLPs to assess the fine scale relationship of these three species from populations where their distributions overlap in northern California. Our results revealed two groups of L. trifoliatum, one nested within a group of L. affine and L. parviflorum and the other grouped with other populations of L. parviflorum, contrary to predictions based on plant morphology alone. The morphological pattern was not supported by the molecular data, suggesting that pink flowers evolved more than once, concurrently with other floral traits such as size and nectary length. It is also possible that this pattern is due to recent evolution or gene flow between the two color morphs. The possible ecological importance of these differences in floral traits (e.g., for pollination) warrants further study, as well as the extent to which these populations are reproductively isolated. Key Words: Amplified fragment length polymorphisms (AFLPs), endemic, floral traits, genetic variation, Lithophragma, Saxifragaceae. Floral morphology has been found to be important in reproductive success (e.g., Galen 1989; Herrera 1993; Guitian et al. 1997), pollinator specialization (Muchhala 2003) and isolation between plant species (e.g., Bradshaw Jr. et al. 1998; Ellis and Johnson 1999; Fulton and Hodges 1999). Floral traits may dictate reproductive success by mediating attractiveness to pollinators. In this way, individual pollinators may exert strong directional selection for particular floral syndromes (Campbell et al. 1997). In part because they often play a role in reproductive isolation, floral traits are also used to differentiate closely related species. Yet, are these floral traits meaning- ful in explaining the genetic relatedness among populations and species? Floral morphology alone can be phylogenetically deceptive due to conver- gent evolution and the gene flow and hybridiza- tion that can result from shared pollinators. The processes of diversification and speciation remain important problems in evolutionary bi- ology because species arise through many genetic mechanisms and their relative importance among different taxa is unresolved (Hewitt 2001). Di- versification in plants is particularly intriguing because they may speciate through hybridization and exhibit reticulate evolution. Incomplete re- productive isolation can limit diversification, but it also may maintain a greater range of floral morphologies, as intermediate morphotypes ‘Current address: Biosystematics, Wageningen Uni- versity and Research Centre, P.O. Box 9101, 6700 HB Wageningen, the Netherlands. would be maintained due to gene flow and hybridization. Hybridization can result in genetic variation, which provides the raw material for rapid adaptation and can play an important role in evolutionary diversification (Rieseberg 1997; Arnold et al. 1999; Rieseberg et al. 2000). The genus Lithophragma (Saxifragaceae) pro- vides an opportunity to evaluate how floral morphology and hybridization contribute to the diversification of plant taxa. The genus has ten named species that differ in floral morphology, hybrid history and geographic range (Taylor 1965). Three of the species within this genus, L. affine A. Gray, L. parviflorum (Hook) Torrey & A. Gray, and L. trifoliatum Eastw. form a tight clade and represent the extremes in geographic distributions within the genus (Taylor 1965; Nicholls and Bohm 1984; Soltis et al. 1992: Kuzoff et al. 1999). Hybridization has been suggested among these three species (Taylor 1965). In addition, results from a phylogeny based on internal transcribed spacer (ITS) sequences of ribosomal DNA showed these species as a para- phyletic group (Kuzoff et al. 1999). The authors thought the most compelling reason for this is that the putative species are not distinct lineages. In the literature, there 1s disagreement regarding their specific status: Taylor (1965) lists them as three separate species, while the Jepson Manual (Hickman 1996) lists two species, L. affine and L. parviflorum vars. parviflorum and trifoliatum. Although existing molecular studies have shown these three species to be genetically very similar, they differ in several morphological traits, including some (i.e., floral scent and color; see Table 1) that are unique 26 TABLE 1. MADRONO [Vol. 53 SUMMARY OF SPECIES DESCRIPTIONS. INFORMATION IN TABLE FROM THE JEPSON MANUAL (HICKMAN 1993) AND TAYLOR (1965). The Jepson Manual (Hickman 1993) lists L. trifoliatum as L. parviflorum var. trifoliatum and L. parviflorum as L. parviflorum var. parviflorum. Flower Floral Leaf Species color shape shape L. affine White Hypanthium Shallowly obconic, lobed inflated above L. parviflorum | Generally Hypanthium Deeply white, long- lobed some obconic, pink 2x longer than wide L. trifoliatum Pink Hypanthium Deeply long- lobed obconic, 3-4 longer than wide within the genus. Northern California is the only place in the range of this genus where plants with pink flowers occur. Along with these pink flowers are other unusual traits for this genus, such as much longer petals and hypanthia (see Table 1). There are also populations that have larger white flowers, resembling L. parviflorum, but that have leaf morphology more similar to L. affine. Populations with mixed traits (generally labeled L. parviflorum) in this region led Taylor (1965) to conclude that there was hybridization in this region. This species complex provides an opportunity to study two central questions regarding the diversification of plant taxa: 1) Do these popula- tions of putative species differ consistently in the floral traits that are used to distinguish the species? 2) Do patterns of floral morphology inform our understanding of phylogenetic rela- tionships among these species? Specifically, does the unique pink color correlate with other floral traits that differ consistently among species? If the pink flowers have evolved only once and are an important reproductive isolating mechanism among these species, we expect the populations with pink flowers to be more closely related to each other than to populations with white flowers and for there to be limited gene flow between populations of different flower color. MATERIALS AND METHODS Geographical Distribution of Lithophragma Lithophragma (Saxifragaceae) is an herbaceous perennial genus with a broad geographical distribution from southern California to southern Floral scent Ploidy (2n) (AWA le 28 Distribution Coastal mountains of CA, a few specimens in outer central Sierra Nevada foothills. Open well-drained grassy clearings or open bluffs in oak or coniferous-oak woodlands from sea level to 2150 m W. NA (BC, WA, ID, MT, OR, Wy, SD, CA, NV, UT, CO); Habitat extremely variable, seacoast bluffs to open gravelly prairielands, to subalpine regions up to 3050 m elevation Restricted to western slope of the Sierra Nevada, with center in Butte Co., CA. In igneous- derived scabland in oak- coniferous woodland from 60—600 m elevation none none 14, 21, 28.530 fragrant British Columbia and from the west coast of North America to South Dakota (Taylor 1965). The genus is thought to have originated within California (Taylor 1965), where most taxa and the most basal taxa occur (Taylor 1965; Soltis et al. 1992; Kuzoff et al. 1999). Molecular data indicate that L. affine, L. parviflorum, and L. trifoliatum are so similar that the limits of the species are uncertain with respect to both the chloroplast and nuclear markers that have been used (Soltis et al. 1992; Kuzoff et al. 1999). These three species may, therefore, not be distinct lineages. Nonetheless, these taxa differ in a variety of morphological traits and geograph- ical distribution, with more differentiation occur- ring among populations and species in California (near the purported center of the distribution) than in more northerly populations. Lithophragma affine is primarily restricted to the coastal mountains of California from Hum- boldt County to Santa Barbara County (Taylor 1965), but a few specimens have been found in the foothills of the central Sierra Nevada, in Tuolumne, Stanislaus, and Amador counties (CalFlora Occurrence Database; UC, Berkeley Jepson Herbarium). Taylor (1965) described L. affine as very polymorphic due to environmental variability and population isolation caused by the topography of the region. Lithophragma parviflorum is the most widely distributed species in the genus. It ranges from southern California to southern British Columbia and from the west coast of North America to South Dakota. Across its wide range, L. parvi- florum shows great morphological variation. 2006] Lithophragma trifoliatum has the narrowest distribution in the genus and is restricted to the western slope of the Sierra Nevada. Taylor (1965) considered this species closely related to L. parviflorum. His data indicated that this species represented a sterile derivative of L. parviflorum that was persisting through vegetative reproduc- tion in a small geographic area. (However, field- collected seeds have germinated in the greenhouse (Hufft unpublished data), but the extent of their viability across all populations is not known.) More recent phylogenies have shown that L. trifoliatum is part of the L. parviflorum/affine clade, but its exact relation to the other two species is not known (Soltis et al. 1992: Kuzoff et al. 1999). Sampled Populations The distributions of all three species, which overlap in northern California, provide an opportunity to study local phylogeographic patterns of this complex, which have been unresolved in the larger scale phylogeographic work that has been done on this genus. Sixteen populations, along with several roadside collec- tions, were located in spring 2001 and 2002 in Mendocino, Tehama, Butte, Plumas, and Lassen Counties (Fig. 1). Populations were chosen within the zone of overlap in northern California that represent the morphological diversity of the three species (Table 2). Populations were identified to species based on the floral characteristics that define them in Jepson (Hickman 1994) and verified by a local expert, Vern Oswald. Plants were haphazardly selected and were chosen at least two meters from the nearest plant used to ensure that they were distinct individuals and not growing from the same underground bulbils. Floral Morphology Although these species have similar floral structures, relative to other species in the genus (Kuzoff et al. 2001), variation in floral traits are used to distinguish among the three species (Taylor 1965; Hickman 1996). Nine morpholog- ical traits were measured on _ field-collected flowers: average petal length, average petal width, corolla gap, floral length, tip to nectary, short angle, average nectary depth, nectary length, and average diaganol (Fig. 2). The second flower from each plant was collected from all study sites (Table 2) and stored in 70% ethanol. Floral traits were measured using a microscope (Wild M8 microscope) fitted with an ocular micrometer. For these nine traits, 144 flowers were analyzed from 14 populations (Table 2) using Principal Components Analysis (PCA). Flower color was not included in the analysis because it is not a quantitative trait and only one HUFFT AND RICHARDSON: MORPHOLOGY AND GENETICS OF LITHOPHRAGMA oA color (white or pink) occurred within a popula- tion. Instead, flower color was used as a grouping variable in the various analyses to determine its usefulness in distinguishing among the species. We used discriminant analysis to determine if flower color, species, or molecular group (based on AFLP data, see below) better differentiated these individuals based on floral morphology (SAS 6.12 1996). Half of the samples in each group were randomly selected to create the discriminant function, with the other half used to test the model. The proportion of test samples classified correctly provided a quantitative mea- sure of the ability of each grouping variable to accurately distinguish among these individuals. Estimates of pairwise population morphological distances were calculated with discriminant anal- ysis (SYSTAT 10.2 2002) for comparison with estimates of genetic distance (see below). Pollinator Observations We performed pollinator observations at six sites in spring 2002 (1 L. affine, 1 white L. parviflorum, 1 pink L. parviflorum, and 3. L. trifoliatum). We chose plants haphazardly and observed all plants within a 1-m quadrat (number of plants per observation=1—13) for 30 min. We performed a total of 112 observation periods (15 L. affine, 61 L. parviflorum, and 35 L. trifolia- tum). We recorded the number of flowers within a quadrat, the number and identity of pollinator visitors and the number of flowers visited. Amplified Fragment Length Polymorphisms (AFLPs) Gene flow within and among populations and patterns of relatedness among individuals and populations was evaluated using genetic finger- printing (amplified fragment length polymor- phisms [AFLPs; Vos et al. 1995] that are pre- dominantly nuclear). DNA was extracted using the method from Doyle and Doyle (1987). Following standard protocols, AFLPs were analyzed (Applied Biosystems manuals 1997). A total of 158 individuals from 16 populations (Table 2) were scored for the presence of 216 markers from two ABI AFLP primers, CAT- ACT (blue) and CAG-AAG (green). Data were analyzed with an AMOVA using ARLEQUIN (Schneider et al. 2000), a Principle Coordinate Analysis (PCoA) using the R Package (Casgrain 2004) and UPGMA using PAUP 4.0b10 (Swaf- ford 2001) and visualized in TreeView (Page 2001). Pairwise population differences were calculated, a one-way AMOVA was performed to measure among population variation and a hierarchical AMOVA was performed to parti- tion the variance into species and floral color effects. The relationship between floral morphol- Og MADRONO : Vt fh nee “e YA Fic. 1. [Vol. 53 a P Tia ft ¢ a hast wash tai California field sites, as numbered in Table 2. White circles = L. affine, white triangles = white-flowered L. parviflorum, black triangles = pink-flowered L. parviflorum, and black stars = L. trifoliatum. ogy and genetic relatedness was first assessed by comparing the output of the PCA and PCoA. The quantitative floral morphology data and AFLP data were then statistically compared using a regression of the AFLP genetic distances and the floral morphology distances, to test for a positive relationship between genetic and floral distances from the discriminant analysis. A log transformation was performed on floral distances to normalize data prior to performing the linear regression. RESULTS Floral Morphology Results of the PCA suggest that populations with pink flowers (L. parviflorum and L. trifolia- tum) can be distinguished from populations with white flowers (L. parviflorum and L. affine) to some degree but the measured floral traits exist along a continuum (Fig. 2). Axis 1 and Axis 2 explained 46% and 23% of the variation, re- 2006] TABLE 2. HUFFT AND RICHARDSON: MORPHOLOGY AND GENETICS OF LITHOPHRAGMA 29 SPECIES, FLOWER CCOLOR AND NUMBER OF SAMPLES UUSED FOR THE MORPHOLOGICAL AND GENETIC ANALYSES FOR EACH POPULATION. * Not used in AMOVA analyses. Population Species Flower Color Flower Count AFLP count 1. Alder Affine White | 6 2. Big Oak Affine White 14 10 3. Hwy101 Affine White 6 E 4. Mendocino] Affine White 0 1] 5. Mendocino2 Affine White 5 10 6. Feather Falls Parviflorum White 10 14 7. Hwy70 Parviflorum White 19 10 8. Plasket Meadows Parviflorum White 0 2 ). Dye Creek Parviflorum Pink l 9 10. Hogsback Parviflorum Pink 60 25 11. Forbestown Trifoliatum Pink 0 3 12. Hog Lake Trifoliatum Pink l 12 13. Hwy 36 Trifoliatum Pink 0 3 14. MilsapBar Trifoliatum Pink 3 17 15. North Table MT Trifohatum Pink 10 14 16. Shingletown Trifoliatum Pink 0 11 Total 139 158 spectively (Table 3). Most populations did not fall out as tight clusters (data not shown). The white and pink flowered individuals separate out mainly along Axis 2. The traits that have the strongest influence on axis 2 are length of nectary (with an eigenvector value of 0.556), nectary depth (0.443) and corolla gap size (0.519). Flower A®,,@B pec @®E ‘Nectary Fic. 2. Diagram illustrating floral measurements. Corolla gap = distance between A and B, flower length = midpoint of A and B to E, tip to nectary = midpoint of A and B to the midpoint of C and D, short angle = B to C, average diagonal = A to C and B to D, average Boo depth = height of nectary, and nectary length = to. color was a very good grouping variable in the DFA, with high classification rates for both colors (Table 4). Although, LZ. parviflorum and L. affine also showed high classification rates, L. trifoliatum proved to be a very bad grouping variable, with the majority of ZL. ¢rifoliatum individuals being classified as L. parviflorum (Table 4). The molecular groups had a_ higher total misclassification of individuals (35%) than either color (13%) or species (26%). Pollinator Observations We recorded 128 insects visiting 324 flowers. These preliminary observations revealed that although the pink populations received more visits (1.15 pollinators/observation period vs. 0.56 for white populations), generalists (Bombylid flies and solitary bees) were visiting all of the plants, indicating the possibility of gene flow between the color morphs. Unlike previous studies of L. parviflorum (Thompson and Pellmyr 1992: Thompson 1999; Thompson and Cunning- ham 2002), the specialist Greya politella was found at only one site (6. Feather Falls). AFLPs The results of the PCoA are shown in Figure 4, with individuals labeled by species and flower color. Lithophragma trifoliatum is split into two groups along axis 2, and the majority of samples with pink flowers are clumped along the same half of axis 1. Additionally, the AFLP results can be seen in the UPGMA phenogram (Fig. 5). Although, populations mostly group together, there is very low resolution of the relationships among populations. Population 15 (North Table Mountain) appears to be the most derived. Although not strongly supported, Population 14 30 MADRONO TABLE 3. [Vol. 53 RESULTS OF PRINCIPAL COMPONENT ANALYSIS OF FLORAL MORPHOLOGY. Variance extracted on the first three components (a) and eigenvectors of floral characters (b). A. Component Eigenvalue l 4.181 2 2.055 3 0.818 B. Character Eigenvector | Average petal length —0.3814 Average petal width —().3831 Corrolla gap =(.1280 Floral length -().3623 Tip to nectary —0.4211 Short angle —0.4184 Average nectary depth (J), 1032 Nectary length —().1388 Average diagonal —0.4216 (Milsap Bar) is more closely related to the two geographically closest populations (6. Feather Falls and 7. Hwy 70) than to other L. trifoliatum populations. The one-way AMOVA results revealed signif- icant genetic differentiation among populations, with 35.9% of the variation partitioned among the populations. The pairwise genetic distances are given in Table 4. Of the 105 comparisons, 6 were not significant. A nested ANOVA of genetic distances grouped by comparisons within species versus Comparisons among species re- vealed that there was no difference between within versus among species comparisons (F- ratio = 0.686, df = 1, P = 0.409), meaning that overall distances between populations of different species were not different from distances of populations of the same species. However, there were significant differences in the comparisons of distances among species and the distances within species (F-ratio = 2.898, df = 4, P = 0.024), with pairs involving L. trifoliatum having the largest within species distances and those involving L. affine having the smallest, The L. trifoliatum-L, parviflorum group was the largest among species distance. However, dis- tances between populations of the same color were more similar than populations of different color (Pooled Variance t = 2.757, df = 151, P = 0.007), No relationship was found between the pairwise genetic and morphological distances (b = —0.022, P = 0.222, adjusted r> = 0.007). The power of this test to detect a positive slope as small as 0.05 was 0,77. DISCUSSION Currently defined species within this group (Hickman 1996; Taylor 1965) do not appear much differentiated by floral morphology. The % of Variance Cum.% of Var. 46.450 46.450 22.834 69.285 9.085 78.369 Eigenvector 2 Eigenvector 3 =O.1771 0.1858 ==O.1658 0.1583 Oa185 = (95339 1p 59 0.1068 —0.1961 = (,0721 —0.2448 —0.0660 0.4426 0.7313 0.5561 0.0728 0.2202 (0.3070 traits that seem to be most important in separating individuals in the PCA are corolla gap, nectary length and nectary depth, all traits that could be important in pollinator preference (Campbell et al. 1997; Fulton and Hodges 1999), In addition, floral color can be an important cue in pollinator discrimination (Wilson and Stine 1996) and also appears to distinguish two main groups of individuals when they are plotted onto the first two principal components. However, the three species are not segregated into discrete groups by the principal components. The pre- dominance of generalist pollinators in this region would suggest that other, possibly neutral mech- anisms, are maintaining floral variation, It has been hypothesized that diversity of ovary posi- tion in species of Lithophragma may be the result of modifications in one or a few genes (Kuzoff et al. 2001), This could also be the cause of the variation in other floral traits, like those mea- sured here. Using genetic markers to analyze these groups, the three species clump together indicating they are Closely related. The degree of overlap suggests these are not distinct species. Location 1s important (1.e., populations clump together in- dicating individuals within a population are closely related), but flower color also has some genetic component (i.e., clumping of pink and white flower color along Axis 1, Fig. 4), Larger genetic distances between L. trifoliatum popula- tions help explain the separation of this species into two groups in the PCoA. This could indicate more variation within populations or multiple lineages. This was also supported in the discrim- inant function analysis of morphological traits, where L. trifoliatum was not found to be a viable group. Flower color was not as good a grouping variable for the genetic relationships as it was for the floral morphology (see separation of pink 2006] HUFFT AND RICHARDSON: MORPHOLOGY AND GENETICS OF LITHOPHRAGMA 31 TABLE 4. DISCRIMINANT FUNCTION ANALYSIS OF FLORAL MORPHOLOGY GROUPED BY SPECIES, COLOR AND MOLECULAR GROUP. Molecular groups were based on AFLP results: Group 1 = 1.Alder, 3.Hwy101, 7. Hwy/70; Group 2 = 6.Feather Falls, 14.Milsap Bar: Group 3 = 2.Big Oak, 4.Mendocinol, 5.Mendocino2; Group 4 = 15.North Table Mountain; Group 5 = 9.Dye Creek, 10.Hogsback, 12.Hog Lake. The total-sample standardized canonical coefficients are shown, along with the percent classifications for the testing samples. * Total error= 26.05%. ° Total error=13.25%. © Total error = 35.38%. Total-Sample Standardized Canonical Coefficients Petal Corrola Flower Tip to Nectary Nectary Species: length Petal width gap length nectary Shortangle depth length Diagonal CANI 0.301 =0.120 —0.104 0.898 1.748 Sa99 0.173 =1,050 0.381 CAN2 =0:991 0.683 1.003 0.925 —5.547 Del —0.464 =G.027 =) 299 Color: CANI1 =00:01S 0.044 — 492 0.838 0.346 0.474 =0:196 =O 37 —().292 Molecular: CANI O337 0.240 =0.173 0.848 =O.177 0.765 0.001 =h2/1 0.030 CAN2 0.824 —0.996 0.653 0.687 0.657 —0.914 0.398 =G,132 0.046 CAN3 —00.790 0.371 —0.033 1.233 1.502 = b.225 =(0,391 0.864 =0915 CAN4 0.281 0.755 0.290 0,332 =0.085 =)039 —0.410 =U109 0.802 Percent Classified into Species: From Species: L. parviflorum L. trifoliatum L. affine Co error No. of samples L. parviflorum 84.09 13.64 22t [591 44 L. trifoliatum 57.14 28:57 14.29 71.43 7 L. affine al 25 0 68.75 31:25 16 Percent Classified into Color: From color: Pink White % error? No. of samples Pink UA ie, 8.11 8.11 57 White E935 80.65 L935 31 Percent Classified into Molecular Group: No. of From group: l 2 3 4 3 % error‘ samples ] 46.15 23.08 15.38 0 15.38 53.8) 13 2 33.55 33.33 33-53 0 0 66.67 6 5 21.43 7.14 64.29 0 7.14 35.71 14 4 0 0 0 0 100 100.00 5 5 0 0 6.67 3:33 90.00 10.00 30 TABLE 5. PAIRWISE POPULATION DIFFERENTIATION EXPRESSED BY @®gy7 (EXCOFFIER ET AL. 1992). Bold values are not significant. Population names refer to population number and species (Aff = L. affine, Par = L. parviflorum, Tn = L. trifoliatum). For L. parviflorum, flower color is also noted (P = pink, W = white). LAff 2Aff 4Aff SAff 6ParW 7ParW 8ParW 9ParP 10ParP 11Tri 12Tri 13Tn 14Tn 15Tn 2Aff 0.412 4Aff O591 - 0202 SAff O 316° {0,236.0 218 6ParW 0.545 0.429 0.263 0.404 7ParW 0.296. »0296.-0.198.