c O Z " " -J J 2 ES SMITHSONIAN INSTITUTION NOUnHiSNI NVINOSHimS SBldVdan LIBRA - ^ ^ V z Z in £ (/) ± CO ;NI NVINOSHIIIAIS S3iaVdan libraries SMITHSONIAN INSTITUTION NOIiH CO z CO * :z c/) *^ 2 ES SMITHSONIAN INSTITUTION NOIinillSNI NVINOSHimS S3 I Vd 8 11 L! B R/ C/) 2 ^ ~ CO ;Nl"^NV!NOSHimS S3ldVdan LIBRARI ES^ SMITHSONIAN~^INSTiTUTION NOIlfl r~ > Z r- Z r- 2 3 CO ' r: (/> £ (/) r: ES SMITHSONIAN INSTITUTION NOIinillSNI NVINOSHill^S S3iyvyan LI BR/ O to X — to Ni NviN0SHims^'S3idvyan libraries 'Smithsonian institution Noun. 2>C^ ^ t/5 ~ <^ CO 2 ' ^ Z — ^ 2 — t ES SMITHSONIAN INSTITUTION NOIiOiliSNI NVINOSHimS S3 I d Vd 3 11^ LI B R) 2 r" 2 r" z *"* _ CO ± — O) \ ^ CO 5N1 NVINOSHimS S3IdVdan libraries SMITHSONIAN INSTITUTION „ NOIIH H (/> Z .vy. to 2 .s^v ^ CO Z CO * Z CO ES SMITHSONIAN INSTITUTION NOIinillSNI NVINOSHlllAIS S 3 I d V d a IT _ L I B R CO r; CO zz . co ^ ^ 2 A. X'^ CO VIITHSONIAn'^INSTITUTION NOlinillSNl"'NVINOSHilWS S3IdVyan LIBRARIE \/IN0SHimS_S3 1 d ViJ a n^LI B R AR I ES SMITHSONIAN^INSTITUTION CO NOllniliS^ VIITHSONIAN INSTITUTION NOIiniliSNI NVINOSHlllMS S3 I d Vd 8 n__ LI B R AR I E in c o O z VINOSHimS S3idVdai1 libraries SMITHSONIAN INSTITUTION NOIiniliSI z o CO Z o Z) CO 2! MITHSONIAN INSTITUTION NOIiniliSNI NVINOSHlll^S S3 I d V d 9 11 I B R A R I E < IVINOSHimS S3ldVd9n LIBRARIES SMITHSONIAN INSTITUTION NOiiHiliS — /rt — CO SMITHSONIAN INSTITUTION NOIiniliSNI NVINOSHimS S3ldVdan LIBRARI NviNOSHims S3idVdan libraries Smithsonian institution NOiiniiis j.^ m frt SMITHSONIAN INSTITUTION NOIiniliSNI NVINOSHllk^S S3ldVdan ~ ^ 5 ..^r-^ z A UJ CO LI B RAR MADRONO A WEST AMERICAN JOURNAL OF BOTANY VOLUME XXXVI 1989 BOARD OF EDITORS Class of: 1989 — Frank Vasek, University of California, Riverside Barbara Ertter, University of California, Berkeley 1990 — Steven Timbrook, Ganna Walska Lotusland Foundation, Montecito, CA Thomas R. Van Devender, Arizona-Sonora Desert Museum, Tucson 1991— James Henrickson, California State University, Los Angeles Wayne R. Ferren, University of California, Santa Barbara 1992— Bruce A. Stein, The Nature Conservancy, Washington, D.C. William L. Halvorson, Channel Islands National Park, Ventura, CA 1993— Jon E. Keeley, Occidental College, Los Angeles, CA Rhonda L. Riggins, California Polytechnic State University, San Luis Obispo, CA Editor— David J. Keil Biological Sciences Department California Polytechnic State University San Luis Obispo, CA 93407 Published quarterly by the California Botanical Society, Inc. Life Sciences Building, University of California, Berkeley 94720 Printed by Allen Press, Inc., Lawrence, KS 66044 ii Dedicated to botanical pursuits, Art Kruckeberg has distinguished himself as an eminent student of edaphic ecology. His work on serpentine endemism is classic. Known throughout western North America as an intrepid field researcher, Art's expertise spans ecology, population biology, systematics, phytogeography, evolution, con- servation biology, and horticulture. Long before it became fashion- able to be interdisciplinary, he was building bridges between these different fields of study. As a respected teacher and mentor, he has challenged and stimulated many students and colleagues. Born and educated in California, he has spent all 40 years of his professional career at the University of Washington. It is a pleasure to pay tribute to his outstanding botanical contributions. — Melinda F. Denton, Department of Botany, University of Washington, Seattle, WA 98195. Photo provided by Douglass M. Henderson, Department of Biological Sciences, University of Idaho, Moscow, ID 83843. iii TABLE OF CONTENTS Alverson, Edward R., Isopyrum stipitatum (Ranunculaceae) in the Willamette Valley, Oregon 217 Argus, George W., Salix scouleriana (Salicaceae) in Sonora, Mexico 135 Armstrong, Wayne P., Noteworthy collections of Wolffia 283 Bartolome, James W. (see McClaran, Mitchel P.) Bayer, Randall J., A systematic and phytogeographic study of Antennaria aromatica and A. densifolia (Asteraceae: Inuleae) in the western North American cordillera 248 Becking, Rudolf W., Segregation of Hastingsia serpentinicola, sp. nov. from Hastingsia alba (Liliaceae: Aspholdeleae) 208 BowcuTT, Frederica S., Reappraisal of the range of Phacelia vallicola (Hy- drophyllaceae) 5 1 Breedlove, Dennis E., and Jose Luis Leon de la Luz, A new species of Daphnopsis (Thymelaeaceae) from Baja California Sur, Mexico 266 Broich, S. L., Chromosome numbers of North American Lathyrus (Fabaceae) 41 BuTTERWiCK, Mary, Noteworthy collections of Ephedra funerea and Melica imperfecta 136 Carter, Annetta, Review of Baja California Plant Field Guide by Norman C. Roberts 287 Chadwick, Ann, and David J. Keil, Noteworthy collection of Prunus fasci- culata \2lv. fasciculata 32 Chambers, Kenton L., The taxonomic relationships of Allocarya corallicarpa (Boraginaceae) 280 Chambers, Kenton, and Jacqueline Greenleaf, Gentiana setigera is the cor- rect name for G. bisetaea (Gentianaceae) 49 Clark, Curtis (see Harrington, Daniel F.) Ceska, Adolf (see Savale, W. F., Jr.) Denison, Stella S., and Dale W. McNeal, A re-evaluation of the Allium sanbornii (Alliaceae) complex 122 Devine, Timothy B. (see Wagner, Warren H., Jr.) DoLAN, Rebecca W., and Lawrence F. LaPre, Taxonomy of Streptanthus sect. Biennes, the Streptanthus morrisonii complex 33 Dominguez-Cadena, Raymundo (see Leon de la Luz, Jose Luis) DoRN, Robert D., Typification of Salix geyeriana (Salicaceae) 135 Ertter, Barbara (see Shevock, James R.) Felger, Richard S. (see Reeder, John R.) Gottlieb, L. D., Roral pigmentation patterns in Clarkia (Onagraceae) 1 Greenleaf, Jacqueline (see Chambers, Kenton) Halvorson, William L., Review of North American Terrestrial Vegetation edited by Michael G. Barbour and W. Dwight Billings 219 Hammer, Samuel, Phytogeographical notes on acidophilous Cladonia species in California 169 Harrington, Daniel F., and Curtis Clark, Reduction in light reflectance of leaves of Encelia densifolia (Asteraceae) by trichome wetting 1 80 Hickman, James C, Points of View: Response to Reveal 218 Hilsenbeck, Richard A., A new species of Siphonoglossa (Acanthaceae) and some infrageneric transfers 198 Hilsenbeck, Richard A. (see Ralston, Barbara E.) Horn, Sally P., Postfire vegetation development in the Costa Rican paramos 93 JoKERST, James D. (see Shevock, James R.) iv JoYAL, Elaine, Noteworthy collections of Astragalus curvicarpus var. subglaber, CoUomia debilis var. larsenii, and Rudbeckia occidentalis var. montana from Oregon 53 JOYAL, Elaine, Myosotis latifolia and not M. sylvatica (Boraginaceae) in Cal- ifornia 131 Keeley, Jon E., Review of The Biogeography of Fire in the San Bernardino Mountains of California— A Historical Study by Richard A. Minnich 287 Kelley, Walter A., Comments and notes on Portulaca in California 281 Kelley, Walter A., and Dieter H. Wilken, Several noteworthy collections from Colorado 285 Keil, David J., Review of A Guide to Wildlife Habitats of California by K. E. Mayer and W. F. Laudenslayer 57 Keil, David J. (see Chadwick, Ann) Knight, Paul (see Spellenberg, Richard) Kruckeberg, a. R., Review of Soil-Plant Relations: An Ecological Approach byD. W.Jeffrey 219 KuijT, Job, A note on the germination and establishment of Phoradendron californicum (ViscaceaQ) 175 Landrum, Leslie R., The generic position of Myrtus alternifolia and notes on Calycolpus (Myrtaceae) 9 LaPre, Lawrence F. (see Dolan, Rebecca W.) Leon de la Luz, Jose Luis (see Breedlove, Dennis E.) Leon de la Luz, Jose Luis, and Raymundo Dominguez-Cadena, Flora of the Sierra de la Laguna, Baja California Sur, Mexico 6 1 LucKOW, Melissa, Review of Trees and Shrubs of Trans-Pecos Texas by A. Michael Powell 56 Mason, Charles T., Infraspecific name changes in Limnanthes (Limnantha- ceae) 50 McClaran, Mitchel P., and James W. Bartolome, Effect of Quercus douglasii (Fagaceae) on herbaceous understory along a rainfall gradient 1 4 1 McLeod, Malcolm G., Review of Inventory of Rare and Endangered Vascular Plants of California edited by James P. Smith and Ken Berg 54 McNeal, Dale W. (see Denison, Stella S.) Nesom, Guy L., Macromeria alba (Boraginaceae), a new species from Tamau- lipas, Mexico 28 Parsons, David J., and Thomas J. Stohlgren, Effects of varying fire regimes on annual grasslands in the southern Sierra Nevada of California 154 Passini, Marie-Franoise, and Nicole Pinel, Ecology and distribution of Pinus lagunae in the Sierra de la Laguna, Baja California Sur, Mexico 84 Patterson, Robert, Taxonomic relationships of Gilia maculata (Polemoni- aceae) 15 Peterson, Paul M., A re-evaluation of Bealia mexicana (Poaceae: Eragros- tideae) 260 Phillips, H. Wayne, Noteworthy collection of Goodyera repens 174 Pinel, Nicole (see Passini, Marie- Fran^oise) PiNKAVA, Donald J., Review of Colorado Flora: Western Slope by William A. Weber 56 Ralston, Barbara E., and Richard A. Hilsenbeck, Taxonomy of the Opuntia schottii complex (Cactaceae) in Texas 221 Reeder, John R., and Richard S. Felger, The Aristida californica - glabrata complex (Gramineae) 187 Reveal, James L., Points of View: On the modern death of David Douglas .. 137 Savale, W. F., Jr., and Adolf Ceska, Noteworthy collections of Wolffia Co- lumbiana 153 ScHiERENBECK, Kristina A., Noteworthy collection of Polygonum marinense 207 v Schmidt, Clifford L., Review of Guide to the Regional Parks Botanic Garden by Walter Knight and Irja Knight 54 Shelly, J. Stephen, Biosystematic studies of Phacelia capitata (Hydrophyl- laceae), a species endemic to serpentine soils in southwestern Oregon 232 Shevock, James R., Barbara Ertter, and James D. Jokerst, Monardella beneolens (Lamiaceae), a new species from the crest of the southern Sierra Nevada, California 271 Spellenberg, Richard, and Paul Knight, A new species of Erigeron (Aster- eaceae: Astereae) from central New Mexico 1 1 5 Stohlgren, Thomas J. (see Parsons, David J.) Vincent, Karl A., Noteworthy collection of Penstemon venustus (Scrophu- lariaceae) from California 53 Wagner, Warren H., Jr., and Timothy B. Devine, Moonworts (Botyrchium: Ophioglossaceae) in the Jonesville area, Butte and Tehama Counties, Cal- ifornia 131 WiLKEN, Dieter H. (see Kelley, Walter A., and Dieter H. Wilken) Wilson, Paul, Noteworthy collection of Pallavicinia lyellii (Hepaticae: Pal- la viciniaceae) from California 53 ZiKA, Peter F., Noteworthy collection of Panicum rigidulum 207 vi VOLUME 36, NUMBER 1 33 JANUARY-MARCH 1989 MADRONO A WEST AMERICAN JOURNAL OF BOTANY Contents Floral Pigmentation Patterns in Clarkia (Onagraceae) L. D. Gottlieb The Generic Position of Myrtus alternifolia and Notes on Calycolpus (Myrtaceae) Leslie R. Landrum Taxonomic Relationships of Gilia maculata (Polemonl\ceae) Robert Patterson Macronema alba (Boraginaceae), a New Species from Tamaulipas, Mexico Guy L. Nesom Taxonomy of Streptanthus sect. Biennes, the Streptanthus morrisonii Complex (Brassicaceae) Rebecca W. Dolan and Lawrence F. LaPre Chromosome Numbers of North American Lathyrus (Fabaceae) Steven L. Broich NOTES Gentiana setigera is the Correct Name for G. Bisetaea (Gentianaceae) Kenton L. Chambers and Jacqueline Greenleaf Infraspecihc Name Changes in Limnanthes (Limnanthaceae) Charles T. Mason Reappraisal of the Range of Phacelia vallicola (Hydrophyllaceae) Frederica Bowcutt NOTEWORTHY COLLECTIONS California Oregon REVIEWS ANNOUNCEMENTS 1 9 15 28 33 41 49 50 51 32, 53 53 54 27, 40, 60 \ .... ^ PUBLISHED QUARTERLY BY THE CALIFORNIA BOTANICAL SOCIETY Madrono (ISSN 0024-9637) is published quarterly by the California Botanical So- ciety, Inc., and is issued from the office of the Society, Herbarium, Life Sciences Building, University of California, Berkeley, CA 94720. Subscription rate: $30 per calendar year. Subscription information on inside back cover. Established 1916. Second-class postage paid at Berkeley, CA, and additional mailing offices. Return requested. Postmaster: Send address changes to James R. Shevock, Botany Dept., California Academy of Sciences, San Francisco, CA 941 18. Editor— Daviu J. Keil Biological Sciences Department California Polytechnic State University San Luis Obispo, CA 93407 Board of Editors Class of: 1989 — Frank Vasek, University of California, Riverside, CA Barbara Ertter, University of California, Berkeley, CA 1990— Steven Timbrook, Ganna Walska Lotusland Foundation, Montecito, CA Thomas R. Van Devender, Arizona-Sonora Desert Museum, Tucson, AZ 1991— James Henrickson, California State University, Los Angeles, CA Wayne R. Ferren, Jr., University of California, Santa Barbara, CA 1992— Bruce A. Stein, The Nature Conservancy, Washington, D.C. William L. Halvorson, Channel Islands National Park, Ventura, CA 1993— Jon E. Keeley, Occidental College, Los Angeles, CA Rhonda L. Riggins, California Polytechnic State University, San Luis Obispo, CA CALIFORNIA BOTANICAL SOCIETY, INC. Officers for 1988-89 President: John L. Strother, University Herbarium, University of California, Berkeley, CA 94720 First Vice President: James Affolter, Botanical Garden, University of California, Berkeley, CA 94720 Second Vice President: James Henrickson, California State University, Los An- geles, CA 90032 Recording Secretary: Rodney G. Myatt, Department of Biological Sciences, San Jose State University, San Jose, CA 95192 Corresponding Secretary: James R. Shevock, Department of Botany, California Academy of Sciences, San Francisco, CA 94 1 1 8 Treasurer: Thomas F. Daniel, Department of Botany, California Academy of Sci- ences, San Francisco, CA 94 1 1 8 Financial Officer: Cherie L. Wetzel, Department of Biology, City College of San Francisco, 50 Phelan Ave., San Francisco, CA 941 12 The Council of the California Botanical Society consists of the officers listed above plus the immediate Past President, Dale McNeal, Department of Biological Sciences, University of the Pacific, Stockton, CA 9521 1; the Editor of Madrono; three elected Council Members: John Mooring, Department of Biology, University of Santa Clara, Santa Clara, CA 95053; Barbara Ertter, Herbarium, Botany Department, Uni- versity of California, Berkeley, CA 94720; Elizabeth McClintock, Herbarium, Bot- any Department, University of California, Berkeley, CA 94720; and a Graduate Student Representative, Valerie Haley, Department of Biological Sciences, San Jose State University, San Jose, CA 95192. THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER. FLORAL PIGMENTATION PATTERNS IN CLARKIA (ONAGRACEAE) L. D. Gottlieb Department of Genetics, University of California, Davis, CA 95616 Abstract The pattern of anthocyanin pigmentation on flower petals of species of Clarkia sections Rhodanthos and Godetia is reviewed. The petals have large spots, blotches, or bands of red-purple color near the base, the center, the upper margin or in several positions, or are unspotted. A recent genetic analysis in C. gracilis revealed that the large central petal spot characteristic of subsp. sonomensis is allelic to a basal petal spot in subsp. gracilis that is normally not expressed because of the action of a modifier gene. Since Clarkia species display only a small number of discrete pigmentation patterns, I suggest that the major components of pattern differences among them are not complex from a genetic standpoint. Novel patterns can be assembled by substi- tuting alleles at relatively few loci. Additional genes presumably contribute to various details of the patterns. In many genera of plants, the flower displays more differences from species to species than any other plant part, varying, for ex- ample, in size, shape, outline, texture, orientation, and color as well as in features of scent, nectar, and pollen. Many floral traits serve to attract and reward specific pollinators. Although the flower has obvious and important adaptive significance, the genetic basis of floral differences between species has been worked out in only a few cases (Gottlieb 1984). Genetic studies might reveal whether new floral phenotypes result from the activities of a relatively small num- ber of (major) genes or originate only after the accumulation of numerous genetic diflerences. In addition, when the same or similar phenotype appears in a number of species, it is important to deter- mine whether this reflects the activities of the same subset of genes or convergence based on the independent appearance of new com- binations of genes. Information on these issues will help us to un- derstand how morphological diflerences evolve in plants. Floral differentiation involving changes in both structural mor- phology and pigmentation patterns is particularly important among species of Clarkia (Onagraceae) native to California. Overall floral diflerentiation in Clarkia is closely associated with the pollination system (MacSwain et al. 1973). Indeed, much of the adaptive ra- diation in the genus has primarily involved the flower, so that if plants of diflerent species were stripped of their flowers, they would be nearly identical in appearance (MacSwain et al. 1973). The genus has long been used as a model system for studies of plant evolu- Madrono, Vol. 36, No. 1, pp. 1-8, 1989 2 MADRONO [Vol. 36 tionary biology, beginning with the elegant biosystematic and cy- togenetic studies of Professor Harlan Lewis and his students and colleagues (Lewis 1953, 1962, 1973; Lewis and Lewis 1955). More recent studies of Clarkia include examination of genetic differentia- tion among species using data from electrophoretic analysis of iso- zymes (Gottlieb and Weeden 1979; Pichersky and Gottlieb 1983; Odrzykoski and Gottlieb 1984; Soltis et al. 1987), and reconstruc- tions of species relationships using restriction endonuclease analysis of chloroplast DNA (Sytsma and Gottlieb 1986a, b). The most common floral type in diploid Clarkia species is the "godetia" flower which characterizes sections Godetia, Rhodanthos, Peripetasma, and Fibula, totalling about two dozen species. The godetia flower is held upright and is shaped like a bowl. The four petals are obovate to fan-shaped and are not much narrowed at the base. Although their structure, size, and shape are generally similar, the pattern of anthocyanin pigmentation, particularly on the petals, varies strikingly. The petals may have large spots, blotches, or bands of reddish-purple color near the base, the center, the upper margin, or sometimes in several of these positions. Alternatively, the petals may have hundreds of very small (2-8 cells) red or purple flecks, particularly near the center, or be unspotted. The characteristic lav- ender to reddish-purple pigments have been identified as glycoside derivatives of malvidin, supplemented with derivatives of cyanidin and delphinidin (Soltis 1986). The large spots or flecks appear to result from locally elevated levels of the same pigments, though in different proportion than in the petal background (Dom and Bloom 1984). Although the petals are generally pigmented throughout, those of many species also have large white areas of no pigmentation. Anthocyanins may also be present or absent on filaments and an- thers, the stigma, and the floral tube. Petal Pattern in Clarkia gracilis The petal pigmentation patterns of species assigned to Godetia and Rhodanthos can be divided into four major types: central spot only, distal spot only, basal band only, and unspotted. In addition, the petals of several species have more than one large pigmented spot. Since nearly all Clarkia species are strongly reproductively isolated, the genetic basis of the petal patterns cannot be directly studied. However, it has been possible to carry out a genetic study between two subspecies of Clarkia gracilis, an allotetraploid species in Rhodanthos, derived from the diploid C amoena subsp. huntiana and an extinct species related to C. lassenensis and C. arcuata (Abdel- Hameed and Snow 1972). Clarkia gracilis includes four interfertile subspecies that, as a group, display three of the four basic petal patterns described above. Sub- 1989] GOTTLIEB: FLORAL PIGMENTATION IN CLARKIA 3 Fig. 1 . (A) Flowers of C. gracilis subsp. gracilis (left), and C. gracilis subsp. sono- mensis (right); (B) Rowers of F2 recombinant with large flower and basal spot, and an F2 recombinant with small flower and central spot. Species sonomensis has large pink petals each with a large central red-purple spot, subspp. albicaulis and tracyi have large pink petals each with an intense red-purple band of color across the base, and subsp. gracilis has small unspotted pink petals. The three subspecies with large petals have a marked protandry (the pollen is relased 3- 5 days before the stigma becomes receptive) and are outcrossing. Subspecies gracilis is predominantly self-pollinating because when its flowers open, the stigma is already receptive and at the same height as the adjacent anthers. The four subspecies have a high genetic identity (Holsinger and Gottlieb 1988) and a moderate amount of chromosomal structure divergence (Abdel-Hameed and Snow 1968, 1972). The petals of subsp. sonomensis are about 2.5 times longer and 4 times wider than those of subsp. gracilis. Numerous genes probably contribute to the size difference because the petal sizes of both par- ents were not recovered in a large F2 (238 plants) or backcross progenies from hybrids between them (Gottlieb and Ford 1988). Although petal size behaves like a classical quantitative trait, the presence/absence of petal spots is controlled by a single gene. Thus a recent genetic analysis (Gottlieb and Ford 1988) revealed that subsp. gracilis has a gene governing basal petal spot that is not normally expressed (Fig. 1). The gene is allelic to one for central spot in subsp. sonomensis, but is not expressed because it is inhibited by a gene at another locus (Gottlieb and Ford 1988). Allelism of central petal spot, basal "band" (see below), as well as unspotted petal had previously been reported in the related diploid C rubi- cunda (Rasmuson 1921; Hiorth 1940). Additional genes modify the size and exact position of the central spot. For example, the width of the central spot can vary from very narrow, with only a few dozen files of pigmented cells, to a broad blotch of color about half or more the width of the petal. The genetics of these modifications have not yet been analyzed. 4 MADRONO [Vol. 36 GODETIA Nit Dav* Pur* Spe/i Ten* RHODANTHOS Arc Amo/ 1 Gra/g* (Gra/g*) Nit Pro* Pur* Spe/p,s Wll Amo/a, c , h, w Gra/s* Wll Arc Fra Gra/a*, t* Las Rub Imb Pur* Put* Fig. 2. Diagram of petal pigmentation patterns in species ofClarkia sections Godetia and Rhodanthos. An asterisk indicates the species is polyploid. Names of the taxa are abbreviated as follows: C. arcuata (ARC); C. amoena subsp. amoena (AMO/a), subsp. caurina (AMO/c), subsp. huntiana (AMO/h), subsp. lindleyi (AMO/1), subsp. whit neyi {AMO C. davyi(D\N)\ C.fmnciscana (FRA); C. gracilis subsp. albicaulis (GRA/a), subsp. gracilis (GRA/g), subsp. sonomensis (GRA/s), subsp. tracyi (GRA/ t); C. imbricata (IMB); C. lassenensis (LAS); C. nitens (NIT); C. prostrata (PRO); C. purpurea (PUR); C. rubicunda (RUB); C. speciosa subsp. immaculata (SPE/i), subsp. polyantha (SPE/p), subsp. speciosa (SPE/s); C. tenella (TEN); C. williamsonii (WIL). Petal Pattern in Related Clarkia Species Knowledge of the genetic basis of petal pigmentation in C gracilis suggests that many of the petal patterns that distinguish other species in Godetia and Rhodanthos may also be governed by single genes. The petal patterns of all the species in the two sections are dia- grammed in Figure 2. Most species are spotless, or have a central spot or a basal band of color. Two species have a wedge-shaped spot that extends from the center of the petal to its upper margin. Four species (C. arcuata, C. nitens, C. purpurea, and C. williamsonii) are polymorphic for petal pattern. Of them, the most polymorphic is C. purpurea, an allohexaploid that has at least five common patterns, all frequently observed in the same population. The many variants of C. purpurea probably reflect its hexaploidy and the consequences of occasional hybridization followed by sorting out of new homozygous genotypes by virtue of its self-pollinating breeding system. Lewis and Lewis (1955, p. 304) reported a cross between the small-flowered C. purpurea subsp. quadrivulnera and the large-flowered subsp. viminea that showed continuous segrega- tion for petal size in an F2 but sharp segregation for petal color and pattern, similar to results obtained in C gracilis (Gottlieb and Ford 1989] GOTTLIEB: FLORAL PIGMENTATION IN CLARKIA 5 1988). Apparently homologous genes affecting floral pigmentation in diploid and tetraploid derivatives have been described in Primula (De Winton and Haldane 1933) and Gossypium (Harland 1935). Clarkia williamsonii often exhibits both central spot and basal band on its petals. Since the two are known to be allelic in another Clarkia (Rasmuson 1921; Hiorth 1940), their combined presence in a true-breeding pattern in a diploid species is consistent with duplication of the coding genes, analogous to duplication of genes encoding isozymes described for a number of diploid Clarkia species (Gottlieb 1977; Pichersky and Gottlieb 1983). Alternatively, the multiple spots may reflect activation of a single gene at diflerent times during petal development. Since clarkias display only a small number of discrete petal pat- terns, and since differences in several of them are allelic, it is plausible that the presence of the same petal pattern reflects utilization of the same subset of genes, and that different patterns result from simple allelic substitutions. Many additional genes presumably contribute to the details of fforal pigmentation, for example, the particular pigment mix, the size and shape of the spots (round versus wedge- shaped) and their exact position on the petal. The inheritance of the major components of fforal pattern, for example, spot presence/ absence and spot position, however, need not be regarded as com- plex. New patterns can be assembled by substituting alleles at a small number of loci, not unlike the situation described for cuticular patterns of chaetae and trichomes in Drosophila (Garcia-Bellido 1983). The Basal Spot Gene in subsp. gracilis The recovery of an allele for basal spot from the unspotted subsp. gracilis poses interesting questions about its origin. The basal band of color in the related subspecies, albicaulis and tracyi, extends com- pletely across the petal base whereas the subsp. gracilis spot is small, more or less round, and does not extend to the edges of the petal. The gene for large central spot in subsp. sonomensis most likely comes from C amoena subsp. huntiana which also has a similar central petal spot, and the gene for basal band characteristic of subspp. albicaulis and tracyi probably derives from a species related to the diploids C lassenensis and C arcuata which have a basal petal band. The two other diploid species in Rhodanthos also have a basal petal band. Natural hybridization between subsp. gracilis and subsp. sono- mensis has been documented (Lewis and Lewis 1955, p. 280), and it is interesting that plants with small petals and central spots that otherwise resemble subsp. gracilis were identified. Were some plants to show the novel basal spot, an observer would not know that its 6 MADRONO [Vol. 36 coding gene already existed, and was simply "released" from inhi- bition. Many loci may include alleles that normally remain unex- pressed, and segregation following hybridization, which is frequent in plants, may place such alleles as well as normally expressed alleles under new patterns of regulation resulting in the abrupt appearance of new forms. Though they might seem like macromutations, the novel phenotypes could be simply explained. Novel phenotypes would also occur when a rare recessive allele that governs a mor- phological trait in an outcrosser becomes homozygous and is ex- pressed following hybridization with a related selfer (Rick and Smith 1953). The progenies between subsp. gracilis and subsp. sonomensis also segregated for several other differences in their pigmentation patterns (Gottlieb and Ford 1988). The background petal color of subsp. sonomensis is frequently uniform. However, in some individuals, the basal quarter of each petal lacks anthocyanin pigments and is bright white, giving the appearance of a "white cup," especially when the flower is newly opened. Presence versus absence of white cup proved to be governed by a single locus, with white cup (representing absence of pigment) recessive. A single gene was also identified that governs presence versus absence of dark red pigmentation on the inner surface of the floral tube, and another gene was found that controls presence versus absence of color on the anthers and fila- ments. In all, five genes that affect pattern of anthocyanin color on the petals and other floral organs were identified in addition to the polygenic control of petal size. The white cup pattern is present in many Clarkia species including C. arcuata, C. bottae, C. davyi, and C imbricata as well as in C. gracilis subspp. albicaulis and tracyi, and its genetic basis may be similar to that found in subsp. sono- mensis. Presence or absence of color on the floral tube also distin- guishes Clarkia species. Conclusions It is not known whether the differences in pigmentation pattern of Clarkia species were selected by different pollinators. MacSwain et al. (1973) believed that the differences in petal spotting may play a role in determining pollinator constancy on individual trips from the nest, thus serving to increase the frequency of intraspecific pol- linations. Some of the novel variants of C gracilis constructed during the genetic study may be useful to test hypotheses concerning floral pattern and pollination system. Appropriate variations can be in- troduced into different habitats and monitored to observe pollinator preferences. In such studies, it would be possible to focus on indi- vidual traits such as spot position or white cup or on different trait combinations. 1989] GOTTLIEB: FLORAL PIGMENTATION IN CLARKIA 7 The discovery that subsp. gracilis has a gene for basal petal spot that is not normally expressed because of the action of a second gene emphasizes the importance of genetic studies for understanding mor- phological differences between species. This may be particularly sig- nificant in plants where interspecific hybridization and allopoly- ploidy are prevalent. Many novel phenotypes in plants may result from new modes of gene regulation brought about by the juxtapo- sition of divergent genetic materials rather than by accumulation of novel genes. The pigmentation pattern on petals and other floral organs com- monly varies among species in many plant genera. Important genetic studies in cultivated plants include the analysis of Japanese morning glories which revealed a large number of single genes that confer sharply distinct pigment patterns on the petals such as blizzard, flecked, lined, striated, speckled, and smeary (Imai 1931). Other ge- netic studies were carried out in Primula sinensis (De Winton and Haldane 1933). Few studies have been done in wild plants although many reports are available of pattern differences between and within species. A recently published example is Platystemon californicus, which exhibits at least six color patterns on its showy flowers that appear to be roughly correlated with geographical distribution (Han- nan 1981). A number of species with flower color polymorphism are listed in Hannan (1981) and Kay (1978). The flower is a complex structure in which many specialized tis- sues and cell types form distinctive organs in a precise and orderly manner. The differentiation of structures is most likely independent of pigmentation pattern, and this is one reason the patterns may be changed by few genes. Although the patterns may have a simple and readily modified developmental basis, pigment patterns are likely to have complex effects on pollination and eventual seed set. Changed patterns appear to be particularly important in annual plants like Clarkia in which most species are outcrossing, and often as many as five or six species grow intermixed beneath the oak trees on the same Sierran hillside. Literature Cited Abdel-Hameed, F. and R. Snow. 1968. Cytogenetic studies in Clarkia, sect, pri- migenia. IV. A cytological survey of Clarkia gracilis. Amer. J. Bot. 55:1047- 1054. and . 1 972. The origin of the allotetraploid Clarkia gracilis. Evolution 26:74-93. De Winton, D. and J. B. S. Haldane. 1933. The genetics of Primula sinensis. II. Segregation and interaction of factors in the diploid. J. Genetics 27:1-44. DoRN, P. S. and W. L. Bloom. 1984. Anthocyanin variation in an introgressive complex in Clarkia. Biochem. Syst. Ecol. 12:311-314. Garcia-Bellido, a. 1 983. Comparative anatomy of cuticular patterns in the genus Drosophila. Pp. 227-259 in B. C. Goodwin, N. Holder, and C. C. WyUe (eds.), Development and evolution. Cambridge Univ. Press, Cambridge. 8 MADRONO [Vol. 36 Gottlieb, L. D. 1977. Evidence for duplication and divergence of the structural gene for phosphoglucose isomerase in diploid species of Clarkia. Genetics 86: 289-307. . 1984. Genetics and morphological evolution in plants. Amer. Naturalist 123:681-709. and V. S. Ford. 1988. Genetic studies of the pattern of floral pigmentation in Clarkia gracilis. Heredity 60:237-246. and N. F. Weeden. 1979. Gene duplication and phylogeny in Clarkia. Evolution 33:1024-1039. Hannan, G. L. 1981. Flower color polymorphism and pollination biology of Platy- stemon californicus Benth. (Papaveraceae). Amer. J. Bot. 68:233-243. Harland, S. C. 1935. Homologous genes for anthocyanin pigmentation in new and old world cottons. J. Genetics 30:465-476. HiORTH, G. 1940. Fine Serie multipler Allele fur Bliitenzeichnungen bei Godetia amoena. Hereditas 26:441-453. HoLSiNGER, K. and L. D. Gottlieb. 1988. Isozyme variability in the tetraploid Clarkia gracilis (Onagraceae) and its diploid relatives. Syst. Bot. 13:1-6. Imai, Y. 1931. Analysis of flower color in Pharbitis nil. J. Genetics 24:203-224. Kay, Q. O. N. 1978. The role of preferential and assortive pollination in the maintenance of flower color polymorphisms. Pp. 1 75-1 90 in A. J. Richards (ed.). The pollination of flowers by insects. Academic Press, London. Lewis, H. 1953. The mechanism of evolution in the genus Clarkia. Evolution 7: 1-20. . 1962. Catastrophic selection as a factor in speciation. Evolution 16:257- 271. . 1973. The origin of diploid neospecies in Clarkia. Amer. Naturalist 107: 161-170. and M. E. Lewis. 1955. The genus Clarkia. Univ. CaUf Publ. Bot. 20:241- 392. MacSwain, J. W., P. H. Raven, and R. W. Thorp. 1973. Comparative behavior of bees and Onagraceae. IV. Clarkia bees of the western United States. Univ. Calif Publ. Entom. 70:1-80. Odrzykoski, I. J. and L. D. Gottlieb. 1984. Duplications of genes coding 6-phosphogluconate dehydrogenase in Clarkia (Onagraceae) and their phyloge- netic implications. Syst. Bot. 9:479-489. PiCHERSKY, E. and L. D. Gottlieb. 1983. Evidence for duplication of the structural genes coding plastid and cytosolic isozymes of triose phosphate isomerase in diploid species of Clarkia. Genetics 105:421-436. Rasmuson, H. 1921. Beitrage zu einer genetischen Analyse zweier Godetia- Axi^n und ihrer Bastarde. Hereditas 2:143-289. Rick, C. M. and P. G. Smith. 1953. Novel variation in tomato species hybrids. Amer. NaturaHst 87:359-373. SoLTis, P. S. 1986. Anthocyanin variation in Clarkia (Onagraceae). Biochem. Syst. and Ecol. 14:487-489. , D. E. SoLTis, and L. D. Gottlieb. 1987. Phosphoglucomutase gene dupli- cations in Clarkia (Onagraceae) and their phylogenetic implications. Evolution 41:667-671. Sytsma, K. J. and L. D. Gottlieb. 1986a. Chloroplast DNA evolution and phy- logenetic relationships in Clarkia sect. Peripetasma (Onagraceae). Evolution 40: 1248-1261. and . 1986b. Chloroplast DNA evidence for the origin of the genus Heterogaura from a species of Clarkia (Onagraceae). Proc. Nat. Acad. Sci. U.S.A. 83:5554-5557. (Received 27 Jul 1988; revision accepted 7 Oct 1988.) THE GENERIC POSITION OF MYRTUS ALTERNIFOLIA AND NOTES ON CALYCOLPUS (MYRTACEAE) Leslie R. Landrum Department of Botany, Arizona State University, Tempe, AZ 85287 Abstract Myrtus alternifolia, a species of long standing uncertain generic affinities, is trans- ferred to Calycolpus. This transfer is based on characteristics of the seed coat, embryo, calyx, anthers, and vessel elements. A revised description of Calycolpus is provided. Resumen Myrtus alternifolia, una especie que desde hace mucho tiempo ha sido de afinidades genericas inciertas, se transfiere a Calycolpus. Este cambio se basa en caracteristicas de la testa de la semilla, el embrion, el caliz, las anteras, y los elementos vasculares. Se provee una descripcion revisada de Calycolpus. Myrtus alternifolia Gleason was described in 1939 and placed in Myrtus L. because Gleason believed it to be congeneric with species of Myrtus {M. myricoides Kunth, M. stenophylla Oliver ex Thum, M. roraimensis N. E. Brown) now considered to belong to Ugni Turcz. McVaugh (1969) in his revision of the Myrtaceae of the Guayana Highland, recognized that Myrtus alternifolia was not a species of Ugni and that it probably did not belong to Myrtus, but lacking any clear evidence of its true affinities, did not transfer it to any genus. Myrtus, the type genus of the Myrtaceae, was once very inclusive, accommodating perhaps a quarter of the family in the early nine- teenth century. Since then a progressively smaller percentage of species has been assigned to Myrtus. I agree with McVaugh (1968) that Myrtus probably does not include any American species. Scott (1 979) and Byrnes (1982) have expressed the opinion that certain Austral- asian species belong to Myrtus, but if those are excluded, the genus consists of only the Mediterranean Myrtus communis L. Myrtus alternifolia differs from M. communis in many morphological char- acters, most notably in embryo structure. The two species are also geographically widely separated. Thus, extraction of Myrtus alter- nifolia from Myrtus is warranted. I am now monographing American genera of the Myrtinae with hard seed coats, the group to which Myrtus alternifolia belongs, and have considered the following genera to be possible acceptors for M. alternifolia: Calycolpus Berg, Mosiera Small, Myrteola O. Berg, Psidium L., and Ugni. Thanks to a collection of M. alternifolia with Madrono, Vol. 36, No. 1, pp. 9-14, 1989 10 MADRONO [Vol. 36 mature seeds recently made by Bruce Hoist of the Missouri Botanical Garden and a survey of seed coat characters in the Myrtinae un- dertaken at Arizona State University (Landrum and Sharp in press), I am able to assign this species to Calycolpus. A description of the species and reasoning for this generic assignment are given below. Calycolpus alternifolius (Gleason) Landrum, comb. noY. —Myrtus alternifolia Gleason, Brittonia 3:173. 1939.— Type: Venezuela, Bolivar, Auyan-tepui, 2200 m, Tate 1344 (holotype, NY!). Shrub to 2 m high, sometimes with trailing stems to 5 m long, the young growth mostly densely lanate to velutinous; hairs simple, to ca. 1.5 mm long, silvery-grey to yellowish-grey, straight or curled, spreading to ascending. Young twigs lanate, the hairs mainly per- sisting until the first bark falls, the older twig bark rough and cracked, light grey to blackish-grey, sometimes tinged with reddish-brown. Leaves opposite, or less often alternate, mostly separated by short internodes, elliptic, narrowly elliptic, obovate, ovate, or suborbic- ular, flat or revolute, 1.6-3.6 cm long, 0.6-2.5 cm wide, 1.4-3 (-4.6) times as long as wide, densely covered with hairs beneath, the upper surface tomentose in immature leaves, later glabrescent; apex acute to abruptly acuminate; base acute to rounded; petiole narrowly channeled, 1-3 mm long, 1-2 mm thick, lanate when young, losing most hairs with age; midvein impressed above, prominent below; lateral veins indistinct, or ca. 8-12 nearly straight, ascending pairs, scarcely visible, the marginal vein about equalling the laterals, about parallel to the margin; blades thickly coriaceous, drying olive- green to nearly black, lustrous above. Flowers pentamerous; pedun- cles 6-30 mm long, ca. 1 mm wide, solitary in the axils of leaves, uniflorous, densely lanate; bracteoles narrowly triangular to linear, 3-4 mm long, caducous at about anthesis; calyx-lobes triangular, 3- 5 mm long, 1.5-3 mm wide, moderately pubescent within, lanate without; petals elliptic-obovate, 8-9 mm long, glabrous or with a ciliate margin; hypanthium lanate, obconic, 3-4 mm long; disk 2- 4 mm across, tomentose; stamens 38-70, 3-5 mm long; anthers oblong, 0.8-1 mm long, with ca. 5-9 nearly equal glands in the connective; style ca. 5 mm long, glabrous to sparsely pubescent; ovary 2- to 3-locular; ovules 10-18 per locule, the placenta, a raised U-shaped pad. Fruit globose, tomentose, crowned with a persistent calyx, ca. 5-7 mm in diam.; seeds ca. 6, ca. 3 mm long, the seed coat lustrous, mostly 1 cell thick, the embryo whitish, C-shaped, the cotyledons not reflexed, accounting for less than V4 the embryo's length. Analysis of Relationships The characters of the following structures have been taken into account in my attempt to assign a generic position to Myrtus alter- 1989] LANDRUM: MYRTUS ALTERNIFOLIA 11 nifolia: 1) seed coat (surface; and outer seed coat thickness and prevalent cell shape in the outer seed coat); 2) embryo (cotyledon to hypocotyl ratio and cotyledons reflexed or not); 3) calyx-lobes (number and fusion); 4) anthers (shape, glandularity, and size in relation to the filament); and 5) vessel elements (with or without scalariform perforation plates). Seed coat. In Psidium the seed coats a) are not lustrous but dull, b) have an external layer of pulpy tissue, c) have outer seed coats normally over 8 cells thick, and d) have outer seed coats in which the prevalent cell shape is elongate (Landrum and Sharp in press). Myrtus alternifolia, Calycolpus, Mosiera, Myrteola, and Ugni all differ in having seed coats that a) are lustrous, b) normally lack an external layer of pulpy tissue, c) have outer seed coats usually 1 to 4 cells thick, and d) have outer seed coats in which the prevalent cell shape is more or less isodiametric to oblong. These characters of the seed coat are the most useful in circumscribing Psidium and exclude M. alternifolia from the genus. Embryo. The cotyledons of Myrteola and Ugni are about as long as the hypocotyl and are not reflexed. In the other genera and Myrtus alternifolia they are shorter than the hypocotyl; in Psidium and in some species of Mosiera and Calycolpus the cotyledons are reflexed. In Myrtus alternifolia they are not reflexed. Based on embryo struc- ture then, M. alternifolia might belong to Mosiera or Calycolpus. Calyx-lobes. Mosiera has 4 calyx-lobes, Myrteola has 4 or 5, and the rest of the genera and Myrtus alternifolia normally have 5. Thus, calyx-lobe number indicates that Myrtus alternifolia does not belong to Mosiera, but might belong to any of the other genera. Usually there is notable fusion of the calyx-lobes beyond the ovary's summit in Psidium, and there is often such fusion in Calycolpus. In Myrtus alternifolia and the other genera here considered, fusion is lacking. This character indicates that Myrtus alternifolia probably does not belong to Psidium. Anthers. Anthers of most species of American Myrtaceae (includ- ing Mosiera and Myrteola) are subglobose and have no gland or a prominent terminal gland. In Myrtus alternifolia, Calycolpus, Ugni, and some species of Psidium the anthers are more or less elongate and have a few to several glands in the connective. In Ugni the mature filament is 1 to 3 times as long as the anther. In other species in this group the mature filament is more than 3 times as long as the anther. Thus, anther glandularity and proportional size indicate that Myrtus alternifolia should probably not be placed in Mosiera, Myrteola, or Ugni. Vessel elements. In Ugni and Myrteola vessels have (or mostly have) scalariform perforation plates (Schmid and Baas 1984). Myrtus 12 MADRONO [Vol. 36 alternifolia does not have scalariform perforation plates. One species of M osier a (cited as Psidium longipes [O. Berg] McVaugh) and Psid- ium pubifolium Burret studied by Schmid and Baas were found to have simple perforation plates. The condition of other species of Mosiera, Psidium, and Calycolpus is not known to me. This character indicates that Myrtus alternifolia is not a member of Ugni nor Myr- teola. In summary Psidium is eliminated as a possible acceptor for Myr- tus alternifolia because of seed coat characteristics. Elimination of Ugni and Myrteola seems warranted because of characters of the anthers, embryo, and vessel elements. Myrtus alternifolia would be out of place in Mosiera because of the glandularity of the anthers and calyx-lobe number. Calycolpus is not eliminated by any char- acteristic considered and the addition of Myrtus alternifolia does not add greatly to its variability. Moreover, Myrtus alternifolia has leaves that are thickly coria- ceous, and although the venation is faint, it can be seen to consist of several nearly straight lateral veins that connect with a discrete marginal vein. This is the common leaf type in Calycolpus. Another factor that favors placing Myrtus alternifolia in Calycol- pus is that it grows on the mountain peaks (tepuis) of the Guayana highlands, the area in which Calycolpus is most diverse. Mosiera, perhaps the second most suitable acceptor for Myrtus alternifolia, is a Caribbean and Central American genus. Recharacterization of Calycolpus Until now Calycolpus has been thought to be morphologically a rather uniform group, recognized by characteristics of the inflores- cence (flowers usually borne in pairs on short axillary shoots), calyx ("open in bud and flower, with 5 broad and often flaring lobes"), and ovary (4- to 5-locular) (McVaugh 1968). Recent studies of Ca- lycolpus indicate that none of these characteristics is as consistent as was previously thought and that seed coat and embryo characters indicate a somewhat more inclusive group than was accepted by McVaugh. In addition to Myrtus alternifolia, I would also include in Calycolpus, Psidiopsis moritziana Berg, which McVaugh (1956) placed in Psidium. Even without Myrtus alternifolia and Psidiopsis moritziana there has been no particular characteristic found in all species of Caly- colpus that is not also found in other genera. Calycolpus, at least at present, must be defined by a set of characters: seed coats lustrous, with no pulpy covering, the outer seed coat 1-4 cells thick with the prevalent cell shape more or less isodiametric; cotyledons shorter than the hypocotyl, reflexed or not; calyx-lobes normally 5, fused 1989] LANDRUM: MYRTUS ALTERNIFOLIA 13 or not beyond the ovary's summit; anthers elongate, with a few to several glands in the connective, about V3 to V\o as long as the fila- ment; vessel elements with simple perforation plates. The following is an updated description of Calycolpus based on studies of the following species: C. alternifolius, C. calophyllus (Kunth) O. Berg, C. cochleatus, C. goetheanus (DC.) Berg, C. legrandii Mattos, C moritzianus (O. Berg) Burret, C revolutus (Schauer) O. Berg, C. roraimensis Steyerm., C. surinamensis McVaugh, and C warsze- wiczianus O. Berg. Calycolpus O. Berg, Linnaea 27:378. 1856. Lectotype species: C. goetheanus (DC.) O. Berg. Designated by Riley (1926). Shrubs or trees up to 10(-15) m high. Hairs whitish or yellowish, unicellular, simple or dibrachiate, up to ca. 1.5 mm long. Leaves persistent, coriaceous, the venation brochidodromous, with several pairs of nearly straight lateral veins that are united by a marginal vein that parallels the leaf margin. Inflorescence a solitary axillary flower or a very short axillary bracteate shoot with 1-3 decussate pairs of flowers. Flowers pentamerous (sometimes tetramerous in one population of C cochleatus); calyx-lobes often flared, often with an apical appendage, the calyx fused beyond the ovary's summit and tearing between the lobes at anthesis or the calyx-lobes free; petals slightly fleshy, white or tinged with red, often drying brown; bracteoles usually small, about triangular, caducous at about anthesis or in C. legrandii leafy and persistent; stamens 35-270, folded cen- terward or more or less erect in the bud; anthers somewhat to mark- edly elongate, with about 4 to 40 glands; ovary 2- to 6-locular, the locular wall sometimes glandular; ovules 8-32(-ca. 80 in C. mor- itzianus), the placenta a U-shaped pad or mound of tissue or an essentially round peltate structure. Fruit subglobose; seeds few to numerous, about reniform, the seed coat hard, lustrous, the external wall 1-4 cells thick, the surface cells rounded to elongate, the central portion of the seed sometimes soft; embryo oily, whitish, C-shaped, the cotyledons reflexed or straight, less than V4 the length of the embryo. Acknowledgments Special thanks go to Bruce Hoist who sought out a specimen of Calycolpus alter- nifolius with mature seeds on the Auyan-tepui in Bolivar, Venezuela. I am also grateful to the curators of the following herbaria for allowing me to include their collections in my studies of Calycolpus: A, ASU, CAS, GH, MICH, MO, NY, UC, and US. Bruce Hoist, David Keil, Rudolph Schmid, and Al Gentry offered helpful suggestions as to how this paper could be improved. Literature Cited Byrnes, N. 1982. The genus Myrtus or Austromyrtus in Australia? Austral. Syst. Bot. Newslett. 31:10-11. 14 MADRONO [Vol. 36 Landrum, L. R. and W. P. Sharp. 1989. Seed coat characters of some American Myrtinae (Myrtaceae): Psidium and related genera. Syst. Bot. 14 (in press). McVaugh, R. 1956. Tropical American Myrtaceae. Notes on generic concepts and descriptions of previously unrecognized species. Field Mus. Nat. Hist., Bot. Ser. 29(3): 145-228. . 1968. The genera of American Myrtaceae— an interim report. Taxon 17: 354-418. . 1969. Myrtaceae. In The Botany of the Guayana Highland— Pt. 8. Mem. New York Bot. Card. 18(2): 5 5-286. Riley, L. A. M. 1926. Revision of the genus Calycolpus. Kew Bull. 1926:145-154, ScHMiD, R. and P. Baas. 1984. The occurrence of scalariform perforation plates and helical vessel wall thickenings in wood of Myrtaceae. Int. Assoc. Wood Anat. Bull. 5(3): 197-2 15. Scott, A. J. 1979. New species and combinations in Myrtaceae from Malesia and Australia. Kew Bull. 33:51 1-515. (Received 21 Jun 1988; revision accepted 25 Oct 1988.) TAXONOMIC RELATIONSHIPS OF GILIA MACULATA (POLEMONIACEAE) Robert Patterson Department of Biology, San Francisco State University, San Francisco, CA 94132 Abstract Gilia maculata is reassigned from its previous placement in Linanthus. It was described originally by Parish in 1892 as Gilia maculata, and placed in Linanthus by Milliken (1904). This species is poorly known because of its rarity and because of its very small size, causing it to be easily overlooked in the field. Rediscovery of a population of G. maculata provided the opportunity to study this taxon critically. Leaf arrangement and shape, indumentum, corolla and calyx morphology, and pollen exine morphology, argue against its unequivocal assignment to Linanthus, and favor its inclusion in Gilia. Gilia maculata Parish (Fig. 1) is a systematic enigma. It is a minus- cule, little-known desert annual that occurs near the western margins of the Little San Bernardino Mountains of southern California (Fig. 2). It has been regarded by most floristic treatments as Linanthus maculatus (Parish) Milliken, although it has few diagnostic features of Linanthus. Although botanists and governmental agencies have sought it because of its potentially rare status, it has seldom been seen or collected, and few specimens are present in herbaria, making study extremely difficult. Furthermore, its relationships with other species of Linanthus or Gilia, as well as with other Polemoniaceae, have never been examined critically. In April 1986, a substantial population of this species was located near the northwest entrance to Joshua Tree National Monument, providing enough material to conduct a more thorough study of the morphological relationships of this species. Taxonomic History of Gilia maculata Parish's (1 892) original description of G. maculata was as follows: "Inch high, diffiisely branched from the base, sparsely pubescent; leaves entire, two lines long, broadly linear, thick and strongly car- inate, obtuse, acerose; earlier flowers nearly sessile in the lower forks, later ones crowded above; calyx lobes nearly equal, much like the leaves but with a narrow hyaline membrane, ciliate; the narrowly campanulate tube of the corolla not exceeding the calyx, the limb rotate, two lines wide; filaments inserted on the base of the tube; anthers exserted; seeds few". Parish noted that the species was . . near G. demissa Gray, from which it differs in its entire leaves, obtuse Madrono, Vol. 36, No. 1, pp. 15-27, 1989 16 MADRONO [Vol. 36 1 mm Fig. 1 . Gilia maculata. A. Habit. B. Enlargement of trichomes. C. Flower at early anthesis. D. Calyx lobe showing hyaline margins. E. Face view of flower showing position of spots at base of corolla lobes. and ciliate calyx-lobes, narrower corolla, and exserted anthers". The implication is that Parish considered the two species related based on an overall resemblance. It is not surprising that Parish did not recognize the new species as a member of Linanthus, because at that time most species recognized currently as Linanthus were included 1989] PATTERSON: GILIA MACULATA 17 7 San Bernardino County Riverside County l-T 10 Km Fig. 2. Location of 1986 collection (M) of Gilia maculata {Bourell et al. 3000). in Gilia. Although the genus name Linanthus dates from 1833, its common use did not begin until Greene's (1892) treatment. Parish's original description is not completely in accord with the holotype or with material collected in the field during this study. Contrary to his description, the leaves of the holotype are, in fact, obovate and certainly not acerose (it is possible that Parish was referring to a mucronate tip, which may be present in some speci- mens). Parish also omitted certain other features that distinguish G. maculata from other genera in the family. He neglected to state in his description that the leaves of this taxon were alternate. He did, however, describe the calyx lobes as ciliate, identifying an important character that sets G. maculata off taxonomically. Milliken (1 904) placed this species in Linanthus, although without any explicit justification. Moreover, her description of L. maculatus is not in complete accord with her inclusive description of the genus Linanthus. She described the leaves of L. maculatus as . . entire, the upper sometimes alternate, oblong . . .", whereas her description of the leaves for the genus reads "... opposite and palmately parted, or rarely entire and linear". Thus, although Milliken's treatment dictated the taxonomic status under which this species has been recognized in all modern floras, it failed to distinguish this species clearly as a member of Linanthus. If, in fact, Milliken's key to genera of Polemoniaceae were used, G. maculata would be identified clearly as a member of Gilia. Brand's (1907) concept of Gilia included most of the tribe Gilieae, including Linanthus. He recognized G. maculata as a member of sect. Campanulastrum, along with G. parryae and G. bella (=Li- nanthus p. and b.), Gilia dactylophyllum i=L. demissus), and three currently recognized species of Gilia, G. campanulata, G. filiformis, and G. micromeria. This treatment is particularly noteworthy be- 18 MADRONO [Vol. 36 cause it is the earliest occasion where G. maculata is alHed with G. campanulata. Phenetically G. maculata is probably most closely related to G. campanulata, although there are still substantial dif- ferences that distinguish them. Grant (1959) later combined sect. Campanulastrum with sect. Giliastrum under the latter name, al- though without providing justification. Subsequent treatments continued to recognize this taxon as Li- nanthus without apparent concern for accuracy of the taxonomic placement. Jepson (1925, 1943) recognized L. maculatus but did not refer directly to its alternate leaves. Only in the genus description did he allude to Linanthus as having leaves "rarely with some up- permost alternate". Interestingly, in his Manual (1925) he placed L. maculatus with L. demissus and L. parryae in subgenus Parrya. Later, in his Flora of California (1943) Jepson included L. bellus, L. concinnus, and L. dianthiflorus in this subgenus, circumscribing what Grant (1959) later referred to as sect. Dianthoides. It is note- worthy that Jepson (1943) made a special comment in the generic description that L. maculatus (among other species) has entire leaves. He also commented on the narrow endemism shown by the range of L. maculatus. It is curious that, with the extra attention given to this species in his Flora, Jepson did not discuss the significance of alternate leaves in this species. Mason omitted L. maculatus from the entire treatment of the Polemoniaceae in Abram's illustrated Flora of the Pacific States (195 1). It is unclear whether this was an oversight, or whether it was due to a belief on Mason's part that this species did not belong in Linanthus. Munz (1959, 1974) included L. maculatus in his treat- ments, describing it accurately as having alternate leaves, but making no other special mention of this character. Because this species is not well known, and because the original diagnosis is scanty and not in complete agreement with the holotype material, an updated description is provided here based on material from the population collected in 1986: Gilia maculata Parish, Bull. Torrey Bot. Club 19:93. 1892. (figs. 1, 3).— Linanthus maculatus (Parish) Milliken, Univ. Calif Publ. Bot. 2:55. 1904.-Type: USA, California, Riverside Co., bor- ders of the Colorado Desert, at Agua Caliente [Palm Springs], W. G. Wright s.n. (holotype: CAS!). Diminutive ephemeral annual 1-3 cm high. Stems branching above the first 1 or 2 leaves, densely hairy with 1- to 4-celled trichomes throughout. Leaves alternate, fleshy, narrowly oblanceolate or ob- long, sessile, mucronate, marginally ciliate with 1 - to 2-celled, white hairs from the base to at least Vi the length (often farther), the blade concave adaxially. Flowers borne in simple or compound cymes, sessile or subsessile, peduncle <1 mm long; calyx lobes narrowly 1989] PATTERSON: GILIA MACULATA 19 oblanceolate or spatulate with mucronate tip, ca. 2 mm long, green, distinct nearly to base (only the adjacent membranes connected at base), glabrous, with membranous ciliate margins extending to the tip, the trichomes 2(-3) cells long, the terminal cell long-acuminate (Fig. IB); corolla campanulate, tube ca. 1.5-2 mm long, yellow or yellow-green, slightly hairy on inner surface, throat <1 mm long, white, lobes broadly ovate-cordate, tips slightly concave, 1-1.5 mm long, white with cerise spot at base, spreading at right angle to the tube or (more commonly) reflexed, venation simple, open; stamen filaments attached to near base of corolla tube, narrowly lanceolate, 1.5-2 mm long; anthers oval, slightly exserted beyond corolla throat; pollen yellow, round, exine reticulate with 10-12 slit-like apertures distributed evenly on the surface of the grains; ovary triangular- ovate, ca. 0.5 mm long, style 1 mm long, stigma lobes <0.5 mm long. Seeds minuscule, dark reddish-brown, non-mucilaginous, 10- 12 per capsule, ±distributed evenly among locules; n=9. Additional specimens. USA, CA, San Bernardino Co., rd from Joshua Tree to Joshua Tree Natl. Mon., ca. 3.5 km S of junction with CAL Hwy 62, 6 Apr 1986, Bourell, Patterson, and Timbrook 3000 (CAS); Coyote Holes, Joshua Tree National Monument, near line of Riverside and San Bernardino counties, 20 Apr 1924, Munz 7941\ Chipmunk Trail, 28 Mar 1968, Stebbins 6650 (CAS!); 17 mi W of 29 Palms on rd to Morongo Valley, 950 m, 6 Apr 1937, Daniels s.n. (CAS!); ca. 5 mi N of Windmill Tank, 3600 ft, 2 Apr 1942, Ripley and Barneby 4273 (CASiy, S miW of 29 Palms, 12 Apr 1935, Keck 3843 (CAS!). Distribution and ecology. Gilia maculata occurs in moderately coarse sand in open areas of Larrea-Yucca brevifolia scrub as a member of the annual spring flora. It is extremely inconspicuous in its gray-green herbage and white corollas, and blends well with the substrate, even when in flower (Fig. 3). It is likely that this may be one reason why the species is so little known in the field and poorly represented in herbaria. The population that was rediscovered in April 1986 occurs at 1000 m elevation. No other populations were found during this study. It remains uncertain as to whether this population is repre- sentative of other populations of this species; however, considerable area of similar habitat occurs throughout the region. The Joshua Tree population consisted of approximately 100 individuals in April 1986; the following year the population was reduced markedly in number, but individuals were found in the same area. Relationships. The decision to place G. maculata into one of the currently recognized genera of Polemoniaceae or to erect a new genus must be weighed carefully. Any decision is completely dependent 20 MADRONO [Vol. 36 Fig. 3. Gilia maculata in the field. Note size compared with coin. on how well the existing genera are known taxonomically. The Pol- emoniaceae have been studied carefully by many authors, but sys- tematic and ecological relationships among most members of the family are not well understood. Although G. maculata is most com- monly included as a member of Linanthus, even superficial consid- eration of morphological characters does not support this alignment. The most commonly used defining feature for the genus Linanthus within the Polemoniaceae has been the presence of opposite leaves that are either a) palmately-divided with linear or narrowly lanceo- late divisions or b) entire and linear. Presumably the entire leaves in certain species (e.g., L. dichotomus and relatives, L. dianthiflorus) represent a reduction of leaf lobes to one. Linanthus sensu stricto never has completely alternate leaves, although occasionally in some species the upper leaves near the inflorescence are subopposite. In- clusion of G. maculata as a member of Linanthus is out of accord with the morphological unity of the latter; it would be difficult to distinguish Linanthus as a discrete genus were G. maculata included. Bentham (1833), Greene (1892), and Milliken (1904), as well as nearly all subsequent authors, recognized the taxonomic importance of leaf morphology in this lineage, and I find no reason to diminish its value. In addition to having alternate arrangement, the oblong-obovate 1989] PATTERSON: GILIA MACULATA 21 leaves of Gilia maculata represent a shape not found in any other species of Linanthus. This character appears to have been neglected as a distinguishing feature, although it is mentioned in several de- scriptions of the species (Jepson 1925, 1943; Munz 1959, 1974). Most species of Linanthus have linear or linear-lanceolate leaf lobes or leaves. The only species of Linanthus that have oblanceolate leaf lobes are in sect. Leptosiphon (e.g., L. oblanceolatus, L. bicolor)\ however, other morphological differences (corolla shape, leaf ar- rangement and divisions, inflorescence structure) between this sec- tion and G. maculata are so strong that similarity in leaf or leaf lobe shape can be regarded as an example of convergence. Not only is the placement of G. maculata in Linanthus difficult based on the circumscription of the latter genus, there is no apparent morphological alliance between the former species and any existing species of Linanthus. Previously suggested relationships with other species of Linanthus are problematic. Parish (1 892) and Jepson (1 925, 1943) proposed an alliance with L. demissus (sect. Dianthoides) presumably founded on a superficial resemblance in habit and co- rolla morphology. Although both taxa are small desert annuals with white campanulate corollas with reddish basal spots on the lobes, other features do not support a close relationship. Pollen exine pat- terns of these two species are strikingly different (Fig. 4), and provide convincing evidence against a taxonomic alliance. Linanthus de- missus has striate regions amid a reticulate exine, a pattern char- acteristic of certain species of Linanthus sect. Dianthoides. Gilia maculata lacks any striations and is uniformly reticulate, a pattern that occurs in certain species of Gilia and in most other Linanthus species. Inasmuch as pollen exine patterns have been extremely useful in helping to understand relationships in the Polemoniaceae (StuchHk 1967a, b; Taylor and Levin 1975; Chuang et al. 1978; Day and Moran 1986; Timbrook 1986; Patterson, Golden, and Vagenas, unpubl.), this divergence suggests a strong taxonomic difference. Irrespective of relationships between G. maculata and other Gilia species as indicated by pollen morphology, taxonomic placement of G. maculata near L. demissus is not defensible. Additionally, L. demissus has palmately divided, opposite leaves (although upper leaves may occasionally be subopposite, the majority of the leaves remain opposite). Other species within sect. Dianthoides, in which G. maculata was placed by Grant (1959), share few if any diagnostic characters with the latter species. Only L. dianthiflorus has simple leaves, but these are linear and opposite. Deeply cleft calyx tubes are present in L. parryae, L. bellus, and L. demissus, but they are not as deeply cleft as in G. maculata— there is always a fused portion, i.e., a calyx tube. Leaves of L. bellus and L. parryae are always opposite and palmately cleft. One feature shared by G. maculata and most members of sect. 22 MADRONO [Vol. 36 Fig. 4. Scanning electron micrographs of pollen grains. A. Gilia maculata. B. Li- nanthus demissus. C. G. campanulata. D. G. inyoensis. Bar represents 20 mhi. Dianthoides is the presence of red marks at the base of the corolla lobes. In this respect the corolla of L. demissus is similar superficially to that of G. maculata; however, it is likely that this represents convergence in corolla color pattern. Presence of red spots on corolla lobes is common in many species of Linanthus, and is present as a character in sects. Pacificus and Leptosiphon, as well as in Dian- thoides. An argument in support of corolla color pattern as indicative of close relationship in this case negates the importance of other features such as leaf arrangement and shape. The latter characters have had a major role in distinguishing genera within the family, but corolla color patterns have rarely been regarded as important generic diagnostics. As Grant (1965) points out, although evolution of floral morphology (including color patterns) has been a major factor in speciation in the Polemoniaceae, it is apparent that similar color patterns have evolved more than once across generic lines. Gilia is the only genus in which G. maculata can readily be in- corporated based on comparative morphology. Gilia is a large, mor- 1989] PATTERSON: GILIA MACULATA 23 phologically diverse, polythetic genus. Its circumscription is difficult, and it is most easily recognized by lacking characters that are present in other genera. Leaves are always alternate in Gilia, but leaf, tri- chome, and floral morphology in Gilia is extremely diverse. The basic chromosome number in the genus is x=9, as it is for Linanthus (Grant 1959; Patterson 1979). Gilia was partitioned into five sections (Table 1) by Grant (1959), each of which is morphologically and ecologically diverse. Gilia maculata has morphological features that ally it with members of sect. Giliastrum Brand. This section ranges from perennials such as G. ripleyi to diminutive annuals like G. campanulata and G. in- yoensis. It is also poorly understood from a taxonomic viewpoint (Grant 1965). One character that distinguishes it from other sections of Gilia is the presence mostly of campanulate or rotate corollas. This feature is present not only in G. maculata, but also in three other small desert annuals: G. campanulata, G. inyoensis, and G. tenerrima. Another character by which these four species are allied is calyx morphology. The calyx is divided into five lobes to near the base (the lobes actually appearing distinct), with membranous margins that extend most to all of the length of the calyx lobes (Fig. ID). This feature is absent in other species of Gilia. Pubescence features also appear to ally these species while illus- trating the complex interrelationships among them. All four species are moderately to densely pubescent on their stems, leaves, and calyx lobes. The trichomes are generally 2- to 4-celled long, and uniseriate. They show a further similarity among these species in that cells appear to alternate in orientation with respect to one another, form- ing a "chain link" structure (Fig. IB). Slightly different trends in cell number exist among different species and on different organs, but irrespective of these differences, trichome morphology provides an additional argument for including G. maculata within sect. Gilias- trum. Despite similarities among Gilia maculata and the other three species cited above, it is notably distinct in other characters. In particular, no other species of Gilia have ciliate leaf margins and calyx lobes. Furthermore, the pollen exine pattern of G. maculata is different from that of any other Gilia species, especially that of G. campanulata or G. inyoensis (Fig. 4). Pubescence type and pollen exine morphology are regarded generally as conservative characters and have been used in numerous instances as taxonomically valuable characters throughout the Polemoniaceae (Grant 1959; Patterson 1977; Timbrook 1986; Gordon-Reedy in press). Therefore, although affinities exist between G. maculata and certain other species of sect. Giliastrum, strong differences remain, rendering the problem of re- lationships with the remainder of the genus far from solved. 