°0:263 O.188 8ParW 0.314 0.365 0.279 0.337 0.508 0.105 9ParP 0.499 0.350 0.224 0.344 0.369 0.253 0.464 10ParP 0.397 =0.286 0.169 0.245 0.213 0.185 0.317 0.094 11Tri 0.463 0.359 0.125 0.327 0.062 0.062 0.271 0.335 0.147 PA al 0.601 0.443 0.351 0.417 0.486 0.334 0.600 0.354 0.213 0.552 13Tri 0.423) 0.323 0.128 0.268 0.409 0.158 0.257 0.185 0.036 0.338 0.326 14Tri 0.579 0.476 0.348 0.462 0.259 0.258 0.569 0.502 0.306 0.328 0.568 0.525 1STri 0.649 0.488 0.369 0.487 0.561 0.424 0.660 0.449 0.295 0.590 0.626 0.528 0.613 16Tri 0.524 0.299 0.237 0.314 0.408 0.301 0.514 0.241 0.178 0.440 0.363 0.271 0.504 0.456 ies) N Axis 2 FIG. 3. MADRONO Axis 1 Axis 2 XK X + | XK + _A A me + FIG. 4. pink; X L. trifoliatum, pink. Axis 1 [Vol. 53 Principal Component Analysis of floral morphology data labeled by species and flower color. © L. affine, white; A L. parviflorum, white; A L. parviflorum, pink; * L. trifoliatum, pink. Principal Coordinate Analysis of AFLP data. + L. affine, white; A L. parviflorum, white; A L. parviflorum, 2006] 10 FIG. 5. into two main groups in Fig. 4, and lack of association in Fig. 5). This could be because the pink morph has evolved more than once, it evolved recently or there is gene flow between the two color morphs. If pollinators were a strong HUFFT AND RICHARDSON: Plas262 MORPHOLOGY AND GENETICS OF LITHOPHRAGMA AIdA189 AldA180 AldA190 Alda 89 HogB3 Hg = Bych434 are eat Mn Aca cea iiake FF15S £5) FF139 as = FE149 [F143 =) Hy ———— [170868 Hess =H MB583 Mes8 a MBe6é FF134 me = FE{sd A HERA HIk412 HIK49 = — faa Hoghsé HogD218 == Hegbs68 —— fa Hogk 102 HFC — HS9P 138 Shin482 oping s3 shing8s Shin476 eae shind8? ai = oe = [Sg86° — Heger’? ee Mndasae Oanagoe : Heese —— Hn B83 MenB268 MenB271 enB269 Oak 503 a8 Oak5 if — Oak 504 50 MenB266 shin () Gake8? un” 7M450 NTM463 NTM456 NTM468 4S Tari MenB272 ” MenB5oe H70A45 Heese He OBe2 NiMa37 Population — Color Species Plasket w pa w aff w aff w aft w aff w aff aff af EVVVVVOUVUDUDUE € par par par aff : a tn w par Feather Falls w par Feather Falls w par Feather Falls w par F is Ww par w Ww w w suvsUEE 2% ar Mendocino1 Milsap Bar Milsap Bar VVVVUUEDVUUDUDD Milsap Bar Milsap Bar Milsap Bar s0uv0, Hwy36 Hogsback Mendocino1 Hogsback n Shingletown Shingletown Shi n Hogsback Hogsback Hogsback Hogsback Hogsback Hogsback Hogsback Hogsback Hogsback Hogsback Hogsback Hogsback Mendocinot Mendocino1 Desens MMMM. Di, Don ECD. ES ED 1D. Diem EET EVVVVVVVVVGOVV VUE VUVUVVUUVUUVUUEDEDE TS EVVVUVVVOVVUUVUUDUUDUE < eststteize o North Table MT North Table MT p North Table North Table M North Table MT A UPGMA phenogram of AFLP data. Bars to the right of figure indicate flower color (black = pink) and species (black = L. trifoliatum, gray = L. parviflorum). White flowers and L. affine are indicated by spaces between the bars. * mark bootstrap values greater than 70%. isolating mechanism we would expect less overlap among the species and between individuals of different flower color in the AFLP data, dicating less gene flow. The lack of a relationship found between the molecular and morphological in- 34 MADRONO data could be due in part to shared pollinators. However, it is also possible that very strong selection has resulted in the morphological di- vergence we see despite the molecular evidence of gene flow. The floral traits are thought to be genetically inherited, but a common garden experiment is necessary to separate the environ- mental component that could be responsible for some of the variation in the data (although no obvious habitat variation is known among the populations). The three taxa studied here differ in the size of their geographic ranges, which might influence the morphological variation observed, as you would expect species with larger ranges to have more morphological variation due to the in- creased environmental variation across their range. An additional aspect of this work was to identify genetic and ecological differences be- tween a narrow endemic and its broad ranging relatives. It is expected that rare plants have low phenotypic variability (Kruckeberg and Rabino- witz 1985), but this is not always the case (Guitian et al. 1997). Given the narrow distribu- tion of L. trifoliatum relative to its two sister species studied here, the expectation would be for it to show less variation in phenotypic traits than L. affine and L. parviflorum. However, all three species show similar amounts of variation for each individual trait measured (data not shown). It is possible that the morphological variation seen over the range of this clade is not due to differences among three species, but rather is just variation within one or two species. In agreement with previous work in this system, this research shows there is strong evidence that these are not three distinct lineages. The addition of the morphological data also supports previous mo- lecular work that L. trifoliatum may not be a true, distinct species. In order to better understand why these species show more variation in morphology than at neutral molecular markers, the relative impor- tance of selection and local adaptation must be determined. More pollinator observations, pref- erence trials and estimates of pollinator travel distance will help us better understand the role of biotic selection on these species. Sorting out the amount of gene flow and the mechanisms responsible for the maintenance of floral varia- tion will also aid our understanding of the roles of selection versus drift in creating this variation in floral morphology. Exploring these diversifi- cation mechanisms will give us insight into their role in creating and maintaining biodiversity on a broader scale. ACKNOWLEDGEMENTS The authors thank J. N. Thompson, I. M. Parker, G. Pogson, and G. Gilbert for support and advice for this project. Additional thanks to K. Dlugosch, C. Hays, W. [Vol. 53 Satterthwaite, P. Shahani D. Kao, J. Hunter and two anonymous reviewers for providing helpful feedback on this manuscript. A California Native Plant Society Helen Sharsmith grant to RAH helped fund this work. LITERATURE CITED APPLIED BIOSYSTEMS. 1997. AFLP plant mapping protocol. Perkin-Elmer Corporation, Foster City, CA. ARNOLD, M. L., M. R. BULGER, J. M. BURKE, A. L. HEMPEL, AND J. H. WILLIAMS. 1999. Natural hybridization: How low can you go and still be important? Ecology 80:371-381. BRADSHAW, JR., H. D., K. G. OTTO, B. E. FREWEN, J. K. MCKAY, AND D. W. SCHEMSKE. 1998. Quantitative trait loci affecting differences in floral morphology between two species of monkeyflower (Mimulus). Genetics 149:367—382. CALFLORA: INFORMATION ON CALIFORNIA PLANTS FOR EDUCATION, RESEARCH AND CONSERVATION. 2005. The CalFlora Database, Berkeley, California. Available at: http://www.calflora.org/. CAMPBELL, D. R., N. M. WASER, AND E. J. MELEN- DEZ-ACKERMAN. 1997. Analyzing pollinator-medi- ated selection in a plant hybrid zone: hummingbird visitation patterns on three spatial scales. American Naturalist 149:295—315. CASGRAIN, P. 2004. The R Package for multidimen- sional and spatial analysis. University of Montreal, Available at http://www.fas.umontreal.ca/biol/ casgrain/en/labo/R/. DoyYLE, J. J. AND J. L. DOYLE. 1987. A rapid DNA isolation procedure from small quantities of fresh leaf tissue. Phytochemical Bulletin 19:11—15. ELLis, A. G. AND S. D. JOHNSON. 1999. Do pollinators determine hybridization patterns in sympatric Satyrium (Orchidaceae) species? Plant Systematics and Evolution 219:137—150, EXCOFFIER, L., P. E. SMOUSE, AND J. M. QUATTRO. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes—appli- cation to human mitochondrial-DNA restriction data. Genetics 131:479-491. FULTON, M. AND S. A. HoDGEs. 1999. Floral isolation between Aquilegia formosa and Aquilegia pubes- cens. Proceedings of the Royal Society London B 266:2247-2252. GALEN, C. 1989. Measuring pollinator-mediated selec- tion on morphometric floral traits: Bumblebees and the alpine sky pilot Polemonium-Viscosum. Evolu- tion 43:882—890. GUITIAN, J., L. NAVARRO, P. GUITIAN, AND J. M. SANCHEZ. 1997. Variation in floral morphology and reproductive success in Petrocoptis grandiflora (Caryophyllaceae). Annales Botanici Fennici 34:35—40. HERRERA, C. M. 1993. Selection on complexity of corolla outline in a hawkmoth-pollinated violet. Evolutionary Trends in Plants 7:9—-13. Hewitt, G. M. 2001. Speciation, hybrid zones and phylogeography-or seeing genes in space and time. Molecular Ecology 10:537-549. HICKMAN, J. C. 1996. The Jepson manual: higher plants of California. University of California Press, Berkeley, California, USA. KRUCKEBERG, A. R. AND D. RABINOWITZ. 1985. Biological aspects of endemism in higher plants. 2006] Annual Review of Ecology and Systematics 16:447—480. KUZOFF, R. K., L. HUFFORD, AND D. E. SOLTIS. 2001. Structural homology and developmental transfor- mations associated with ovary diversification in Lithophragma (Saxifragaceae). American Journal of Botany 88:196—205. , D. E. SOLTIS, L. HUFFORD, AND P. S. SOLTIS. 1999. Phylogenetic relationships within Litho- phragma (Saxifragaceae): Hybridization, allopoly- ploidy, and ovary diversification. Systematic Bot- any 24:598-615. MUCHHALA, N. 2003. Exploring the boundary between pollination syndromes: Bats and hummingbirds as pollinators of Burmeistera cyclostigmata and B. tenuiflora (Campanulaceae). Oecologia 134:373- 380. NICHOLLS, K. W. AND B. A. BOHM. 1984. The flavonoids of Lithophragma (Saxifragaceae). Cana- dian Journal of Botany 62:1636—1639. PAGE, R. D. M. 2001. TreeView. Available at: http:// taxonomy.zoology.91la.ac.uk/rod/rod.html RIESEBERG, L. H. 1997. Hybrid origins of plant species. Annual Review of Ecology and Systematics 28:359-89. , Ss. J. E. BAIRD, AND K. A. GARDNER. 2000. Hybridization, introgression, and linkage evolu- tion. Plant Molecular Biology 42:205—224. SAS 6.12. 1996. SAS Institute Inc., Cary, NC. SCHNEIDER, S., D. ROESSLI, AND L. EXCOFFIER. 2000. Arlequin: a software for population genetics data HUFFT AND RICHARDSON: MORPHOLOGY AND GENETICS OF LITHOPHRAGMA 3D analysis. Genetics and Biometry Laboratory, Uni- versity of Geneva, Switzerland. SOLTIS, D. E., P. S. SOLTIS, J. N. THOMPSON, AND O. PELLMYR. 1992. Chloroplast DNA variation in Lithophragma (Saxifragaceae). Systematic Botany 17:607—-619. SWAFFORD, D. L. 2001. PAUP 4.0b10. Sinauer Associates, Inc., Sunderland, MA. SYSTAT V. 10.2. 2002. Systat Software, Inc., Richmond, CA. Available at: http://www.systat.com. TAYLOR, R. L. 1965. The genus Lithophragma (Saxi- fragaceae) University of California Publications in Botany, Berkeley, CA. THOMPSON, J. N. 1999. The evolution of species interactions. Science 284:2116—2118. AND B. M. CUNNINGHAM. 2002. Geographic structure and dynamics of coevolutionary selection. Nature 417:735—738. AND O. PELLMYR. 1992. Mutualism with pollinating seed parasites amid co-pollinators: Constraints on specialization. Ecology 73:1780— 1791, Vos, P., R. HOGERS, M. BLEEKER, M. REIANS, D. L. T. VAN, M. HORNES, A. FRIJTERS, J. PoT, J. PELE- MAN, M. KUIPER, AND M. ZABEAU. 1995. AFLP: a new technique for DNA fingerprinting. Nucleic Acids Research 23:4407-4414. WILSON, P. AND M. STINE. 1996. Floral constancy in bumble bees: handling efficiency or perceptual conditioning? Oecologia 106:493-499. MADRONO, Vol. 53, No. 1, pp. 36-45, 2006 ADDITIONS TO THE VASCULAR FLORA OF WASHINGTON FROM A BIODIVERSITY STUDY ON THE HANFORD NUCLEAR RESERVATION KATHRYN A. BECK Beck Botanical Services, 1708 McKenzie Ave., Bellingham, WA 98225 calypso@openaccess.org FLORENCE E. CAPLOW Washington Natural Heritage Program, Washington Department of Natural Resources, P.O. Box 47014, Olympia, WA 98504-7014 ABSTRACT During a three-year botanical inventory of the Hanford Nuclear Reservation in south-central Washington, we located three previously undescribed entities (Astragalus conjunctus S. Watson var. rickardti Welsh, K.A. Beck, and Caplow, Eriogonum codium Reveal, Caplow, and K.A. Beck, and Physaria douglasii (S. Watson) O’Kane and Al-Shehbaz subsp. tuplashensis (Rollins, K.A. Beck, and Caplow) O’Kane and Al-Shehbaz, four species new to Washington (Cistanthe rosea (S.Watson) Hershk., Gilia leptomeria A. Gray, Loeflingia squarrosa Nutt., and Myosurus clavicaulis M.E. Peck), and one previously described species that had generally gone unnoticed by the American botanical community (Festuca washingtonica E.B. Alexeev). The botanical inventory was a part of the Hanford Biodiversity Project, which was funded by the Department of Energy and administered by The Nature Conservancy of Washington. Key Words: Hanford Nuclear Reservation, Astragalus conjunctus var. rickardii, Eriogonum codium, Physaria douglasii subsp. tuplashensis, Cistanthe rosea, Gilia leptomeria, Loeflingia squarrosa, Myosurus clavicaulis, Festuca washingtonica. The 1450 km* Hanford Nuclear Reservation (Hanford Site) is located in Benton, Grant, and Franklin Counties in south-central Washington (Fig. 1). It was acquired in 1943 by the USS. Government for the production of weapons- grade plutonium, and it has been administered by the U.S. Department of Energy (DOE), or its predecessors. Portions of the Hanford Site have also been administered by the U.S. Fish and Wildlife Service (FWS) and the Washington State Department of Wildlife. The majority of the Hanford Site has been closed to the public and to grazing and agriculture since 1943. In the last two decades, the mission of the Hanford Site has changed to nuclear waste clean- up, environmental restoration, research and de- velopment. In 1992, the DOE and The Nature Conservancy of Washington (TNC) entered into a Memorandum of Understanding that called for a cooperative and coordinated inventory of plants, animals, and ecologically significant areas of the Hanford Site. The Hanford Biodiversity Project began in 1994 and continued through 1998 (Soll et al. 1999). The authors conducted the botanical portion of the Biodiversity Project from 1994 through 1997. A description of previous botanical work and a checklist of vascular plant collections: from the Hanford Site 1s in Sackschewsky and Downs (2001). In 1999, the Hanford Reach of the Columbia River was designated a Wild and Scenic River. In 2000, President Clinton declared 78,780 ha of the Hanford Site the ““Hanford Reach National Monument” and placed administration of the National Monument under the jurisdiction of the FWS. The new Monument includes the Hanford Reach of the Columbia River, the lands north of the Columbia River, and the 31,080 ha Fitzner- Eberhardt Arid Lands Ecology Reserve (ALE) (Fig. 1). The goal of this paper is to summarize the botanical portion of the Biodiversity Project focusing on the most significant findings. The Hanford Site is a unique area from both a bio- geographical and historical perspective. It sup- ports unusual plant assemblages, many rare plant species, and a high degree of endemism, due to its underlying geology, landscape setting, topogra- phy, and climate. In addition, it is the largest area in the inland Northwest to be intentionally closed to grazing, agriculture, and residential develop- ment over a period during which human activity drastically altered much of the inland Northwest. STUDY AREA The Hanford Site is located within the Columbia Basin Ecoregion, an area that histor- ically included over 6 million ha of steppe and shrub-steppe vegetation across most of central and southeastern Washington State, as well as portions of north-central Oregon (Franklin and Dyrness 1973). The Columbia Basin Ecoregion consists primarily of shrubs, perennial bunch- 2006] Washington State Saddle Mountains 4 Map Legend Hanford Reach National Monument (USFWS) "9 mo rip Fi : 77-44 JS. Department of Energy FP ra] 12 Kilometers. EG. 4. topographic features. grasses and a variety of forbs (Franklin and Dyrness 1973). It has undergone substantial loss and degradation in the post-European era due to intensive livestock grazing, introduction of in- vasive non-native plants, the advent of dryland wheat farming and irrigated agriculture, and altered fire regimes. Because the Hanford Site has been closed to all agricultural activities, including grazing, for nearly sixty years, its shrub-steppe plant ecosystem has been preserved in a condition and to an extent, that exists nowhere else (Soll et al. 1999). The Hanford Site BECK AND CAPLOW: ADDITIONS TO WASHINGTON FLORA FROM THE HANFORD SITE — 37 Hanford Site (Benton, Grant, and Franklin counties, Washington): management areas and major supports many areas of high quality plant communities. The plant community portion of the Biodiversity Project identified 48 high quality occurrences of 17 terrestrial plant community types, and 6 high quality riparian wetland communities (Soll et al. 1999). The Hanford Site is within the Central Basin climatological region, which is the hottest and driest climatological region in Washington. Av- erage annual precipitation ranges from 35 cm at the summit of Rattlesnake Mountain to 16 cm near the Columbia River (Downs et al. 1993). 38 MADRONO The Hanford Site is topographically variable (Fig. 1). Elevations within the Hanford Site range from 110 m along the Columbia River to 1100 m at the crest of Rattlesnake Mountain. Adjacent to the east side of the river are the steep bluffs of the Ringold Formation (the White Bluffs), which rise in places to over 185m above the river. A number of long basalt anticlinal ridges traverse Hanford, including: Rattlesnake Mountain and the Rattlesnake Hills, Yakima Ridge, Umtanum Ridge, Saddle Mountains, and Gable Mountain. The only free-flowing section of the Columbia River in the U.S., known as the Hanford Reach, flows for 76 km from the northwest to the southeast through the northern portion of Hanford. No perennial creeks drain to the Columbia River from the Hanford Site. Large dune fields occur on both sides of the Columbia River within the Hanford Site. The Hanford Site has a number of unique habitats and substrates, which include springs, sand dunes, vernal pools, riverine wet- lands, caliche soils, basalt ridgetops, basalt-de- rived sand dunes, and alkaline areas. METHODS The authors made over 15 visits to the Hanford Site between 1994 and 1997 to survey for vascular plant taxa considered rare in Washington at the time (Washington Natural Heritage Program 1994). We spent a total of 285 days in the field between late March and early September. We surveyed more than 19,000 ha. Field inventory methodology was based on habitat and flowering times of rare plant taxa potentially present in the area. The list of plants potentially present in the area was acquired from the Washington Natural Heritage Program, and the flowering times and habitat were obtained from the five-volume Vascular Plants of the Pacific Northwest (Hitch- cock et al. 1955-1969). We focused inventories in areas with high quality vegetation associations, high diversity, and unusual substrates. Search intensity varied, depending on the quality of the habitat and the likelihood that the particular habitat could support rare plant populations. The lightest search intensity was a walked transect through the area, and the most intensive search intensity was a series of tightly spaced transects in which the entire area was visually examined, often several times during the growing season. A list of all plants was compiled, and specimens were deposited primarily at the University of Washington Herbarium (WTU). In some cases, taxonomic experts requested specimens, or 1so- types were collected for other herbaria, and so specimens were also deposited at NY, US, WS, BRY, CAS, COLO, GH, K, MARY, MO, NY, RM, RSA, TEX, UC, and OSU. Nomenclature generally follows relevant Flora of North Amer- [Vol. 53 ica treatments and the Integrated Taxonomic Information System (ITIS) database (Integrated Taxonomic Information System 2004). Existing specimens at the University of Washington Herbarium (WTU) and the Washington State Marion Ownbey Herbarium (WS) were used for comparison. RESULTS We identified more than 400 species of vascular plants and numerous populations of 29 rare plants considered rare in Washington (Soll et al. 1999) during our three year inventory of the Hanford Site, of which three were previously undescribed, four were new records for the state, and one had long been ignored. These eight are the focus of this paper. The three previously undescribed entities are Astragalus conjunctus S. Watson var. rickardii S.L. Welsh, K.A. Beck and Caplow, Eriogonum codium Reveal, Caplow, and K.A. Beck, and Physaria douglasii (S. Watson) O’Kane and Al-Shehbaz subsp. tuplashensis (Rollins, K.A. Beck and Caplow) O’Kane and Al-Shehbaz. The three species new to Washing- ton are Cistanthe rosea (S. Watson) Hershk.., Gilia leptomeria A. Gray, Loeflingia squarrosa Nutt., and Myosurus clavicaulis M.E. Peck). The previously described species that had generally gone unnoticed by the American botanical community is Festuca washingtonica E.B. Alex- eev). Each entity is discussed below, and the larger findings of the study are also summarized. Previously Undescribed Entities Astragalus conjunctus S. Watson var. rickardii S.L. Welsh, K.A. Beck & Caplow (Fabaceae) The section Conjuncti Barneby in Astragalus L. is well represented in south-central Washington and north-central Oregon. Astragalus conjunctus S. Watson, A. leibergii M.E. Jones, A. reventi- formis (Rydb.) Barneby, and A. hoodianus Howell all occur in this region (Barneby 1964). Although clearly belonging to the section Conjuncti, a num- ber of Astragalus collections from Rattlesnake Mountain on the Hanford Site and from the nearby Horse Heaven Hills could not be satis- factorily assigned to any of the above species. We subsequently described a new variety, A. con- junctus var. rickardii (Welsh et al. 1997). In addition to Rattlesnake Mountain and Horse Heaven Hills, the variety is also known from northern Oregon. The var. rickardii has erect, sessile, strigulose pods, campanulate to short-cylindric calyces, and a banner reflexed at about 45°. This combination of characteristics is not found in other members of the section (Welsh et al. 1997). Astragalus conjunctus var. rickardii was found between 450 and 1070 m elevation in deep soil on 2006] the upper portions of mostly north-facing slopes. It was primarily found in bunchgrass communi- ties that are typically not currently grazed. Common associated species were Artemisia tri- dentata Nutt., Pseudoroegneria spicata (Pursh.) A. Love, Poa secunda J. Presl., Phlox longifolia Nutt., Balsamorhiza careyana A. Gray, and Crepis atribarba A. Heller. The variety is named in honor of Dr. William Rickard, a shrub-steppe ecologist and one of those responsible for the establishment of the Arid Lands Ecology Reserve. Astragalus con- junctus var. rickardii is currently included on the Watch list in Washington (Washington Natural Heritage Program 2005). Eriogonum codium Reveal, Caplow & K.A. Beck (Polygonaceae) During our 1995 botanical inventory of the Umtanum Ridge area of the Hanford Site we found a population of plants in the genus Eriogonum that did not resemble any known taxon. We subsequently published a new spe- cies, Eriogonum codium (Reveal et al. 1997). Eriogonum codium, or Umtanum desert buck- wheat, is included in the subgenus Eucycla Nutt. The combination of cymose-umbellate inflores- cences and sparsely tomentose flowers and achenes set E. codium apart from all other Eucycla taxa. Eriognum codium forms low, shrubby mats up to 1 m in diameter, and bears yellow flowers in cymose-umbellate inflorescences atop short scapes. A single plant may have as many as 300 inflorescences. Inflorescences produce flowers from June through September. Our growth ring counts on several dead individual plants suggest potential life spans of greater than 100 years (Dunwiddie et al. 2001). Eriogonum codium is known from one population that occurs in a discontinuous band 2.5km long by 30m wide. The population census in 1997 found 5200 adult plants in five small subpopulations on flat or gently sloping substrates above the steep north and northwest- facing slopes of Umtanum Ridge, at an elevation of approximately 375m. A census of the population in 2005 found only 4418 adult plants in the population (a 15% decline in 8 years). The population was restricted to the exposed, barren basaltic flow top of the Lolo Flow, one of the many Columbia River basalt flows (Goff 1981). It was found with Grayia spinosa (Hook.) Mogq., Artemisia tridentata Nutt., Salvia dorrii (Kellogg) Abrams, Krascheninniko- via lanata (Pursh) A.D.J. Meeuse & Smit, Hesperostipa comata (Trin. & Rupr.) Barkworth, Poa secunda J. Presl., Pseudoroegneria spicata (Pursh.) A. Loéve, Astragalus caricinus (M.E. Jones) Barneby, Bromus tectorum L., Mentzelia albicaulis (Douglas ex Hook.) Douglas ex. Torr. BECK AND CAPLOW: ADDITIONS TO WASHINGTON FLORA FROM THE HANFORD SITE = 39 & A. Gray, and Cryptantha pterocarya (Torr.) Greene. In 1996, a wildfire killed 20% of the adult plants in the population. There were no signs of either fire tolerance or resprouting after the fire (Dunwiddie et al. 2001). A demographic moni- toring study begun in 1997 has revealed aspects of the biology of this species that may affect its long-term viability. Despite abundant germina- tion, only one Eriogonum codium seedling has survived to flowering since the commencement of studies in 1997. Eriogonum codium is considered Endangered in Washington (Washington Natural Heritage Program 2005), and a Candidate for federal listing under the Endangered Species Act (U.S. Fish and Wildlife Service 1999). Physaria douglasii (S. Watson) O’Kane and AI- Shehbaz subsp. tuplashensis (Rollins, K.A. Beck and Caplow) O’Kane and AlI-Shehbaz (Brassicaceae) In 1883, T.S. Brandegee and F. Tweedie, who were employed as botanists for the Northern Transcontinental Survey (Rose 1904), collected a Lesquerella trom the White Bluffs of the Columbia River. This fragmentary collection, dated July 1883, caused difficulties in a number of early treatments of Lesquerella (Gray and Watson 1895; Piper 1906), and was attributed to L. douglasii Wats. (Payson 1922; Rollins and Shaw 1973; Rollins 1993). L. douglasii is generally restricted to the valleys of the Kootenay and Columbia Rivers from northern Oregon to southern British Columbia (Rollins and Shaw 1973). In July of 1994, we collected flowering material from a Lesquere/la on the White Bluffs that appeared to have both significant morpho- logical and phenological differences from L. douglasii. These plants were subsequently de- scribed as a new species: Lesquerella tuplashensis (Rollins et al. 1996). A recent treatment (Al Shehbaz and O’Kane 2002) has united Lesquerella with Physaria and united L. douglasii and L. tuplashensis under the single species name Physaria douglasii (S. Wat- son) O’Kane & Al-Shehbaz. Lesquerella tupla- shensis was retained as a subspecies of P. douglasii and became P. douglasii subsp. tupla- shensis (Rollins, K.A. Beck & Caplow) O’Kane & Al-Shehbaz, but not on the basis of new research or study of the two taxa. We are conducting morphometric and common garden experiments to further clarify the rank of this entity. At this time, ITIS still lists Lesquerella tuplashensis as the accepted name (ITIS 2005). Physaria douglasii subsp. tuplashensis differs from P. douglasii subsp. douglasii in the following ways: the trichomes of the silique valves of P. douglasii subsp. tuplashensis are stipitate when viewed under a 10 lens, while those of P. douglasii subsp. doug/asii are generally sessile and AO MADRONO appressed to the silique valve surface. The cauline leaves and the basal leaves of P. douglasii subsp. tuplashensis are broader and more imbricate than those of P. douglasii subsp. douglasii (Rollins et al. 1996). In addition, P. douglasii subsp. tuplashensis blooms in late May and early June and again in favorable years in late July and early August, while P. douglasii subsp. douglasii on riverine cobble on the Hanford Site blooms from late April to early May. We have never observed or collected P. douglasii subsp. douglasii in flower in late summer, nor have we seen any herbarium specimens of late summer flowering P. douglasii subsp. douglasii. Physaria. douglasii subsp. tuplashensis was restricted to a narrow 17 km band on the upper portion of the White Bluffs of the Columbia River. The White Bluffs are composed of lacustrine sediments of the Ringold Formation (Newcomb 1958), capped by a cemented, highly alkaline (pH 8.4+) calcium carbonate paleosol (a ““caliche” soil). Most of the population occurred on this caliche paleosol. Common associated plant species were Grayia spinosa, Artemisia tridentata, Achnatherum contractum (B.L. John- son) Barkworth, Bromus tectorum, Eriogonum microthecum Nutt. var. laxiflorum Hook., and Cryptantha spiculifera (Piper) Payson. Physaria douglasii subsp. tuplashensis 1s Threatened in Washington (Washington Natural Heritage Program 2005) and a Candidate for listing under the Endangered Species Act (U.S. Fish and Wildlife Service 1999). Both agencies list the entity as Lesquerella tuplashensis. ‘““Tuplash” is the traditional Wanapum Indian name for the White Bluffs. The Wanapum tribe occupied the area of the White Bluffs until the first decades of the 20th century. They still actively work to protect the many native cultural sites on the Hanford Site. Plants with Substantial Range Extensions Cistanthe rosea (S. Watson) Hershkovitz (Portu- lacaceae) Cistanthe rosea is a small annual previously known from central Oregon to central California east of the Sierra Nevada, east to Nevada, and in the Pacific Northwest in Butte County, Idaho (Hitchcock et al. 1955-1969). It generally occurs in sagebrush desert to arid montane forest. The Hanford Site population is the first report for this species in Washington. It represents a 340 km extension of its previous northern limit in Harney County, Oregon. In 1997, we located a total of 12 small clusters of Cistanthe rosea north of Gable Mountain in the central portion of the Hanford Site, in an area approximately 1.5 Xx 1.2 kilometers. With an estimated total of 150 plants, the population was not large. The plants were growing in flat or [Vol. 53 gently sloping areas on dark _ basalt-derived sand substrate within low swales in relatively dense big sagebrush. Cistanthe rosea grew with Artemisia tridentata, Bromus tectorum, Descurai- nea pinnata (Walter) Britt., Loeflingia squarrosa Nutt., and Mimulus suksdorfii A. Gray. Cistanthe rosea 18 currently included on the Threatened list in Washington (Washington Natural Heritage Program 2005). It occured in an area that is not formally protected from Hanford activities. Gilia leptomeria A. Gray (Polemoniaceae) Gilia leptomeria is known from open, sandy or rocky areas in the Great Basin and Rocky Mountain states of the North American west (Kartesz 2003). In Washington, an accurate identification of G. /leptomeria was difficult because Hitchcock and Cronquist (1973) did not recognize that there were two expressions within the concept of G. Jleptomeria. Day (1993) segregated G. /ottiae, a more frequently encountered gilia of the sandy shrub-steppe in eastern Washington, from G. leptomeria, which is primarily known from southern Oregon south through California and east to Utah. The Jepson Manual (Hickman 1993) contains a key that differentiates G. lottiae from G. leptomeria. The Hanford Site populations are the first reported for G. /eptomeria in Washington, and our collections were confirmed by Alva Day. It represents a 340 km extension of its previous northern limit in Malheur County, Oregon. We found Gilia leptomeria in a number of localities on the Hanford site, including the White Bluffs, Umtanum Ridge, north of Gable Moun- tain, and the gravelly bluffs north of the Columbia River in the vicinity of Vernita Bridge. The Hanford Site populations were small, at least in the years we saw them. When we surveyed them in 1995 and 1997, most of the eight populations located had 100 plants or less, for an estimated total of less than 1000 plants. We observed G. leptomeria growing in a variety of habitats and substrates, including basalt dunes, caliche soil, gravelly slopes, and shrub-steppe. Commonly observed associates of G. /eptomeria included Artemisia tridentata, Poa secunda, Bro- mus. tectorum L., Astragalus caricinus, Salvia dorrii, and a number of other annual species. It was found growing with several other plants considered rare in Washington. Currently, G. leptomeria is included on the Threatened list for Washington (Washington Natural Heritage Pro- gram 2005). Some of the populations occurred in areas that are not formally protected from Hanford activities. Loeflingia squarrosa Nutt. (Caryophyllaceae) This is the first report of Loeflingia squarrosa for Washington. It is generally found at less than 2006] 1200 m in California and northern Baja Califor- nia, and from southeastern Oregon (Hickman 1993). Our collections represent a 420 km exten- sion of its previous northern limit near Malheur Lake and Frenchglen in Harney County, Oregon (Peck 1961). We found Loeflingia squarrosa in a number of localities on the Hanford site. The eight Hanford Site populations were relatively large in the years we saw them. When we surveyed them in 1995 and 1997, we estimated a total of at least 8000 plants. At Hanford, L. squarrosa typically grew in flat or gently sloping areas on dark basalt derived sand substrate within low swales in relatively dense big sagebrush. One population was grow- ing in a vernal pool on Umtanum Ridge. Commonly observed associates of L. squarrosa included Artemisia tridentata, Poa secunda, Bro- mus tectorum, Cryptantha circumscissa (Hook. & Arn.) I.M. Johnst., Mimulus suksdorfii, and a variety of other annual species. Currently, L. squarrosa 1s included on the Threatened list for Washington (Washington Natural Heritage Pro- gram 2005). Most populations of L. squarrosa occurred in areas that are not formally protected from Hanford activities. Myosurus clavicaulis Peck (Ranunculaceae) In 1997, we found a population of unusual Myosurus plants in a vernal pool on the east end of Umtanum Ridge on the Hanford Site. Plants from this site resemble Myosurus plants of certain vernal pool populations in coastal southern California, northern Baja California, and Harney County, Oregon. Plants from these populations share most of the morphological characteristics of Myosurus minimus L., but have short, curved scapes, so that the heads of the achenes are immersed in the leaves (Whittemore 1997). Since 1997, M. clavicaulis has been found at several additional vernal pool sites in eastern Washing- ton by other workers (Washington Natural Heritage Program database). Stone (1959) suggested that this form of Myosurus resulted from past hybridization be- tween the two species Myosurus minimus L. and M. sessilis S. Watson. during a time when ™M. sessilis had a wider range than it does now. Currently, M. sessilis is not known to occur north of Umatilla Co., Oregon (Hitchcock and Cron- quist 1973; Whittemore 1997). At the coastal southern California, northern Baja California, Oregon, and Washington sites, this expression occurs as a self-sustaining entity, independent of M. sessilis and often independent of M. minimus. The taxonomic status of these plants is obscured by the presence of morphologically identical plants in the Central Valley of California. The Central Valley plants are non-persistent products of on-going hybridization between M. minimus and M. sessilis (Whittemore 1997). When he BECK AND CAPLOW: ADDITIONS TO WASHINGTON FLORA FROM THE HANFORD SITE = 41 examined the Hanford Site collections in 1997, Whittemore suggested that we call them Myo- surus Clavicaulis Peck to distinguish them from typical M. minimus and plants from M. minimus x sessilis hybrid swarms. The Washington populations represent a more than 375 km extension of its previous northern limit in Harney County, OR. When surveyed in 1997, there were estimated to be over 1000 plants in the Hanford Site population. The population was found on the south and southeast-facing sides of the farthest west vernal pool (20 * 20 m) in a set of vernal pools on Umtanum Ridge. Plants grew on the drying edges of the pool, with Veronica peregrina L. ssp. xdlapensis (Kunth) Pennell, Camissonia andina (Nutt.) Raven, Myo- surus apetalus Gay, Epilobium minutum Lindl. ex Lehm., Artemisia rigida (Nutt.) A. Gray, mosses and lichens. Myosurus clavicaulis on the Hanford Site is self-sustaining and does not occur within close proximity of either M. sessilis or M. minimus. Currently, M. clavicaulis is included on the Sensitive list for Washington (Washington Natural Heritage Program 2005). Previously Unrecognized Species Festuca washingtonica Alexeev In 1995, we collected an unusual Festuca high on the north-facing slopes of Rattlesnake Moun- tain in the Arid Lands Ecology Reserve (ALE). We sent this collection to B. Wilson at Oregon State University, who identified it as Festuca washingtonica Alexeev. The history of this taxon is as follows: in 1960, J. G. Smith collected what he called F. rubra L. in Peavine Canyon, Chelan County, Washington. A portion of the collection was eventually sent to Leningrad. From that lone specimen, E.B. Alexeev (1982) described F. washingtonica, with the note (translated from Russian) “in habit and anatomical structure of the leaf blades this species resembles F. rubra. However, within the limits of the very poly- morphic latter species, we do not know of a single taxon with leaf blades that are externally scabrous as in F. washingtonica.” (Alexeev 1982). Alexeev’s new fescue species was largely ignored until our collections of it in 1995. Based on our collections from the Hanford Site and from other locations, B. Wilson wrote an expanded description of F. washingtonica, (Wil- son 1999). Festuca washingtonica is a relatively large, bright to deep green, cespitose fescue with flat leaves | to 6mm wide, lemmas scabrous or pubescent, and ovary apex typically pubescent. According to Wilson (1999), “Festuca washingto- nica differs from F. viridula in having closed sheaths, extravaginal shoots, and abaxial to adaxial sclerenchyma strands in the leaves, illustrated by Alexeev (1982). Festuca washing- 42 MADRONO tonica may belong in the subgenus Festuca, although analysis of seed proteins suggests problems with that classification (Aiken et al. 1998). Festuca washingtonica appears to be endemic to seasonally moist habitats in deep soil of lightly grazed or ungrazed shrub-steppe communities east of the Cascade Range in Washington, including the ALE Reserve on the Hanford Site. On the ALE Reserve, plants occurred just below the top of Rattlesnake Mountain in rocky silt loam at an elevation of 1100 m. This was within an Artemisia tripartitalFestuca idahoensis plant association with Lupinus arbustus Douglas ex Lindl. ssp. calcaratus (Kellogg) D. Dunn, Melica bulbosa Geyer, Poa cusickii Vasey, P. secunda J. Presl., and Senecio integerrimus Nutt. Since 1995, Festuca washingtonica has been collected from other counties in eastern Wash- ington, including Yakima, Kittitas, Chelan and Okanogan Counties, suggesting that the taxon may have been overlooked and/or mis-identified by past researchers. Most sites where it grows are protected from livestock, either administratively or topographically. To our knowledge, it has not been collected outside of Washington. Other Findings The Hanford Site supports approximately 725 different kinds of vascular plants (Sackschewsky and Downs 2001). Of these, 29 (4%) are listed by the state of Washington as rare (Washington Natural Heritage Program 2005). During our three-year botanical inventory of the Hanford Site, we identified more than 500 species of plants and 112 populations of 26 state-listed rare plants (Soll et al. 1999). Two of these rare plants are endemic to the Hanford Site: Physaria douglasii subsp. tuplashensis and Eriogonum codium. Four are narrow regional endemics that also occur outside the boundaries of the Hanford Site: Astragalus columbianus Barneby, A. conjunctus var. rickardii, Lomatium tuberosum Hoover, and Penstemon eriantherus Pursh var. whitedii (Piper) A.Nelson. Four are more widespread regional endemics that are also known from elsewhere in the Columbia Basin: Camissonia pygmaea (Doug- las ex Lehm.) Raven, Cryptantha leucophaea (Douglas ex Lehm.) Payson, Erigeron piperianus Cronquist, and Rorippa columbiae (Suksd. ex B.L.Rob.) Suksd. ex Howell. There are also a number of regional endemics known from the Hanford Site that are not considered rare in Washington. Eighteen species found by the authors on the Hanford Site are widely distributed in North America but are disjunct or peripheral in Washington. These are considered rare and are tracked by the Washington Natural Heritage Program: Ammannia robusta Heer & Regel, [Vol. 53 Anagallis minima (L.) Krause, Astragalus geyeri A. Gray, Camissonia minor (A. Nelson) Raven, Castilleja minor (a. Gray) Gray ssp. minor, C. spiculifera (Piper) Payson, Cistanthe rosea, Cryp- tantha scoparia A. Nelson, Cuscuta denticulata Engelm. var. denticulata, Cyperus bipartitus Torr., Eatonella nivea (D.C. Eaton) A. Gray, Gilia leptomeria, Hypericum majus (Gray) Britt., Lipocarpha aristulata (Covill) G.C. Tucker, Loe- flingia squarrosa, Mimulus suksdorfii Gray, Myo- surus Clavicaulis Peck, Oenothera caespitosa Nutt., and Rotala ramosior (L.) Koehne. Festuca washingtonica is abundant enough state-wide that it is not tracked by the Washington Natural Heritage Program. DISCUSSION AND CONCLUSION The findings of the botanical portion of the Hanford Biodiversity Project have reinforced the importance of the Hanford Site for conservation of shrub-steppe vascular plants. The very high plant species richness of the Hanford Site is likely the result of a combination of geographic and historical factors. The Columbia River corridor, which bisects the Hanford Site, is known to support a large number of regional and narrow endemics (Hitchcock and Cronquist 1973). The Hanford Reach, the only undammed portion of the Colum- bia River, supports emergent wetland communities that include populations of several rare species that were once widely collected but are now uncommon elsewhere on the river (Soll et al. 1999). In addition, the unique climate of the Hanford Site supports a number of species that are more common in desert areas of the Great Basin to the south and east (Kartesz 2003). The diversity of unique substrates and habitats on the Site provide habitat for a large number of state-listed rare plants and endemics, including those reported above. Much of the biodiversity of the Hanford Site might have been lost without its unique history. If the Hanford Site had not been acquired by the U.S. Government in 1943, it would most certainly resemble the rest of the lower Columbia Basin where population growth, large-scale irri- gation projects, livestock grazing, and noxious weeds have significantly reduced and/or degraded available habitat for many plants. The relatively large Hanford Site, taken with the similar sized Department of Defense Yakima Training Center to the west, constitute the largest remaining blocks of relatively undisturbed shrub-steppe in the Columbia Basin Ecoregion (Soll et al. 1999). Of the state-listed rare plants found on the Hanford Site, 18 (62%) are annuals, and 3 of these have not been found elsewhere in Wash- ington. Because of the relatively high number of rare annuals on the Hanford Site, it is an important location for the conservation of rare annuals in Washington. Annuals tend to occur in 2006] [ABLE IL. BIODIVERSITY PROJECT. SUMMARY OF ADDITIONS TO THE VASCULAR FLORA OF WASHINGTON FROM THE BECK AND CAPLOW: ADDITIONS TO WASHINGTON FLORA FROM THE HANFORD SITE = 43 HANFORD Species Previously known range Washington counties where currently known Washington Natural Heritage Program status (2005) Astragalus conjunctus Not known var. rickardii Cistanthe rosea Oregon, California, Idaho, Nevada, Wyoming Not known Chelan County, Washington California, Oregon, Idaho, Colorado, New Mexico, Arizona, Nevada, Wyoming, Montana Not known Eriogonum codium Festuca washingonica Gilia leptomeria Physaria douglasii subsp. tuplashensis Loeflingia squarrosa ssp. squarrosa Myosurus clavicaulis Western states not including Pacific NW California, Oregon moist microhabitats in open sandy or gravelly soils, and many of these sites elsewhere in the state have been disturbed and compacted by grazing and invaded by Bromus tectorum. At the time that the Hanford Biodiversity Study began, the conservation of the biodiversity of the Hanford Site was not assured. There was a powerful local movement to distribute most of the land north of the Columbia River to private landowners for agricultural development. The future protection of the ALE Reserve was also in question. Since that time, the Hanford Reach has been declared a Wild and Scenic River and much of the Hanford Site, including the ALE Reserve, has been set aside as the Hanford Reach National Monument. However, the central portion of the Site, which supports a number of rare plant populations, is still primarily dedicated to nuclear waste storage and clean-up. The FEriogonum codium population, although formally within the Hanford National Monument, is not in an area currently managed by the USFWS. Even with protection, active, on-going man- agement will be necessary to maintain the biodiversity of the Hanford Site. In particular, noxious weeds (e.g., Bromus tectorum L., Salsola kali L., Centaurea solstitialis L. and others) and wildfire will bring adverse changes to natural vegetation communities over time. Nonetheless, the future of the vascular plant diversity of the Hanford Site appears considerably better than it did in 1994, when the Hanford Biodiversity Project was initiated. ACKNOWLEDGMENTS We would like to thank the following organiza- tions and people for their gracious assistance: The U.S. Department of Energy, The Nature Conservancy of Washington, The Pacific Northwest National Labs, Peter Dunwiddie, Pam Camp, Curt Soper, Benton Watch Benton Threatened Benton Endangered Benton, Yakima, Kittitas, None Chelan, Okanogan Franklin, Grant, Benton Threatened Franklin Threatened Benton Threatened Adams, Spokane, Benton Sensitive Bobby Tomanawash of the Wanapum people, Mike Sackschewsky, Barbara Wilson, Rupert Barneby, Stan Welsh, Ron Hartman, Alva Day, Susan Aiken, Allan Whittemore, and Jim Reveal. This paper is dedicated to the memory of Reed Rollins, whose lifelong knowledge of the Cruciferae, kindness, patience, and good humor were an inspiration to both of us. LITERATURE CITED AIKEN, S. G., S. E. GARDINER, H.C. M. BASSETT, B. L. WILSON AND L. L. CONSAUL. 1998. Implications from SDS-PAGE analyses of seed proteins in the classification of taxa of Festuca and Lolium (Poaceae). Biochemical Systematics and Ecology 26:511—533. ALEXEEV, E. B. 1982. New and little-known fescues (Festuca L.) of North America. Bjulleten Moskovs- kogo ObSéestva Ispytatele} Prirody, Otdel Biologi- ceskij [Bulletin of the Moscow Society of Natural- ists, Biological Division] 87(2): 109-118. AL-SHEHBAZ, I. AND S. O’KANE. 2002. Lesquerella is untied with Physaria (Brassicaceae). Novon 12:319-329. BAER (NORTHERN STATES BURNED AREA EMERGEN- CY REHABILITATION) TEAM. 2000. 24 Command Fire burned area emergency rehabilitation (BAER) plan. United States Department of the Interior, Washington DC. BARNEBY, R. C. 1964. Atlas of North American Astragalus. Memoirs of the New York Botanical Garden 13:1—1188. DAUBENMIRE, R. 1970. Steppe vegetation of Washing- ton. Washingon State Agricultural Experiment Station, Technical Bulletin 62:1—131. Day, A. G. 1993. New taxa and nomenclatural changes in Allophyllum, Gilia, and Navarretia (Polemonia- ceae). Novon 3:331—340. Downs, J. L., W. H. RICKARD, C. A. BRANDT, L. L. CADWELL, C. E. CUSHING, D. R. GEIST, R. M. MAZAIKA, D. A. NEITZEL, L. E. ROGERS, M. R. SACKSCHEWSKY AND J. J. NUGENT. 1993. Habitat types on the Hanford Site: wildlife and plant species of concern. PNL-894, Pacific Northwest National Laboratory, Richland, WA. A4 MADRONO DUNWIDDIE, P. W., K. A. BECK AND F. E. CAPLOw. 2001. Demographic studies of Eriogonum codium in Washington. Pp. 59-69 in S. H. Reichard, P. W. Dunwiddie, J. Gamon, A. R. Kruckeberg, and D. L. Salstrom (eds.), Conservation of Washing- ton’s rare plants and ecosystems. Washington Native Plant Society, Seattle, WA. FRANKLIN, J. F. AND C. T. DYRNESS. 1973. Natural vegetation of Oregon and Washington. General Technical Report PNW-8, U.S. Forest Service, Portland, OR. GOFF, F. E. 1981. Preliminary geology of eastern Umtanum Ridge, south central Washington. RHO-BWI-C-21, Rockwell Hanford operations, Richland, WA. GRAY, A., S. WATSON AND B. L. ROBINSON. 1895. Synoptical flora of North America I pt. I, Fascicles 1. Polypetalae from the Ranunculaceae to the Frankeniaceae. Continued and edited by Benjamin Lincoln Robinson. Ivison, Blakeman, Taylor, and Co., New York, NY. HICKMAN, J. C. (ed.). 1993. The Jepson manual: higher plants of California. University of California Press, Berkeley, CA. HITCHCOCK, C. L., A. CRONQUIST, M. OWNBEY AND J. W. THOMPSON. 1955-1969. Vascular plants of the Pacific Northwest. University of Washington Press, Seattle, WA. AND A. CRONQUIST. 1973. Flora of the Pacific Northwest: an illustrated manual. University of Washington Press, Seattle, WA. INTEGRATED TAXONOMIC INFORMATION SYSTEM ON- LINE DATABASE. Available at: http://www. itis. usda.gov. Accessed: December 12, 2005. KARTESZ, J. T. 2003. A synonymized checklist and atlas with biological attributes for the vascular flora of the United States, Canada, and Greenland, 2nd ed. In J. T. Kartesz and C. A. Meacham (eds.), Synthesis of the North American flora, Version 2.0. J.T. Kartesz & Phylosystems Corp., Rich- mond, CA. NATURE CONSERVANCY OF WASHINGTON. 1998. Bio- diversity inventory and analysis of the Hanford Site 1997 annual report. The Nature Conservancy of Washington, Seattle, WA. Newcoms, R. C. 1958. Ringold Formation of the Pleistocene Age in type locality: the White Bluffs, Washington. American Journal of Science 256:328—340. PACIFIC NORTHWEST NATIONAL LABORATORY. 1997. Climatological data, Hanford Meteorology Sta- tion. Pacific Northwest National Laboratory, Richland, WA. Available at: http://etd.pnl.gov: 2080/HMS/. PAYSON, E. B. 1922. A monograph of the genus Lesquerella. Annals of the Missouri. Botanical Garden 8:103—236. Peck, M. E. 1961. A manual of the higher plants of Oregon. Oregon State University Press, Portland, OR. PIPER, C. V. 1906. Flora of the State of Washington. Contributions of the U.S. National Herbarium 11:1-637. REVEAL, J. L., F. CAPLOW AND K. BECK. 1997 [1995]. Eriogonum codium (Polygonaceae: Eriogonoideae), a new species from southcentral Washington. Rhodora 97:350—356. [Vol. 53 ROLLINS, R. C. 1993. The Cruciferae of Continental North America. I-XVI, 1-976. Stanford University Press, Stanford, CA. AND E. C. SHAW. 1973. The Genus Lesquerella (Cruciferae) in North America. Harvard University Press, Cambridge, MA. , K. A. BECK AND F. E. CAPLOw. 