24 MADRONO [Vol. 36 •2 I 5 S ^ 00 o ^ 1/3 C 73 < o > 13 ^ S u tin fQ T3 (U 3 Oh ^ < - > < - P(i "S en eeply per -o C ■<-> cd o o CO c ual >. liar < S Simi C/5 C/3 a (U > > ^ a a C (U o O — o ^ 13 <2 [1. CQ 3 l-i (U § o ■t-> cd ■*-> o u O (U 2 o (/) o cd " ^ C O o o 1989] PATTERSON: GILIA MACULATA 25 A case for and against a new genus. It is inevitable that, as more taxonomic information has become available in the Polemoniaceae, reassessments have appeared, often necessitating recognition of new taxa above the level of species. Day and Moran (1986) recently accumulated evidence in favor of reassigning the former Ipomopsis gloriosa to a new genus, Acanthogilia; the combination of characters in this taxon precluded unequivocal placement in any previously existing genus. Timbrook (1986) similarly reaffirmed the generic status of Loeseliastrum, formerly a section of Langloisia. A strong case might be made for a similar treatment of G. maculata, inasmuch as it does not ally very closely with any known member of Gilia, and certainly not with Linanthus or any other existing genus in the family. Morphologically it represents a mosaic of features from dif- ferent genera, lacking all of the defining characters of even the more variable genera in the family. However, a large number of unsolved questions remain about relationships within Gilia as well as among Gilia and other genera. Other genera in the Gilieae are reasonably well-circumscribed and distinct, even though they may share a suite of characters with Gilia. Based on information presently available, Gilia maculata does not possess any character or combination of characters that clearly set it apart at the generic level. Further studies of character distribution in this species and in the remaining species of Gilia may provide an alternative insight on this problem. Taxonomic importance of Gilia maculata. Questions of evolu- tionary and taxonomic importance remain in which G. maculata may provide some insight. Its previous placement within and with- out Linanthus reemphasizes the point that there is a great deal that is not understood clearly about the relationship between Linanthus and Gilia. For example, just as G. maculata has been moved from Linanthus in this study, Moran (1977) removed L. uncialis from Gilia. In neither case are morphological features problematic or difficult to measure; rather, both of these species are poorly known, being uncommon and inconspicuous in the field, and poorly rep- resented in herbaria. This underscores the need for considerable caution in assessing taxonomic relationships when some or all mem- bers of a group are not completely understood. It is also noteworthy that both of these species have been placed in Linanthus sect. Dian- thoides and Gilia sect. Giliastrum. This situation suggests that an evolutionary connection between Gilia and Linanthus might be sought among these two sections. Such an hypothesis is attractive in exploring further the relationships between these genera. The position of G. maculata in the California flora also remains un- known. Its isolation in the Little San Bernardino Mountains, a region not particularly well-associated with isolated and endemic plant species, is not readily explained; neither is its geographic disjunction 26 MADRONO [Vol. 36 by over 300 km from its postulated nearest relatives, G. inyoensis and G. campanulata. In addition, not only is there minimal infor- mation about population size in this species, but virtually nothing is known about its reproduction (e.g., pollination, seed production, dispersal). Prior to attempting to answer questions about evolution in this species, considerably more information must be gathered regarding the ecology, distribution, and reproductive biology of G. maculata. Acknowledgments Thanks are due to the following persons: Mona Bourell and Steve Timbrook for helping to find G. maculata in the field and for valuable discussions; Mary Ann Showers and Rich Simpson for preparation of figures; Ginger Vagenas for SEM assistance; Peggy Fiedler for reviewing an early draft of the manuscript; and especially to Alva Day for her thoughtful consult and help in interpreting relationships within Gilia. Part of this work was facilitated by a SFSU Professional Research and Devel- opment Award. Literature Cited Bentham, G. 1833. Polemoniaceae. Bot. Reg. 19:t. 1622. Brand, A. 1907. Polemoniaceae. A. Engler, Das Pflanzenreich 4 (250). Chuang, T., W. C. Hsieh, and D. H. Wilken. 1978. Contribution of pollen mor- phology to systematics in Collomia (Polemoniaceae). Amer. J. Bot. 65:450—458. Day, a. G. and R. Moran. 1 986. Acanthogilia, a new genus of Polemoniaceae from Baja California, Mexico. Proc. Calif Acad. Sci. 44:1 1 1-126. Gordon-Reedy, P. 1989. Trichome patterns and geographic variation in Lepto- dactylon californicum (Polemoniaceae). Madrofio (in press). Grant, V. 1959. Natural history of the Phlox family: systematic botany. M. NijhofT, The Hague. . 1965. Flower pollination in the Phlox family. Columbia Univ. Press, New York. Greene, E. L. 1892. Some American Polemoniaceae. Pittonia 2:251-260. Jepson, W. L. 1925. A manual of the flowering plants of California. A.S.U.C. Bookstore, Berkeley, CA. . 1 943. Polemoniaceae. In A flora of California 3: 1 3 1-222. A.S.U.C. Book- store, Berkeley, CA. Mason, H. L. 1951. Polemoniaceae. In L. R. Abrams, Illustrated flora of the Pacific states 3:396-474. Stanford Univ. Press, Stanford, CA. MiLLiKEN, J. 1904. A review of California Polemoniaceae. Univ. Calif. Publ. Bot. 2:1-71. Moran, R. 1977. New or renovated Polemoniaceae from Baja California, Mexico. Madrono 24:141-159. MuNZ, P. A. 1959. A California flora. Univ. Cahfomia Press, Berkeley, CA. . 1974. A flora of southern California. Univ. California Press, Berkeley, CA. Parish, S. B. 1892. New California plants. Bull. Torrey Bot. Club 19:93. Stuchlik, L. 1967a. Pollen morphology in the Polemoniaceae. Grana Palynol. 7: 146-240. . 1967b. Pollen morphology and taxonomy of the family Polemoniaceae. Rev. Palaeobot. Palynol. 4:325-333. Taylor, T. N. and D. A. Levin. 1975. Pollen morphology of Polemoniaceae in 1989] PATTERSON: GILIA M AC U LATA 27 relation to systematics and pollination systems: scanning electron microscopy. Grana 15:91-112. TiMBROOK, S. 1 986. Segregation of Loeseliastrum from Langloisia (Polemoniaceae). Madrono 33:157-174. (Received 9 Nov 1987; revision accepted 15 Oct 1988.) ANNOUNCEMENT Relocation of UC and JEFS to Interim Quarters During March and April 1989, the UC and JEPS collections will be temporarily relocated to a site several miles from the UC-Berkeley campus; the staff will move in May or June. The herbaria will be housed at the off-campus site until renovation of new quarters in the Life Sci- ences Building is completed at the end of 1992. Interim quarters will be fully functional and accessible to researchers. We expect to retain our phone numbers, and mail addressed to "University Herbarium" or "Jepson Herbarium" (but not "Department of Botany") will be deliv- ered to our new location. The move is being coordinated in such a way as to minimize disrup- tion of research needs. Loans will generally be unaffected, other than potential minor delays in processing. We do ask that shipments to UC of routine exchange, returned loans, and similar low-priority transac- tions be kept to a minimum until June 1989, so that our staff can concentrate on the move. Visitors during March and April 1989 should contact us in advance to determine whether or not their groups have been moved, and what needs to be done to bring the researcher and specimens together. Except for the day or two that any group of specimens is in transit, they should be accessible at one place or the other, but special arrangements will need to be made to provide access to the new quarters until June 1989. The location of the interim quarters is at 6701 San Pablo Avenue, two blocks south of Ashby Avenue at the junction of Berkeley, Oakland, and Emeryville. The herbaria will occupy a minor portion of a huge warehouse owned by the University, commonly referred to as the Mer- chant Building. The facility is easily accessed by automobile from In- terstate Highway 80 at the Ashby Exit. For public transportation, take a bus from the Ashby BART Station west along Ashby Avenue to San Pablo Avenue. Eastern Hemisphere collections of spermatophytes (except Astera- ceae, Apiaceae, Myrtaceae, and Ranunculaceae) will continue to be housed at the annex established five years ago adjacent to campus. After renovation is complete, however, these collections will be reintegrated into the main herbarium. Until then, visitors who expect to see these specimens should make arrangements in advance. MACROMERIA ALBA (BORAGINACEAE), A NEW SPECIES FROM TAMAULIPAS, MEXICO Guy L. Nesom Department of Botany, University of Texas, Austin, TX 78713 Abstract Macromeria alba is described from Gomez Farias area in the Sierra Guatemala of west-central Tamaulipas. It is most closely related to M. notata from the high moun- tains of Coahuila and Nuevo Leon to the north. Resumen Se describe Macromeria alba de la region de Gomez Farias en la Sierra Guatemala en la oeste-central de Tamaulipas. La especie nueva parece tener afinidades estrechas con M. notata, la cual se encuentra en las montafias altas del norte de Coahuila y Nuevo Leon. Continued curation of the Boraginaceae at TEX-LL has brought to light an undescribed species of Macromeria. It is the second new borage from the Sierra Guatemala of west-central Tamaulipas (see Nesom 1988)— both made by Dr. Alfred T. Richardson, presently at Texas Southmost College in Brownsville. This species is the first addition to the genus since Johnston's revision (1954), making a total of nine currently recognized species. The genus ranges from the southwestern United States to Guatemala with two species-rich areas, one in northeastern Mexico and the other in southwestern Mexico. Macromeria alba Nesom, sp. nov. (Fig. 1)— Type: MEXICO, Ta- maulipas, Mpio. Gomez Farias, area W of Rancho del Cielo in the sierra, ca. 5-7 km NW of Gomez Farias, between San Jose and La Perra [just S of Agua del Indio, area of pine-oak], 30 May 1969, A. Richardson 1263 (holotype, TEX). M. notatae simile sed foliis brevipetiolatis, lobis calycis brevior- ibus, et corollis albis lobis multo longioribus differt. Perennials to 2 m tall. Stems with ascending-appressed hairs 0.3- 1.5 mm long. Leaf blades lanceolate-elliptic to ovate-lanceolate, 2.5- 10 cm long, 8-30 mm wide, with primary veins slightly impressed above, raised beneath, lighter-colored beneath; apices acute to acu- minate, bases obtuse to rounded and abruptly forming a stipe-like petiolar base 1-2 mm long, margins entire, usually narrowly revo- lute, appressed-ciliate; the lower surfaces evenly strigose with closely Madrono, Vol. 36, No. 1, pp. 28-32, 1989 1989] NESOM: NEW MACROMERIA FROM MEXICO Fig. 1 . Habit of Macromeria alba (from Richardson 366). 30 MADRONO [Vol. 36 Fig. 2. Distribution of Macromeria alba and M. notata. appressed hairs mostly 0.5-0.9 mm long, upper surfaces usually with minute, pustulate trichome bases without emergent trichomes, or trichomes, when present, never more than 0.2 mm long. Flowers in axils of well-developed leaves on intemodes 5-10 mm long, mature fruits separated on intemodes up to 40 mm long; pedicels 5-8 mm long in flower, to 1 3 mm long in fruit; calyx lobes linear- triangular, 0.5-1 mm wide, 6-8 mm long in flower, to 11 mm long in fruit; corollas white, prominently spreading-hairy on the outside, 42-47 mm long, gradually ampliate from near the base, regular, 4-6 mm wide (pressed) at the throat, lobes erect or ascending, lanceolate to ovate-lanceolate or triangular, 6-9 mm long, inner surface of each with a band of stipitate-glandular hairs beginning in the throat and extending halfway to the apex, corollas otherwise glabrous inside; style persistent, as long as or slightly longer than the filaments, barely exserted, stigma subterminal, separated by sterile tip of the style; anthers glabrous, 2 mm long, medio-fixed, filaments as long as the 1989] NESOM: NEW MACROMERIA FROM MEXICO 31 corolla. Nutlets ovoid, smooth and shiny, white or brownish, 2 mm wide, 2.5-3 mm long. Known only from the region of Rancho del Cielo near Gomez Farias, Sierra de Guatemala, Tamaulipas, at ca. 4800-6300 ft in elevation (Fig. 2). Paratypes: MEXICO, Tamaulipas, Mpio. Gomez Farias, area of Rancho de Cielo, ca. 5-7 km NW of Gomez Farias: Agua Linda trail, 5 Jun 1969, Richardson 1367 (TEX); between Indian Springs and Agua Linda tumoff, 26 Jun 1968, Richardson 366 (TEX). Macromeria alba is clearly most similar to M. notata I. M. John- ston and keys to that species in Johnston's study (1954) of the genus. Both species have corollas with erect or ascending lobes and with prominent lines of stipitate glands (described in M. notata as "weakly invaginate elongate densely glanduliferous plaits") on each lobe ex- tending from inside the throat below each lobe to beyond the middle of it. As noted by Johnston, M. longiflora D. Don and M. pringlei Greenman also have glandular corollas with erect lobes, although the glands are not positioned similar to those of M. alba and M. notata. The epithet "alba" refers to the corolla color (as noted by the collector); in the other species of Macromeria, corollas usually range from yellow to light yellow or yellow-green. Differences be- tween the new species and its closest relative are presented in the following couplet. A. Stem pubescence of hairs variable in length, all ascending-appressed; leaves usu- ally abruptly narrowed at the very base to a stipe-like petiole 1-2 mm long; pedicels 5-8 mm long in flower; calyx lobes 7-8 mm long in flower; corollas white, fun- nelform, gradually opened to the throat, the lobes 6-9 mm long, lanceolate to ovate-lanceolate M. alba A. Stem pubescence a mixture of short, arching-appressed hairs and longer, straight, spreading ones; leaves basally attenuate and sessile, not at all petiolate; pedicels 2- 4 mm long in flower; calyx lobes 10-15 mm long in flower; corollas yellow with greenish lobes, slightly funnelform, abruptly flaring at the throat, the lobes 3- 5 mm long, widely to very widely ovate M. notata The pubescence of the upper leaf surface in Macromeria alba mostly consists of minute, pustulate bases without emergent tri- chomes, or when trichomes are present they are never more than 0.2 mm long. In M. notata the trichomes are 0.5-0.1 mm long. In M. notata, the flowers are tightly clustered at the branch tips, com- pared to the more distantly separated flowers and fruits of M. alba. In addition, label data indicate that plants of M. alba grow to a height of about 2 m, whereas the collections of M. notata are all of plants 0.2-0.5 m tall. Macromeria alba apparently is endemic to the area just west of Gomez Farias in the Sierra Guatemala of west-central Tamaulipas. All collections were made in an area of pine-oak woodland (Al Richardson pers. comm.). Macromeria notata is a species apparently restricted to high mountains of Coahuila (Sierra de Viga) and Nuevo 32 MADRONO [Vol. 36 Leon (Sierra Infemillo— the type, Sierra de la Marta, and Cerro Potosi). The closest known localities of the two species are about 200 kilometers apart. Acknowledgments I greatly appreciate the helpful comments of Jim Miller and other reviewers. Literature Cited Johnston, I. M. 1954. Studies in the Boraginaceae, XXVI. Further reevaluations of the genera of the Lithospermeae. J. Arnold Arbor. 35:1-81. Nesom, G. L. 1988. A synopsis of the American species of Omphalodes Toum. (Boraginaceae). Sida 13:25-30. (Received 5 Jul 1988; revision accepted 16 Nov 1988.) NOTEWORTHY COLLECTION California Prunus fasciculata (Torrey) A. Gray var. fasciculata (Rosaceae).— San Luis Obispo Co., E foothills of La Panza Mts. on hill above Dominez Rd between Del Rosa and Doris Trail in California Valley tract area, unit 32, T30S R19E, NW'A of sect. 31, 645 m, ca. 80 individuals among rocks, assoc. with Ericameria linearifolia and grasses, 30 Mar 1988, Douglas Chadwick s.n. (OBI). Previous knowledge. The desert almond is widespread in transmontane deserts of SE CA, and ranges E to NV, UT, and AZ. In CA it is occasional to locally abundant in desert portions of Transverse and Peninsular ranges, desert-facing slopes of the southern Sierra Nevada, and various of the transmontane desert ranges. Significance. First record for S Coast Ranges of cismontane CA; disjunct by ca. 165 km from nearest population (in S Sierra Nevada near Onyx, Kern Co.; Twis- selmann, Fl. Kern Co., Calif, 1967). Prunus fasciculata is represented in coastal areas of San Luis Obispo and Santa Barbara cos. by var. punctata Jepson, the sand almond, a taxon restricted to coastal dune formations (Hoover, Vase. PI. San Luis Obispo Co., Calif., 1970; Smith, Fl. Santa Barbara Region, Cahf., 1976). The population of P. fasciculata var. fasciculata is separated from the nearest San Luis Obispo Co. populations of var. punctata by ca. 78 km and by the principal ridges of the La Panza and Santa Lucia Mts. In its hot, dry climate the California Valley and adjacent regions of the Carizzo Plain resemble the Mojave Desert much more than they do the coastal dune areas.— Ann Chadwick and David J. Keil, Biological Sciences Department, California Polytechnic State University, San Luis Obispo, CA 93407. TAXONOMY OF STREPTANTHUS SECT. BIENNES, THE STREPTANTHUS MORRISONII COMPLEX (BRASSICACEAE) Rebecca W. Dolan Holcomb Research Institute and Biological Sciences Department, Butler University, 4600 Sunset Avenue, Indianapolis, IN 46208 Lawrence F. LaPre Tierra Madre Consultants, 4178 Chestnut Street, Riverside, CA 92501 Abstract The Streptanthus morrisonii complex is a six-taxon group of closely related ser- pentine rock outcrop endemics from Lake, Napa, and Sonoma counties of California, USA. Two new subspecies {S. morrisonii subsp. kruckebergii and S. brachiatus subsp. hoffmanii) from Lake County, California, are described. The relationship of these taxa to others in the section is reviewed and descriptions and a key are provided. Floristic surveys of serpentine rock outcrops conducted for en- vironmental impact reports for geothermal and gold-mining oper- ations in Lake, Napa, and Sonoma counties of California, have revealed new data on many rare and unusual plants. During these surveys previously undescribed populations of plants were discov- ered that are clearly members of the section Biennes of the genus Streptanthus (the Streptanthus morrisonii F. W. Hoffman complex), and yet do not match the morphology or geographical distribution of the described taxa (Hoffman 1952). Some Streptanthus taxa re- stricted to serpentine are known for extreme local differentiation (Kruckeberg 1956, 1958). We undertook a study of the biochemistry, morphology, and distribution of the section to evaluate the existing taxonomy and to determine the taxonomic status of newly-discov- ered populations. Hoffman (1952) first addressed the taxonomy of "biennial" (i.e., monocarpic perennial) Streptanthus. He collected and described two species, one with three subspecies. These plants grow on serpentine outcrops of limited access and had not previously been collected. Hoffman placed his taxa in the subgenus Euclisia Nutt. ex Torrey & A. Gray {Streptanthus with zygomorphic flowers, nonbracteate inflorescences, and one or two pairs of stamens with connate or partially connate filaments), that was monographed by Morrison in Madrono, Vol. 36, No. 1, pp. 33-40, 1989 34 MADRONO [Vol. 36 1941. The biennial Streptanthus were recognized as section Biennes by Kruckeberg and Morrison (1983). Members of the Streptanthus morrisonii complex have glabrous and glaucous vegetative parts. Their most distinctive feature is cab- bage-like rosette leaves that are broad, palmately-lobed, fleshy or succulent, and often mottled on the adaxial surface. Succulent rosette leaves indicate the biennial life history characteristic of the group. Some related Streptanthus that grow on serpentine also possess fleshy rosette or basal leaves. This tendency toward succulence appears to be one of the suite of traits shared by serpentine endemics (Krucke- berg 1984a, b). The newly discovered populations differ from the described taxa of the complex in morphological traits and/or geographic range. Plant habit, flower color, and leaf characteristics are the most sig- nificant discriminating traits. These features and genetic relationship as revealed by starch gel electrophesis of enzyme variants (unpubl. data) support the taxonomy of the section as developed by Hoffman (1952) with the addition of two new subspecies; one each for Strep- tanthus brachiatus and S. morrisonii. The relationship of these sub- species to other members of the complex is presented in the following taxonomic treatment. Type localities for the taxa are mapped in Figure 1. Additional collection sites are in the immediate vicinity of the type localities for these extremely restricted endemics. Streptanthus Nutt. sect. Biennes Kruckeb. & J. Morrison, Ma- droiio 30:242. 1983. Glacous and glabrous biennials, low (20 cm) to tall (75-125 cm), the first year rosettes of petiolate, broadly spatulate, and coarsely dentate leaves. Flowers in openly branched racemes or panicles, zygomorphic; calyces flask-shaped; sepals glabrous or setose, yellow to purple, carinate; petals white to salmon-colored, crisped, unequal, recurved; stamens in 3 unequal pairs (upper, lateral, and lower), the upper with connate filaments, strongly recurved upward, the lower set partially connate and recurved downwards. Siliques erect, di- varicate or reflexed, usually torulose; seeds only weakly winged at tip; cotyledons accumbent. Key to Taxa of Section Biennes of Streptanthus A. Plants short ( 1 0-30 cm) and much branched near the base. ... 1 . S. brachiatus B. Calyces glabrous, rose-purple; endemic to the immediate vicinity of Socrates Mine, Sonoma County la. S. brachiatus subsp. brachiatus B. Calyces usually pubescent, yellow or dark purple; endemic to the Sulphur Creek drainage on Lake/Sonoma county line lb. S. brachiatus subsp. hoffinanii A. Plants tall (30-100 cm) and remotely branched 2. S. morrisonii C. Calyces densely pubescent with long (2 mm) hairs, dark purple; endemic to the headwaters of East Austin Creek, Sonoma County 2a. S. morrisonii subsp. hirtijlorus 1989] DOLAN AND LaPRE: STREPTANTHUS SECT. BIENNES 35 • Type localities i — i — i 0 16 Fig. 1. Map of type localities of the Streptanthus morrisonii complex. C. Calyces glabrous to slightly pubescent, greenish yellow to golden yellow. D. Upper and lower surfaces of rosette and lower cauline leaves usually green; upper connate filaments uniformly orange or orange-yellow; endemic to drainage of Big and East Austin creeks, Sonoma County 2b. S. morrisonii subsp. morrisonii D. Upper surfaces of rosette and lower cauline leaves heavily mottled with purple-brown, lower surface uniformly purplish; upper connate filaments uniformly yellow. 36 MADRONO [Vol. 36 E. Upper cauline leaves 2-4 times as long as wide 2c. 5. morrisonii subsp. elatus E. Upper cauline leaves 1-2 times as long as wide 2d. S. morrisonii subsp. kruckebergii 1 . Streptanthus brachiatus F. W. HofFm., Madrono 1 1 :230. 1952. Rosette leaves gray-green, mottled with purple-brown above, uni- formly purple beneath. The original stem extending the second year and producing more or less brachiate branches bearing short-peti- olate and sessile, undulate, auriculate, orbicular to orbiculate and oblong-spatulate, prominently veined leaves with entire or coarsely serrate margins or with the margins entire basally and serrate api- cally, passing into narrowly-lanceolate usually toothed bracts. Flow- ers in discrete racemes, bracteate or not. Calyces glabrous or sparsely pubescent and reticulate with fine lines, rose-purple with yellowish base or purple or yellow. Sepals broadly lanceolate. Upper connate filaments orange-colored with two longitudinal, purple lines. Chro- mosome number unknown. la. Streptanthus brachiatus F. W. HofFm. subsp. brachiatus.— Type: USA, California, Sonoma-Lake county line, E of Pine Flat, exposed serpentine ridge near Contact Mine, 3000 ft, 5 Jun 1949, Kruckeberg and Hoffman 1905 (holotypes, UC!). Flowers glabrous; calyces rose-purple with yellowish bases; mature upper connate filaments orange with two longitudinal purple lines. Streptanthus brachiatus subsp. brachiatus is known only from the immediate vicinity of Socrates Mine on the Sonoma-Lake county line (Fig. 1). According to H. K. Sharsmith (specimen annotation 8, Oct. 1952), the type locality is near the junction of Socrates Mine Rd, with Pine-Flat-Middletown Rd on ridge W of canyon of Big Sulphur Creek, Sonoma County. The Napa-Lake county line is on ridge east of canyon of Big Sulfur Creek. Additional collections. USA, CA, Lake Co., near Contact Mine, E of Pine Flat, on the Sonoma-Lake county line, 3000 ft, Kruckeberg and Hoffman 1905 (UC); same area, Hoffman 3436 (UC); summit of ridge about 0.5 mi S of Mercury ville on rd to Big Geysers, Hoff- man 3379 (UC); Sonoma Co., near junction of Socrates Mine Rd with Pine-Flat-Middletown Rd, Mayacamas Mts., 3200 ft, Shar- smith 4129 (UC). lb. Streptanthus brachiatus F. W. Hoffman subsp. hoffmanii Dolan & LaPre, subsp. nov.— Type: USA, California, Lake County near Sonoma county line on Bear Ridge Rd V4 mi S of three- way junction with Ridge Rd and Davies Rd, on serpentine outcrop near a geothermal expansion joint, 2 May 1985, LaPre s.n. (holotype, UC; isotypes, CAS, RSA, UC). 1989] DOLAN AND LaPRE: STREPTANTHUS SECT. BIENNES 37 Herbae 10-30 cm altae probe basin ramossimae; calyces plerum- que pubescentes, lutei vel atropurpurei. Flowers glabrous to pubescent and variable within a population. Calyces purplish green to greenish yellow. Mature upper connate filaments yellowish with two dark-colored longitudinal lines. This taxon occurs on isolated serpentine rock outcrops, occasion- ally scattered in adjacent chaparral, near the Lake-Sonoma county line (TION R7W and R8W) primarily in geothermal development areas, from the junction of Ridge Road, Da vies Road, and Bear Ridge Road off Socrates Mine Road, S to Buck Rock and SE to Mount St. Helena (Fig. 1). Populations are morphologically uniform within single outcrops but much local differentiation is present be- tween outcrops, even those in close proximity. Calyx color varies most prominently (from purple to yellow) along with stature (from short to tall) along the line from the northwest to southeast. Pop- ulations in the southeast nearest the location of S. morrisonii subsp. elatus tend to converge on morphological characteristics of that taxon. Species growing on the serpentine outcrops with Streptanthus brachiatus subsp. hoffmanii include the rare plants Eriogonum ner- vulosum and Allium falcifolium. Growing on the margins of the outcrops in the more weathered serpentine are Pinus sabiniana, Arctostaphylos viscida, Cupressus sargentii, Quercus durata, Sola- rium parishii, Fremontodendron californicum subsp. napense, and Ceonothus jepsonii. This taxon is named in honor of Freed Hoffman, an amateur botanist, of Guerneville, CA, who specialized in serpentine flora. He was the first to collect Streptanthus in "The Geysers" region. 2. Streptanthus morrisonii F. W. Hoffm., Madrono 1 1:225. 1952. Rosette leaves uniformly gray-green above and beneath, or the lower surface somewhat purple-tinged, or the upper surface heavily mottled with purplish or brownish blotches and the lower surface purple. The stem, in the second year, extended and producing au- riculate-spatulate to auriculate-ovate, sessile, clasping, entire or few- toothed leaves, these passing into auriculate-lanceolate acute, sessile leaves and awl-shaped bracts. Flowers scattered along the flowering stems or concentrated towards the tips of branches. Calyces densely pubecent, with a few scattered hairs, or entirely glabrous, yellow to purple. Sepals ovate-lanceolate. Upper connate filaments uniformly yellow or orange with two longitudinal purple lines when the calyx is purple. Chromosome number unknown. 2a. Streptanthus morrisonii F. W. Hoffm. subsp. hirtiflorus F. W. Hoflm., Madrofio 11:228. 1952.-Type: USA, California, Sonoma County, on bluffs and cliff talus, serpentine soil, above 38 MADRONO [Vol. 36 Dorr's Cabin, headwaters of East Austin Creek, 17 Jun 1948, Hoffman 2344 (holotype, UC!). Flowering stems strict or much branched and diffuse, up to 80 cm tall. Juvenile leaves heavily mottled with purple-brown above, uni- formly purple beneath; upper stem leaves auriculate-spatulate to auriculate-ovate, sessile, clasping, entire or few toothed. Flowers abundant, scattered along the flowering branches. Calyces red-pur- ple, abundantly clothed in long hairs (2 mm long) which gives the plant a grayish appearance. Petals dull white with purplish veins. Upper connate filaments orange, with two longitudinal, purple lines. Streptanthus morrisonii subsp. hirtiflorus grows on serpentine bluffs and talus slopes with western exposure. This rare serpentine endemic occupies an area of not over 1 00 m^ on west-facing serpentine bluffs and slopes at the headwaters of East Austin Creek, a short distance above Dorr's Cabin, Sonoma Co., California (Fig. 1). It has not been collected elsewhere. 2b. Streptanthus morrisonii F. W. Hoffm. subsp. morrisonii.— Type: USA, California, Sonoma Co., serpentine outcrop, head of Big Austin Creek at Layton Mine, 26 Sep 1946, Hoffman 1020 (holotype, UC!). Flowering stems strict, 20-60 cm tall. Juvenile and adult leaves gray-green on both surfaces, or slightly purplish beneath, without maculation. Upper stem leaves similar to those of subsp. hirtiflorus. Flowers discretely produced toward the tips of the ascending or divergent branches. Calyces greenish yellow becoming golden yellow with age, glabrous or with a few scattered hairs. Petals creamy white to light salmon with brownish or orange-colored veins. Upper con- nate filaments uniformly orange. This taxon occurs on serpentine outcrops in "The Cedars" area of northern Sonoma County, along the drainage of Big Austin Creek and its tributaries (Fig. 1). Additional collections. USA, CA, Sonoma Co., headwaters of Big Austin Cr. at Layton Chromite Mine, Hoffman 1020 (UC); Layton Mine, Austin Cr., Hoffman 1027 (UC); near headwaters of Devil Cr., The Island: tributary of upper East Austin Cr., Hoffman 2995 (UC); trail from Gray Cr. to The Island, headwaters of East Austin Cr., Hoffman 3360 (UC); The Cedars, headwaters of East Austin Cr., 700-2000 ft, Raiche 30581 (JEPS). 2c. Streptanthus morrisonii F. W. Hoffm. subsp. elatus F. W. Hoffm., Madrofio 1 1:228. 1952. -Type: USA, California, Napa- Lake county line, Va mi W of White's Point, Table Mountain Rd, ca. 5 mi E of Mountain Mill House, 3 May 1 947, Kruckeberg 1438 (UC!). 1989] DOLAN AND LaPRE: STREPTANTHUS SECT. BIENNES 39 Flowering stems strict, remotely branched, 35-105 cm tall. Ju- venile leaves with upper surface mottled with purplish brown and lower surface uniformly purple, blades long-petioled, obovate or flabelliform, prominently veined, with margins entire basally and coarsely dentate distally. Upper stem leaves oblong-spatulate, cym- biform, clasping. Flowers produced toward the tips of ascending branches. Calyces greenish, turning golden yellow with age, glabrous or sparsely pubescent. Petals white, turning yellowish with age. Up- per connate filaments uniformly greenish yellow. Known only from several closely spaced serpentine outcrops near Three Peaks and White's Point on the Lake-Napa county line (Fig. 1). Additional collections. USA, CA, southern Lake Co., along ridge from White Pt, near Napa-Lake county line, 2.7 mi E of Mt. Mill House, 2500 ft, Hoffman 2906 (UC); Hoffman 2872 (UC); rosettes grown from seed collected at White Pt, Hoffman s.n. (UC); Hoffman Cr., about 1 mi E of Mirabel Park, Raven 010745 (UC). 2d. Streptanthus morrisonii F. W. Hoffm. subsp. kruckebergii Do- lan & LaPre, subsp. no v.— Type: USA, California, Lake Co., Dunnigan Hill in Knoxville Recreation Area (TUN R5W, sect. 11), on serpentine outcrop, 8 Jun 1985, LaPre s.n. (holotype, UC; isotypes, CAS, RSA, UC). Herbae 30-100 cm altae, remote ramosae, folia rosularia maculata purpureobrunneis in superioribus paginis, uniformiter purpurea in paginis inferis; folia caulina superiora l-2plo longiora quam latiora; calyces glabrae vel leviter pubescentes, viridiflavae; superiora fila- menta connata uniformiter lutea. Flowering stems remotely branched, 20-1 1 5 cm tall. Juvenile leaves green with punctations above, uniformly purple beneath. Upper stem leaves oblong, spatulate, cymbiform, clasping, often deciduous before flowering. Flowers produced toward the tips of ascending branches. Calyces yellowish green, turning bright yellow with age. Petals creamy white. Upper connate filaments uniformly greenish yellow. This new subspecies is a morphologically uniform taxon. The plant occurs on scattered serpentine outcrops near the Lake-Napa county line, primarily in the Knoxville Recreation Area (Tl IN R4W), Dun- nigan Hill region, and associated watersheds (Fig. 1). Species associated with Streptanthus morrisonii subsp. krucke- bergii include Eriogonum nervulosum. Allium falcifolium, Streptan- thus breweri, and S. hesperidis. Pinus sabiniana, Arctostaphylos vis- cida, Cupressus sargentii, Quercus durata, and Ceonothus jepsonii grow in the adjacent chaparral. This taxon is named in honor of Dr. Arthur R. Kruckeberg, leading expert on the serpentine flora of the western United States. 40 MADRONO [Vol. 36 Acknowledgments This work was supported by Contract No. YA551-CT4-340080 between the U.S. Department of the Interior, Bureau of Land Management and Tierra Madre Con- suhants of Riverside, CA. Developers and utilities operating at the BLM's Known Geothermal Resource Area contributed funds to the project. Arthur Kruckeberg, Jim Bartel, and John Willowby provided helpful comments on the manuscript. Latin diagnoses were provided by Bert Steiner, Butler University. Holcomb Research In- stitute, Butler University, provided technical support. Literature Cited Hoffman, F. W. 1952. Studies in Streptanthus. A new Streptanthus complex in California. Madroiio 11:221-223. Kruckeberg, A. R. 1956. Variability in fertility of hybrids between isolated pop- ulations of the serpentine species, Streptanthus glandulosus Hook. Evolution 1 1 : 185-211. . 1958. The taxonomy of the species complex, Streptanthus glandulosus Hook. Madrofio 14:217-227. . 1984a. California's serpentines: flora, vegetation, geology, soils, and man- agement problems. Univ. Calif. Publ. Bot. 78:1-180. . 1984b. California's serpentine. Fremontia 1 1: 1 1-17. and J. L. Morrison. 1983. New Streptanthus taxa (Cruciferae) from Cali- fornia. Madrono 30:230-244. Morrison, J. L. 1941. A monograph of the section Euclisia Nutt., of Streptanthus. Ph.D. dissertation. Univ. of California, Berkeley. (Received 29 Mar 1988; revision accepted 22 Nov 1988.) ANNOUNCEMENT Fifth Annual Southwestern Botanical Systematics Symposium "Endemism" For information write to: Rancho Santa Ana Botanic Garden, Bo- tanical Systematics and Evolution Symposium, 1500 N. College Ave., Claremont, CA 9171 1; phone (714) 625-8767. Date: 19-20 May 1989. CHROMOSOME NUMBERS OF NORTH AMERICAN LA THYR US (FABACEAE) S. L. Broich Department of Crop Science, Oregon State University, Corvallis, OR 97331 Abstract Chromosone counts are reported for 1 8 populations of eight perennial Lathyrus species endemic to North America. Included are first counts of In^lA for L. holo- chlorus (Piper) C. Hitchc, L. delnorticus C. Hitchc, L. glandulosus Broich, L.jepsonii E. Greene subsp. jepsonii and L. vestitus Nutt. in Torrey & A. Gray subsp. vestitus, and 2«=28 for L. nevadensis S. Watson subsp. nevadensis. Chromosome counts of 2n=\4 for L. jepsonii subsp. californicus (S. Watson) C. Hitchc, L. vestitus subsp. bolanderi (S. Watson) C. Hitchc, L. polyphyllus Nutt. in Torrey & A. Gray and L. sulphureus Brewer ex A. Gray agree with those reported previously. Karyotypes of diploid species are symmetrical and similar to one another. Among the species studied here there does not appear to be the reduction in chromosome size and DNA amount reported in the literature for annual, autogamous Mediterranean Lathyrus species. Lathyrus L. is a genus of approximately 150 species of herbaceous perennial and annual papilionoid legumes (Fabaceae: Faboideae: Vicieae). The genus is distributed primarily in temperate Europe, Asia, North America, and South America, and in North Africa (Senn 1938; Kupicha 1981, 1983). There are about 26 species of Lathyrus endemic to North America (Hitchcock 1952; Welsh 1965; Bameby and Reveal 1971; Broich 1983, 1986, 1987; Nelson and Nelson 1983; Welsh et al. 1987); chromosome numbers have been reported for 15 of these species (Senn 1938; Hitchcock 1952; Ledingham 1957;Brunsberg 1965; Raven etal. 1965; Taylor and Mulligan 1968; Love and Love 1982; Ward 1983). The purpose of this paper is to report new observations of chro- mosome number and morphology of species of Lathyrus endemic to North America and to place these observations within the context of the genus world-wide. Materials and Methods Seeds of native Lathyrus were collected in July of 1979, 1980, and 198 1 . In addition to the author's collections, seeds of L. jepsonii E. Greene subsp. jepsonii were obtained from W. Roderick (Tilden Park Botanical Garden, Berkeley, CA), of L. vestitus Nutt. in Torrey & A. Gray and L. laetiflorus E. Greene (=L. vestitus subsp. vestitus sensu Broich, 1987) were obtained from Mary Allcott (Santa Barbara Botanic Garden, Santa Barbara, CA) and of L. vestitus from Mon- Madrono, Vol. 36, No. 1, pp. 41-48, 1989 42 MADRONO [Vol. 36 terey County, CA, were obtained from Dr. J. R. Griffin (Hastings Natural History Reservation, Carmel Valley, CA). Seeds were scarified with a razor blade and stored in rolls of damp germination paper (Dillard Paper Co., Doraville, GA) in a refrig- erator at ca. 5°C for 2 months. Five to six rolls were then placed vertically in a glass jar containing 100 ml of tap water, covered with a clear plastic bag and placed in a growth chamber on a cycle of 1 8 hours light at 22°C and 6 hours darkness at 1 8°C. After germination, seedlings were transplanted to the greenhouse into a soil mixture of equal parts of sand, peat, and soil. The number and morphology of mitotic chromosomes were stud- ied by examining root tip squashes. Root tips were pretreated with distilled water saturated with para-dichlorol-benzene at 10-1 5°C for 4 hours, fixed in 95% ethanol : glacial acetic acid (3:1; v:v), hydro- lyzed in 1 N HCl for 20 minutes at 60°C, stained in Feulgen (Dar- lington and La Cour 1975) and stored in 70% ethanol in a refrigerator (ca. 5°C). Stained root tips were squashed in 45% acetic acid and examined and photographed on a Zeiss phase-contrast microscope; slides were not made permanent. Voucher specimens, deposited at Oregon State University Her- barium (OSC), were made from two sources: specimens of plants taken from populations where seeds were later collected (field vouch- ers), and specimens of the plants from which root tips were taken - (greenhouse vouchers). The species of Lathy rus studied here did not flower under greenhouse conditions, therefore the greenhouse vouch- er specimens are of vegetative stems only. Results Table 1 presents a summary of new chromosome counts for Pacific Coast Lathyrus. First counts of 2n=\A were determined for L. glan- dulosus, L. holochlorus, L. delnorticus, L. jepsonii subsp. jepsonii and L. vestitus subsp. vestitus, and a count of 2«=28 for L. nevadensis subsp. nevadensis. Additional counts of 2« = 1 4 for L. jepsonii subsp. californicus, L. polyphyllus, L. sulphureus, L. vestitus subsp. bolan- deri agree with the reports of Hitchcock (1952). Karyotypes of all species examined are symmetrical and fall into classes lA and IB described by Stebbins (1971). Chromosome com- plements of these species are similar to one another; there is less than 25% difference in total haploid chromosome length among all diploid species examined. Chromosomes within a species are also similar to one another; the ratio of longest to shortest chromosome within a given species ranged from 1.4 to 1.7. The genome of each diploid species consists of 3-4 metacentric chromosomes decreasing in length from 7.1 to 5.5 micrometers and 4-3 submetacentric chro- mosomes also decreasing in length from ca. 7.0 to 5.0 micrometers. 1989] BROICH: LATHYRUS CHROMOSOME NUMBERS 43 Table 1. New Chromosome Counts of Pacific Coast Species of Lathyrus. An asterisk indicates first count(s) for that taxon. *L. delnorticus C. Hitchc. 2«= 14. CA, Del Norte Co., Panther Rat Campground, Six Rivers National Forest, T16N R3E sect. 22, Broich 642 (OSC); along French Hills Rd, 0.5 km S of jctn with Hwy 199, T17N RIE sects. 24-25, Broich 654 (OSC). *L. holochlorus (Piper) C. Hitchc. 2«=14. OR, Benton Co., along Oak Creek Rd ca. 0.4 km S of entrance to McDonald State Forest, TllS R5W sect. 19, Broich 1298 (OSC); Linn Co., along Hwy 99E opposite Linn-Benton Community Col- lege, TllS R4W sect. 36, Broich 630 (OSC). *L. glandulosus Broich. 2«=14. CA, Humboldt Co., 0.6 km E of the Freshwater- Kneeland Rd on rd to Maple Cr., Broich 772 (OSC); ca. 6.4 km S of the Kneeland School on rd to Bridgeville, Broich 777 (OSC). *L. jepsonii E. Greene suhsp. jepsonii. 2«=14. CA, Contra Costa Co., Brown's Island near Pittsburg. Plants grown in greenhouse from seed provided by W. Roderick, Tilden Park Bot. Card., Berkeley, CA, Broich 1278 (OSC). L. jepsonii E. Greene subsp. californicus (S. Watson) C. Hitchc. 2n=l4. CA, Trinity Co., 1.3 km E of Dinsmore's on Hwy 36, T30N R5E sect. 3, Broich 1166 (OSC). *L. nevadensis subsp. nevadensis. 2/?=28. OR, Benton Co., ca. 0.2 km S of entrance to McDonald State Forest, T121S R5W sect. 19, Broich 608 (OSC). L. polyphyllus Nutt. in Torrey & A. Gray. 2«=14. CA, Siskiyou Co., 3.9 km N of Happy Camp on rd to Takilma, Oregon, Broich 1182 (OSC). OR, Linn Co., along Peoria Rd, T12S R4W sect. 8, Broich 615 (OSC). Benton Co., McDonald State Forest, ca. 300 m N of the Oak Creek Entrance, Tl IS R5W sect. 19, Broich 1103 (OSC); along Peterson Rd, T12S R6W sect. 35, Broich 603 (OSC). L. sulphureus Brewer ex A. Gray. 2«=14. OR, Josephine Co., 0.8 km S of Waldo on FS rd 40S03, T40S R8W sect. 28, Broich 1131 (OSC). *L. vestitus Nutt. in Torrey & A. Gray subsp. vestitus. 2«=14. CA, Monterey Co., S slope of Junipero Serra Peak, Los Padres National Forest, T21S R5E sect. 4, plants grown in greenhouse from seed provided by J. R. Griffin, Hastings Natural History Reservation, Carmel Valley, Broich 1277 (OSC). Santa Barbara Co., plants grown in greenhouse from seed provided by Mary Allcott, Santa Barbara Botanic Garden, Broich 1267 (OSC). Ventura Co., ca. 64 km S of Ventucopa on Hwy 33, Los Padres National Forest, Broich 808 (OSC). L. vestitus subsp. bolanderi (S. Watson) C. Hitchc. 2«=14. CA, Del Norte Co., Panther Flat Campground, Six Rivers National Forest, T 1 7N R3E sect. 22, Broich 643 (OSC). Chromosomes in the tetraploid L. nevadensis were also metacentric to submetacentric and of approximately the same length as those of diploid species. On average, 5-10 good metaphase spreads were observed per root tip prepared, but in most cases only 1-2 photographs per plant were taken for measurement. Differences in degree of contraction were observed on slides and also among the photographs taken. Given the small sample size for each species and the karyotype similarity among species studied, interspecific differences could not be detected over the possible sources of error involved in karyotype measure- ments (Bentzer et al. 1971). 44 MADRONO [Vol. 36 o a O > U (J •l-H U o **** N •r^ IT) On 0^ O O o o (N o o CJ m oo On ON o o o Si o m OS o o o Si o (/3 cd I I (N ■*->■*-< oo C3 CO cd ir> ON o o o X o o o 3 3 ^ 'sT I C ~ 00 (N Tt Tt 00 JH OO £2 - — (N g (N g o o ;3 * 00 oo 3 I I I I I 1^ CO _5J "cj c o CO o X CJ ■(-• X ?5 (J (3 > > C X o CJ X C/5 <-> X ^ CJ "o !/5 J3 > X . <3 C3 pq c:> Cj >3 c u o CJ X CJ ■t-' X 0 o CO ■(-' C3 CO CO -O X cu 3 3 1/5 (yi 00 Cij J3 cd X CO T3 cd (-( m c< ^ w ^ CO N S s; -is X CO X 13 » (U O V. 1^ a a CO CO X X C/5 CO > > W 4J c o 03 X ^ ^5 c^ d C/3 X C/2 . Co d d c/5 in X X oo 1/5 1989] BROICH: LATHYRUS CHROMOSOME NUMBERS 45 OS 00 OS o u O '> CI. > :0 cd (U > :0 hJ IT) 6 J=l c hJ do" OS d c C/3 fN rs| (N (N Csl (N iTi m in in On ^ ^ ^ On OS o O o o I p:^ p:^ ^ pe: > :0 in OS '-3 OS c c C/2 I 1;!^ J3 « cd cd >n (N OS OS o C/3 cd in OS CJ m ffi u K (2 2 ^ Q u D g H O u u m < o o 3 3 C rt Tt Tj- o C3 I I 00 » I I * I/) a GO o . 1/5 o ^ X 1/5 Co Cd 1/5 !/) XI X) ;/5 (/5 2 O < o H "IIh lis 2 O X o 00 (U 00 > cd >> (/5 cd 10 pa ^ 5 X X! 3 to u o <■ <^ >. fc o H x: o 'O XI t-H O PQ -3 ^ u a> o (U ■t-> X 3 lis >■ -i: Q ci. CX ci 1/5 0/5 (/5 XXX 3 3 3 (/5 C/5 C/5 u X o • ^ X u o s.-g ^ s • pa o a a C/5 C/5 X X 3 3 5/5 46 MADRONO [Vol. 36 Discussion Lathyrus is widespread in temperate regions of both the Old and New World. Bassler (1973) and Raven and Axelrod (1978) have suggested that the genus originated in the Arcto-Tertiary geoflora of the Eocene. Lathyrus now consists of approximately 75% perennials and 25% annuals organized into 13 sections (Kupicha 1983). Six sections consist exclusively of perennials, six sections of annuals and one section includes both perennials and annuals. All species en- demic to North America are perennials and included in the section Orobus (L.) Godron in Gren. & Godron, which contains about one- third of all Lathyrus species. Lathyrus L. is predominantly diploid at 2n=2x=14. Kupicha (1977), in a summary table of counts for 56 species, reports five species which deviate from this number; Fedorov (1969) listed five polyploid species and two aneuploids (one 2n=\2\ one 2n=\6) of 61 species reported. The Fedorov list, however, does not include counts reported by Hitchcock (1952). When information from Hitchcock (1952) and more recent compilations (Moore 1 973; Gold- blatt 1981, 1984, 1985) are taken into account, a total of seven polyploid taxa have been reported in Lathyrus. All polyploid taxa are perennial and belong to the section Orobus except L. patensis L. (2^=14, 28, 42) which has been placed in sect. Pratensis Bassler (Kupicha 1983). With the new determinations reported here, a sample of chro- mosome numbers is now known for 18 of the 26 Lathyrus species endemic to North America (Table 2). North America appears to be a center for polyploidy in Lathyrus: four of the seven known poly- ploid species (L. venosus Muhl., 2/i=28; L nevadensis S. Watson, 2^=28; L. littoralis (Nutt. ex Torrey & A. Gray) Endl., 2/7=28; L. lanzwertii Kellogg, 2n=\4, 28) are endemic to the continent; two of the remaining three {L. japonicus Willd., 2n=l4, 28; L. palustris L., 2/7=42) have circumboreal distributions and are native to North America. The complete extent and significance of polyploidy in North American Lathyrus have yet to be studied in detail. Variation in the amount of genome DNA among Lathyrus species has also been studied (Rees and Hazarika 1969; Narayan 1982). Annual, autogamous species, which have evolved in the Mediter- ranean region, exhibit a threefold decrease in chromosome size cor- related to a fourfold decrease in the amount of nuclear DNA per diploid nucleus. In contrast, all western North American species of Lathyrus are perennial. Of those occurring along the Pacific Coast, L. vestitus subsp. bolanderi, L. holochlorus, and L. polyphyllus have been found to be self-incompatible (Broich 1983). L. vestitus is re- ported as having the greatest amount of nuclear DNA of the 21 species studied by Narayan (1982), and if chromosome size can be 1989] BROICH: LATHYRUS CHROMOSOME NUMBERS 47 taken to indicate, approximately, nuclear DNA amounts within a genome, the other species studied here have similar high amounts of DNA in comparison to the annual species of the Mediterranean Region. New chromosome observations reported here, therefore, corroborate the correlation between reduced DNA amounts and the evolution of an annual habit reported for Lathyrus (Rees and Ha- zarika 1969) and for higher plants in general (Price 1976). Acknowledgments I wish to thank Kenton Chambers for his support of this project and Mary Alcott, W. Roderick, and especially James R. Griffin for graciously responding to requests for seed of California Lathyrus species. Funds for this project were provided by the Oregon State University Herbarium and by National Science Foundation grant DEB- 7911543. Literature Cited Barneby, R. C. and J. L. Reveal. 1971. A new species of Lathyrus (Fabaceae) from the Death Valley region of California and Nevada. Aliso 7:361-364. Bassler, M. 1973. Revision der eurasiatischen Arten von Lathyrus L. sect. Orobus (L.) Gren. & Godron. Feddes Repert. 84:329-347. Bentzer, B., R. V. BoTHMAR, L. Engstrand, M. Gustafsson, and S. Snogerup. 1971. Some sources of error in the determination of arm ratios of chromosomes. Bot. Not. 124:65-74. Broich, S. L. 1983. A systematic study of Lathyrus vestitus Nutt. ex T. & G. (Fabaceae) and allied species of the Pacific Coast. Ph.D. thesis. Oregon State University, Corvallis, OR. . 1986. A new species of Lathyrus (Fabaceae) from northwestern California. Madroiio 33:136-143. . 1 987. Revision of the Lathyrus vestitus-laetiflorus complex (Fabaceae). Syst. Bot. 12:139-153. Brunsberg, K. 1965. The usefulness of thin-layer chromatographic analysis of phenolic compounds in European Lathyrus L. Bot. Not. 1 18:377-402. Darlington, C. D. and L. F. LaCour. 1975. The handhng of chromosomes, 6th ed. John Wiley and Sons, New York. Fedorov, a. a. 1 969. Chromosome numbers of flowering plants. Izdatel'stvo Nauk, Leningrad, U.S.S.R. Goldblatt, p. 1981. Index to plant chromosome numbers, 1975-1978. Missouri Botanical Garden, St. Louis, MO. . 1984. Index to plant chromosome numbers, 1979-1981. Missouri Botanical Garden, St. Louis, MO. . 1985. Index to plant chromosome numbers, 1982-1983. Missouri Botanical Garden, St. Louis, MO. Hitchcock, C. L. 1952. A revision of the North American species Lathyrus. Univ. Wash. Publ. Biol. 15:1-104. KupiCHA, F. K. 1981. Tribe 2 1 . Vicieae (Adans.) DC. Pp. 377-38 1 in R. M. Polhill and P. H. Raven (eds.). Advances in legume systematics. Royal Botanic Gardens, Kew, U.K. . 1983. The infrageneric structure of Lathyrus. Notes Roy. Bot. Gard. Edin- burgh 41(2):209-244. Ledingham, G. F. 1957. Chromosome numbers of some Saskatchewan Legumi- nosae with particular reference to Astragalus and Oxytropis. Can. J. Bot. 35:657- 666. 48 MADRONO [Vol. 36 Love, A. and D. Love. 1 982. lOPB chromosome number report LXXV. Taxon 3 1 : 344-360. Moore, R. J. 1973. Index to Plant chromosome numbers, 1967-1971. Regnum Vegetabile No. 90. . 1972. Index to Plant chromosome numbers, 1972. Regnum Vegetabile No. 91. Narayan, R. K. J. 1982. Discontinuous DNA variation in the evolution of plant species: the genus Lathyrus. Evolution 36:877-891. Nelson, T. W. and J. P. Nelson. 1983. Two new species of Leguminosae from serpentine of Humboldt County, California. Brittonia 35:180-183. Price, H. J. 1976. Evolution of DNA content in higher plants. Dot. Rev. 42: 27-52. Raven, P. H. and D. I. Axelrod. 1978. Origin and relationships of the California flora. Univ. Calif Publ. Dot. 72:1-134. , D. W. Kyhos, and A. J. Hill. 1965. Chromosome numbers of spermato- phytes, mostly Califomian. Aliso 6:105-1 13. Rees, H. and M. H. Hazarika. 1969. Chromosome evolution in Lathyrus. Chro- mosomes Today 2:158-165. Senn, H. a. 1938. Experimental data for a revision of the genus Lathyrus. Amer. J. Bot. 25:67-78. Stebbins, G. L. 1971. Chromosome evolution in higher plants. Addison-Wesley, Reading, MA. Taylor, R. L. and G. A. Mulligan. 1 968. The flora of the Queen Charlotte Islands II: cytological aspects of the vascular plants. Queen's Printer, Ottawa, Canada. Ward, D. E. 1983. Chromosome counts from New Mexico and southern Colorado. Phytologia 54:302-309. Welsh, S. L. 1965. Legumes of Utah. III. Lathyrus L. Proc. Utah Acad. Sci. 42: 214-221. , N. D. Atwood, L. C. Higgins, and S. Goodrich. 1 987. A Utah flora. Great Basin Naturalist Mem. 9, Brigham Young Univ., Provo, UT. (Received 13 Jul 1988; revision accepted 7 Dec 1988.) NOTES Gentiana setigera is the Correct Name for G. bisetaea (Gentianaceae).— Gentiana bisetaea Howell (Fl. N.W. Amer., 445, 1901) is a localized but fairly well understood species found in Darlingtonia bogs and seeps on serpentine hillsides in southwestern Oregon. The type locality given by Howell is ". . . eastern base of the Coast Mountains near Waldo, Oregon," the town of Waldo being a once-thriving gold-mining community in the southern part of the upper Illinois River Valley in Josephine County. Today the species is known from numerous bogs between Eight Dollar Mountain, 19 km N of Waldo, S to Gasquet Mountain, Del Norte Co., CA, and at scattered sites westward in the rugged Siskiyou Mountains to Curry Co., OR (files of Oregon Natural Heritage Data Base, Portland, and Siskiyou National Forest, Grants Pass). In 1941 M. E. Peck (Man. Higher PI. Oregon, 1st ed., 558) synonymized G. bisetaea with G. setigera A. Gray, a California taxon, but in the second edition of his book (607, 1961) and in L. R. Abrams' "Illustrated Flora of the Pacific States" (3:358, 1 95 1) G. bisetaea is treated as a distinct species. Because of its restricted range and specialized habitat, this gentian has been considered for possible listing as an endangered or threatened species (R. J. Meinke, "Threatened and Endangered Plants of Oregon: An Illustrated Guide," U.S. Fish & WildHfe Service, 160, 1982). The name Gentiana setigera A. Gray, in the usage of California botanists, has for over 60 years been applied to quite a different species from G. bisetaea (see descriptions and illustrations in Jepson, Fl. Calif. 3:9 1 , 1939; Abrams, loc. cit.; Munz, A California Fl., 442, 1959). However, in an unpubHshed manuscript, C. T. Mason, Jr. (1981) stated that G. setigera is actually the earliest name for G. bisetaea, and that a new name is needed for the species that has been confused with G. setigera in the various floras. In order to determine the correct application for the name G. setigera, we have reviewed pertinent literature and reexamined the holotype specimen {Bolander, No. 840 of the Kellogg and Harford distribution, GH!). Additionally, one of us (J.G.) visited the type locality (Red Mountain, Mendocino Co., CA) in company with Joann Holm, U.S. Bureau of Land Management, and collected a suite of specimens of the one Gentiana species found there (Greenleaf 1458, 4 Oct 1983, 2 sheets JEPS, 4 sheets OSC). We have seen only one further collection from the type locality (C P. Bonsall s.n., 1933, JEPS!). As discussed below, the critical morphological features of the holotype as well as the recently collected topotypes strongly support Mason's contention that G. setigera and G. bisetaea are synonymous. The misunderstanding by Jepson, Abrams, and others about the identity of G. setigera may have been due both to ambiguities in Gray's published description of the taxon and to these authors' unfamiliarity with the Oregon populations, named G. bisetaea by Howell. Gray's original description in Latin (Proc. Amer. Acad. 11: 84, 1876) and his later ones in English (Synoptical Fl. No. Amer. 2[1]:121, 1878; Bot. Calif. 1:482, 1880) fit the Bolander type-specimen very well, except that they do not emphasize enough the strikingly decumbent stems and overlook entirely the tuft of basal rosette leaves. One oblanceolate, acute rosette leaf is nearly hidden by a stem but is at least 5 cm long; another one, clearly exposed, is 4 cm long and 1 cm wide. At least four more rosette leaves are present and are 2-3 cm long. A basal tuft of leaves from the caudex is very characteristic of G. bisetaea but is poorly developed in the plants previously assigned to G. setigera. The flowering stems of the Bolander type resemble G. bisetaea in having closely spaced and fairly numerous lower leaf- pairs, mostly with well-developed blades, with longer intemodes distally and nar- rower-bladed leaves at the upper nodes. The lower cauline leaves have unusually broad blades (examples of length : width in cm are 2.0:1.5, 2.4:1.7, 2.3:1.6, 2.1:1.6), differing in this respect from the majority of G. bisetaea plants in Oregon. However, Madrono, Vol. 36, No. 1, pp. 49-52, 1989 50 MADRONO [Vol. 36 the 1983 population sample from Red Mountain, while notably broad-leaved, in- cludes some individuals that are indistinguishable from typical G. bisetaea both in habit and in leaf shape. Gray described the upper two pairs of stem leaves of G. setigera as forming an involucre to the solitary terminal flower. The 1983 collections from Red Mountain show this to be a variable trait, however; most plants have only the uppermost leaf- pair subtending the flower— as is typical of G. bisetaea. The variation in form of the appendages of the corolla sinuses is identical in G. bisetaea and G. setigera, both as described by Gray from the Bolander type and as noted in the 1983 samples from Red Mountain. We observed that the flowers of many plants on Red Mountain were paler blue than those in the Illinois Valley area, especially on the outer surface of the corolla. Those gentians from northwestern California and adjacent Oregon, to which the name G. setigera has been misapplied, differ from the plants described above in having strictly erect or ascending stems, a poorly developed basal rosette, broad cauline leaves nearly alike (except the lowest 2-3 pairs) and at equally spaced nodes up the stem, often several flowers at the apex, and corolla sinuses often with more numerous capillary appendages. Further study may show these plants to be distinct from the closely related G. calycosa Grisebach and worthy of species status. The Red Mountain population of G. setigera i=G. bisetaea) is about 225 km south of the nearest sites in Del Norte Co. and southwestern Oregon. It occurs in a wet meadow on a serpentine ridge at ca. 1065 m elevation. As presently understood, this species is rare in California, and due to the misuse of its name for a different taxon, its listing in the "Inventory of Rare and Endangered Vascular Plants of California" (Smith and Berg, CNPS Spec. Publ. No. 1, 4th ed., 58, 1988) should be reevaluated. In Oregon G. setigera is threatened by prospective nickel mining, although due to economic considerations it seems unlikely that extraction and smelting of nickel ore will occur in the near future. The nomenclatural change from G. bisetaea to G. setigera has little effect on the biological status of the species, as only a single widely disjunct population in California is being added to its previously known occurrences. We thank James Hickman for his help and encouragement with this study. Travel funds were provided by the Oregon State University Herbarium.— Kenton L. Cham- bers and Jacqueline Greenleaf, Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR 97331. (Received 23 Nov 1987; revision accepted 15 Oct 1988.) Infraspecific Name Changes in Limnanthes (Limnanthaceae).— In anticipation of a new edition of Jepson's "Manual of the Flowering Plants of California", it is necessary to make certain nomenclatural changes to provide uniformity throughout the genus Limnanthes. The International Code (Voss, 1983, Regnum Veg. Ill) provides no definitions for the taxa, subspecies and variety, and accordingly no distinction is made other than sequence if both are used. At the time of my Limnanthes monograph (Mason, 1952, Univ. Calif Publ. Bot. 25:455-512) I chose variety as the rank for the infraspecific taxa. In 1973 Arroyo (Brittonia 25:177-1 9 1) described several new taxa which she called subspecies. These appear to be taxonomically the same as my varieties. The following changes are made to elevate the several varieties to subspecies, and thereby standardize the taxonomy. Limnanthes douglasii R. Br. subsp. sulphurea (C. Mason) C. Mason, stat. nov.— Limnanthes douglasii var. sulphurea C. Mason, Univ. Calif Publ. Bot. 25:477. 1952. Limnanthes douglasii R. Br. subsp. nivea (C. Mason) C. Mason, stat. nov.— Li manthes douglasii var. nivea C. Mason, Univ. Calif Publ. Bot. 25:477. 