1996 [1995]. An undescribed species of Lesquerella (Cruciferae) from the state of Washington. Rhodora 97:201—207. Rose, J. N. 1904. William M. Canby, Botanical Gazette 37:385-387. SACKSCHEWSKY, M. R. AND J. L. DOwns. 2001. Vascular plants of the Hanford Site. PNNL- 13688, Pacific Northwest National Laboratory, Richland, WA. SOLL, J., J. A. HALL, R. PABST AND C. SOPER. 1999. Biodiversity inventory and analysis of the Hanford Site Final Report: 1994-1999. The Nature Conser- vancy of Washington, Seattle, WA. STONE, D. E. 1959. A unique balanced breeding system in the vernal pool mouse-tails. Evolution 13:151-174. U.S. FISH AND WILDLIFE SERVICE. 1999. Endangered and threatened wildlife and plants; review of plant and animal taxa that are candidates or proposed for listing as endangered or threatened; notice of review. Federal Register 64(205): 57534-57547. WASHINGTON NATURAL HERITAGE PROGRAM. 1994. Endangered, threatened and sensitive vascular plants of Washington—with working lists of rare non-vascular species. Department of Natural Re- sources, Olympia, WA. WASHINGTON NATURAL HERITAGE PROGRAM. 2005. List of plants tracked by the Washington Natural Heritage Program. Available at: http://www. dnr.wa.gov/nhp/refdesk/lists/plantrnk.html. WELSH, S. L., F. CAPLOW AND K. BECK. 1997. New variety of Astragalus conjunctus S. Watson from Benton County, Washington. Great Basin Natu- ralist 57:352—354. WHITTEMORE, A. T. 1997. Myosurus Linnaeus, Sp, PI. 1: 2284. 1753; Gen. Pl. ed. 5, 137. 1754. Flora of North America 3:135—138. WILSON, B. L. 1999. Festuca taxonomy in the Pacific coast states. Unpublished Ph.D. thesis. Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR. APPENDIX | RELEVANT HERBARIUM COLLECTIONS BY THE AUTHORS OF RARE PLANTS ON THE HANFORD SITE Astragalus conjunctus var. rickardii Welsh, K.A. Beck & Caplow (FABACEAE). TYPE: U.S.A., Washington, Benton Co. Hanford Site, on northeast-facing slopes of Rattlesnake Mountain, with Artemisia tridentata, Poa sandbergii, 1036 meters, TIIN R26E sect. 30 NW'% of SW %, 29 May 1995, Kathryn Beck & Florence Caplow 95-083, (Holotype: BRY!; Isotypes: NY!, US!, WTU!, WS!):; Benton County, Horse Heaven Hills, Chandler Butte, BLM owned, in silt loam, upper north-facing slopes, with Agropyron spicatum, Poa cusickii, Artemisia tridentata, Poa sandbergii, 600 meters, T9N R26E sect. 22 SE% of NW%, 14 April 1995, Florence E. Caplow & Kathryn A. Beck 95-022 (BRY, WTUV). 2006] Cistanthe rosea S. Watson (PORTULACACEAE). Benton Co., Hanford Site north of Gable Mountain in basalt derived sands, in a small swale, with Artemisia tridentata, Bromus tectorum, Descurainea pinnata, Loe- flingia squarrosa ssp. squarrosa and Mimulus suksdorfii, 12% slope, south exposure, 150 meters, TI3N R26E sect. 2SE“%s of S%, 28 May 1995, Florence E. Caplow & Kathryn A. Beck 95-078 (WTU). Eriogonum codium Reveal, Caplow & K.A. Beck (POLYGONACEAE). TYPE: U.S.A., Benton Co., Hanford Site, on the northern edge of Umtanum Ridge west of Washington Highway 24 overlooking the Columbia River about 38 air miles northwest of Richland, on volcanic soil, with Grayia_ spinosa, Artemisia tridentata, Salvia dorrii, Hesperostipa comata, and Pseudoroegneria spicata, 350 meters, TI3N R24E sect. 13, 27 June 1995, Reveal, Caplow and Sackschewsky 7484 (Holotype: US; Isotypes: BM, BRY, CAS, COLO, GH, K, MARY, MO, NY, RM, RSA, TEX, UC, WS, WTU and elsewhere); Also in reddish to black, hard-packed basalt gravel, with Artemisia tridentata, Salvia dorrii, Poa sandbergii, Bromus tectorum, Phacelia linearis, 300°, 3% slope, TI3N R24E sect. 13 NE% of SW'%, 31 May 1995, Florence E. Caplow & Kathryn A. Beck 95-084 (WTU, WS). Festuca washingtonica E.B.Alexeev (POACEAE). Benton Co., Arid Lands Ecology Reserve, Hanford Site, Rattlesnake Mountain, in a narrow strip just below the top of the mountain, in the “snowmelt” zone, in rocky silt loam, with Artemisia tridentata, Lupinus laxiflorus var. calcaratus, Poa nevadensis, Senecio integerrimus, Festuca idahoensis, Melica bulbosa, and Poa cusickii, 45°, 35%, 1130 meters, TIIN R26E sect. 30 NW of SW%, 4 June 1995, Florence E. Caplow & Kathryn A. Beck 95088 (OSU, WTU); Chelan Co., Wenatchee National Forest, Wenatchee Mountains, on lower west-facing dry slopes of a small tributary in Peavine Canyon, with Pinus ponderosa, Symphoricarpos albus, Salix sp., Acer douglasii, Prunus emarginata, Purshia tridentata, Lupinus leucophyllus, Aster foliaceus, Agropyron spicatum, and Poa_ pratensis, 270°, 5%, 630 meters, T22N RI9E sect. 15 SW'% of SE%, Kathryn A. Beck & Florence E. Caplow 96-052 (OSU, WTU). Gilia leptomeria A. Gray (POLEMONIACEAEBE),. Grant Co., Hanford Site, Saddle Mountain NWR, on upper slopes of gravelly bluff north of Vernita Bridge on the Columbia River, with Salvia dorrii, Camissonia pygmaea, and Eatonella nivea, south exposure, 30%, 215 meters, TI4N R25E sect. 31 SE%, 15 May 1995, Florence E. Caplow & Kathryn A. Beck 95-035 (WTU): Franklin Co., Hanford Site, Wahluke Wildlife Area, White Bluffs of the Columbia River, near top of bluffs, in caliche soil, with Eriogonum microthecum, Poa secunda, Bromus tectorum, Astragalus caricinus, and Eurotia lanata, western exposure, 285 meters, T13N R27E sect. 25, 19 May 1995, Florence E. Caplow & Kathryn A. Beck 95-055 (WTU); Also on. sparsely vegetated ground in sand and caliche, with Chrysotham- nus nauseosus, Bromus tectorum, Astragalus caricinus, Artemisia tridentata, Astragalus succumbens, and Gilia sinuata, 220°. 5%, T13N R28E sect. 19 SE% of NW, 31 May 1997, Florence E. Caplow & Kathryn A. Beck 97-042 (WTU): Benton Co., Hanford Site Umtanum Ridge, ridge-top, at edges of steep north-facing slopes, BECK AND CAPLOW: ADDITIONS TO WASHINGTON FLORA FROM THE HANFORD SITE — 45 in fine reddish to blackish basalt, gravel and pumice, with Artemisia tridentata, Eriogonum codium, Bromus tectorum, Poa secunda, and Salvia dorrii, 350 meters, TI3N R24E sect. 13 NE% of SW’4, 31 May 1995, Florence E. Caplow & Kathryn A. Beck 95-076 (WTU); Additional unvouchered populations were located. Plants were determined by Alva Day, 1995. Physaria douglasti subsp. tuplashensis Rollins, K.A. Beck & Caplow (BRASSICACEAE). TYPE: U.S.A. Washington, Franklin County: White Bluffs, above the Columbia River, caliche soil at edge of eroding bluff, with Artemisia tridentata, Astragalus caricinus, Cryp- tantha_ spiculifera, Eriogonum microthecum, and Poa sandbergii, T13N R27E sect. 11 W1/2, 20 July 1994, Kathryn A. Beck & Florence Caplow 94-001 (Holotype: GH; Isotype: WTU). Franklin Co., Hanford Site, Wahluke Wildlife Area, White Bluffs northeast of the Columbia River, near powerlines, on the leading edge of the bluffs, in hard caliche soils, with Pod secunda, Astragalus caricinus, Cryptantha spiculifera, Eriogonum microthecum var. laxiflorum, 300 meters, TI3N R27 E sect. 24 NE% of NE™%, 19 May 1995, Florence E. Caplow & Kathryn A. Beck 95-053 (WTU, GH). Upper slopes near mouth of Ringold Canyon, in loose caliche soils, with Poa secunda, Bromus tectorum, Amsinckia tessellata, 220 meters, TIZ2N R28E sect. 11 NW'% of SW, 1 June 1995, Florence E. Caplow & Kathryn A. Beck 95-086 (GH). Loeflingia squarrosa Nutt. ssp. squarrosa (CARY O- PHYLLACEAE). Benton Co., Hanford Site, north of Gable Mountain in basalt-derived, stabilized sand dunes, on a sandy berm of a little-used road, with Artemisia tridentata, Poa secunda, Bromus tectorum, Cryptantha circumcissa, and Mimulus suksdorfii, 145 meters, T13N R26E sect. 13, 16 May 1995, Florence E. Caplow & Kathryn A. Beck 95-039 (WTU): Benton Co., Hanford Site, between two main portions of Gable Butte, in shrub-steppe at edges of vernal pool in rocky basalt layer, with Artemisia tridentata, Gnaphalium palustre, Camissonia andina, Juncus bufonius, Epilobium minutum, Moss spp., and lichen spp., 175 meters, T13N R26E sect. 18 SW'%4 of SW '%4, 20 May 1997, Florence E. Caplow & Kathryn A. Beck 97-023 (WTUV); also south of Gable Mountain, plants growing in shrub-steppe, in barren basalt-derived sand, along a little-used sandy road, with Artemisia tridentata, Ambrosia acanthicarpa, Pectocarya_ linearis, Bromus tectorum, and Mimutlus suksdorfii, 160 meters, TI3N R26E sect. 25 SE'% of NE™%, 29 May 1997, Florence Caplow & Kathryn A. Beck 97-041 (WTU). Plants were verified by Ronald Hartman, 1995. Additional unvouchered populations were located, including ones at the eastern end of Umtanum Ridge, in a vernal pool and also in a swale amidst shrub-steppe. Myosurus clavicaulis Peck (RANUNCULACEAE). Benton Co., Hanford Site, top of east end of Umtanum Ridge, plants growing on dried edges of vernal pool in depression of basalt layer, with Veronica peregrina var. xalapensis, Camissonia andina, Myosurus apetalus, Epilobium minutum, Artemisia rigida, moss spp., and lichen spp:, 315°, 1% slope, 195 meters, TI3N R23E sect. 17 Y% of %, 6 May 1997, Florence E. Caplow & Kathryn A. Beck 97-010 (US, WTU); same location, 7 May 1997, Florence E. Caplow & Kathryn A. Beck 97- 014 (WTU), verified by Alan Whittemore. MADRONO, Vol. 53, No. 1, pp. 46-54, 2006 A RECONSIDERATION OF BRODIAEA MINOR (BENTH.) S. WATSON AND BRODIAEA PURDYI EASTWOOD (THEMIDACEAE), WITH THE RESURRECTION OF BRODIAEA NANA HOOVER ROBERT E. PRESTON Jones & Stokes, 2600 V Street, Sacramento, CA 95818 rpreston@jsanet.com ABSTRACT A review of taxonomic literature, examination of existing herbarium specimens, and a morpholog- ical study of field-collected material demonstrates that species circumscriptions have been misapplied for the small-flowered Brodiaea species with spreading perianth lobes and floral tubes narrowed above the ovary. The results of these studies demonstrate that Niehaus’s (1971) concept of B. purdyi, polyploid plants that occur in woodland habitats in the northern Sierra Nevada foothills, applies to the taxon originally described as B. minor, placing B. purdyi in synonymy with B. minor. Niehaus’s concept of B. minor, diploid plants occurring in vernal pool terrain, applies only to those populations originally described as Brodiaea nana, which is resurrected at species rank. Key Words: Themidaceae, Brodiaea, taxonomic revision, California. The genus Brodiaea (Themidaceae) consists of approximately 14 or 15 species, almost entirely restricted to the California Floristic Province (Niehaus 1971, 1980; Keator 1993; Pires 2002). Brodiaea has a rich taxonomic history and has been placed variously in Liliaceae, Amaryllida- ceae, and Alliaceae (Hoover 1939; Keator 1967, 1989; Niehaus 1971, 1980). Recent phylogenetic studies, however, place Brodiaea and relatives not with Allium but with Hyacinthaceae and other families (Fay and Chase 1996; Fay et al. 2000; Pires et al. 2001; Pires and Sytsma 2002). As a result, Brodiaea has been reassigned to the family Themidaceae or a more inclusive Aspar- agaceae (Angiosperm Phylogeny Group 2003). These studies have focused on relationships among families and genera and have not ad- dressed relationships within Brodiaea, which remain poorly resolved despite having been monographed twice (Hoover 1939; Niehaus 1971). Species circumscriptions and relationships among species historically have been difficult to elucidate, largely because study of fresh material is crucial for comparison of the diagnostic floral features, which are obliterated when specimens are pressed and dried (Greene 1886; Hoover 1939). Pires (2002) points out the need to prepare open flowers when making herbarium specimens, but even with fresh material, making a determi- nation with confidence can often be frustrating. The small-statured species with spreading perianth lobes and floral tubes that are narrowed above the ovary exemplify this taxonomic diffi- culty. Current floristic treatments of Brodiaea (Keator 1993; Pires 2002) recognize two species, Brodiaea minor (Benth.) S. Watson and Brodiaea purdyi Eastwood, based on Niehaus’s (1971) monograph of the genus. Niehaus differentiated between the two species based on morphology, cytology, and ecology. However, it is often not possible to assign specimens unambiguously to one or the other species, using the current taxonomic keys (Oswald 1994; personal observa- tion). In this paper, I show that the frustration with species determinations using current floristic treatments is not due simply to an inadequate diagnostic key, but stems from a more funda- mental error. I provide a morphometric analysis supporting Niehaus’s recognition of two taxa at species rank, but I demonstrate that Niehaus misapplied the name B. minor and did not correctly circumscribe all populations under the correct species concepts. I discuss the source of Niehaus’s error and clarify the nomenclature. In addition, I discuss the relationship of these two species with other members of the genus. METHODS I examined herbarium specimens of B. minor and B. purdyi, as circumscribed by Niehaus (1971), in the principal collections of both species (herbaria consulted: JEPS, UC, CHSC, DAV) and photographs of the types of B. minor, B. purdyi, and B. nana. 1 sampled 36 populations throughout the ranges, based on localities pro- vided on the specimen labels. I collected fresh material and dissected one flower from 10 plants in each population, using flowers at approxi- mately the same stage of anthesis, to minimize variation due to any change in flower size from the beginning to the end of anthesis. I measured 11 floral characters and noted the shape and position of the floral parts. I employed principal components analysis, using the SYSTAT 11 2006] statistics package (SYSTAT Software, Rich- mond, CA), to reduce the number of variables and simplify the morphological comparison. The analysis was performed using the mean floral measurements from each population. Factor scores for the first two principal components were then plotted to determine whether discrete groups of populations could be recognized. RESULTS AND DISCUSSION Taxonomic Review Theodor Hartweg collected the type of Bro- diaea minor. In the spring of 1847, he had traveled to California on a mission to collect botanical specimens for the Horticultural Society of London (Hartweg 1848). Hartweg made numerous collections during his stay at the ranch of ““Mr. L.”’, in the northern Sacramento Valley (undoubtedly Peter Lassen, who homesteaded in southern Tehama County, near the present town of Vina (Swartzlow 1964)). During a visit to the foothills east of the ranch, he collected specimens that later became the type of Brodiaea grandiflora Sm. var. minor Benth. Sereno Watson (1879) later raised var. minor to species rank. Greene (1894) apparently initiated some con- fusion by applying the name B. minor to all of the small-flowered brodiaeas in the Central Valley and adjacent Sierra Nevada foothills. Subse- quently, Alice Eastwood (1896) described a “‘new’’ species from the northern Sierra Nevada foothills, Brodiaea purdyi, noting the long, narrow perianth lobes as the distinguishing feature of this species. Jepson (1922) recognized that Greene had encompassed several different taxa under the name B. minor and that B. purdyi was synonymous with B. minor, as originally described by Bentham. Jepson applied the name Brodiaea synandra (Heller) Jepson to the small- flowered plants of the Central Valley that, like B. minor, had the perianth tube narrowed above the ovary. Unfortunately, Jepson did not have access to the type specimen of B. synandra, which actually is conspecific with the earlier-published Brodiaea leptandra (E. Greene) Baker. Jepson also repeated Greene’s error, citing specimens now assigned to several different species, in- cluding B. coronaria and B. terrestris, within his circumscription of B. synandra. Hoover (1936, 1939) eventually sorted out the nomenclatural confusion. Hoover (1939) was the first to monograph the genus and developed most of the species concepts that are still used to circumscribe the taxa. Brodiaea species have traditionally been differentiated on the basis of the shape and position of the floral parts, and Hoover followed this tradition by recognizing species when there were discrete differences in morphology and recognizing varieties when taxa PRESTON: BRODIAEA MINOR REVISED 47 differed primarily in the size of the floral parts. For the small-flowered plants along the east side of the Central Valley, he proposed the name Brodiaea nana, because of their small stature, having scapes less than 5 cm tall. Hoover (1939) subsequently reduced B. nana to a variety of B. minor, citing his observations that the two taxa intergraded morphologically. Niehaus (1971) expanded on Hoover’s work with Brodiaea by incorporating observations from anatomy, cytology, palynology, flavonoid chemistry, ecology, and hybridization studies. Although his study tended to support Hoover’s taxonomic framework, he expanded some of the morphologically-based species concepts in Bro- diaea to include data from cytology and ecology. Niehaus recognized two small-statured, small- flowered species with spreading perianth lobes and floral tubes that are narrowed above the ovary, one consisting of populations of diploid (n = 6) plants growing in vernal pool terrain along the eastern edge of the Central Valley, the other consisting of tetraploid (n = 12) and octaploid (n = 24) populations occurring in foothill habitats, often on gabbro or serpentine. To the former species, which essentially followed Hoover's concept of B. nana, he applied the name B. minor, placing B. nana in synonymy. He resur- rected the name B. purdyi to apply to the latter species. Recent floristic treatments of Brodiaea (Keator 1993; Pires 2002) mostly followed Niehaus’s treatment of the genus and maintained both B. minor and B. purdyi at species rank, although Keator noted that B. purdyi might merit placement as a subspecies of B. minor. Morphological Study The morphological study found that plants from the 36 sampled populations could be unambiguously assigned to one of two groups, but not to the groups represented by B. minor and B. purdyi as circumscribed by Niehaus (1971). First, two groups were differentiated by the shape and position of the stamens and staminodes. The first group of populations (Group A) had stamen morphology that was unusual for the genus and most similar to that described for B. pallida (Hoover 1938), including the presence of prom- inent papillae on the abaxial surface of the anthers (Fig. la). The connective broadened towards the apex, which was widely V-notched (Fig. la, b), and the filaments were short and abaxially winged (V- or Y-shaped in cross-section [Fig. lc]). The staminodes were short, broad, and erect, with slightly inrolled margins (Fig. Id, 2a). The styles were about 1.5 times longer than the ovary (Fig. le). Stamen morphology in the second group of populations (Group B) was not remarkably different from that in many other Brodiaea A8 MADRONO [Vol. 53 A An Adidas Srdhenesctschtats act Sivas sh ree s) oC ER NT J Fic. |. Comparison of inner floral parts. A-E Group A (Brodiaea nana Hoover). F—J Group B (Brodiaea minor (Benth.) S. Watson). A, F. Stamen (adaxial view). B, G. Stamen (abaxial view). C, H. Filament (cross-section). D, I. Staminode. E, J. Pistil. The scale bar represents a length of 5 mm. species. Abaxial papillae were present on the anthers but were not prominent (Fig. If). The connective was uniformly wide to only slightly broader at the apex, which was narrowly notched (Fig. If, g), and the filaments were longer and laterally winged (T-shaped in cross-section [Fig. lh]). The staminodes were longer and narrower with strongly inrolled margins (Fig. 11) 2006] PRESTON: BRODIAEA MINOR REVISED 49 FIG. 2. Comparison of flowers. A, B Brodiaea nana Hoover. A. Top view. B. Lateral view, showing constriction above the ovary. C, D Brodiaea minor (Benth.) S. Watson. C. Typical form. D. Narrow-lobed form described as B. purdyi Eastw. and were erect to recurved at the tip (Fig. 2c, d). The styles were about 1.75 times the length of the ovary (Fig. 1)). The principal components analysis reduced the floral variables to two factors (Table 1). The first principal component, which explained almost 79% of the variation, appeared to be a size factor, primarily loading on length of the floral parts. The second principal component also appeared to be a size factor, but loading on the size of the perianth lobes (length and width). The plot of the two principal components also separated the populations into two groups that corresponded precisely with Groups A and B, but not to B. minor and B. purdyi as currently circumscribed (Fig. 3). Group A corresponds closely to Hoover’s (1936) original circumscription of Brodiaea nana and includes populations along the eastern edge of the Central Valley, ranging from Butte County to Merced County, where the type was collected. Group A includes all of the populations Niehaus (1971) determined to be diploid. Group B consists of populations Niehaus assigned to B. purdyi but also includes populations he assigned to B. minor. Populations comprising Group B range from the Sierra Nevada foothills to the northern Sacramento Valley in Butte and Te- hama Counties, encompassing the type localities of both B. minor and B. purdyi. Group B includes the populations Niehaus (1971) determined to be tetraploid and octaploid. Therefore, Group B corresponds to B. minor as circumscribed by Jepson (1922) and Hoover (1939), rather than Niehaus’ (1971) later circumscription. Brodiaea nana Resurrected The results of the morphological analysis show that B. nana should be recognized as a taxon distinct from B. minor, and on both morpholog- ical and cytological grounds, B. nana warrants recognition at the rank of species. Hoover (1936) originally described B. nana at species rank, but he later (1939) reduced it to a variety of B. minor, citing his observation that specimens from Sacramento County were intermediate between B. minor and B. nana. Hoover did not elaborate 50 MADRONO [Vol. 53 TABLE |. RESULTS OF PRINCIPAL COMPONENTS ANALYSIS ON MEANS OF ELEVEN VARIABLES FROM SIXTEEN POPULATIONS OF B. NANA (GROUP A) AND 20 POPULATIONS OF B. MINOR (GROUP B). Variable POol PC 2 Pedicel Length =0,25] 0.428 Perianth Tube Length 0.917 0.189 Perianth Lobe Length 0.531 0.529 Width, Inner Lobe —0.411 0.859 Width, Outer Lobe 0.088 0.925 Staminode Length 0.958 0.071 Filament Length 0.924 C127 Anther Length 0.930 —0.024 Ovary Length 0.951 —0.025 Style Length 0.983 0.011 Ovule Number =O5758 0.299 Eigenvalue 6.446 2.204 Variation Explained 58.60% 20.04% OB. minor m@ B. purdyi Fic. 3. Two-dimensional scatter diagram of first and second principal component scores based on population | means of floral characters from B. minor and B. purdyi (sensu Niehaus [1971]). Group A corresponds to B. nana | Hoover, and Group B corresponds to B. minor (sensu Jepson (1922) and Hoover (1939). 