1952. Limnanthes douglasii R. Br. subsp. rosea (Benth.) C. Mason, stat. noY.— Limnanthes rosea Benth., PI. Hartw. 302. 1848. 1989] NOTES 51 Limnanthes gracilis Howell subsp. parishii (Jepson) C. Mason, stat. now.— Lim- nanthes versicolor (E. Greene) Rydb. var. parishii Jepson, Fl. Calif. 2:41 1. 1936. Limnanthes alba Hartweg in Benth. subsp. versicolor (E. Greene) C. Mason, stat. nov. —Floerkea versicolor E. Greene, Erythea 3:62. 1895. —Charles T. Mason, Jr., University of Arizona Herbarium, Tucson, AZ 85721, (Received 21 Jul 1988; revision accepted 2 Dec 1988.) Reappraisal of the Range of Phacelia vallicola (Hydrophyllaceae). — Dis- covery of disjunct populations of Phacelia vallicola Congdon ex Brand (Hydrophyl- laceae) in Nevada and Sierra cos., California, inspired a search of major California herbaria for more information on the plant's distribution and habitat requirements. Until now P. vallicola was known from the W side of the Sierra Nevada in Mariposa, Tuolumne, Placer, and El Dorado cos. (Lee, Noteworthy collections, Madroiio 30: 129, 1983). The collections listed below from Butte, Madera, Nevada, Shasta, and Sierra cos. extend the known range more than 200 km (Fig. 1). Specimens examined. CA, Butte Co., overlook on the Skyway 4.2 km S of Neal Rd, T22N R3E NW'A of SE'A sect. 30, on the side of Tuscan outcrops on the rim of Little Butte Cr. Canyon, foothill woodland-chaparral ecotone, 335 m, 16 Apr 1986, Oswald 1069 (CHSC); Hwy 70 at SE entrance of Grizzley Dome Tunnel, along base of granite cliff in decomposed granite, 17 Apr 1976, Lickey 75 (CHSC); E of Feather Falls ca. 13 km, on dry bare lava cap, yellow pine f, 1000 m, 4 Jun 1982, Ahart 3534 (CHCS); Lumpkin Ridge, E of village of Feather Falls, T21N R7E sect. 36, in open on Lovejoy basalt, yellow pine f., 20 May 1981, Schlising 4060 (CHSC); ca. 12.9 km NE of Feather Falls, on dry bare broken black lava, yellow pine f , 1280 m, 12 May 1987, Ahart 5627 (CHSC). Madera Co., Canyon of Nelder Cr. ca. 16 km N of Oakhurst, slopes E of creek, T6S R21E SE'A of SE'A sect. 30, small domal granite outcrops in brushy yellow pine f., 9 May 1984, Jokerst 1999 (CHSC). Nevada Co., NW of Emigrant Gap 8 km, 0.8 km N of South Yuba River, T17N Rl IE SW'/4 sect. 2, foothill woodland, on metamorphic parent material, 1 May 1985, Bowcutt 499 (UC); W of Lake Spaulding dam 0.8 km, along Bowman Lake Rd 2.9 km from Hwy 80, 12.8 km SE of Washington, 0.2 km W of South Yuba River, T17N Rl IE SWA sect. 2, foothill woodland, on metamorphic rock outcrop and talus 1418 m, 1 May 1985, Bowcutt 501 (UC). Placer Co., N of Long Canyon Cr., E of Blacksmith Flat, T13N R13E NW'/4 of SWA sect. 7, on rock talus and bedrock outcrops in openings in S facing slopes, mixed coniferous f and chaparral, 1 128 m, 19 Jun 1984, Jokerst 2057 (CHSC). Shasta Co., 4.8 km E of Hwy 5, along Gilman Rd near Shasta Lake, 11 May 1983, Lennon s.n. (DAV, JEPS). Sierra Co., Lavezzola Creek Canyon, 12 km NE of Downieville, along U.S. Forest Service trail, T21N Rl IE SWy4 of NW'A sect. 33, 1390 m, 5 Aug 1985, Bowcutt 649 (UC); N of Pacific Mine ca. 0.4 km, ca. 8 km E of LaPorte, on dry bare rocky soil in yellow pine f., 28 Jul 1982, Ahart 3690 (CHSC); on dry rocky ridge above Foote Rd, ca. 8 km S of Alleghany, yellow pine f, 1219 m, 4 Jun 1978, Ahart 1770 (CHSC). Habitats. The following habitat description is based on herbarium label data and field observations over three years. The herbarium data presented above do not include the forty specimens collected in Mariposa and Tuolumne cos. that are housed at CAS, CHSC, HSU, JEPS, and UC. However, the range of habitats and elevations for these counties is represented in the description. Phacelia vallicola is found on granitic, metamorphic, and volcanic rock outcrops and talus slopes in foothill wood- land, yellow pine forest, mixed coniferous forest, and chaparral communities. Ac- cording to Munz (UC Press, 1973), Phacelia vallicola also occurs in red fir forest; however, I have not seen any specimens collected from this plant community. The species' known elevational range is from 335 to 2134 m. The plant is often common but scattered where it occurs. Most known populations are on U.S. Forest Service lands. Possible threats include hydroprojects and mining. Indirect impacts could result 52 MADRONO [Vol. 36 42*N Lat. Fig. 1 . Distribution of Phacelia vallicola Congdon. ex Brand. Locations based on herbarium specimens and the Uterature. due to logging of adjacent lands. The rock outcrops and talus slopes that support Phacelia vallicola support a sparse cover of trees and thus are not prime timber lands. Significance. Mariposa phacelia, Phacelia vallicola, is currently on the watch list of the inventory of California's rare and endangered plants [Smith and Berg, CNPS Spec. Publ. No. 1, 4th ed., 1988]. Based on the presented range extension and wide habitat requirements, Phacelia vallicola is too widespread to be considered a plant of concern. Based on the current knowledge of this plant's distribution, Mariposa phacelia would more appropriately be called Congdon's phacelia. J. W. Congdon, a physician in the foothill mining town of Mariposa and an active plant collector, discovered the plant in the late 1800's. I thank Dr. Gregory Lee for verifying the identifications of several specimens. Thanks also goes to Carrie Anne Shaw for assistance in the San Francisco Bay Area herbaria and to the curators of the cited institutions for access to specimens.— Frederica S. Bowcutt, California Dept. of Parks and Recreation, Box 942896, Sacramento, CA 94296-0001. (Received 5 Oct 1988; revision accepted 4 Nov 1988.) NOTEWORTHY COLLECTIONS California Pallavicinia lyellii (Hook.) S. Gray (Hepaticae: Pallaviciniaceae). —Humboldt Co., Areata, 0.5 mi up Jolly Giant Creek from Humboldt State University dormitories, in a moist recess of a rotten log, next to a creek, in a second growth redwood forest, T6N RIE sect. 28, 100 m, 2 Oct 1987, Wilson 1534 (NY, ORE confirmed D. H. Wagner). Significance. A species common in eastern North America; the nearest previously published record came from Minnesota. The new site is removed from present human activities and disturbances, providing no reason to assume that the plant's presence is the result of a recent anthropogenic introduction.— Paul Wilson, Department of Ecology and Evolution, State University of New York, Stony Brook, NY 1 1794. California P*ENSTEMON VENUSTUS Douglas cx Lindlcy (Scrophulariaceae). —Lassen Co., Sier- ra Nevada, SW side of Fredonyer Butte; on several roadcuts along CA hwy 36 through Susan River alluvium in pineland ca. 1 1.5 mi (18.5 km) WSW of Susan ville, T29N RlOE sect. 15, ca. 5200 ft (1585 m), collected in flower 25 Jun 1987, Vincent 4436 (NY), and in fruit 5 Aug 1987, Vincent 4500 (NY). Significance. First report for CA. Observed first in 1986, ca. 30 individuals were seen in 1987 on two roadcuts, so presumably naturalized. Native to drainages of the Columbia and Snake rivers of adjacent WA, OR, and ID, often on similar alluvial soils.— Karl A. Vincent, New York Botanical Garden, Bronx, NY 10458. Oregon Astragalus curvicarpus (Sheldon) Macbride var. subglaber (Rydb.) Bameby (Fabaceae).— Baker Co., W side of Powder River Canyon, NE of Baker in recently burned Artemisia tridentata/Festuca idahoensis-Agropyron spicatum community (T7S R40E sect. 13?), ca. 300 m, 19 Jun 1985, E. Joyal 883 (Baker-BLM, NY, OSC). (Verified by R. Bameby, NY.) Significance. Extension E of ca. 1 50 km from the lower Deschutes and John Day River drainages. Collomia debilis (S. Watson) E. Greene var. larsenii (A. Gray) Brand (Polemoniaceae).— Grant Co., Wallowa-Whitman Natl. For., Elkhom Crest trail S of Anthony Lake, common in talus along trail (T7S R37E sect. 30?), ca. 2500 m, 1 Sep 1985, Joyal 1030 (CS, OSC). (Verified by D. Wilken, CS.) Significance. Extension E of ca. 300 km from the Cascade Range of Oregon. RuDBECKiA OCCIDENT ALis Nutt. var. MONTANA (A. Gray) Perdue (Asteraceae).— Baker Co., Elkhom Range, Hunt Mt., Pine Cr., common along stream bank in Pseu- dotsuga menziesii forest opening (T8S R38E sect. 21), ca. 1700 m, 23 Jul 1985, E. Joyal 1004 (OSC); 20 Jul 1986, E. Joyal 1225 (Baker-BLM, CAS, NY, OSC, UTC). (Verified by A. Cronquist, NY.) Significance. First record for OR and an extension NW of ca. 900 km from S UT and W CO. — Elaine Joyal, The Nature Conservancy, 1815 N. Lynn St., Arlington, VA 22209. Madrono, Vol. 36, No. 1, p. 53, 1989 REVIEWS Guide to the Regional Parks Botanic Garden. By Walter Knight with Irja Knight. East Bay Regional Park District, 11500 Skyline Blvd., Oakland, CA 94619-2443. 1988. 490 pp.. Soft cover. $20.00. Any walk through the East Bay Regional Parks Botanic Garden at Tilden Park in Berkeley, CA is a delight. However, if you seek a particular taxon then the new revision of Guide to the Regional Parks Botanic Garden provides a valuable assist. Hereafter, I will refer to the revision as the Guide. The garden is divided into ten sections representing major botanic regions of California. The planted beds in each section are identified by color coded stakes. Individual plants are labeled with the same sectional color and also by number (e.g., the SIERRAN plants in Section 6 are in Bed numbers 601-662 and are color coded Blue). The Guide includes a California map illustrating the general location of the ten major botanic regions. A map of the Botanic Garden helps you to find a specific section in the garden. Inconveniently, these two important maps are 32 pages apart. The first 38 pages are numbered by Roman Numerals; the remaining pages of the Guide are not numbered. Access to the plant data is provided by an index at the end of the text. The index, nicely highlighted by buff-colored pages, is an alphabetical listing of scientific and common names, the latter in bold face. I would have preferred the scientific names to be bold face type. Numbers following the names refer to bed numbers in the several sections (e.g., Ceanothus roderickii 610, 647). Bed 647 contains eight species including 84.320 Ceanothus roderickii Knight. PINE HILL CEANOTHUS. The number 84.320 means that this specimen was the 320th acquisition made in 1 984. Other data include: family name, rare & endangered status, ~ location, habitat, flowering time, morphology, habit, and county. I tested the accuracy of the Guide by locating many plants and found no errors in my test sample. Besides the ten major sections there are three special habitats: the ANTIOCH DUNES area (yellow labels), the COASTAL DUNES area (yellow labels), and the POND area (small brown labels). White labels either indicate self-sown plants native to the park or plants not native to California. The Guide begins with pertinent "General Information" and a good "Using the Guide" portion. Also included are "more history", reports, a weed list, literature cited, Indian (Native American) culture literature, glossary, and illustrations of plant structural attributes. If the editors of a future revision would place much of the supporting information in appendices, then the two critical maps would be together. This would make the Guide even more effective. The East Bay Regional Parks Botanic Garden is a beautiful, functional asset to those of us who enjoy plants. The Guide to the Regional Parks Botanic Garden greatly enhances the value of the garden.— Clifford L. Schmidt, Department of Biological Sciences, San Jose State University, San Jose, CA 95192-0100. Inventory of Rare and Endangered Vascular Plants of California. Edited by James P, Smith and Ken Berg, illustrated by Loran May. California Native Plant Society, Special Publication Number One (4th ed.), Sacramento. Sep. 1988. 168 pp. Softbound. $19.95. The California Native Plant Society has, since its inception in the 1960's, been "Dedicated to the Preservation of California Native Flora". A central feature of this Madrono, Vol. 36, No. 1, pp. 54-58, 1989 1989] REVIEWS 55 dedication was, and is, the listing of plants that are threatened in some way. Starting with a card file of G. Ledyard Stebbins the society began to develop lists of Rare and/ or Endangered Plants. These were, at first, informal lists, but by 1974 they had been refined to the point that they were incorporated into the first edition of the Inventory. This first edition was edited by W. Robert Powell and was made possible by the contributions of time and talent by many professional and lay botanists. A number of eminent botanists were invited to join a Scientific Advisory Committee, while others combed herbaria and searched for the plants in the field. This Inventory quickly became the source book for anyone who needed information about endangered plants. An edition of the Inventory is based on the best information available at the time of publication. Therefore the subsequent editions have been evolutionary rather than revolutionary in nature. This undertaking of the Native Plant Society was the first of its kind. It became evident at the map-in of 1974 that information was entirely lacking or incomplete for many taxa throughout the state. It was six years before the backlog of information became so great that it was necessary to publish the second edition, although supplementary lists had been circulated. The hiring of a staff botanist and the acquisition of a computer made it easier for the society to perfect the lists. In fact there were a couple of supplements between the second edition of 1980 and the third edition of 1984. Supplements were not published between the third and fourth editions. Why publish a new edition now? Taxa formerly thought to be extinct have been rediscovered. Some thought to be rare were determined to be fairly common. Other taxa that were once found in relatively large numbers are now presumed to be extinct or nearly so. Many field studies have been undertaken under the direction of the several agencies involved with data collection. These have been done by volunteers, many of whom have been giving of their time and expertise since the beginning of the program. A number of experts in certain taxa have had time to complete studies which have resulted in changes of one kind or another. The resulting informational changes have been incorporated into the new edition. Changes in format have also been made. The one most obvious change has been made possible through the acquisition of computer software. It allows the society to convert data files to camera-ready copy which is produced in two columns with the type across the short dimension of the page. That is, the text may be read with the book held upright. Previous editions have had the data for each taxon spread across the length of the page. This change should meet with universal approval. A change which may prove more controversial is the incorporation of all the lists into one unit. This unit includes all taxa considered, in an alphabetical sequence. The lists are still the same: #1 A, Presumed extinct in California; #1B, Rare or endangered in California and elsewhere; #2, Rare and endangered in California, more common elsewhere; #3, Need more information; #4, Plants of limited distribution, a watch list. The assembled lists appear in Appendix II as long, continuous, alphabetical listings of scientific names only. When one needs to find a certain taxon, this arrangement of the taxa into one unit has certain advantages. Any taxon may be found easily, regardless of its listing. There are even plants in the main body which have been considered but rejected because they did not warrant inclusion in one of the lists. There are a few other minor changes. Edition 3 had each botanical name written entirely in upper case. Edition 4 returns to properly written scientific names using italics. The strategically placed drawings of Loran May enhance the appearance of the text, and the Prisma Color Pencil drawing of Barbara Adair makes this cover the most beautiful one yet. The book is very well edited. I found only a few very minor typographical errors. The new type-setting capabilities have improved the appearance of the text. Even last minute changes in the data have been incorporated. All in all the book reflects the many hours of work which have gone into its production. The editors, the many contributors and the Native Plant Society may be justly proud of the results of their 56 MADRONO [Vol. 36 efforts. — Malcolm G. McLeod, Biological Sciences Department, California Poly- technic State University, San Luis Obispo, CA 93407. Colorado Flora: Western Slope. By William A. Weber. Colorado Associated Uni- versity Press, Boulder, CO. 1987. 530 pp. Hardbound. $19.50. ICBN 0-87081-167-3. This handy-sized field guide is an expression of the author's broad knowledge of plants worldwide and years of field experience. The format is easy to use by the amateur (or the professional) and includes a glossary, 64 color plates and more than 100 line drawings. The flora includes all vascular plants (ferns and fern allies, gym- nosperms, and angiosperms) of the "entire hydrologic Western Slope of Colorado— from the Continental Divide to the Utah, Wyoming, and New Mexican borders." Each major group is presented in alphabetical sequence, in turn, by family, genus, and species. Families bear descriptions, but lower level taxa are mostly described only in the keys. This guide is very similar to its earlier version. The Rocky Mountain Flora (Weber, W. A., 1972, Colorado Assoc. Univ. Press, Boulder), but now includes more taxa in its larger geography, more numerous and better detailed illustrations, and more re- alistic family treatments of ferns and gymnosperms. The generic treatment will prob- ably disturb certain users. Many genera have been subdivided (even more so than in the earlier version), apparently due to the author's belief in narrow generic concepts and his worldwide knowledge of certain plant groups. He may be justified for these changes, but for those familiar with floras of the surrounding states, many new generic names will be hard to translate because the full direct synonyms used in the earlier version are no longer present; usually there is only a note to indicate in which genus it was formerly included. I believe that it is the duty of the author to defend the position (however correct) from commonly used scientific names. In general, I am much concerned about the diverse treatments of generic circum- scription by taxonomists. Obviously the taxonomist must have freedom of expression and judgement, but I am concerned when certain genera are split and reunited re- peatedly generation after generation. One suggestion might be that generic delimi- tation should (must?) include major differences in reproductive structures, not to be distinguished by vegetative characters alone, e.g., Berberis vs. Mahonia; Euphorbia vs. Chamaesyce, Poinsettia, etc.; Fouquieria vs. Idria; Potentilla vs. Argentina; Opun- tia vs. Cylindropuntia; and so on. I recommend this flora despite nomenclature inconveniences; the book size is right; the quality is good, and the price is right. — Donald J. Pinkava, Department of Botany, Arizona State University, Tempe, AZ 85287-1601. Trees and Shrubs of Trans- Pecos Texas. By A. Michael Powell. 1988. Big Bend Natural History Association, Big Bend National Park, TX. 536 pp. $19.95 (paper- bound) ISBN 0-912001-14-3. Trees and Shrubs of Trans-Pecos Texas is a complete and professional treatment of the woody plants west of the Pecos River in Texas. It begins with a short description of the area, including climate, soils, topography, and major vegetation types. A map of the counties and major topographic features clearly defines the area covered in the manual. Each of the five vegetation types in the Trans-Pecos is described using both common and scientific names for the dominant plant species and illustrated with photographs. The introduction is followed by a floristic treatment that includes keys to families, genera, and species as well as family and generic descriptions. Species accounts consist of fairly detailed distribution information and usually some inter- esting facts about the plant, ranging from economic uses to newly discovered localities. Most species are illustrated in fine pen and ink drawings showing vegetative, floral, and fruiting features. Common names and a short glossary are provided to aid novices. The manual is quite comprehensive and includes many slightly woody perennial herbs in addition to trees and shrubs. The nomenclature is up to date with only a 1989] REVIEWS 57 few exceptions (e.g., Petalostemum, a genus that Bameby sank into Dalea in 1977, is discussed as being poorly differentiated from Dalea). The dichotomous key to families is indented, easy to read, and numbered such that it is very easy to backtrack. This is a big advantage because my major criticism of the book is that Powell borrowed the family key from Correll and Johnston's Manual of the Vascular Plants of Texas. This key can be notoriously difficult and, while we are spared the "not as above" couplets, the dreaded "spirally coiled embryo" couplet still must be traversed to arrive at the Chenopodiaceae. Generic and species keys are much better, probably because many of them are tailored for the Trans-Pecos. The illustrations are very good and might make up for the defects in the key as one can leaf through the manual looking for the plant in question. A remarkable number of the species are illustrated (the Agavaceae and Cactaceae with photographs rather than line drawings) and although the drawings have been reduced in size to conserve space, the salient features can still be seen. There are many small errors scattered throughout the text and the book would have benefited from more careful editing. Thus, we find Cercocarpus is misspelled as Cercocarcarpus, Figures 2 and 3 in the introduction seem to have been switched, and Krameria parvifolia is found in every country rather than county. In his preface, Powell states that the book is intended for use by both scientists and non-scientists. It is obvious that he has made a serious attempt to avoid overusing technical terminology and to include common names. As a consequence, both the introduction and floristic treatments are eminently readable. Were it not for the family key, I would whole-heartedly recommend the book to any amateur botanist. The book is handsomely bound with a durable plastic-coated cloth cover and sewn pages and should withstand many seasons in the field. Trees and Shrubs of Trans-Pecos Texas is a valuable addition to the growing body of regional floras and will be a useful tool for all those interested in this beautiful and fascinating area. — Melissa LucKOW, Department of Botany, University of Texas, Austin, TX 78713. A Guide to Wildlife Habitats of California. K. E. Mayer and W. F. Laudenslayer, Jr. (eds.). 166 pp., paperbound. California Department of Forestry and Fire Protec- tion, Sacramento. 1988. Copies may be obtained from WHR Coordinator, Depart- ment of Fish and Game, 1701 Nimbus Road, Suite D, Rancho Cordova, CA 95670. $13.02 including tax and shipping. This is the first of several publications from the California Wildlife-Habitat Re- lationships (WHR) System designed to serve as ecosystem-oriented resources for wildlife biologists and managers in California. The Guide describes various wildlife habitats that constitute the WHR classification system that was developed by the California Interagency Wildlife Task Group. Its goal is to identify and classify existing vegetation types important to wildlife. One objective of the WHR system is to address the problem of confusion among vegetation/habitat-type classifications prepared for different purposes (e.g., wildlife biology, range management, forestry, etc.) by pro- viding a framework that can be used by all. The introduction describes the purposes for which the book has been designed, how it was constructed, and how it is to be used. It includes a tabular comparison of the WHR classification with others published for California vegetation. Contributions by various specialists have been tightly edited for standardization of format and information content. "Habitats" (communities) are grouped as "Tree-dominated," "Shrub-dominated," "Herbaceous-dominated," "Aquatic," and "Developed." The one- to two-page written summary of each habitat type begins with "Vege- tation," a section broken into paragraphs on structure (physiognomy), composition including dominants and major associated species, and a comparison of other clas- sification schemes. A section entitled "Habitat Stages" is a summary of current knowledge (often very little) about successional relationships. The "Biological Setting" includes a brief description of "habitat" that lists ecotonal relationships of the habitat 58 MADRONO [Vol. 36 type being described and "wildlife considerations" that summarizes the importance of the community to wildlife. "Physical Setting" describes the characteristics of the physical environment within which the habitat type occurs. "Distribution" includes a paragraphic summary with geographical range and elevational zonation. A map depicts the community's distribution or potential distribution in California. Some maps are much more generalized than others; for instance, the distribution of the "Blue-oak" habitat type is broken into many tiny patches whereas that of "Joshua Tree" is a single large blotch that includes many areas where this species does not occur. There is a single high-resolution color photograph of each community. Line drawings of various birds, mammals, reptiles, and amphibians that occur in California are dispersed through the text, generally one per community; these do not appear to be indexed in any way. A curious feature of the WHR classification is the unevenness of "habitat" defi- nition. This is noted in the introduction as a "Caveat for Users," but the rationale for the differences in inclusiveness is not always readily apparent. For example, "Blue Oak" is treated as a separate habitat from "Blue Oak-Digger Pine" though the two intergrade extensively and share the same suite of associates. On the other hand, "Coastal Scrub" is defined so broadly that communities with no species in common at all are included together. Within "Coastal Scrub" the plants of beaches and dunes are grouped together with species of other very different environments and are not even mentioned in the community discussion. One would hope that management practices for dune communities would be different from those of areas with more stable substrates and different ecological constraints. A "habitat" entitled "Chamise- Redshank Chaparral" is depicted as occurring throughout most of cismontane Cal- ifornia although redshank {Adenostoma sparsifolium) is found no farther north than San Luis Obispo County and forms communities quite different in physiognomy and stature than chamise {Adenostoma fasciculatum). The book has a collective "Literature Cited" section. Common names are used in the text, and are cross-referenced to species names in a "Species List" at the end of the book. This is a one-way cross-reference; a user who knows a species name but not the common name can spend a lot of time looking. Additionally, some common names are of peculiar derivation, and may not be the names used commonly by people in the field— always a problem with common names. A shortfall of the book is the lack of an index. Despite its shortcomings, the Guide is sure to be a useful addition to the information available to wildlife biologists and wildlife managers. The price is very affordable. I hope that the editors will readily accept input from the botanical community and prepare a revised edition before too much time has passed. — David J. Keil, Biological Sciences Department, California Polytechnic State University, San Luis Obispo, CA 93407. ANNOUNCEMENT New Publications Albee, B. J., L. M. Shultz, and S. Goodrich. 1988. Atlas of the vascular plants of Utah. 670 pp. ISBN 0-940378-09-4. Utah Museum of Natural History, Salt Lake City, UT 841 12. $26.00 (clothbound) + $4.00 shipping and handling. [The atlas includes distribution data in the form of dots on detailed relief maps for each of the 2438 species growing in Utah in more than one location and without benefit of cultivation. Data for the publication were obtained by critical ex- amination of some 400,000 specimens housed in the various herbaria in Utah. Phenological and ecological data are provided for each species, an annotated list of 384 species restricted in distribution is included in an appendix, and alternate scientific names are provided in the index.] Bowers, J. E. 1988. A sense of place. The life and work of Forrest Shreve. University of Arizona Press, 1230 N. Park, #102, Tucson, AZ 85719. $19.95 (clothbound). [A biography of pioneer plant ecol- ogist, Forrest Shreve, whose work laid the foundation for many sub- sequent ecological studies in western North America.] BuRBRiDGE, J. 1989. Wildflowers of the southern interior of British Columbia and adjacent parts of Washington, Idaho, and Montana. 400 pp., approx. ISBN 0-7748-0320-7. University of British Colum- bia Press, 6344 Memorial Road, Vancouver, B.C. V6T 1 W5, Canada. $29.95 (clothbound) or $19.95 (paperbound) + postage and handling ($1.60 in Canada; $3.50 in the U.S.). [A field guide illustrated with 335 color photographs, diagrams, and 1 map.] Ferren, W. R. and D. A. Pritchett. 1 988. Enhancement, restoration, and creation of vernal pools at Del Sol Open Space and Vernal Pool Reserve, Santa Barbara County, California. Environmental Research Team, The Herbarium, Department of Biological Sciences, Univer- sity of California, Santa Barbara, CA 93106. Environmental Report No. 13. 169 pp. plus two booklets. $15.00 (paperbound). Klinka, a., V. J. Krajina, A. Ceska, and A. M. Scagel. 1989. In- dicator plants of coastal British Columbia. 330 pp., approx. ISBN 0-7748-0321-5. University of British Columbia Press, 6344 Memo- rial Road, Vancouver, B.C. V6T 1W5, Canada. $36.95 + postage and handling ($1.60 in Canada; $3.50 in the U.S.). [419 selected vascular plants, bryophytes, and lichens of coastal British Columbia are described and illustrated in color. Indicators with similar values are grouped into indicator species groups that are used to evaluate site quality. Information is presented on geographical distribution, life-form, shade tolerance, and other ecological characteristics. Three methods are presented for use of indicator plants for site diagnosis.] ANNOUNCEMENT Minor Changes in Madrono Format In response to comments from authors and reviewers, I polled past Editors, and members of the Editorial Board regarding various editorial matters. Beginning with this issue, readers will see several minor stylistic changes in Madrono's format that resulted from the various comments. The changes are summarized below. For mailing addresses within the United States, the postal abbrevi- ation of the state will be included whether the name of the state is mentioned in the institutional title or not. Postal abbreviations will also be used in specimen citations, except in the case of types where the name of a state will be spelled out in full. To the extent possible, citation of authors for scientific names will follow the Draft index of author abbreviations compiled at the Royal Botanic Gardens at Kew (1980). Authors who do not have access to a copy of the Draft index should note this in the cover letters with their manuscript submissions to alert the Editor to non-standard author ab- breviations. In adopting the Draft index as a standard. Madrono joins Systematic Botany and the Jepson Manual Project among others. There is a minor change in punctuation of literature citations within the text of an article where two or more papers are cited together. Examples of the new style follow: (Jones 1976, 1977), (Jones 1976; Smith 1987), (Jones 1976, 1977a, b; Smith 1987). Note the use of semicolons where commas formerly appeared. Citation of nomenclatural types has been somewhat inconsistent in the past. Type citations will now appear immediately after the names that they typify and before the diagnosis and/or description of the taxon (as on p. 18, this issue). In all cases the specific kind of type will be used; the word "holotype," for instance, will be used and not implied by the word "type." Unless the reasons for lectotype choice are partic- ularly complicated, the rationale for lectotypification will accompany the type citation. Dichotomous keys now appear in 8-point type (as on p. 3 1 , this issue) instead of 10 point. Compass directions now appear as capital letters without punctuation rather than as lower case with punctuation. Observant readers may note a few other changes that are not men- tioned here. I apologize in advance to authors whose manuscripts are caught in the transition. — David J. Keil, Editor. Volume 36, Number 1, pages 1-60, published 3 May 1989 Subscriptions — Membership Membership in the Cahfomia Botanical Society is open to individuals ($22 per year; students $ 1 2 per year for a maximum of seven years). Members of the Society receive MadroRo free. Family memberships ($25) include one five-page publishing allotment and one journal. Emeritus rates are available from the Corresponding Secretary. Institutional subscriptions to Madrono are available ($30). Membership is based on a calendar year only. Applications for membership (including dues), orders for sub- scriptions, and renewal payments should be sent to the Treasurer. Requests and rates for back issues, changes of address, and undelivered copies of Madrono should be sent to the Corresponding Secretary. Information for Contributors Manuscripts submitted for publication in Madrono should be sent to the editor. All authors must be members, and membership is prerequisite for review. Manuscripts and review copies of illustrations must be submitted in triplicate for all articles and short items intended for NOTES AND NEWS. Follow the format used in recent issues for the type of item submitted and allow ample margins all around. All manuscripts MUST BE DOUBLE SPACED THROUGHOUT. For ar- ticles this includes title (all caps, centered), author names (all caps, centered), addresses (caps and lower case, centered), abstract, text, acknowledgments, literature cited, tables (caption on same page), and figure captions (grouped as consecutive paragraphs on one page). Order parts in the sequence listed ending with figures, and number each page. Do not use a separate cover page, "erasable" paper, or footnotes. Manuscripts prepared on dot matrix printers may not be considered. Table captions should include all information relevant to tables. All measurements should be in metric units. Line copy illustrations should be clean and legible, proportioned (including cap- tions) to the Madrono page, and designed for reduction to % original size. Scales should be included in figures, as should explanation of symbols, including graph coordinates. Symbols smaller than 1 mm after reduction are not acceptable. Maps must include latitude and longitude references. Halftone copy should be designed for reproduction at actual size. In no case should original illustrations be sent prior to the acceptance of a manuscript. When needed they should be mounted on stiff card- board and sent flat. No illustrations larger than 22 x 28 cm will be accepted. Presentation of nomenclatural matter (accepted names, synonyms, typification) should follow the format used for Rhus integrifolia in Madrono 22:288. 1974. Institutional abbreviations in specimen citations should follow Holmgren, Keuken, and Schofield, Index Herbariorum, 7th edition. Abbreviations of serial titles should be those in Botanico-Periodicum-Huntianum (Lawrence et al., 1968, Hunt Botanical Library, Pittsburgh). If the correct abbreviation cannot be determined, the full serial title should be used. Titles of books should be given in full, together with the place and date of publication, publisher, and edition, if other than the first. All members of the California Botanical Society are allotted five free pages per volume in Madrono. Joint authors may split the full page number. Beyond that number of pages a required editorial fee of $65.00 per page will be assessed. The purpose of this fee is not to pay directly for the costs of publishing any particular paper, but rather to allow the Society to continue publishing Madrono on a reasonable schedule, with equity among all members for access to its pages. Printer's fees for illustrations and typographically difficult material @ $35.00 per page (if their sum exceeds 30 percent of the paper) and for author's changes after typesetting @ $4.50 per line will be charged to authors. At the time of submission, authors must provide information describing the extent to which data in the manuscript have been used in other papers that are published, in press, submitted, or soon to be submitted elsewhere. CALIFORNIA BOTANICAL SOCIETY STATEMENT OF OWNERSHIP, MANAGEMENT, AND CIRCULATION (Required by Title 39, U.S.C. 3685) MADRONO, A West American Journal of Botany, is published quarterly at Berke- ley, California. Annual subscription price is $25.00. The Publisher is the California Botanical Society, Inc., Life Science Building, Uni- versity of California, Berkeley, CA 94720. The editor is David J. Keil, Biological Sciences Department, California Polytechnic State University, San Luis Obispo, CA 93407. The owner is the California Botanical Society, Inc., Life Science Building, Uni- versity of California, Berkeley, CA 94720. There are no bondholders, mortgagees, or other security holders. The average number of copies distributed of each issue during the preceding 12 months is 1 108; the number of copies of the single issue closest to the filing date is 1 132. The net press run is 1250 copies. I certify that the statements made by me above are correct and complete. 24 February 1989 DAVID J. KEIL, Editor VOLUME 36, NUMBER 2 APRIL-JUNE 1989 i MADRONO A WEST AMERICAN JOURNAL OF BOTANY Contents Flora of the Sierra de la Laguna, Baja California Sur, Mexico Jose Luis Leon de la Luz and Raymundo Dominguez-Cadena 61 Ecology and Distribution of Pinus lacunae in the Sierra de la Laguna, Baja California Sur, Mexico Marie- Franqois Passini and Nicole Pinel 84 Posthre Vegetation Development in the Costa Rican Paramos Sally P. Horn 93 A New Species of Erigeron (Asteraceae: Astereae) from Central New Mexico Richard Spellenberg and Paul Knight 1 1 5 A Re-evaluation of the Allium sanbornii (Alliaceae) Complex Stella S. Denison and Dale McNeal 1 22 NOTES Myosotis latifolia and not M. sylvatica (Boraginaceae) in California Elaine Joyal 1 3 1 MooNwoRTS (Botrychium: Ophioglossaceae) in the Jonesville Area, Butte and Tehama County, California Warren H. Wagner, Jr. and Timothy B. Divine 1 3 1 Salix scouleriana J. Barratt ex Hook. (Salicaceae) in Sonora, Mexico George W. Argus 135 Typihcation of Salix geyeriana (Salicaceae) Robert W. Dorn 135 NOTEWORTHY COLLECTIONS Arlzona 136 COMMENTARY Points of View: On the Modern Death of David Douglas James L. Reveal 137 ANNOUNCEMENTS ^ Temporary New Address for Editor of Madrono 121 New Publications a 140 i PUBLISHED QUARTERLY BY THE CALIFORNIA BOTANICAL SOCIETY Madrono (ISSN 0024-9637) is published quarterly by the California Botanical So- ciety, Inc., and is issued from the office of the Society, Herbarium, Life Sciences Building, University of California, Berkeley, CA 94720. Subscription rate: $30 per calendar year. Subscription information on inside back cover. Established 1916. Second-class postage paid at Berkeley, CA, and additional mailing offices. Return requested. Postmaster: Send address changes to James R. Shevock, Botany Dept., California Academy of Sciences, San Francisco, CA 941 18. Editor— David J. Kjeil Biological Sciences Department California Polytechnic State University San Luis Obispo, CA 93407 Board of Editors Class of: 1989 — Frank Vasek, University of California, Riverside, CA Barbara Ertter, University of California, Berkeley, CA 1990— Steven Timbrook, Ganna Walska Lotusland Foundation, Montecito, CA Thomas R. Van Devender, Arizona-Sonora Desert Museum, Tucson, AZ 1991— James Henrickson, California State University, Los Angeles, CA Wayne R. Ferren, Jr., University of California, Santa Barbara, CA 1992 — Bruce A. Stein, The Nature Conservancy, Washington, D.C. William L. Halvorson, Channel Islands National Park, Ventura, CA 1993— Jon E. Keeley, Occidental College, Los Angeles, CA Rhonda L. Riggins, California Polytechnic State University, San Luis Obispo, CA CALIFORNIA BOTANICAL SOCIETY, INC. Officers for 1988-89 President: John L. Strother, University Herbarium, University of California, Berkeley, CA 94720 First Vice President: James Affolter, Botanical Garden, University of California, Berkeley, CA 94720 Second Vice President: James Henrickson, California State University, Los An- geles, CA 90032 Recording Secretary: Rodney G. Myatt, Department of Biological Sciences, San Jose State University, San Jose, CA 95192 Corresponding Secretary: James R. Shevock, Department of Botany, California Academy of Sciences, San Francisco, CA 94118 Treasurer: Thomas F. Daniel, Department of Botany, California Academy of Sci- ences, San Francisco, CA 94 1 1 8 Financial Officer: Cherie L. Wetzel, Department of Biology, City College of San Francisco, 50 Phelan Ave., San Francisco, CA 941 12 The Council of the California Botanical Society consists of the officers listed above plus the immediate Past President, Dale McNeal, Department of Biological Sciences, University of the Pacific, Stockton, CA 9521 1; the Editor of Madrono; three elected Council Members: John Mooring, Department of Biology, University of Santa Clara, Santa Clara, CA 95053; Barbara Ertter, Herbarium, Botany Department, Uni- versity of California, Berkeley, CA 94720; Elizabeth McClintock, Herbarium, Bot- any Department, University of California, Berkeley, CA 94720; and a Graduate Student Representative, Valerie Haley, Department of Biological Sciences, San Jose State University, San Jose, CA 95192. THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER. FLORA OF THE SIERRA DE LA LAGUNA, BAJA CALIFORNIA SUR, MEXICO Jose Luis Le6n-de la Luz and Raymundo Dominguez-Cadena Centre de Investigaciones Biologicas de Baja California Sur, Apdo. Postal 128, La Paz, Baja California Sur, Mexico Abstract The Sierra de la Laguna is the main high mountain range in the southern portion of the arid state of Baja California Sur (Mexico). It is high and narrow, rising boldly from coastal lowlands, with many precipitous and rocky slopes. Its peaks reach up to 2200 m. Above 1500 m the Sierra is occupied by the only montane woodland community in the state and it is believed to have been an island in both the strict and biological senses. Our botanical survey in the oak-pine woodland community of the Sierra de la Laguna, between 1982 and late 1987, reports 66 families of vascular plants, including 228 taxa. These are distributed in four groups by habit: 8% are trees, 10% are shrubs or subshrubs, 43% are short-lived perennial herbs, and 14% are hydrophytes. About 1 7% of this flora may be considered endemic. Almost a century ago, a survey in approximately the same area (by T. S. Brandegee) yielded 146 taxa of vascular plants. The main purpose of this study is to document botanical resources in support of the establishment of this plant community as a preserve, in accordance with some MAB-UNESCO fundamentals. Resumen La Sierra de la Laguna se ubica en la porcion mas elevada de las montanas del sur del arido estado de Baja California Sur (Mexico). La serrania se eleva sobre de una inmensa planicie costera y se caracteriza por lo pronunciado de sus declives. Las cimas alcanzan hasta 2200 m de altitud. Por encima de los 1500 m se encuentra la unica comunidad propiamente boscosa de la entidad, entre esta y los 1000 m se ubica un encinar; esta comunidad puede ser considerada como una isla tanto en el sentido estricto como biologico. Las exploraciones efectuadas han rendido hasta la fecha 228 especies de plantas vasculares comprendidas en 66 familias. Estas se han ubicado en las siguientes formas de vida: 8% arboles, 10% arbustos o subarbustos, 43% de herbaceas, y 14% de hidrofitas. Casi el 1 7% de la flora considerada es endemica. Hace casi un siglo T. S. Brandegee realize una investigacion floristica en la misma area reportando solo 146 taxa. El objetivo central de este trabajo consiste en actualizar el conocimiento floristico de la serrania en referencia, actualmente en vias de decretarla como una Reserva de la Biosfera de acuerdo con los planteamientos de MAB- UNESCO. The flora of the Baja California peninsula has been the concern of American botanists for over a century and a half Early collections were made by J. Xantus, L. Belding, T. S. Brandegee, C. A. Purpus, E. A. Goldman, and many others (Johnson 1958). Recent publica- tions include those of Shreve and Wiggins (1964), Wiggins (1980), and Gould and Moran (1981). During the present century, several California botanists have contributed greatly to our knowledge of Madrono, Vol. 36, No. 2, pp. 61-83, 1989 62 MADRONO [Vol. 36 the peninsular flora: M. E. Jones, A. Eastwood, I. M. Johnston, R. Moran, and A. Carter. Most of these investigations have focused on the central and northern areas of Baja California. In the southern tip of the peninsula, botanical explorations have been rare. This area was designated as a biogeographical unit by Bryant (1891) as "Cape Region of Baja California". Nelson (1921) and Shreve (1937) published early botanical descriptions of the arid and tropical portions of the region. More recently, Gilmartin and Neighbours (1978) undertook field work in hopes of preparing a flora of this region, but their project was never completed. Within the Cape Region, the Sierra de la Laguna resembles an island. A vegetation of mesic affinities is now restricted to its highest elevations. The only reports on its plant composition were published at the end of the last century (Brandegee 1891, 1892a,b, 1894, 1903). The oak-pine woodland community that occurs in the upper ele- vations contains a high proportion of endemic taxa. Increasing hu- man settlement in the vicinity has resulted in destructive use of this natural resource, and thus the plant community may be at risk of losing its natural balance. Study Area Topography. The mountains of the Cape Region extend in a south- north direction, from 23°00' to 23°35'N lat. The range is crossed by the Tropic of Cancer. About 500 square km of these mountains are estimated to be above 1000 m elevation with the highest peak at 2200 m (El Picacho). The Sierra de la Laguna is located in the northern part of the Cape Region mountains. It includes five major canyons (Fig. 1), having an estimated surface of about 100 square km. Geology and soils. The Sierra de la Laguna is composed totally of massive intrusive rocks, granites and syenites for the most part. It is an extension of a great batholith of Upper Jurassic or Lower Cretaceous age, which underlies most of the peninsula and presum- ably also parts of the Gulf of California (Beal 1948; Durham and Allison 1960). Most of these rocks are moderately coarse-grained and subject to rapid disintegration. The soils are sandy, with a thin layer of litter; the content of loam and clay is low. On slopes, foot- hills, and alluvial plains there are no differentiated soil layers. The soils are classified as Regosols (FAO-UNESCO system modified by Mexican government; Secretaria de Programacion y Presupuesto 1 98 la). At the bottom of brooks and canyons some permanent pools occur on the hard rock bed. The courses of the canyons are a direct consequence of active erosion along their escarpments (Hammond 1954; Lopez Ramos 1973). The eastern face of the Sierra is more precipitous than the western slopes. Similar patterns are present also 1989] LEON AND DOMINGUEZ: SIERRA DE LA LAGUNA 63 I— I bO •^ _J 1-1 O CI rr <^ S • i-H G 00 c l/i >. 73 a • ^ 73 (U cS 00 o^ n ^ cd ^ CX O o o o w o cd O oo OS cr • l-H o cx o OS cd r-~ o ^ (N c/) ^ 1/5 m cd ■(-> d s cx o 00 ■c oo W — c XI Z2 00 OS +-> o o o cd ■(-> w I Id CX cd OO o On 43 Cd •.-( (— ) ^ Oh ^ — 1 OS (U 00 cd W o C/2 ■HJ 00 O H-i <4i in •a ^x c cx cd > cd u op ^ cx o ' 1-H ■t-l Cd C • i-i c D Co -Si (U o C cd • 1-H B O Q 000 ir^ ^ OS 000 m (N 00 000 00 fN ^ 000 vo cn (N 000 vo 00 'O CO cx ■*-H c o • r—t +-> Cd ■t-> (U 00 > Vh > o o ■(-< c Oh O O O 00 m "T^ o o Vi (U a 1/3 c CO c o ■(-J o <5j o ■*-> 3 s; I I o U Q 1/3 CO ^ t 5 b 5^ o s; t« ill ^ is oq Q C^) bo C -5; 6<) * »^ o U 2 ^ 1/3 Si Q s; ^ O O O ci. to 43 t« I u I o 90 MADRONO [Vol. 36 2073 Fig. 3. Tall dense forest of Pinus lagunae and Quercus devia. 1, Pinus lagunae; 2, Quercus devia; 3, Arbutus peninsularis; 4, Nolina beldingii; 5, Lepechinia hastata; 6, Rumfordia connata; 7, Acacia peninsularis; 8, Muhlenbergia sp. growing with Erythea brandegeei, Populus brandegeei, and Salix sp., in an association similar to the "bosque en galeria" described by Rzedowski (1978). Two-hundred-year-old pines are found at 1500 m with low, open stands of Quercus tuberculata and Dodonaea vis- cosa. On the east slope, small Pinus lagunae (< 5 m) grow in isolated groves between 1 200 and 1 500 m. Extensive forests grow with Quer- cus devia above 1600 m. The southern limit of pine is near Canon San Jorge at 1700 m. Pinus lagunae is absent from Quercus devia stands, on open, sunny southern exposures from Canon de Agua Caliente (Fig. 2, transect 1-1'). It is also absent from southern exposures in the Sierra San Lazaro and San Lorenzo at 1 700 m. Quercus devia may withstand more xeric conditions than Pinus lagunae. Pinus lagunae Forest Types Ecologic (Table 1) and floristic surveys (Table 2) show Pinus la- gunae vegetation to consist of 1) tall, open Pinus lagunae forests; 2) tall, dense Pinus lagunae forests with Quercus devia; and 3) low. 1989] PASSINI AND PINEL: FINUS LAGUNAE 91 Open Pinus lagunae forests with understory of Quercus devia and Quercus tuberculata. 1 . The tall, open Pinus lagunae forests can be seen above 1 700 m, on level, sandy basins (Fig. 2, sec. 2) or on the rounded summits of Sierra de la Laguna. Trunks average 8 m apart. Pinus lagunae has a pyramid shaped crown, 12 to 18 m high; first branches are about 2 m above ground. There is little shrub growth but the her- baceous layer (0-1 m high) is diverse especially after the rainy season. Dominant species, including Bidens sp., Castilleja bryantii, Dalea sp., Desmodium neomexicanum, Helianthemum glomeratum, Hy- pericum peninsulare, Linanthus nuttallii, and Stachys coccinea, dis- appear between December and July. They are replaced by a "pas- tizal" of Gramineae, including Agrostis exarata, Agrostis semiverticillata, Aegopogon tenellus, Aristida schiediana, Bouteloua hirsuta, Lycurus phleoides, Muhlenbergia microsperma, Muhlenber- gia rigida, and Piptochaetium fimbriatum. 2. The tall, dense Pinus lagunae and Quercus devia dominant formations prevail on steeper upper flanks of the Sierra de la Laguna (Fig. 3). Trunks average 5 m apart. Individual trees of Pinus lagunae have longer boles than in open forest areas on the summit. The first branches are at 3 to 7 m above ground. Pine diameters vary from 32 to 58 cm (n = 50; mean = 45 cm). Annual rings from wood samples taken by a Pressler auger show that 34-cm-diameter pines average 60 years of age. Numerous young 0.5-2-m trees can be seen in open spaces, 50-200 per hectare. Quercus devia has an average height of 10 m and an average diameter of 40 cm. Three hundred specimens form 35% coverage on the experimental site. Contiguous cover of this species protects shrub layer growth during the dry season. It sheds its leaves in late February, flowers in March, and is in full growth by the end of March. Acorns ripen in September. Other trees in this type include Arbutus peninsularis (madrono, 5 m) and Nolina beldingii (2 to 4 m). A dense shrub layer is dominated by Lepechinia hastata, Rumfordia connata, Tagetes lacera, Callian- dra brandegeei, and Acacia peninsularis. The herbaceous layer, though less dense and less rich than in the previous type, includes a number of Gramineae: Agrostis s^., Aristida schiediana, Piptochaetium fimbriatum, Muhlenbergia emersleyi, Per- eilema crinitum, Schizachyrium sp., and Stipa sp. Species other than the Gramineae are the same as those encountered in tall Pinus la- gunae dominant forests. 3. Low open forests of Pinus lagunae, Quercus devia, and Quercus tuberculata are found on the lower edge of the previous type. It represents the ecotone between Pinus lagunae and Quercus devia forests. Pines have a more stunted habit, growing no higher than 1 2 m. The shrub layer includes Croton sp., Dodonaea viscosa, and Rhus integrifolia. The herbaceous layer is mainly made up of Gramineae. 92 MADRONO [Vol. 36 Summary Pinus lagunae appears above 1 200 m in Sierra de la Laguna which represents one of the lowest altitudinal limits observed among pines in the "cembroides group". Pinus lagunae stands fall into three types: a low thin forest of Pinus lagunae and Quercus tuberculata below 1500 m, a tall dense stand of Pinus lagunae and Quercus devia, found throughout the Sierra above 1500 m on the summit, and a tall open Pinus lagunae forest. Pinus lagunae display maximum growth rate and seed production in the latter type. Acknowledgments This study was part of the research program on xeric pine formations led by Passini. Field studies were conducted by Nicole Pinel, from October 1984 until July 1985, in collaboration with the "Centro de Investigaciones Biologicas" de La Paz (C.I.B.) and the "Instituto Nacional de Investigaciones Forestales" (I.N.I.F.), Todos Santos (Baja California Sur). The authors are grateful to Annetta Carter, Richard A. Minnich, and David J. Keil, for their numerous helpful comments. Literature Cited Bagnouls, F. and H. Gaussen. 1953. Saison seche et indice xerothermique. Bull. Soc. Hist. Toulouse 88:193-239. Bailey, D. K. 1983. A new allopatric segregate from and a new combination in Pinus cembroides Zucc. at its southern limits. Phytologia 54:89-99. Durham, J. W. and E. Allison. 1960. The geologic history of Baja California and its marine faunas. In Symposium: The biogeography of Baja California adjacent seas, Syst. Zool. 9:47-91. GoDRON, M., et al. 1 968. Releve methodique de la vegetation et du milieu. C.N.R.S., Paris, 292 pp. Passini, M.-F. 1982. Les forets de Pinus cembroides au Mexique, etude phytogeo- graphique et ecologique. Ed. Recherches sur les Civilisations, Paris. 374 pp. . 1987. The endemic pinyon of Lower California: Pinus lagunae M.-F. Pas- sini. Phytologia 63:337-338. Pinel, N. 1985. La formation a Pinus cembroides var. lagunae dans la Sierra de la Laguna, Basse Califomie, Mexique. Rapport stage de Diplome d'Etudes Appro- fondies, Toulouse. Reygadas and Velazquez. 1981. Aporte energetico y climatologico en la Sierra de la Laguna, Baja California Sur. Informe general del Centro de Investigaciones Biologicas, La Paz, pp. 235-243. Robert, M.-F. 1973. Contribution a I'etude des forets de Pinus cembroides dans Test du Mexique. These de 3eme cycle, Montpellier. 1 3 1 pp. Robert-Passini, M.-F. 1981. Deux nouveaux pins pignons du Mexique. Bull. Mus. Hist. Natn., Paris, 4, 3, sect. B, Adansonia 1:61-73. RzEDOwsKi, J. 1978. Vegetacion de Mexico, Ed. Limusa, Mexico. 432 pp. Shreve, F. and I. L. Wiggins. 1964. Vegetation and flora of the Sonoran desert. Vol. I. Stanford University Press. 839 pp. (Received 4 Jan 1988; revision accepted 28 Oct 1988.) POSTFIRE VEGETATION DEVELOPMENT IN THE COSTA RICAN PARAMOS Sally P. Horn Department of Geography and Graduate Program in Ecology, The University of Tennessee, Knoxville, TN 37996-1420 Abstract Postfire vegetation development was studied at four recent bum sites within the grass- and shrub-dominated paramos of the Cordillera de Talamanca, Costa Rica. The bamboo Swallenochloa subtessellata and the ericaceous shrubs Vaccinium con- sanguineum and Pernettia coriacea resprout vigorously after fire, but the shrub Hy- pericum irazuense suffers high mortality and reestablishes by seed. Herbs and shrubs are slow to colonize openings created by fire, and bare patches of ground persist for a decade or more following burning. Regenerating bamboo clumps regain prefire heights of 1-2 m within ten years, but associated shrubs require more than a decade to regain comparable prebum statures. Resumen Se estudio el desarrollo de vegetacion en cuatro sitios quemados despues de unos incendios en los paramos dominados por gramineas y arbustos de la Cordillera de Talamanca, Costa Rica. El bambu Swallenochloa subtessellata y las ericaceas Vac- cinium consanguineum y Pernettia coriacea se retofian vigorosamente despues de quemarse, pero el arbusto Hypericum irazuense sufre alta mortalidad, y se reestablece por semillas. Colonizacion por hierbas y arbustos avanza lentamente, y sitios de tierra sin cubierta vegetal persisten por una decada o mas. Los bambus quemados reponen su altura original de 1-2 m dentro de diez anos, pero los arbustos necesitan mas de una decada para recuperar estaturas comparables. The grass- and shrub-dominated vegetation found above timber- line in the Cordillera de Talamanca, Costa Rica, shows close bo- tanical affinity with the Andean paramos, and is generally regarded as representing the northernmost extent of paramo vegetation in the neotropics (Weber 1959; Cuatrecasas 1979; Lauer 1981). Like their South American counterparts, the Costa Rican paramos have been subjected to considerable human disturbance, including repeated burning. Carelessly tossed matches and cigarettes, arson, and heli- copter and airplane crashes are among the sources of recent paramo fires. The high, incidence of thunderstorm activity in Costa Rica (World Meteorological Organization 1953, 1956) suggests that light- ning may also serve as an ignition source; however, there are as yet no documented cases of lightning-set fires. Madrono, Vol. 36, No. 2, pp. 93-114, 1989 94 MADRONO [Vol. 36 The impact of burning on neotropical paramo vegetation has re- ceived little attention. Grubb (1 970) discussed changes in herbaceous cover following human-set fires in the paramo of Cerro Antisana, Ecuador, and Smith (198 1) described damage to rosettes of Espletia shultzii Wedd. caused by burning in a Venezuelan paramo. In Costa Rica, Janzen (1973a) studied vegetation recovery following a 1969 fire on Cerro Asuncion. He provided data on regeneration rates for the dominant woody dicots, but did not examine differences in sur- vivorship. More recently, Williamson et al. (1986) monitored pat- terns of postfire vegetation recovery on nearby Cerro Zacatales. Their study revealed evidence of differential shrub mortality, which they related to fire history. Chaverri and associates (1976) established quadrats within the paramo of Chirripo National Park, Costa Rica following a major wildfire in March of 1976, but the results of their long-term study have not yet been published. Study Areas Studies of postfire regeneration were carried out in the highlands surrounding Cerro Chirripo (3819 m), the highest point in Costa Rica, and near Cerro Buenavista (349 1 m) along the Inter-American highway crossing in the northwestern end of the Cordillera de Tala- manca (Fig. 1). The granitic rocks that form the backbone of the Cordillera are exposed near Cerro Chirripo but mantled by basalts and pyroclastic deposits in the Buenavista massif (Weyl 1957). Soils in both areas are generally well-drained, rich in organic matter, and acidic, with pH values as low as 4.0 (Otarola 1976). Soil depths range from a few centimeters on the summits of the peaks to over 50 cm near the upper forest limit. Glaciers occupied the upper por- tions of valleys in the Chirripo highlands during the Pleistocene (Hastenrath 1973), leaving behind a scenic ice-carved landscape that was part of the impetus to declare the area a national park. The slightly lower Buenavista massif was apparently not glaciated (Has- tenrath 1963). Meteorological data from the Buenavista highlands (Table 1) show a mean annual temperature of 7.6°C and an annual rainfall total of about 2500 mm. Nearly 90% of the total precipitation falls during the May to November wet season. Afternoon clouds characteristi- cally enshroud the Talamancan highlands, and high atmospheric humidity moderates the dry season. But for weeks or months during the dry season the condensation belt lies below timberline, and the paramos experience clear, dry weather. Grasses, sedges, and some herbaceous dicots die back at this time, and ground litter dries out, providing the fuel for fires. The vegetation of the Chirripo and Buenavista paramos has been described by Weber (1959). Grasses, perennial herbs, and evergreen 1989] HORN: COSTA RICAN PARAMOS 95 Fig. 1. Map showing the location of study areas in the Cordillera de Talamanca, Costa Rica. Shading indicates the approximate present distribution of paramo vege- tation. Triangles in inset map are major peaks along the crest of the range. Based on field observations and the 1:50,000 topographic maps published by the Instituto Geografico Nacional. shrubs cover all areas above 3300 m, and extend as low as 3150 m in some areas. The montane forests that replace the paramo at its lower elevational limit are dominated by the oak Quercus costari- censis Liebm. The most characteristic paramo species is the bamboo Swallenochloa subtessellata (Janzen 1983), which often occurs in nearly monospecific stands. Intermixed with the bamboo grow sev- eral small-leaved, evergreen shrubs, with the families Hypericaceae, Compositae, and Ericaceae particularly well represented. Grasses, sedges, herbaceous dicots, and club mosses occur sparsely in the understory of the shrub layer, and in dense swards where the canopy is more open. Vegetation patterns in both areas reflect a long history of human disturbance. The presence of large cut stumps and charred snags of Escallonia poasana and Arctostaphylos arbutoides (Lindl.) Hemsl. on Cerro Zacatales and Cerro Asuncion indicates that small trees or large shrubs once grew higher on the slopes of the Buenavista massif Clearing and burning has enlarged the Buenavista paramo and may even have initially created it (Janzen 1973a, b, 1983). During 1985 I observed only a few cattle in the Buenavista paramo, but grazing 96 MADRONO [Vol. 36 Table 1. Climatic Data, Cerro Paramo Meteorological Station (3475 m), BuENA VISTA Highlands, Costa Rica. Temperature data are from the period 1971- 1979; precipitation data are from the period 1 97 1-1 984. Compiled from unpublished records on file in the Departamento de Estudios Basicos, Instituto Costaricense de Electricidad, San Jose, Costa Rica (station 073080). The station elevation is incor- rectly listed in the records as 3365 m. Temperature (in °C) Prppi HI ta ti on 1 XXX XXXXXX f Mean daily minimum Mean daily maximum Monthly mean Mean monthly total January 3.0 10.5 6.8 31.7 February 3.2 11.3 7.3 26.0 March 3.7 12.2 8.0 28.4 April 4.2 12.2 8.2 105.1 May 4.8 11.6 8.2 367.4 June 4.6 11.0 7.8 344.0 July 4.2 10.3 7.3 213.0 August 4.3 10.6 7.5 365.2 September 4.5 10.6 7.6 384.2 October 4.6 10.7 7.7 367.6 November 4.2 10.3 7.3 216.2 December 3.4 10.3 6.9 78.1 Annual mean 4.1 11.0 7.6 Annual total 2526.9 seems to have been more significant in the past, and some of the clearing and burning may have been carried out to encourage her- baceous forage. In the more isolated Chirripo massif the absence of trees is gen- erally regarded as the natural condition (Hartshorn 1983). Clearing and grazing (by horses) is minor and is restricted to a few heavily used campsites, but human-set fires occur frequently throughout the paramo. The charcoal stratigraphy of a 1 10-cm sediment core from a glacial lake revealed that fires due to human activity or lightning have affected the highland for over 4000 years (Horn 1989a). Methods Postfire vegetation recovery was monitored at four paramo bum sites in late 1984 and early 1985 using line and belt transects. The number and arrangement of transects varied depending on the size of the bum and available field time, but these variations do not affect the analysis. The slow rate of organic matter decay in the Talaman- can paramos was an asset to the study, because persistent fire-killed shmb and bamboo stems could be used to reconstmct the species composition and general stature of the prebum woody vegetation. Shrub and herb cover were measured separately along random transects using the line intercept method (Bauer 1943). The cover 1989] HORN: COSTA RICAN PARAMOS 97 estimates for the herb layer included all herbs, ferns, club mosses, true mosses, and the low ericaceous shrub Pernettia prostrata, which is generally only 10-20 cm high; the cover estimates for the shrub layer included all larger shrubs and bamboo. At the Conejos, Za- catales, and Sabila bum sites I used six 1 00-m transects to estimate shrub cover (canopy projection) and five (Conejos) or six 20-m tran- sects to estimate herb cover. At the smaller Tower 65 site I used twenty 20-m transects for shrub cover and six for herb cover. At the larger bum sites, data on shrub and bamboo density, fire response, and prefire and postfire plant stature were collected in 2-m-wide belt transects centered on the lines used for the cover analysis. The total area sampled was 400 m^ at the Zacatales site, 600 m^ at the Sabila site, and 1200 m^ at the Conejos site. Bamboos at the latter site were tallied in all transects but measured only in the first three. At the Zacatales site, resprouting plants were measured in additional belt transects covering 800 m^. The Tower 65 data were collected in five contiguous 5 x 20-m belt transects. Following methods adapted from Williamson et al. (1986), I clas- sified all living and dead shmbs and bamboo within the sample areas with prefire or postfire heights of at least 40 cm into one of four fire response categories: 1) "dead," for plants that had been killed by the fire; 2) "resprouter," for plants that had suffered crown loss but had subsequently resprouted; 3) "postfire colonist," for plants that showed no evidence of having bumed in the fire and that presumably had become established after the fire occurred; and 4) "fire survivor," for plants that had survived the fire with minimal crown loss. Plants in the latter category showed evidence of scorching, which provided a means to separate them from postfire colonists. Dead plants could be identified to species based on stem architecture and the color and texture of their bark and wood. Dead plants killed in earlier fires were distinguished based on bark weathering and were tallied sep- arately. Dead plants not killed by buming were seen rarely; these were also tallied separately. For the woody dicots, each well-defined cluster of stems was re- garded as a separate individual, except in cases where root connec- tions were clearly present. For the clump-forming dwarf bamboo Swallenochloa subtessellata, each distinct clump was counted as one plant. Closely spaced clumps were counted as different plants if they were separated by at least 75 cm of ground that was devoid of dead or live culms. In practice, these criteria probably overestimated the number of separate plants, since underground stems and root sys- tems can extend for several meters. However, without excavating every plant it was impossible to be certain whether such underground connections existed, so decisions as to what constituted an individual plant had to rest on the spacing of aboveground stems. The highest leaf or stem of all plants in the "postfire colonist," 98 MADRONO [Vol. 36 "resprouter," and "fire survivor" classes was measured to the nearest cm, and the prefire height of all "dead" and "resprouter" plants was estimated by measuring the highest dead stem. The heights of broken stems were recorded separately and excluded from calculations. The number of dead and/or live stems at the base of each shrub and the basal diameter of the largest of each type of stem were recorded. Where present, older dead stems were measured and tallied sepa- rately. For the bamboos, live culms were measured and classed by abundance (<50, 50-100, >100). Published and unpublished reports (cited below) and field evi- dence provided dates for some of the fires; age estimates for other bums and for earlier fires at all bum sites were made by examining growth rings in stems of resprouting Vaccinium consanguineum shrubs. Ring counts in stems regenerating after fires of known age suggest that this plant produces annual rings under the seasonal precipitation regime that characterizes the Talamancan highlands. Because not all stems resprout in the first year after buming, the ring counts indicate minimum fire recurrence intervals. Voucher specimens were deposited at UC, WIS, CR, and lA (grasses only). Site Descriptions Tower 65 site. The Tower 65 bum site was located at an elevation of 3310 m on the crest of a ridge extending northeastward from Cerro Buenavista. About a year before sampling a small fire had bumed roughly one- tenth hectare of vegetation located just to the east of Tower 65 of the new Cartago-San Isidro electrical transmis- sion line. The bum area extends to the power lane that was cleared in 1983, and shmb stumps in the lane are charred, indicating that the fire postdated the cutting of the shrubs. It probably resulted from accidental or intentional ignition of the vegetation when the trans- mission tower and lines were installed in early 1984. Growth rings in stems of fire-killed Vaccinium consanguineum suggest that the vegetation was at least 1 6 years old at the time the fire occurred. Conejos site. This bum site was located within the broad basin at the head of the glaciated valley known as the Valle de los Conejos in the Chirripo highlands (Fig. 2). The site lies on a south-facing slope between 3480 and 3500 m. The area last bumed during the >5000 ha fire that swept through the highlands in March 1976 (Chaverri et al. 1976). At the time of the surveys the regenerating vegetation was nine years old. According to Weston (quoted in Koh- kemper 1976), the Valle de los Conejos had previously bumed in 1961, a date consistent with the maximum of fifteen rings found on fire-killed stems of Vaccinium consanguineum. Zacatales site. The Zacatales site was located on the steep south- facing slope of Cerro Zacatales in the Buenavista highlands. Tran- 1989] HORN: COSTA RICAN PARAMOS 99 Fig. 2. General view of the head of the glaciated Valle de los Conejos, nine years after the major 1976 fire. The Conejos study site is visible in the background on the left side of the photograph. The peak in the center is Cerro Piramide (3807 m). The dominant woody species is the dwarf bamboo Swallenochloa subtessellata. sects were run between 3340 and 3370 m, within the same area sampled by WilHamson et al. (1986) in 1982. Site visits showed that the fire, which appears to have covered about 10 ha, occurred be- tween early February and late March of 1973 (Williamson et al. 1986). When surveyed in late 1984 and early 1985 the vegetation was about twelve years old. Fire-killed stems in two stages of decay were common at the site, indicating that the area had burned twice in recent decades. The more intact dead stems of Vaccinium con- sanguineum showed a maximum of twelve growth rings, suggesting that at the time of the 1973 fire the vegetation was at least this old. Older dead stems that had been killed by the next to last fire were too decayed to make reliable ring counts. Sdbila site. The Sabila site was located a few kilometers to the east of the Zacatales site on the southeastern face of Cerro Sabila. Elevations within the area sampled range from 3370 to 3410 m. Ring counts made in 1985 on regenerating Vaccinium consanguin- eum stems revealed a maximum of twelve rings, suggesting that the 15 ha fire occurred no later than 1973. The exact date of burning is 100 MADRONO [Vol. 36 not known. I counted 29 growth rings on the largest dead stem of one regenerating V. consanguineum, suggesting that the vegetation was at least this old at the time of the last fire. Results Postflre cover. Table 2 shows cover data for the shrub and herb layers at the four paramo bum sites. The data are grouped together only to facilitate comparison; the arrangement is not meant to imply a successional sequence. Cover ranged from minimal at the Tower 65 site, where most of the ground was still bare one year after burning (Fig. 3), to nearly continuous after twelve years of regeneration at the Zacatales site. Patches of bare ground ranging in size from 0. 1 to 0.5 m^ were present at both the Conejos and Sabila sites, even in areas not shaded by the shrub canopy. Standing and downed fire- killed shrub and bamboo stems were conspicuous at all four sites (Fig. 4). Fire response. Burning was complete or nearly so within all bum sites, and most plants had suffered total crown loss. The overall incidence of basal resprouting by the larger shmbs and bamboo ranged from 40% at the Zacatales site to 83% at the Conejos site, with differences largely attributable to variations in prebum species composition. Chi-square tests indicated significant heterogeneity in fire re- sponses (Table 3). The bamboo Swallenochloa subtessellata showed the highest resprout rate, with live shoots present within 99-100% of all bumed clumps examined. Also exhibiting vigorous resprouting following buming were the common ericaceous shrubs Vaccinium consanguineum (90-98% resprout rate) and Pernettia coriacea (93- 96% resprout rate). The shrub Hypericum irazuense, in contrast, suffered high mortality at the four sites; only 4-14% of bumed in- dividuals had suckered following buming. Also exhibiting low re- sprout success were the shmbs Hypericum strictum and Senecio firmipes (0 and 4%, respectively, with data only from the Tower 65 site). Low to intermediate frequencies of basal resprouting were shown by the shmbs Rapanea pittieri (1 5-25%) and Escallonia poa- sana (50-57%). The association between prefire plant stature and fire response was tested using a median test (Sachs 1984). For each species, ob- servations on prefire height (and, later, on prefire stem diameter) of both dead and resprouting plants were pooled together and ordered to determine the common median. The observations were then sort- ed according to whether they were larger or smaller than the common median, and a Fisher exact test (Sokal and Rohlf 1981) was applied to the resulting two-way table to test the null hypothesis that whether 1989] HORN: COSTA RICAN PARAMOS 101 I •e o 3 o OS (L> I/) ^ 2 is c \^ O O o o a «5 .a • 1-1 "O ^ «^ ex c ^ o (U o cd o 00 W On H ^ 5 o <; o 06 cd o cd O U O T3 < O > n m rN ^ V « 00 ■"^ V V U CJ K Co Co O 1^ Cd C/2 I N C . 5 ^' Co s; CO ^ O o X 'g I N D ^ CX 'hJ V ^ Cd o o ^ o cx i3 > o G '•5 (U > o CJ "cd ■*-> O H o 00 o ^ 00 >n in (N G G d G G Y V V o^ o4 in vo r- vo o CX jd > o 00 c ''3 X (U i-T > o o ■t-> o H tt] ^ ^ Qc; Co u 102 MADRONO [Vol. 36 Fig. 3. View within the Tower 65 bum site one year after the fire. Large dead shrub in the center of the photograph is Hypericum irazuense; regenerating clumps of the bamboo Swallenochloa subtessellata and resprouting shrubs of Vaccinium consan- guineum and Pernettia coriacea surround the dead plant. The herb in the right fore- ground is Senecio oerstedianus. Note the abundance of standing dead wood and the extent of bare ground. 1989] HORN: COSTA RICAN PARAMOS 103 a shrub's prefire height or stem diameter fell above or below the median was unrelated to its survival. The median tests revealed little association between fire response and either prefire height or stem diameter for the most common shrubs. At the Tower 65 site there were more small-diameter Ra- panea pittieri in the "dead" category and more large-diameter R. pittieri in the "resprouter" category than would be expected under the null hypothesis of no association (p = 0.026), suggesting that larger diameter individuals were more likely to resprout than smaller diameter individuals. No association was apparent between prefire height and fire response. At the Sabila site, taller stems were asso- ciated with higher survival for Hypericum irazuense (p = 0.03), but the large number of broken stems that had to be excluded from calculations makes the finding suspect. Prefire stem diameter and fire response were not significantly associated. At the Sabila site the fire response of Escallonia poasana was significantly associated with both prefire height and prefire stem diameter. Larger plants were more often in the "dead" category and smaller plants were more often in the "resprouter" category than would be expected if no association existed (p = 0.05 for stem height, p = 0.002 for stem diameter), suggesting that smaller plants were more likely to resprout after burning than larger plants. Few postfire shrub colonists other than Hypericum irazuense were observed at the sites. Abundant H. irazuense recruitment (all from seed) was observed at the Zacatales and Sabila sites (Fig. 5), but had not occurred at the younger bum sites. Seedling establishment by associated woody species was rarely observed. The common erica- ceous shrubs Vaccinium consanguineum and Pernettia coriacea, and the bamboo Swallenochloa subtessellata produce widely diverging roots and rhizomes, and most postfire colonists probably represented sprouts produced at new points along pre-existing root/rhizome net- works rather than new genetic individuals established from seed. Changes in species composition. The varying rates of shrub and bamboo mortality and seedling establishment at the four bum sites led to changes in species composition (Fig. 6). Most notable was the marked reduction in the density of Hypericum irazuense at the Tow- er 65 and Conejos sites. The narrow-leaved congener Hypericum strictum also showed a marked decline in density at the Tower 65 site, where no resprouting individuals and no seedlings were ob- served one year after buming. Postfire growth rates. Table 4 gives prefire and postfire heights for the bamboo Swallenochloa subtessellata and its five most common shrub associates. Among these species Swallenochloa consistently showed the highest postfire growth rate. Prior to the fires the bamboo 104 MADRONO [Vol. 36 averaged between one and two meters in height, with the tallest plants at the Tower 65 site and the smallest at the Conejos site. Height recovery was 18% one year after burning at the Tower 65 site; 98% nine years after burning at the Conejos site; 113% twelve years after burning at the Zacatales site; and 1 29% twelve or more years after burning at the Sabila site. At the Tower 65 site, the tender new shoots of the bamboo had been grazed by native rabbits (Sil- vilagus sp.) and by at least one cow, such that the data in Table 4 may underestimate the actual first year growth increment of the plant. Little or no evidence of grazing was apparent at the older bum sites, but may have occurred in the past. Grazing damage on plants other than Swallenochloa subtessellata was rarely observed. Prior to burning, the ericaceous shrub Vaccinium consanguineum averaged around a meter in height. Percentage height recovery was 26% at the Tower 65 site, 71% at the Conejos site, 87% at the Zacatales site, and 88% at the Sabila site. The more prostrate eri- caceous shrub Pernettia coriacea showed percentage height recovery of 24% at the Tower 65 site, 97% at the Zacatales site, and 79% at the Sabila site. 1989] HORN: COSTA RICAN PARAMOS 105 Table 3. Fire Responses of Woody Species at Paramo Burn Sites. Table includes only species with a sample size greater than 20 at individual sites. Chi-square statistic indicates heterogeneity of fire responses. Abbreviations in brackets are used for the taxa in Figure 6. Tower 65 site Response category Fire Species Dead Resprouter survivor Swallenochloa subtessellata [Ss] 0 68 3 Vaccinium consanguineum [Vc] 2 113 10 Pemettia coriacea [Pc] 7 88 15 Hypericum irazuense [Hi] 100 4 15 Hypericum strictum Kunth [Hs] 51 0 11 Senecio firmipes Greenman [Sf ] 22 1 0 Rapanea pittieri Mez [Rp] 12 4 5 = 419, df = 12, p < 0.001. Conejos site Response category Postfire Species Dead Resprouter colonist Swallenochloa subtessellata [Ss] 5 386 4 Vaccinium consanguineum [Vc] 2 19 5 Hypericum irazuense [Hi] 80 5 1 = 463, df ^ = 4,p < 0.001. Zacatales site Response category Fire Postfire Species Dead Resprouter survivor colonist Vaccinium consanguineum [Vc] 2 48 10 15 Pemettia coriacea \?c\ 3 72 14 22 Hypericum irazuense [Hi] 240 34 43 434 x' = 429, df = 6, p < 0.001. Sabila site Response category Fire Postfire Species Dead Resprouter survivor colonist Swallenochloa subtessellata [Ss] 0 41 0 8 Vaccinium consanguineum [Vc] 1 65 2 6 Pemettia coriacea [Pc] 5 100 4 15 Hypericum irazuense [Hi] 147 23 13 98 Rapanea pittieri [Rp] 17 3 1 2 Escallonia poasana [Ep] 9 12 0 1 X^ statistic cannot be calculated due to low cell counts. 106 MADRONO [Vol. 36 Fig. 5. Seedlings of Hypericum irazuense growing up amidst shrubs of the same species killed by fire twelve years earlier at the Zacatales site. Cerro Zacatales (3399 m) is visible in the background. The stick is one meter long. The shrub Hypericum irazuense averaged 90-160 cm tall at the time of the four fires investigated; percentage height recovery for the rare individuals that resprouted was 15% at the Tower 65 site, 64% at the Conejos site, 71% at the Zacatales site, and 70% at the Sabila site. Hypericum irazuense seedlings at the latter two sites were 1989] HORN: COSTA RICAN PARAMOS 107 ZACATALES SPECIES SPECIES SPECIES SPECIES Fig. 6. Prefire (first bar) and postfire (second bar) density of common shrubs and bamboo at the bum sites. See Table 3 for key to species abbreviations. Note change of scale for the Zacatales data. about 10% smaller on average than resprouts. Height recovery by resprouting shrubs of Rapanea pittieri and Escallonia poasana at the older bum sites was faster on both an absolute and percentage basis; both plants had more than regained their prefire stature at the Sabila site and Escallonia had also exceeded its average prefire size at the Zacatales site. Discussion The slow rates of postfire vegetation recovery and litter breakdown documented at the bum sites are consistent with observations made by Janzen (1973a) following a 1969 fire on Cerro Asuncion. Three years after the Asuncion fire, Janzen found that there were still large areas of uncolonized ground, and that most of the shmb stems killed by the fire were still intact and standing. Surveys at the Conejos, Zacatales, and Sabila sites revealed that bare patches of ground and upright fire-killed stems can persist for nine or more years following buming. The slow rates of growth and colonization that characterize these tropical high montane environments stand in marked contrast to the usual situation in the lowland tropics, where fire-killed wood decomposes quickly, and openings are rapidly colonized by dense stands of sucker sprouts and the seedlings of fast growing, weedy secondary species (Janzen 1973a). Pattems of postfire vegetation development in the Talamancan paramos support the initial floristics model of succession (Egler 1 954). Following disturbance, most paramo shrubs and bamboo replace themselves directly, usually by resprouting from protected buds. Woody perennials comprising the mature vegetation are present in 108 MADRONO [Vol. 36 Table 4. Prefire and Postfire Heights of Resprouting Shrubs and Bamboo at Paramo Burn Sites. Values listed are means, (sample sizes), and standard deviations. Sample sizes for prefire heights are smaller than those for postfire heights because of the exclusion of broken dead stems from the calculations. This table shows only data for the six most common woody species. For data on prefire stature of dead plants, and postfire stature of fire-surviving plants and postfire colonists see Horn (1986). Height (cm) at Tower 65 site Height (cm) at Conejos site 1 year after burning 9 years after burning Species Prefire Postfire Prefire Postfire Swallenochloa 186.2 (65) 34.4 (68) 104.6 (100) 102.8(91) subtessellata SD 54.4 SD 17.4 SD 39.3 SD 45.8 Vaccinium 83.5 (103) 21.7 (103) 107.1 (17) 75.6(18) consanguineum SD 37.5 SD 11.5 SD 51.9 SD 25.8 Pernettia coriacea 67.4 (87) 16.3 (88) SD 24.6 SD 5.3 Hypericum irazuense 129.8 (4) 20.0 (4) 143.0 (4) 91.0 (4) SD 60.1 SD 18.7 SD 43.8 SD 44.9 Rapanea pittieri 131.3(4) 19.0 (4) SD 60.2 SD 18.7 Escallonia poasana 84(1) 14(1) Height (cm) at Zacatales site Height (cm) at Sabila site 1 2 years after burning > 1 2 years after burning Species Prefire Postfire Prefire Postfire Swallenochloa 124.9(11) 140.9 (14) 147.8 (28) 190.6 (41) subtessellata SD 51.6 SD 64.8 SD 47.0 SD 52.8 Vaccinium 76.7 (157) 66.7 (167) 116.6 (52) 103.0 (65) consanguineum SD 34.7 SD 20.7 SD 46.4 SD 35.3 Perettia coriacea 46.9 (104) 45.7 (134) 82.4 (52) 65.0(100) SD 12.7 SD 11.4 SD 31.8 SD 25.4 Hypericum irazuense 91.1 (28) 65.1 (78) 159.4 (7) 110.9(23) SD 29.9 SD 23.2 SD 29.9 SD 46.7 Rapanea pittieri 148.5 (2) 116.5 (2) 137.0 (2) 172.0 (3) SD 2.1 SD 19.1 SD 5.7 SD 61.2 Escallonia poasana 83.2 (21) 96.3 (30) 86.7 (6) 148.3(12) SD 34.1 SD 39.4 SD 40.5 SD 43.6 the bum site the first year after fire, and Hmited colonization by other species takes place once the first year population is established. Basal resprouting by shrubs and bamboo. The fire response data confirm and extend the results of Williamson et al. (1986), who reported that over 90% of the ericaceous shrubs on Cerro Zacatales had resprouted following burning, as compared to only 1 1% of Hy- pericum irazuense shrubs. Patterns described here differ markedly, however, from the trends evident following the 1969 Asuncion fire studied by Janzen (1973a). Postfire vegetation development at Cerro 1989] HORN: COSTA RICAN PARAMOS 109 Asuncion was characterized by abundant regeneration of Hypericum irazuense (Janzen's Hypericum caracasanum) following crown loss. Williamson et al. (1986) contrasted the fire response of H. irazuense at the Asuncion and Zacatales sites, and suggested that the higher mortality at Cerro Zacatales was due to the depletion of root reserves by successive fires. While there is evidence of more frequent burning at the Zacatales site, the discovery of high Hypericum irazuense mortality on Cerro Sabila, where recent fire recurrence intervals appear similar to those at Janzen's Asuncion site, suggests that other environmental conditions, such as perhaps soil moisture conditions at the time of the fire, may explain the much higher resprout success at the Asuncion site. The near elimination of the bamboo Swallenochloa subtessellata from the Zacatales site was cited by Williamson et al. (1986) as a second consequence of repeated burning. This interpretation con- tradicts my data from the site, which showed little evidence of Swal- lenochloa mortality, and is at variance with the results from the other bum sites, where vigorous postfire sprouting by the bamboo has allowed it to maintain or even increase its dominance. I doubt that the recent (post- 1950) fire recurrence interval at the site (esti- mated by Williamson and associates to be on the order of ten years) has been too short for between-fire recovery of root reserves, because at higher elevations in the Chirripo massif bamboo clumps that burned in 1 976 and again in 1 982 have produced abundant resprouts (Horn unpubl. data). If Swallenochloa was once more important at the Zacatales site, factors other than fire were probably responsible for its decline. Gill (1981) has discussed the possible importance of the prefire age of a plant in determining its regenerative capacity following crown loss. Although higher survival might be expected among larger (and presumably older) plants with well-developed root reserves, median tests revealed no consistent associations between the size of paramo shrubs and their resprout success. For the shrub Escallonia poasana, susceptibility to fire seems to increase with size. Higher fire mortality among older shrubs has been reported by Stohlgren (1985) for Adenostoma fasciculatum H. & A. in the California chap- arral. Stohlgren suggested that the older shrubs may contain more dead wood from previous fires, and that this might have made them more likely to be killed by fire. This explanation would not apply at the Sabila site, however, because the large individuals of Escal- lonia poasana killed by the fire showed no evidence of having burned previously. Growth and recovery of resprouting shrubs and bamboo. The data on shrub and bamboo heights summarized in Table 4 and by Janzen (1973a) and Williamson et al. (1986) suggest that the growth of 110 MADRONO [Vol. 36 woody perennials is most rapid during the first few years after burn- ing. The higher initial growth rates may in part reflect improved soil fertility following the release of nutrients by burning. The consistently greater heights of shoots of Swallenochloa sub- tessellata as compared to those of the common woody dicots confirm Janzen's (1983) statement that the bamboo exhibits one of the fastest rates of regrowth in paramo vegetation. It is worth stressing, how- ever, that on an absolute basis the growth of the bamboo is quite slow; the field data support Janzen's (1983) observation that about 8-10 years are required for burned plants to regain their prefire adult statures. Associated woody dicots may require a decade or more to reach their prefire adult size of 1-2 m. Measurements of stem di- ameters (reported in Horn 1986) indicate that for all species except Escallonia poasana, the recovery of basal stem diameter lags far behind height recovery. Paramo regeneration, while slow in comparison to regeneration in lowland tropical forests, proceeds at rates comparable to those measured in tropical montane forests in Costa Rica and Puerto Rico. Ewel (1980) artificially cleared montane rain forest from eight plots near Ojo de Agua (3000 m) in the Cordillera de Talamanca to mon- itor regrowth; after one year the average height of the three tallest plants was only 70 cm. One year after burning at the Tower 65 site the three tallest plants (all clumps of Swallenochloa subtessellata) averaged 85 cm in height. Byer and Weaver (1977) noted similarly slow rates of growth on artificial clearings within elfin woodland in the Luquillo mountains of Puerto Rico. One year after clearing, the mean maximum heights for woody sprouts in three clearings was 33.6 cm, a value quite close to the mean maximum height of 34.4 cm for Swallenochloa subtessellata at Tower 65. Postfire regeneration at the Tower 65 site occurred more slowly than it did following a fire in a montane mire (elev. 2690 m) near Tres de Junio in the Cordillera de Talamanca (Horn 1989b). In one year of growth the fastest growing shrub in the mire, Hesperomeles heterophylla, had produced resprouts averaging 68 cm in height, or about twice the mean height of the highest culms of Swallenochloa subtessellata at Tower 65. Shrub and herb colonization. The long persistence of bare patches of ground within the study bums reflects both the slow production and growth of new shoots from surviving rootstocks and an ex- tremely low rate of seedling colonization. The observed paucity of herb or shrub seedlings one year after the Tower 65 fire, and their rarity three years after the Asuncion fire (Janzen 1973a), suggests that soil seed pools are small or that seeds suffer high mortality from fire. Unlike fire-prone shrub communities in the mediterranean- climate regions, the Talamancan paramos do not harbor a dormant 1989] HORN: COSTA RICAN PARAMOS 111 seed bank of herbaceous species that proHferate following burning. The paramo shrubs and bamboo similarly show little reliance on either in-soil seed storage or fire-stimulated germination. Instead, postfire seedling colonization by woody perennials depends on seed production by plants that survived the fire, or on the influx of seeds from plants in surrounding, unbumed areas. In some cases, seedling colonization may not occur until plants that regenerated after the fire from basal buds reach maturity and begin seed production within the bum site. The dominance of tussock grasses and sedges in the postbum herbaceous cover may be due primarily to the ability of these plants to survive fires, rather than to an ability to recolonize burned areas as seedlings. The perennating buds of these graminoids are generally protected from fire by their position underground or within persis- tent leaf bases, and many of the plants are able to resprout rapidly after burning. Chaverri et al. (1976) noted that burned sedges within the Chirripo paramo had produced sprouts up to 15 cm long within one month of the 1976 fire. Among herbaceous and semi-woody dicots, Castilleja talamancensis, Rubus eriocarpus, and Acaena cy- lindrostachya also resprout vigorously following burning, and their importance in the herbaceous cover at recent bum sites is probably due as much to fire survival as to postfire seedling establishment. Among the woody perennials at the bum sites only Hypericum irazuense had established abundant seedlings. Twelve years after the last fire at the Zacatales site, seedling recmitment had more than compensated for the heavy fire-induced mortality of this species. Most of the seedlings were closely spaced, however, and natural thinning may reduce their density. Williamson et al. (1986) tallied few seedlings over 40 cm high in their 1982 survey at the site; they reported that most of the Hypericum recruits seen were only 10-25 cm high. That most of the seedlings were at least twice this high in 1985 suggests that the plants observed by Williamson and associates had only recently become established, and hence that the biggest pulse of seedling establishment did not occur until several years after the 1973 fire. Substantial Hypericum irazuense recruitment was also apparent at the Sabila bum site, but had not occurred at the Conejos site. This presumably reflects the very large size of the 1976 Chirripo fire and the shortage of flowering plants to reseed the bum site. If re- population of bumed areas by H. irazuense depends primarily on seed influx from adjacent, unbumed vegetation, the rate of seedling colonization should be lower the larger the fire and the more com- plete the bum. The near complete reliance on vegetative regeneration by the com- mon woody perennials in the Costa Rican paramos means that openings created by fire are slow to be colonized. Even the prolific 112 MADRONO [Vol. 36 resprouter Swallenochloa subtessellata seems incapable of rapid spread. Nearly all of the bamboo clumps in the bum sites contained charred culms, indicating that they had been present prior to the fires. Some of the clumps had enlarged following burning by sprout- ing from lateral positions within the clumps. A few plants had be- come established after the fires, probably through the production of shoots at more distant points along extensive, preexisting networks of rhizomes. But even at the older sites, the bamboo had not yet spread into many of the gaps created by burning. The relatively poor colonizing ability of Swallenochloa subtessel- lata is due largely to the fact that the species rarely sets seed. Although it is always possible to find some plants in flower, the grass specialist Richard Pohl (1980) reported that he has never detected filled cary- opses in herbarium specimens, nor observed seedlings in the field. Janzen (1983) has suggested that this may reflect a pollen shortage among plants flowering out of phase with the rest of a gregariously flowering population. Low pollen viability may also contribute to the low seed set. Pollen grains in anthers taken from two specimens in the University of California, Berkeley, herbarium (UC 1434192, UC Ml 11383) were misshapen and did not stain in lactophenol cotton blue, indicating that at the time the plants were pressed the pollen grains contained no protoplasm. The possibility that existing populations of Swallenochloa subtessellata in the Costa Rican para- mos are male sterile deserves further attention. Regenerative strategies and postfire vegetation change. The repro- ductive modes of the common woody perennials in the Costa Rican paramos resemble the regenerative strategies identified by Zedler et al. (1983) in fire-prone chaparral shrublands in California. The bam- boo Swallenochloa subtessellata and the ericaceous shrubs Vaccin- ium consanguineum and Pernettia coriacea exemplify the "sprouter- nonseeder" mode: most burned plants survive the fire by resprouting from the base, but few or no seedlings are produced. The shrub Hypericum irazuense follows the "sprouter-seeder" pattern: the shrub resprouts infrequently, but supplements vegetative regeneration with more abundant seedling recruitment. However, the stored seed re- serves and fire-stimulated germination that makes this strategy so eflective in chaparral communities seem to be lacking in the para- mos. Insufficient data are available to characterize the reproductive behavior of the less common paramo shrubs, but most appear to rely heavily on vegetative reproduction. The shrub Hypericum stric- tum suffered complete mortality at the only site where it was com- mon (Tower 65), and may be an obligate seeder. The data from the four paramo bum sites indicate that the more fire-resistant sprouters are clearly favored in the first 5-10 years after buming. The apparent absence of in-soil seed storage puts the fire- 1989] HORN: COSTA RICAN PARAMOS 113 sensitive species Hypericum irazuense at a distinct disadvantage during the first several postfire years. Delayed seedling recruitment may in time compensate for heavy fire-induced mortality, but a fire- fi-ee interval of at least two decades will likely be required for recruits to obtain statures comparable to those of the fire-killed plants they have replaced. Acknowledgments I thank Roger Horn for field assistance, and Daniel Janzen, Jon Keeley, and Richard Corlett for critical reviews of the manuscript. Roger Byrne, Herbert Baker, Theodore Oberlander, and James Parsons provided helpful comments on an earlier version of this paper. Funding was provided by the Institute of International Education, the Association of American Geographers, and the Center for Latin American Studies of the University of California, Berkeley. Literature Cited Bauer, H. L. 1 943. The statistical analysis of chaparral and other plant communities by means of transect samples. Ecology 21:45-60. Byer, M. D. and P. L. Weaver. 1977. Early secondary succession in an elfin woodland in the Luquillo mountains of Puerto Rico. Biotropica 9:35-47. Chaverri, a., C. Vaughan, and L. J. Poveda. 1976. Informe de la gira efectuada al macizo de Chirripo al raiz del fuego ocurrido en Marzo de 1976. Re vista de Costa Rica 11:243-279. CuATRECASAS, J. 1979. Comparaciou fitogeografica de paramos entre varias Cor- dilleras. Pp. 89-99 in M. L. Salgado-Labouriau (ed.). El medio ambiente paramo. Ediciones Centro de Estudios Avanzados, Merida, Venezuela. Egler, F. 1954. Vegetation science concepts. I. Initial floristic composition, a factor in old-field vegetation development. Vegetatio 14:412-417. EwEL, J. 1 980. Tropical succession: manifold routes to maturity. Tropical Succession (Supplement to Biotropica 12): 2-7. Gill, A. M. 1981. Fire adaptive traits of vascular plants. Pp. 208-230 in H. A. Mooney et al. (tech. coords.). Fire regimes and ecosystem properties. U.S. Dept. of Agriculture Forest Service General Technical Report WO-26. Grubb, p. J. 1970. The impact of man on the paramo of Cerro Antisana, Ecuador. J. Applied Ecol. 7(2):7P-8P. Hartshorn, G. S. 1983. Plants: introduction. Pp. 1 18-157 in D. H. Janzen (ed.), Costa Rican natural history. Univ. Chicago Press, Chicago. Hastenrath, S. 1963. Zumeinflub der massenerhebung auf den verlauf der klima- und vegetationstufen in Mittelamerika und im sudlichen Mexico. Geogr. Ann. 45:76-83. . 1973. On the Pleistocene glaciation of the Cordillera de Talamanca, Costa Rica. Z. Gletscherkd. 9:105-121. Horn, S. P. 1986. Fire and paramo vegetation in the Cordillera de Talamanca, Costa Rica. Ph.D. dissertation. Univ. California, Berkeley. . 1 989a. Prehistoric fires in the Chirripo highlands of Costa Rica: sedimentary charcoal evidence. Revista de Biologia Tropical (in press). . 1 989b. Effect of burning on a montane mire in the Cordillera de Talamanca, Costa Rica. Brenesia (in press). Janzen, D. H. 1973a. Rate of regeneration after a tropical high elevation fire. Biotropica 5:117-122. . 1973b. Sweep samples of tropical foliage insects: descriptions of study sites, with data on species abundances and size distributions. Ecology 54:659-678. 114 MADRONO [Vol. 36 , 1 983. Swallenochloa subtessellata (chusquea, batamba, matamba). Pp. 330- 331 in D. H. Janzen (ed.), Costa Rican natural history. Univ. Chicago Press, Chicago. KoHKEMPER M., M. 1976. Sexta expedicion al macizo del Chirripo. Instituto Geo- grafico Nacional de Costa Rica. Informe Semestral Enero-Junio:63-69. Lauer, W. 1981. Ecoclimatological conditions of the paramo belt in the tropical high mountains. Mountain Res. Develop. 1:209-221. Otarola D., C. E. 1976. Caracterizacion y clasificacion de algunos suelos de la Cordillera de Talamanca. Thesis. Universidad de Costa Rica. San Jose, Costa Rica. PoHL, R. 1980. Rora costaricensis: family number 15, Gramineae. Fieldiana, Bot., n.s. 4:1-608. Sachs, L. 1984. Applied statistics: a handbook of techniques, 2nd ed. Springer- Verlag, New York. Smith, A. P. 1981. Growth and population dynamics of Espletia (Compositae) in the Venezuelan Andes. Smithsonian Contrib. Bot. 48:1-45. Sokal, R. R. and F. J. Rohlf. 1981. Biometry, 2nd ed. W. H. Freeman and Company, San Francisco. Stohlgren, T. J. 1985. Fire-caused mortality in chamise chaparral. Pp. 385-387 in J. E. Lotan et al. (tech. coords.). Proceedings— symposium and workshop on wilderness fire. U.S. Dept. of Agriculture Forest Service General Technical Report INT- 182. Weber, H. 1959. Los paramos de Costa Rica y su concatenacion fitogeografico con los andes suramericanos. Instituto Geografico Nacional, San Jose, Costa Rica. Weyl, R. 1957. Contribucion a la geologia de la Cordillera de Talamanca de Costa Rica (Centro America). Instituto Geografico Nacional, San Jose, Costa Rica. Williamson, G. B., G. E. Schatz, A. Alvarado, C. S. Redhead, A. C. Stam, and R. W. Sterner. 1986. Effects of repeated fires on tropical paramo vegetation. Trop. Ecol. 27:62-69. World Meteorological Organization. 1953. World distribution of thunder- storm days. I. Tables. World Meteorological Organization, Geneva. . 1956. World distribution of thunderstorm days. II. Tables of marine data and world maps. World Meteorological Organization, Geneva. Zedler, p. H., C. R. Gautier, and G. S. McMaster. 1983. Vegetation response to extreme events: the effect of a short interval between fires in California chap- arral and coastal scrub. Ecology 64:809-818. (Received 25 Aug 1988; revision accepted 24 Jan 1989.) A NEW SPECIES OF ERIGERON (ASTERACEAE: ASTEREAE) FROM CENTRAL NEW MEXICO Richard Spellenberg Department of Biology, New Mexico State University, Las Cruces, NM 88003-0001 Paul Knight State Forestry, Energy, Minerals, and Natural Resources Department, Villagra Building, Santa Fe, NM 87503 Abstract Erigeron acomanus is an inhabitant of sandy slopes beneath sandstone cHffs in central New Mexico. It is morphologically similar to E. tener of the Intermountain Region of the western United States. Erigeron acomanus has 1 6-30 white rays, has 4-10 leaves on the flowering stem, and forms leafy mats 10-70 cm in diameter. Pappus bristles are 20-25 in number and are about the length of the disk corolla. Erigeron acomanus lacks the peg-like glandular hairs present on the stems of E. tener. Since the passage of the Endangered Species Act of 1973, each of us has been involved in numerous surveys to determine the presence of endangered or threatened plant species on lands in New Mexico administered by federal land management agencies. On two such surveys, at separate locations, we independently discovered an Erig- eron in central New Mexico that could not be identified in local keys (e.g., Martin and Hutchins 1981) or Cronquist's (1947) revision of the genus. Morphologically it most closely resembles E. tener (A. Gray) A. Gray, a widespread species from the Intermountain Region, but E. acomanus is readily recognized by several morphological features. It is also disjunct from E. tener by more than 500 km (Fig. 1). We describe it here as a new species. Erigeron acomanus Spellenberg and Knight, sp. nov. (Fig. 2).— Type: USA, New Mexico, McKinley Co., ca. 3.2 km N of Prewitt at base of cliffs in canyon, T14N R12W sect. 24 NW V4 of NE 1/4, 35°26'03"N, 108°03',29"W, elev. 2120 m, 14 Jul 1983, Knight 2689 (holotype, UC; isotypes; ARIZ, ASU, COLO, NMC, NY, RM, TEX, UNM, US). Plantae perennes rosulis foliaceis, tegetes 10-70 mm diametrum formantes. Laminae foliorum oblanceolatae vel spathulatae, 8-23 mm longae, 2-7 mm latae, pubescentiae modice in superficiebus ambabus, laminae ad basim in petiola sensim attenuatae. Caules Madrono, Vol. 36, No. 2, pp. 115-121, 1989 116 MADRONO [Vol. 36 Fig. 1 . Western United States showing distribution of Erigewn acomanus, E. tener, E. hessii, and E. scopulinus. 4.5-15 cm alti, 4-10 folia et 1 capitulum ferentes. Capitula 8-10 mm diametro, ca. 5 mm alta. Ligulae 16-30, albae, 4.5-9 mm longae, 1.3-1.8 mm latae. Setae pappi 20-25, 1.5-2.5 mm longae. Chro- mosomatum gametophytorum numerus 9. Taprooted perennial forming a mat 10-70 cm in diameter, the caudex branches covered by persistent leaf bases. Leaves mostly basal, spreading or ascending, 8-30 in rosettes at ends of caudex branches; blades oblanceolate to narrowly obovate or spatulate, 8- 23 mm long, 2-7 mm wide, round or obtuse at the tip, evenly tapered to a petiole-like base 3-13 mm long, moderately puberulent on both surfaces with fine, appressed, slightly wavy hairs ca. 0.3 mm long, the leaves on the flowering stem similar but becoming progressively smaller toward the capitulum. Flowering stems erect, 4.5-15 cm tall, bearing 4-10 leaves. Capitula solitary, pendulous in bud, erect in flower and fruit. Flowering stem strigose, usually with a single bract- like leaf about the length of the phyllaries immediately beneath the involucre, this separate from the next leaf on the stem by a space of 2-15 mm. Involucre ca. 5 mm high, 7-10 mm across (when pressed). Phyllaries 25-38, lanceolate, 2.5-4 mm long, 0.5-0.8 mm wide, purplish especially on the margins, greenish on the back, glan- 1989] SPELLENBERG AND KNIGHT: ERIGERON ACOMANUS 117 Fig. 2. Erigeron acomanus Spellenberg & Knight (drawn from the holotype, Knight 2689, unless otherwise indicated). (A) Involucre; (B) Cypsela {Knight 3613); (C) Disk flower; (D) Variation in leaves (upper set Spellenberg 4911, 4912, lower set from type); (E) Habit; (F) Late diakinesis in pollen parent cell, 2/2=9ii, {Spellenberg and Ward 9493). dular-puberulent and lightly to moderately white-strigose. Ligules 16-30, white, 4.5-9 mm long, 1.3-1.8 mm wide. Disk corollas 2.5- 3 mm long, yellowish, the triangular lobes often blushed with ma- roon. Cypselas somewhat flattened, ca. 2 mm long, 2-nerved, lightly 118 MADRONO [Vol.36 hirsute; pappus of 20-25 fine barbellate bristles 1.5-2.5 mm long, 0.8-1 times the length of the disk corolla; pappus bristles intermin- gled with a few shorter setae that are continuous with the hairs on the upper portion of the fruit. Chromosome number: n=9. Paratypes: USA, NM, McKinley Co., 5 mi N of US Highway 66 in Prewitt on road past power plant up Casomero Draw, 35°26'45"N, 108°02'40"W, elev. 7000 ft, 22 Jun 1985, Barrie 1412 (COLO, NMC, TEX, US); ca. 2 mi N of Prewitt, T14N R21E sect. 24 NE 1/4, elev. 6900 ft, 20 Jul 1983, Fletcher 7074 (ALB); ca. 5 mi N of Prewitt, 7000-7100 ft, 30 Aug 1982, Knight 1717 (TEX); 13 Aug 1986, Knight 3416 (UNM); 3 Sep 1987, Knight 3613 (UNM); Prewitt, behind Plains Electric power plant, 6 Jun 1985, Spellenberg and Corral 8231 (NMC); ca. 2 mi N of Prewitt, T14N R12E sect. 24 NE 1/4, 6800 ft, 31 May 1986, Spellenberg 8497 (NMC); 31 May 1988, Spellenberg and Ward 9493, voucher for chromosome count 2n=9ii (NMC); Valencia Co. (this area now in Cibola Co.), just E of Laguna Indian Reservation in Blue Water Canyon, ca. 1 3 air mi SE of Acoma Pueblo, T6N R7W boundary of sects. 25-36, 24 Sep 1977, Spellenberg 4911, 4912 (NMC, NY). Distribution and habitat. Erigeron acomanus is named for the Acoma Indian Pueblo. It was first discovered at the base of the cliffs that delimit the eastern boundaries of the lands of that tribe {Spel- lenberg 4911, 4912). At this site the species is restricted to shaded sandy slopes that build up beneath cliffs of Zuni Sandstone (Dane and Bachman 1957), which is derived from an eolian dune of upper Jurassic age (Maxwell 1 975). At the type locality, near Prewitt, plants are also restricted to protected sandy slopes beneath cliffs of Entrada Sandstone of Jurassic age, which are capped by a gypseous limestone segment of the Todilto Formation (Smith 1954). Both sandstones degrade to produce a fairly coarse-grained alkaline sand. At present, E. acomanus is known from only these two small populations and is being considered for federal protection by the U.S. Fish and Wildlife Service. There were perhaps 100 plants in the Blue Water Canyon population {Spellenberg 4911, 4912) in ca. V4 hectare. At the type locality there are probably a few thousand plants along canyon bases in ca. 2 km^. However, the strata on which it occurs, and other related Mesozoic formations, extend intermit- tently along a 150-km east-west oriented crescent in west-central New Mexico (Green and Pierson 1977). Other populations of the species can be expected to occur in sheltered sandy places in this area. Where it is known to occur, common plant associates are Pinus edulis Engelm., Juniperus monosperma (Engelm.) Sarg., Artemisia tridentata Nutt., Ribes cereum Dougl., Oryzopsis hymenoides (Roe- mer & Schultes) Ricker, Bouteloua gracilis (Kunth) Lagasca, Pen- stemon barbatus (Cav.) Roth, Yucca angustissima Engelm., Gutier- rezia sarothrae (Pursh) Britt. & Rusby, Aletes sessiliflorus Theobald 1 989] SPELLENBERG AND KNIGHT: ERIGERON ACOMANUS 1 1 9 & Tseng, and other species of either Intermountain or Madrean affinities. Morphology and relationships. Erigeron acomanus is readily dis- tinguished from other species of Erigeron in New Mexico by its monocephaUc, leafy stems, round or obtuse leaf tips, white rays, and mat-forming habit. In Martin and Hutchins (1981) it keys with difficulty to near E. vetensis Rydb., but E. vetensis differs by its greater number of rays that are pink or blue and its herbage that is more or less densely glandular and at most sparsely hirsute (Cron- quist 1947). In Cronquist, again with some difficulty, E. acomanus keys to E. tener of the Intermountain Region or to E. cronquistii Maguire, a very local species of northern Utah placed near to E. tener by Cronquist. Each unknown to the other, we sent our collec- tions to Dr. Guy Nesom for identification. He suggested that we had disjunct material of E. tener. Erigeron acomanus has white rays, has a mat-forming habit, and lacks peg-like glandular trichomes similar in shape to the "type C" trichomes illustrated by Nesom (1978) for another species. In con- trast, labels on specimens of E. tener, when the color is given, con- sistently note the rays to be bluish or pinkish, descriptions consistent with those of Cronquist (1 947) and Hitchcock et al. (1 955). Erigeron tener also differs by its densely caespitose habit and by the presence of golden, peg-like, glandular trichomes on the stem. Erigeron aco- manus also differs from E. tener by a number of quantitative features, most notably E. acomanus having generally shorter petioles, a great- er number of leaves on the flowering stem, and a consistently mono- cephalic habit (Table 1). These differences, especially when consid- ered with the disjunction in geographic ranges and differences in habitat, support the delimitation of E. acomanus as a species distinct from E. tener. Pappus bristles and habit readily distinguish E. aco- manus from E. cronquistii. In the former there are 20-25 bristles that are about as long as the disk corolla; in the latter there are 1 2- 20 bristles that are conspicuously shorter than the disk corolla (Cron- quist 1947). In little more than a decade two other narrowly endemic species of Erigeron have been described from New Mexico: E. hessii Nesom (1978) and E. scopulinus Nesom & Roth (1981), both more southern in distribution (Fig. 1). The relationship of E. acomanus may ulti- mately be shown to be with these. All are monocephalic and white- rayed, and perennate from a highly-branched subrhizomatous or rhizomatous caudex. Erigeron acomanus is more similar in habit and pubescence to E. scopulinus than to E. hessii. The three are, nevertheless, easily distinguished. Erigeron hessii, superficially sim- ilar to E. acomanus when pressed, is known only from about 3000 m in the Mogollon Mts. It has acute leaves, easily distinguished minute "type B" trichomes, and coarse "type A" trichomes (Nesom 120 MADRONO [Vol. 36 Table 1 . Comparison of Erigeron acomanus with E. tener. Values are for the mean ± one standard deviation, with the range given in parentheses, p values for the differences between means are derived from Student's t-test. In making leaf measurements, the base of the blade was arbitrarily determined to be at the point where the width of the petiolar base becomes twice that of its narrowest point. Erigeron acomanus is represented from two sites; E. tener from 48 sites throughout the range of the species. Erigeron acomanus Erigeron tener \^ii50 cm yr ' average rainfall, and no difference was apparent at sites with <50 cm yr ' rainfall. Variation in species composition between grassland and understory was independent of rainfall gradient. Differences in individual species presence and abundance between grassland and understory were found at all sites. We conclude that variation in canopy effect on biomass resulted from changes in relative production between understory and open grassland, not from differences in relative composition. Winter deciduous Quercus douglasii Hook. & Am. (blue oak) forms a patchy canopy over and within a continuous annual herbaceous layer on more than a million hectares in California (McClaran and Bartolome 1989). Average rainfall in blue oak woodland varies geo- graphically from 30 to 100 cm yr~^ (Griffin 1977). In high rainfall areas understory biomass was lower than open grassland (Bartolome 1 986; Jansen 1987), and understory biomass increased after thinning or removal of oak canopy (Johnson et al. 1959; Heady and Pitt 1979; Kay 1987). In low rainfall areas understory biomass was greater than in open grassland (Duncan and Reppert 1960; Holland 1980). Com- position differences between grassland and understory were reported for several sites (Heady and Pitt 1979; Holland 1980; Bartolome 1986; Jansen 1987). Because Q. douglasii cover was not constant among these study sites, relative influence of rainfall and tree cover on herbaceous biomass and composition cannot be interpreted. Madrono, Vol. 36, No. 3, pp. 141-153, 1989 142 MADRONO [Vol. 36 In general, differences between understory and grassland com- position have been attributed to negative competition effects or pos- itive site modification effects (e.g., Parker and Muller 1982; Yavitt and Smith 1983; Schott and Pieper 1985). Biomass differences be- tween understory and grassland also have been attributed to these positive and negative effects. Examples of positive effects described for relatively xeric areas (e.g., Tiedemann and Klemmedson 1977; Patten 1978), contrast with negative effects in mesic regions (e.g., Halls and Schuster 1965; Jameson 1967; Ford and Newbould 1977). Negative effects of light and water competition were suggested in tree-grass models (Walker et al. 1981; McMurtie and Wolf 1983) and the interpretation of empirical biomass data (Knoop and Walker 1985). The suggestion that species interactions vary along environmental gradients (Begon et al. 1986) provides a framework to organize the diversity of canopy effects by Quercus douglasii. We used this frame- work to identify our objective and generate a null hypothesis. To evaluate effects of rainfall and Q. douglasii canopy on herbaceous understory, we used the null hypothesis that the effect of 50% Q. douglasii canopy on the similarity of grassland and understory bio- mass and composition does not vary along a rainfall gradient. Methods Study sites. We selected five sites along an average annual rainfall gradient of 40 to 90 cm yr ' (Tables 1 and 2, Fig. 1). Livestock had grazed all sites each year for the last five years prior to 1985. Slopes on all sites were less than 1 5 percent with either a west or a south aspect. At each site we sampled herbaceous biomass and composition under three canopies and their adjacent open grassland areas. All canopies consisted of three to five Quercus douglasii projecting ap- proximately 50% cover over 100-200 m^. Trees were 7-10 m tall, 30-60 cm in diameter at 1.35 m height, and 2-5 m apart. The sample canopies were randomly selected from 5-6 available canopies with the above characteristics at each site. To study ungrazed herbaceous vegetation we randomly placed two pair (pair = one understory and one grassland) of 1-m^ pyramidal cages (Frishnecht and Conrad 1965) in each of the three understory and grassland areas per site. Biomass. We estimated live herbaceous biomass at the time of greatest standing biomass (PEAK) in early to late May 1986 and 1987 because phenology varies with latitude. The estimates were made within livestock exclosure cages by weighing vegetation clipped in a 0.0625-m- plot to 1 cm height, and dried at 65°C for 48 hr. Cages were again randomly located after PEAK sampling in 1986 to accommodate 1987 sampling. 1989] McCLARAN AND BARTOLOME: HERBACEOUS UNDERSTORY (3D (U pa c x: (U oo in ON C/3 O O 5 oo a ON ~. ^ I: GO S o -2 ? s ^ s O ,1-H Uh — " 1/5 ^ c3 T3 G < DC u Q < o H < U o c On 00 c o 3 5^ ci3 o o 03 C w .2 3 ■4— » 00 C o X o u < ^ >< ZJ O ^ O X 2 2 D. OJ X! cd o = S Jii 3 S S ^ a _o C 03 03 O n c*^ rj- oo m ^ ^ >V-> O — I m m OO •(-» a w cd ■(-' OO u >■ 'c c "a o X >3 .Si U oo C/) 03 ■t-< OO w >■ oo tin U . c ^ 2 Id 03 ^ O C ■-i O §,s oo oo 144 MADRONO [Vol. 36 Table 2. Cumulative Seasonal Rainfall at Study Sites During 1985-1986 AND 1986-1987, AND Long-term Average. ' Station records for Hopland, Sierra, and San Joaquin; nearest National Atmospheric and Oceanic Administration (NAOA) station, Mt. Hamilton for San Felipe and Paso Robles for Sinton. - Only trace amounts of rainfall occur between May and September in this Mediterranean climate. Site' Rainfall (cm) 1 Sep-30 Nov 1 Sep-28 Feb 1 Sep-30 Apr^ Length of record 1985 1986 X 1985- 1986 1986- 1987 X 1985- 1986 1986- 1987 X Hopland 26.2 7.8 21.1 92.5 42.0 73.8 112.0 59.3 92.6 34 yr Sierra 21.6 8.0 18.2 67.1 34.6 55.2 84.8 48.1 70.6 25 yr San Felipe 15.9 9.5 10.1 46.7 25.2 39.7 61.8 35.7 52.7 106 yr San Joaquin 14.9 2.3 8.7 41.4 19.5 33.6 54.3 27.5 46.3 53 yr Sinton 3.8 2.0 5.9 29.2 8.6 27.5 35.1 17.6 41.8 101 yr We analyzed PEAK biomass with a 3-way analysis of variance (ANOVA) to assess year, site, and canopy effects. When main effects or interactions were significant, we employed Duncan's multiple range test to determine year, site, and canopy differences (Steel and Torrie 1980). Composition. We estimated aerial cover by species in each cage immediately prior to 1986 PEAK biomass sample. Estimates were made by recording the first species hit when lowering a fine-tipped metal pin 1 00 times per cage. Pins were mounted in a ten pin point- frame (Heady and Rader 1958) that was independently placed in ten different locations per cage to amass 100 hits. Taxonomic no- menclature follows Munz (1968). Bareground was recorded when the pin did not hit vegetation. To assess site and canopy effects on total plant cover, grass and forb cover, and cover of species that were present on all five sites we used a 2-way ANOVA. When main effects or interactions were significant, we employed Duncan's mul- tiple range test to determine year, site, and canopy differences (Steel and Torrie 1980). We normalized cover data with an angular (arcsin) transformation before ANOVA and Duncan's analysis (Steel and Torrie 1980). We used a cover-weighted similarity index to compare species composition in understory and grassland between study sites. Sim- ilarity was estimated with a cover weighted modification of Soren- sen's ( 1 948) similarity index (SI) that Barbour et al. ( 1 987) described as: SI = 2 2Q /(A + B) \all 1 / 1989] McCLARAN AND BARTOLOME: HERBACEOUS UNDERSTORY 145 Fig. 1 . Location of five study sites in California. where A is total percent cover in stand A, B is total percent cover in stand B, and C is the lowest percent cover of species i from either stand A or B. To estimate between site change in species composition along the gradient we calculated a similarity index between adjacent study sites for both understory (UNDER) and open grassland (OPEN). Rogers (1980) used this technique to describe composition changes between adjacent sites along a gradient. This research design stresses the difference between grassland and understory within each site, and compares these within-site differ- ences among sites to describe a change in canopy effect along a rainfall gradient. The between-site differences in absolute grassland or absolute understory biomass and composition that also are avail- able from this research design are mostly expected and less germane to our study objective. We chose a relatively small sample size of six plots from three canopies and grasslands at each site to enable completion of biomass and composition sampling at a site in one day. This pace allowed us to minimize phenologic differences among sites by sampling adjacent sites on subsequent days. The ANOVA procedures compensate for small sample sizes by requiring greater differences between population means to reach significant differences (Steel and Torrie 1980). We could not complete the composition 146 MADRONO [Vol. 36 sampling in 1987, and we recognize the importance of yearly vari- ability in annual grassland composition (Pitt and Heady 1978). How- ever, we include the 1986 composition data because it reveals in- teresting relationships between biomass and composition along the rainfall gradient, and it represents unique empirical data for com- parison with future work. Results Biomass. Significant canopy x site, and year x site interactions were found for herbaceous biomass at PEAK sampling period from 3-way ANOVA (Table 3). Canopy x site interaction shows a change in OPEN biomass relative to UNDER biomass along the gradient: as rainfall increased OPEN biomass exceeded UNDER (Fig. 2). Year X site interaction shows less PEAK biomass in 1987 than 1986 at all sites except San Felipe (Fig. 2). Absence of significant canopy x year, and site x canopy x year interactions illustrates the consis- tency of this pattern between wet and dry years (Table 3), even though absolute biomass values differed between years (Fig. 2). Composition. Fifty-six species were recorded on all five sites com- bined. Nearly half were found only in grassland or understory; 13 species were recorded only in grassland and 1 4 species only in under- - story. A complete species list is available from the senior author. Aerial cover of only seven herbaceous species was present on all of the five sites (Table 4). Avena barbata, Bromus diandrus, Lupinus bicolor, forb, and total cover varied among sites (Table 5). Avena barbata cover was greater at San Felipe and Hopland than at Sinton and San Joaquin. Bromus diandrus cover was greater at Sinton than San Felipe, and L. bicolor was greater at San Felipe than at Hopland. Forb cover was greater at Sierra and Hopland than at San Joaquin. Total cover was greater at San Felipe and Hopland than at other sites. Bromus diandrus, Erodium cicutarium, Lupinus bicolor, forb, and total cover varied between grassland and understory (Table 5). Bro- mus diandrus cover was greater under canopy, and the other taxa and taxa groups were greater in grassland. Significant canopy x site interactions were found for Bromus mollis, Festuca megalura, and grass cover (Table 5). Bromus mollis was greater in grassland than understory at all sites except San Joa- quin and Sierra. Festuca megalura cover was only greater under canopy than in grassland at San Felipe. Understory grass cover was greater than grassland cover at Sierra and San Joaquin. Similarity of grassland and understory composition between ad- jacent sites varied along the rainfall gradient (Fig. 3). Similarity index (SI) values for grassland and understory between Hopland, Sierra, and San Felipe were similar. Between San Felipe and San Joaquin 1989] McCLARAN AND BARTOLOME: HERBACEOUS UNDERSTORY 147 7 Hopland Sierra San Felipe San Joaquin Sinton Locat ion IZZl Under ^ Open Fig. 2. Understory (open bars) and open grassland (diagonal-hatched bars) standing biomass for five study sites in PEAK 1986 (A) and PEAK 1987 (B). Vertical bars represent ±one standard error of the mean. grassland SI increased while understory decreased. Between San Joa- quin and Sinton understory SI increased to a level similar to the grassland. Table 3. Probability Values for ANOVA F Statistic Describing Site, Canopy, Year, and Interaction Effects on Peak Herbaceous Biomass. Significance of single variables is of no consequence if the interaction of these variables is significant. Variables and interactions Site Canopy Site X Site x canopy Year year Canopy x year Site X canopy x year 0.0001 0.0001 0.0001 0.001 0.0001 0.09 0.183 148 MADRONO [Vol. 36 ^ 2 ^ p < t oo ^ C 1/5 < C/2 lU u c U o U c t/5 O 5 H IZ) t/5 o B < ^ H Q O D m w < K c o c • ^ CO c a* a o c 1/5 1/5 ■*— > C/5 (U c/5 T3 c o X c a O C C a O (U c O Ph c CO c O u C a (U a O c/) 'o a CO mm (Noo ooo om m/^ om r--^ r-oo — < (N fN en (N od en m -H as \D rn Tj-' (N m >r^ON oO'^ OO i^oo OO oofN oo/") OO — < d d in 0^ m r- OO o ^ m OO rn On On OO m o o d d o o d d OO (N rs| o ir^(N r~-io oooo ro— > oo r--^ (NfN rooo mo r--^ osn (Nm (N-^ mn 5 C3 ! cicut megat bicolc 'odium Festuca . ipinus X3 S-H O 1/3 1/3 o O H 1989] McCLARAN AND BARTOLOME: HERBACEOUS UNDERSTORY 149 Table 5. Probability Values for ANOVA F Statistic Describing Site, Canopy, AND Interaction Effects on Total and Absolute Cover of Common Herbaceous Species and Species Groups during Peak Sampling Period, 1986. Significance of single variables is of no consequence if the interaction of these variables is significant. Variables and interactions Species/species groups Site Canopy Site X canopy Avena barbata 0.01 0.24 0.48 Bwmus mollis 0.001 0.001 0.01 Bromus diandrus 0.06 0.001 0.10 Erodium cicutarium 0.31 0.001 0.50 Festuca megalura 0.16 0.32 0.22 Lupinus bicolor 0.06 0.002 0.07 Medicago hispida 0.16 0.37 0.62 GRASS 0.005 0.076 0.01 FORB 0.04 0.001 0.12 TOTAL 0.02 0.001 0.23 Discussion Variation in similarity of herbaceous biomass in grassland and under approximately 50% Quercus douglasii canopy along a rainfall gradient lead to our rejection of the null hypothesis for biomass. Change in similarity occurred at approximately 50 cm yr ~^ average rainfall; similarity in biomass was greater below 50 cm yr '. This pattern was repeated in both wet and dry years. We did not reject the null hypothesis for composition because grassland and under- story similarity varied independently of the rainfall gradient. Neutral canopy effects characterized response of PEAK herba- ceous biomass to Quercus douglasii at lower rainfall sites. PEAK biomass was not different between grassland and understory at San Joaquin and Sinton. Negative canopy effects characterized response of PEAK herbaceous biomass to Q. douglasii at higher rainfall sites. PEAK biomass was higher in grassland than understory at Hopland, Sierra, and San Felipe. Variation in Quercus douglasii effect on herbaceous biomass was assumed in recent mangement guidelines describing optimum stock- ing levels of oak (Passof et al. 1985). These guidelines suggested 1) reducing tree cover below 50% on sites with > 50 cm average rainfall to increase herbaceous production, and 2) maintaining tree cover at approximately 50% on sites with < 50 cm average rainfall to combine high herbaceous production with oak browse, acorn crops, wildlife habitat, and potential fuelwood. Our results document a shift in understory biomass around a 50% tree cover and 50 cm yr"' rainfall intersection. However, our results do not address impacts of tree removal from 50% cover to lower levels, reduction to 50% from higher levels, or unmanipulated oak cover at levels other than 50%. 