2006] TABLE 2. PRESTON: BRODIAEA MINOR REVISED 5] COMPARISON OF FLORAL CHARACTERS FOR BRODIAEA MINOR AND BRODIAEA NANA. Measurements were made on fresh material, from one flower per plant and 10 plants per population, from 20 populations of B. minor and 16 populations of B. nana. Measurements in mm. Brodiaea minor Brodiaea nana Character mean range mean range Pedicel 21.4 7-45 24.5 6-56 Perianth tube 8.6 6.5-11.5 13 5.0—9.0 Perianth lobes 15.0 9.8—20.5 14.3 10.0—21.0 Width, inner lobes 4.7 3.0-7.0 Dee) 4.0-8.0 Width, outer lobes 3.7 2.8—5.0 3.8 3.0-5.0 Staminode on 6.2—12.5 i 6.0—9.0 Filament Diz ].0—3.5 1.4 1.0—2.0 Anther Dee 3.5-7.0 4.0 3.0—5.0 Ovary 4.9 3.2-7.0 55 2.5—5.0 Style 8.8 6.0-12.0 Deo 4.0-7.5 Ovule number |e gee ]2—24 222 12-33 on which features were intermediate. The ranges for all floral part measurements do overlap, but on average, all floral parts of B. nana are smaller than those of B. minor (Table 2). It is more noteworthy that the shapes of the staminodes, stamens, and pistils consistently differentiate B. nana from B. minor (Fig. 1), because Brodiaea species traditionally have been recognized on the basis of the shape and position of the floral parts. Recognizing B. nana at species rank is also consistent with Niehaus’s (1971) expanded spe- cies concepts in Brodiaea. Niehaus’s (1971) diploid chromosome counts, a major criterion for re-establishing B. nana at species rank (albeit as B. minor), were all based on populations of B. nana as circumscribed by Hoover and confirmed as such by the present morphological study. The distribution of Brodiaea nana, documented by herbarium specimens and confirmed by visits to the collection localities, ranges from Merced County north to Chico, in Butte County (Fig. 4). In addition, several disjunct populations of B. nana occur on volcanic mudflows adjacent to Payne’s Creek and Battle Creek, in northern Tehama County and southern Shasta County. The distributions of B. nana and B. minor overlap in Butte and Tehama Counties, but the two species are almost never sympatric. However, I collected both species growing together at one location in Chico, Butte County. Brodiaea nana occurs in vernal swales, shallow vernal pools, and on the margins of deeper vernal pools. Brodiaea minor Revisited and Brodiaea purdyi Reconsidered It is clear that Niehaus’s (1971) concept of B. minor applied only to those populations circum- scribed by Hoover’s B. nana. It also clear that many of the populations Niehaus assigned to B. minor are morphologically indistinguishable from populations he assigned to B. purdyi. Niehaus’s concept of two species, one consisting of diploid populations occurring in vernal pool terrain, the other of polyploid populations occurring in foothill habitats, appears to have been only partially correct, as some populations of B. minor occur in vernal pool terrain. Moreover, he misapplied the names when circumscribing the populations that made up the two species. How did this error come about? First, the flowers of both species are super- ficially similar (Fig. 2a, c), and many of the floral parts overlap in size (Table 2). Niehaus used scape length (=10 cm = B. minor, =10 cm = B. purdyi) and petal width (S—~7 mm = B. minor, 4-5 mm = B. purdyi) as key characters for separating the two species. The type of B. minor (Hartweg 2002 [Isotype, NY]) has short scapes, and Niehaus evidently presumed that this population was assignable to the same taxon as Hoover’s B. nana, and that the correct name for the taxon, therefore, was B. minor. However, the type locality of B. minor occurs in blue oak-foothill pine woodland (Hartweg 1848), not in vernal pool terrain. Moreover, scape length is not a reliable character for differentiating between Brodiaea taxa. Scape length varies both within and among Brodiaea populations and may be environmentally plastic, to some degree (Doalson 1999). The type specimen of B. purdyi (Purdy s.n. [CAS]), which illustrates this variation quite nicely, consists of three plants, one with a short scape, one with a long scape, and one with an intermediate- length scape. Petal width also overlaps between B. minor and B. nana (Table 2) and is not reliable for differentiating between them. The results of this paper demonstrate that Niehaus’s (1971) concept of B. purdyi, the poly- ploid small-flowered species, with spreading perianth lobes and floral tubes that are narrowed above the ovary, and that occurs in woodland habitats in the northern Sierra Nevada foothills, applies to the taxon originally described as B. minor and as recognized by Jepson (1922) and N N FIG. 4. Hoover (1939), placing B. purdyi in synonymy with B. minor. Currently, there is no basis for recognizing B. purdyi as a separate taxon. Eastwood (1896) noted that the original collections of B. purdyi were remarkable for their relatively long, narrow perianth lobes (Fig. Id). In all other respects, however, including the shape and relative posi- tion of the floral parts, populations cannot be differentiated reliably. Moreover, there is sub- stantial variation in perianth lobe length among populations of B. minor, and plants with long, narrow lobes appear to be at one end of a continuum of variation in lobe length (personal observation). As recognized in this study and as documented by herbarium specimens, Brodiaea minor (in- MADRONO [Vol. 53 100 50 Miles (approximate) Distribution of Brodiaea minor (m), Brodiaea nana (a), and Brodiaea pallida (%) in California, USA. cluding B. purdyi) ranges along the eastern margin of the northern Sacramento Valley, from | Shasta County to Butte County, into the Sierra Nevada foothills, and south to Amador County | (Fig. 4). Most populations occur in vernal pool terrain, oak woodland, or chaparral, with a few _ populations occurring in dry montane meadows at higher elevations. Although some populations — occur on gabbro or serpentine, B. minor does not appear to be restricted to those substrates. Species Relationships Relationships between Brodiaea species are — poorly understood. Hoover (1939) recognized a series of infrageneric groups, based on floral | morphology. He proposed four informal sections, t 2006] including section ‘‘Stellares’”’, within which he placed B. stellaris, B. pallida, and B. minor (including B. nana). Niehaus (1971) added B. insignis to this group. Section “‘Stellares” is composed of small-flowered species with rotate corollas, broad staminodes, and short filaments that are more or less channeled on the abaxial side. The strong morphological similarity between B. minor and B. nana, as shown in this study, supports a close relationship between these two species. Niehaus (1971) found that the flavonoid chemistry and floral vasculature of the two species was also very similar. Brodiaea pallida and B. nana appear to be closely related, as well. Both species are diploid (n = 6) and have similar flavonoid chemistry (Niehaus 1971), and their ranges overlap (Fig. 4). Their floral morphology is also quite similar. The perianth tube in B. pallida is not or only slightly narrowed above the ovary, but in both species the staminodes are erect and the margins only slightly inrolled. The anthers have prominent abaxial papillae, the connective broadens towards the apex (see Fig. 16 in Niehaus [1971]), and the filaments are abaxially winged, although the wings in B. nana are not as pronounced as in B. pallida. Hoover (1938, 1939) discussed at length the unusual morphology of the staminodes and stamens in B. pallida. His statement that these features were quite different from those of B. minor and his later treatment of B. nana as a variety of B. minor suggests that he was unaware of the similarities between B. pallida and B. nana. Brodiaea stellaris and B. insignis appear to be less closely related to B. nana, B. minor and B. pallida. Flowers of B. insignis are at least superficially similar to those of B. nana and B. minor (unpublished data), although the floral tube is not constricted and the chromosome number (n = 16) and flavonoid compounds are substantially different than those two species (Niehaus 1971). In contrast, B. stellaris is a diploid (n = 6) with similar flavonoid chemistry to B. nana, B. minor, and B. pallida, but it is morphologically quite different from these spe- cies. Hoover (1939) originally grouped B. pallida with B. stellaris because the filaments of both Species are prominently winged abaxially. In most other respects—shape of the perianth tube and lobes, staminodes, stamens, and ovary, and the relative proportion of these floral parts—B. ste/laris is very different (unpublished data). The following key to the species of section “Stellares” serves to differentiate between the species. la. Staminodes hooded at the tips, the margins not or only slightly incurving, connate at the base with the stamens; filaments with prom- PRESTON: BRODIAEA MINOR REVISED ae, inent apical appendages abaxially; North Coast Ranges, on serpentine ........6.4. B. stellaris Ib. Staminodes not hooded at the tips, the margins incurving to strongly inrolled, not connate at base with stamens; lateral mar- gins of filaments winged, but appendages lacking. 2a. Perianth tube not narrowed above the ovary; filaments dilated at base; style shorter than ovary; southern Sierra Nevada foothills ee nit fee, vos RS Saale tee Gd teh Gp eee eee B. insignis 2b. Perianth tube narrowed above the ovary; filaments not or only slightly broader at base than at apex; style longer than the ovary. 3a. Staminodes erect to spreading, margins strongly inrolled; stamens narrowly notched at apex, lacking prominent papillae abaxially; filaments winged laterally, T-shaped in cross- SechlON: .¢ 6 Fd bo he fe oe B. minor 3b. Staminodes erect, margins not to. slightly inrolled; stamens broadly V-shaped at apex, with prominent abaxial papillae; filaments winged abaxially, V- or Y-shaped in cross section. 4a. Perianth lobes paler towards the base; perianth tube slightly narrowed above the ovary; staminodes as broad as or broader than the outer perianth lobes; central Sierra Nevada foothills, in swale, serpentine soil . . ate aie ta Aad a EO, AE te RAN fake Ae een de tay B. pallida 4b. Perianth color uniform; perianth tube strong- ly narrowed above the ovary; staminodes narrower than outer perianth lobes; eastern edge of Central Valley, in vernal pools and SWAleSt uy act Mate ares at tase pd he he B. nana Questions for Further Study Although this study may have resolved the taxonomy of B. minor and B. nana, many phylogenetic questions remain. Brodiaea minor consists of populations of both tetraploids and octaploids. Niehaus (1971) postulated that B. minor was derived from diploid B. nana. Whether B. minor was derived via autopolyploidy or allopolyploidy is unclear. Whether octaploid B. minor was derived from B. nana or from tetraploid B. minor is also unclear. Moreover, the possibility exists that B. minor is a complex of polyploid populations of multiple origins, rather than a tetraploid lineage and an _ octaploid lineage. The relationships between B. pallida, B. minor, and B. nana and other Brodiaea species remain uncertain. Reliance on morphological data alone has proved of limited usefulness in resolving relationships between and among Brodiaea spe- cies. Although groups of species can be recog- nized on the basis of unique characters, the phylogenetic relationships among the groups are still ambiguous. Niehaus (1971) provided some cytological, anatomical, and flavonoid data that may provide evidence for elucidating relation- ships, but little has been done to follow up on Niehaus’s work. Niehaus’s suggestion that eco- 54 MADRONO logical data might be useful has also not been pursued. Recent studies based on molecular data have proved useful for understanding relationships within the Themidaceae and may point a way towards resolving species relationships within Brodiaea (Pires and Sytsma 2002). Independent data sets derived from molecular data may help determine which morphological characters are plesiomorphic, which are derived, and which, like the “winged” filaments of B. stellaris and B. pallida, may be homoplasic. Molecular data may also be useful for differentiating between entities that have been derived via autopolyploidy or allopolyploidy (Rieseberg and Ellestrand 1993). Brodiaea remains a nearly untapped source for investigations on polyploidy, hybridization, and edaphic relationships. ACKNOWLEDGMENTS I thank Chris Pires for helpful discussions of species relationships in Brodiaea and for providing excellent digital photos of the Brodiaea types; Dale McNeal and an anonymous reviewer for providing constructive criticism of the draft manuscript; and Jones & Stokes for support of the manuscript preparation. LITERATURE CITED ANGIOSPERM PHYLOGENY GROUP. 2003. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants; APG IT. Botanical Journal of the Linnean Society 141:399_436. BURBANCK, M. P. 1941. Cytological and taxonomic studies in the genus Brodiaea. Botanical Gazette 103:247-265. DOALSON, M. C. 1999. Morphological variation and reproductive biology of a native California geo- phyte, Brodiaea californica (Liliaceae). M.S. thesis. California State University, Chico, CA. EASTWOOD, A. 1896. Description of some new species of California plants. Proceedings of the California Academy of Sciences, 2°" Series 6:422—430. FAY, M. F. AND M. W. CHASE. 1996. Resurrection of Themidaceae Salisb. for the Brodiaea alliance, and recircumscription of the Alliaceae, Amaryllidaceae and Agapanthoideae. Taxon 45:441-451. , P. J. RUDALL, S. SULLIVAN, K. L. STOBART, A. Y. DE BRUUN, F. QAMARUZ-ZAMAN, W. P. HONG, J. JOSEPH, W. J. HAHN, J. G. CONRAN, AND M. W. CHASE. 2000. Phylogenetic studies of Asparagales based on four plastid DNA _ loci. Pp. 360-371 in K. L. Wilson and D. A. Morrison (eds.), Monocots: systematics and evolution. CSIRO, Collingwood, Victoria, Australia. [Vol. 53 GREENE, E. L. 1886. Studies in the botany of California and parts adjacent. I. Some genera which have been confused under the name Brodiaea. Bulletin of the California Academy of Sciences 2:125—144. HARTWEG, T. 1848. Journal of a mission to California in search of plants. Journal of the Horticultural Society of London 3:217—-228. HOOVER, R. F. 1936. New California plants. Leaflets of Western Botany 1:225—230. 1938. New California plants. Leaflets of Western Botany 2(8): 128-133. . 1939. A revision of the genus Brodiaea. American Midland Naturalist 22:551—574. JEPSON, W. L. 1922. A flora of California, Vol. 1, pt. 6, Cyperaceae (Eleocharis) to Orchidaceae. Associat- ed Students Store. University of California, Berke- ley, CA. KEATOR, G. 1967. Ecological and taxonomic studies of the genus Dichelostemma. Ph.D. dissertation. Uni- versity of Californica, Berkeley, CA. —_—., 1989. The brodiaeas. Four Seasons 8:4—11. . 1993. Brodiaea. Pp. 1180-1183 in J. C. Hick- man (ed.), The Jepson manual: higher plants of California. University of California Press, Berke- ley, CA. NIEHAUS, T. F. 1971. A biosystematic study of the genus Brodiaea (Amaryllidaceae). University of California Publications in Botany 60:1—67. . 1980. The Brodiaea complex. Four Seasons 6(1): 11-21. OSWALD, V. H. 1994. Manual of the vascular plants of Butte County, California. California Native Plant Society, Sacramento, CA. PIRES, J. C. 2002. Brodiaea. Pp. 321-328 in Flora of North America Editorial Committee (eds.), Flora of North America north of Mexico, Vol. 26. Liliidae: Liliales and Orchidales. Oxford University Press, New York, NY. , M. F. FAy, W. S. DAvVis, L. HUFFORD, J. Rova, M. W. CHASE, AND K. J. SYTSMA. 2001. Molecular and phylogenetic analyses of Themidaceae (Asparagales). Kew Bulletin 56: 601-626. AND K. J. SYTSMA. 2002. A_ phylogenetic evaluation of a biosystematic framework: Brodiaea and related petaloid monocots (Themidaceae). American Journal of Botany 89:1342—1359. RIESEBERG, L. H. AND N. C. ELLESTRAND. 1993. What | can molecular and morphological markers tell us about plant hybridization? Critical Reviews in | Plant Sciences 12:213—241. | SWARTZLOW, R. J. 1964. Lassen: his life and legacy. — Loomis Museum Association, Lassen National | Park, Mineral, CA. | WATSON, S. 1879. Revision of the North American © Liliaceae. Proceedings of the American Academy | of Arts and Sciences 14:213—288. -MADRONO, Vol. 53, No. 1, pp. 55-59, 2006 FACTORS AFFECTING UNDERSTORY ESTABLISHMENT IN COASTAL SAGE SCRUB RESTORATION MATT V. TALLUTO, KATHARINE NASH SUDING, AND PETER A. BOWLER Department of Ecology and Evolutionary Biology, University of California, Irvine, CA 92697-2525 mtalluto@uci.edu ABSTRACT Coastal sage scrub (CSS) is a target for restoration because it provides habitat for numerous special- status species and it has been impacted by urbanization, agriculture and invasion by non-native species. Many restoration designs have neglected the herbaceous understory component of CSS, although it may comprise the majority of vascular plant species in a natural CSS stand. The omission of an understory may promote invasion by non-native plants and reduce overall success. This study investigated the role of native seed addition, non-native species removal, gaps in the shrub canopy, and soil moisture, upon establishment of a native understory. Native biomass increased significantly with seed addition, and the abundance of experimentally seeded native species was positively correlated with soil moisture. Natives were not affected by competition with non-natives or the presence of gaps. Although all seeded native species germinated, only two of seven established successfully, perhaps due to very low rainfall. Non-native species were negatively affected by the addition of native seeds and had greater growth in gaps. We conclude that planting shrubs in a dense configuration to reduce gap size may reduce non-native species abundance in the understory while having little effect on the native understory. Seeding may be all that 1s required to establish a native understory, and may also be an effective method of suppressing non-native species. Key Words: coastal sage scrub, restoration, competition, invasive species, seed limitation. Coastal sage scrub (CSS) is one of the most endangered habitats in southern California. Estimates of CSS loss vary, however it is likely that CSS currently occupies less than half of its historical distribution (Westman 1981; O’Leary 1995). Primary causes of CSS loss are urbaniza- tion, agriculture, and the degradation and re- placement of natural stands by the invasion of non-native species, mostly annual grasses from the Mediterranean region (Freudenberger et al. 1987; O’Leary and Westman 1988; O’Leary 1995; Minnich and Dezzani 1998). These ecosystem stressors, as well as the importance of CSS as habitat for numerous rare, threatened, or endan- gered plants and animals, make the community a priority for conservation (Davis et al. 1994). Regional multi species habitat conservation plans throughout southern California anticipate addi- tional loss of CSS and require both preservation and restoration of CSS to mitigate for this loss. Thus, development is resulting in increasing numbers of CSS restoration sites, making suc- cessful restoration strategies essential for effective CSS conservation. Typical CSS has a dense shrub canopy 0.5— 1.5m in height and a sparse herbaceous un- derstory of primarily annual species concentrated in gaps between shrubs. Invasive non-native Species are increasingly common in CSS unders- tories (O’Leary 1995). Non-natives are undesir- able in restoration projects because they reduce the success of planted shrubs (Eliason and Allen 1997) and compete with native understory herbs. Although many restoration designs target only the native shrub canopy, restoration of the understory as well is a more logical approach (Bowler 2000). Thus, understanding the ecolog- ical factors controlling understory structure and composition 1s important. Previous studies have suggested numerous mechanisms involved in competition between non-native and native plant species. Competition with non-native annual grasses for soil moisture often limits the success of native perennials (Melgoza et al. 1990; Eliaason and Allen 1997; Humphrey and Schupp 2004; but see Seabloom et al. 2003a). Competition with non-natives for soil moisture could also have an impact upon native understory annuals, although this has not been directly tested in CSS. Reduced light availability beneath shrubs may also reduce understory growth, resulting in a relatively higher density of understory plants in gaps, and shading from non-native grasses growing in gaps may affect these native plants (Thompson and Harper 1988; Dyer and Rice 1999). Coastal sage scrub understories share many species with California’s coastal grasslands. In two coastal grassland experiments in California, Seabloom and colleagues found that native perennial grasses (2003a) and annual forbs (2003b) were strongly seed-limited. Following seed addition, natives successfully established despite competition from non-natives. These 56 TABLE |. MADRONO [Vol. 53 SEEDING DENSITY, % EMERGENCE, AND RELATIVE SUCCESS OF NATIVE ANNUAL PLANTS ADDED. Biomass values (+1 SE) were calculated from seed addition plots only. Seed emergence data were obtained from S&S Seeds (Carpinteria, CA). Species N seeds added/plot Amsinckia menziesii (Boraginaceae) 280 Cryptantha muricata (Boraginaceae) 990 Lasthenia californica (Asteraceae) 6240 Lepidium nitidum (Brassicaceae) 970 Plantago erecta (Plantaginaceae) 360 Lupinus bicolor (Fabaceae) 190 Lupinus truncatus (Fabaceae) 60 results suggest that seed limitation also could explain the failure of many restored CSS communities to develop a native understory. This experiment’s objective was to identify the primary factors limiting establishment of native understory herbs during CSS restoration. Limi- tation likely results from several interacting factors. Through an experimental restoration, we addressed the following four questions: |. Is seed addition alone sufficient to restore a CSS native understory? 2. Do non-native and native understory plants compete with one another? 3. Does the reduced light environment beneath mature shrubs reduce the growth of understory plants? 4. Does soil moisture affect native understory establishment and competition with non-natives? METHODS Site Description This experiment was conducted on an existing CSS restoration site adjacent to the University of California, Irvine Arboretum and UC Natural Reserve System’s San Joaquin Freshwater Marsh Reserve. Prior to restoration in 2002, this site was an abandoned agricultural field that had no resident native taxa and was dominated by non-native species including Brassica nigra (Brassicaceae), Foeniculum vulgare (Apiaceae), Cynara cardunculus (Asteraceae), and annual grasses, primarily of the genus Bromus. (Nomen- clature follows Hickman 1993) These same species dominated the area surrounding the site. Annual rainfall in the area is approximately 300 mm yr”. During October-November 2002, Artemisia californica (Asteraceae) shrubs were planted in circular clusters approximately 3 m in diameter. Shrubs were planted at an average density of 2.5 plantings m *. To minimize mortality during establishment, plantings were watered weekly until February 2003, and were hand-weeded during the first spring (2003). No herbicides were applied at any time during restoration. Mean end of season % Seed Emergence biomass (g m °*) 54 0.38 + 0.15 26 0.00 80 8.03022 2:51 >I 0.00 88 0.00 85 0.00 91 0.00 Experimental Design This experiment was conducted during Febru- ary—May 2004, 1.5 years after the A. californica clusters were planted. Native propagule abun- dance was manipulated by adding native seeds of five native annual forbs and two annual legumes (Table 1). Presence of non-native species was manipulated by clipping all non-native species at ground level. Seed addition and non-native removal were combined in all possible combina- tions, resulting in the following four plot types: 1. Native seeds added, non-native species removed; 2. Native seeds added, no removal; 3. No seed addition, non-native species removed; and 4. No seed addition and no removal. Experimental plots were situated a minimum of 25cm from the edge of each A. californica cluster and from other plots. The plots were 50cm xX 50cm, with treatments extending an additional 10 cm beyond the plot boundary. Plot locations were assigned randomly within eight replicate blocks, each of which was located within a single cluster of shrubs and included all four treatment combinations. Within blocks, plot orientations were assigned non-randomly to include a single shrub immediately outside the plot at one end and a gap in the shrub canopy (no shrub canopy directly overhead) at the other end. This plot orientation allowed investigation of the importance of gaps in the distribution of un- derstory herbs beneath a CSS canopy. Each plot was divided into two 25cm X 50cm subplots: one under the canopy and one in the adjacent gap. Experimental Treatments Seed addition. We amended plots with seed from five native annual forb and two annual | legume species at a density of 4.5 g m * per taxon (Table 1). All seeds were obtained commercially from S&S Seeds (Carpinteria, CA). Plots were seeded on 19 February 2004. Prior to seed | addition, the soil was disturbed by hand-raking | to a depth of 2 cm. Control plots were similarly | disturbed. After seed addition, the loosened soil | 2006] TALLUTO ET AL.: COASTAL SAGE SCRUB UNDERSTORY RESTORATION 57 9 B. 14 Native 8 ie 7, [] Non-native a_ 10 6 iS = 8g 5 0) 4 2 6 5 3 a : 2 2 1 0 0 Control S SR R Canopy Gap Fic. 1. Biomass of native (shaded bars) and non-native (open bars) species in (A): plots with native seed addition only (S), seed addition and non-native species removal (SR), non-native species removal only (R), and no treatment (control), and (B): subplots beneath the shrub canopy and in gaps. In control and R plots, Artemisia californica was the only native present. In S and SR plots, most of the native biomass was from Lasthenia californica. Bars are means + | SE. was spread over the seeds to minimize losses from wind, runoff, and predation. Because rainfall was below average during the experiment, supple- mental water was provided in late March and early April to prevent excessive mortality. Seeded species were harvested on 12 May 2004, at the end of the growing season. To determine if seed addition had an effect on any preexisting native understory, all volunteer native species were collected as well. Aboveground biomass was dried for 48 hr at 60°C and weighed. Non-native species removal. All non-native species were removed from weeded plots by clipping at ground level. Early germinating species were removed on 9 March, as soon as plants were identifiable. A second removal was performed on | April to remove later germinating species. No non-native species were observed in any weeded plots after the second removal. Non- natives from unweeded plots were harvested, dried, and weighed with native species at the end of the growing season. Soil moisture and light availability. Photosyn- thetically active radiation (PAR) was measured