150 MADRONO [Vol. 36 0.7 0.65- 0.25 H Hopland Sierra Sierra San Felipe San Felipe San Joaquin San Joaquin Sinton Fig. 3. Similarity of understory (crosses) and open grassland (solid boxes) species composition between adjacent sites along the rainfall gradient. We suggest that differences in species composition between grass- land and understory are independent of the rainfall gradient. This interpretation is supported by canopy effects that are consistently positive or negative throughout the gradient. Nearly half the species were found in only grassland or understory; Bromus diandrus cover was greater in understory than grassland; and Erodium cicutarium, Lupimis bicolor, forb, and total cover were greater in grassland than in understory. Furthermore, there was no gradient related distri- bution of sites where Q. douglasii effect on composition varied from negative to neutral to positive. Understory grass and Bromus mollis cover were greater than in grassland at all sites except San Joaquin and Sierra; understory F. megalura cover was greater than in grass- land only at San Felipe. In addition, the shift from comparable similarity in grassland and understory composition between adjacent high-rainfall sites, to disparity between San Felipe and San Joaquin was not repeated between San Joaquin and Sinton. Because com- position was measured only in 1986, interpretations should recog- nize yearly variation in annual grassland composition (Pitt and Heady 1978). However, relative yearly variation between grassland and understory composition is not known. Inconsistent trends in biomass and composition similarity be- tween understory and grassland along the gradient suggests that bio- mass and composition change independently. We propose that the 1 989] McCLARAN AND BARTOLOME: HERBACEOUS UNDERSTORY 1 5 1 gradient related effect of Quercus douglasii on herbaceous biomass results from changes in relative understory and grassland production, not from changes in relative composition. We hypothesize that understory production increases relative to grassland in low rainfall areas because Quercus douglasii shading improves available soil moisture for herbaceous growth. However, it is probable that Q. douglasii ameliorates other potentially limiting factors such as soil nutrients and extreme temperatures in low rainfall areas. Our hypothesis is based on annual grassland production and rainfall relationships. Especially important is maximum production in spring (Pitt and Heady 1978). Where PEAK understory biomass is not lower than that of grassland, we suggest shading by oak foliage prolongs availability of soil moisture in spring enabling understory biomass to reach grassland levels despite negative effects of light reduction. If this hypothesis has merit then availability of soil mois- ture in late spring will be greater under 50% Q. douglasii canopy than grassland only on sites with <50 cm yr"' average rainfall. We conclude that the effect of 50% Quercus douglasii canopy on herbaceous understory biomass does vary along a rainfall gradient, but species composition does not. Above 50 cm yr ' average rainfall understory biomass is less than open grassland, but below 50 cm yr'^ there is no difference in biomass. Acknowledgments This research was funded in part by the CaHfornia Department of Forestry. A. Dennis, D. Duncan, J. Dunne, L. Huntsinger, N. McDougald, A. Murphy, D. Spring- steen, and W. Weitkamp assisted in plot location and sampling; and M. Blumler, C. Gonzales, B. Roundy, and E. L. Smith carefully reviewed earlier versions of this manuscript. Literature Cited Barbour, M. G., J. H. Burk, and W. D. Pitts. 1987. Terrestrial plant ecology, 2nd ed. Benjamin/Cummings Publ. Co., Menlo Park, CA. Bartolome, J. W. 1986. Herbaceous productivity in oak woodland. Trans. West- sern Sect. WildHfe Soc. 22:1 12-1 16. Begon, M., J. C. Harper, and C. R. Townsend. 1986. Ecology: individuals, pop- ulations, and communities. Sinauer Associates Inc., Sunderland, MA. Duncan, D. A. and J. N. Reppert. 1 960. A record drought in the foothills. U.S.D.A. For. Serv. Misc. Paper PSW-46. Ford, E. D. and P. J. Newbould. 1977. The biomass and production of ground vegetation and its relation to tree cover through a deciduous woodland cycle. J. Ecology 65:201-212. Frishknecht, N. C. and P. W. Conrad. 1965. Adaptable, transportable utilization cages. J. Range Managem. 18:33-34. GowANS, K. D. 1958. Soil survey of the Hopland Field Station. California Agri- cultural Experiment Station, Berkeley, CA. Griffin, J. R. 1977. Oak woodland. Pp. 383-415 in M. G. Barbour and J. Major (eds.). Terrestrial vegetation of California. Wiley Interscience, New York. 152 MADRONO [Vol. 36 Halls, L. K. and J. L. Shuster. 1965. Tree herbage relations in pine-hardwood forests of Texas. J. Forest. (Washington) 63:282-283. Heady, H. F. and L. Rader. 1958. Modifications on the point frame. J. Range Managem. 11:95-96. and M. D. Pitt. 1979. Reactions of northern California grass- woodland to vegetational conversions. Hilgardia 47:51-73. Herbert, F. W. and E. L. Begg. 1969. Soils of the Yuba area, California. Coop- erative Project, Department of Soil and Plant Nutrition, University of California, Davis and County of Yuba, CA. Holland, V. L. 1980. Effect of blue oak on rangeland forage production in central California. Pp. 314-318 in T. R. Plumb (tech. coord.). Ecology, management and utilization of California oaks. U.S.D.A. Gen. Tech. Report PSW-44. Jameson, D. A 1 967. The relationship of tree overstory and herbaceous understory vegetation. J. Range Managem. 20:247-249. Jansen, H. C. 1987. The effect of blue oak removal on herbaceous production on a foothill site in the northern Sierra Nevada. Pp. 343-350 in T. R. Plumb and N. H. Pillsbury (tech. coords.). Multiple use management of California's hard- wood resources. U.S.D.A. Gen. Tech. Report PSW-100. Johnson, W., C. M. McKell, R. A. Evans, and L. J. Berry. 1959. Yield and quality of annual range forage following 2,4-D application on blue oak trees. J. Range Managem. 9:18-20. Kay, B. L. 1987. Long-term effect of blue oak removal on forage production, forage quality, soil, and oak regeneration. Pp. 351-357 in T. R. Plumb and N. H. Pillsbury (tech. coords.). Multiple use management of California's hardwood resources. U.S.D.A. Gen. Tech. Report PSW-100. KiMCHANG, K. 1965. Detailed soil survey of San Joaquin Experimental Range, O'Neals, California. In D. A. Duncan (ed.). Comprehensive network site de- scription—San Joaquin. U.S. Biol. Program Tech. Rep. No. 296, Natural Re- sources and Ecology Laboratory, Colorado State University, Fort Collins, CO. Knoop, W. T. and B. H. Walker. 1985. Interactions of woody and herbaceous vegetation in a southern African savanna. J. Ecology 73:235-253. LiNDSEY, W. C. 1 974. Soil survey of the eastern Santa Clara County area, California. U.S. Soil Conservation Service. . 1983. Soil survey of San Luis Obispo County, California. Paso Robles area. U.S. Soil Conservation Service. McClaran, M. p. and J. W. Bartolome. 1 989. Fire related recruitment in stagnant Quercus douglasii populations. Can. J. For. Res. (in press). McMurtie, R. and L. Wolf. 1 983. A model of competition between trees and grass for radiation, water and nutrients. Ann. Bot. (London) 52:449-458. MuNZ, P. A. 1968. A California flora and supplement. Univ. California Press, Berkeley. Parker, V. T. and C. H. Muller. 1982. Vegetational and environmental changes beneath isolated live oak trees {Quercus agrifolila) in a California annual grass- land. Amer. Midi. Naturalist 107:69-81. Passof, p. C, W. J. Clawson, and E. L. Fitzhugh. 1985. Preliminary guidelines for managing California's hardwood rangelands. Coop. Ext. Publ. No. 21413, Univ. California. Patten, D. T. 1978. Productivity and production efficiency of an upper Sonoran desert ephemeral community. Amer. J. Bot. 65:891-895. Pitt, M. D. and H. F. Heady. 1 978. Responses of annual vegetation to temperature and rainfaU patterns in northern California. Ecology 59:336-350. Rogers, R. S. 1980. Hemlock stands from Wisconsin to Nova Scotia: transition in understory composition along a floristic gradient. Ecology 61:178-193. ScHOTT, M. R. and R. D. Pieper. 1985. Influence of canopy characteristics of one- seed juniper on understory grasses. J. Range Managem. 38:328-331. 1989] McCLARAN AND BARTOLOME: HERBACEOUS UNDERSTORY 153 SoRENSEN, T. 1948. A method of establishing groups of equal amplitude in plant sociology based on similarity of species content. Biol. Skr. 5:1-34. Stangenberger, a. G. and J. I. Mallory. 1981. Index to forest and range soil series in California. Soil- Vegetation Survey, Calif Dept. Forest. Div. Agric. Sci., Univ. California, Berkeley. Steel, R. G. D. and J. H. Torrie. 1980. Principles and procedures of statistics: a biometrical approach, 2nd ed. McGraw-Hill, New York. TiEDEMAN, A. R. and J. O. Klemmedson. 1977. Effect of mesquite trees on vege- tation and soils in the desert grassland. J. Range Managem. 3:361-367. Walker, B. H., D. Ludwig, C. S. Holling, and R. M. Peterman. 1981. Stability of semi-arid savanna grazing systems. J. Ecology 69:473-498. Yavitt, J. B. and E. L. Smith. 1983. Spatial patterns of mesquite associated her- baceous species in an Arizona grassland. Amer. Midi. Naturalist 109:89-93. (Received 25 Apr 1988; revision accepted 8 Apr 1989.) NOTEWORTHY COLLECTIONS British Columbia WoLFFiA COLUMBIANA Karstcn (Lemnaceae).— Canada, British Columbia, Victo- ria, water filled ditch N of Blenkinsop Lake with Spirodella polyrrhiza (L.) Schleiden, 48°29.2'N, 123°21.6'W, elev. 20 m, 27 Jun 1988, A. & O. Ceska 23775 (V); Victoria, Swan Lake, accumulated along the shore in a stand of Phalaris arundinacea L. near the wharf at the Swan Lake Nature Sanctuary, 48°27.9'N, 123°22.2'W, elev. 15 m, 14 Dec 1988, A. Ceska & C. Dorworth 25540 (V). Identification. Wolffia columbiana differs from W. borealis (Engelm.) Landolt by its globose shape without a flat top, from W. arrhiza (L.) Horkel by fewer stomata (up to 10). Wolffia globosa (Roxburgh) Den Hertog & van der Plas is smaller than W. columbiana and its fronds are slightly elongated. Our identification was confirmed by W. P. Armstrong and E. Landolt. Significance. The first report from British Columbia. In North America W. colum- biana is frequent in SE USA, and in SE Canada, rare in California [Armstrong, Madrofio 31:123-124. 1984; Armstrong and Thorne, Madrono 31:171-179, 1984; Armstrong Fremontia 13(1):1 1-14, 1985] and Oregon. In western Canada reported from Saskatchewan [as W. arrhiza: Looman, Can. Field. Nat. 97:220-222, 1983] and from Alberta [Griffiths, Alberta Naturalist 18(1): 18-20, 1988]. For the world distri- bution data and map see pp. 371-376 in Landolt. The family of Lemnaceae— a monographic study. Veroff. Geobot. Inst. ETH, Stiftung Rubel 71:1-566. 1986.-W. F. Savale, Jr. and Adolf Ceska, % Botany Unit, Royal British Columbia Museum, Victoria, BC, V8V 1X4, Canada. EFFECTS OF VARYING FIRE REGIMES ON ANNUAL GRASSLANDS IN THE SOUTHERN SIERRA NEVADA OF CALIFORNIA David J. Parsons Research Division, Sequoia and Kings Canyon National Parks, Three Rivers, CA 93271 Thomas J. Stohlgren Cooperative National Parks Resources Studies Unit, University of California, Davis, CA 95616 Abstract Effects of up to three successive spring and fall bums on composition and biomass of the predominantly non-native grasslands of the southern Sierra Nevada foothills were evaluated. Fall and spring burning regimes increased the number and biomass of both alien and native forb species. No native grass species became established following the treatments. Thus, whereas the biomass of alien grass species can be reduced by repeated burning, they will be replaced by increases in both alien and native forbs. Changes seen following one or two burns (spring or fall) were not sustained following cessation of burning treatment. The annual grasslands that characterize much of California, in- cluding the low elevation foothills of the southern Sierra Nevada, are dominated by species native to the Mediterranean Basin (Wester 1981). Prior adaptation to grazing by livestock has favored alien species in the replacement of native grasses and forbs (Barry 1972; Jackson 1985; Macdonald et al. 1988). Analysis of both historical accounts and microfossil remains indicate that areas that once sup- ported either extensive native annual grasslands or Stipa bunchgrass prairie are now dominated by alien annual grasses (Barry 1972; Heady 1977; Bartolome et al. 1986). The abilities to withstand drought and grazing have combined to assure the continued dom- inance of annual Mediterranean grasses. The timing and intensity of precipitation and grazing pressure, including the amount of nat- ural mulch retention, have been shown to significantly influence vegetative production and species composition on annual grassland sites in central California (Talbot et al. 1939; Duncan and Wood- mansee 1975; Bartolome et al. 1980). Hervey (1949) quantified the effects of an early summer burn in temporarily favoring broadleaved forbs over grasses on a coastal foothill range and Larson and Duncan (1982) have documented a near doubling of production two years following a fall burn on the San Joaquin Experimental Range near Fresno. Yet, although variable patterns of climate, fire and grazing Madrono, Vol. 36, No. 3, pp. 154-168, 1989 1989] PARSONS AND STOHLGREN: GRASSLAND FIRE REGIMES 155 result in considerable year to year fluctuation in nonnative annual species composition and biomass (Bentley and Talbot 1958; Heady 1958; Pitt and Heady 1978), those studies showed no indication of successful reestablishment of the native flora. Land management agencies charged with the preservation of nat- ural ecosystems face the dilemma of either accepting the loss of a significant component of the native flora or attempting to restore a more native species composition. The former, as has been suggested by Heady (1977), requires a compromise in management objectives, essentially declaring alien species as naturalized. The latter requires both a sophisticated knowledge of grassland ecology and a potentially intensive restoration program, if such is even possible. In Sequoia National Park, several thousand hectare of annual grassland are protected as both part of the Park and the Sequoia and Kings Canyon International Biosphere Reserve. National Park Service manage- ment policy calls for protection and restoration of this ecosystem. Although effective protection from livestock grazing is provided, the question remains as to whether native species can ever successfully be restored to the area. An active program of restoring periodic fire to an area where fires have been effectively suppressed for most of a century (Parsons 1981; Bancroft et al. 1985), together with evidence that the frequency and season of fire can influence species composition and production in other grassland communities (Hover and Bragg 1981; Towne and Owensby 1984), led us to test the effects of varying fire regimes on species composition and biomass. Similar experiments, using a com- bination of seeding, fertilization, and burning, provided mixed re- sults in restoring degraded grasslands in San Diego County (Garcia and Lathrop 1984). The objectives of this study were to investigate the eflfects of fre- quency and season of burning on the relative composition and dom- inance of native and alien species in the annual grassland commu- nities of the southern Sierra Nevada. Such data are essential to understanding the effects of past and present management practices as well as in evaluating the possibility of using fire as a tool to reestablish native species. This is critical to understanding the short- and long-term impacts of reestablishing natural fire regimes (Parsons etal. 1986). Study Area The study area is located on a gentle eastern exposure at an ele- vation of 700 m in the rolling foothills of the drainage of the Middle Fork of the Kaweah River, Sequoia National Park, Tulare County, California. The region is characterized by the hot, dry summers typical of Mediterranean climates (Aschmann 1973). Annual pre- 156 MADRONO [Vol. 36 cipitation, which is concentrated in January, February, and March, averaged 139, 73, 133, and 202 percent, respectively, of the long- term annual average of 68.1 cm during the four study years. Vege- tation in the area consists of annual grasslands beneath scattered blue oak {Quercus douglasii Hook. & Arn.). Grazing by domestic cattle and horses occurred on the site during the late 1800's and early 1900's but has now been absent for at least 60 years. The area has no record of burning after a 1960 wildfire burned much of the vicinity (Stocking 1966). Park fire maps record no fires in the area between 1925 and 1960. Prior to the fire protection provided by the creation of Sequoia National Park in 1 890, lightning fires, together with intentional ignitions by aboriginals and sheep- herders (for a brief period between about 1 860 and 1 890), are thought to have regularly burned the Sequoia foothills (Vankat and Major 1978; Parsons 198 1). Previous studies of vegetation in the area have focused primarily on the overstory oak woodland (Baker et al. 198 1; McClaran 1986), and nearby chaparral communities (Stohlgren et al. 1984; Rundel et al. 1987). Soils in the area are characterized by fine to coarse textured sandy loam Ultic Haploxeralfs (Huntington and Akeson 1988). They are derived from granitic bedrock and are moderately deep and moderately well drained. Methods An area approximately 2.0 ha was selected as representative of the foothill annual grasslands of Sequoia National Park. Seven 10-m by 1 0-m study sites were identified within this area in positions that maximized chances of successful fire control. The sites were each assigned one of seven fire treatments based on fire logistical concerns. Treatments included a single fall burn (1980; Fl), two successive fall burns (1980 and 1981; F2), three successive fall burns (1980, 1981, and 1982; F3), a single spring burn (1980; SI), two successive spring burns (1980 and 1981; S2), three successive spring burns (1980, 1981, and 1982; S3), and an unburned control (C). All fires were carried out as prescribed burns under pre-established prescrip- tions previously tested to assure both containment and high fuel consumption. Fall burns were conducted in late October or early November, near the end of the natural fire season (Parsons 1981). During the fall burns fine fuel (cured grass) moisture contents ranged from 10 to 15%, air temperatures from 18 to 21°C, and relative humidities from 40 to 65%. Flame lengths averaged 0.6 to 1.0 m and rate of spread 2.5 to 5.0 m/min. The spring burns were carried out in early to mid-June, after the annual vegetation had dried but before most natural ignitions would normally have occurred (Parsons 1981). Fine fuel moistures ranged from 7 to 9%, air temperatures from 22 to 27°C and relative humidities from 41 to 46%. Flame lengths ranged 1989] PARSONS AND STOHLGREN: GRASSLAND FIRE REGIMES 157 from 0.6 to 1.5 m and rate of spread from 4.0 to 20 m/min. All burns resulted in essentially total consumption of vegetative bio- mass. Burning during the hot dry summer was avoided due to the threat of the fire escaping the area. Vegetation sampling was carried out annually within five random- ly distributed O.OS-m^ circular plots within each burn treatment (new plots were selected each year). Beginning in the spring of 1 980, before the first burn, and continuing through 1983, one growing season following the final burn, all of the current year's vegetative growth was clipped at a height of 1.0 cm above ground level in each plot. Clipped samples were separated by species and oven dried for 24 hours at 94°C to determine dry weight. Sampling was carried out in the spring at peak biomass, before significant seed loss or senescence had occurred. The sampling schedule varied as a function of that year's phenology, falling between 4 May and 24 May. Data collected from the 35 randomly placed 0.05-m- plots sampled during the spring of 1980 were used to characterize the pre-study species com- position and biomass of the area. It is recognized that the lack of true replicate treatments may limit interpretation of the data. Due to logistic constraints related to the burn operation the decision was made to focus on multiple treat- ments at the expense of replication. Possible block eflfects are min- imized by also considering the results as a percent of the 1980 pre- burn condition of the same plots. Statistical analysis of the vegetation data included two-way analysis of variance (ANOVA) to compare the effects of treatment, or burning regime, by year for each of three vegetative groups (alien grasses, alien forbs, and native forbs; no native grasses were encountered) for biomass and species richness. If the F-test ratios were significant (p < 0.05) for either factor, Tukey's multiple range test (SAS Institute Inc. 1985) was used to detect significant differences (p < 0.05) within the vegetative groups for that factor. We recognize the study design does not fully meet the underlying assumption of an analysis of variance since plots were not randomly assigned treatments (due to fire control logistics constraints). But because of the homogeneity of the pre-burn con- dition and the severity of the treatments applied, we present the results of the ANOVA as supportive information. To assess species-specific responses to different burning regimes, percent of total biomass for each major species was calculated fol- lowing the three successive fall or spring burns and compared to that both preceding treatment (1980) and for the 1983 control. Results and Discussion The pre-burn 1980 vegetation of the study sites consisted of eigh- teen species of grasses and forbs with a mean total biomass of 334.5 g/m^ (Table 1). The 35 sample plots averaged 5.6 species per plot. 158 MADRONO [Vol. 36 C3 B o -a c 00 •S 3 0 ^ 1 ^ 00 o c c --^ 13 O <^ S Qc: . o ^ ^ C " u c C/5 ^ 00 a" Q 2:J < ►J 1/5 -> 3 d ^ £ o V O o Ph O Ph o o O ^ O t/3 < q H o PJ GO +1 00 s o c cr Ph o o (/5 a or) od (N (N rO 0^ q 5 2;. s: S: £: -S ^ 5 S a Qq Qq (s!) (U > o o X Cu Co C ■(— > 03 o u ^ 03 S: 2P S 03 o >3 ,^3 >3 >3 3 C3 >3 5 too ^< < ■is 1989] PARSONS AND STOHLGREN: GRASSLAND FIRE REGIMES 159 Table 2. Effect of Treatment by Year on Number of Species and Biomass (g/ 0.05 m-) OF Alien Grasses. Only means with different letters within columns (a, b, c) or within rows (i, j, k) are significantly different at p < 0.05. Treatment codes are C = control, Fl = one fall burn, F2 = two fall burns, F3 = three fall burns, SI = one spring burn, S2 = two spring burns, S3 = three spring burns. Year Treatment 1 o o r» 19o0 1 n o 1 1 98 1 1 O O T 1 982 1 n o '7 1983 No. species c 3 0 3 8 3 0 3 0a Fl 3.0 2.4 3.6 3.0 a F2 3.8 3.2 3.2 3.6 a F3 3.4 3.6 3.4 2.6 b SI 4.4 3.4 3.4 3.4 a S2 3.8 4.4 4.4 3.8 a S3 3.0 3.2 Biomass 3.8 4.0 a C 14.4 13.3 17.8 i 10.4 c/j Fl 15.0 17.0 11.2 9.1 F2 20.0 i 14.1 6.7j 12.3 c/j F3 19.2 i 17.3 i 9.4 2.3 a/j SI 13.2 14.4 18.1 17.3 b S2 18.3 i 11.3 8.6j 8.6j S3 15.0 11.9 8.6 8.1 c Avena fatua, a grass introduced from Europe, dominated all plots, constituting 75.0% of the total biomass (range = 63.2 to 89.7%). Three alien grass species, A. fatua, Bromus mollis, and B. diandrus, occurred in nearly every plot and together accounted for 95.5% (range = 90.3 to 99.6) of the total biomass. Broadleaved forbs oc- curred only sporadically, constituting 5.9 g/m^ or 1.8% (range = 0 to 10.1%) of the total mean biomass. Alien forbs averaged less than one species and 2.8 g/m- whereas native forbs averaged 1.3 species and 3.1 g/m^. Non-native species dominated all plots in both fre- quency and biomass. Although a total of nine native species were encountered, only Brodiaea elegans was found in more than 20% of the plots, and all nine together accounted for less than one percent of the total biomass (Table 1). This pre-treatment composition is similar to that found in other ungrazed or lightly grazed California annual grassland sites (Heady 1977). The only detected pre-burn significant differences between plots located in the different treatment areas were for number of species and biomass of alien forbs in the site to receive a single spring burn (SI) and biomass of native forbs in the site to receive two successive spring burns (S2). The only significant differences found between years for the control plot were for biomass of alien grasses (1982- 1983), and number of species of alien forbs (1980-198 1) (see Tables 160 MADRONO [Vol. 36 Table 3. Effect of Treatment by Year on Number of Species and Biomass (g/ 0.05 m-) of Alien Forbs. Only means with different letters within columns (a, b, c) or within rows (i, j, k) are significantly different at p < 0.05. Treatment codes are C = control, Fl = one fall burn, F2 = two fall burns, F3 = three fall burns, SI = one spring burn, S2 = two spring burns, S3 = three spring burns. Year 1 Icairiiclll 1980 1 Q8 1 1 70 1 1 Q89 1 70Z 1 Q9l'K No. species - c 0.4 a/i 2.4 a/j 1.2a 1.2a Fl 0.6 a/i 2.4 a/j 3.6 b/j 2.4j F2 1.0 i 5.4 b/j 4.2 b/j 3.6 b/k F3 0.8 a/i 3.2j 3.8 b/j 4.0 b/j SI 2.6 b 3.6 3.8 b 2.8 S2 0.6 a/i 2.6 a/j 3.4 b/j 3.2j S3 0.8 a/i 3.6j 4.0 b/j 4.0 b/j Biomass C 0.1 a 0.3 a 0.8 a 0.8 a Fl 0.01 a/j 1.9j 7.6 b/i 1.9j F2 0.1 a/j 2.9 J 7.3 b/i 2.7 b/j F3 0.03 a/i 4.2 b/j 3.6j 5.4 b/j SI 0.7 b 1.5 2.8 1.2 S2 0.03 a/i 0.7 i 1.7 a/j 3.2 b/k S3 0.1 a/i 2.1 i 8.1 b/j 4.3 b/j 2-4). This shows strong similarity between the pre-burn character of the seven treatment areas as well as a consistent year to year character for the unburned control. Two-way ANOVA of treatment and year for each vegetation group detected significant effects (p < 0.05) of both treatment and year on all but the number of alien grass species. Treatment effects influenced the biomass of the different vegetation groups as well as the number of native and alien forb species. Together, the burn treatments resulted in the appearance of 18 additional native forb species, five additional alien forb species, and no new grass species. The most important of these species are dis- cussed in the text. Figures 1 and 2 summarize the effects of one, two, and three successive fall and spring burns on the relative biomass of alien grasses, alien forbs and native forbs. Under both burning regimes the biomass of alien grasses is decreased relative to that of both alien and native forbs. Tables 2-4 detail the effects of the six ex- perimental burning regimes on number of species and biomass of these three vegetative groups. Alien grasses appear to be minimally affected by the fall burning regimes. Species richness (and composition) is not inffuenced by successive fall burns (Table 2). Whereas biomass of alien grasses is 1 989] PARSONS AND STOHLGREN: GRASSLAND FIRE REGIMES 1 6 1 Table 4. Effect of Treatment by Year on Number of Species and Biomass (g/ 0.05 m-) OF Native Forbs. Only means with different letters within columns (a, b, c) or within rows (i, j, k) are significantly different at p < 0.05. Treatment codes are C = control, Fl = one fall burn, F2 = two fall burns, F3 = three fall burns, SI = one spring burn, S2 = two spring burns, S3 = three spring burns. Year Treatment 1980 1981 1982 1983 No. species c 0.6 1.2 a 0.8 a 1.6 a Fl 1.0 3.0 2.2 1.0 a F2 2.0 3.0 3.4 3.6 b F3 0.4 i 3.0j 1.8 4.2 b/j SI 1.4 3.6 b 3.2 1.6 a S2 2.6 i 5.8 b/j 4.4 b/j 2.2 i S3 1.0 i 4.4 b/j Biomass 3.2 j 3.2j C 0.1 a 0.7 a 0.37 0.5 a Fl 0.01 a/i 6.4j 0.8 i 0.4 a/i F2 0.2 a 3.1 6.2 2.5 F3 0.02 a/i 6.2 j 1.5 2.1 SI 0.6 a/i 12.4 b/j 2.1 i 0.9 i S2 0.6 b/i 5.9 j 3.3 0.8 i S3 0.1 a/i 5.3j 2.9 3.5 b decreased over pre-treatment levels by two fall burns, it is only after three such burns that biomass (11% of pre-burn) also differs from numbers found in the control plots. Decreases in abundance of Avena fatua and Bromus diandrus account for most of the biomass change. Both the number of species and biomass of alien forbs increased over pre-burn and control levels following two or more fall burns (Table 3). Alien forb biomass increased as much as 12,000% fol- lowing two fall burns (F2; 1982) and 18,000% following three fall burns (F3; 1983) over pre-burn levels in the same plots. Although an increased number of species is still evident following a year of recovery in the F2 plot, biomass has dropped markedly (Table 3). A single species, Centaurea melitensis L., which was not encountered in any plots during pre-burn sampling and only rarely found in the control plots in succeeding years, accounts for the majority of the alien forb response. Other alien forb species that increased with fall burning include Silene gallica L., Galium parisiense, and Hypo- choeris glabra L. Whereas native forbs increased in both species richness and bio- mass following fall burning, only the number of species following three successive fall burns (F3) differed from both pre-burn and control levels (Table 4). Native forb biomass increased sharply fol- 162 MADRONO [Vol. 36 lowing an initial burn and maintained moderately high levels in the following years. Lotus subpinnatus and Orthocarpus attenuatus A. Gray were the native forb species exhibiting the largest and most consistent increases following fall burning. Although up to three successive fall burns clearly influenced relative species composition and dominance, including increasing the relative importance of both native and alien forbs at the expense of alien grasses (Fig. 1) it is uncertain what fire return interval would be required to maintain such changes or whether they would revert to pre-burn levels fol- lowing a short time without fire. By the end of the study in 1983, it is only in the plot burned for three successive years (F3) that alien grasses do not dominate total biomass. Increases in the forb groups following one or two burns in the other fall treatments either have returned or have begun to approach pre-burn levels. Additional time would be required to determine if the relative suppression of alien grasses achieved with three fall burns could be maintained. Whereas total biomass tended to increase following an initial fall burn, it returned to near control levels in succeeding years. The minimal influence of repeated fall burning on total productivity counters other findings that the amount of mulch residue strongly influences productivity in California's annual grasslands (Bartolome et al. 1980). Spring burning, although probably not as important a part of historical southern Sierra Nevada foothill fire regimes as summer and fall burns, does show some potential for altering composition of annual grasslands. Total species richness was increased by suc- cessive spring burns, a difference due entirely to increased numbers of forbs. Total biomass was not influenced by spring burning (Tables 2-4). Spring burning showed little effect on alien grasses, other than a substantial but not statistically significant decrease in biomass (Table 2, Fig. 2). Species richness and biomass of alien forbs increased following two or three spring burns. Biomass (Fig. 2) increased by as much as 135-fold (from 0.1 to 8.1 g/0.05 m^) in the 1982 F3 plot (Table 3). Alien forb species, including Silene gallica L., Erodium botrys (Cav.) Bertol., and Hypochoeris glabra L., increased significantly following spring burns. An initial spring burn dramatically increased both the number of species and biomass of native forbs. However, these increases either returned to near pre-burn levels with the cessation of fire (SI and S2) or stabilized at slightly lower levels even with repeated annual burning (F3; Table 4). Trifolium microcephalum, Lupinus benthamii, and Lotus subpinnatus are the major native forbs that increased following spring burning. Increases in L. benthamii were not sus- tained following the initial burn. Table 5 lists those species that were inffuenced most by three 1989] PARSONS AND STOHLGREN: GRASSLAND FIRE REGIMES 163 ^^^^ 1980 1981 1982 1983 Year Fig. 1. Percentage of total plot biomass for alien grasses, alien forbs, and native forbs following 1, 2, and 3 successive fall burns. Data from 5 O.OS-m^ plots located in F3 (three fall bums) treatment. successive fall or spring burns. Burning in either season resulted in dramatic reductions in dominance of the omnipresent Avena fatua. This invasive European grass dominates much of the annual grass- lands throughout California. Three successive fall burns reduced A. fatua to 5.3% of the total biomass whereas three successive spring burns reduced it to 12.4% (Table 5). A second common introduced grass, Bromus diandrus, was also reduced to minimal presence (0.2 and 1.3%, respectively) by successive fall or spring burns. The third dominant grass of the unburned grassland, Bromus mollis, was not significantly affected by burning in either season, contributing be- tween 10 and 27% of the total biomass of both pre- and post-burn plots. Both the F3 and S3 treatments shifted the relative dominance of both species number and biomass from grasses to forbs. Successive fall burns resulted in a dramatic increase of the alien forb Centaurea melitensis from non-existent in 1980 to 46.3% of the total biomass in 1983. Alien and native forbs together accounted for 8.2 of the 164 MADRONO [Vol. 36 1980 1981 1982 1983 Year Fig. 2. Percentage of total plot biomass for alien grasses, alien forbs, and native forbs following 1 , 2, and 3 successive spring burns. Data from 5 0.05-m^ plots located in S3 (three spring burns) treatment. average 10.8 (76%) species per plot (as compared to 1.2 of the 4.6 or 26% in the same plots before burning and 2.8 of 5.8 or 48% of the 1983 control plots) and 7.5 of the average 9.8 g/0.05 m^ (76%) (as compared to 0.05 of 19.2 g/0.05 m^ or 0.2% before burning and 1.3 of 11.7 or 11% of the control plots). Five species of forbs (3 alien and 2 native) accounted for 71.6% of the total post-treatment biomass. The same five species were completely absent from these plots in 1980 and together accounted for only 0.3% of the total biomass in the 1983 control plots (Table 5). A total of six native and six alien forb species were encountered following the three fall burns. While three successive spring burns also increased the relative importance of forbs over grasses, the magnitude of the shift was less and the species composition different from that observed following fall burning. The five most common forbs following three successive spring burns included two native and three alien species. Together these contributed 43.6% of the total biomass (Table 5). A total of 1989] PARSONS AND STOHLGREN: GRASSLAND FIRE REGIMES 165 Table 5. Species Showing Major Change in Relative Biomass Following Three Successive Fall (F3; 1983) and/or Three Successive Spring (S3; 1983) Burns as Compared with Preburn (1980) and Control (C; 1983) Plots. Data presented as percentage of total plot biomass. Growth form codes are A = alien, N = native, F = forb, G = grass. Growth Control Preburn Postburn Ivji 111 1 y o ^ / Fall burn increasers Centauria melitensis AF 0.1 0.0 46.3 Lotus subpinnatus NF 0.0 0.0 10.8 Silene gallica AF 0.1 0.0 5.3 Hypochoeris glabra AF 0.1 0.0 4.8 Orthocarpus attenuatus NF 0.0 0.0 4.4 Fall burn decreasers Avena fatua AG 39.0 89.7 5.3 Bwmus diandrus AG 12.0 10.8 0.2 Spring burn increasers Erodium botrys AF 0.0 0.0 15.8 Trifolium microcephalum NF 0.1 0.5 11.0 Silene gallica AF 0.1 0.0 7.5 Lotus subpinnatus NF 0.0 0.0 6.9 Festuca megalura AG 0.0 0.1 6.2 Centauria melitensis AF 0.1 0.0 2.4 Spring burn decreasers Avena fatua AG 39.0 76.5 12.4 Bromus diandrus AG 12.0 12.9 1.3 five native and six alien forb species were encountered. The most common post-burn species following three successive spring fires were Erodium botrys, an alien introduced from Europe, and the native Trifolium microcephalum. Neither of these increased sub- stantially following fall burns (Table 5). Alien and native forbs to- gether accounted for 7.2 of the 1 1.2 species per plot (64%) (as op- posed to 1.8 of 4.8 or 38% before burning and 2.8 of 5.8 or 48% of the control plots) and 7.8 of the 15.9 g/0.05 m^ (49%) (compared with 0.2 of 15.2 g/0.05 m^ or 1.3% preburn and 1.3 of 1 1.7 or 1 1% of the control plots) biomass. Conclusions The annual grasslands that characterize much of the foothills of the southern Sierra Nevada are dominated by species introduced from Europe and other Mediterranean climate areas. The near com- plete dominance of these species has been largely attributed to their resistance to disturbance associated with grazing, erosion and agri- culture. The role of fire in the competitive interaction between native 166 MADRONO [Vol. 36 and alien species in California's grasslands is uncertain. In natural areas such as national parks and nature preserves there is interest in reestablishing the native herbaceous flora. The experiments re- ported here indicate that both the frequency and seasonality of fire can influence grassland species composition and biomass. Both spring and fall burning increased the total number of species. Repeated burns both decreased the relative dominance of introduced grasses and increased the diversity and dominance of native and alien forbs. Neither single nor repeated (up to three times) fall or spring burns resulted in the establishment of additional species of native grasses in the study area. This may, in part, be attributed to the fact that native grass species and thus seed sources are rare in this community. Although both fall and spring burns favored forb establishment at the expense of grasses, they had minimal effect on total biomass when maintained over three years. Fall burns tended to increase the number and biomass of alien forbs more than that of native species whereas spring burns favored both about equally. Thus, although the number and biomass of alien grasses may be reduced by regular and repeated burning (especially late in the season), both alien and native forbs will be increased. In the case of both fall and spring burning, alien grasses quickly regained their pretreatment dominance when burning was halted following one or two treatments. From a management perspective, this means that whatever gains might be realized from a program of regular burning could be quickly lost if that program were sus- pended. Frequent burning will almost certainly be needed to main- tain long-term changes if such is even possible. Acknowledgments We thank Steve DeBenedetti and Nate Stephenson for assisting with field work and preliminary data summaries. Sylvia Haultain-Tweed provided final data sum- maries. Mitch McClaran, Ray Ratliff, Steve DeBendetti, James Bartolome, and Jon Keeley critically reviewed an earlier version of the manuscript. Neil Willits provided statistical consultation. Literature Cited AscHMANN, H. 1973. Distribution and peculiarity of Mediterranean ecosystems. Pp. 1 1-20 in F. di Castri and H. A. Mooney (eds.), Mediterranean type ecosys- tems—origin and structure. Springer- Verlag, New York. Baker, G. A., P. W. Rundel, and D. J. Parsons. 1981. Ecological relationships of Quercus douglasii (Fagaceae) in the foothill zone of Sequoia National Park, Cal- ifornia. Madrofio 28:1-12. Bancroft, L., T. Nichols, D. Parsons, D. Graber, B. Evison, and J. van Wag- TENDONK. 1985. Evolution of the natural fire management program at Sequoia and Kings Canyon National Parks. Pp. 174-180 in Proceedings, Wilderness Fire Symposium. U.S.D.A. Forest Service Gen. Tech. Report INT- 182. 1989] PARSONS AND STOHLGREN: GRASSLAND FIRE REGIMES 167 Barry, W. J. 1972. California prairie ecosystems (Vol. I): the Central Valley prairie. California Department of Parks and Recreation, Sacramento. 8 1 pp. Bartolome, J. W., S. E. Klukkert, and W. J. Barry. 1986. Opal phytoliths as evidence for displacement of native Californian grassland. Madrono 33:217- 222. , M. C. Stroud, and H. F. Heady. 1980. Influence of natural mulch on forage production on differing California annual range sites. J. Range Managem. 33:4-8. Bentley, J. R. and M. W. Talbot. 1948. Annual plant vegetation of California foothills as related to range management. Ecology 29:72-79. Duncan, D. A. and R. G. Woodmansee. 1975. Forecasting forage yield from precipitation in California's annual rangeland. J. Range. Managem. 28:327-329. Garcia, D. and E. W. Lathrop. 1984. Ecological studies on the vegetation of an upland grassland (Stipa pulchra) range site in Cuyamaca Rancho State Park, San Diego County, California. Crossosoma 9(7): 5- 12. Heady, H. F. 1958. Vegetational changes in the California annual type. Ecology 39:402-416. . 1977. Valley grassland. Pp. 491-533 in M. G. Barbour and J. Major (eds.), Terrestrial vegetation of California. John Wiley and Sons, New York. Hervey, D. F. 1949. Reaction of a California annual-plant community to fire. J. Range Managem. 2:11-121. Hover, E. I. and T. B. Bragg. 1981. Effect of season of burning and mowing on an eastern Stipa-Andropogon prairie. Amer. Midi. Naturalist 105:13-18. Huntington, G. L. and M. A. Akeson. 1988. Soil resource inventory of Sequoia National Park, central part, California. Dept. Land, Air and Water Resources, Univ. California, Davis. Jackson, L. E. 1985. Ecological origins of California's Mediterranean grasses. J. Biogeogr. 12:349-361. Larson, J. R. and D. A. Duncan. 1 982. Annual grassland response to fire retardant and wildfire. J. Range Managem. 35:700-703. Macdonald, I. A. W., D. M. Graber, S. DeBenedetti, R. H. Groves, and E. R. Fuentes. 1988. Introduced species in nature reserves in Mediterranean-type climate regions of the world. Biol. Conserv. 44:37-66. McClaran, M. p. 1986. Age structure of Quercus douglasii in relation to livestock grazing and fire. Ph.D. dissertation. Univ. California, Berkeley. 1 19 pp. Parsons, D.J. 1981. The historical role of fire in the foothill communities of Sequoia National Park. Madrono 28:1 12-120. , D. M. Graber, J. K. Agee, and J. W. van Wagtendonk. 1986. Natural fire management in national parks. Environ. Managem. 10:21-24. Pitt, M. D. and H. F. Heady. 1978. Responses of annual vegetation to temperature and rainfall patterns in northern California. Ecology 59:336-350. Rundel, p. W., G. a. Baker, D. J. Parsons, and T. J. Stohlgren. 1987. Postfire demography of resprouting and seedling establishment by Adenostoma fascicu- latum in the California chaparral. Pp. 575-59 1 in J. D. Tenhunen, F. M. Catarino, O. L. Lange, and W. C. Oechel (eds.). Plant response to stress, functional analysis in Mediterranean ecosystems. Springer- Verlag, New York. SAS Institute, Inc. 1985. S AS/ST AT guide for personal computers, version 6. Cary, NC. Stocking, S. 1966. Influences of fire and sodium-calcium borate on chaparral vegetation. Madrofio 18:193-203. Stohlgren, T. J., D. J. Parsons, and P. W. Rundel. 1984. Population structure of Adenostoma fasciculatiim in mature stands of chamise chaparral in the south- ern Sierra Nevada, California. Oecologia 64:87-91. Talbot, M. W., H. H. Biswell, and A. L. Hormay. 1939. Fluctuations in the annual vegetation of California. Ecology 20:394-402. 168 MADRONO [Vol. 36 TowNE, G. and C. Owensby. 1 984. Long-term effects of annual burning at different dates in ungrazed Kansas tallgrass prairie. J. Range Managem. 37:392-397. Vankat, J. L. and J. Major. 1978. Vegetation changes in Sequoia National Park, California. J. Biogeogr. 5:377-402. Wester, L. 1981. Composition of native grasslands in the San Joaquin Valley, California. Madroiio 28:231-241. (Received 16 Dec 1988; revision accepted 11 May 1989.) CALIFORNIA BOTANICAL SOCIETY Meeting Program 1989-1990 'TLANT CONSERVATION RESEARCH NEEDS FOR THE 1990'S" 8:00 P.M. University of California, Berkeley 159 Mulford Hall DATE SPEAKER AND TOPIC 19 OCT Mr. Niall McCarten, Dept. of Integrative Biology, Univ. >„alllOrillcl, DcIKClCy "Plant extinction rates in the California flora: outlook for the future" 16 NOV Mr. Timothy Krantz, Botanical Consultant, Haward, CA "Rare and endemic plants of the Big Bear Preserve, San Bernardino County" 18 JAN Mr. James Bartel, U.S. Fish and Wildlife Service, Sac- ramento, CA "Rarity or endangerment: a call for a consensus on prior- ities" 1 7 FEB* Dr. Michael Soule, Dept. of Environmental Studies, Univ. California, Santa Cruz "A zoologist's perspective on plant conservation biology" 15 MAR Dr. Thomas Griggs, The Nature Conservancy "Restoration of Riparian Systems" 19 APR Dr. Bruce Pavlik, Dept. of Biology, Mills College, Oak- land, CA "Genetic and ecological aspects of rare plant reintroduc- tion: the case of Amsinckia gmndiflord" 17 MAY** Ms. RoxANNE BiTTMAN, Natural Diversity Data Base, Cal- if Dept. of Fish and Game, Sacramento "Plant conservation research needs for the 1990's" * Annual Banquet— location to be announced. ** Meeting to be held at University of California Botanical Garden, Strawberry Canyon, Berkeley. PHYTOGEOGRAPHICAL NOTES ON ACIDOPHILOUS CLADONIA SPECIES IN CALIFORNIA Samuel Hammer^ San Francisco State University, San Francisco, CA 94132 Abstract Several species in the lichen genus Cladonia (Ascomycotina: Lecanorales) and their restricted ranges in California are discussed. Cladonia carneola, C. phyllophora, C. cervicornis subsp. cervicornis, and C. crispata s. str. grow in small populations on azonally occurring, extremely acidic soils (pH 2.9-4.0) in Mendocino and Amador counties. The taxa at these localities are reproductively isolated from other Cladonia populations. Isolation is more marked at the Amador County sites, where only one subsection of the genus is represented. Cladonia cervicornis specimens from acidic substrata in Amador County contain the p-depside atranorin, an unusual and prim- itive chemical constituent for this species. In contrast to vascular plant species, which are represented by restricted endemic taxa at these sites, the Cladonia populations discussed in this paper belong to cosmopolitan taxa whose ranges are very restricted in California. These taxa are represented by relictual populations at the sites studied. Of 32 Cladonia taxa in California (Hammer 1988) four taxa show distinct distributional patterns on acidic soils. Two areas in northern California are noteworthy for their azonal, conspicuously acidic soil types. Several sites in Mendocino County are characterized by shal- low lateritic soils underlain by a hardpan layer that is impenetrable to plant roots. These soils and the pygmy forest that grows on them have been discussed by a number of authors (Jenny et al. 1969; Kruckeberg 1969). Unique patterns of endemism in vascular plants of the pygmy forest were noted by Mason (1946a, b) and McMillan (1956). Jenny et al. briefly mentioned the lichens of the pygmy forest and Malachowski (1975) treated the macrolichen flora of the area. The lone Formation, which is composed of outcroppings of a unique exhumed oxisol, lies in the foothills of the Sierra Nevada in Amador County. It shares the acidic properties and underlying hard- pan of the pygmy forest. Its geological history and soil characteristics were discussed in Singer and Nkedi-Kizza ( 1 980). Gankin and Major (1964) and Stebbins and Major (1965) discussed the flora of the lone Formation, focusing on the endemic plant Arctostaphylos myrtifolia C. Parry that grows in extremely restricted populations on lone Formation outcroppings. Arctostaphylos myrtifolia is taxonomically closely related to A. nummularia A. Gray, which is endemic to the ' Present address: The Farlow Herbarium, Harvard University Herbaria, 20 Di- vinity Avenue, Cambridge, MA 02138. Madrono, Vol. 36, No. 3, pp. 169-174, 1989 170 MADRONO [Vol. 36 Fig. 1. Acidophilous Cladonia taxa found in Pygmy Forest (Mendocino County) and lone Formation (Amador County) sites. pygmy forest. These narrowly restricted species are both endemic to conspicuously acidic soils in California. The lichen flora of the lone Formation was mentioned in Gankin and Major but not treated in detail. Materials and Methods Over 3000 collections of Cladonia from California were studied from fresh collections and herbarium specimens over a 2-year pe- riod. Several hundred collections of Cladonia were studied from five sites in the pygmy forest and the lone Formation. The pygmy forest sites in Mendocino County lie in or adjacent to Jackson State Forest, approximately 39°15'N, within 3 km of the Pacific coast. The two collecting sites in Amador County are approximately 30 km east of Sacramento, at approximately 38°30'N, in the western foothills of the Sierra Nevada (Fig. 1). Collection citations from both sites are in Hammer (1988). Preliminary examination of specimens was performed using a dissecting microscope and routine chemical spot tests (White and James 1985). Cladonia phyllophora Hoffm., C. crispata (Achar.) Flotow, and C. carneola (Fries) Fries were positively determined in this way. Several thalli of C. cervicornis (Achar.) Flotow subsp. cer- vicornis exhibited an equivocal reaction to the application of potas- sium hydroxide and were subsequently tested for chemical constit- 1 989] HAMMER: ACIDOPHILOUS CLADONIA IN CALIFORNIA 1 7 1 uents using thin layer chromatography (TLC). Specimens that performed unequivocally in the spot test (i.e., only fumarprotoce- traric acid detected), were not subjected to TLC. Solvent systems "A" and "C" were used (Culberson 1974; Culberson et al. 1981; White and James 1985). Thirty-one thalli of C cervicornis were tested by TLC to detect atranorin. Purified atranorin from a chemical supply company was used as a standard. Results All four taxa discussed in this paper were collected at the pygmy forest sites. All of these taxa were rare in the pygmy forest, growing on bare soil or under Arctostaphylos shrubs. Only C. cervicornis was represented by more than 10 collections at the pygmy forest local- ities. The lone Formation yielded only specimens of C phyllophora and C cervicornis. These species were observed as the dominant terricolous lichens at the lone Formation sites, where they were found growing in abundance under A. myrtifolia. Atranorin was detected in 1 6 of the 3 1 specimens of C cervicornis subsp. cervicornis that were tested by TLC. This substance was found in addition to the depsidone fumarprotocetraric acid, which is typ- ically present in C cervicornis. Atranorin was detected in most (80%) of the Amador County specimens, but was rare in specimens from Mendocino County (>10%), where it was present in only trace amounts. Discussion The four taxa of Cladonia examined in this paper demonstrate distinct distributional patterns in relation to their isolation on acidic substrata. Cladonia carneola is found at one locality outside of the pygmy forest in the Klamath Region of northwestern California. Here it comprises part of a relictual flora that is at least as old as the Pleistocene (Hammer 1989) and may be of Tertiary origins (Ra- ven and Axelrod 1978). Cladonia crispata s. str. is restricted to two sites in the pygmy forest, and is not found elsewhere in the state. Cladonia phyllophora, rare in the pygmy forest and the Klamath Region of northwestern California, is co-dominant with C. cervi- cornis subsp. cervicornis at the lone Formation sites. Cladonia cer- vicornis s. lat. is distributed widely in California, and subsp. cervi- cornis, of which a few coastal specimens contained atranorin, is present along the Pacific coast from San Diego County to Del Norte County (Fig. 1). Analysis of the chemical data on C. cervicornis from the lone Formation sites supports the hypothesis that these specimens rep- resent an ancestral population. This population has experienced long geographical and reproductive isolation from other populations in 172 MADRONO [Vol. 36 Table 1 . Cladonia Taxa Present at Pygmy Forest and Ione Formation Sites. Asterisks indicate acidophilous taxa discussed in this paper. Pygmy forest (Mendocino County) lone formation (Amador County) Sect. Cladonia Subsect. Cocciferae (Del. in Duby) Mattick C. polydactyla (Florke) Sprengel s. lat. C. transcendens (Vainio) Vainio C. macilenta Hoffm. C. hellidiflora (Achar.) Schaer. Subsect. Ochwleucae (Fries) Mattick *C. carneola (Fries) Fries Subsect. Thallostelides (Vainio) Mattick *C. phyllophora Hoffm. *C. cervicornis (Achar.) Flotow subsp. cervicomis C. cervicornis subsp. verticillata (Hoffm.) Ahti C. pyxidata (L.) Hoffm. C. chlorophaea (Florke) Sprengel C. fimhriata (L.) Fries C. subulata (L.) Wigg. Sect. Perviae (Fries) Mattick Subsect. Chasmariae (Achar.) Mattick Series Furcatae E. Dahl C. furcata (Hudson) Schrader Series Squamosae E. Dahl C. carassensis Vainio s. lat. C. squamosa (Scop.) Hoffm. var. squamosa C. squamosa (Scop.) Hoffm. var. sub- squamosa (Nyl. ex Leighton) Vainio *C. crispata (Achar.) Flotow s. str. Subsect. Thallostelides *C. phyllophora *C. cervicornis subsp. cervicornis *C. chlorophaea western North America. Atranorin was present in most specimens of C. cervicornis subsp. cervicornis from the lone Formation sites. It was rare in over 100 specimens of C cervicornis from outside of Amador County. Culberson (1986) discussed the loss of atranorin as a derived trait in the C. chlorophaea group, which is closely related to C. cervicornis (subsect. Thallostelides). The influence of gene flow in populations of lichens on their chemical constituents was dem- onstrated by Culberson et al. (1988). Cladonia cervicornis in the pygmy forest has remained in contact with contiguous populations that range along the Pacific coast of California from the Mexican border to Oregon. Gene flow and the subsequent evolution of derived chemotypes (presence of fumarprotocetraric acid only and loss of atranorin in most individuals), may be inferred for these popula- tions. The Amador County population of C. cervicornis has been geographically isolated from coastal populations. Its concomitant 1989] HAMMER: ACIDOPHILOUS CL^/)6>A^L4 IN CALIFORNIA 173 reproductive and genetic isolation may be inferred from its retained primitive chemical characteristics. Geographic isolation at the lone Formation sites has influenced the diversity of Cladonia species as well as their chemical charac- teristics. In contrast to the pygmy forest, where 1 7 species and four subsections are represented, only one subsection and three species are represented at the Amador County sites (Table 1). The pygmy forest of Mendocino County and the lone Formation of Amador County are areas where low soil pH is correlated with unique distributions of phanerogams. Cryptogams such as the lichen genus Cladonia demonstrate unique patterns of distribution in these areas as well. Vicariance and subsequent isolation on specialized soils has led to speciation and endemism in many vascular plants such as Arctostaphylos nummularia and A. myrtifolia. Many widely distributed taxa of the genus Cladonia are relictual at these localities, and possess primitive characters. Where contact with widespread populations has been maintained, as in coastal populations, derived characters may be observed. Acknowledgments I thank Dr. Harry D. Thiers, under whom this work was begun, and Dr. Nancy Carnal for their reading of an earlier version of this manuscript. Thanks to Dr. V. Thomas Parker, who sparked my interest in the lone Formation. I am grateful to Dr. John W. Thomson for verification of specimens. I thank Dr. Teuvo Ahti for his discussions concerning taxonomic concepts. I am indebted to Barbara Lachelt and Herbert Saylor for access to their collections. 1 thank the curators of CAS, FH, HSC, LAM, NY, SBM, and SFSU for their assistance. The California Botanical Society is acknowledged for its financial support in the form of a graduate student research fellowship. Dr. Donald H. Pfister is thanked for his comments regarding this manu- script. This paper was presented in part at a graduate seminar in biogeography at Harvard University. Literature Cited Culberson, C. 1974. Conditions for the use of Merck silica gel 60 F 254 plates in the standardized thin-layer chromatographic technique for lichen products. J. Chromatogr. 97:107-108. . 1986. Biogenetic relationships of the lichen substances in the framework of systematics. Bryologist 89:91-98. , W. L. Culberson, and A. Johnson. 198 1. A standardized TLC analysis of the jS-orcinol depsidones. Bryologist 84:16-29. , , and . 1988. Gene flow in lichens. Amer. J. Bot. 75:1 135- 1139. Gankin, R. and J. Major. 1964. Arctostaphylos myrtifolia, its biology and rela- tionship to the problem of endemism. Ecology 45:792-808. Hammer, S. 1988. A taxonomic survey of the lichen genus Cladonia in California. M.A. thesis, San Francisco State University, San Francisco, CA. . 1989. Cladonia carneola: two new localities in western North America. Bryologist 92:126-127. Jenny, H., R. J. Arkley, and A. M. Schultz. 1969. The pygmy forest-podosol ecosystem and its dune associates of the Mendocino coast. Madrofio 20:60-74. 174 MADRONO [Vol. 36 Kruckeberg, a. R. 1 969. Soil diversity and the distribution of plants, with examples from western North America. Madrofio 20:129-154. Malachowski, J. A. 1975. Macrolichens of the pygmy forest. M.S. thesis. Chico State University, Chico, CA. Mason, H. H. 1946a. The edaphic factor in narrow endemism I. The nature of environmental influences. Madrofio 8:209-226. . 1 946b. The edaphic factor II. The geographic distribution of plants of highly restricted patterns of distribution. Madroiio 8:241-257. McMillan, C. 1956. The edaphic restriction of Cupressus and Pinus in the coast ranges of central California. Ecol. Monogr. 26:177-212. Raven, P. H. and D. I. Axelrod. 1978. Origin of the California flora. Univ. Cal. Publ. Bot. 72:1-134. Singer, M. J. and P. Nkedi-Kizza. 1980. Properties and history of an exhumed Tertiary oxisol in California. Soil Sci. Soc. Amer. Proc. 44:587-590. Stebbins, G. L. and J. Major. 1965. Endemism and speciation in the California flora. Ecol. Monogr. 35:1-35. White, F. J. and P. W. James. 1985. A new guide to microchemical techniques for the identification of lichen substances. Brit. Lich. Soc. Bull. 75(Suppl.):l-41. (Received 19 Jan 1989; revision accepted 9 May 1989.) NOTEWORTHY COLLECTIONS Montana GooDYERA REPENS (L.) R. Br. (Orchidaceae). — Judith Basin Co., Little Belt Mts., Sandpoint Cr. of Lost Fork of Judith R., 42 km SW of Stanford, T12N RlOE SEV4 sect. 10, 1860 m, 3 Aug 1987, Phillips 870803-34 (MRC) (verified: J. S. Shelly, P. F. Stickney, MRC). Approximately 200 plants were found in a 400-square-meter area on a north-facing slope in dense shade of an old-growth Douglas-fir/lodgepole pine forest. The site is an Abies lasiocarpa/ Linnaea borealis habitat type on limestone substrate. The orchids were found in an unburned island in the center of a large 1985 burn, growing in thick mats of the mosses Drepanocladus uncinatus, Hylocomnium splendens, and Pleuroziiim schreberi (det: J. C. Elliott). Other associated species include Linnaea borealis, Pyrola secunda, Galium boreale, Clematis columbiana, Thalictnim occidentale, Smilacina stellata, Juniperus communis, Pseudotsuga men- ziesii, Abies lasiocarpa, and Picea engelmannii. Significance. Second record for MT. Previously known from a single collection (6 Aug 1980, /. DeSanto, GNP) from Glacier National Park near Upper Kintla Lake 370 km northwest of the location reported here. This site is also disjunct by 600 km from next closest known U.S. station in the Black HiUs, SD. In addition to MT, western U.S. distribution includes CO, NM, and AZ, but the species is not known from the adjacent states of WY and ID. — H. Wayne Phillips, Lewis and Clark National Forest, P.O. Box 871, Great Falls, MT 59403. A NOTE ON THE GERMINATION AND ESTABLISHMENT OF PHORADENDRON CALIFORNICUM (VISCACEAE) Job Kuijt Department of Biological Sciences, University of Lethbridge, Lethbridge, Alberta TIK 3M4, Canada Abstract During germination, the radicle of Phoradendron califomicum elongates greatly, eventually forming a minute disk or wedge from which penetration of the host is effected. Subsequently, the epicotylar pole and the entire radicle wither and die, the young plant having an exclusively endophytic existence for a brief period. All aerial shoots are formed adventitiously from endophytic strands. This pattern of establish- ment corresponds to that in Arceuthobium, but in Phoradendron is not known from other species, although a few species in Phoradendron and Viscum show transitional patterns. The stage in a mistletoe's life cycle which leads from seed ger- mination to the full establishment of the endophyte is of great bi- ological interest. It is surprising, therefore, that this stage has not been scrutinized in the case of the common desert mistletoe, Pho- radendron californicum Nutt., even though the interaction of the seedling with its host has received some notice (Cannon 1 904; Tinnin et al. 1971). The present note documents some unusual aspects of this early development. Fruits were collected on about 20 May 1 982, near Julian, San Diego County, California, the host being Acacia greggii A. Gray. The material illustrated in Figure 4 derives from the same locality and host two years earlier. The most important avian disseminator of P. californicum is the Phainopepla, Phainopepla nitens, which tends to concentrate elim- inated seeds in small clusters on the host branches near the fruiting parasite (Cowles 1936; Crouch 1943). My attention was drawn some years ago to the fact that no erect seedlings could be located even where live host branches were littered with innumerable germinating mistletoe seeds. Furthermore, those seedlings were characterized by an exceedingly long and slender radicle strikingly reminiscent of that of Arceuthobium and not at all like the type which one finds in North America and tropical America in other species of Phoradendron and in the closely related genus Dendrophthora (unpubl. obs.), where radicles tend to be short and the seedlings phanerocotylous. No seedling could be found in which the cotyledons had been withdrawn from the endosperm. Madrono, Vol. 36, No. 3, pp. 175-179, 1989 176 MADRONO [Vol. 36 Observations With such questions in mind, I closely followed germination on several potential but unproven host species {Cassia sp., Callistemon sp., Neriurn oleander L.) for up to 18 months in the University of Lethbridge greenhouse. Unfortunately, none of the numerous seed- lings became established, and the complete transition from the in- dependent stage to parasitic dependence could thus not be followed. Radicles, both in the field and in the greenhouse, are generally bright red and uniformly 0.3 mm in diameter. They often grow out in a nearly straight fashion until about 3 mm long (Fig. 1). Since most seeds become attached to the host in a sideways manner, the initial growth-direction tends to parallel the host surface. Eventually, a hinge-like movement apparently localized just outside the endo- sperm brings the tip of the rigid radicle into contact with the host, where a disk is formed (Fig. 2). The form of the disk is adapted to whatever space is available at that point, and may thus be more like a flattened wedge or peg. When the seed initially adheres so as to point towards the host, the growth pattern tends to be more regular in that the radicle advances towards the host surface. Conversely, if the seed becomes attached by the end opposite the radicle, the latter curves immediately upon emergence and grows towards the host surface. In one instance, the radicle reached a length of 7 mm and lived for 18 months (Fig. 3), but field conditions probably do not allow such records. Fortunately, young specimens were collected which unequivocally showed the subsequent development (Fig. 4). The epicotylar end, indeed, does not emerge from the endosperm, and all external por- tions shrivel up and die after penetration has occurred. In the mean- time, the endophytic strands expand into the host tissues and pro- duce a cluster of aerial shoots near the original point of entry. I have no information on how long a period passes before flowering takes place, but I assume this to be not before the second growing season. In other cases, individual shoots were noted as they emerged from a vigorous endophytic strand visible by a line of raised host bark radiating out from an older plant. The vigor and frequency of shoots which emerge are likely to be a function of the size of the host branch and perhaps host species. Discussion The young mistletoe radicle has often been described as being negatively phototropic, thus curving directly towards the host branch no matter on what side it is attached (Kuijt 1969), after which the haustorial disk is formed and the cotyledons and epicotyl are with- drawn from the endosperm. This is undoubtedly true for the great majority of mistletoes (see, for example, Kuijt 1982). It is notewor- 1989] KUIJT: PHORADENDRON GERMINATION 177 4 c m Figs. 1-3. Germination stages of Phoradendron califomicum, respectively of the following ages: 1,1, and 1 8 months. 1 . The radicle has grown out straight, parallel to the host surface. 2. The entire radicle, while still straight, has been lowered, the haustorium forming at the tip. 3. The haustorial disk is attached to the host surface, and the exhausted endosperm has been lifted by the epicotylar pole of the seedling. Germination in greenhouses. University of Lethbridge. 4. The earliest shoots formed of a vigorous, young individual of P. califomicum, collected on Acacia greggii, Julian, California. The original seedling still adheres (arrow) but is dead. thy, however, that Tubeuf (1923, p. 414) already noted an insen- sitivity of the radicle to external growth stimuli in the early ger- mination stages of Viscum album L., a situation which seems to be comparable to that in Phoradendron califomicum. Clearly, the rad- 178 MADRONO [Vol. 36 icle passes through physiologically very different stages during ger- mination. Vegetative reproduction from endophytic strands is common in many Viscaceae, including some other species of Phoradendron like P. villosum Engelm., some Viscum species (Tubeuf 1923; Kuijt 1986) and Dendrophthora (Kuijt 1987b), and the entire genus Arceutho- bium (Kuijt 1960). However, the complete replacement of the nor- mal shoot system derived from the epicotyl by adventitious shoots has not before been reported in Phoradendron. This pattern of es- tablishment has earlier been called the Arceuthobium pattern (Kuijt 1986), since it has long been known for that genus. It has recently also been shown to occur in Viscum minimum, but there appears to be still in a transitional stage from a normal pattern in that the plant does not entirely "withdraw" into the host upon entry, but rather leaves an external haustorial cushion (the disk) from which some of the early aerial shoots are formed (Kuijt 1986). The Ar- ceuthobium pattern has evolved independently in a single species of Loranthaceae {Tristerix aphyllus (DC.) Barlow & Wiens; Reiche 1 904; Mauseth et al. 1985), and is possibly present in Lepidoceras peru- vianum Kuijt of Eremolepidaceae (Kuijt 1988). It may also be ex- pected in the genus Phacellaria (Santalaceae) (Kuijt 1969). This represents at least 3 documented cases of remarkably parallel evo- lution. In one group of Phoradendron species, all of which are hyperpar- asitic on other mistletoes, a separate transitional pattern has evolved (Kuijt 1987a). While we do not know anything about the seedlings of this group, it is clear that at an early stage a haustorial cushion is formed very much like that of Viscum minimum. It is from this cushion that most (or perhaps all) shoots originate. Whether the primary shoot, and therefore the epicotyl of these species function normally remains to be established. Sprouting from the haustorial disk may, of course, be present in other species as well but is likely to be infrequent. It has not been observed in P. californicum. In summary, then, I have shown that the seedling of P. califor- nicum is cryptocotylar and of the Arceuthobium-type of development in that all aerial shoots are adventitious, the primary apical shoot meristem aborting. The "seed coat" of this mistletoe forms a shiny capsule around the living endosperm, and it is worth mentioning that the only North American member of Loranthaceae s.s. to be- come similarly adapted to deserts, Psittacanthus sonorae {S. Watson) Kuijt, has evolved a comparable sheathing capsule and cryptocotylar germination pattern (Kuijt 1973). In fact, these two mistletoes in Baja California may be found together and even in hyperparasitic combination (Kuijt 1971). The above-mentioned Viscum minimum also is a distinctively xeric species, being parasitic upon extremely succulent Euphorbia species in South Africa, as is Tristerix aphyllus 1989] KUIJT: PHORADENDRON GERMINATION 179 on several desert cacti in Chile. The desert environment thus seems to have brought about parallelisms in the pattern of establishment of diverse mistletoes. The germination pattern in Arceuthobium, similarly, may point to an early evolutionary origin in a desert en- vironment. Acknowledgments I am indebted to Mr. Wayne Armstrong, Palomar College, for assistance in the field. Financial support originated from the Natural Science and Engineering Research Council of Canada. Literature Cited Cannon, W. A. 1904. Observations on the germination Phoradendron villosum and P. califomicum. Bull. Torrey Bot. Club 31:435-443. Cowles, R. B. 1936. The relation of birds to seed dispersal of the desert mistletoe. Madrono 3:352-356. Crouch, J. E. 1 943. Distribution and habitat relationships of the Phainopepla. Auk 60:319-333. KuiJT, J. 1960. Morphological aspects of parasitism in the dwarf mistletoes {Ar- ceuthobium). Univ. Calif Publ. Bot. 30:337-436. . 1969. The biology of parasitic flowering plants. Univ. Calif Press, Berkeley & Los Angeles. . 1971. Transfer of Phrygilanthus sonorae to Psittacanthus. Madrorio 21: 13-14. . 1973. Further evidence for the systematic position oi Psittacanthus sonorae (Loranthaceae). Madroiio 22:177-185. . 1982. Seedling morphology and its systematic significance in Loranthaceae of the New World, with supplementary comments on Eremolepidaceae. Bot. Jahrb. 103:305-342. . 1 986. Observations on establishment and early shoot emergence of Viscum minimum (Viscaceae). Acta Bot. Neerl. 35:449-456. . 1987a. Novelties in Mesoamerican mistletoes (Loranthaceae and Visca- ceae). Ann. Missouri Bot. Card. 74:51 1-532. . 1987b. Miscellaneous mistletoe notes, 10-19. Brittonia 39:447-459. . 1988. Monograph of the Eremolepidaceae. Syst. Bot. Monogr. 18:1-60. Mauseth, J. D., G. Montenegro, and A. M. Walckowiak. 1985. Host infection and flower formation by the parasite Tristerix aphyllus (Loranthaceae). Can. J. Bot. 63:567-581. Reiche, K. 1904. Bau and Leben der chilenische Loranthacee Phrygilanthus aphyl- lus. Flora 93:271-297. TiNNiN, R. O., C. L. Calvin, and R. L. Null. 1971. Observations on the estab- lishment of seedlings of Phoradendron califomicum on Prosopis juliflora. Phy- tomorphology 21:313-320. TuBEUF, C. 1923. Monographic der Mistel. Oldenbourg, Berhn. (Received 24 Oct 1988; revision accepted 1 1 Apr 1989.) REDUCTION IN LIGHT REFLECTANCE OF LEAVES OF ENCELIA DENSIFOLIA (ASTERACEAE) BY TRICHOME WETTING Daniel F. Harrington and Curtis Clark ^ Biological Sciences, California State Polytechnic University, Pomona, CA 91768 Abstract Encelia densifolia is a desert shrub endemic to a single mountain range in the central western portion of the Baja California peninsula. When dry, its pubescent leaves reflect about as much light as other pubescent Encelia species, such as E. actoni and E. palmeri, although not as much as the densely tomentose E. farinosa. When the leaves are wet, their trichomes absorb water, and reflectively is decreased to a level comparable to nearly glabrous Encelia species, such as E. californica and E. frutescens. This response to wetting is likely to be advantageous to the plants; they inhabit a region of summer fogs, where light intensity and air temperature are reduced and relative humidity is increased by the same conditions that reduce leaf reflectance. Resumen Encelia densifolia es un arbusto del desierto, endemico a una sola sierra en la comarca occidental del centro de la peninsula de Baja California. Cuando secas, sus hojas pubescentes reflejan casi tanto como otras especies pubescentes de Encelia, tal como E. actoni y E. palmeri, aunque no tanto como la muy tomentosa E. farinosa. Cuando las hojas son mojadas, sus tricomas absorben el agua, y la reflectividad disminue a un nivel analogo a especies pelonas, tal como E. californica y E. frutescens. Esta respuesta a lo mojando es probable que sera ventajosa a las plantas. Habitan una comarca de nieblas estivales, donde la intensidad de la luz y la temperatura del aire se reducen y la humedad relativa se aumenta por las mismas condiciones que reducen la reflectividad de las hojas. A number of studies have been conducted in recent years con- cerning the adaptive nature of leaf pubescence. The genus Encelia has been one of the most intensely investigated, particularly E. far- inosa and the role the pubescence plays in the adaptation of this species to its desert climate (Ehleringer and Clark 1988, and refer- ences cited therein). Ehleringer and others have shown that leaf hairs are important to the success of the plants in their hot, dry habitat, because pubescence increases reflection of solar radiation, lowering leaf temperatures and rates of water loss (Ehleringer and Clark 1988). A newly described species of Encelia, E. densifolia Clark & Kyhos, is similar to other species of the genus in that a dense pubescence of multicellular uniseriate hairs covers both surfaces of the leaves ' Correspondence should be addressed to this author. Madrono, Vol. 36, No. 3, pp. 180-186, 1989 1 989] HARRINGTON AND CLARK: ENCELIA LEAVES 1 8 1 (Clark et al. 1988). Although the trichomes form a continuous cover over the surface of the leaves, we have found that the leaves are not as reflective as those of other species such as E. palmeri and E. farinosa (Ehleringer and Clark 1988). Moreover, unlike those of E. farinosa, the trichomes of E. densifolia absorb water easily and in doing so lose most of their reflective capacities. Encelia densifolia is endemic to canyons in the Picachos de Santa Clara, Baja California Sur, Mexico, where fog is common in the summer, and Clark et al. (1988) suggested that the loss of reflectivity of the leaves could be an adaptation to increase light capture on foggy days. Our purpose in this study was to compare the differences in leaf light reflectance between dry and wet leaves of Encelia densifolia, to compare these results with those of the sympatric and well-studied species E. farinosa, and finally to determine what physical features of the trichomes of these two species account for the diflerences. Materials and Methods All studies were conducted with plants grown from seed and main- tained outdoors at California State Polytechnic University, Pomona (methods follow Kyhos et al. 1981). All the seeds were obtained from plants growing in native habitats in Mexico, in the state of Baja California Sur (vouchers at CSPU): Encelia densifolia — FicsL- chos de Santa Clara, 13.6 mi NW of San Ignacio-Abreojos road at a point 24.7 mi NE of Punta Abreojos, 24 Mar 1981, accession 184, Clark 585. Encelia farinosa — Femex station on Mex Hwy 1 at San Ignacio, 23 Mar 1981, accession 183; S of Bahia Concepcion at Microondas Rosarito, 25 Mar 1981, accession 186; S end of Bahia Concepcion, near the beach, 27 Mar 1981, accession 190. We measured leaf reflectance by placing the individual leaves (adaxial side exposed) in the sample slot of the integrating sphere of a Shimadzu/Bausch & Lomb Spectronic 210 U.V. recording re- flectance spectrophotometer. Reflectances were measured at the vi- olet (425 nm) and red (670 nm) peaks of the photosynthetic action I spectrum (Jensen and Salisbury 1 984, p. 8 1 ). Magnesium oxide pow- der (Wako Pure Chemical Industries, supplied with the integrating sphere) was used as a reference. To insure accuracy of the reference, two blanks were prepared and calibrated against each other; if either deviated from the range of 99%-101% compared to the other, both were discarded, and new references were prepared. Ten ontogenetically mature leaves were selected at random from each of ten plants of Encelia densifolia during the period from 7 to 13 October, 1986, and six leaves were used from each of twelve plants ofE. densifolia on 2 February 1987. During the fall, individual leaves of E. densifolia were sometimes not large enough to cover the opening in the integrating sphere (18 mm diameter); in those 182 MADRONO [Vol. 36 cases, two leaves were overlapped, giving sample sizes less than 10. For comparison, ten leaves from each of three plants of E. farinosa were taken in 2 February. For each leaf, reflectance was measured, the leaf was wetted with deionized water, excess water was allowed to drain, and reflectance was measured again. The same wetting procedure was used to generate five reflectance spectrograms over the range of 400-700 nm. We tested for the presence of cuticle on the trichomes of both species with Sudan Black B (a histochemical stain for lipids) used on freehand sections and epidermal peels of fresh leaves. We estimated the rate of water absorption by the trichomes by dropping two drops of water in separate locations on each of four leaves of both species, and measuring the elapsed time until each drop was absorbed. To assess the eflect of cuticle removal, we per- formed three experiments. In each, leaves were dipped in 95% ace- tone for varying times, allowed to air-dry, and the absorption rate test was performed as described above. In the first experiment, four leaves of each species were dipped for 1 sec while still attached to the plants, and allowed to recover for 86.4 ksec (1 day). Encelia densifolia leaves were not used in the other two experiments, since water absorption in the first experiment was nearly instantaneous. In the second experiment, four leaves were removed from the plant, dipped in acetone for 1 sec, and measured immediately upon drying. In the third experiment, five leaves were removed from the plant, dipped for 300 sec, and measured immediately upon drying. In all cases, timing of the absorption of an individual water drop was discontinued if it had not been absorbed after 60 sec. Fog simulation was accomplished with a Fogg-It "Waterfog" Noz- zle, rated at one-half gallon per minute (31.5 ml/sec). We sprayed the fine mist upwind from the plants to mimic actual fog conditions. Effects on the leaves were assessed visually. All statistical analyses were performed with the MINITAB sta- tistical package on the California State University Central Cyber System computer. Results We found significant differences in reflectance between wet and dry leaves of Encelia densifolia. Over the entire spectrum from 400 nm to 700 nm, the reflectance of wet leaves was 50-60% of that of dry leaves (Fig. 1), whereas in E. farinosa there was no appreciable change (data not shown). At both 425 nm and 670 nm, wet E. densifolia leaves were significantly less reflective than dry leaves, for both the fall and winter samples. In contrast, there were no significant differences between wet and dry E. farinosa leaves at both wave- lengths (Fig. 2). These results also held for each individual plant of both species (two-sample t-test, p < 0.05). 1989] HARRINGTON AND CLARK: ENCELIA LEAVES 183 40-1 (D 10- 0 ~ ' ' "1 T ■ T 1 1 1 1 I I I I I 1 1 I 1 I I 1 I I 1 1 1 1 r " 1 1 1 1 400 500 600 700 Wavelength (nm) Fig. 1 . Reflectance spectra of a representative leaf of E. densifolia, before and after wetting of the trichomes. Although it was not our intent to assess seasonal changes in re- flectivity, because the plants are freely watered in cultivation, we did find differences between the October and February samples. At 425 nm, dry leaves were significantly more reflective in October (two sample t-test, p < 0.05), but there was no significant diflerence between wet leaves. At 670 nm, both wet and dry leaves were sig- nificantly more reflective in February (p < 0.01). In the wetting experiment, most areas on an untreated leaf of E. densifolia were saturated by water in under 1 sec. With untreated E. farinosa leaves, saturation generally had not occurred at 60 sec, when measurement was discontinued, and on one leaf neither water droplet had been absorbed after more than 2 ksec. Even when such long wetting times are recorded as 60 sec, there is a significant diflerence in wettability between the two species (Mann- Whitney U-test, p < 0.0001). Although cuticle of the epidermal cells of both species was stained by Sudan Black B, the trichomes of both species remained generally unstained, with only a very few of the basal cells staining. Wettability decreased in acetone-dipped leaves, however, to the extent that E. farinosa leaves treated for five minutes with acetone were, with the exception of a single leaf, just as absorbent as leaves of E. densifolia (Fig. 3). As expected, simulated fog caused an immediate reduction in leaf reflectance of Encelia densifolia; the fine water droplets were quickly absorbed by the trichomes. With E. farinosa, on the other hand, the 184 MADRONO [Vol. 36 40 n 30- 20- 10 40- 30 20 10 T ♦ i T ♦ i T ♦ i T ♦ i T ♦ i T ♦ i dry wet dry wet October February densifolla dry wet February farinosa Fig. 2. Reflectance of wet and dry leaves of E. densifolia and E. farinosa. Means are marked by diamonds, and error bars represent ± 1 SD. Means were significantly different (two-sample t-test, p < 0.001) for wet and dry E. densifolia leaves in both seasons. 1989] HARRINGTON AND CLARK: ENCELIA LEAVES 185 100.0 o o (0 c "5 10.0 o E T3 (D CO Q. LjJ Fig. 3 by the 1.0- 0.1 O — 1 — control 1* control 1* 1 densifolia farinosa duration of acetone rinse (sec) Univariate scatterplot of elapsed time until a drop of water was absorbed leaf trichomes of E. densifolia and E. farinosa (see text for explanation). timing stopped at 60 sec o o o o o 300 water droplets beaded up on the surface of the trichome layer, even- tually coalescing and running olf the leaf. Discussion When dry, the leaves of Encelia densifolia are as reflective as other pubescent-leaved Encelia species, such as E. actoni and E. palmeri. When they are wet, their reflectance decreases to a level comparable to nearly glabrous species, such as E. californica and E. frutescens (Ehleringer and Clark 1988). The well-studied E. farinosa, which is sympatric with E. densifolia, shows no changes in reflectance be- tween dry and wet leaves, and in fact the leaf trichomes are not wettable. Although histochemical staining showed no cuticle layer on the trichomes of either species, the hydrophobic nature of E. farinosa trichomes was sharply reduced by an acetone rinse. This is consistent with the removal of cutin or some other lipophilic substance. We believe these diflerences can be accounted for by the unusual environmental conditions in the habitat of E. densifolia. The region is characterized by frequent fog from the Pacific Ocean during the summer, especially in May, June, and July, the driest months of the year (Wiggins 1980). During fog conditions the leaf trichomes be- 186 MADRONO [Vol. 36 come wet and the leaves lose most of their reflectivity. This decreased reflectivity comes at a time when it is most beneficial (light intensity is reduced by the fog) and least harmful (air temperatures are reduced and relative humidity is very high). In the same environment, the leaves of E. farinosa maintain their reflectance regardless of atmospheric conditions. That E. farinosa and pubescent species of other genera thrive in the area is an indi- cation that the adaptation exhibited by E. densifolia is not a general requirement. On the other hand, E. farinosa is the most widely distributed species in the genus (Shreve and Wiggins 1964), while E. densifolia is not found outside of the fog zone. Although there are undoubtedly other factors involved in this restriction, we believe that the unique adaptation of E. densifolia helps it survive as a relict in a generally unsuitable region, rather than giving it an advantage over other desert shrub species. Acknowledgments This study represents a Senior Project carried out by D. F. H. at California State Polytechnic University, Pomona, in partial fulfillment of the requirements for the Bachelor of Science degree from California Polytechnic State University, San Luis Obispo. Research was supported in parts by grants to C. C. from the Cal Poly Kellogg Unit Foundation and the Affirmative Action Faculty Development Program. We thank L. M. Blakely and J. A. Tres for their assistance. Literature Cited Clark, C, D. W. Kyhos, and N. Charest. 1988. A new Encelia (Asteraceae: Heliantheae) from Baja California. Madroiio 34:10-15. Ehleringer, J. and C. Clark. 1988. Evolution and adaptation in Encelia (Aster- aceae). Pp. 221-248 in L. D. Gottlieb and S. K. Jain (eds.), Plant evolutionary biology. Chapman & Hall, London. Jensen, W. A. and F. B. Salisbury. 1984. Botany, 2nd ed. Wadsworth Publishing Company, Belmont, CA. Kyhos, D. W., C. Clark, and W. C. Thompson. 1981. The hybrid nature of Encelia laciniata (Compositae: Heliantheae) and control of population composition by post-dispersal selection. Syst. Bot. 6:399-411. Shreve, F. and I. L. Wiggins. 1964. Vegetation and flora of the Sonoran Desert. Stanford Univ. Press. Wiggins, L L. 1980. Flora of Baja California. Stanford Univ. Press, Stanford, CA. (Received 14 Dec 1988; revision accepted 30 May 1989.) THE ARISTIDA CALIFORNICA-GLABRATA COMPLEX (GRAMINEAE) John R. Reeder Herbarium, University of Arizona, Tucson, AZ 85721 Richard S. Felger Office of Arid Land Studies, University of Arizona, Tucson, AZ Abstract Two closely related taxa of the Aristida californica complex differ primarily in having either pubescent or glabrous culms. The pubescent form, A. californica, was described by Thurber in 1880. Eleven years later plants with glabrous culms were named A. californica var. glabra! a by Vasey. Hitchcock elevated Vasey's variety to species rank in 1924, a disposition which has been accepted generally by subsequent botanists. To determine whether these taxa merit specific rank, collections from throughout the ranges of both were analyzed utilizing five spikelet characters. Measurements of length of both glumes, lemma, awn column, and free awns revealed a high degree of overlap; for the lemma, the overlap was complete. Chromosome counts (2«=22) are reported for the first time for collections of both taxa. Even though their geographical ranges are reasonably separate, and they tend to occur at different elevations, we conclude that it is more realistic to treat them as varieties of a single species. Resumen Dos taxa cercanas, pertenecientes al complejo de Aristida californica, difieren prin- cipalmente en la superficie del culmo; una es pubescente y la otra es glabra. La forma pubescente, A. californica, fue descrita por Thurber en 1880. Once anos despues, plantas con culmos glabros fueron llamadas A. californica var. glabrata por Vasey. La ultima taxon fue elevada al nivel de especie por Hitchcock en 1924. Para determinar la categoria mas correcta, hemos analizada colecciones de la dis- tribucion entera para ambas taxa, con la base de cinco caracteres de la espiguilla. Las medidas de longitud de ambas glumas, la lemma, la columna de las aristas, y las aristas libres, revelaron un alto grado de traslapo; para la lemma este fue completo. Se reporta por primera vez el numero cromosomico (2/?=22) para las ambas taxa. Aunque sus distribuciones geograficas estan separadas razonablemente y, en general, ellas tienden a ocupar altitudes diferentes, es nuestra opinion que una mejor taxo- nomia resulta cuando estas taxa estan consideradas como dos variedades de una sola especie. Aristida californica was described by Thurber (1880), who cited collections from the Colorado Desert {Schott) and Fort Mohave {Cooper). There is no problem regarding the identity of the plant, which is described as being tufted, the culms pubescent, branched above, and with the awn column articulated with the lemma body. Madrono, Vol. 36, No. 3, pp. 187-197, 1989 188 MADRONO [Vol. 36 from which it separates at maturity. Thurber stated: "It is the only species of the section with articulated, caducous awns (Arthrather- um) thus far found in North America." This comment is not strictly correct, but Thurber can be forgiven. Although two other North American species, A. tuberculosa Nutt. (1818) and A. desmantha Trin. & Rupr. (1842), have the characteristics of the section, this fact had not been recognized when Thurber described A. calif ornica in 1880. The section Arthratherum was established by Reichenbach (1 828), based on the genus Arthratherum P. Beauv. (1812). Neither Rei- chenbach nor Trinius and Ruprecht (1842), who monographed the genus, had recognized either A. tuberculosa or A. desmantha as be- longing to the section. Moreover, Trinius and Ruprecht indicated that members of the section Arthratherum are confined to Africa, Asia, and Australia. In that work we find A. tuberculosa and A. desmantha listed under "§.I. Aristida (genuina)." [=Sect. Aristida]. Hitchcock (1924) seems to have chosen the Cooper specimen as "lectotype," indicating that it is in the U.S. National Herbarium. Among the specimens he cites, one finds: Fort Mohave, Cooper 2217. Moreover, Henrard (1926) states that a fragment of the specimen collected by Dr. Cooper "was kindly received from Prof. Hitchcock." An illustration of Aristida californica in Henrard's revision (p. 66) bears the caption: ''From cotype {Fort Mohave, Dr. Cooper 2217).'" Neither Hitchcock nor Henrard mentions having examined a Schott collection. Although Hitchcock did not specifically state he was choosing the Cooper specimen as lectotype, it is clear that he con- sidered it to be "type material." To prevent any future ambiguity regarding the type of this species, we here designate Cooper 2217 (US-81008) as lectotype of Aristida californica Thurber. Ten years after the publication of Aristida californica, Vasey (1 89 1) described a form with glabrous culms, giving it the name var. gla- brata. It was based on a collection made by T. S. Brandegee at San Jose del Cabo, Baja California Sur, Mexico, in 1890. Vasey noted that along with the glabrous culms, this taxon differs from the type [A. californica] in its larger size, more spreading and branching habit and shorter-awned flowering glumes [lemmas], yet appears to be too near for a new species. In his revision of Aristida, Hitchcock (1924) elevated Vasey's var. glabrata to the rank of species. He gave a complete description of the taxon, and noted that it differs from A. californica in the glabrous culms, shorter awn column, and longer, more densely flowered pan- icles. This transfer has been accepted, without comment, by all sub- sequent botanists. The key character has been pubescent versus gla- brous culms, although diflerences in glume and awn column length are frequently mentioned. In the same publication, Hitchcock ( 1 924) described a third mem- 1989] REEDER AND FELGER: ARISTIDA 189 ber of Sect. Arthmtherum, A. peninsularis, based on Palmer 501 collected in Nov 1887 at Bahia de Los Angeles, Baja California Norte, Mexico. It was said to have pubescent culms like A. califor- nica, but to differ in being annual, and having larger glumes, lemmas, and awns. Henrard (1926) accepted this species without expressed reservations, as have most other botanists. The sole exception is Gould and Moran (1981), who treat it as a synonym of A. californica. They comment that since no distinctly annual plants have been found (in the type locality), A. peninsularis must be only an annual appearing form of A. californica. We quite agree with this conclusion; all plants of this complex, from any area, are apparently perennial, although they may flower during their first year of growth. Regarding Hitchcock's statement that the spikelet parts are larger in his A. peninsularis, with the exception of the lemma (which he gives as "about 8 mm"), we found his measurements fall well within the range we determined for^. californica. We examined several spec- imens from the type locality and none had lemmas longer than 7 mm. We did not see the type. The present study addresses the question of whether or not the two taxa, Aristida californica and ^4. glabrata, are sufficiently distinct to merit specific designation. Vegetatively, the plants exhibit no important differences except for the culms, which are completely glabrous or variously pubescent. The indument, when present, may be rather long and somewhat matted, or extremely short. Various gradations between these two extremes are encountered when a large suite of specimens is examined. Some culms which are devoid of hair may be quite scabrous and, therefore differ but slightly in ap- pearance from others with very short hairs. Nevertheless, one ex- periences little difficulty in scoring plants as either "glabrous" or "pubescent," and this reasonably consistent feature has been the key character used to separate the two taxa. The basis for naming her- barium specimens also, obviously has been this one feature, and little importance seems to have been placed on whether the hairs are long or short. Along with the pubescent or glabrous culms, au- thors frequently indicate slight differences in lengths of glumes, and especially the awn column, and free awns. Thus Hitchcock (1951) states that the awn column in A. californica is 15 to 20 mm long in contrast to that in A. glabrata which he indicates is 6 to 14 mm. In Gould and Moran (1981) the awn length in A. californica is said to be 2.5 to 4.5 cm versus 1.5 to 4 cm in A. glabrata. Methods At ARIZ there is a rather large collection representing this com- plex, which includes gatherings made by both authors over a period of years. In addition, specimens were borrowed from ASU, CAS- 190 MADRONO [Vol. 36 DS, RSA-POM, SD, and UC-JEPS. Our sample consisted of 145 plants with some pubescence on the culms, and 84 in which the culms were completely glabrous. Because plants of the two taxa are very similar vegetatively, we focused on the spikelet: glumes, lemma, awn column, and free awns. To determine the amount of variability in size of spikelet parts on a single individual, we measured ten spikelets on each of four specimens— two with pubescent culms and two with glabrous culms. Care was taken to select spikelets which we judged to be mature, as evidenced by fully developed caryopses within the lemma. Summary statistics were calculated for the 229 specimens, using mean, standard deviation, standard error of the mean, and skewness. Some field work was carried out especially for this study, primarily to gain more information on range and to collect cytological material. Results Among the specimens borrowed from UC-JEPS we found two types. One of these was Cooper s.n. collected at Fort Mohave in 1860-61 (UC-37301). This is one of the two collections cited by Thurber in his original description of Aristida californica and is thus a syntype; the specimen at US was designated as lectotype earlier in this paper. UC-3730 1 is, therefore, an isolectotype of^. californica Thurb. This specimen has moderately "woolly" culms, but the hairs are not as dense and long as in many other examples of this species. The first glume measures 7-9 mm, the second ca. 15. The lemma is 6.5-7 mm, with a column 14-15 mm, and the awns are 3-4 cm in length. An apparent isotype of Aristida californica var. glabrata Vasey is UC- 12 1421. This specimen was collected by T. S. Brandegee {s.n.) at San Jose del Cabo, Baja Californica Sur, Mexico, 1 Oct 1890. [The Chase Index cites this collection as Brandegee 34.] The culms of the UC plant are completely glabrous. The first glume measures 5 mm in length; the second ca. 10 mm. The lemma is 6-6.5 mm, with a column ca. 8 mm, and awns only ca. 2 cm. Clearly, plants representing the types of these two taxa exhibit other differences besides pubescent versus glabrous culms. In var. glabrata the glumes are somewhat shorter, and this is also true for the awn column and the spreading awns. Table 1 summarizes the results obtained from measuring spikelet parts on a single individual; Table 2 presents the data from all specimens studied, and this information is also shown in graphic form in Figure 1. Chromosomes Since we found no published chromosome numbers for members of this complex, along with collecting herbarium specimens, one of 1989] REEDER AND FELGER: ARISTIDA 191 LEM CP CG Fig. 1 . Summary of measurements in mm of glume 1 , glume 2, lemma, awn column, and free awn for the two taxa of the Aristida californica complex. Plants with pubescent culms (CP); those with glabrous culms (CG). The diagrams illustrate the range, mean, and the mean ± 1 standard deviation. US (Reeder) preserved young inflorescences in the standard 3:1 ab- solute alcohol-acetic acid mixture for cytological examination. The anthers were later squashed in aceto-carmine. Whether or not the plants had pubescent or glabrous culms, all were determined to be diploid (2^=22). The collections are listed below. Collection num- bers are those of John R. and Charlotte G. Reeder. Vouchers are at ARIZ. Aristida californica Thurb. var. californica: USA. AZ: Pima Co., Cabeza Prieta Natl. Wildlife Refuge, fsO m, 5 Mar 1977, 683 5\ CA: Imperial Co., 20 km W of Glamis along CA-78, sea level, 1 Oct 1987, 8160. Table 1 . Summary of Ten Measurements for Each of Five Spikelet Characters ON Four Different Specimens of the Aristida californica Complex. Measure- ments in mm. All specimens are at ARIZ. Culms pubescent Culms glabrous Reeder 8238 Felger 16712 Reeder 7221 Wiggins 7239 AZ, YUMA MEX, SON MEX, BCN MEX, SON 1st glume 7.0-8.5 6.5-8.0 5.5-8.0 7.0-9.2 2nd glume 14.0-16.0 13.0-15.0 11.5-13.5 13.0-13.5 Lemma 6.0-6.5 5.0-6.0 5.0-6.3 6.2-7.0 Awn column 11.5-16.0 12.0-18.0 12.0-16.0 12.0-16.0 Free awn 35.0-40.0 40.0-45.0 22.0-27.0 32.0-38.0 192 MADRONO [Vol. 36 Table 2. Summary Statistics for the Five Key Characters of the Aristida CALiFORNiCA COMPLEX. Plants with pubescent culms (CP), n = 145; those with gla- brous culms (CG), n = 84. Measurements in mm. Character Taxon Mean SD SE Min. Max. Skewness 1st glume CP 7.50 0.835 0.069 6.0 10.0 0.466 CG 6.63 0.912 0.100 4.0 9.2 -0.348 2nd glume CP 14.34 1.506 0.125 12.0 23.0 1.814 CG 12.24 1.402 0.153 8.0 15.0 -0.298 Lemma CP 6.06 0.537 0.045 5.0 7.0 0.016 CG 6.20 0.567 0.063 5.0 7.0 -0.284 Awn Col. CP 14.31 3.250 0.270 8.0 26.0 0.768 CG 10.60 2.760 0.301 4.0 16.0 -0.138 Awn (free) CP 35.12 5.644 0.469 25.0 50.0 0.058 CG 26.07 6.211 0.678 13.0 40.0 -0.304 Aristida californica Thurb. var. glabrata Vasey: USA. AZ: Cochise Co., on Willcox Rd. just E of jet. with Caseabel Rd., 975 m, 5 Jun 1987, 804 1\ Pima Co., Santa Rita Range Reserve along Box Canyon Rd., 1310 m, 29 May 1987, 803 5\ E end of Rincon Mts, 1220 m, 4 Jun 1987, 8038\ Pinal Co., Pinal Pioneer Pkwy., 3 km S of Bradley Wash, 1310 m, 3 Oct 1986, 7994\ Yavapai Co., along US-93, 3.5 km NW of Santa Maria River crossing, 550 m, 2 Oct 1987, 8162\ along US-93, ca. 4.5 km NW of jet. with US-89, 670 m, 2 Oct 1987, 8163. Discussion Table 1 needs little explanation. It reveals, as we suspected, that there is a considerable amount of variation in lengths of spikelet parts on a single individual. This was taken into consideration as measurements were made on the entire suite of specimens examined in the study. For each sample numerous spikelets were measured in order to arrive at a value which seemed representative. Perusal of Table 2 reveals that in each case there is a considerable amount of overlap in the measurements of the spikelet parts; with the lemma, this overlap is complete. With respect to glume I, we determined that 89% of CP and 88% of CG plants fell in the zone of overlap, which is 6-8 mm. Glume II had 99% of CP and 77% of CG plants in the overlap zone of 12-18 mm. The awn column showed 79% of CP and 88% of CG plants in the overlap zone of 8- 16 mm. Finally, the spreading awns, with an overlap zone of 25-40 mm, revealed a similar picture: 90% of CP and 65% of CG plants fell in this zone. It is evident that lengths of these spikelet features are not good key characters to use in separating species. For each feature (except the lemma) the longest measurement is from a plant 1989] REEDER AND FELGER: ARISTIDA 193 with pubescent culms, whereas the shortest is from one in which the culms are glabrous. Ecology and Geographical Distribution The pubescent taxon occurs in southeastern California, Baja Cal- ifornia Norte and Sur, southwestern Arizona, and in Sonora south- ward along coastal dunes to Bahia Colorado (south of Tastiota) and farther south from near the mouth of the Rio Mayo to Topolobampo in northwestern Sinaloa (Fig. 2). This is one of the few grasses re- ported to be endemic to the Sonoran Desert, but, as the collections from Sinaloa indicate, it actually occurs south of the Sonoran Desert. It is also in the Mojave Desert of California. Densely pubescent populations inhabit sandy soils, flats and dunes, in the lowland regions of the northern part of the Sonoran Desert of northwestern Sonora, extreme southwestern Arizona, southeast- ern California, and northeastern Baja California Norte. This area was designated by Shreve (1951) as the Lower Colorado Valley phytogeographic region of the Sonoran Desert. Here rainfall is largely winter-spring, almost entirely so in the western part, and this is also true for the Mojave Desert of southern California. There may be some summer rainfall in the more southern and eastern portion of the Lower Colorado Valley, although precipitation in this extremely arid region is unpredictable (Ezcurra and Rodriguez 1986). The pubescent form also occurs southward into regions of rela- tively greater and more predictable rainfall in the Baja California peninsula and in southern Sonora and northwestern Sinaloa consid- erably south of Guaymas, the usually accepted boundary of the Sonoran Desert (Felger and Lowe 1976). The southern Sonora and Sinaloa populations are disjuncts, presently known from coastal dunes in the delta regions of the Rio Mayo and Rio Fuerte (we predict that it should also occur in coastal dunes in the delta region of the Rio Yaqui). Plants from these southern populations are not as densely pubescent as are those from the Lower Colorado Valley and Mojave Desert. Aristida glahrata generally occupies regions peripheral to that of A. californica, in areas of higher elevation (Fig. 3) and/or higher precipitation, and mostly where summer-fall rainfall is greater and more predictable. In Arizona and Sonora A. glabrata extends into grassland and chaparral well above the desert. In addition, this species often occurs on soils which are more rocky and gravelly than those occupied by A. californica. Throughout their distribution the two taxa are essentially allopatric, but in a few cases they have been collected in close proximity. Both have been found in Baja California Norte, in the San Matias Pass, at only slightly different elevations [Reeder and Reeder 7221, 700 m, A. glahrata (ARIZ); and Thome 194 MADRONO [Vol. 36 Fig. 2. Map showing the range for the two varieties of Aristida californica: var. californica (solid circles); var. glabrata (open circles). Triangles represent type local- ities. and Charlton 60184, 600 m, A. californica (RSA)]. Another some- what similar situation is encountered in the vicinity of La Paz, Baja CaUfornia Sur. There are several collections of A. glabrata from that region, and we have seen one gathering of ^. californica from this 1989] REEDER AND FELGER: ARISTIDA 195 Sea Level • • • • • • • • • • • • • • o • o — • o o 'o • o o 8b p ° 0 o 00----Q. -- ""—^ -XX — ca- - cCOcR--o -- O OO .0-------0 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Latitude Fig. 3. Elevational distribution for Aristida californica var. californica (solid circles) and var. glabrata (open circles). same area [Carter 2726 (UC)]. Both taxa have also been found on Tiburon Island in the Gulf of California. Aristida californica [Felger and Russell 7009\ Felger et al. 12537 (ARIZ)] is common on the beach dunes of the arid eastern and northern coasts (Felger and Lowe 1976); one collection of .4. glabrata [Felger et al. 17351 (ARIZ)] was made from the large interior central valley, a considerable dis- tance inland, at an elevation of 120 m. The interior of this large island supports a much denser, less xerophytic vegetation and un- doubtedly has higher rainfall than does the arid coast (Felger 1966; Felger and Lowe 1976). A noteworthy gathering of A. glabrata is Wiggins 1 7335 (RSA) from Isla San Marcos, situated near the eastern coast of Baja Californica Sur slightly south of Santa Rosalia. All collections of this complex on the Baja California peninsula between the San Pedro Martir in the north and La Paz in the south have pubescent culms {A. californica). Within the Pinacate Region and Gran Desierto of northwestern Sonora, one sees a trend of more densely pubescent plants in more xeric habitats. Populations farther inland in the Gran Desierto tend to have stems more densely white-pubescent than those growing under the more equable conditions along the coast. It seems apparent, therefore, that the distribution of pubescent and non-pubescent forms correlates with vegetational and ecological factors. Aristida glabrata occupies regions of more predictable and higher precipitation, especially summer-fall rainfall, and is replaced by A. californica in hotter, drier climates with cool season rains. Presuming the Sonoran Desert to be more recent than thornscrub, one may speculate that the origin of pubescent forms pre-dates the Sonoran Desert. Such plants may have evolved as coastal dune- 196 MADRONO [Vol. 36 adapted plants along the great river deltas of subtropical scrub re- gions of northwestern Sinaloa and southwestern Sonora. Summary The two taxa, Aristida calif or nica and A. glabrata, are clearly closely related, and have the same chromosome number {ln=22). Vegetatively they are very similar, with essentially the same growth habit. Measurements of the spikelet parts show a high degree of overlap, which is complete with respect to the lemma. Nevertheless, the measurements are skewed; the longest for each character (except the lemma) was always found on plants with pubescent culms. Even though their geographical ranges are reasonably separate, and they tend to occur at different elevations, we conclude that it is more realistic to treat them as varieties of a single species. Some might prefer to designate them as subspecies, but we make no distinction between subspecies and variety when only one infraspecific level is recognized. Acknowledgments We are indebted to the curators at ASU, CAS-DS, RSA-POM, SD, and UC-JEPS for the loan of herbarium specimens. Grateful appreciation is also extended to Janice Bowers and Tony Burgess for assistance in computer mapping; to J. Mark Porter and L. J. Toolin for helpful discussions; to Charlotte Reeder for assistance with the literature and critically reading the manuscript; and to C. T. Mason, Jr., for providing space at ARIZ and arranging for loans from other herbaria. Literature Cited EzcuRRA, E. and V. Rodriguez. 1986. Rainfall patterns in the Gran Desierto, Sonora, Mexico. J. Arid Environ. 10:13-28. Felger, R. S. 1966. Ecology of the gulf coast and islands of Sonora, Mexico. Ph.D. dissertation. Univ. Arizona. 570 pp. Felger, R. S. and C. H. Lowe. 1976. The island and coastal vegetation and flora of the northern part of the Gulf of California. Los Angeles County Mus. Contr. Sci. 285. 59 pp. Gould, F. W. and R. Moran. 1981. The grasses of Baja California, Mexico. Mem. San Diego Soc. Nat. Hist. 12. 140 pp. Henrard, J. T. 1926. A critical revision of the genus Aristida being a preliminary study and an introduction to the Monograph, vol. 1. Meded. Rijks Herb. No. 54, Leiden. VIII + 220 pp. Hitchcock, A. S. 1924. The North American species Aristida. Contr. U.S. Natl. Herb. 22:517-586 + index. . 1951. Manual of the grasses of the United States. (Ed. 2, revised by A. Chase.) U.S. Dept. Agric. Misc. Publ. 200. 1051 pp. NuTTALL, T. 1818. The genera of North American plants, and a catalogue of the species, to the year 1817. Printed for the author by D. Heartt, Philadelphia. Vol. 1, viii + 312; Vol. 2, 254 pp. + index and erratum, additions. Palisot de Beauvois, a. M. F. J. 1812. Essai d'une nouvelle Agrostographie; ou nouveaux genres des Graminees. Paris. Ixxiv + 182 pp. + xxv plates. Reichenbach, H. G. L. 1828. Conspectus Regni Vegetabilis per Gradus Naturales Evoluti. xiv + 294 pp. Lipsiae. 1989] REEDER AND FELGER: ARISTIDA 197 Shreve, F. 1951. Vegetation of the Sonoran Desert. Publ. Carnegie Inst. Washington 591:i-xii, 1-192. Reprinted in F. Shreve and I. L. Wiggins, Vegetation and Flora of the Sonoran Desert. Stanford Univ. Press. 1964. Thurber, G. 1880. Gramineae. Pp. 253-328 in S. Watson, Botany [of California] vol. 2. John Wilson and Son, Boston. Trinius, C. B. and F. J. Ruprecht. 1 842. Gramina Agrostidea. III. Callus obconicus (Stipacea). Mem. Acad. Imp. Sci. St.-Petersbourg Ser. 6, Sci. Nat. 7(2): 1-1 89. [1843]. Preprinted as Species Graminum Stipaceorum (1842). Vasey, G. 1891. Gramineae. Pp. 177-179 in T. S. Brandegee, Flora of the Cape Region of Baja California. Proc. Calif. Acad. Sci. II. 3:108-182. (Received 12 Jan 1989; revision accepted 11 May 1989.) ANNOUNCEIVIENT The 1989 Jesse M. Greenman Award The 1 989 Jesse M. Greenman Award has been won by Carol A. Todzia for her publication "Chloranthaceae: Hedyosmum,'' which appeared in Flora Neotropica Monographs, volume 48. This monograph is derived from a Ph.D. dissertation submitted to the University of Texas, under the direction of Dr. Beryl B. Simpson. The genus Hedyosmum is com- prised of 40 species of predominantly montane, neotropical shrubs and trees. The comprehensive monograph, which includes four newly de- scribed species, reexamines previous treatments of the genus and pre- sents new data on anatomy, morphology, ecology, and geography. Syn- opses of the taxonomic history, palynology, cytology, and uses are also provided. This Award is named for Jesse More Greenman (1867-1951), who was Curator of the Missouri Botanical Garden Herbarium from 1919 until 1943. A cash prize of $500 is presented each year by the Garden, recognizing the paper judged best in vascular plant or bryophyte sys- tematics based on a doctoral dissertation published during the previous year. Nominations for papers published during 1989 are now being ac- cepted for the 22nd annual award, which will be presented in the summer of 1990. Reprints of such papers should be sent to: Greenman Award Committee, Research Division, Missouri Botanical Garden, P.O. Box 299, St. Louis, MO 63166-0299, U.S.A. In order to be considered for the 1990 award, reprints must be received by 1 June 1990. A NEW SPECIES OF SIPHONOGLOSSA (ACANTHACEAE) AND SOME INFRAGENERIC TRANSFERS Richard A. Hilsenbeck' Department of Botany, The University of Texas at Austin, Austin, TX Abstract Siphonoglossa mexicana (Acanthaceae), a new species from Mexico, is described and illustrated. The new species is most closely related to S. ramosa and S. canbyi of Siphonoglossa sect. Siphonoglossa. Characters, variation patterns, and reproductive biology of the species are discussed with direct reference to an earlier treatment that combined 5*. ramosa, S. discolor, and S. hondurensis under the name S. sessilis. Siphonoglossa ramosa and 5". sessilis are retained as distinct species, and the new combinations, S. ramosa var. discolor and 5". ramosa var. hondurensis, are proposed. Additionally, S. calcarea of northern Colombia is reduced to a variety of S. sessilis. Resumen Siphonoglossa mexicana Acanthaceae, una nueva especie de Mexico, es descrita y ilustrada. La nueva especie parece estar mas relacionada a S. ramosa y S. canbyi de Siphonoglossa seccion Siphonoglossa. Characteres, modelos de variacion y biologia reproductiva de la especie son discutidas con referenda a un tratamiento anterior que combina S. ramosa, S. discolor y S. hondurensis bajo el nombre S. sessilis. Siphonoglossa ramosa y S. sessilis se mantienen como especies distinctas, y las nuevas combinaciones, S. ramoso var. discolor y S. ramosa var. hondurensis, son propuestas. Adicionalmente, S. calcarea de el norte de Colombia se reduce a variedad de .S. sessilis. In connection with a monographic treatment and chemosyste- matic investigation of the genus Siphonoglossa Oerst. (Acanthaceae), I have discovered several new taxa. Two of these new species, S. durangensis (sect. Siphonoglossa) and S. linearifolia (sect. Pentaloba) were previously described by Henrickson and Hilsenbeck (1979). A third new species is proposed here. Siphonoglossa mexicana Hilsenbeck, sp. nov. (Fig. 1).— Type: MEX- ICO, Sinaloa, Imala, shady wooded ravine in valley, 500 ft, 29 Nov 1939, H. S. Gentry 5099 (holotype, CAS; isotypes, GH, MO, NY, US). A speciebus affinibus caulibus saepe brunneis, pilis caulium con- fertis brevibus uniformibus 0. 1-0.3 mm longis; foliis tenuibus mem- ' Present address: Department of Biology, Sul Ross State University, Alpine, TX 79832. Madrono, Vol. 36, No. 3, pp. 198-207, 1989 1989] HILSENBECK: SIPHONOGLOSSA 199 2 cm Fig. 1. Siphonoglossa mexicana Hilsenbeck. A. Branch showing flowering spikes and detail of stem and bract pubescence. B. Corolla showing characteristic long tube and "crowfoot" pattern on anterior lip. C. Anther, posterior view. D. Open fruit showing retinacula. E. Seeds showing characteristic muricate encrustations of the testa. branaceis; floribus oppositis (raro alternis) in spicis dispositis, brac- teis (3.5-)4-6 mm longis hispidis trichomatibus pro parte simplicibus pro parte glandularibus confertis sparsisve; fructis 7-9(-ll) mm longis; seminibus ca. 2 x 2 mm, testis incrustatis incrustationibus imbricatis proximaliter acutis (0.2-)0. 3-0.4 mm latis differt. Erect, ascending or clambering perennial herb to subshrub, branched above, (3-)5-10 dm high, from a rhizomatous rootstock; 200 MADRONO [Vol. 36 Stems terete, striate, greenish to brownish above, covered with a short, even pubescence of erect to recurved hairs 0.1-0.2(-0.3) mm long, usually dark brown and becoming woody below. Leaf blades thin, membranaceous, lance-ovate to ovate, (2.5-)4-6(-7.5) cm long, (1. 5-) 2-3. 5 (-4) cm wide; tip acute to long acuminate; base often slightly oblique becoming attenuate to a petiole (4-)8-10(-15) mm long; margin entire to slightly repand with hairs to 0.5 mm, olive green above, paler and duller beneath, densely lineolate, nearly gla- brous above except along veins with scattered hairs to 0.2 mm, hispid beneath with scattered hairs 0.4-0.7 mm long. Inflorescence of nar- row terminal and axillary spikes; flowers usually opposite, nearly sessile; bracts linear lanceolate, (2.5-)4-6 mm long, 0.6-1 mm wide at base, hispid with mixed simple and glandular trichomes; brac- teoles linear to subulate, 3-3.5 mm long, hispid. Calyx deeply 4- parted, 4.5-5.5 mm long; segments lanceolate 4-5 mm long (oc- casionally with a greatly reduced fifth posterior segment to 1 mm long), hispid and usually densely covered with glandular trichomes. Corolla pink to pale purple, zygomorphic, (19-)22-28 mm long; posterior lip entire, erect to slightly recurved, 4.5-6 mm long, 1.5- 2 mm wide; anterior lip 3-lobed, spreading, the middle lobe 5.5-8 mm long, 2.5-4 mm wide with a purple to reddish "crow-foot" pattern on the wrinkled palate at the narrow throat, the lateral lobes 5- 7 mm long, 2-3.5 mm wide; tube (13-) 15-20 mm long, terete, pubescent with scattered hairs on exterior surface. Style 16-23 mm long, slightly exserted from beneath posterior lip; stigma linear, pos- terior lobe greatly reduced. Stamens 2, barely exserted; filaments 2.5-3.5 mm long, inserted near base of lateral lobes; anthers 2-celled, 1.2-1.5 mm long, the anther sacs slightly superposed and oblique, the upper sac 1.1-1.2 mm long, the lower sac 0.9-1 mm long, the bases blunt; connective 0.3-0.4 mm wide. Fruit a clavate, medially constricted capsule with elliptic head, 7-9(-12) mm long, with the base a solid stripe 2-4 mm long, light brown or tan, often glandular pubescent; seeds 4, (1.8-)2 x 2(-2.2) mm, light yellowish-tan prior to maturity, dark brown when mature, muricate, the irregular and pointed encrustations (0.2-)0.3-0.4 mm in diameter, forming some- what overlapping plates. Chromosome number n=ll (Hilsenbeck 790, 797). Paratypes: MEXICO, Guerrero, Distr. Mina, Puerto de Oro, 500 m, 9 Nov 1936, Hinton 9831 (GH, K, LL, MICH, TEX); 1 km al N de Xalitla, 740 m, 18 Nov 1975, Lopez H. s.n. (ENCB). Jalisco, ca. 5 mi N ofTomatlan, ca. 200 ft, 17 Mar 1982, Daniel 2079 {TEX). Morelos, Pedregal de Cuemavaca, 24 Dec 1960, Paray 3137 (ENCB); caiion de Lobos, 1270 m, 4 Dec 1970, Vazquez 2782 (MEXU). Puebla, ca. 8 mi SE of Izucar de Matamoros on Mexico Hwy 190, 1 100 m, 24 Nov 1980, Hilsenbeck 797 (TEX); 2 km SE de Petlan- cingo, sobre la carretera a Huajuapan, 1450 m, 29 Nov 1972, /. 1989] HILSENBECK: SIPHONOGLOSSA 201 Rzedowski 28928 (ENCB); ca. 20 km SE of Izucar de Matamoros, 1150 m, 18 Feb 1965, McVaugh 22485 (ENCB, MICH). Veracruz, ca. 4 km off Mexico Hwy 140, E of Palo Gacho, on rd to Actopan, ca. 200 m, 22 Nov 1980, Hilsenbeck 790 (TEX); Actopan, 150 m, 5 Nov 1970, Ventura A. 2761 (ENCB); Plan del Rio, 300 m, 10 Sep 1974, Ventura A. 10510 (ENCB); Sinaloa, Ymala, 25 Sep-8 Oct 1891, Palmer 1712 (US). No locality (but probably Guerrero), 1791, Haenke 988 (F). Siphonoglossa mexicana has a 4-parted calyx, flowers disposed in spikes with subulate bracts and bracteoles, an entire posterior corolla lip, blunt anther sacs, and ellipsoidal capsule heads. These features characterize sect. Siphonoglossa and, therefore, this species clearly belongs in the type section. The new species is related most closely to two other Mexican species: S. ramosa Oersted, the type species of the genus found primarily in Puebla and Oaxaca of southern Mexico, and S. canbyi (Greenman) Hilsenbeck of Tamaulipas and Nuevo Leon in northeastern Mexico. Siphonoglossa mexicana has been misidentified consistently as S. ramosa or S. canbyi. Because Siphonoglossa had never been monographed and no keys have been written, uncertainty as to what characters delimited these species has existed until now (Table 1). Siphonoglossa mexicana is the most widespread species of the genus in Mexico. It extends from the state of Veracruz on the Gulf Coast westward across the northern edge of the Isthmus of Tehuan- tepec and then northward through Guerrero and Michoacan to Si- naloa. This species is somewhat variable throughout its range and is therefore difficult to accurately identify, but can be distinguished from other taxa of Siphonoglossa by the combination of: 1) thin, membranaceous, often long-petiolate leaves; 2) flowers opposite on the spikes with bracts mostly 4.5-6 mm long; 3) bracts and calyx hispid pubescent often with a dense covering of glandular trichomes; 4) usually smaller (7-9 mm) fruit; and 5) seeds with muricate, over- lapping encrustations (0.2-)0. 3-0.4 mm wide. In some populations of S. mexicana, the flowers become somewhat alternately arranged in the spike and the glandular trichomes on the inflorescence become sparse, but the other characters cited serve to distinguish this species. Additional morphological (Hilsenbeck 1983) and chemical data (Hilsenbeck unpubl.), as well as recent collections primarily from western Mexico (Daniel pers. comm.), strongly suggest that S. mex- icana intergrades with S. ramosa var. ramosa in Oaxaca and Puebla. Furthermore, these same data suggest that S. mexicana also inter- grades extensively with S. canbyi in southern Veracruz and across the northern Isthmus of Tehuantepec into Guerrero with some pop- ulations of S. mexicana even as far north as Sinaloa exhibiting variation in pubescence, bracts, capsule, and seed morphology in the direction of S. canbyi. The significance and extent of this vari- 202 MADRONO Q Z < CO CO i Co O I CO z ^ H U CQ c/D U U z -J pa < Co CO Co d I d > l-l ;3 CJ 5 3 — o 00 00 aj c c <^ - 00 o - O c O C CJ ■*-» •l-H "ob t/5 1/5 ^ e £ .X c u ^ ^ ■'3 d c/f (U ■<— > 00 c s s J2 o is c« — ' 00 a d 00 c • T-H c« CM ^ . 3 1/3 (/5 OO (U £ +- > , S.£ £6 £ S £ Ci'^ 00 ^ o £ O C (U +i- c« r; c !_ frt ON 1 :3 -S +-> (U UJ £ •1— > C/5 .2 ^ -I— • (U -5 Oct 1951, Williams 18320 (¥)\ Morazan: vicinity of Suyapa, region of La Tra- vesia, 1 100-1200 m, Sep-Dec, Standley 12453 (F); near San Fran- cisco, 800 m, 21 Aug 1949, Standley 22988 (F, NY); along Rio Yeguare near San Francisco, 850 m, 21 Aug 1949, Williams 15912 (F); along Rio Yeguare, E of El Zamorano, 850 m, 10 and 15 Dec 1946, P. C. Standley 1090 (F); near El Jicarito along road toward El Pedregal, ca. 900 m, 6 Ju 1949, Standley 20845 (F); mountains above El Jicarito, near Rio Caparrosa, 900-1300 m, 26 May 1951, Standley 28492 (F); Margenes de la Quebrada de las Burras entre Suyapa'y Tegucigalpa, 1050 m, 11 Dec 1948, Molina R. 1836 (F, GH); vicinity of Suyapa, 1100-1200 m, Sep-Dec 1948, Standley 15394 (F); Dpto. de Francisco, Morazan, 1 Aug 1975, Garcia 104 (MO); Olancho: vicinity of Catacamas, 450-500 m, 1 8-26 Mar 1 949, Standley 18432 (F); vicinity of Catacamas, 450-500 m, 18-26 Mar 1949, Standley 18408 (F). Variety hondurensis may be distinguished from var. ramosa by a combination of the dark brown stems, the more ovately-lanceolate leaves, and the usually shorter corollas. In the holotype, an inflo- rescence is not readily apparent, but the beginning (or vestige?) of a spike can be seen at two places on the specimen. Almost all of the other specimens seen have several conspicuous flowering spikes. Leonard (1958) described Siphonoglossa calcarea based on a single collection from La Paz, Department of Magdalena, in extreme north- eastern Colombia. This taxon is clearly conspecific with S. sessilis; almost every qualitative and quantitative character of S. calcarea falls within the range of variation of S. sessilis. These two taxa share a similar seed coat morphology, and in both the flowers are always solitary in the leaf axils. Siphonoglossa calcarea differs from S. ses- silis by its reduced stature and its hirtellous stem pubescence, minor character differences which may have been environmentally induced by the "very dry limestone soil" from which the type specimen was collected. So similar are these two taxa, that I have some reservations in recognizing, at any rank, the material from La Paz. Because there 206 MADRONO [Vol. 36 are slight differences between the two taxa, however, and because S. calcarea is outside the known range of S. sessilis, I choose to recognize it at the varietal level. By doing so, I hope that the variation of this population will be brought to the attention of botanists and further collections will be encouraged. Siphonoglossa sessilis (Jacq.) Oersted ex Radlk. var. calcarea (Leon- ard) Hilsenbeck, comb, nov . — Siphonoglossa calcaraea Leon- ard, Contr. U.S. Natl. Herb. 31:402. 1958. -Type: COLOM- BIA, Magdalena, on very dry limestone soil at La Paz, 200 m, 14 Jan 1944, O. Naught 3929 (holotype, US!). Siphonoglossa sessilis var. sessilis may be distinguished readily from other taxa of Siphonoglossa by a combination of the solitary (rarely 2), sessile or very short-peduncled, axillary flowers, the nearly glabrous short bracteoles (1.5-3 mm) and sepals (2.5-3.5 mm), and the nearly glabrous fruits which are 8-10 mm long. Siphonoglossa sessilis var. calcarea may be distinguished from the typical variety by a combination of the densely hirtellous stems, petioles and brac- tlets, and the somewhat shorter (7-8 mm long), densely hirtellous capsules. The single collection that represents the holotype is the only known material of this variety. How Leonard overlooked the similarities between S. calcarea and S. sessilis may be explained partially by his belief that S. sessilis belonged to Justicia. Thus, he may have never considered the two taxa in the same light. Leonard (1958) placed Siphonoglossa in the Odontonemeae, where Lindau (1894, 1895) had placed it. Henrick- son and Hilsenbeck (1979) and Hilsenbeck (1979), however, have shown that Siphonoglossa belongs in the Justicieae. The generic concept of Siphonoglossa and its circumscription relative to Justicia will be discussed in a future paper dealing with generic concepts and delimination in the Justicieae. Acknowledgments I sincerely thank James Henrickson and Tom F. Daniel for helpful discussions and encouragement, and Daniel for providing collections of the new species from Jalisco and Michoacan, Mexico, where it was previously unknown. I also thank Marshall C. Johnston for the Latin translation of the diagnosis, Julia Larke for the preparation of the illustration, and Jacquelyn Kallunki for editorial assistance. This study was partially supported by the National Science Foundation, Grant DEB 8014249. Literature Cited Blake, S. F. 1922. New plants from Guatemala and Honduras. Contr. U.S. Natl. Herb. 24:1-32. Gibson, D. N. 1972. Studies in American plants (Guatemala). Fieldiana Bot. 34: 57-87. . 1974. Acanthaceae. In P. C. Standley, L. O. Williams, and D. N. Gibson (eds.). Flora of Guatemala. Fieldiana Bot. 24 X, nos. 3, 4:438-461. 1989] HILSENBECK: SIPHONOGLOSSA 207 HiLSENBECK, R. A. 1979. Tribal and subtribal realignment of Siphonoglossa (Acan- thaceae). Abstract Bot. Soc. Amer. Misc. Publ. 157:59. . 1983. Systematic studies of the genus Siphonoglossa sensu lato (Acantha- ceae). Ph.D. dissertation. Univ. Texas, Austin. Henrickson, J. and R. A. Hilsenbeck. 1979. New taxa and combinations in Si- phonoglossa (Acanthaceae). Brittonia 31:373-378. Leonard, E. C. 1958. The Acanthaceae of Colombia. Contr. U.S. Natl. Herb. 31: 1-781. LiNDAU, G. 1894. Beitrage zur Systematik der Acanthaceen. Bot. Jahrb. 18:36-64. . 1895. Acanthaceae. In A. Engler (ed.). Die Naturlichen Pflanzenfamilien 4, 3b:274-354. Oersted, A. S. 1854. Mexicos og Centralamerikas Acanthaceer. Vidensk. Medd. Naturl. For. Kjobenhavn. 6:113-181. Radlkofer, L. 1883. Ueber den systematischen Werth der Pollenbeschaffenheit bei den Acanthaceen. Sitzungsber. Math.-Phys. CI. K5nigl. Bayer. Akad. Wiss. Munchen II. 13:256-314. Standley, p. C. and J. Steyermark. 1952. In P. C. Standley and L. O. Williams, Plantae Centrali-Americanae, IV. Ceiba 3:101-132. (Resubmitted 6 Dec 1988; revision accepted 10 May 1989.) NOTEWORTHY COLLECTIONS California Polygonum marinense Martens and Raven (Polygonaceae). — Marin Co., Escalle Marsh, ca. 1 50 m SE of Bon Air Bridge along the SW shore of Corte Madera Creek, T 1 N R6W sect. 1 6, NE 'A, in salt marsh with Distichlis spicata and Salicornia virginica, 15 Jun 1987, Schierenbeck s.n. (JEPS). Two populations of about 20 and 25 plants, respectively. Both populations ca. 8 m from the shoreline edge of vegetation. Iden- tification confirmed by J. Hickman. Significance. Rediscovery of population last seen 23 April 1944 by J. Howell, previously thought to have been extirpated. One of two known extant occurrences. This population is threatened by housing development. The other population on Pt. Reyes near the end of Schooner Bay just north of Sir Francis Drake Highway was not found during the last visit in 1984 by R. Fowler. — Kristina A. Schierenbeck, Department of Botany, Washington State University, Pullman, WA 99164-4230. Oregon Panicum rigidulum Bosc ex Nees (Poaceae). — Douglas Co., Umpqua River, 7 mi [1 1.3 km] S of Elkton, T23S R7W sect. 30, 40 m, 20 Sep 1988, Zika 10635 (OSC). Growing with Leersia oryzoides, Eleocharis acicularis, and E. palustris in damp ground on the east riverbank. Significance. First record for Oregon. Previous West Coast reports from California and British Columbia; native from the Great Lakes to the Atlantic, where it is com- monly associated with the same taxa. P. rigidulum has long been known as P. agwstoides Sprengel, an illegitimate name (Voss, Rhodora 68:435-463, 1 966). Recent eastern authors (Dore and McNeill, 1 980, Grasses of Ontario, Mitchell, 1986, A Checklist of New York State Plants, Kartesz and Kartesz, 1980, A Synonymized Checklist of the Vascular Flora of the United States, Canada, and Greenland, agree that the author of P. rigidulum is Bosc ex Nees, and do not recognize the former Fernaldian varieties of P. agrostoides.—FETER F. Zika, BLM, P.O. Box 10226, Eugene, OR 97440. SEGREGATION OF HASTINGSIA SERPENTINICOLA, SP. NOV. FROM HASTINGSIA ALBA (LILIACEAE: ASPHODELEAE) Rudolf W. Becking College of Natural Resources, Humboldt State University, Areata, CA 95521 Abstract A new taxon, Hastingsia serpentinicola, is segregated from Hastingsia alba. The range of Hastingsia serpentinicola is sympatric with that of H. alba except for the northern Sierra Nevada where no H. serpentinicola has been found. Hastingsia ser- pentinicola is exclusively limited to serpentine soils. It occupies dry open hillsides and does not occur within the permanently wet, boggy habitats of H. alba. There is no evidence of hybridization. The segregation of Hastingsia serpentinicola necessi- tated an emended description of H. alba. A systematic study of Hastingsia of the Klamath Mountains of Northern California (Becking et al. 1 982; Becking 1986) has revealed the existence of a new species herein described. Hastingsia serpentinicola Becking, sp. nov.— Type: USA, Oregon, Josephine Co., Cave Junction, Eight Dollar Mt. BLM Darling- tonia bog, 450 m, T38S R8W sect. 28 SWV4 of SWV4, long. 123°39'25"W, lat. 42°14'00"N, 28 May 1985, R. Becking 850500 (holotype, CAS; isotypes, GH, HSC, OSC, UC, US). Herba typice inferioris. Bulbus oblongatus, frequenter attenuatus, 23-40 mm longus, 14-21 mm diametro. Scapus unus per annuum, 28.6-51.4 cm altus. Folia angusta, linearia, 19.6-34.5 cm longitu- dinis maximae et 4-6 mm latitudinis maximae. Racemi terminales cum 24-35 floribus, 3.8-12 cm longi, erecti, solitarii, rarifer rami- ficantes. Tepala angustiae, albae, palUdae viridae vel flavae, 5-6 mm longae, 1-2 mm latae, pauce oblanceolatae, acuminatae, fortifer re- flectae cum staminibus extrudantibus. Capsula oblonga, 5-8 mm longa, 4-6 mm lata. Perennial small herbs. Bulb oblong, 23-40 mm long and 14-21 mm wide. Scape slender, solitary, 286-514 mm tall, its basal pe- duncle thickness 1-3 mm. Leaves grasslike, distinctly keeled, grayish green, glabrous, with 19.6-34.5 cm maximum leaf length, and 4-6 mm maximum leaf width. Terminal racemes slender, relatively open, with 24-35 flowers. Racemes 3.8-12 cm long, usually solitary, oc- casionally with 1-3 much shorter lateral ascending racemes. Tepal segments all nearly alike, light greenish to light yellowish-white, linear 5-6 mm long, 1-2 mm wide with a greenish-yellow or light purple central vein. Tepals sharply reflexed fully about % or more of their length. Stamens exserted, straight, 5-6 mm long; anthers Madrono, Vol. 36, No. 3, pp. 208-216, 1989 1989] BECKING: HASTINGSIA SERPENTINICOLA 209 light purple becoming brownish-yellow when shedding pollen. Cap- sule oblong, 5-8 mm long and 4-6 mm wide. Seed wrinkled, black, 4-5 mm in length (Fig. 1). Paratvpes. USA, OR, Curry Co., Collier Creek, 28 Jun 1929, Leach s.n., 2411, 2412 (ORE); Josephine Co., W of O'Brien, 5 Jul 1939, Hitchcock and Martin 5166 (CAS, NY, UC); CA, Del Norte Co., Gasquet, Old Gasquet Toll Rd, 5 Jun 1965, Roderick s.n. (Jeps) [n=26, n=27. Cave (1970) chromosome voucher 6775]; Humboldt Co., Willow Creek, Horse Mt., 23 Jul 1979, Overton and Butler 9473 (HSC); Lake Co., Hulville, 2 Aug 1902, Heller 6013 (DS, GH, NY, POM); Mendocino Co., Laytonville, Red Mt., Jun 1867, Bolander s.n. (DS); Trinity Co., Peanut, Philpot Creek, 13 Jun 1951, Baci- galupi and Constance 3392 (UC). Distribution and ecology. Hastingsia serpentinicola is limited to the Klamath Mountains and North Coast Ranges from low to high elevations. It is almost exclusively limited to ultramafic or serpentine rock outcroppings. It occupies open sites in the Klamath Mountains that are moist in the spring but dry out rapidly in the early summer (Fig. 3). It occurs along the edges of wet bogs or in the drier islands within the bog environment. At high elevations (above 1800 m) on exposed serpentine ridges it favors mesic habitats. Hastingsia alba (Durand) S. Watson, Proc. Amer. Acad. Arts 14: 242, 1879, emend. Becking. — Schoenolirion album Durand, J. Acad. Nat. Sci. Philadelphia (Ser. #2) 3:103, 1855. The Has- tingsia alba holotype material is very fragmentary, consisting only of a few raceme stalks and flowers collected from several individual plants. This "holotype" material was later divided into 3 portions and deposited by Durand himself into three different herbaria (P-DU, PH!, NY!). The Paris, France (P-DU) material was judged to be too fragile for overseas shipment and has not been available (S. Barrier, letter 17 Jun 1987). Because there is no longer a single intact holotype, there are consequently no isotypes (G. Zijlstra ICBN (U), letter 1 Jun 1987; B. C. Stone (PH), letter 12 Aug 1987). The PH material is designated here- with as lectotype because Durand described his "holotype" ma- terial during his tenure at the PH herbarium. The P-DU and NY material become isolectotypes. The PH material still con- tained a single fully intact flower, representing "a portion of the type material collected by H. Pratten in the vicinity of Deer Creek, California, 1851", according to the typed label infor- mation. Perennial herb, often robust. Bulb 26-56 mm long, 17-31 mm wide. Scape robust, 40.4-89 cm tall, basal thickness of the peduncle 3-5 mm. Leaves glaucous green, changing with age to light green. 