in each subplot within 2 hr of solar noon on 29 April. Values were recorded as the ratio of light measured below the canopy to incident light measured directly above the canopy. Soil mois- ture in the top 12 cm of soil was also recorded using a Hydrosense TDR probe (Campbell Scientific, Logan, UT). Deeper readings were not taken to avoid excessive disturbance within the plots and because the rocky soil made probe insertion difficult. Three soil moisture readings were taken from each subplot, and the results were averaged prior to analysis. Soil moisture was recorded on 11 May, one day prior to harvesting all aboveground biomass in the plots. Statistical Procedures To meet assumptions of normality and homo- geneity of variance, a natural log transformation on all biomass measurements was performed, and soil moisture measurements were rank _ trans- formed. Because light intensity measurements were normally distributed, they were not trans- formed. To determine how shrubs influenced light and soil moisture, a paired t-test was conducted to compare light intensity between canopy and gap subplots, and a Wilcoxon signed rank test was used to test for a soil moisture difference between subplot types. We used analysis of covariance to examine effects of seed addition, non-native removal, light, and soil moisture on the biomass of native herbs. A similar analysis was used to test effects of native seed addition, light, and soil moisture on non-native biomass within non-removal plots. RESULTS AND DISCUSSION Native Seed Addition Native seed addition significantly increased native understory biomass. Native biomass in seed addition plots averaged 8.41 + 2.45 gm °, compared to 0.03 + 0.02 gm ° in control plots (Fi so = 310, P < 0.001; Fig. la). Natives averaged 2.6% of total herbaceous biomass in 58 MADRONO unseeded, unweeded plots and 88.5% in seeded, unweeded plots. Although all seven seeded species were ob- served germinating, only Lasthenia californica (Asteraceae) and Amsinkia menziesii (Boragina- ceae) survived long enough to produce flowers (Table 1). Lasthenia californica produced 98% of native biomass, while A. menziesii produced the remaining 2%. The only volunteer natives ob- served were A. californica seedlings. No. substantial populations of native CSS herbs were observed within several hundred meters of the study site, and it is likely that seed dispersal beyond this distance is very low (Van Dorp et al. 1996; Jongejans and Schippers 1999). Furthermore, the absence of any native herba- ceous growth in unseeded plots suggests that no native seed bank remained on the site. Past disturbance and several decades of non-native species dominance has likely eliminated any native seed bank that may have been present. Because an impoverished native seed bank and few local seed sources may be typical of many potential CSS restoration sites, seed addition of understory herbaceous plants is likely to be a necessary component of many restoration projects. Competition In unweeded plots, native seed addition caused a small but significant decline in non-native biomass, from 1.68 gm ~~ to 1.19 gm? (Fj 29 = 4.76, P = 0.04; Fig la). The small magnitude of this change is likely due to low non-native biomass throughout the planted area. The high biomass of L. californica in seeded plots suggests that this species, if seeded at high densities, may be effective at controlling non-native species. There were no significant effects of non-native removal on native biomass, regardless of seeding treatment. However, due to low non-native biomass and low survival of most seeded native species, it is difficult to make any conclusions about a competitive effect of non-native species on natives. Strong competitive effects of non- native species in CSS and similar systems have been demonstrated in the past, and may have been observed in this system had non-natives been more abundant (D’Antonio and Vitousek 1992: Eliason and Allen 1997; but see Seabloom et al. 2003b). Shrub Effects As expected, light intensity was significantly lower beneath the canopy (mean 47 + 3.0%) than it was in gaps (74 + 3.0%). There was no significant difference in soil moisture between canopy and gap subplots. Native understory biomass did not differ significantly between the [Vol. 53 shrub canopied and gap plots, suggesting that any competitive effect of shrubs on the natives may have been balanced by a facilitative effect. Non-native biomass was significantly less be- neath the canopy (Fj. = 5.80, P = 0.03; Fig. 1b). This was likely due to the large reduction in light intensity beneath the canopy, although it may also be due to belowground effects not measured in this study. Soil Moisture There was a significant positive relationship between soil moisture and native biomass (Fj 59 = 5.54, P = 0.02), but not between soil moisture and non-native biomass. Although soil moisture was limiting for both natives, the importance of soil moisture may have been exaggerated by the very low precipitation throughout the growing season. Low soil moisture may also explain the high mortality observed for most seeded native species. These species were observed germinating, but died during mid-season when rainfall also declined sharply below weekly averages. Finally, the lack of a difference in native growth between low light (canopied) and high light (gap) subplots would be expected if soil moisture was more limiting than light, and shading by shrubs reduced water stress for the native understory. Conclusions This study suggests that in restored dense- | canopy CSS, soil moisture is an important limiting factor for native understory species, while light is more limiting for non-natives. — Non-natives were also more abundant in gaps than they were beneath the shrub canopy, suggesting a competitive effect of the shrubs. Thus, the presence of a mature, dense shrub layer may effectively exclude a large fraction of non- native species without adversely affecting some native species. Competition with the restored native under- story (which consisted primarily of L. californica) significantly reduced non-native biomass. Thus, the addition of seeded native understory herbs may be an effective secondary restoration strat- egy, particularly given that long distance dispers- al of native understory herbs is likely to be a rare event. Seeding may be accomplished at a fairly | low cost and can successfully establish some native herbs and reduce the abundance of non- | native species. ACKNOWLEDGEMENTS We thank the University of California Natural | Reserve System’s San Joaquin Freshwater Marsh Re- serve for permitting us to conduct research on lands it manages, and for letting us use its equipment. The UCI Arboretum also provided equipment used in this study. 2006] Comments and field assistance from Bradford Haw- kins, Erin Hayes, Emily Grman, and Robert Arkle, and comments from two anonymous reviewers, greatly improved this manuscript. This study was supported in part by Grant 01-130 from the California Coastal Conservancy. LITERATURE CITED BowLeR, P. A. 2000. Ecological restoration of coastal sage scrub and its potential role in habitat conservation plans. Environmental Management 26:S85—S96. 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 Systematics 23:63—87. DAVIS, F. W., P. A. STINE, AND D. M. StToms. 1994. Distribution and conservation status of coastal sage scrub in southwestern California. Journal of Vegetation Science 5:743—756. Dyer, A. D. AND K. J. Rice. 1999. Effects of competition on resource availability and growth of a California bunchgrass. Ecology 80:2697—2710. ELIASON, S. A. AND E. B. ALLEN. 1997. Exotic grass competition in suppressing native shrubland _ re- establishment. Restoration Ecology 5:245—255. FREUDENBERGER, D. O., B. E. FISH, AND J. E. KEELEY. 1987. Distribution and stability of grass- lands in the Los Angeles basin. Bulletin of the Southern California Academy of Sciences 86:13—26. HICKMAN, J. C. (ed.). 1993. The Jepson manual: higher plants of California. University of California Press, Berkeley, CA. HUMPHREY, L. D. AND E. W. SCHUupP. 2004. Compe- tition as a barrier to establishment of a native perennial grass (Elymus elymoides) in alien annual grass (Bromus tectorum) communities. Journal of Arid Environments 58:405—422. TALLUTO ET AL.: COASTAL SAGE SCRUB UNDERSTORY RESTORATION 59 JONGEJANS, E. AND P. SCHIPPERS. 1999. Modeling seed dispersal by wind in herbaceous species. Oikos 87:362-372. MELGOZA, G., R.S. NOWAK, AND R. J. TAUSCH. 1990. Soil water exploitation after fire: competition between Bromus tectorum (cheatgrass) and two native species. Oecologia 83:7—13. MINNICH, R. A. AND R. J. DEZZANI. 1998. Historical decline of coastal sage scrub in the Riverside-Perris plain, California. Western Birds 29:366—391. O’LEARY, J. F. 1995. Coastal sage scrub: threats and current status. Fremontia 23(4): 27-31. AND W. E. WESTMAN. 1988. Regional distur- bance and effects on herb succession patterns in coastal sage scrub. Journal of Biogeography 15:775-786. SEABLOOM, E. W., W. S. HARPOLE, O. J. REICHMAN, AND D. TILMAN. 2003a. Invasion, competitive dominance, and resource use by exotic and native California grassland species. Proceedings of the National Academy of Sciences 100:13384— 13389. , E. T. BORER, V. L. BOUCHER, R. S. BURTON, K. L. COTTINGHAM, L. GOLDWASSER, W. K. GRAM, B. E. KENDALL, AND F. MICHELI. 2003b. Competition, seed limitation, disturbance, and reestablishment of California native annual forbs. Ecological Applications 13:575—592. THOMPSON, L. AND J. HARPER. 1988. The effect of grasses on the quality of transmitted radiation and its influence on the growth of white clover 7vi- folium repens. Oecologia 75:343—347. VAN Dorp, D., W. P. M VAN DEN HOEK, AND C. DALEBOUDT. 1996. Seed dispersal capacity of six perennial grassland species measured in a wind tunnel at varying wind speed and height. Canadian Journal of Botany 74:1956—-1963. WESTMAN, W. E. 1981. Diversity relations and succession in Californian coastal sage scrub. Ecology 62:170—184. MADRONO, Vol. 53, No. 1, pp. 60-64, 2006 OBSERVATIONS OF FROND GROWTH AND DEVELOPMENT IN PENTAGRAMMA TRIANGULARIS SUBSP. TRIANGULARIS (PTERIDACEAE) OF SOUTHERN CALIFORNIA GARY B. PERLMUTTER North Carolina Department of Environment and Natural Resources, Raleigh Regional Office, 3800 Barrett Drive, Raleigh, NC 27609 gary.perlmutter@ncmail.net ABSTRACT Frond growth and development of Pentagramma triangularis subsp. triangularis was examined in canyon populations in Ventura County, California during the winter growing season of 2003. Fronds developed from emerging croziers to full size within 35 days and reached sexual maturity by 70 days. Frond growth occured in two stages with stipe elongation nearly complete when the crozier began to unfurl, about a week before full maturation of the entire frond. Discernable developmental stages are described as well as their respective ages. Fern growth occurred both prior to and after the study frond monitoring period, indicating multiple growth flushes throughout the growing season, possibly related to rain events. Based on these preliminary results, suggestions are offered for further study. Key Words: frond growth, Pentagramma triangularis, phenology, Pteridaceae, southern California. Pentagramma_ triangularis (Kaulf.) Yatsk., Windham & E.Wollenw. (Pteridaceae) grows in western North America (Yatskievych and Wind- ham 1993) inhabiting rock crevices within a var- iety of habitats including woodlands, chaparral and deserts (Smith 1998). Plants of this species complex, commonly called goldback and silver- back ferns, produce 10-35 fronds (7—40 cm) from a short, slender rhizome. Each frond consists of a brown to black stipe and a deltate- pentagonal, 1—2 pinnate-pinnatifid blade as wide as long (4-9 cm). The abaxial surface of blades are covered with silver or gold farina, hence their common names. Plants are winter- green perennials producing fronds at first rains, which persist until hot weather, when their blades curl due to drought (Yatskievych et al. 1990; Smith 1998). The ecology and life history of these ferns are largely unexplored. Except for having a growing season that coincides with winter rains, frond phenology of P. triangularis is largely undocu- mented. This lack of knowledge regarding frond phenology is representative of the pteridophytes in general, with few studies reported in the literature (e.g., Sato 1982; Willmot 1989; Sharpe and Jernstedt 1990; Johnson-Groh and Lee 2002). The objectives of this study were to measure the growth and to describe the developmental stages of P. triangularis subsp. triangularis fronds. This report presents the first detailed phenolog- ical description for Pentagramma and_ lays a foundation for further study of P. triangularis life history, which could provide a better un- derstanding of this fern’s adaptation to its seasonally dry habitat. METHODS Pentagramma_ triangularis subsp. triangularis ferns were monitored in two small populations along the Cozy Dell Trail in Sheldon Canyon, Ventura County, California (34°28'39’N, 119°17’09"W). The study area, located 22 km from the coast at an elevation of 427 m, is in the Transition Climate Zone and is influenced by a mixture of maritime and continental air masses with hot, dry summers and mild, rainy winters (Bailey 1966; Hickman 1993). Total rainfall for the 2002-2003 rainy season (from October through April) in Stewart Canyon, 3.2 km ESE of the site, was 454 mm with most falling in November, December and March. Temperatures taken at Matilya Dam, 2.4 km WNW of the study site for the same period, included a winter minimum of 6.3°C in December and _ spring maximum of 22.8°C (Ventura County Watershed Protection District, unpublished data). However, this rainy season was atypical in being warm, with a dry spell in January (Fig. 1). The study area is a steep, north-facing slope with cobble and boulder talus of sedimentary rocks (Norris and Webb 1990) in coastal oak woodland/mesic chaparral dominated by a patchy canopy of Quercus agrifolia Née (Fagaceae) with a dense shrub layer of Ceanothus cuneatus (Hook.) Nutt. (Rhamnaceae) and Heteromeles | arbutifolia (Lindl.) M. Roem. (Rosaceae). Un- derstory seed plants included Toxicodendron diversilobum (Torr. & A. Gray) Greene (Anacar- | diaceae), Diplacus (= Mimulus) longiflorus Nutt. | (Phrymaceae), and undetermined grasses and forbs. Other ferns in this area included Poly- podium glycyrrhiza D.C. Eaton (Polypodiaceae), 2006] -_ O —_ co) = =) — i») en o ok = ) — PERLMUTTER: PENTAGRAMMA FROND DEVELOPMENT 61 140 14120 100 80 60 40 20 0 Rainfall (mm) Oct Nov Dec Jan Feb Mar Apr May Jun FIG. 1. Rainfall and temperature patterns for the October 2002—April 2003 rainy season, extended to June 2003, at Stewart Canyon and Matilijja Dam meteorological stations, respectively (data source: Ventura County Watershed Protection District). Pallea andromediflora (Kaulf.) Fée (Pteridaceae), and Dryopteris arguta (Kaulf.) Maxon (Dryop- teridaceae). Exposed rock surfaces harbored Selaginella bigelovii Underw. (Selaginellaceae) plus foliose and crustose lichens. The study populations, separated by 0.2 km, consisted of 24 and 25 ferns each, with densities of approximately 1.6 and 6.3 ferns m °, re- spectively (densities appeared determined by amount of rocky edge microhabitat). Ten ferns per population were haphazardly selected for study. Data from the two populations were lumped for analysis because no _ significant differences were observed between them. On 5 January 2003, I selected for study the crosier on each fern that was closest to the soil surface (=10 cm height), and thus recently emerged, and marked its stipe with correction fluid for growth monitoring. Weekly measure- ments of marked fronds included stipe and rachis lengths (while in crozier, head diameter was measured to represent this variable); for de- velopmental stage classification, the frond’s de- velopment (i.e., shape, color, farina, and presence of sporangia) was also described. Lengths were measured until fronds had reached their full size (i.e., no increase for three consecutive measure- ments), but developmental observations contin- ued until the frond senesced. Three fronds failed to grow and develop completely, and for these a replacement crosier on the fern was marked and monitored. Two of these were on one plant that senesced prematurely, in March; this fern was on an exposed rockface. Growth curves were developed from the stipe and rachis length means. Other phenological observations of these plants were also made at irregular intervals, including after the period of frond growth monitoring. Voucher specimens were deposited at Santa Barbara Botanic Garden Herbarium (SBBG). RESULTS Frond Growth and Development From 5 January through 26 March 2003 development of study fronds passed through seven discernable stages from crozier to mature sporophyll. Growth curves are shown in Figure 2, while the mean (+1 SD) frond sizes and their respective percent mature sizes for each develop- mental stage are listed in Table | and included in the text below. Frond stages are described below. (1) Young Crozier (YC): Fronds first appeared light green, with stipe lengths from emer- gence (0 cm) to 5.5 cm. Crozier diameters averaged 0.32 =: 0.11 cm, (2) Mature Crozier (MC): Stipe lengths aver- aged 9.3 + 3.7 cm or 64 + 16% of the full length and darkened from the base toward the tip. Croziers grew to 0.44 + 0.14 cm diameter with pinnae visible. Fronds were in this stage for 7-14 days. (3) Unfurling Frond (UF): Blade development began with croziers opening at 14—28 days after emergence, lasting seven days or less. Stipes elongated on average to 94 + 4% of their mature length and rachises reached a meanvot 304 9.5% of mature size. 62 MADRONO [Vol. 53 Stipe Growth Length (cm) Rachis Growth Length (cm) Fic. 2. Mean (+ 1 SD) growth curves of frond stipes and rachises from 18 Pentagramma triangularis subsp. triangularis ferns monitored in Sheldon Canyon, California, USA during 05 January—16 March, 2003. (4) Immature Frond (IF): By 7-21 days, blade developed. Stipes reached their full blades were fully open at this stage length while rachises elongated to an with thin shiny, yellowish green pinnae average of 80 + 15% (range 53-100%) of and pinnules that expanded as the full size. TABLE 1. MEAN + 1 SD LENGTHS PLUS PERCENT OF MEAN MATURE SIZES OF PENTAGRAMMA TRIANGULARIS SUBSP. TRIANGULARIS FROND DEVELOPMENTAL STAGES. Observations made in Sheldon Canyon, Ventura County, California, USA, 05 January—16 March with Sporophyll Observations until 01 June 2003. Number of observations included multiple examinations of 18 monitored ferns (one frond per plant). Monitoring of stipe growth ceased once fronds reached full size. See text for stage definitions. Frond age Stipe Rachis Stage (Days) Length (cm) % Mature Length (cm) % Mature No. Observations YC O-7 3:3 = 1,6 26+ 9 0.3 2404 6+ 1 8 MC < 7-14 03. Se 64 + 16 0.4 + 0.1 roa 16 UF 14-28 13.1 + 4.3 O9<8256 20 Ici 30 2 9 8 IF 14-35 14.7. 2 3.5 100 wae Nieuste ie 81 + 16 25 MF 21-63 14.4 + 43 100 5971s 100 56 IS 63-77 N/A N/A N/A N/A 14 MS 70+ N/A N/A N/A N/A 48 | | 2006] (5) Mature Frond (MF): Blades developed a dull green color with light farina abaxially by 21-35 days. During frond maturation some stipes developed a bent stature. (6) Immature Sporophyll (IS): Sporangia were first observed at 56-70 days as a light brown dusting on the abaxial surface. (7) Mature Sporophyll (MS): Fronds appeared dark green adaxially; farina were mixed with dark brownish sporangia on_ the abaxial surface. Growth-monitored fronds were first observed in this stage by 63— Eiidays. Not all crosiers were of equal length at the start of monitoring, thereby lowering precision of this growth study. By 26 July, 175 days after moni- toring began, all study fronds had curled adaxially in senescence, exposing the farinose abaxial surface. Duration of frond (i.e., stipe and _ rachis) growth averaged 19.8 + 5.7 days. Stipe elonga- tion averaged 12.4 + 5,3 days, which was a significantly shorter growth period than for the frond entire (t-test, P = 0.0004). Growth duration of the rachis was equivalent to growth duration of the frond entire, because it included rachis growth while in the crosier head. Phenology At the onset of this study (5 January 2003), ferns had young as well as older fronds, the latter already in the mature sporophyll stage. On 15 December 2002, ferns in the area also were observed exhibiting the full range of frond developmental stages from young crosier to mature sporophyll. During this study, crosiers appeared on the study plants as late as 23 March, when most of the growth-monitored fronds were already in the mature sporophyll stage. Young crosier observations appeared to follow rain events of >50 mm within approximately 14 days, but data were insufficient for statistical analysis. By 26 July, however, all fronds on plants in the study area had senesced. DISCUSSION The growth of monitored fronds was complet- ed within 35 days; reproductive maturation of fronds was completed within 70 days. Frond development passed through a series of discern- able stages, which are useful for describing a frond’s growth and reproductive state. Ob- servations of older fronds at later stages growing _ concurrently with younger fronds indicated that frond production and growth occurs in successive flushes. Determining how many flushes of new fronds can occur within a season and how they _ are triggered will probably require observations PERLMUTTER: PENTAGRAMMA FROND DEVELOPMENT 63 over several years, especially considering the variable nature of the southern California climate (Nilsen and Muller 1981). While this note describes the frond growth, development and phenology in P. triangularis subsp. ¢triangularis, the results are based on observations at a single site during only part of one year; thus, further study is needed to confirm and elaborate on these findings. In particular, populations should be monitored for consecutive years, and observations of reproductive develop- ment should include microscopic examination of freshly collected sporophylls. Questions for fu- ture investigation include: 1) what environmental factors (e.g., rain, temperature, daylength) trigger onset of fern growth and how quickly do ferns respond to these factors; 2) do flushes of growth coincide with periodic rain events within a season, and if so, how many flushes can occur in a year; 3) how long is the growing season for this species, and how long do individual fronds live before senescence;, 4) what triggers sporulation, and how long is sporangial development prior to sporulation; 5) how does habitat and microhab- itat affect the phenology; and 6) how does the phenology of P. triangularis subsp. triangularis compare with that of co-occurring fern taxa. ACKNOWLEDGMENTS I thank Ruth Kirkpatrick, John C, Hunter and two anonymous reviewers for their helpful comments on earlier drafts of this manuscript. David Laak of the Ventura County Watershed Protection District pro- vided climatic data for the study period. LITERATURE CITED BAILEY, H. P. 1966. Weather of southern California. University of California Press, Berkeley, CA. HICKMAN, J. C. (ed.). 1993. The Jepson manual: higher plants of California. University of California Press, Berkeley, CA. JOHNSON-GROH, C. L. AND J. M. LEE. 2002. Phenol- ogy and demography of two species of Botrychium (Ophioglossaceae). American Journal of Botany 89:1624—-1633. NILSEN, E. T. AND W. H. MULLER. 1981. Phenology of the drought-deciduous shrub Lotus scoparius: climatic controls and adaptive significance. Eco- logical Monographs 51:323—341. Norris, R. M. AND R. T. WEBB. 1990. Geology of California, 2nd ed. John Wiley and Sons, Inc. New York, NY. SATO, T. 1982. Phenology and wintering capacity of sporophytes and gametophytes of ferns native to northern Japan. Oecologia 55:53-61. SHARPE, J. N. AND J, A. JERNSTEDT. 1990. Leaf growth and phenology of the dimorphic herbaceous fern Danea wendlandii (Marattiaceae) in a Costa Rican rain forest. American Journal of Botany 77:1040—-1049. SMITH, C. 1998. A Flora of the Santa Barbara region, California, 2nd ed. Santa Barbara Botanic Garden and Capra Press, Santa Barbara, CA. 64 MADRONO [Vol. 53 WILLMOT, A. 1989. The phenology of leaf life spans in woodland populations of the ferns Dryopteris filix-mas (L.) Schott and D. dilatata (Hoffm.) A. Gray in Derbyshire. Botanical Journal of the Linnean’ Society 99:387 Editorial Committee (ed.), Flora of North Amer- ica, north of Mexico, Vol. 2. Pteridophytes and gymnosperms. Oxford Univeristy Press, New York, NY. , AND E. WOLLENWEBER. 1990. A BoD: YATSKIEVYCH, G., M. D. WINDHAM. 