210 MADRONO [Vol. 36 Fig. 1 . Hastingsia serpentinicola. a, Fruiting plant; b, flowering plant; c, stamens; d, frontal view of flower; e, pistil; f, part of flowering raceme; g, longitudinal section of flower; h, fruiting raceme; i, part of fruiting raceme; j, cross-section of capsule; k, longitudinal section of capsule; 1, seeds; m, longitudinal section of seed. All bar scales = 10 mm. 1989] BECKING: HASTINGSIA SERPENTINICOLA 211 more fleshy, keeled, 27.8-53 cm long and 7-14 mm wide. Racemes compact, 8-20.3 cm long, densely flowered with 62-75+ flowers, often 2-3 branched, erect and terminal. Perianth segment pure white, sometimes slightly yellowish-white, obtuse, rotate at about half of the tepal length, giving the opened perianth a distinct star-shaped appearance. Basal portion of the perianth bell-shaped when closed. Outer 3 tepals linear, 6-8 mm long and 1 mm wide, ending in a blunt, slightly swollen white tip. Inner 3 tepals ovate, 6-8 mm long and 2 mm wide, often with a crisped margin. Stamens only slightly protruding beyond the perianth, anthers often purplish. Filaments 6-7 mm long. Outer 3 stamens opposite the ovate tepals, usually elongate, and open first; the 3 inner stamens elongate on following days and become almost of equal length to the outer stamens when their anthers dehisce. If the flower has just opened, it appears that for the first days the 3 outer stamens have long filaments whereas the inner stamens have short filaments. Capsule oblong, broadly 3-lobed, 6-9 mm long, and 5-8 mm wide. Seed wrinkled, black, 4- 6 mm long (Fig. 2). Representative specimens. USA, OR, Curry Co., Vulcan Lake, 23 Jul 1978, Dawn 47 (OSC); Josephine Co., Cave Junction, Illinois River bridge USFS Road 3843, 17 Jun Becking 820600 {yi%C)\ CA, Butte Co., Jonesville, 22 Jul 1914, Hall 9769 (NY); Del Norte Co., Steven Mt., 3 Aug 1980, Baker 3577 (HSC); Humboldt Co., Hoopa, Mifl Creek Lakes, 1 Aug 1979, Clifton and Griswold 11944 (HSC); Nevada Co., Willow Springs, 26 Jun 1961, Cruden 1035 (JEPS) [^=26, Cave 1 966, 1 970]; Plumas Co., N Fork Feather River, Mosquito Creek, 1 Jul 1965, See s.n. (JEPS) [n=26. Cave 1970]; Shasta Co., Delta, Sacramento River, 1 8 Jun 1 923, Bethel s. n. (CAS); Siskiyou Co., Callahan, French Creek Rd, 12-16 Jun 1948, Ferris and Lorraine 11730 (ORE, NY, RSA, DS); Tehama Co., Deer Creek Canyon, 17 Jul 1911, Eggleston 7265 (NY); Trinity Co., Denny, New River-Trinity River confluence, 17 May 1975, Sullivan 65 (HSC). Distribution and ecology. Hastingsia alba is not predominantly a species of the Klamath Mountains Geological Province (Ramp and Peterson 1979; Borine 1983; Ferlatte 1974). It has been collected in numerous other locations in the North Coast Ranges and the north- ern Sierra Nevada. In both situations it occurs on serpentine, on granite, and diorite. It occupies most commonly open rocky seepage areas with a year-round water supply, or bogs or wet meadows, especially at high elevations. Its elevation ranges from 500-2300 m. At high elevations, H. alba is often stunted and has smaller bulbs, shorter scapes, and shorter and narrower leaves. The other floral and capsule distinctions, however, are retained (Fig. 3). 212 MADRONO [Vol. 36 Fig. 2. Hastingsia alba, a. Fruiting plant; b, flowering plant; c, top of flowering raceme; d, longitudinal view of flower; e, longitudinal section of flower; f, cross- section of leaf; g, flowering plant, robust form; h, top of flowering raceme; i, part of fruiting raceme; j, longitudinal section of capsule; k, cross-section of capsule; 1, seeds and longitudinal section of seed. All bar scales = 10 mm. 1989] BECKING: HASTINGSIA SERPENTINICOLA 213 • HASTINGSIA ALBA • HASTINGSIA SERPENTINICOLA Fig. 3. Distribution of Hastingsia serpentinicola and H. alba. Comparison of Hastingsia Serpentinicola and H. Alba Statistical analyses. Comprehensive t-test analyses for unequal variances were applied to the 442 measured specimens selecting the 47 characters for comparison at the individual character level. Twelve most significant characters out of the total 33 significant characters at the 0.01 probability level are identified in Table 1. Two discriminant analyses (Nie et al. 1975; Dixon et al. 1983) were used independently to classify the 196 herbarium specimens of Hastingsia serpentinicola and the 246 specimens of H. alba. Thir- ty-nine quantitative and eight qualitative characters were selected in these tests by the computer programs. All the specimens missing one or more characters of this character set were disqualified from the discriminant analysis. Because of this potential for many un- warranted exclusions, the number of characters had to be selectively limited to increase the number of qualifying cases. Rarely one finds fully developed flowers and mature capsules together on the same individual herbarium specimen or field specimen. Five character groupings were selected using various combinations of bulb, scape, foliage, flower and capsule characters (Table 2). The segregations into the two species were supported by both computer programs in all the character groupings tested (canonical r% = 0.824-0.984, 86- 100% correct classification). Foliage and bulb characters of Hastingsia alba performed the poorest among herbarium specimens when either discriminant pro- gram was applied. Collectors tend to choose small individuals to 214 MADRONO [Vol. 36 Table 1 . Distinctions between Hastingsia serpentinicola and H. alba. All char- acters listed had significant differences at the 0.0 1 probability level by t-test analysis. Leaf width was measured selectively on 2-3 undamaged leaves at about the middle of the leaf length to establish the maximum and minimum dimension. "Raceme open" length was measured from the top of the terminal raceme to the lowest lateral raceme branch or the lowest flower in the terminal raceme. "Raceme dense" length was measured from the top of the terminal raceme to where the interspatial distances between the flower pedicels exceeded the length of the perianth. Density of flowers counted the numbers per 5 cm length of the terminal raceme and expressed this count per 10 cm raceme length. The densest (top most portion) and the least sparse (basal portion) raceme portions were selected for counting the maximum/minimum range of flower density respectively. The number of cases involved is indicated in paren- theses. H. serpentinicola H. alba Bulb length 23.2-40.2 mm (150) 26.3- -55.9 mm (164) Bulb diameter 13.5-21.1 mm (150) 17.3- -31.3 mm (164) Scape length 28.6-51.4 cm (196) 40.4- -89.0 cm (237) Leaf length max. 19.6-34.5 cm (194) 27.8- -52.9 mm (241) Leaf width max. 3.5-5.7 mm (194) 7.4- -13.6 mm (241) Tepal length max. 4.9-6.3 mm (189) 5.5- -7.5 mm (213) Tepal width max. 0.9-1.9 mm (189) 1.7- -2.3 mm (213) Raceme open length 4.5-26.7 cm (192) 13.7- -39.8 cm (243) Raceme dense length 3.8-12.0 cm (192) 8.0- -20.3 mm (243) Flower density max. 28.3-48.5 mm (160) 38.6- -63.6 mm (207) Flower density min. 21.7-41.9 mm (131) 32.4- -56.2 mm (205) Capsule length max. 4.7-7.5 mm (061) 5.7- -9.3 mm (098) Capsule width max. 3.6-6.0 mm (061) 4.5- -7.7 mm (098) press and often remove the dead and shriveled black basal leaves. Extraction of bulbs at depths of 25-45 cm often leaves the tunica in the soil. Only older specimens develop the typical tunica, whereas younger and smaller specimens do not. The bulbs are often cut or squashed making it most difficult to obtain representative measure- ments. Capsules are often immature. Only fully matured capsules with black seeds are dimensionally representative. In segregations based upon herbarium specimens, H. serpentinicola was classified more consistently than H. alba (85-100% correct classification). In mature and fresh specimens, however, such misclassifications were rarely encountered. Distinction between species. In past treatments (Jepson 1921, 1936; Abrams 1923; Mason 1957; Munz 1959; Peck 1961; Ferlatte 1974) Hastingsia serpentinicola has not been segregated from H. alba. Stature and size differences were attributed to harsh environmental conditions. However, Roderick, in 1965, annotated his collections of//, serpentinicola (JEPS) as distinct from H. alba. He further noted that plants now attributed to H. serpentinicola occur "almost always on serpentine on well-drained soil never in a bog". Both Hastingsia species were cultivated for many years in the Botanical Garden at UC Berkeley, CA. Growing both in the same 1989] BECKING: HASTINGSIA SERPENTINICOLA 215 Table 2. Discriminant Analyses using 246 Specimens of Hastingsia alba and 196 Specimens of H. serpentinicola. The first two lines per category represent the SPSS90 program results, the third line the BMDP83 results, using identical data sets, /y* # char = the number of characters selected by the program for classification into the grouping. 2) (#*) = the number of most significant characters selected within the character grouping for the final classification function. 3) # cases = the number of cases selected with complete character sets within the grouping. 4) can- on. r% = canonical correlation coefficient of the final classification function. 5) H. a. and H. s. = Hastingsia alba and H. serpentinicola, respectively. 6) % class (#) = per- cent of correct classification (number of cases involved). /J J = Jackknife classification using Mahalanobis (BMDP83). 8)\^ = Wilk's Lambda (SPSS90) classification. Percent correctly classified Character char, 3# "canon. a. H. s. Total groupings ^(#*) cases r% ^% class (#) % class (#) % class Bulb, scape. 10 298 0.824 86.3 (153) 100.0(145) 93.0 foliage (2*) 'J: 86.3 (153) 100.0 (145) 93.0 10 298 0.824 «W: 87.9 (224) 100.0 (194) 93.5 (4*) Foliage, scape 8 411 0.824 87.8 (222) 100.0 (189) 93.2 (4*) J: 87.4 (222) 100.0 (189) 93.2 8 411 0.824 W: 87.9 (224) 100.0(194) 93.5 (4*) Foliage, scape. 26 232 0.984 99.3 (135) 99.0 (097) 99.1 flower (3*) J: 99.3 (135) 99.0 (097) 99.1 26 232 0.985 W: 97.9 (188) 84.9 (179) 91.6 (7*) Foliage, scape, 14 92 0.873 95.0 (060) 100.0 (032) 96.7 capsule (3*) J: 95.0 (060) 100.0 (032) 96.7 14 92 0.885 W: 98.3 (064) 100.0 (032) 98.9 (5*) Bulb, scape, foliage, 47 30 0.972 100.0 (018) 100.0 (012) 100.0 flower, capsule (6*) J: 100.0 (018) 100.0 (012) 100.0 47 37 0.999 W: 95.5 (022) 86.7 (015) 91.9 (18*) rich garden soil, the cuhivated specimens produced flowers and viable seeds (W. Roderick, letter 24 Dec 1985) and retained their distinctive characters, suggesting no environmental influences upon these morphological characters. Hastingsia serpentinicola did not flourish well, however, in Berkeley, with its mild winters and hot dry summers. Herbarium specimens of these cultivated Hastingsia plants served as voucher specimens for pollen collection and chro- mosome counting (Cave 1966, 1970). Acknowledgments Curators of the various herbaria (CAS, DAY, DS, GH, HSC, JEPS, NY, ORE, OSC, PH, PUA, ROPA, RSA, SOC, UC) are acknowledged for their hospitality during visits and loan of specimens. Acknowledged are the constructive comments of Dr. F. Raymond Fosberg (NAT), Dr. M. G. McLeod, Dr. Robert Peet (NC), Dr. Wm. A. Weber (CO), 9 anonymous reviewers and 2 editors. 216 MADRONO [Vol. 36 Literature Cited Abrams, L. 1923. An illustrated flora of the Pacific states, vol. 1. Stanford Univ. Press, Stanford, CA. Becking, R. W. 1986. Hastingsia atropurpurea (Liliaceae, Asphodeleae), a new species from southwestern Oregon. Madrofio 33:175-181. Becking, R. W., J. A. Lenihan, and E. Muldavin. 1 982. Final report. Schoenolirion bracteosum. Ecological investigations. Contract #43-9A4-0- 1 645. Six Rivers Nat. Forest, Eureka, CA. BoRiNE, R. 1983. Soil survey of Josephine County, Oregon, USDA Soil Conser- vation Service, Grants Pass, OR. Cave, M. S. 1966. In Documented chromosome numbers of plants. Madrono 18: 245-246. . 1970. Chromosomes of the California Liliaceae. Univ. Calif. Publ. Bot. 57: 8, 10. Dixon, W. J., (chief ed.) et al. 1983. BMDP statistical software 1983 printing with additions. 18.4 Stepwise Discriminant Analysis P7M:5 19-537. Univ. Calif. Press, Berkeley, CA. Durand, E. 1855. Plantae Prattenianae Californicae. J. Acad. Nat. Sci. Philadelphia (Ser. 2)3:103. Ferlatte, W. J. 1974. A flora of the Trinity Alps of northern California. Univ. Calif. Press, Berkeley. Jepson, W. L. 1921. A flora of California, vol 1 , part 6. Assoc. Students Store, Univ. California Berkeley. . 1923-1925, 1936. A manual of the flowering plants of California. Assoc. Students Store, Univ. California Berkeley. Mason, H. L. 1957. A flora of the marshes of California. Univ. Calif. Press, Berkeley. Munz, p. a. 1959. A California flora. Univ. California Press, Berkeley. NiE, N. H., C. H. Hull, J. G. Jenkins, K. Steinbrenner, and D. L. Bent. 1975. SPSS: statistical package for the social sciences, pp. 434-467. McGraw-Hill, Inc, New York. Peck, M. E. 1961. A manual of the higher plants of Oregon. Metropolitan Printing Co., Portland, OR. Ramp, L. and N. V. Peterson. 1979. Geology and mineral resources of Josephine County, Oregon. Dept. Geology & Mineral Resources, State of Oregon. Bull. 100. Watson, S. 1879. Revision of the North American Liliaceae. Proc. Amer. Acad. Arts 14:213-288. . 1885. Hastingsia bracteosa S. Wats. Proc. Amer. Acad. Arts 20:377. (Resubmitted 12 Sep 1988; revision accepted 9 Mar 1989.) NOTE ISOPYRUM STIPITATUM A. GrAY (RaNUNCULACEAE) IN THE WILLAMETTE VaLLEY, Oregon.— This inconspicuous perennial herb has been collected from only three localities in the Willamette Valley, where it is at the northern limit of its known distribution. The species is also known from the Klamath-Siskiyou region of Oregon and California, southward into the Cascades and northern Coast Ranges of California, and is disjunct at its southern limit in the East Bay region of Alameda and Santa Clara Counties (Calder and Taylor, Madroiio 17:69-76, 1963). In Rare, Threatened and Endangered Plants and Animals of Oregon (Oregon Natural Heritage Data Base, 1989) /. stipitatum is currently included on the review list, which comprises species for which more information is needed before their status can be determined. The first Willamette Valley collections were from Yamhill County, where /. stip- itatum was collected on 5 occasions between 1957 and 1959 along Willamina Creek, north of Willamina [7 Mar 1957, Mendenhall s.n. (OSC)]. The species is still extant in this locality, though the number and extent of colonies has declined in recent years (E. Mendenhall, pers. comm.). A 1958 collection from Polk County [31 Jan 1958, Lofgren s.n. (OSC)] was taken from Buell County Park, along Mill Creek about 12 km SE of the Yamhill County locality. The most recently discovered population, along the Marys River south of Corvallis, Benton County [30 Mar 1980, Chambers 4602 (OSC)] appears to have been extirpated. The circumstances surrounding the Polk County occurrence, which is still extant [7 Apr 1988, Alverson 1306 (OSC)], seem sufficiently unusual to warrant a description. In the Willamette Valley /. stipitatum appears to occur primarily in rich deciduous woods that occupy alluvial stream bottoms. Herbarium labels and field observations show that Acer macrophyllum and Fraxinus latifolia dominate the tree canopy of such sites, with a diverse herb layer typified by Delphinium trolliifolium, Hydro- phyllum tenuipes, Viola glabella, Thallictrum occidentale, and Trillium albidum. The diminutive /. stipitatum occupies open microsites amongst the generally thick her- baceous cover. At Buell Park, /. stipitatum could not be found in such natural v/ood- lands, but instead occurred in sizable patches in the rough lawn of an adjacent picnic area and playground, over an area of about 0.5 ha. Isopyrum was most abundant and vigorous under the canopies of scattered trees of Fraxinus and Acer, where competing grasses were relatively sparse. The remnant trees of Fraxinus and Acer suggest that this area was also alluvial deciduous woodland at one time. It is plausible that /. stipitatum was present at the site before the park was established, and has persisted, or perhaps even increased, with the removal of competing native shrubs and herbs. Interestingly, the habit of persisting in lawns and pastures has been reported for Isopyrum biternatum, a species of eastern North America (Korling, Eastern De- ciduous Forest, 1973). The small stature of the plants and the early blooming season (mostly February and March) are perhaps partially responsible for the paucity of known occurrences of /. stipitatum in northwestern Oregon. Additional populations may possibly occur undetected elsewhere in the Willamette Valley, particularly in alluvial stream bottom habitats. However, care should be taken to ensure that the few known populations are not needlessly destroyed. I thank Elizabeth Mendenhall for providing access to /. stipitatum colonies, as well as information on their original discovery, and the Mazamas Research Committee for providing funds supporting field work— Edward R. Alverson, Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR 97331. (Re- ceived 22 Nov 1988; revision accepted 30 May 1989.) Madrono, Vol. 36, No. 3, p. 217, 1989 COMMENTARY Points of View: Response to Reveal The report of David Douglas's death (Reveal, Madrono 36:137-140, 1989), to paraphrase Mark Twain, has been exaggerated. Nothing that editors can do, fortu- nately, will change the honor owed to and bestowed upon that intrepid explorer. Reveal argues that botanical honor accrues from authorship of names. I believe that this view is both faulty and dangerous. It is faulty because many standardly used names were not proposed by the foremost expert in the group. As one common kind of example (portrayed here in an extreme form), consider that the greatest contribution to botanical knowledge in a group was made by a botanist who, after years of detailed study, reduced many unjustifiable specific epithets to varietal rank. Soon after, another botanist, compiling a manual, decided to use those names and circumscriptions at subspecific rank. Accordingly, he recombined all the names of the expert botanist, leaving the expert's name un- attached as author to any epithets in the group to which he devoted his life. This unfairness is required by the rules of nomenclature, which have no concern with recognition of individuals. Furthermore, Reveal's assumption of honor-to-the-author is dangerous because it leads individuals to wish their names attached to a large number of taxa as a measure of their great contribution to systematic botany. Over the long run, no urge could be more counterproductive to the accepted goal of stability of scientific names. Citation of persons in addition to the publishing author (as allowed through proper use of "in" and "ex") serves a historical function but not a nomenclatural one. Given serious constraints on space, the editors of The Jepson Manual chose to leave ex- plication of botanical history to authors and works able to focus adequately on this important and fascinating subject. We chose instead to concentrate on allowing the easiest possible identification of California plants, with the most accurate possible morphological and range descriptions. In The Jepson Manual, David Douglas will be listed as author only rarely— for those names that he actually published. But his name will be repeated many times, commemorated in common names and specific epithets appropriate to his pioneering work.— James C. Hickman, Jepson Herbarium, University of California, Berkeley, CA 94720. ANNOUNCEMENT Temporary New Address for Editor of Madrono From 1 September 1989 to 18 December 1989 Dr. David J. Keil, Editor of Madrono, will be on sabbatical leave. Manuscripts, proofs, and other correspondence should be sent to: Dr. David J. Keil Department of Botany Arizona State University Tempe, Arizona 85287 Madrono, Vol. 36, No. 3, p. 218, 1989 REVIEWS Soil- Plant Relations: An Ecological Approach. By D. W. Jeffrey. Croom Helm, London and Sydney, and Timber Press, Portland, OR. 1987. 295 pp., $26.95 (paper); $33.95 (hardbound). The below-ground environment for plants is too often overlooked by botanists. Despite the plethora of research papers, symposia and books on the soil-plant inter- face, it largely remains the domain of the specialist. It is refreshing, therefore, to encounter a generalist's guidebook to this important earthy subject. Jeffrey's book looks at the soil-plant system, not from a greenhouse or agricultural view, but from an ecological one. The book's three parts follow a logical and didactic progression: From the essentials of the soil-plant context, to soils and mineral nutrition, and finishing with a selection of case histories. The first of the three sections (Part I, "A plant-centered biological complex") gives the reader the fundamentals of plant and soil physiology: ion uptake, inorganic mineral nutrition of plants, water uptake. Further along in this section are chapters on mycorrhizal and other symbioses, biomass recycling, and a precis on fire in the soils-vegetation mix. Part II focuses on the soil component of the soil-plant syndrome. Good accounts of soils formation, the microenvironment of soils (matrix temperature and nutrient supply) now follow. Ch. 1 1 deals with the critical issues of nutrient availability and toxic ions, and Ch. 12 explores techniques used in testing for soil variables. Part III is a refreshing departure. Rather than an attempt at rounding out an encyclopedic coverage, of yet other topics, Jeffrey adopts the case-history approach. Each one of the seven chapters deals with a specific and significant issue in soils- vegetation studies: 1 ) Autecology of contrasting species, 2) Restoration of derelict land, 3) Heathlands and other nutrient-poor ecosystems, 4) Arctic tundra, 5) Salt marshes, 6) Calcareous, and 7) serpentine plant-soil relations. It is in this section that Jeffrey builds an ecological edifice from the substance of the earlier chapters. The selection of case-histories is judicious, and tells a fascinating story. All in all the book ably fills a significant niche in telling of the all-important relations between soils and plants. The North American co-publisher. Timber Press, is to be commended in supporting this worthy contribution. — A. R. Kruckeberg, University of Washington, Seattle, WA 98195. North American Terrestrial Vegetation. Edited by Michael G. Barbour and W. DwiGHT Billings. Cambridge University Press, New York. 1988. 434 pp., $49.50 (hardbound). ICBN 0-521-26198-8. At Last! Barbour and Billings have put together the state-of-the-art compilation of the vegetation of North America. Though they discuss in the preface how they started in 1982 on a three year project to produce this book, the need for such a volume has been talked about for the past twenty years. They are to be commended for the successful completion of a very difficult task. The success of this publication is in large part due to the team of writers that Barbour and Billings were able to gather. The thirteen chapters are authored by a veritable "who's-who" of vegetation scientists on this continent: Bliss— Arctic, Elliott- Fisk— boreal forest, Peet— Rocky Mountains, Franklin — Pacific Northwest, Bar- Madrono, Vol. 36, No. 3, pp. 219-220, 1989 220 MADRONO [Vol. 36 bour— California forests and woodlands, Keeley and Keeley— Chaparral, West— in- termountain shrublands and woodlands, MacMahon— warm deserts, Sims— grass- lands, Greller— deciduous forests, Christensen— coastal plain, Hartshorn— tropical and subtropical forests, and Billings— alpine. Each of the chapters deals in general with the topics of vegetation structure and composition, response to disturbance, variations due to environmental gradients, autecology of selected species, vegetational history, and suggestions for future re- search. These chapters are individualized, however; each one with a slightly different emphasis based on the authors' experience and bias and upon the main thrust of the research accomplished in the specific vegetation type. Examples are the emphasis on species in the chapter on chaparral, the emphasis on forest types in the chapter on Pacific Northwest, and the emphasis on community-environment relations in the chapter on arctic vegetation. This book could not possibly have reviewed all the work in vegetation science in North America, no one book could do that. For instance, Peet, in his excellent review of the forests of the rocky mountain region, mentions Populus tremuloides only in the context of its being a "montane serai forest" type. Mueggler, in a study of aspen forest types in the rockies (USDA Forest Service, Gen. Tech. Report INT-250), however, lists 14 major, 12 minor, and 33 incidental aspen community types. No one should approach this book with the idea that it will answer all questions on any given vegetation, it won't. We will still have need of more in-depth, regional works. This is an overview look at the plant cover of the continent and as such it is excellent. Even when one comes to this book in the right frame of mind, there are some disappointments. Most of the chapters treat the subject as if we were still in the days of Daniel Boone or even George Washington. I found myself on a number of occasions wanting to see the authors discuss the intrusion of man into the system with the resulting isolation of "natural" or "pristine" areas. As we all know from our treks about the country-side, it is getting harder and harder to find good examples of many plant communities in natural condition. This issue is so little dealt with that it hinders the book's usefulness to those that are traveling outside their normal sphere of ex- perience. The second disappointment to me is the lack of reference to major regional mapping, such as Brown and Lowe's Biotic communities of the Southwest (1980, USDA Forest Service Gen. Tech. Rep. RM-78) and the many classification efforts that are ongoing, especially here in the west. More recognition should have been given to these kinds of work and it would have made the book more useful to have them referenced. In their preface the authors state that their intended audience is knowledgeable laypeople, advanced undergraduates, graduate students, and professional ecologists in both basic and applied fields. It's stretching things a bit to think that a layperson will pay the price for a fairly large book without glossy pictures and maps to tell them where to go and see, but this book is a must for every field biologist of whatever ilk. In 1960 we all ran around with Oosting's Plant Ecology, in 1990 we'll have Barbour and Billings' North American Terrestrial Vegetation. — William L. Halvorson, Chan- nel Islands National Park, Ventura, CA 93001. Volume 36, Number 3, pages 141-220, published 17 October 1989 Subscriptions — Membership Membership in the CaHfomia Botanical Society is open to individuals ($22 per year; students $ 1 2 per year for a maximum of seven years). Members of the Society receive Madrono free. Family memberships ($25) include one five-page publishing allotment and one journal. Emeritus rates are available from the Corresponding Secretary. Institutional subscriptions to Madrono are available ($30). Membership is based on a calendar year only. Applications for membership (including dues), orders for sub- scriptions, and renewal payments should be sent to the Treasurer. Requests and rates for back issues, changes of address, and undelivered copies of Madrono should be sent to the Corresponding Secretary. Information for Contributors Manuscripts submitted for publication in Madrono should be sent to the editor. All authors must be members, and membership is prerequisite for review. Manuscripts and review copies of illustrations must be submitted in triplicate for all articles and short items intended for NOTES AND NEWS. Follow the format used in recent issues for the type of item submitted and allow ample margins all around. All manuscripts MUST BE DOUBLE SPACED THROUGHOUT. For ar- ticles this includes title (all caps, centered), author names (all caps, centered), addresses (caps and lower case, centered), abstract, text, acknowledgments, literature cited, tables (caption on same page), and figure captions (grouped as consecutive paragraphs on one page). Order parts in the sequence listed ending with figures, and number each page. Do not use a separate cover page, "erasable" paper, or footnotes. Manuscripts prepared on dot matrix printers may not be considered. Table captions should include all information relevant to tables. All measurements should be in metric units. Line copy illustrations should be clean and legible, proportioned (including cap- tions) to the Madrono page, and designed for reduction to % original size. Scales should be included in figures, as should explanation of symbols, including graph coordinates. Symbols smaller than 1 mm after reduction are not acceptable. Maps must include latitude and longitude references. Halftone copy should be designed for reproduction at actual size. In no case should original illustrations be sent prior to the acceptance of a manuscript. When needed they should be mounted on stiff card- board and sent flat. No illustrations larger than 22 x 28 cm will be accepted. Presentation of nomenclatural matter (accepted names, synonyms, typification) should follow the format used for Rhus integrifolia in Madrono 22:288. 1974. Institutional abbreviations in specimen citations should follow Holmgren, Keuken, and Schofield, Index Herbariorum, 7th edition. Abbreviations of serial titles should be those in Botanico-Periodicum-Huntianum (Lawrence et al., 1968, Hunt Botanical Library, Pittsburgh). If the correct abbreviation cannot be determined, the full serial title should be used. Titles of books should be given in full, together with the place and date of publication, publisher, and edition, if other than the first. All members of the California Botanical Society are allotted five free pages per volume in Madrono. Joint authors may split the full page number. Beyond that number of pages a required editorial fee of $65.00 per page will be assessed. The purpose of this fee is not to pay directly for the costs of publishing any particular paper, but rather to allow the Society to continue publishing Madrono on a reasonable schedule, with equity among all members for access to its pages. Printer's fees for illustrations and typographically difficult material @ $35.00 per page (if their sum exceeds 30 percent of the paper) and for author's changes after typesetting @ $4.50 per line will be charged to authors. At the time of submission, authors must provide information describing the extent to which data in the manuscript have been used in other papers that are published, in press, submitted, or soon to be submitted elsewhere. CALIFORNIA BOTANICAL SOCIETY VOLUME 36, NUMBER 4 OCTOBER-DECEMBER 1989 ?3 MADRONO A WEST AMERICAN JOURNAL OF BOTANY Contents Taxonomy of the Opuntia schottii Complex (Cactaceae) in Texas Barbara E. Ralston and Richard A. Hilsenbeck 3IOSYSTEMATIC STUDIES OF PhACELIA CaPITATA (HyDROPHYLLACEAE), A SpECIES Endemic to Serpentine Soils in Southwestern Oregon J. Stephen Shelly K Systematic and Phytogeographic Study of Antennaria aromatic a and A. DENSIFOLIA (ASTERACEAEI InULEAE) IN THE WESTERN NORTH AMERICAN CORDILLERA Randall J. Bayer ^ Re-evaluation of Bealia mexicana (Poaceae: Eragrostideae) Paul M. Peterson K New Species of Daphnopsis (Thymelaeaceae) from Baja California Sur, Mexico Dennis E. Breedlove and Jose Luis Leon de la Luz HONARDELLA BENEOLENS (LaMIACEAE), A NeW SpECIES FROM THE CrEST OF THE t Southern Sierra Nevada, California James R. Shevock, Barbara Ertter, and James D. Jokerst TOTES The Taxonomic Relationships of Allocarya corallicarpa (Boraginaceae) Kenton L. Chambers Comments and Notes on Portulaca in California Walter A. Kelley jJOTEWORTHY COLLECTIONS California Colorado -EVIEWS .NNOUNCEMENTS RRATUM 221 232 248 260 266 271 280 281 283 285 287 231, 247, 265, 270, 279, 286, 288, 289 282 OMMENTARY Editor's Report for Volume 36 EVIEWERS OF MANUSCRIPTS ^IDEX for VOLUME 36 EDICATION ABLE OF CONTENTS FOR VOLUME 36 ATES OF PUBLICATION I 289 290 291 ii iv 290 UBLISHED QUARTERLY BY THE CALIFORNIA BOTANICAL SOCIETY I Madrono (ISSN 0024-9637) is published quarterly by the California Botanical So- ciety, Inc., and is issued from the office of the Society, Herbarium, Life Sciences Building, University of California, Berkeley, CA 94720. Subscription rate: $30 per calendar year. Subscription information on inside back cover. Established 1916. Second-class postage paid at Berkeley, CA, and additional mailing offices. Return requested. Postmaster: Send address changes to James R. Shevock, Botany Dept., California Academy of Sciences, San Francisco, CA 941 18. £'<^zY<9r— David J. Keil Biological Sciences Department California Polytechnic State University San Luis Obispo, CA 93407 Board of Editors Class of: 1989 — Frank Vasek, University of California, Riverside, CA Barbara Ertter, University of California, Berkeley, CA 1990— Steven Timbrook, Ganna Walska Lotusland Foundation, Montecito, CA Thomas R. Van Devender, Arizona-Sonora Desert Museum, Tucson, AZ 1991— James Henrickson, California State University, Los Angeles, CA Wayne R. Ferren, Jr., University of California, Santa Barbara, CA 1992 — Bruce A. Stein, The Nature Conservancy, Washington, D.C. William L. Halvorson, Channel Islands National Park, Ventura, CA 1993— Jon E. Keeley, Occidental College, Los Angeles, CA Rhonda L. Riggins, California Polytechnic State University, San Luis Obispo, CA CALIFORNIA BOTANICAL SOCIETY, INC. Officers for 1989-90 President: Robert W. Patterson, Biological Sciences Department, San Francisco State University, San Francisco, CA 94132 First Vice President: Niall F. McCarten, Department of Integrative Biology, Uni- versity of California, Berkeley, CA 94720 Second Vice President: Mary L. Bowerman, 970 Second Street, Lafayette, CA 94549 Recording Secretary: Rodney G. Myatt, Department of Biological Sciences, San Jose State University, San Jose, CA 95192 Corresponding Secretary: Barbara Ertter, University Herbarium, University of California, Berkeley, CA 94720 Treasurer: Mona Bourell, Department of Botany, California Academy of Science, San Francisco, CA 94118 Financial Officer: Barrett Anderson, Department of Botany, California Academy of Science, San Francisco, CA 94118 The Council of the California Botanical Society consists of the officers listed above plus the immediate Past President, John Strother, University Herbarium, Uni- versity of California, Berkeley, CA 94720; the Editor of Madrono; three elected Council Members: James Shevock, Department of Botany, California Academy of Science, San Francisco, CA 941 18; Elizabeth McClintock, University Herbarium, University of California, Berkeley, CA 94720; Bruce Pavlik, Department of Biology, Mills College, Oakland, CA 94613; and a Graduate Student Representative, James D. Morefield, Rancho Santa Ana Botanic Garden, Claremont, CA 91711. THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER. TAXONOMY OF THE OPUNTIA SCHOTTII COMPLEX (CACTACEAE) IN TEXAS Barbara E. Ralston* and Richard A. Hilsenbeck Department of Biology, Sul Ross State University, Alpine, TX 79832 Abstract Morphologic, field, and chromosomal studies of Opuntia series Clavatae including the three major taxa, O. schottii, O. grahamii, and a series of populations from Texas originally described as a putative hybrid by Anthony, suggest that O. schottii and O. grahamii are distinct and do not hybridize. The plants once considered to be hybrids are herein described as a new species, O. aggeria, most closely related to O. grahamii and O. moelleri, a species from northern Coahuila, Mexico. Resumen Estudios de campo, morfologicos y de cromosomas de Opuntia series Clavatae incluyendo las tres taxa mayor, O. schottii, O. grahamii, y una serie de poblaciones de Texas describidos originalmente como hibridos putativos por Anthony, sugieren que O. schottii y O. grahamii son distinctos ye que no se cruzan. Las plantas las cuales se consideran originalmente como hibridos se describen aqui como una nueva especie, O. aggeria, mas relacionado a O. grahamii y a O. moelleri, una especie del norte de Coahuila, Mexico. In Opuntia the series Clavatae (sensu Benson 1982), subgenus Cylindropuntia, is composed of 1 7 taxa in North America, the plants forming low mats or clumps. Two species, Opuntia schottii Engelm. and O. grahamii Engelm., colloquially known as club chollas, are common in southwestern Texas from the Rio Grande Plain into the Chihuahuan Desert. They have been reported to hybridize in the Big Bend Region of Texas in southern Brewster County (Anthony 1956; Benson 1982). These two species, their putative hybrid, and a single disjunct population of O. emoryi Engelm., constitute the O. schottii complex in Texas as circumscribed by Benson (1982) and Ralston (1987). Previous studies have differed in their treatments of O. schottii and O. grahamii (Britton and Rose 1919; Anthony 1956; Benson 1982; Weniger 1984). Previously, chromosome num- bers for these two species were reported as n=\ \ and n=22, re- spectively (Weedin and Powell 1978; Pinkava et al. 1 985), indicating that speciation in the group may involve polyploidy {x= 1 1 , Benson 1982; Grant 1981). This study uses morphologic, chromosomal, and breeding system data to clarify the taxonomic and phylogenetic re- ' Present address: Department of Biology, Northern Arizona University, Flagstaff, AZ 86011. Madrono, Vol. 36, No. 4, pp. 221-231, 1989 222 MADRONO [Vol. 36 lationships of the three major club chollas in Texas, and provides keys and descriptions for all four species (includes O. emoryi) of series Clavatae in Texas. The results of this study provide a clearer taxonomic arrangement concerning the O. schottii complex. The taxonomic and cytogenetic portions of the study are here presented. Taxonomic History Opuntia schottii and O. grahamii were described from collections made during the U.S. and Mexican Boundary Survey of 1851-1853 (Englemann 1856). The type localities for O. schottii and O. grahamii were given as "near the mouth of the San Pedro and Pecos", and "near El Paso", respectively. Britton and Rose (1919) maintained these species. Since that time, however, several authors have altered the tax- onomy at both the specific and generic level. Anthony (1956) de- scribed the putative hybrid, O. schottii x O. grahamii from popu- lations in southern Brewster County, Texas. Benson (1969) reduced O. grahamii to a variety of O. schottii, apparently based on the overlapping ranges and intergrading morphology as reported by An- thony (1956). In his treatment of the complex, Weniger (1984) re- tained Engelmann's taxonomy and disputed Benson's claim of range overlap and intergradation between O. grahamii and O. schottii. Segregate genera that include these taxa have also been proposed (see synonymy), but we find no grounds, morphologic, chromo- somal, or chemical, to support these alternative generic dispositions. Habitat and Distribution Taxa of the O. schottii complex grow in loosely consolidated ig- neous or calcareous desert alluvium, as well as on limestone out- crops. The plants grow on flats or gentle slopes and may be found both in the open or in the shade of desert shrubs, predominantly Larrea tridentata, Prosopis glandulosa, and Acacia spp. In Texas, the O. schottii complex extends from extreme south- central New Mexico, southeastward along the Rio Grande to the Gulf of Mexico (Fig. 1). Opuntia schottii occupies the southern and eastern reaches of this range, from southern Brewster County to Cameron County, whereas O. grahamii occupies the more western and northern regions of the range (i.e., southern Brewster County to El Paso County, and the southeastern edge of New Mexico). The ranges do overlap in southern Brewster County with morphologic intergradation between the species reported there (Anthony 1956; Benson 1982). The two species also occur along the Rio Grande River in adjacent Mexico (Benson 1982). Our study, however, ex- amined the members of the complex as they exist in Texas. The widely disjunct O. emoryi occurs as a single population in extreme 1989] RALSTON AND HILSENBECK: OPUNTIA 223 Fig. 1 . Distribution of taxa in the Opuntia schottii complex in Texas; Opuntia schottii (open stars), O. grahamii (closed stars), O. aggeria (closed circles), and O. emoryi (closed square). southern Presidio County and is not involved in the major taxo- nomic problem surrounding the complex in Texas. Methods and Materials Population samples were collected throughout the geographic range of the complex in Texas, with particular emphasis on the area of reported intergradation in Big Bend National Park (BBNP). Vouch- ers are deposited in SRSC. Loans of herbarium specimens, including types, were obtained from ASU, LL, MICH, MO, POM, RSA, and TEX. Vegetative and floral characters were measured from dried and living material. Bud material for meiotic counts was collected in the field. The buds were fixed in modified Camoy's solution (chloroform, absolute ethanol, and glacial acetic acid, 4:3:1, v:v:v). Results The data in Table 1 disclose that Opuntia schottii is easily distin- guished from O. grahamii by spine length and width, branching architecture, root-type, areole diameter, and relative prominence of tubercles. In Texas, O. schottii grows primarily east of the Pecos MADRONO [Vol. 36 C/3 O u X) in I in 1/5 o H o I o in C/5 n I in B o c o -2 o (N 00 OO I I I >n ^ ^ I I in 00 O I o in in o OS OO I I (U in I o o in I m in (N 7 7 O O I in 13 — ( -t— < o I in o o I I 00 in I I 13 ■(— > O ON I in in ■*-> £ o ^ 4J ^ ^ O c C3 < OD C/2 C C/3 / 1989] RALSTON AND HILSENBECK: OPUNTIA 225 River and flowers much later than O. grahamii, which is found west of the Pecos in the United States. Populations in southern Brewster County, mostly within BBNP, designated by Anthony (1956) as putative hybrids, only occasionally exhibit intermediate character- istics or measurements between O. schottii and O. grahamii (Table 1). Measurements of O. emoryi and O. moelleri A. Berg, are also provided; the relationship of this latter species to the taxa in the complex is addressed below. Field work in southwestern Texas indicates that the ranges of the principal taxa within the complex overlap only in southern Brewster County (Fig. 1). Two herbarium specimens identified as O. schottii {Weedin and Weedin 237, and Worthington 6910.5, both SRSC) suggested that this species was found in BBNP and as far west as El Paso. These specimens are now properly identified as O. grahamii, based on spine, joint, and root morphology. Chromosome numbers, including previously published counts, are listed in Table 2. Chromosome numbers for O. grahamii and O. schottii, are n=22 (Weedin et al. 1989; Pinkava et al. 1985). The «=11 number previously reported for O. schottii is now correctly attributed to the putative hybrid populations from BBNP. Chro- mosome counts made by us for Anthony's putative hybrid popu- lations in and around BBNP (Table 2) reveal that all populations are n=\\. This number is known within the series Clavatae only for these populations, and for the Mexican species O. moelleri (Pinkava and Parfitt 1982). The geographic and morphologic data show that although O. gra- hamii and O. schottii are marginally sympatric in BBNP, they do not intergrade. Additionally, the chromosomal data disclose that the putative hybrid exists at the diploid level, whereas the former two species are tetraploids. These data suggest, therefore, that the O. schottii complex is best treated as three species: O. schottii and O. grahamii, which show no evidence of hybridization, and Anthony's putative hybrid that is herein described as new. Taxonomy Key to Opuntia schottii Complex North of the Mexican Border a Joints ovoid to obovoid; new growth emerging near apex of previous year's growth; spines mostly terete 2. Opuntia grahamii 2l Joints more or less clavate; new growth emerging from sides or bases of previous year's growth; spines mostly flattened. b Spines 7-9 per areole, pink to white/gray; areoles 3-4 mm wide; roots tuberous. 1 . Opuntia aggeria b' Spines 8-16 per areole, yellow to red/brown; areoles 5-7 mm wide; roots fibrous. c Plants to 8 cm high; joints 4.5-6.5 cm long; tubercles 15-20 mm long, 6- 8 mm wide 3. Opuntia schottii c' Plants to 15 cm high; joints 7-15 cm long; tubercules 35-50 mm long, 10- 15 mm wide 4. Opuntia emoryi 226 MADRONO [Vol. 36 Table 2. Chromosome Numbers for Taxa in the Opuntia schottii Complex AND O. MOELLERi. VouchcFs are deposited in SRSC unless otherwise indicated. R = Ralston. Hap- loid num- ber Species in) Locality and Voucher 1 1 1 1 TX Rrpw^//7 Anthony, Amer. Midi. Nat. 55: 239. 1956 (Fig. 2). -Type: USA, Texas, Brewster County, Big Bend National Park, on Tornillo Flats, 2800 ft, 30 Jul 1948, M. Anthony 856 (holotype, MICH!). Plants forming low mound to 10 cm high, 1 m wide. Roots thick- ened, tuberous. Branches creeping; new growth emerging from lateral areoles of previous year's growth. Joints 4-7 cm long, 2.5-3 cm in diamater. Tubercles 10-20 mm long, 8-10 mm wide, 5-7 mm high, green; areoles circular, 3-4 mm wide. Spines 7-9, mostly flattened, pink to white/gray; 3-4 spines per areole 5.5-9 cm long; 4-5 spines per areole spreading, 5-25 mm long; 2-4 radial spines deflexed; glochids numerous to 5-10 mm long. Flowers 5-7 cm long and 4- 5 cm wide. Petaloids in 3-4 whorls, grading from yellow-green with 1989] RALSTON AND HILSENBECK: OPUNTIA 227 Fig. 2. Opuntia aggeria Ralston & Hilsenbeck. A. Habit showing tuberous root, characteristic branching pattern, and distribution of spine clusters. B. Detail of spine cluster. Illustrated from live specimen Ralston 114. central pink tinge on the outer whorls to bright yellow in innermost series, to 25 mm long, 20 mm wide, spatulate, apiculate. Filaments green, to 8 mm long. Style cream, to 3 cm long. Pericarpel narrowly obconic, to 55 mm long, 20 mm wide with areoles bearing glochids. Fruits gray, dry at maturity, to 5 cm long. Seeds brown to cream, to 5 mm in diameter. n=\\. Flowering late March to April. 228 MADRONO [Vol. 36 Pamtypes. USA, Texas, Brewster County, E of Nine-Point Mesa, 3 Aug 1948, M. Anthony 909 (MICH); 15 mi N of Terlingua, along road to Alpine, 14 Sep 1948, M. Anthony 1181 (MICH); flats just N of Santa Elena Canyon, BBNP, 15 Sep 1948, M. Anthony 1246 (MICH). The specific epithet is chosen to describe the clumped or aggre- gated growth habit of this mound-forming species. Phenetically, O. aggeria appears most closely related to O. grahamii by its spine morphology, tuberous root system, and areole diameter, as well as to a species located in northern Mexico, O. moelleri. Comparison of O. aggeria to both O. grahamii and O. moelleri is given in Table 1 . Opuntia moelleri is distributed in Coahuila, Mexico (Britton and Rose 1919; Bravo-Hollis 1978). Morphology (particularly the tu- berous roots), geographic distribution, and the fact that O. aggeria and O. moelleri are the only known diploids in series Clavatae, suggest that these two species, through past hybridization, may be the progenitors of the more northerly distributed, tetraploid O. gra- hamii. Chromosome counts from 12 populations of O. aggeria are all n=\l (Table 2). If O. aggeria was the product of hybridization between O. schottii and O. grahamii as Anthony suggested, it would likely be a tetraploid, or if diploid, accompanied by possible hybrid sterility (cf Ralston 1987). Opuntia aggeria is, however, highly fertile as determined by pollen stainability (Ralston and Hilsenbeck un- publ.) and, being a diploid, would be more or less reproductively isolated from the other two species of club cholla with which it co- occurs. Moreover, O. aggeria only occasionally exhibits characters intermediate between O. grahamii and O. schottii, whereas a true hybrid might be expected to show definite intermediacy, particularly a vegetatively propagated clonal entity as are many of the chollas, including O. aggeria (Grant 1981). The data thus show that the predominant club cholla occurring in BBNP (i.e., Anthony's putative hybrid) does not represent the product of hybridization between O. schottii and O. grahamii and should be formally recognized at the specific level. 2. Opuntia grahamii Engelm., Proc. Amer. Acad. 3:304. — Cor^^- nopuntia grahamii (Engelm.) F. Knuth, Kactus ABC, 116 1935. — Opuntia schottii Engelm. var. grahamii (Engelm.) L. Benson, Cactus & Succ. J. (Los Angeles) 41:124. 1969. — Gru- sonia grahamii (Engelm.) H. Robinson, Phytologia 26:176. 1973. — Type: USA, Texas, sandy soil in the bottom of the Rio Grande, near El Paso. 1851, Wright Opuntia no. 10 (lectotype, MO!). Plants forming low sprawling mounds, 8 cm high, to 3 dm wide. Roots thickened, fleshy, tuberous. Branches creeping, with new 1989] RALSTON AND HILSENBECK: OPUNTIA 229 growth added apically, ascending; joints obovate, 3.5-5 cm long, 1.5-3 cm diameter; tubercles broad, not prominent, to 6 mm wide, 8-12 mm long, 4-6 mm high, green; areoles circular, 3-4 mm wide. Spines 7-14 per areole in upper half of joint. Spines mostly terete, straw-colored, with pink tinge; spine sheaths caducous, to 3 mm long; glochids numerous, increasing in number toward base of joint, to 5 mm long on old joints; 3-4 larger spines per areole, 1.5-5 cm long, spreading; 1-9 shorter spines, 5-25 mm long, spreading; 2-4 of these shorter spines deflexed. Flowers to 5 cm long, 4 cm wide; petaloids in 3-4 whorls grading from yellow with central pink tinge in outer ones to bright yellow in innermost series, to 20 mm long, 15-20 mm wide, spatulate, apiculate; filaments yellow green to 10 mm long; style cream, to 25 mm long; pericarpel obconic, 25-30 mm long and to 20 mm in diameter, with numerous glochids in areoles. Fruits and seeds unknown except in type illustration. n=22. Flowering early May through early June. Many features distinguish O. schottii and O. grahamii (Table 1). Any intergradation through hybridization is now unlikely, as the ranges of the two species are marginally sympatric because of dif- fering ecological preferences, and they differ in phenology as well. 3. Opuntia schottii Engelm., Proc. Amer. Acad. 3:304. 1856.— Corynopuntia schottii (Engelm.) F. Knuth. Kactus-ABC. 114. 1935. — Grusonia schottii (Engelm.) H. Robinson. Phytologia 26:176. 1973. — Type: USA, Texas, Rio Grande, near mouth of Pecos and San Pedro, Sep 1853, ^. Schott s.n. (lectotype, MO!). Plants forming extensive mats to 8 cm high, 5 m wide. Roots fibrous. Branches sprawling, forming long chains, new growth emerg- ing from lateral areoles of previous year's growth; joints to 6.5 cm long, 3 cm diameter; tubercles prominent, 1 5-20 mm long, 6-8 mm wide, and 6-8 mm high, green; areoles circular, to 7 mm wide. Spines 8-14, flattened, reddish brown; spine sheaths to 5 mm long; 3-4 spines per areole 4-6 cm long, with 1 prominent central spine; 2-8 spines per areole shorter, to 30 mm long, spreading, 2-4 spines per areole deflexed; glochids not abundant, to 5 mm long. Flowers 5.5- 6.5 cm long, to 3 cm wide; petaloids in 3-4 whorls grading from yellow green with central pink tinge in outer ones to bright yellow in innermost series, to 22 mm long, 10 mm wide, spatulate, apic- ulate; filaments yellow, to 10 mm long; style cream, to 25 mm long; stigma lobes 5-7, pink tinged; pericarpel narrowly obconic, to 30- 45 mm long, 25 mm wide, with glochids in areoles. Fruits fleshy, yellow, to 45 mm long; areoles on fruits bearing spines and glochids to 5 mm long, fruits often persisting to following year. Seeds cream to brown, to 4 mm wide, with beaked aril. n=22. Flowering mid- June to early July. Opuntia schottii appears most closely related, particularly through 230 MADRONO [Vol. 36 its fibrous root system and flattened spine morphology, to O. emoryi, a species predominantly distributed in Arizona (however, see below). Morphology (Table 1), differing phenology, and its occupation of more mesic habitats primarily east of the Pecos River, easily dis- tinguish O. schottii from the other species of club chollas in Trans- Pecos, Texas. 4. Opuntia emoryi Engelm., Proc. Amer. Acad. 3:303. 1856.— Cactus emoryi Lemaire. Cactees 88. 1868. — (9. stanlyi Engelm. [in Emory, Notes Mil. Reconn., 157, fig. 9. 1848, nom. prov.] ex B. D. Jackson. Index Kewensis 2:358. IS95. — Corynopuntia stanlyi Knuth. Kactus-ABC. 114. \935. — Grunsonia stanlyi (Engelm.) H. Robinson. Phytologia 26:176. 1973.— Type: Mex- ico, arid soil south and west of El Paso, especially between the sandhills and Lake Santa Maria, 1852, Bigelow s.n. (lectotype, MO!, seeds only). Plants forming low sprawling mats to 1 5 cm high, 4 m wide. Roots fibrous. Branches forming chains; new growth emerging from areoles of previous year's growth. Joints 7-15 cm long, 5 cm in diameter; tubercles prominent, 35-50 mm long, 10-15 mm wide, 10-12 mm high, green; areoles circular to 7 mm wide. Spines 1 1-16, flattened yellow to red/brown; 6-8 spines 3.5-7 cm long, 5-8 spines 10-25 mm long. Glochids sparse to 5 mm long. Flowers 5.5-6.5 cm long, and to 3 cm wide; petaloids in 3-4 whorls grading from yellow green with central pink tinge outermost to bright yellow innermost, 25 mm long, 15 mm wide, spatulate, apiculate; filaments yellow, to 10 mm long; style cream, to 25 mm long; stigma lobes 5-7, pink tinged. Pericarpel narrowly obconic, 30-45 mm long, 20 mm wide, with areoles bearing glochids. Fruits and seeds not known for Texas pop- ulation. n=22. Flowering May to early June. In Texas, O. emoryi appears most closely related to O. schottii. Although O. emoryi is primarily known from Arizona, a disjunct population has been recently documented in the Big Bend region of West Texas near Candelaria in southern Presidio County (Weedin and PoweU 1978; Ralston 1987; Ralston and Hilsenbeck in prep.). Opuntia emoryi, also a tetraploid species, is peripheral to the tax- onomy of O. aggeria but is nonetheless an important, recent addition to the Texas flora. The larger size of the plants, including the much larger joints and tubercles, distinguishes O. emoryi from other species in the complex. Further study within the complex, however, and within series Clavatae, that takes into full account the northern Mexico and Arizona taxa is warranted. Acknowledgments The authors wish to thank Drs. A. M. Powell, A. D. Zimmerman, and D. J. Pinkava for their helpful advice throughout this project. Ms. Julia Larke is thanked for the 1989] RALSTON AND HILSENBECK: OPUNTIA 231 illustration and Ms. Rena Gallego for technical assistance. Support for this study was provided by a Texas State Legislature Chihuahuan Desert Studies Grant #1141- 30212-00 awarded to R.A.H. Literature Cited Anthony, M. 1956. The Opuntiae of the Big Bend region of Texas. Amer. Midi. NaturaHst 55:225-256. Benson, L. 1969. The cacti of the United States and Canada— new names and nomenclatural combinations— L Cactus & Succ. J. 41:124-128. . 1982. The cacti of the United States and Canada. Stanford Univ. Press, Stanford, CA. Bravo-Hollis, Helia. 1978. Las cactaceas de Mexico, 2nd ed. Universidad Na- cional Autonoma de Mexico, Ciudad Universitaria, Mexico, DP. Britton, N. L. and J. N. Rose. 1919. The Cactaceae, Vol. 1. Carnegie Inst. Wash. 248. Engelmann, G. 1856. Synopsis of the Cactaceae of the territory of the United States and adjacent regions. Proc. Amer. Acad. Arts 3:259-31 1. Grant, V. 1981. Plant speciation, 2nd ed. Columbia Univ. Press, New York. PiNKAVA, D. J. and B. D. Parfitt. 1982. Chromosome numbers in some cacti of western North America-IV. Bull. Torrey Bot. Club 109:121-128. , M. A. Baker, B. D. Parfitt, and M. W. Mohlenbrock. 1985. Chromosome numbers in some cacti of western North America— V. Syst. Bot. 10:471-483. Ralston, B. E. 1987. A biosystematic study of the Opuntia schottii complex (Cac- taceae) in Texas. M.S. thesis. Sul Ross State Univ. Weedin, J. F. and A. M. Powell. 1978. Chromosome numbers in Chihuahuan Desert Cactaceae: Trans-Pecos Texas. Amer. J. Bot. 65:531-537. , , and D. O. Kolle. 1989. Chromosome numbers in Chihuahuan Desert Cactaceae. IL Trans-Pecos Texas. Southw. Naturalist 34:160-164. Weniger, D. 1984. Cacti of Texas. Univ. Texas Press, Austin. (Received 18 Jan 1989; revision accepted 5 Jun 1989.) ANNOUNCEMENT New Publications Crawford, R. M. M., Studies in plant survival: Ecological case histories of plant adaptation to adversity, Blackwell Scientific Publications, Osney Mead, Oxford OX2 0EL,'l989, x, 296 pp., illus., ISBNs 0-632- 01475-X (hardbound), 0-632-01477-6 (paperbound), prices un- known. [= Studies in Ecology, Vol. 1 1 . Discusses many plant examples for Arctic, montane, desert, coastal, and other areas.] Cullmann, W., E. Gotz and G. Groner, The encyclopedia of cacti, trans, by K. M. Thomas, Timber Press, 9999 SW. Wilshire, Portland, OR 97225, 1986 (publ. 1987), 340 pp., illus. (most color), endpaper maps, ISBN 0-88192-100-9 (hardbound), $49.95. [Publ. in Britain by Alphabooks, Sherborne, same title, 1986. Translation of Kakteen, 2. Aufl., Eugen Ulmer GmbH & Co., 1984. With excellent photos, clear descriptions, and many keys to taxa.] BIOSYSTEMATIC STUDIES OF PHAGE LI A CAP IT AT A (HYDROPHYLLACEAE), A SPECIES ENDEMIC TO SERPENTINE SOILS IN SOUTHWESTERN OREGON J. Stephen Shelly^ Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR 97331 Abstract Studies of the distribution, edaphic restriction, cytology, and morphology of Pha- celia capitata were undertaken to assess the relationship of the species within the Phacelia magellanica polyploid complex. Phacelia capitata is restricted to serpentine or related substrates in Coos, Douglas, and Jackson counties in southwestern Oregon. Chromosome counts of /7=1 1, the base number for the complex, were obtained from 19 of 20 populations sampled. Two populations contained tetraploid («=22) indi- viduals, one of these being a mixed population of diploids and tetraploids. Morpho- logical studies uphold the distinctiveness of P. capitata at the diploid level. The origin of the tetraploids is problematic, but some specimens suggest that P. capitata has hybridized in the past with P. hetewphylla ssp. virgata and plants resembling P. hastata. Owing to their scarcity and geographic restriction these hybridization events appear to have been recent, and they do not challenge the status of P. capitata as one of the most morphologically and ecologically distinct members of the P. magellanica complex. The species appears to be an example of an edaphic neoendemic. Ultramafic soils are those derived from the igneous rocks perido- tite and dunite and their metamorphic derivative, serpentinite. The spectacular influences that these substrates have on the plants that inhabit them are well known, and are of great interest to community ecologists and biosystematists. Serpentine vegetation is frequently dwarfed, and xerophytism is common. Species composition is often unique, and endemism and range disjunction are characteristic of serpentine floras (Kruckeberg 1969, 1984; Whittaker 1954, 1960). In southwestern Oregon, serpentine substrates occur over large areas in Curry and Josephine counties, and in smaller and more widely scattered outcrops to the north and east in Coos, Douglas, and Jackson counties (Fig. 1). Phacelia capitata Kruckeberg is re- stricted to the northeastemmost serpentine outcrops in this region. Phacelia capitata is a member of the P. magellanica (Lam.) Cov. polyploid complex, a wide-ranging group of biennial and perennial species having affinities with the South American species P. secunda Gmel. (=P. magellanica). The members of this polymorphic group ' Present address: Montana Natural Heritage Program, State Library, 1515 E. 6th Avenue, Helena, MT 59620. Madrono, Vol. 36, No. 4, pp. 232-247, 1989 1989] SHELLY: PH ACE LI A CAPITATA 233 • capitata 2n Fig. 1 . Geographic distribution and chromosome numbers of Phacelia capitata and P. corymbosa in southwestern Oregon. Uhramafic formations are outlined (Wells and Peck 1961). Phacelia corymbosa locations are plotted from herbarium specimens housed at OSC, except for the tetraploid location at Sexton Mountain, Josephine County (Heckard 1960). form a polyploid pillar complex (Stebbins 1971) that has a foun- dation comprising seven species with a base chromosome number of x=ll, including P. capitata (Heckard 1960 and pers. comm.). The latter was described by Kruckeberg (1 956). In his detailed study, 234 MADRONO [Vol. 36 however, Heckard (1960) did not assess the interrelationships of P. capitata within the P. magellanica complex. Owing to its restriction to serpentine soils (or closely similar de- rivatives) over a small geographic area, P. capitata has received attention as a potentially threatened species. Most recently it has been placed on the "watch list" of Oregon rare, threatened, and endangered plant species (Oregon Natural Heritage Data Base 1987). In addition, the species is a Category 2 federal candidate, under review for possible listing by the U.S. Fish and Wildlife Service (U.S. Department of Interior 1985). My study was undertaken to assess the distribution, edaphic re- striction, cytology, and morphological characteristics of P. capitata. I also studied natural populations of other species in the complex that occur in southwestern Oregon {P. corymbosa Jepson, P. hastata Douglas ex Lehmann, and P. heterophylla Pursh), to evaluate the interrelationships of P. capitata within the P. magellanica complex. Methods Distribution. I compiled locations for 13 P. capitata populations from specimens deposited at various herbaria (CAS/DS, HSC, JEPS, ORE, OSC, UC, and WTU), from the Oregon Rare, Threatened, and Endangered Plant files housed in OSC, from U.S. Bureau of Land Management offices in Roseburg and Medford, Oregon, and from the Oregon Natural Heritage Data Base in Portland. My field studies were conducted during May-July 1982, April 1983, and April-June 1984. I surveyed as many locations as possible, in both serpentine and non-serpentine areas of southwestern Oregon, to de- termine the location and status of P. capitata populations. Edaphic restriction. To characterize the edaphic environment of the ultramafic areas occupied by P. capitata, I collected soil samples from four fully serpentinized locations, and from one less serpen- tinized site. In addition, I sampled two serpentine locations for P. corymbosa, and one non-serpentine location for P. heterophylla. The soil analyses were conducted by the Soil Testing Laboratory at Or- egon State University, using the extraction and analysis methods of Berg and Gardner (1978). Cytology. Chromosome counts were made from pollen mother cells. Immature helicoid cymes were field-fixed in a mixture of 3 parts 95% ethanol : 1 part glacial acetic acid (V:V) for 24 hours. They were then washed in two overnight changes of 70% ethanol, and frozen in a third change of the same. After warming at 55-60°C for 24-48 hours in an alcoholic hydrochloric acid-carmine stain (Snow 1963), the cymes were rinsed in distilled water and studied imme- diately, or stored in 70% ethanol in a freezer. Anthers were dissected 1989] SHELLY: PHAGE LI A CAP IT AT A 235 onto slides, squashed in 45% acetic acid, and semi-permanently mounted. Morphological studies. A total of 530 living individuals from 26 natural populations in southwestern Oregon was studied. For P. capitata, 420 plants from 20 populations were examined. For com- parison with P. capitata, 1 1 0 plants representing other members of the P. magellanica complex were studied: 80 of P. corymbosa (4 populations), 20 of P. heterophylla ssp. virgata (E. Greene) Heckard (1 population), and 10 of P. hastata (1 population). Data for 1 0 morphological characteristics were collected from each plant studied: blade length (mm), leaf width (mm), width of corolla opening (mm), corolla length (mm), calyx length (mm), stem height (cm), number of leaf lobes, number of floral branches below the terminal cymes, type of leaf lobing, and glandulosity. The five leaf and flower characters were measured with a Zeiss 10 x handlens containing a 10 mm scale. Stem height was measured from the base of the flowering stem to the tip of the terminal cymes. Qualitative assessments of the type of leaf lobing and the degree of glandulosity were made in the following manner: Type of leaf lobing 0 — leaves entire. 