1993. Penta- gramma. Pp. 149-151 in Flora of North America reconsideration of the genus Pityrogramma (Adian- taceae) in western North America. American Fern Journal 80:9-17. MADRONO, Vol. 53, No. 1, pp. 65—68, 2006 HABITAT, GEOLOGIC, AND SOIL CHARACTERISTICS OF SHASTA SNOW- WREATH (NEVIUSIA CLIFTONIT) POPULATIONS LEN LINDSTRAND III Fisheries/Wildlife Biologist, Terrestrial Biology Program Manager, North State Resources, Inc., 5000 Bechelli Lane, Suite 203 Redding, CA 96002 lindstrand@nsrnet.com JULIE KIERSTEAD NELSON Forest Botanist, Shasta-Trinity National Forest, 3644 Avtech Parkway, Redding, CA 96002 jknelson@fs.fed.us ABSTRACT Following our recent discovery of seven new Shasta snow-wreath populations during 2003 and 2004, we conducted a GIS analysis using location data from the new sites, and from all previously known sites to determine the habitat, geologic, and soil characteristics of each Shasta snow-wreath population location. Previously considered a species only associated with limestone substrates, our new information shows that 47% of all known Shasta snow-wreath sites occur on non-limestone geologic or soil types. Key Words: Shasta snow-wreath, Neviusia cliftonii, limestone, Shasta Lake, geology, soil, Rosaceae. Shasta snow-wreath (Neviusia cliftonii She- vock, Ertter & Taylor) is a recently discovered shrub of the Rosaceae: Kerrieae. It is endemic to northern California in the vicinity of Shasta Lake, Shasta County. The environmental conditions and geographic isolation of the species suggest that it is one of the remnant taxa of an old, formerly more widespread genus (Shevock et al. 1992). A fossil occurrence of a closely related plant in a Pacific Northwest Eocene flora supports this hypothesis (DeVore et al. 2004). The species’ type locality and the subsequent two populations discovered were on limestone substrates (Shevock et al. 1992; Taylor 1992; Shevock 1992). Shasta snow-wreath was there- fore thought to represent a limestone endemic species, and is described in The Jepson Manual (Hickman 1993) as a species occurring in habitats associated with limestone rock forma- tions. Its closest relative, Neviusia alabamensis Gray, of the southeastern U.S., commonly occurs on limestone, but is also found on sandstone and shale substrates (Yocom and Little 1975: Patrick et al. 1995). Subsequent observations of Shasta snow-wreath in the Shasta Lake area demon- Strate that it too occurs on soils of non-carbonate origin. PREVIOUSLY KNOWN OCCURRENCES The California Department of Fish and Game’s California Natural Diversity Database (CNDDB) and the California Native Plant Society Inventory of Rare and Endangered Plants contain records of ten known Shasta snow- wreath locations (excluding the new locations discussed in this article). Habitat information reported for these ten previously known loca- tions indicates that eight (80%) occur within habitats associated with limestone rock forma- tions. Most of those first ten Shasta snow- wreath discoveries were made during the year following the initial type locality discovery and publication of the species name, when efforts to find additional Shasta snow-wreath occurrences focused on other limestone areas near Shasta Lake as potentially suitable habitat (Shevock 1992). NEW OCCURRENCES During field investigations in 2003 and 2004, North State Resources, Inc. personnel discovered seven new Shasta snow-wreath locations in the vicinity of Shasta Lake (Lindstrand and Nelson 2004) (Fig. 1). Specifically, these discoveries occurred during vegetation and habitat mapping conducted along the margins of Shasta Lake, and were opportunistic discoveries, rather than results of a systematic plant survey. These new locations occur primarily along drainages in dense, shady montane hardwood-conifer and ponderosa pine forests, but also in open foothill pine-blue oak woodland habitat. The new snow- wreath populations range in size and aerial extent from several plants in a relatively small area, to extensive stands consisting of thousands of plants blanketing both sides of a stream corridor for at least a quarter mile. Of these seven new occurrences, only one is associated with limestone substrate. 66 MADRONO At” | 12215! Shasta Dam—~™ Fic. 1. ANALYSIS OF GEOLOGY AND SOIL CHARACTERISTICS A geographic information system (GIS) anal- ysis of all known Shasta snow-wreath occur- rences was performed in order to quantify the distribution of occurrences on carbonate and non-carbonate substrates. All ten previously documented occurrences listed in the CNDDB were extracted to a GIS shapefile as point data. We then converted our seven new occurrences to point data and added that data to the shapefile. The shapefile was intersected with a digital geologic map of the Shasta Lake area (USDA Forest Service 2004), and the distribution of Shasta snow-wreath occurrences by geologic map unit and general geologic type was de- termined. The Shasta snow-wreath occurrence points were then intersected with the digital Order 3 soil survey of the area (USDA Forest Service 1983), and the distribution of Shasta snow-wreath occurrences by Order 3 soil map unit was determined. Each new population was also field-checked for evidence of unmapped limestone outcrops, since there are occasional exposed limestone inclusions within the Shasta Lake area that are too small to be included in landscape-scale geology or soil maps. Shasta Snow-wreath Occurrences @ New Location A. Previously Recorded Location S 0 1 4 Kilometers Data courtesy of CA Dep't of Fish and Game (CNDDB), and North State Resources, Inc Known Shasta snow-wreath occurrences and distribution. Shasta County, California. RESULTS Nine of the 17 known Shasta snow-wreath occurrences intersected the mapped extent of limestone bedrock and are found within limestone geologic types (Table 1). The remaining eight occurrences are found within primarily metavol- canic and metasedimentary geologic types. Three of the 17 known Shasta snow-wreath occurrences intersected the mapped extent of Order 3 soil map units with limestone parent material (Table 2). However, six additional occurrences are located immediately adjacent to, or downstream of limestone outcrops; where there is a high likelihood that colluvial or fluvially transported limestone substrate is present; or are located within limestone inclusions occurring within the Order 3 soil mapping unit. Collective- ly, nine Shasta snow-wreath occurrences are either located within, or are immediately adjacent to soil types containing limestone parent materi- al. The remaining eight occurrences are found within primarily metasedimentary/metavolcanic Order 3 soil map units. DISCUSSION Of our seven new Shasta snow-wreath discov- ery locations, only one intersected the mapped 2006] LINDSTRAND AND NELSON: SHASTA SNOW-WREATH HABITAT 67 TABLE 1. DISTRIBUTION OF NEVIUSIA CLIFTONIT OCCURRENCES BY GEOLOGIC TyPE.'—Also contains limestone fragments and strata. Geologic No. Shasta snow- map unit Formation Rock type Age wreath occurences Cb Baird Metasedimentary Carboniferous 2 Cbmv Baird Metavolcanic Carboniferous l De Copley Greenstone Metavolcanic Devonian l Pmd Quartz Diorite — Dikes Intrusive Permian l Pmml McCloud Limestone Carbonaceous Permian l Pmn Nosoni Metasedimentary/ Permian l metavolcanic Trh Hosselkus Limestone Carbonaceous Triassic 4 Trm Modin Metavolcanic! Triassic 3 Trp Rit Metasedimentary Triassic 3 TABLE 2. DISTRIBUTION OF NEVIUSIA CLIFTONIT OCCURRENCES BY ORDER 3 SOIL TYPE. Order 3 soil No. Shasta snow- map unit Dominant soil type Dominant parent material wreath occurrences 102 Holland Family Metasedimentary/metavolcanic | 105 Holland Family Metasedimentary/metavolcanic 5 117 Holland Family, deep Metasedimentary/metavolcanic l 178 Marpa Family Metasedimentary/metavolcanic | 179 Marpa Family Metasedimentary/metavolcanic l 180 Marpa Family Metasedimentary/metavolcanic | 183 Marpa Family Metasedimentary/metavolcanic l 195 Millsholm Family Sedimentary l 204 Neuns Family Metasedimentary/metavolcanic l 222 Neuns Family Metasedimentary/metavolcanic I 250 Rock Outcrop, limestone Limestone 3 extent of limestone bedrock. With the addition of these seven new sites, there are now seventeen documented Shasta snow-wreath occurrences. Following our GIS analysis of the geologic and soil characteristics at each location, nine of the 17 Shasta snow-wreath occurrences intersect the mapped extent of limestone bedrock, or occur immediately adjacent to or downstream of limestone outcrops. The remaining eight loca- tions, including four of the five most extensive populations, are found in non-limestone habitats. These seven new Shasta snow-wreath discov- eries have nearly doubled the number of known occurrences. Additionally, the geology and soil type analysis of these locations show that nearly one-half (47%) of the known species locations occur in habitats not associated with limestone rock formations nor soils formed from limestone parent material. Though these new occurrences have filled some gaps in the known distribution, they are within the previously recorded species range. Most of the documented Shasta snow-wreath occurrences are located within the eastern half of the Shasta Lake region in the Pit River, Squaw Creek, and McCloud River drainages. One occurrence lies within the western half of the Shasta Lake region in the Sacramento River drainage. Given our analysis of the geologic and edaphic character- istics at the known Shasta snow-wreath sites, the previous assumptions regarding the species geo- logic and edaphic associations, and the fairly limited geographic extent of previous survey efforts, only a small fraction of potential habitat for this species has been surveyed, and it 1s highly likely that additional populations occur. The conservation implications from these new Shasta snow-wreath discoveries may be described as two-fold. These new populations, in terms of both numbers of known occurrences and more widespread geologic/edaphic associations, show that the species may not be as rare or narrowly distributed as initially thought. However, six of the known populations have already been at least partially flooded by the creation of Shasta Lake. Additionally, these new discoveries also show that the species is still likely a narrow endemic to the Shasta Lake region. Endemism within this region is already fairly well documented by the presence of several endemic plant and wildlife species including Shasta eupatory (Ageratina shastensis), Shasta salamander (Hydromantes shastae), Shasta sideband snail (Monadenia trog- lodytes troglodytes), and Wintu sideband snail (M.t. wintu), and is likely a function of geologic and climatic factors. The geology of this region 1s considered ancient, particularly relative to sur- rounding regions, and was not affected by 68 MADRONO glaciation, nor was overlain by volcanic material. Additionally, this region lies within an area of high annual precipitation at relatively low eleva- tions, producing a combination of mesic condi- tions and mild temperatures. These geologic and climatic factors in the Shasta Lake region result in conditions favorable for a diverse flora and fauna, including several endemic species. The boundaries of Shasta snow-wreath’s geo- graphic and elevational range have yet to be determined. It is clear from our analysis of geologic and edaphic characteristics at the known snow- wreath population sites that non-limestone sub- strates cannot be excluded as suitable habitat, and that field inventories for Shasta snow-wreath within the species known distribution should include a wider range of substrates, aspects, and vegetation types than was thought suitable for the species. ACKNOWLEDGEMENTS We thank the U.S. Bureau of Reclamation Mid- Pacific Region office for their support. LITERATURE CITED DEVORE, M. L., S. M. MOORE, K. B. PIGG, AND W. C. WEHR. 2004. Fossil Neviusia leaves (Rosaceae: Kerriae) from the lower-middle Eocene of southern British Columbia. Rhodora 106:197—209. [Vol. 53 HICKMAN, J. C. (ed.). 1993. The Jepson Manual: higher plants of California. University of California Press, Berkeley, CA. LINDSTRAND III, L. AND J. K. NELSON. 2005. Noteworthy Collections. Madrono 52:126—127. PATRICK, T. S., J. R. ALLISON, AND G. A. KRAKOw. 1995. Protected plants of Georgia: an information manual on plants designated by the State of Georgia as endangered, threatened, rare, or un- usual. Georgia Natural Heritage Program, Wildlife Resources Division, Georgia Department of Natural Resources, Social Circle, GA. SHEVOCK, J. R. 1992. How rare is Shasta snow-wreath? Fremontia 22(3): 7—10. B. ERTTER, AND D. W. TAYLOR. 1992. Neviusia cliftonii (Rosaceae: Kerriae), An intriguing new species from California. Novon: 285—289. TAYLOR, D. 1992. A new discovery in California. Fremontia 22(3): 3-4. USDA FoREsT SERVICE. 1983. Shasta-Trinity National Forest Order 3 Soil Survey, United States De- partment of Agriculture, Forest Service, Pacific Southwest Region, in cooperation with the Uni- versity of California Agricultural Experiment Station. . 2004. Northern Province bedrock geology layer (ArcInfo coverage prov6bedrk). Klamath National Forest, Yreka, CA. Yocom, H. A. AND E. L. LITTLE, JR. 1975. Research Note SO-211, Southern Forest Experiment Sta- tion, USDA Forest Service, Asheville, NC. | MADRONO, Vol. 53, No. 1, pp. 69-71, 2006 ADDITIONAL TAXONOMIC STUDIES OF ARCEUTHOBIUM PENDENS (VISCACEAE): A RARE DWARF MISTLETOE FROM CENTRAL MEXICO ROBERT L. MATHIASEN School of Forestry, Northern Arizona University, Flagstaff, AZ 86011 Robert.Mathiasen@nau.edu CAROLYN M. DAUGHERTY Department of Geography, Planning, and Recreation, Northern Arizona University, Flagstaff, AZ 86011 ABSTRACT Additional measurements of male and female plants and staminate flowers are reported for Arceuthobium pendens. Measurements for fruits and seeds of this rare mistletoe are reported for the first time. Additional information on phenology and host reactions to infection is reported also. RESUMEN Se reportan mediciones adicionales de plantas masculinas y femeninas y de flores masculinas de Arceuthobium pendens. Las mediciones de frutos y semillas de esta rara especie de muérdago enano son reportadas por primera vez. Informacion adicional sobre fluoracion y dispersion de semillas y reacciones del hospedero a la infecci6n son también reportadas. Key Words: pendent dwarf mistletoe, Arceuthobium pendens, Pinus cembroides subsp. orizabensis, Pinus discolor. Arceuthobium pendens Hawksworth & Wiens (Viscaceae) was described in 1980 from the Sierra San Miguelito in southwestern San Luis Potosi, Mexico (Hawksworth and Wiens 1980). This dwarf mistletoe has only been discovered in two additional populations much farther south in Mexico and in close proximity to each other: northwest of Perote near Frijol Colorado in western Vera Cruz and on Cerro Pizarro in northeastern Puebla (Chazaro and Olivia 1987; Hawksworth and Wiens 1996; Hawksworth et al. 2002). Because this dwarf mistletoe is only known from three populations, it is currently considered to be one of the rarest dwarf mistletoes in Mexico (Hawksworth and Wiens 1996). For this species, information on several mor- phological, phenological, and other attributes are limited or unavailable. When Hawksworth and Wiens (1980) described Arceuthobium pendens, they provided morphological data on the size and color of male and female plants and the size of staminate flowers. Because Hawksworth and Wiens’ original specimens collected in 1979 from San Luis Potosi, as well as specimens collected by D. K. Bailey and T. Wendt in 1980 from Vera Cruz, were collected in early March, Hawksworth and Wiens did not have specimens of male plants with open flowers or female plants with mature fruits and seeds (Hawksworth and Wiens 1980). However, they were able to obtain measurements of staminate flowers from specimens collected by M. F. Robert in September 1971 that had male plants with mature staminate flowers. Hence, they concluded that A. pendens probably flow- ered in September, but acknowledged that this was only an estimate based on the Robert specimens. Information on the period of seed dispersal for A. pendens reported by Hawksworth and Wiens (1996) was based on a report by Chazaro and Olivia (1987) who first reported the occurrence of A. pendens in Puebla (Cerro Pizarro) and indicated that it dispersed seed from June to September. Because of the small amount of data provided by Hawksworth and Wiens (1980, 1996) we collected many additional specimens of A. pendens from near Frijol Colorado in Vera Cruz in late July 2005. This allowed us to complete many addi- tional measurements of male and female plants. Since many male plants had begun flowering and several female plants had started seed dispersal, we were also able to measure male flowers and collect the first data on the size of mature fruits and seeds as well as estimate the periods of peak seed dispersal and flowering for A. pendens in Vera Cruz. These additional and new morpho- logical measurements allowed us to make further comparisons between A. pendens and A. divar- icatum Engelm., the only other dwarf mistletoe that parasitizes pinyons as principal hosts. A total of 60 infected branches with plants of A. pendens (30 males and 30 females) were collected from 30 severely infected Pinus orizabensis trees approximately 8 km north of Frijol Colorado along the road to Los Humeros in western Vera Cruz, Mexico (19°37’'N, 97°23'W, elev. 2740 m). 70 MADRONO One male and one female infection with mature plants were collected from the lower part of the crown of each tree. Infections were collected haphazardly in that the first observed and easily accessed male and female infection with mature plants was sampled. The approximate location of this population is illustrated in Fig. 16.85 of Hawksworth and Wiens (1996). The plants were placed in paper bags, and all measurements were made within 24 hr of collection. The dwarf mistletoe plant characters measured were pri- marily those used by Hawksworth and Wiens (1996) in their monograph of Arceuthobium. The following morphological characters were re- corded: 1) height and color of the tallest male and female plant from each infection collected; 2) mature fruit length, width, and color; 3) seed length, width, and color; 4) staminate flower diameter; 5) number, length, and width of staminate perianth lobes; 6) anther distance from the perianth lobe tip; 7) anther diameter; and 8) pre-flowering lateral staminate spike length and width. Dwarf mistletoe plants were measured using a Plasti-cal digital caliper accurate to 0.1 cm and all other measurements were made with a Bausch and Lomb 7X hand lens equipped with a micrometer accurate to 0.1 mm. Specimens of male and female plants of A. pendens collected from the Vera Cruz population have been deposited at the Deaver Herbarium (ASC), Northern Arizona University, Flagstaff, AZ. The heights of male and female plants (Table 1) were much greater than those reported by Hawks- worth and Wiens (1980, 1996) for plants of A. pendens collected in the Sierra San Miguelito of San Luis Potosi. Hawksworth and Wiens (1980, 1996) reported that the maximum plant size for male plants of A. pendens was 22 cm, but we measured male plants 32 cm in height and many male plants (30%) were greater than 25 cm in height. Furthermore, while Hawksworth and Wiens (1996) commented that few female plants of A. pendens were larger than 8 cm, we com- monly found female plants (50%) greater than 15 cm in height and the tallest female plant we measured was 26cm. While male and female plants were both predominantly light green, the base of several older plants was dark green, and for some of the largest male and female plants, the base was dark brown, indicating that as the plants age their bases gradually change color. Although Hawksworth and Wiens (1980) did not report how many plants of each sex they measured, they did comment that they only found two infected pinyons at the Sierra San Miguelito type locality in San Luis Potosi. Therefore, they may not have been able to measure many plants from that location. Because many trees were infected at the Vera Cruz location, we were able to measure a fairly large sample of male and female plants. Therefore, [Vol. 53 our measurements are probably more representa- tive of the mean sizes and ranges for male and female plants of A. pendens. Additional morpho- logical measurements should be completed from the third population from Cerro Pizarro in Puebla to see how they compare to the Vera Cruz and San Luis Potosi populations. ~The mean diameter of male flowers we obtained (2.6 mm) (Table 1) is approximately the same as that reported by Hawksworth and Wiens (2.5 mm). Most male flowers were 3-merous, but rarely 4- merous staminate flowers were observed on the terminal ends of staminate spikes. Our measurements for perianth lobes, anthers, and staminate spikes, mature fruits and seeds (Table 1) are the first reported for A. pendens. Most fruits were the same color as female plants, but some female plants had fruits that were dark green while the plants were light green. Seeds were light green and lacked the character- istic yellow cap that many species of Arceutho- bium have on their mature seeds (Hawksworth and Wiens 1996). Because several male plants had started flower- ing and several others had most of their flowers open in late July, it appears that the flowering period for A. pendens is probably from early July-September. Based on these observations, the peak flowering period is probably in August in Puebla and Vera Cruz and not in September as estimated for the San Luis Potosi population by Hawksworth and Wiens (1980). Seed dispersal had already started for several female plants in late July and some female plants had nearly completed seed dispersal. Therefore, the seed dispersal period reported by Chazaro and Olivia (1987) of June-September is a good approximation. However, because many female plants had immature fruits in late July in the Vera Cruz population, the seed dispersal period may extend into early October for this population. Typically, infections by A. pendens were non- systemic and did not induce witches’ brooms, but when brooms were formed on trees in Vera Cruz, they were always associated with systemic infec- tions as reported by Hawksworth and Wiens (1980, 1996) for the San Luis Potosi population. We observed that the systemic witches’ brooms formed on some infected trees were induced by both male and female plants, but most of these brooms were associated with male plants as reported by Hawksworth and Weins (1980, 1996): we only observed three systemic witches’ brooms induced by female plants versus over 50 induced by male plants. In a few cases, systemic witches’ brooms induced by male plants had one or even two female plants infecting one of the systemically male-infected branches within the broom. We did not see any female-induced systemic witches’ brooms with male plants, but this condition probably occurs as well. 2006] TABLE 1. MORPHOLOGICAL CHARACTERS MEA- SURED FOR 30 MALE AND 30 FEMALE PLANTS (PLANT HEIGHTS) AND FOR 50 STAMINATE FLOWERS, FRUITS, AND SEEDS OF ARCEUTHOBIUM PENDENS FROM VERA CRUZ, MEXICO. Standard Character Mean Deviation Range Mean Plant Height (cm) Male 23 4.9 16—32 Female 16 4.2 7-26 Staminate Flower 26 0.1 DoDD Diameter (mm) Perianth Lobe LZ 0.1 L.0=1.3 Length (mm) Perianth Lobe fl 0.1 0.9-1.2 Width (mm) Anther Diameter 0.6 0.1 0.5—0.8 (mm) Distance of Anther 0.4 Ou] 0.20.6 From Tip of Perianth Lobe (mm) Staminate Spike 8 255 3-13 Length (mm) Staminate Spike 0.9 0.1 0.8-1.1 Width (mm) Fruit Length (mm) 3.4 0.4 2.64.1 Fruit Width (mm) 1.8 0.2 1.42.2 Seed Length (mm) 23 0.2 1.9-2.7 Seed Width (mm) 0.9 0.1 0.7-1.1 The mean diameter of staminate flowers and the mean length and width of fruits and seeds for A. pendens are approximately the same as those reported for A. divaricatum, the other dwarf mistletoe that parasitizes pinyons as _ principal hosts (Hawksworth and Wiens 1996). Although the size of male flowers, fruits, and seeds are similar for these two dwarf mistletoes, the morphology of the male and female plants 1s different for these taxa. Male and female plants of A. pendens are much larger (means 23 and 16cm, respectively; maximum plant height 32 cm), than A. divaricatum (mean for all plants 8 cm; maximum height 13 cm) (Hawksworth and Wiens 1996). In addition, the light green color of male and female plants of A. pendens is in marked contrast to the dark greenish-brown plants of A. divaricatum. The male plants of A. pendens are commonly pendent as reported by Hawksworth and Wiens (1980), which is a clear contrast to the habit of male plants of A. divaricatum (Hawks- worth and Wiens 1996). Furthermore, the flavo- noid chemistry of A. pendens and A. divaricatum is quite different (Hawksworth and Wiens 1980) and the molecular evidence strongly supports the segregation of these dwarf mistletoes as separate species (Nickrent et al. 2004). Although Hawksworth and Wiens (1980) initially classified the host of A. pendens in Vera Cruz as P. cembroides Zuccarini, they later MATHIASEN AND DAUGHERTY: STUDIES OF ARCEUTHOBIUM PENDENS ot classified this pinyon population as P. orizabensis (Hawksworth and Wiens 1996: Hawksworth et al. 2002) based on a taxonomic treatment of this pinyon population proposed by Bailey and Hawksworth (1992). In San Luis Potosi, A. pendens was not observed parasitizing Pinus cembroides, although trees of this species were observed in the vicinity of the infected Pinus discolor D. K. Bailey & Hawksworth observed there (Hawksworth and Wiens 1980). While Hawksworth and Wiens (1996) and Hawksworth et al. (2002) have classified the host of Arceutho- bium pendens at the Vera Cruz and Puebla locations as Pinus orizabensis, monographs of Pinus for Mexico and Central America by Perry (1991) and Farjon and Styles (1997) classify these pinyon populations as P. cembroides subsp. orizabensis D. K. Bailey. Whether or not A. pendens infects P. cembroides further north in Mexico needs to be determined and _ further research on the host range of A. pendens 1s definitely needed in central Mexico. However, finding additional populations of this cryptic and probably very rare dwarf mistletoe may prove difficult as suggested by Hawksworth and Wiens (1980). ACKNOWLEDGMENTS The assistance of Dr. Vidal Guerra de la Cruz, INIFAP, Tlaxcala, Mexico with the translation for the Resumen and for his help with accommodations in Tlaxcala is greatly appreciated. LITERATURE CITED BAILEY, D. K. AND F. G. HAWKSWORTH. 1992. Change in status of Pinus cembroides subsp. orizabensis (Pinaceae) from central Mexico. Novon 2:306—307. CHAZARO, B. M. AND R. H. OLIVIA. 1987. Lorantha- ceae del centro de Veracruz y zona limitrofe de Puebla. I. Cactaceas y Succulentas Mexicanus 32:55—-60. FARJON, A. AND B. STYLES. 1997. Pinus (Pinaceae). Flora Neotropica, Monograph 75, New York Botanical Gardens, New York, NY. HAWKSWORTH, F. G. AND D. WIENS. 1980. A new species of Arceuthobium (Viscaceae) from central Mexico. Brittonia 32:348—352. - AND ———. 1996. Dwarf Mistletoes: Biology, Pathology, and Systematics. USDA Forest Service, Agriculture Handbook 709, Washington, D.C. , AND B. W. GEILS. 2002. Arceuthobium in i North America. Pp. 29-56 in B. W. GEILS, J. CIBRIAN TOVAR, , AND B. Moopy (tech. coords.), Mistletoes of Norch American conifers. General ee poole Bes GIR- 98; U oD Forest 2 Fort Collins, CO. NICKRENT, D. L., M. A. GARCIA, M. P. MARTIN, AND R. L. MATHIASEN. 2004. A phylogeny of all species of Arceuthobium (Viscaceae) using nuclear and chloroplast DNA sequences. American Journal of Botany 91:125—138. PERRY, J. P. 1991. The pines of Mexico and Central America. Timber Press, Portland, OR. MADRONO, Vol. 53, No. 1, pp. 72—76, 2006 A NEW SPECIES OF SILENE IN THE S/JLENE HOOKERI COMPLEX (CARYOPHYLLACEAE) FROM THE KLAMATH MOUNTAINS OF SHASTA-TRINITY NATIONAL FOREST, TRINITY COUNTY, CALIFORNIA THOMAS W. NELSON AND JANE P. NELSON 4244 Fairway Drive, Eureka, CA 95503-6413 janetomnel@humboldt!.com SUSAN A. ERWIN P.O. Box 1190, Weaverville, CA 96093 serwin@fs.fed.us ABSTRACT Silene salmonacea T.W. Nelson, J. P. Nelson and S.A. Erwin is here described as new and illustrated; endemic to the Klamath Mountains of Shasta -Trinity National Forest west of Clair Engle Lake. Si/ene salmonacea is compared with Silene hookeri Nutt. and Silene bolanderi A. Gray. A key to the three members of the S. Hookeri complex is included. Key Words: Silene, serpentine, Trinity County, Weaverville Formation, Shasta-Trinity National Forest, Clair Engle Lake, California. DESCRIPTION Silene salmonacea T.W. Nelson, J. P. Nelson and S. A. Erwin, sp. nov. (Figs. 1—3). Perennes; rhizomata tenuia ramose. Ad radicem palarem profundam affixa; caules erecti 5-14 mm longi canescentes; folia congesta, paribus 3—4, spathu- lati-oblanceolata, 25—35 mm longa, 4—6 (—8) mm lata; inflorescentia terminalis floribus 2—3 (4); calyx distincte 10 nervatis; corolla limbo salmo- naceo 1n lobos 4 fere equales diviso, unguis viridis basi albus; appendices petaloideae virides vel albae, lineares, 2 in quoque petalo, 1.8—2 mm longae 0.5-1 mm latae; antherae longiexertae; stylus inclusa. Perennial; stems 5—14 mm long; erect from thin branching rhizome system arising from deep, thickened tap root, gray-green, canescent; cauline leaves 3-4 pairs, crowded, spathulate to oblan- ceolate, 25—35 mm long, 4-6 (—8) mm wide, gray- green, the lowest much reduced in size; inflores- cence terminal, 2—3 (—4) flowered; calyx gray- green, canescent, distinctly 10 nerved, 18-23 mm long, teeth ciliate, lanceolate, 4-7 mm long, expanding in fruit; limb of corolla salmon-orange colored (Fig. 3), palmately partite into 4 nearly equal obtuse lobes, the lobes 8-10 mm long, inner 2 lobes 2.4—2.5 mm wide, the outer 2 lobes 1.752 mm wide, the limb considerably wider than claw; claw glabrous, green becoming white at base; petal appendages 2, linear, 1.8—2 mm long, 0.5-1.0 mm wide, light green to white, contrasting with salmon colored limb; filaments 16-17 mm long, long exerted; styles 3, ca. 9 mm long, included; ovary 4mm long, green; seeds reniform, reddish-brown, 2.18 * 1.5 mm, strong- ly papillate, the papillae in parallel rows (Fig. 2). Type: USA, CA, Trinity County, T36N R8W_ sect. 22 SEY%4 NW '%. UTM 4534856N 0519803E. Elev. 1043 m (3421 ft). Shasta-Trinity National | Forest. NW of Forest Service Road 36N25 (County Road 123) ca. % mile NW from bridge over Cement Creek. Serpentine hillside recently disturbed by logging. 3 June 2004. 7. W. Nelson | & S. A. Erwin 9218 Isotypes: CAS, RM, UC, MO, MICH.) (HOLOTYPE: HSC; | Paratypes: USA, CA, Trinity County: T36N | R8&8W sect. 22 SE % NE %. Elev. 1020 m | (3400 ft). About 400 m NW of crossing of | Cement Creek by Forest Service Rd. 36N25. 4. June 2002. D W. Taylor 18097, (HSC); Haylock Ridge above Smith Gulch, Trinity Lake, T34N ROW sects. 1 & 12. 11 June 1978. S. Horner s.n. | (Shasta-Trinity National Forest Herbarium); | T34N ROW sect. 12. UTM 45128106N 0514653E. Elev. 774 m (2580 ft). Along Forest Service Road 34N80 between roads 34N73 & 34N71. Open areas in mixed evergreen forest and road banks. 8 June 2005. 7. W. Nelson & S. A. Erwin 9342 (HSC); T34N R4W sect. 6. UTM 4519536N 0514929E. Elev. 845 m (2815 ft). Along Forest Service Road 34N80 in opening in mixed evergreen woodland. 8 June 2005. 7. W. | Nelson & S. A. Erwin 9345 (HSC). Duplicates are to be distributed. Susan Erwin found two more populations each with less than five plants; so no collec- tions were made. The locations are: Trinity County, west of Clair Engle Lake on the iron rich soil of the Weaverville Formation. T34N | R8W sect. 19. Elev. 900 m (2952 ft) at the junction of roads 34N94 & 34N17Y and the other at T34N R9W sect. 9. Elev. 930 m (3650 ft) along road 113. 2006] NELSON ET AL.: SILENE SALMONACEA 73 Fic. 1. Illustration of Silene salmonacea; A. Mature plant at anthesis. B. Flower in longitudinal section. C. Petal showing appendages. 74 MADRONO [Vol. 53 Fic. 2. Silene salmonacea appears to be very rare, as it is only known from the type locality and five other sites. The populations range in number from 3 to 250 plants. It has been found on serpentine soils and the iron rich soils of the Weaverville Formation of the low hills of the Klamath Mountains west of Clair Engle Lake, where it occurs on either open serpentine or openings in mixed evergreen forests at elevations from 845 to 1043 m. Incomplete descriptions (often based on a single specimen) have added some confusion to the treatment to the Silene hookeri complex. Nuttall (1838) described Silene hookeri based on a single specimen collected by Dr. Gardiner, and no type or lectotype has ever been designated. Gray (1868) named S. bolanderi but with no mention of its lack of petal appendages. Tidestrom and Dayton (1929) described S. ingrami based on a single incomplete specimen. They also failed to describe any petal appendages. Peck (1932) described S. pulverulenta based on a single in- complete specimen. His prolog was mostly written in Latin with only a brief commentary in English that included no comments about the petal appendages. Subsequent to these species descriptions, Abrams (1944) reduced S. bolanderi to a subsp. Photograph of the seeds of Silene salmonacea. Seeds are approximately 2 mm in length. of S. hookeri and S. ingrami and S. pulverulenta to synonyms of subsp. hookeri. In their revision of Silene, Hitchcock and Maguire (1947) recog- nized three subsps. of S. hookeri: subsps. hookeri, bolanderi and pulverulenta. Munz and Keck (1959) and Morton (2005) recognized only subsps. hookeri and bolanderi. Wilken (1993) only recognized a highly variable S. hookeri. During our studies of the S. hookeri complex, we concluded that based on the different characters shown in Table 1, S. bolanderi should be recognized as a separate species rather than a subspecies of S. hookeri. Thus we consider, the S. hookeri complex to consist of S. hookeri, S. bolanderi and S. salmonacea. This is also the concept being followed by Hartman (personal communication) in his forthcoming revision for the Jepson Manual. These three taxa of the S. hookeri complex are easily distinguished. Silene bolanderi is separated from the other two because it lacks petal appendages and is the only member with white flowers. In addition to differences in petal color, S. hookeri can be easily separated from S-. salmonacea by the fact that the anthers are included while those of S. salmonacea are long exerted. Additional details are included in Table 1. 2006] NELSON ET AL.: SILENE SALMONACEA FIG. 3. A KEY TO THE CALIFORNIA SILENE HOOKERI COMPLEX A. Petal appendages absent; limbs white or tinted green Silene bolanderi A’ Petal appendages 2; limbs pink or salmon-orange B. Anthers included, petal lobes either 2 each with a small lateral tooth or 4 and unequal; limb and claw pink adaxially, white abax- lally Silene hookeri B’ Anthers excluded, petals with 4 nearly equal parallel lobes; limbs salmon-orange, claws white becoming green at base Silene salmonacea TABLE 1. Silene salmonacea Petal colors Limb salmon-orange, Claw white becoming green at base 4, equally palmately dissected nearly to base, tip obtuse Petal appendages 2 Stems |—2 a» Lobes of limb reclining to erect from thin rhizomes Anthers Long exerted Limb and claw both 4, Photograph of Silene salmonacea at type locality. ACKNOWLEDGMENTS We wish to thank Dr. Guy Nesom for the Latin description and Ms Christana Paleno for the illustration. We also wish to thank Dr. Ronald Hartman (RM) and Dr. Rich Rabeler (MICH) for their constructive suggestion on improving this paper. We especially thank Dr. James P. Smith, Jr. and Robin Bencie for access to the excellent collections at HSC. We also thank Sydney Carothers for assistance in the field and for providing the photograph of the seeds of S. salmonacea. We are grateful to Dr. Dean Wm. Taylor for bringing this very different and rare Si/ene to our attention. MORPHOLOGICAL COMPARISON OF SPECIES IN THE S7LENE HOOKERI COMPLEX. Silene bolanderi Silene hookeri Limb and claw pink adaxially, white abaxially 2 lobes each with a small lateral tooth or 4 unequal lobes white tinted green equally palmately dissected nearly to base, tip acute 0 2 1—3, erect from thin 3-5, reclining from root rhizomes crown Long exerted Included 76 MADRONO LITERATURE CITED ABRAMS, L. 1944. Illustrated flora of the Pacific states, Vol. Il. Polygonaceae to Krameriaceae. Stanford University Press, Stanford University, CA. GRAY, A. 1868. Proceedings of the American Academy of Arts and Sciences 7:330. HITCHCOCK, C. L. AND B. MAGUIRE. 1947. A revision of North American species of Si/ene University of Washington Publications in Biology 13:1—73. Morton, J. K. 2005. Silene. Pp. 166-214 in Flora of North America Editorial Committee, Flora of North America north of Mexico, Vol. 5. Magno- liophyta: Caryophyllidae, pt. 2. Oxford University Press, New York, NY. [Vol. 53 Munz, P. A. AND D. D. KECK. 1959. Silene. Pp. 285— 292 in A California flora. University of California Press, Berkeley, CA. NUTTALL, T. in Torrey & Gray. 1838. Fl. N. Amer. E193: Peck, M. E. 1932. New plants from Oregon. Torreya 32:147-153. TIDESTROM, I. AND W. A. DAYTON. 1929. A new Silene from the Umpqua National Forest. Univer- sity of Washington Publications in Biology 42:207—208. WILKEN, D. 1993. Silene. Pp. 488-493 in J. C. Hickman (ed.), The Jepson manual: higher plants of Cali- fornia. University of California Press, Berkeley, CA. MADRONO, Vol. 53, No. 1, pp. 77-84, 2006 LEPECHINIA ROSSIT (LAMIACEAE), A NARROW ENDEMIC FROM THE WESTERN TRANSVERSE RANGES OF SOUTHERN CALIFORNIA STEVE BOYD AND ORLANDO MISTRETTA Herbarium, Rancho Santa Ana Botanic Garden, Claremont, CA 91711 steve.boyd@cgu.edu ABSTRACT Lepechinia rossii (Lamiaceae) is described as a new species narrowly endemic to the western Transverse Ranges of southern California. It is a member of section Ca/ycinae, which includes four additional species endemic to California and adjacent Baja California, Mexico—L. calycina, L. cardiophylla, L. fragrans, L. ganderi—and L. mexicana, an anomalous, and probably unrelated, species from central Mexico. Lepechinia rossii is most readily distinguished from other members of section Calycinae by geniculate inflorescence axes, bent at 60—90° angles relative to the subtending stems, and by large, foliaceous inflorescence bracts which are generally equaling or exceeding their adjacent flowers in length, and little reduced distally. At present, two populations are documented, one in the Liebre Mountains (Los Angeles County) and one in the Topatopa Mountains (Ventura County), both occurring in chaparral, on public lands administered by the U.S. Forest Service. Conservation concerns include habitat degradation by off-highway vehicle activity, power line maintenance, petroleum exploration and extraction, and anthropogenic changes in fire frequency. RESUMEN Lepechinia rossii (Lamiaceae) es descrita como una nueva especie con endemismo limitado a la cadena Transverse Ranges (occidental) del Sur de California. Esta especie es un miembro de la seccion Calycinae, la cual incluye cuatro especies adicionales endémicas a California y zonas adyacentes a California en Baja California, Mexico—L. calycina, L. cardiophylla, L. fragrans, L. ganderi—y L. mexicana, anomalos y probablemente no relacionadas, especies de Mexico central y sur-central. Lepechinia rossii es facilmente diferenciado de los otros miembros de la secci6n Calycinae porque sus ejes de la inflorescencia son curvados en un angulo de 60—90° con respecto al tallo y porque las bracteas de la inflorescencia son largas y foliaceas, las cuales son generalmente iguales o exceden en longitud a sus flores adyacentes y porque el tamano de las bracteas a lo largo del eje de la inflorescencia apenas se reduce en longitud. En la actualidad, dos poblaciones son documentadas, una en la Liebre Mountains (Condado de Los Angeles) y la otra en la Topatopa Mountains (Condado de Ventura), ambas se encuentran en chaparrales, sobre terrenos publicos administrados por el Servicio Forestal de los Estados Unidos. Problemas de conservacion incluye degradacion del habitat por actividad todoterrenos, mantenimiento de lineas de electricidad, exploracion y extraccion de petroleo, y cambios antropogénicos en la frecuencia de incendios. Key Words: California, Calycinae, endemic, Lamiaceae, Lepechinia, Liebre Mountains, Topatopa Mountains, Transverse Ranges. INTRODUCTION Lepechinia Willd. (Lamiaceae) is a heteroge- nous genus comprised of ca. 55 species of suffruticose perennials, shrubs, and small trees (Epling 1948; Mabberly 1997). Most taxa occur in the mountains of South America, with a few species extending into North America (Mexico, CA), and disjunctly, into the Pacific Ocean archipelagos of Revillagigedos (Mexico) and Hawaii (Epling 1948). The most recent broad- scale floristic treatments for California (e.g., Munz 1959; Averett 1993), and southern Cali- fornia (Munz 1974), recognize four species of Lepechinia in the State. These include L. calycina (Benth.) Epling; L. cardiophylla Epling; L. fragrans (E. Greene) Epling; and L. ganderi Epling. The latter three taxa are found only in the southern third of the State, from the Trans- verse Ranges and northern Channel Islands southward, with two (L. cardiophylla, L. ganderi) reaching adjacent northwestern Baja California, Mexico (Fig. IA). All four taxa are placed by Epling (1948) within his section Calycinae, along with L. mexicana (S. Schauer) Epling from central Mexico. We describe here a fifth species of Lepechinia section Calycinae from Califor- nia—Lepechinia rossii S. Boyd & O. Mistretta— based on collections made in the Liebre and Topatopa mountains, two units of southern California’s western Transverse Ranges. AIl- though the affinity of ZL. rossii with other Californian members of Lepechinia section Caly- cinae 1s evident in a number of variously shared morphological characters (e.g., habit, leaf shape, leaf vestiture, calyx shape, corolla size and shape)—Epling’s “‘living mosaic’ (1944)—it 1s a suite of inflorescence characters that most 78 MADRONO [Vol. 53 Kilometers 180 360 FIG. 1. Lepechinia rossii. A) map showing relative distribution of Lepechinia sect. Calycinae taxa in California and northern Baja California, Mexico. Lepechinia rossii—open stars; L. calycina—closed circles; L. fragrans—closed squares; L. cardiophylla—open triangles; L. ganderi—open diamonds. B) illustration of portion of an upper stem | bearing two inflorescences, showing typical leaves, bent primary inflorescence axes, broadly ovate to suborbicular, overlapping floral bracts, and pendant flowers. 2006] readily set L. rossii apart from all other members of the genus. These include orientation of the inflorescence axes, size, shape, and orientation of the floral bracts, and the degree to which these bracts are reduced apically. Although two non-flowering, historical speci- mens of Lepechinia from the Topatopa Moun- tains, B.W. Evermann (Pine Creek near Sespe; 24 Mar 1917 [CAS #25345]), or R. Hoffmann (Sespe Canyon; 21 Mar 1927 [SBBG #6403; 4#46404]), are likely the earliest collections of this plant, from a practical standpoint L. rossii was “discovered” in the Fall of 1991. While conduct- ing botanical surveys for the Angeles National Forest in late September of that year, the second author collected a sterile, partially deciduous Lepechinia on Red Mountain, between San Francisquito and Elizabeth Lake canyons, in the Liebre Mountains region of northwestern Los Angeles County. Upon seeing this specimen, Timothy S. Ross, then a senior curatorial assistant in the herbarium at Rancho Santa Ana Botanic Garden (RSA), immediately noted that Lepechinia was otherwise unknown in the Liebre Mountains, and the find was therefore of interest from a floristic standpoint. At the time, Ross speculated Mistretta’s plant was likely L. fra- grans, as that species 1s found in the San Gabriel and Santa Monica mountains, regions of the Transverse Ranges lying to the southeast and southwest of the Liebre Mountains, respectively. In the spring of 1992, Ross and the first author visited the Red Mountain area and found a population of Lepechinia growing on the mountain’s northern slope and upper ridgelines (ca. 500—1000 individuals). So distinctive were the plants in flower (geniculate inflorescence axes, with large, foliaceous, upwardly directed floral bracts, little reduced apically and generally longer _ than their adjacent flower) that it was instantly clear they were not L. fragrans, nor did they appear to fit any of the other Californian taxa in the genus. In Fall of 1995, while examining Lepechinia specimens from Santa Barbara Bota- nic Garden (SBBG), Ross encountered an E.R. Blakley collection from Tar Creek in the Topa- topa Mountains (Ventura County; Los Padres National Forest), collected in 1994, that appeared consistent with the undescribed taxon from Red Mountain. Ross and the authors visited Blakley’s Tar Creek site in Spring of 1995 and confirmed the Topatopa Mountains plants were the same undescribed entity as those of the Liebre Moun- tains. The Tar Creek site is ca. 40 km west of Red Mountain, and as a tributary of Sespe Canyon, in the general vicinity of the sterile 1917 and 1927 collections mentioned above. In hght of the central role our friend and colleague, Timothy S. Ross has played in the discovery and understanding of this new Lepe- chinia, it is our pleasure to name this species in his BOYD AND MISTRETTA: LEPECHINIA ROSSI, NEW TRANSVERSE RANGES ENDEMIC — 79 honor. For the vernacular, we recommend the plant be called Ross’ pitcher sage. DESCRIPTION Lepechinia rossii S. Boyd & O. Mistretta, sp. nov. (Fig. 1B)—Type: USA, California, Los Angeles Co., Transverse Ranges, Liebre Moun- tains region: Head of Ruby Canyon on northern flank of Red Mountain, between Elizabeth Lake and San Francisquito canyons; 34°35'33’"N, 118°29'29"W [NAD 27]; 305 m (1000 ft); 11 May 2004, S. Boyd & T. Morgan 11169 (holotype RSA; isotypes CAS, GH, SBBG, UC, UCR, US). Paratypes: USA, California, Los Angeles Co., Transverse Ranges; San Gabriel Mtns region: Red Mountain [technically Liebre Mountains]; 25 Sep 1991, O. Mistretta s.n. (RSA). Transverse Ranges; Liebre Mountains region: Ruby-Clear- water Truck Trail, south of Ruby Canyon, north of Red Mountain; near 34.59788°N, 118.52618°W [NAD 83] (Warm Springs Moun- tain 7.5 quad); T6N R16W sect. 24, SW'% of NE"); 788 m (2585 ft); 24 May 2005, L. Gross et al. 2311 (RSA). Ventura Co., Topa Topa [=Topatopa] Mountains, southern flank of Tar Creek, ca. | air mile southeast of confluence with Sespe Canyon, along an old dirt road leading down to Sespe Canyon off of Squaw Flat Road, about the base of hill “2582” at the boundary of the Sespe Condor Sanctuary; T5N R20W [sec- tions] unsurveyed; ca. 732 m (2400 ft); 12 Jun 1996, S. Boyd et al. 8849 (RSA). Tar Creek, on edge of old road down to Green Cabins on Sespe Creek; 549 m (1800 ft); 12 Jun 1994, E.R. Blakley 7611 (SBBG). Differt a Lepechinia calycina, L. cardiophylla, L. fragrans, et L. ganderi inflorescentia e basi 60— 90° geniculatus et inflorescentiae bracteis folia- clis, plus minusve ultra flores. Shrub, often forming clonal stands following disturbance or fire, generally less than 1.5 m tall with numerous ascending to erect branches from base and strongly aromatic herbage (Fig. 2); stems weak, + brittle, those developing from short-shoots formed in upper axils of previous season’s growth (vs. root or stem suckers) somewhat thickened towards base, with numer- ous, closely spaced leaf scars, growth of current season pale green, minutely glandular-puberulent with short-stipitate and subsessile capitate-glan- dular trichomes (appearing + papillate at 20 x magnification), and scattered multicellular, clear, kinked, irregularly branched, nonglandular tri- chomes (to 1.5 mm long), older branches with bark becoming reddish brown and shredding in age; leaves opposite, with petioles ca. 5—20 (—30) mm long, often slightly winged distally, blades bright, light green or yellowish green, ovate to deltate-ovate, ca. 3-13 cm long, truncate to subcordate at base, margins irregularly and 80) MADRONO [Vol. 53 FIG. 2. Type plant of Lepechinia rossii growing in relatively open area surrounded by chaparral vegetation, showing typical rounded crown, geniculate inflorescence axes, and prominent, ascending bracts. shallowly serrulate to dentate, upper (adaxial) surface shallowly bullate, lower (abaxial) surface with prominent, raised, reticulate venation, ves- titure as on stems, and with scattered, golden, sessile, hemispherical glands set in shallow pits (especially below); inflorescence terminal on growth of current season, geniculate, bent + 60-90° relative to subtending stem and thus arching or spreading, axis shallowly curved between nodes, appearing scalloped (Figs. 1B, 2; 3A), unbranched, or more often with two short branches arising at lowest node, especially on vigorous stems; bracts foliaceous (Figs. 1B, 3B), sessile, ascending, broadly ovate to suborbicular, 2.5—8 cm, generally longer than subtended flower and not strongly reduced in size distally, therefore appearing imbricate towards apex of inflores- cence, margins entire or the lowest 2-4 pairs shallowly serrulate to dentate, surfaces + similar to leaves but less ruggose adaxially, raised veins of abaxial surface visually prominent, and long multicellular hairs sparse or absent; flowers solitary in bract axils, pendent on minutely glandular-puberulent pedicels 12-13 mm long (Figs. 3B, 4A); ca/yx at anthesis + campanulate, the tube 10-12 mm long, finely raised-reticulate veined between 12-15 thicker longitudinal veins from base (Figs. 1B, 4B), minutely glandular- puberulent externally, + glabrous internally, the lobes generally erect or slightly spreading, broad- ly deltate, 4-6 mm long X 4-6 mm wide, the apex abruptly short apiculate, frequently one or more lobes with single, apiculate, deltate marginal tooth 0.5-1.5 mm long X 0.5—-1.5 mm wide, fruiting calyx enlarging, becoming somewhat inflated, papery; corolla overall broadly tubular (Figs. 1B, 3B, 4B), 33-39 mm long, abruptly narrowed in the proximal 8—9 mm, the point of narrowing marked internally by a ring of short, glandular hairs, broad portion of tube exerted from calyx ca. 15 mm, somewhat angled exter- nally below point of stamen attachment, throat 10-11 mm wide, limb 5-lobed, strongly bilateral, the two lateral and two upper (adaxial) lobes short, 3.5-4.5 mm long X 4.5-6 mm wide, rounded apically, spreading to recurved, the lower (abaxial) lobe much larger, 10-12 mm long