1— leaves pinnatifid (lobes with a broad base at the midvein). 2 — leaves pinnately compound (lobes reaching the midvein, with entire leaflets distinguishable). Glandulosity (when viewed with 40 x Zeiss dissecting microscope) 0— eglandular. 1 — a few scattered glandular trichomes among the longer tapering trichomes, mainly in and near the inflorescence. 2— densely and conspicuously glandular, especially on interme- diate-length trichomes in and near the inflorescence, on the pedicels, and on calyces. The mean values for each of the ten morphological characteristics were calculated for each population and subjected to a cluster anal- ysis using CLUSTER, a hierarchical, agglomerative, combinatorial program (Keniston 1978). The Bray-Curtis dissimilarity index was used, and the data were standardized using division by the attribute maximum, in which all values for a given morphological attribute are divided by the maximum value observed for that attribute. This removes "high score bias," and allows use of quantitative and qual- itative values together. A group average fusion strategy was used; this method produces only moderately sharp clustering, but more directly reflects the relationships originally expressed by the dissim- ilarity measure (Boesch 1977). I compared the stem vestiture of P. capitata with that of P. cor- 236 MADRONO [Vol. 36 ymbosa, another serpentine member of the P. magellanica complex, using scanning electron microphotographs prepared by the Oregon State University Scanning Electron Microphotography Lab. Results Distribution. Twenty-two extant populations of P. capitata were located and studied (Table 1, Fig. 1). These populations are asso- ciated with the northeastemmost serpentine outcrops in southwest- em Oregon, in central and southwestern Douglas, southern Coos, and northwestern Jackson counties (Fig. 1). The populations are generally located on slopes with southerly exposures, and occur at elevations from 100 to 1460 m. Six populations of P. corymbosa, two of P. heterophylla ssp. vir- gata, and one of P. hastata were also studied (Table 1). At only one location was P. capitata found growing in biotic sympatry with any other member of the P. magellanica complex; at the Boomer Hill (west) site in Douglas County (the "sub-serpentine" location), it occurs along a disturbed road bank in a mixed population with the plants identified herein as P. hastata. Edaphic restriction. Table 2 presents the results for eight soil anal- yses. Soils at the Boomer Hill (west) site (P. capitata and P. hastata) appear to be derived from metamorphic rocks similar to serpentine, but darker blackish-orange in color. The Tiller sample is from a non- serpentine area adjacent to the serpentine outcrops along Elk Creek in Douglas County. The remaining six samples are for soils derived from fully serpentinized parent materials; the rocks are a shiny, deep greenish-black color. The serpentine soils are all characterized by low levels of phos- phorus and low Ca/Mg ratios; these properties have been well-doc- umented (Kruckeberg 1984; Proctor and Woodell 1975). The sub- serpentine sample has a higher Ca/Mg ratio than the more typical serpentine soils, but it is within the range of values reported by Proctor and Woodell (1975) for such soils. Although large amounts of iron are often found in serpentine soils (Proctor and Woodell 1975), the results do not indicate this for the samples obtained in my study. The highest concentrations of iron are found in the sub-serpentine and non-serpentine samples. The remaining analyses (pH, K, Cu) do not indicate any consistent differences between the serpentine and non-serpentine samples. Ser- pentine soils from the P. capitata and P. corymbosa sites are similar in chemical composition. Cytology. At 1 9 of 20 sites, P. capitata occurs at the base (diploid) chromosome level of n=\ \ (Table 1). Of particular interest are the Elk Creek and Callahan Creek populations, both of which contain 1989] SHELLY: PH ACE LI A CAPITATA 231 Table 1 . Study Populations and Chromosome Numbers for the Phacelia ma- GELLANiCA COMPLEX, SOUTHWESTERN OREGON. All collcction numbers are those of the author (JSS), except where noted. Voucher specimens are deposited in the Oregon State University Herbarium (OSC). Site names denote approximate locations; specific data are available from the author. An asterisk indicates a representative voucher from the same population, the actual chromosome count having been obtained from a separate, unvouchered individual. P. capitata Kruckeberg. /7=1 1— OR, Coos Co.: Bridge, A. Kruckeberg 2703 (WTU, isotype). Douglas Co.: Beatty Creek, 419; Bilger Creek, JSS and M. Nelson 724; Boomer Hill (W), JSS and R. Holmes 766; Boomer Hill (E), JSS and R. Holmes 781; Cow Creek (central), 405; Cow Creek (W), 411; Cow Creek (E), 681; Doe Creek, 706; The Drew, JSS and M. Nelson 734*; Elk Creek, JSS and M. Nelson 726; Lee Creek, 399; Little River, 690*; Myrtle Creek (N), 414; Peel, 354*; SW flank of Red Mountain, JSS and M. Nelson 736; Salt Creek, JSS and M. Nelson 725*; Weaver Road, 689. Jackson Co.: Goolaway Gap, JSS and M. Nelson 738. P. capitata Kruckeberg. «=22.— OR, Douglas Co.: Callahan Creek, 440; Elk Creek, L. Heckard 2930 (JEPS), JSS 433. P. capitata Kruckeberg. Chromosome number unknown— OR, Douglas Co.: Doe/ Thompson Ridge, JSS and R. Holmes 783; Rice Creek, JSS and R. Holmes 792. P. corymbosa Jepson. «=22— OR, Jackson Co.: Grave Creek, 441. Josephine Co.: Eight Dollar Mountain (S), 700; Eight Dollar Mountain (N), 702; Rough and Ready Creek, 698; Waldo, 697; Wimer Road, L. Constance and R. Bacigalupi 3394 (WTU). P. hastata Douglas ex Lehmann (see comments in Results section). /7=22— OR, Douglas Co.: Boomer Hill (W), JSS and R. Holmes 767. P. heterophvlla Pursh ssp. virgata (E. Greene) Heckard. «=11— OR, Douglas Co.: Cow Creek, 714; Elk Creek, JSS and M. Nelson 728. tetraploid individuals; the Elk Creek population also contains dip- loids. Thus, all seven species in the P. magellanica complex that occur as basic diploids are now known to have morphological coun- terparts at the tetraploid {n=22) level. Two previous tetraploid counts for P. corymbosa in Oregon were cited by Heckard (1960). These, plus five additional counts obtained during this study, indicate that P. corymbosa may occur primarily at this level in Oregon. Diploids of this species are known to occur in the North Coast Ranges of California, and in one locality in the Sierra Nevada (Heckard 1960). Three diploid counts for P. heterophylla in southwestern Oregon were previously recorded (Heckard 1 960). The two additional counts reported here indicate that this species (as ssp. virgata) is primarily diploid in Oregon, although it occurs at the tetraploid level elsewhere in its range, mainly as ssp. heterophylla (Heckard 1960). The tetraploid count reported here for P. hastata is apparently the first from southwestern Oregon. Although it was initially known only as a tetraploid, diploid individuals of this species have recently been discovered in northern Nevada (L. Heckard pers. comm.). It should be noted that the population identified throughout this paper 238 MADRONO [Vol. 36 z o o u O z u H U H O C/5 U H on H X o w o 1/3 u >- < o < :3 S 00 00 00 o o C/5 3 a o D c C/5 "G r-) (U C (u (u (u C (U c c c (U t/5 C/5 C/5 1/5 2:i X •1^: -J-/ £ oo Q (U O C C ;=! ID pa W Oh o o o in o On 0^ On (N (U (U C C • »— ( • 1— I (U lU && (U (U 1/5 t/5 B S _ (u * ■ I o in in in oo C 1/5 o (U U 1989] 0.3 n SHELLY: PHACELIA CAPITATA 239 0.2- < ifi 0.1 I <<<<<<<<<<<<<<<< C/5 o ^ o c ?3 q I o I in O I O l-i c« (U •*-< +-< 03 CJ I/) o l/T o o 5i (U r2 U O x =1 •rH c T) •'-I V3 u O c ^ O 03 o.S r^ U( 03 (U 03 X u 03 ■rH fll ---H Oj c/5 03 iS 'c3 ""^ O (N --^ C I ^ 03 ^ O ^H ^ C/l ^ ^x S o V3 03 kH (D t/3 o 6 x; t£ C ^ 03 g M o 03 Oh 03 I o o t-( ^ O 03 Is .-3 i-H 35 o O 03 C t/3 .S m ^ 2 - 03 G 60 a, 03 c — ' C3 ^ ^ c o ^ ^ I- 03 (/3 1^ ^ rH ■B § S " 03 «5 > ^ n 03 JS .5 T3 «^ S (U . 1/3 Vh Wh QJ o (/3 c 03 03 o I I 3 £ 03 o V3 o > o a > o TO ^ 2 (u (U en TO C 1 T3 O C/3 (U (U • g-g ^ o Sh 03 •t^ C 5 (3 a 244 MADRONO [Vol. 36 panied by introgression, has led to very complicated, and sometimes indecipherable, patterns of variation in the P. magellanica complex. This is especially true, he states (Heckard 1960, p. 33), because . . the initial differences are not of a magnitude to furnish char- acters which can be conveniently measured. The result of this in- tergradation is an extensive assemblage of interrelated plants within which lines must be drawn somewhat arbitrarily in order to delimit taxonomic units." Of the seven species in the P. magellanica complex that occur predominantly or partially as diploids, two were considered in detail in this study: P. capitata and P. corymbosa. Each of these is a good example of a "morphological and ecological extreme," having very distinct morphological characteristics and special ecological attri- butes in the form of virtually complete restriction to serpentine soils. However, of all seven species that are now known to occur at both the diploid and tetraploid levels in the P. magellanica complex, P. capitata has the most restricted geographical range. Phacelia capitata was found growing sympatrically with other members of the P. magellanica complex at only two locations. The only instance of biotic sympatry with another member of the group was found at the Boomer Hill (west) site, where, as described above, it occurs along a roadbank in a mixed population with the plants resembling P. hastata. Chromosome counts from this site showed that P. capitata occurs as a diploid and that the P. hastata-like plants are tetraploids, a situation that ordinarily limits the formation of fertile interspecific hybrids. Some collections (i.e.. Shelly and Holmes 768, 77 1, OSC) are intermediate, however, in their possession of relatively narrow, predominantly entire, silvery leaves. Attempted chromosome counts for these intermediate plants were unsuccessful, as all the buds were past meiosis. On the basis of their morphological intermediacy, however, it appears that hybridization occurred in the past. Of interest at this site is the less stringent nature of the soil, which is intermediate between the highly serpentinized habitats to which P. capitata is otherwise restricted, and a non-serpentine soil profile. This, plus the disturbed nature of the roadside habitat, are possibly the factors which allowed P. capitata and the P. hastata- like plants to become sympatric here. The Elk Creek population of P. capitata, containing diploids and tetraploids, further exemplifies the problems involved in determin- ing the origins of hybrid individuals in the P. magellanica complex. On the basis of field observations, the most likely other diploid species involved in this intergradient population is P. heterophylla. This species is neighboringly sympatric with P. capitata, occurring on non-serpentine soils along the highway both north and south of the serpentine outcrops above Elk Creek. Surveys throughout the 1989] SHELLY: PHAGE LI A CAPITATA 245 area did not reveal any other morphologically distinct members of the P. magellanica complex. Numerous P. capitata plants at both Elk Creek and the nearby Callahan Creek site are morphologically suggestive of past hybridization with P. heterophylla, particularly in the possession of pinnately compound leaves, which are character- istic of the latter species. The cluster analysis upholds the morphological distinctiveness of diploid populations of P. capitata in comparison with P. corymbosa (4/7), P. hastata (4/?), and P. heterophylla (2n). The clustering of the two tetraploid P. capitata populations with the tetraploids resem- bling P. hastata is most likely due to similar tendencies toward longer, entire leaves, larger corollas, similar stem height, and an intermediate degree of glandulosity amongst the individuals. Despite the evidence for past genetic interaction of P. capitata with P. heterophylla and with plants resembling P. hastata, it is appropriate to interpret these instances of limited interspecific gene exchange as exceptional events. Phacelia capitata remains a unique entity in the P. magellanica complex, as evidenced by the morpho- logical, ecological, and almost wholly diploid cytological distinc- tiveness of the species. Although the exact evolutionary history of P. capitata remains unknown, it seems most appropriate to categorize it as a serpentine neoendemic species. Such species are newly arisen "insular" taxa, often demonstrating a pattern of adaptive radiation into specialized habitats (Kruckeberg 1984). Additionally, because P. capitata would have evolved at the diploid level within the P. magellanica complex, it conforms well to the diploid speciation model for the establish- ment of an edaphic endemic species as described by Kruckeberg (1986). In this scheme non-serpentine populations, which may be preadapted for serpentine tolerance, give rise via disruptive selection and subsequent genetic divergence and isolation to new biological species. The extent to which this process involved a diploid ancestor shared with P. corymbosa, if any, is enigmatic. In fact, no very close relative of P. capitata is apparent within the P. magellanica complex. No evidence of genetic interaction between P. capitata and P. corymbosa was found, despite their virtually complete restriction to similar serpentine substrates. The two species were found to be fully allo- patric, P. capitata replacing P. corymbosa on the northeasternmost ultramafic outcrops in southwestern Oregon. In addition, P. cor- ymbosa is not known to occur as a diploid in Oregon. The north- ernmost known diploid location for P. corymbosa is in Siskiyou County, California, approximately 175 km south of the southern- most location of P. capitata (Heckard 1960). Lastly, the two species are both distinct morphological entities in the P. magellanica com- 246 MADRONO [Vol. 36 plex, and are very different from one another. If a common diploid ancestor was shared by the two species, it appears to have long been extinct. Field studies revealed that in many places P. capitata has increased in abundance following habitat disturbance. Owing to this ruderal response, the species is not currently threatened or endangered with extinction. It is still important, however, to protect undisturbed populations of such narrowly endemic species so that features of their population ecology, habitat requirements, and evolutionary uniqueness may be studied in natural environments. Acknowledgments This study was completed in partial fulfillment of the requirements for the degree of Master of Science, Oregon State University. I thank Kenton Chambers for his guidance. Support, research, and travel funds were provided by the Department of Botany and Plant Pathology and the Oregon State University Herbarium. Field as- sistance from Mark Nelson and Russ Holmes, and helpful advice from Robert Meinke, is gratefully acknowledged. LaRea Johnston assisted in the preparation of the pho- tographs. Ginger King provided invaluable assistance in the preparation of the figures. I thank the Editor, L. R. Heckard, A. R. Kruckeberg, and an anonymous reviewer for helpful comments on the manuscript. Literature Cited Berg, M. G. and E. H. Gardner. 1978. Methods of soil analysis used in the Soil Testing Laboratory at Oregon State University. Agricultural Experiment Station, Corvallis. Special Report No. 321. BoESCH, D. F. 1977. Application of numerical classification in ecological investi- gations of water pollution. U.S. Environmental Protection Agency, Corvallis, Oregon. Ecological Research Series Report No. EPA-600/3-77-033. Heckard, L. R. 1960. Taxonomic studies in the Phacelia magellanica polyploid complex. Univ. Calif Publ. Bot. 32:1-126. Keniston, J. A. 1978. Program CLUSTER: an aid to numerical classification. Oregon State University, Marine Science Center, Newport (unpubl.). Kruckeberg, A. R. 1956. Notes on the Phacelia magellanica complex in the Pacific Northwest. Madrono 13:209-221. . 1969. Plant life on serpentinite and other ferromagnesian rocks in north- western North America. Syesis 2:15-1 14. . 1984. California serpentines: flora, vegetation, geology, soils, and manage- ment problems. Univ. California Press, Berkeley. . 1986. An essay: the stimulus of unusual geologies for plant speciation. Syst. Bot. 11:455-463. Oregon Natural Heritage Data Base. 1987. Rare, threatened, and endangered plants and animals of Oregon. The Nature Conservancy, Portland. Proctor, J. and S. R. J. Woodell. 1975. The ecology of serpentine soils. Adv. Ecol. Res. 9:255-366. Snow, R. 1963. Alcoholic hydrochloric acid-carmine as a stain for chromosomes in squash preparations. Stain Tech. 38:9-13. Stebbins, G. L. 1971. Chromosomal evolution in higher plants. Edward Arnold Ltd., London. U.S. Department of Interior. 1985. Endangered and threatened wildlife and plants; review of plant taxa for listing as endangered or threatened species; notice of review. Federal Register 50:39525-39584. 1989] SHELLY: PHAGE LI A CAP IT AT A 247 Wells, F. G. and D. L. Peck. 196L Geologic map of Oregon west of the 121st meridian. U.S. Geol. Survey Misc. Geol. Inv. Map 1-325. Whittaker, R. H. 1954. The ecology of serpentine soils. IV. The vegetational response to serpentine soils. Ecology 35:275-288. . 1960. Vegetation of the Siskiyou Mountains, Oregon and California. Ecol. Monogr. 30:279-338. (Received 8 Aug 1988; revision accepted 28 Jun 1989.) ANNOUNCEMENT Second Annual Conference The Society for Ecological Restoration The Society for Ecological Restoration announces its second annual conference, to be held at the Sheraton International Hotel, at O'Hare, in Chicago, 29 April-3 May 1990. The program will include several special sessions to explore the state of the art as it applies to key en- vironmental issues, and a full program of contributed papers and post- ers, special lectures, workshops, field trips, and other special events designed to facilitate communication among restorationists and with decision makers and the general public. Special programs will include: Prairie Restoration, Restoration and Global Climate Change, Setting Standards for Monitoring Restoration Projects, Restoration and Recovery of Endangered Species, and Res- toration Philosophy. Field trips will include visits to the prairie resto- ration project at Fermi National Laboratory, the Des Plaines River and wetland restoration project, Indiana Dunes National Lakeshore projects, Chicago's urban prairies, and the University of Wisconsin Arboretum and Society offices in Madison. The Society invites submission of abstracts of papers dealing with all aspects of ecological restoration. Special consideration will be given to papers directly related to the special sessions listed above, but papers dealing with any aspect of ecological restoration are welcome. These may include political, administrative, social, economic, and philosoph- ical, as well as purely scientific and technical aspects. Forms for sub- mission of abstracts may be obtained from the Society's office: S.E.R., 1207 Seminole Highway, Madison, WI 53711, (608) 263-7889. The deadline for submission of Abstracts for Contributed Papers is 1 5 Jan- uary 1990. A SYSTEMATIC AND PHYTOGEOGRAPHIC STUDY OF ANTENNARIA AROMATICA AND A. DENSIFOLIA (ASTERACEAE: INULEAE) IN THE WESTERN NORTH AMERICAN CORDILLERA Randall J. Bayer University of Alberta, Department of Botany, Edmonton, Alberta T6G 2E9, Canada Abstract Antennaria aromatica and A. densifolia are closely related species occurring in disjunct, unglaciated regions of the Rocky Mountain cordilleran system. The species are morphologically distinct and are readily distinguished based on a number of characters. They both occur on predominantly limestone talus slopes from subalpine to alpine elevations. Indications from the current distribution of the taxa and the known extent of glaciation suggest that they once had more extensive ranges and that these were subsequently reduced by the Wisconsinan glaciation. Glaciation has played a major role in the phytogeographic history of the species, as well as in their evolu- tionary divergence. A population of A. densifolia from Montana, disjunct from the main group of populations in the Northwest Territories and Yukon, is a noteworthy collection. Antennaria aromatica Evert and A. densifolia A. Pors. are two narrowly restricted endemics of the Alpinae, a group of arctic/alpine Antennaria centered in western North America. Both occur on talus, primarily of limestone origin, but in different regions of the western North American cordillera. Antennaria densifolia, first described from the MacKenzie Mountains of the Northwest Territories, is apparently most common on the east slope of the MacKenzie Moun- tains and in the Ogilvie Mountains and southern Richardson Moun- tains of the Northwest Territories and Yukon Territory. Evert (1984) recently described A. aromatica from the Rockies of Montana and Wyoming (type from the Beartooth Pass near Quad Creek, Mon- tana). Examination of herbarium material and the meager literature indicates that the morphological, ecological and distributional as- pects of these two species are not well understood. Distributions of the species indicate that they once may have been more widespread and that their phytogeographic history, as well as their evolutionary divergence, was influenced greatly by Pleistocene glaciation. Com- parison of specimens of both species indicates a close morphological similarity. Recently, a single disjunct population of A. densifolia has been discovered in a remote area of Montana. This, coupled with the ostensible morphological and ecological similarity of the two species, makes a detailed analysis of them appropriate. Madrono, Vol. 36, No. 4, pp. 248-259, 1989 1989] BAYER: ANTENNARIA 249 Table 1. List of 36 Characters and Character States used in the Morpho- METRic Analysis of Antennaria aromatica and A. densifolia. Numbers following each character indicate the scale used. All qualitative state determinations are in mm. Basal rosette characters: 1. Length of entire basal leaf. 2. Maximum width of the basal leaves. 3. Length, along the mid- vein, from leaf tip to the maximum width. 4. Shape of the anterior margin, i.e., length from tip to widest point in the leaf. 5. Number of leaves per basal rosette. Stolon characters: 6. Number of leaves per stolon. 7. Length of the largest leaf 8. Width of the largest leaf 9. Length of the smallest leaf 10. Width of the smallest leaf 1 1. Stolon length. 12. Number of stolons per basal rosette. Cauline (flowering) stem characters: 13. Rowering stem height. 14. Number of leaf nodes per cauline stem. 15. Width of the longest leaf 16. Length of the longest leaf 17. Width of the shortest leaf 18. Length of the shortest leaf 19. Presence of a scarious flag-like structure at the apex of the upper leaves, 0.0 = absent, 1.0 = present. Pistillate capitulescence characters: 20. Height of the involucre. 21. Number of heads per capitulescence. 22. Phyllary length. 23. Phyllary width. 24. Corolla length. 25. Pappus length. 26. Achene length. 27. Phyllary colors, 1.0 = green base, white tips, 2.0 = green base, rose middle, white tips, 3.0 = green base, black/brown middle, white tips, 4.0 = green base, brown middle, rose tips, 5.0 = green base, brownish tips, 6.0 = brown base, white tips, 7.0 = brown base, rose tips, 8.0 = brown base, umber tips, 9.0 = brown base, dark brown tips, 10.0 = brown base, black or very dark green tips. 28. Number of florets per head. Staminate capitulescence characters: 29. Height of the involucre. 30. Number of heads per capitulescence. 31. Phyllary length. 32. Phyllary width. 33. Corolla length. 34. Pappus length. Miscellaneous characters: 35. Presence of stalked glands on the surfaces of stems, leaves, etc., 0.0 = absent, 1.0 = present. 36. Presence of staminate individuals in the population, 0.0 = staminates absent (i.e., population all pistillate), 1.0 stami- nates present. Methods Specimens deposited at ALA, ALTA, CAN, DAO, ID, MONTU, RM, and UAC were examined for the morphological and phyto- geographic information. Chromosome counts were obtained from several collections of A. aromatica and one of A. densifolia utilizing the Feulgen chromosome staining methods described in Bayer (1 984). A principal components analysis (PCA) was used to quantify the morphological differences between the species. Forty specimens rep- resenting the range of morphological diversity from throughout the known range of each species were selected for morphometric anal- ysis. Thirty-six vegetative and reproductive features (Table 1) were measured on twenty specimens of each taxon. In most cases spec- imens, having both staminate and pistillate individuals, were used, but complete specimens could not always be included because of the scarcity of good material. The original data matrix is available from the author upon request. The NTSYS-pc program (Numerical taxonomy and multivariate analysis systems for the IBM-PC microcomputer and compatibles; 250 MADRONO [Vol. 36 vers. 1.2; F. James Rohlf 1987) was used to compute the PCA. The STAND subroutine was used to standardize the data such that each character had a mean of zero and a standard deviation of unity. A similarity matrix of product-moment correlations was derived using the SIMINT subroutine of NTSYS-pc. The EIGEN subroutine was employed to compute the eigenvalue and eigenvector matrices. Three factors were extracted by the EIGEN subroutine. The OTU's were subsequently projected onto axes (the eigenvectors) using the PRO J subroutine of NTSYS-pc, thus concluding the principal components analysis. A 3-dimensional graph of the OTU's onto the first three principal components was plotted using the MOD3DG subroutine of NTSYS-pc. Results Distribution and cytology. The distribution of all available her- barium records for both taxa is presented in Figure 1 . Additionally, the distribution of herbarium records having staminate, as well as pistillate specimens, and those having only pistillate specimens, is indicated. Although absence of staminate plants, as determined from gender ratios in natural populations, is a dependable indicator that the plants in the population are gametophytic apomicts (Bayer and Stebbins 1983), the lack of staminate plants on herbarium specimens provides only weak evidence because it could be a consequence of the failure of the collector to gather both genders in a particular population. The base chromosome number in Antennaria is x=\4. The only previous chromosome report (diploid) for A. densifolia is a Yukon population (Chmielewski and Chinnappa 1988). One population of A. densifolia from Montana has been determined as diploid (Table 2). Three populations of A. aromatica have been counted as diploid, whereas seven are tetraploid (Fig. 2, Table 2). One population oiA. aromatica from the Big Horn Mountains has been counted as hexa- ploid, or 2^2=84 (Fig. 2, Table 2), which is a new number for the species. The three diploid populations of A. aromatica, including the type locality, are from the southern part of the species range (Fig. 2). The tetraploid populations are more widespread, occurring from Fig. 1 . Distribution of Antennaria aromatica (circles) and A. densifolia (squares) in western North America (each symbol may depict one or more collections). Open symbols indicate collections containing pistillate and/or staminate individuals, while closed symbols signify collections having only pistillate plants. The continental divide is indicated by the dot/dash line and the Wisconsinan glacial maximum, following Prest (1984), is shown by "T" shaped symbols. Bar = 500 km. 252 MADRONO [Vol. 36 Table 2. Synopsis of Chromosome Numbers for Eleven Populations of Antennaria aromatic a and Two Populations of^. densifolia. Presented are state/ province, county or mountain range, and voucher number. Report reference codes: a = current study (vouchers at ALT A), b = Bayer (1984), c = Bayer and Stebbins (1987), and d = Chmeilewski and Chinnappa (1988). * = new number for the species. ** = population presumably asexual, containing only pistillate individuals. State or province County or mountain range Voucher number Somatic number Refer- ence codes A. aromatica: Montana Carbon Co. 8092 28 b Montana Gallatin Co. MT-628 28 a Wyoming Park Co. WY-836 28 a Montana Cascade Co. MT-747 56 - a Montana Gallatin Co. 8105** 56 b Montana Judith Basin Co. MT-754 56 a Montana Lewis and Clark Co. MT-890 56 a Montana Madison Co. MT-634** 56 a Montana Teton Co. M-302 56 c Wyoming Park Co. WY-626** 56 a Wyoming Big Horn Co. WY-813** 84* a A. densifolia: Montana Granite Co. MT-725 28 a Yukon Ogilvie Mtns. 2642/2643 28 d the northern part of the range in Teton Co., Montana south to Madison Co., Montana and Park Co., Wyoming (Fig. 2). Morphometries. The 3 -dimensional graph displaying the first three factors from the principal components analysis (PCA) is presented as Figure 3. The first three axes account for 57.06% of the variation and the first 10 factors have individual eigenvalues greater than 1.000, indicating that the variables are not highly correlated. Factor 1 has highest loadings for the characters of presence/absence of glands (Table 1; character 35), pistillate phyllary length (character 22); and various leaf length characters (characters 1, 7, and 9). Stolon length (character 1 1), number of stolons/basal rosette (character 12), and the presence/absence of staminate clones (character 36) have high loadings in factor 2. Factor 3 has high loadings for staminate corolla length (character 33) and length of the basal rosette leaves (character 1). Antennaria aromatica and A. densifolia form two dis- tinct groups (Fig. 3) and disassociate best along factor 1. The diploid and tetraploid cytotypes of A. aromatica appear to form two group- ings along factor 2. A trend toward the separation of sexual and asexual populations of A. aromatica, as expressed by presence/ab- sence of staminate plants, is evident on factor 2. Although two 1989] BAYER: ANTENNARIA 253 Fig. 2. Distribution of cytotypes of Antennaria awmatica and A. densifolia in Mon- tana and Wyoming. Open circles = diploid, closed circles = tetraploid, and star = hexaploid populations of A. awmatica. The open square is a diploid population of A. densifolia. The continental divide is indicated by the dot/dash line and the Wis- consinan glacial maximum, following Prest (1984), is shown by "T" shaped symbols. Bar = 250 km. subgrouping are apparent within the A. densifolia cluster (Fig. 3), it is not certain whether these two represent different ploidy levels. Discussion The direct relationship of A. awmatica to A. densifolia is indicated by similarity of morphology and ecology. They are part of a larger complex of arctic/alpine Antennaria (Alpinae group) and conse- quently are related to A. umbrinella Rydb. andyl. pulchella E. Greene (Bayer 1987b). Using a taxonomic species concept, A. awmatica and A. densifolia can be considered discrete species because they are morphologically distinct. The PCA demonstrates this conclusively, and indicates that the presence of stalked glands (and the coincident citronella odor in the living plants) in A. awmatica and the complete lack of these glands in A. densifolia distinguishes them reliably. Three other characters can be used to separate the two species. Most in- dividuals of^. densifolia possess flat, scarious, linear-lanceolate tips, termed "flags", at the ends of the upper cauline leaves (character 18; Table 1), however most A. awmatica lack these flags. The basal 254 MADRONO [Vol. 36 Fig. 3. PCA composed of 20 OTU's of Antennaria aromatica (circles) and 20 of A. densifolia (squares). Open symbols indicate collections containing pistillate and/or staminate individuals, whereas closed symbols signify collections having only pistil- late plants. An isotype of A. aromatica (ALTA) and the holotype specimen of A. densifolia (CAN) are marked with "T". Specimens of known chromosome number are labeled "2" for diploid and "4" for tetraploid. The one disjunct diploid population of A. densifolia from Montana is the open square labeled "2". rosette leaves (character 1) m A. aromatica are 6.0-16.0 mm long (mean 9.4 ± 2.9), whereas those of A. densifolia are 3.0-6.0(-7.0) mm long (mean 4.7 ± 1.1). Antennaria aromatica has pistillate phyllaries that are (4.0-)5.0-7.0 mm long (mean 5.6 ± 0.76) and those of A. densifolia are 3.0-5.0 mm long (mean 4. 1 ± 0.58). These three characters are not as reliable as presence/absence of glands to separate the taxa, but the character suite is relatively reliable. Antennaria aromatica was first described (Evert 1984) from the Beartooth Mountains of Montana, but has since been collected from several additional sites in Montana, Wyoming, and Alberta. The species is known only from ca. 40 collections (including several duplicate collecting efforts), but is perhaps more common within its rather restricted range than the scant herbarium record would in- dicate because of the relative inaccessibility of habitats where these plants are known to occur. Several older collections exist in herbaria, but they were not recognized as belonging to a new species. They are most often misidentified as A. alpina (L.) Gaertner, A. media E. Greene, or A. umbrinella. Since the description of A. aromatica, there has been some con- fusion about its identity and geographic range. In 1898, Greene described Antennaria pulvinata from Moose Mountain in the Front 1989] BAYER: ANTENNARIA 255 Range of the Rockies west of Calgary, Alberta (Bayer 1989). This taxon is entirely pistillate and resembles^, aromatica closely, except it lacks the glandular hairs. These apomicts are polyploids that un- doubtedly have A. aromatica as one of their sexual progenitors, but probably have other sexual parents as well (Bayer in press a). I have included these apomicts, and others like them, in the A. rosea E. Greene polyploid complex, calling them A. rosea subsp. pulvinata (E. Greene) R. Bayer (Bayer 1989). Recently, Weber (1987) and Hartman and Rottman (1988) reported ^. aromatica from Colorado, but I would instead identify these plants as A. rosea subsp. pulvinata. Authentic A. aromatica is a cushion plant that occurs on talus slopes, almost always composed of limestone. The populations occur from just below treeline to alpine, but, in my experience, are often found just at the treeline with the krummholz vegetation. The plants are viscid, and the younger foliage has a strong citronella odor. The populations are usually sexual, having approximately equal propor- tions of staminate and pistillate clones (Fig. 1). Some of the common associates of A. aromatica are Androsace chamaejasme, Anemone multiflda, Aquilegia jonesii, Arctostaphylos ma-ursi, Arenaria obtu- siloba, Astragalus gilviflorus, Astragalus kentrophyta, Cymopterus hendersonii, Draba oligosperma, Dryas octopetala, Eriogonum an- drosaceum, Eritrichium nanum, Haplopappus uniflorus, Hedysarum alpinum, Hulsea algida, Ivesia gordonii, Paronychia sessiliflora, Pe- dicularis parryi, Phlox caespitosa, Phlox hoodii, Physaria didymo- carpa, Potentilla fruticosa, Ribes oxycanthoides, Valeriana edulis, and Zygadenus elegans. All populations of A. aromatica occur east of the continental di- vide, predominantly in areas that were unglaciated during the Wis- consinan (Fig. 1). Antennaria aromatica could be restricted to the Front Ranges east of the divide as a result of climatic factors, these mountains receiving less rainfall than those to the west in western Montana and eastern Idaho. The main range of the species is from north-central Wyoming, near Cody, north to Glacier National Park, Montana (Fig. 1). Two minimally disjunct populations of A. aro- matica occur in the Alberta Rockies, which extends the range of the species to the Mountain Park area east of Jasper, Alberta (Fig. 1). These two collections (ALTA #27162 and #32291) are the first of this species from Canada. The collection from Mountain Park is especially significant because this is not only an area with large areas of limestone outcrops, but is an area that has been identified as a glacial refugium (Packer and Vitt 1974; Prest 1984) and is known for its large number of species disjunctions. Antennaria aromatica occurs as both diploid and polyploid races (Table 2). The PCA demonstrates that the two groups (diploids and polyploids) are slightly distinct morphologically (Fig. 3). Although the two cytotypes cannot be reliably distinguished on morphological 256 MADRONO [Vol. 36 grounds, the diploids are mostly smaller in all respects when com- pared to the polyploids (pistillate involucres in the 3 confirmed diploids are 5.0-6.0 mm high, whereas those of four tetraploids are (6.0-)6.5-7.5 mm high). Similarly to other species of Antennaria, populations of the A. aromatica diploid cytotypes have equal pro- portions of staminate and pistillate clones (Bayer unpubl. obs.). Some of the tetraploid populations have both staminate and pistillate plants, whereas others have only pistillate plants (Table 2, Fig. 2). The only known hexaploid population is entirely pistillate (Fig. 2, Table 2). Antennaria aromatica conforms to a pattern that appears in several other species of Antennaria, such as A. marginata and A. media, one in which populations of the diploid cytotype within a species are sexually reproducing. Populations of the polyploid cytotype, which are for the most part morphologically indistinguishable from the diploids, are either sexually or asexually reproducing. The diploids frequently have a much more restricted range than the polyploids and are often found strictly in unglaciated regions, whereas the poly- ploids have wide ranges in both glaciated and unglaciated areas (Bayer and Stebbins 198 1). Asexual polyploids are also the ones that usually colonize the areas farthest away from the center of the species range and/or into glaciated terrain. The distribution of cytotypes within A. aromatica (Fig. 2) corresponds well to this pattern. Antennaria densifolia was originally described by Porsild (1945) from the east slope of the MacKenzie Mountains, Northwest Ter- ritories. Porsild (1945) stated that it superficially resembled A. pul- vinata E. Greene and A. compacta Malte. In Porsild's keys to species for arctic Canada (Porsild 1950; Porsild and Cody 1980), the means of distinguishing A. compacta from A. densifolia is by the lack of staminate individuals in A. compacta, and its longer, more linear, basal leaves. The habitat of A. densifolia is similar to that of A. aromatica; i.e., calcareous talus, ranging from subalpine through treeline to alpine. Specimen data suggest that A. densifolia occurs most often in alpine situations, and A. aromatica, most often in the subalpine or at tree- line. In constructing the distribution maps of A. densifolia, I identified some specimens lacking staminate plants as A. densifolia although by using Porsild's key (Porsild 1950) they are A. compacta. These specimens have the shorter, cuneate leaves of^. densifolia and prob- ably represent the disregard of staminate plants on the collectors' part. Most of these collections were made near other sites where staminates are known to occur (Fig. 1). About 40 populations con- taining both staminate and pistillate individuals are now confirmed, but many of these sites are very close together along the only roads in the areas, the Dempster highway over the Ogilvie and Richardson mountains and the north Canol road over the MacKenzie Mountains 1989] BAYER: ANTENNARIA 257 (unfortunately, the road is now abandoned and impassable in the region where A. densifolia is known). The large number of specimens from the southern MacKenzie Mountains in the Nahanni National Park were probably collected during the initial vegetation surveys of this wilderness park. The distribution of A. densifolia (Fig. 1) is in two slightly disjunct areas, the eastern slopes of the MacKenzie Mountains, and the northern Ogilvie and southern Richardson mountains. Although A. densifolia has not been collected in the remote interlying region, the valley-glaciated Wernecke Mountains, it is possible that it does occur there because dolomitic limestone outcrops exist in this range (Oswald and Senyk 1977). The northern Ogilvies and southern Richardsons had only limited alpine glacia- tion during the Wisconsinan (Prest 1984). The eastern slope of the MacKenzies were part of the northern end of the western Canadian ice-free corridor (Rutter 1984), but perhaps had restricted alpine glaciers during the Wisconsinan. The majority of the A. densifolia populations containing staminate plants are located within the un- glaciated regions, only three populations being located in previously glaciated terrain (Fig. 1). One very noteworthy collection was our {Bayer, DeLuca, and Lebedyk MT-725 at ALT A and RM or Lackschewitz 4611 at MON- TU) recent discovery of a population of ^. densifolia in the Anaconda Range of Granite Co., Montana (Figs. 1 and 2) found growing in gravelly, limestone talus on open alpine tundra. This site represents not only the first collection of the species for the United States, but a substantial (1850 km) disjunction from the nearest population in the Northwest Territories. Morphologically the plants from Mon- tana are well within the range of typical A. densifolia from Canada (Fig. 3) and are diploid, the same as the only available count for northern Canadian A. densifolia (Table 2). Although only diploid counts have been obtained for^. densifolia, it is probable that some tetraploid populations exist, as two weakly distinct subgroups within A. densifolia are present in PCA (Fig. 3). It is possible that this species repeats the pattern of morphologically indistinguishable diploid and sexual polyploid races present within species. Additional chromosome counts are needed to confirm or reject this hypothesis. Antennaria densifolia is similar to A. aro- matica, in that a group of apomicts {A. compactd) derived from A. densifolia are morphologically very similar to their parent. In a similar manner, there are polyploid apomicts related to A. aromatica {A. rosea subsp. pulvinata). The taxa can be identified by means of the following key. Key to Antennaria aromatica, A. compacta, and A. densifolia A Basal leaves linear to linear-oblanceolate, two or more times longer than wide, arising from densely tufted basal caudices A. compacta 258 MADRONO [Vol. 36 A' Basal leaves cuneate to cuneate-spathulate, less than two times longer than wide, densely caespitose and arising from short, prostrate stolons. B Stalked glands present on flowering stalks, phyllaries, and leaves; odor of citronella present in living plants; flags usually lacking from upper cauline leaves, basal leaves mostly 6.0 mm or more long; pistillate phyllaries mostly 5.0 mm or more long A. aromatica B' Stalked glands absent from flowering stalks, phyllaries, and leaves; living plants odorless; flags usually present on upper cauline leaves; basal leaves mostly less than 6.0 mm long, pistillate phyllaries mostly less than 5.0 mm long. . . A. densifolia The distributions o^A. aromatica and^. densifolia (Fig. 1) provide clues to the phytogeographic and evolutionary history of the two species. It is probable that both species may once have had much wider ranges, and the advance of the Cordilleran and Laurentide ice sheets during the Wisconsinan obliterated the greater part of their former ranges. Sexual dioecism evidently prevents swift migration, as few sexual species ofAntennaria have extensive ranges in glaciated terrain (Bayer and Stebbins 1987). The solitary population of A. densifolia in Montana near the southern end of the ice-free corridor (Fig. 1) may indicate that the species once occurred more widely as far south as Montana and was eliminated from the intervening areas of British Columbia and Alberta. The main body of populations of A. densifolia remaining today are in the unglaciated area at the northern end of the corridor (Fig. 1). Antennaria aromatica exists at the southern end of the corridor, but the two Alberta populations (Fig. 1) that rest within the corridor itself indicate that the species may have once had a much larger range. The populations near the ice margin and within the ice-free corridor probably survived in situ, because, as pointed out by Packer and Vitt (1974), suitable habitats for other endemic calceophiles exist in the surrounding glaciated region, yet the species that are found within the refugial area in Mountain Park, Alberta have not migrated to them. Antennaria aromatica and A. densifolia are obviously very closely related and it is probable that they evolved from the same common ancestor (Bayer in press b). Glaciation was probably the vicariant event that separated the series of populations, and the ensuing iso- lation perhaps facilitated further divergence. Acknowledgments I thank the curators at ALA, CAN, COLO, DAO, ID, MONT, MONTU, and RM for their assistance in obtaining specimens for study. This research was supported by a grant from the Natural Sciences and Engineering Research Council of Canada (NSERC grant #A3797) and a grant from the University of Alberta Central Research Fund (NSERC). I thank Montana botanists Klaus Lackschewitz and Peter Lesica for sending me their additional collections of A. aromatica. I also appreciate the duplicate material of A. aromatica contributed by Ronald Hartman. I am most grateful to Brett Purdy for his helpful input in this study and for compiling Figures 1 and 2. I thank Meredith Lane, LuDean Marvin, John Packer, Richard Pimentel, Brett Purdy, and 1989] BAYER: ANTENNARIA 259 Ledyard Stebbins for providing helpful comments on the manuscript. C. C. Chinnappa is gratefully acknowledged for allowing me to verify the identification of his diploid collection (UAC) of A. densifolia from Yukon. Literature Cited Bayer, R. J. 1 984. Chromosome numbers and taxonomic notes for North American species of Antennaria (Asteraceae: Inuleae). Syst. Bot. 9:74-83. . 1987a. Evolution and phylogenetic relationships of the Antennaria (Aster- aceae: Inuleae) polyploid agamic complexes. Biol. Zentralbl. 106:683-698. . 1987b. Morphometric analysis of western North American Antennaria Gaertner (Asteraceae: Inuleae). I. Sexual species of sections Alpinae, Dioicae, and Plantaginifoliae. Canad. J. Bot. 65:2389-2395. . 1989. A taxonomic revision of the Antennaria rosea (Asteraceae: Inuleae: Gnaphaliinae) polyploid complex. Brittonia 41:53-60. . In press a. Investigations into the evolutionary history of the Antennaria rosea (Asteraceae: Inuleae) polyploid complex. PI. Syst. Evol. . In press b. A phylogenetic reconstruction of Antennaria Gaertner (Aster- aceae: Inuleae) Canad. J. Bot. and G. L. Stebbins. 1981. Chromosome numbers of North American species of Antennaria Gaertner (Asteraceae: Inuleae). Amer. J. Bot. 68:1342-1349. and . 1983. Distribution of sexual and apomictic populations of Antennaria parlinii. Evolution 37:555-561. and . 1987. Chromosome numbers, patterns of distribution, and apomixis in Antennaria (Asteraceae: Inuleae). Syst. Bot. 12:305-319. Chmielewski, J. G. and C. C. Chinnappa. 1 988. Taxonomic notes and chromosome numbers in Antennaria Gaertner (Asteraceae: Inuleae) from arctic North Amer- ica. Arctic and Alpine Res. 20:1 17-124. Evert, E. F. 1984. A new species of Antennaria (Asteraceae) from Montana and Wyoming. Madrofio 31:109-112. Hartman, E. L. and M. L. Rottman. 1988. The vegetation and alpine vascular flora of the Sawatch Range, Colorado. Madrono 35:202-225. Oswald, E. T. and J. P. Senyk. 1977. Ecoregions of Yukon Territory. Canadian Forestry Service, Pacific Forest Research Center, Victoria. Packer, J. G. and D. H. Vitt. 1974. Mountain Park: a plant refugium in the Canadian Rocky Mountains. Canad. J. Bot. 52:1393-1409. PoRSiLD, A. E. 1945. The alpine flora of the east slope of MacKenzie Mountains, Northwest Territories. Nat. Mus. Canad. Bull. 101, Ottawa. . 1950. The genus Antennaria in northwestern Canada. Canad. Field-Nat. 64:1-25. and W. J. Cody. 1980. Vascular Plants of the Continental Northwest Ter- ritories. Nat. Mus. Canad., Ottawa. Prest, V. K. 1984. The late Wisconsinan glacier complex. In R. J. Fulton (ed.), Quaternary stratigraphy of Canada— a Canadian contribution to IGCP project 24. Geological Survey of Canada paper 84-10, Ottawa. RoHLF, F. J. 1987. NTSYS-pc: numerical taxonomy and multivariate analysis sys- tem for the IBM PC microcomputer (and compatibles). Applied Biostatistics, Inc., Setauket. RuTTER, N. W. 1 984. Pleistocene history of the western Canadian ice-free corridor. In R. J. Fulton (ed.), Quaternary stratigraphy of Canada— a Canadian contri- bution to IGCP project 24. Geological Survey of Canada paper 84-10, Ottawa. Weber, W. A. 1987. Colorado flora: western slope. Colorado Associated Univ. Press, Boulder. (Received 25 Jan 1989; revision accepted 28 Jun 1989.) A RE-EVALUATION OF BEALIA MEXICANA (POACEAE: ERAGROSTIDEAE) Paul M. Peterson Department of Botany, National Museum of Natural History, Smithsonian Institution, Washington, DC 20560 Abstract Based on morphological and cytological evidence, Bealia is recognized as a genus with a single species, B. mexicana. This genus may be related to Dasyochloa and Erioneuron as the three share a base chromosome number of x=8 and a number of morphological features. A key distinguishing Bealia, Dasyochloa, Erioneuron, and Muhlenbergia is presented. A full description, illustration, and citation of specimens examined are given for Bealia mexicana. Resumen En base a la evidencia morfologica y citologica Bealia es reconocido como un genero monotipico, conteniendo la especie B. mexicana. Bealia parece estar relacio- nado a Dasyochloa y Erioneuron por el hecho de que los tres generos comparten el mismo numero basico de cromosomas (.x=8) y otros tantos caracteres morphologicos. El presente articulo provee una clave para los generos Bealia, Dasyochloa, Erioneuron y Muhlenbergia. Presentase ademas una descripcion comprensiva de Bealia mexi- cana, asi como su ilustracion y citacion de ejemplares estudiados. During investigations of the annual species of Muhlenbergia Schreber (Peterson 1988a, b, 1989a,b; Peterson and Rieseberg 1987; Peterson et al. 1989; Peterson and Annable 1 990) it became apparent that a segregate genus should be recognized for Muhlenbergia biloba. New information, particularly from chromosome cytology, coupled with numerous unusual morphological characters, supports its place- ment in Bealia. The purpose of this paper is to reinstate Bealia as a genus, discuss possible relationships to other eragrostoid genera, and give a complete taxonomic account. The binomial Bealia mexicana was first published, without a de- scription or a diagnosis, by Vasey (1889) and was based on material collected by Pringle 819 from the mountains of Chihuahua and Brandegee 16 from Santa Margarita Island, Baja California Sur. Charlotte Reeder (1956) later showed that the specimen Brandegee collected from Santa Margarita Island represented a new species, Muhlenbergia brandegei C. Reeder. I agree with C. Reeder that M. brandegei should currently reside in Muhlenbergia, however, the placement of the Chihuahuan material is the topic of this paper. The first valid publication of Bealia mexicana was that of Scribner in Real (1896) who based it on Pringle 8 1 9 from Chihuahua, Mexico. Madrono, Vol. 36, No. 4, pp. 260-265, 1989 1989] PETERSON: BEALIA MEXICANA 261 Jones (1912) transferred the species to Epicampes, naming it E. mexicana, but Hitchcock (1 9 1 3) and subsequent workers have treat- ed it as a species of Muhlenbergia, M. biloba. The only other combination published in Bealia, B. speciosa (Va- sey) Beal, was based on Palmer 30 from southwestern Chihuahua. Hitchcock (1913). Soderstrom (1967), and I agree with Vasey in treating it as Muhlenbergia speciosa Vasey. Jones (1912) also trans- ferred this species to Epicampes, naming it E. speciosa. Epicampes has been recognized as a section of Muhlenbergia subgenus Podo- semum since Soderstrom's (1967) work. Bealia mexicana has a base chromosome number of x=8 (2/7=16) and relatively large chromosomes when compared with other species Muhlenbergia (Peterson 1988a). This base number has not been reported for species oi Muhlenbergia (J. Reeder 1967, 1968; Peterson 1988a). The common base chromosome number in Muhlenbergia is based on 10, although M.filiformis (Thurb.) Rydb. and M. vaginata Swallen possess 9 pairs. Scribner and Vasey recognized the distinctive morphological fea- tures that distinguish Bealia mexicana from most other members of Muhlenbergia. It has deeply bilobed lemmas with obtuse lobes and pilose to villous glumes that are single-nerved and longer than the lemma. Only M. argentea Vasey and M. lucida Swallen from southern Chihuahua have deeply bilobed lemmas with obtuse lobes. These two species are currently being investigated as potential mem- bers of the genus. In my investigations of the annual species of Muhlenbergia, A UPGMA cluster analysis utilizing 80 morphological and chemical characters depicted B. mexicana as a distinct species (Peterson 1 988a). The cluster phenogram shows B. mexicana with a large phenetic distance of 1 .4 and places it near M. crispiseta and M. peruviana, whose intra-cluster phenetic distance is very small at 0.55 (pheno- gram distance range is 0.3 to 1.7). These later two species are in the section Clomena and superficially resemble B. mexicana, but differ in having glabrous, three-nerved and three-toothed second glumes. The glumes of B. mexicana are pilose to villous, single-nerved, and entire near the apex. Bealia mexicana shows close affinities with Dasyochloa Willd. ex Rydb. (interpreted as a monotypic North American genus) and Erio- neuron Nash (includes four species in North America) by sharing a base chromosome number of x=S and relatively large chromosomes (Tateoka 1961; Peterson 1988b). The lemmas of Dasyochloa, Eri- oneuron, and Bealia are similar. All are three-nerved, emarginate to bilobed, often awned, and pilose with hairs that are associated with either the nerves, margins, and/or lower two-thirds of the lem- ma. However, Bealia differs from Dasyochloa and Erioneuron in having spikelets with a single floret, soft membranous spikelets. 262 MADRONO [Vol. 36 lemmas with crisped curled to flexuous awns, glabrous rachillas, fusiform caryopses, and an annual habit. The unusual morphological characters in B. mexicana (Fig. 1) of a deeply bilobed lemma with obtuse lobes, pilose to villous glumes that are single-nerved and longer than the lemma, and a chromosome number of «=8, support the original treatment of Bealia as a distinct genus. The following key distinguishes among Bealia, Dasyochloa, Erioneuron, and Muhlenbergia using gross morphological features. A Spikelets with one floret, rarely two. B Lemmas deeply bilobed, the lobes rounded to obtuse, the lobes 1-1.4 mm long; awn crisped-curled to flexuous, borne between the lobes Bealia B' Lemmas not deeply bilobed, sometimes minutely bifid, then the teeth usually acuminate to aristate, the teeth less than 1 mm long; awn usually straight to flexuous or awnless Muhlenbergia A' Spikelets with two to many florets. C Panicle short and capitate, usually exceeded by a fascicle of leaves; plants low, creeping, usually stoloniferous Dasyochloa C Panicle exserted, open or contracted; plants spreading, erect or decumbent but not stoloniferous Erioneuron Bealia Scribner in Hackel, True Grasses 103. 1890. Bealia mexicana Scribner in Beal, Grasses N. Amer. 2:267. 1896.— Bealia mexicana Scribner in Vasey, nom. nud. Proc. Calif Acad. Sci. II. 2:212. \SS9.— Bealia mexicana SchhnQv in Hackel, nom. nud. True Grasses 103. \%90.—Epicampes mexicana (Scribner in Beal) M. E. Jones, Contr. W. Bot. 14:7. 1912.— Muhlenbergia bilobaK. Hitchc, Contr. U.S. Natl. Herb. 17:294. 1913.-Type: MEXICO. Chihuahua: thin soil of dry porphyry mts., near Chi- huahua, 7 Oct 1886, Pringle 819 (lectotype selected herewith, US!; isolectotypes, MO!, NME!, NY!, UC!, US!, VT!, WIS!). Slender, tufted annuals (Fig. la). Culms much branched at the lower nodes, often caespitose, scabrellous-pubescent, striate, gla- brous to minutely pubescent below the nodes, 9-35 cm tall; 0.3-0.5 mm diam. just below the inflorescence, intemodes hollow, short, 2- 10 mm long. Sheaths keeled, often striate, scaberulous to almost glabrous, longer than the internodes, usually 2-7.5 cm long; margins wide and scarious. Ligules membranous (Fig. lb), 1.5-3.4 mm long, apex acute or rounded, often irregularly toothed; margins entire, decurrent, usually splitting to form auricles no longer than the ligule (Fig. Ic). Blades flat, involute upon drying, 1-7 cm long, 0.6-1.4 mm wide, scaberulous below and short pubescent above, midnerve and margins whitish-thickened on the abaxial surface. Inflorescence an open panicle (Fig. Ic), 3-10 cm long, 1.5-3.2 cm wide; branches 1 or sometimes 2 per node, 1.8-4.5 cm long, sinuously ascending or flexuous and spreading up to 40° from the culm axis; pedicels 1989] PETERSON: BEALIA MEXICANA 263 Fig. 1 . Bealia mexicana. Chihuahua, Mexico {Peterson andAnnable 5800). a. Habit, b. Ligule. c. Inflorescence, d. Spikelet. e. Glumes, f. Floret, g. Lemma, h. Palea. i. Stamens, pistil, and lodicules. j. Caryopsis. 264 MADRONO [Vol. 36 slender, scabrous and minutely glandular, 1-5.5 mm long, loosely ascending; nodes per inflorescence 6-16. Spikelets erect to loosely spreading, 1 floret per spikelet (Fig. Id, f), disarticulating above the glumes. Glumes equal in length (Fig. le), grayish-green to almost bicolored with the lower Va grayish-green and upper % whitish-green, obtuse, loosely pilose to villous, especially on lower %^ faintly 1 -nerved, 3.2-4.8 mm long, as long or longer than the lemma. Lem- mas 3-nerved, 2.9-3.5 mm long loosely to densely appressed-villous on the lower % (Fig. Ig), hairs tawny, up to 0.7 mm long; lateral nerves evident on lower half; apex deeply bilobed, 1-1.4 mm deep, lobes rounded to obtuse; awn born between the 2 lobes, scabrous and crisped-curled to flexuous, 4-6.5 mm long. Paleas 2-nerved (Fig. Ih), 2.6-3.4 mm long, about as long as lemma, apex obtuse, loosely to densely appressed-villous on the lower %, hairs like those on the lemma. Stamens 3 (Fig. li), anthers 1.2-2.3 mm long, purplish. Caryopsis (Fig. Ij) ca 1.8 mm long, fusiform, olive-brown. Chro- mosome number, n=%. Habitat, distribution, and phenology. Shallow, sandy, whitish soils derived from calcareous parent material on flat escarpments asso- ciated with rock outcrops in pinyon-juniper woodlands and yellow pine forests, 2000-2300 m elevation. Central Chihuahua to northern Durango, Mexico, known from very few localities. Flowering from Sep through Oct. Specimens cited. MEXICO. Chihuahua: Majalca, 16 Sep 1935, LeSueur Mex-026 (GH, MO, UC); 15 mi E of El Vergel on road to Parral, 21 Oct 1959, Correll and Gentry 23270 (GH, MO, UC); Sierra Madre Occidental, 1.2 mi W of Cumbres de Majalca, 22.6 mi W of Hwy 45, Campamento in Parque Nacional, 20 Sep 1986, Peterson and Annable 4529 (ARIZ, ENCB, GH, MEXU, MICH, MO, NMC, NY, RSA, TAES, UC, US, UTC, WIS, WS), 22 Sep 1988, Peterson and Annable 5800 (US); 12 mi S of Villa Matamoros, 27 Sep 1988, Peterson and Annable 5976 (US). Durango: Barranca below Sandia Station, 12 Oct 1905, Pringle 10147 (GH, UC, MO); Sierra Madre Occidental, ca. 6 mi SW of El Ojito and 40 mi SW of Parral on Hwy 24, 24 Sep 1986, Peterson and Annable 4570 (ARIZ, ENCB, GH, MEXU, MICH, MO, NMC, NY, RSA, TAES, UC, UNLV, US, UTC, WIS, WS). Acknowledgments This study was supported by grants from the National Science Foundation to Amy Jean Gilmartin and PMP (BSR-86121 1), Sigma Xi, WSU, and the Smithsonian Institution. Special thanks are given to Carol R. Annable for assistance in the field and Alice Tangerini for providing the illustration. I am grateful to Carol R. Annable, Mary E. Barkworth, Amy Jean Gilmartin, Stephen L. Hatch, and David J. Keil for critically reading the manuscript and Pedro Acevedo for preparing the Spanish ab- stract. 1989] PETERSON: BEALIA MEXICANA 265 Literature Cited Beal, W. J. 1896. Grasses of North America, Vol. 2. Henry Holt, New York. Hitchcock, A. S. 1913. Mexican grasses in the United States National Herbarium. Contr. U.S. Natl. Herb. 17:181-389. Jones, M. E. 1912. New species and notes. Grass notes. Contr. W. Bot. 14:1-21. Peterson, P. M. 1988a. Systematics of the annual Muhlenbergia (Poaceae). Ph.D. dissertation. Washington State Univ., Pullman. . 1988b. Chromosome numbers in the annual Muhlenbergia (Poaceae). Ma- droiio 35:320-324. . 1989a. Lemma micromorphology in the annual Muhlenbergia (Poaceae). Southw. Naturalist 34:61-71. . 1989b. Muhlenbergia majalcensis (Poaceae: Eragrostideae), a new species from Chihuahua, Mexico. Syst. Bot. 14:316-319. and C. R. Annable. 1 990. Systematics of the annual Muhlenbergia (Poaceae: Eragrostideae). Syst. Bot. Monographs (in press). , , and V. R. Franceschi. 1989. Comparative leaf anatomy in the annual Muhlenbergia (Poaceae). Nordic J. Bot. 8:575-583. and L. H. Rieseberg. 1987. Flavonoids of the annual Muhlenbergia. Bio- chem. Syst. Ecol. 15:647-652. Reeder, C. G. 1956. Muhlenbergia brandegei, a new species from Baja California, Mexico and its relationship to Muhlenbergia biloba. Madroiio 13:244-252. Reeder, J. R. 1967. Notes on Mexican grasses VL Miscellaneous chromosome numbers. Bull. Torrey Bot. Club 94:1-17. . 1 969 Notes on Mexican grasses VIIL Miscellaneous chromosome numbers. Bull. Torrey Bot. Club 95:69-86. SoDERSTROM, T. R. 1967. Taxonomic study of subgenus Podosemum and section Epicampes of Muhlenbergia (Gramineae). Contr. U.S. Natl. Herb. 34:75-189. Tateoka, T. 1961. A biosystematic study of Tridens (Gramineae). Amer. J. Bot. 48:565-573. Vasey, G. 1889. Gramineae. Pp. 2 1 0-2 1 4 in T. S. Brandegee, A collection of plants from Baja California. Proc. Calif. Acad. Sci. IL 2:117-216. (Received 30 Jan 1989; revision accepted 3 Jul 1989.) ANNOUNCEMENT New Publication Fryxel, p. a., Malvaceae of Mexico, Systematic Botany Monographs, Vol. 25, pp. [i-ii], 1-522, color fp. 13 Dec 1988, ISSN 0737-8211, ISBN 0-9 1 286 1-25-8 (hardbound), $40.00 U.S., $42.00 foreign, post- paid (from Systematic Botany Monographs, University of Michigan Herbarium, Ann Arbor, MI 48109-1057, with checks payable to "ASPT"). [A monumental opus, on 55 gen. (incl. Allosidastrum, gen. nov.), 372 spp. (184 endemic), lOinfraspecifictaxa, with introductory sections on endemism, diversity, familial subdivisions, specialized characters, common names, the massive taxonomic part (448 pp.!), and concluding biblio., appendices (chief references on important collectors of Mexican plants; new sections in Hibiscus, by O. J. Blan- chard), species list, and indices to specimens and scientific names.] A NEW SPECIES OF DAPHNOPSIS (THYMELAEACEAE) FROM BAJA CALIFORNIA SUR, MEXICO Dennis E. Breedlove Department of Botany, California Academy of Sciences, San Francisco, CA 94118 Jose Luis Leon de la Luz Centro de Investigaciones Biologicas de Baja California Sur, La Paz, Baja California Sur, Mexico 23000 Abstract Daphnopsis lagunae, a new species from the highest ridges of the Sierra de la Laguna, Baja Cahfornia Sur, Mexico, is described. This is a locally restricted species which is the first record for the genus and the family (Thymelaeaceae) for the peninsula of Baja California. Resumen Se describe una nueva especie de los picachos de la Sierra de la Laguna, Baja California Sur, Mexico: Daphnopsis lagunae. Se trata de una especie con habitat sumamente restringido, es tambien la unica representante del genero y de la familia (Thymelaeaceae) para la peninsula de Baja California. Fruiting specimens of this Daphnopsis were first collected by M. E. Jones in 1930. Subsequently, Breedlove and Axelrod in 1977 and Leon de la Luz in 1985 made several additional fruiting collections. Although the plant was obviously related to Daphnopsis no attempt could be made to further identify it without flowers. Leon de la Luz, in the course of his studies on the vegetation of the Sierra de la Laguna (Leon de la Luz and Dominguez 1989), returned to the region three times in spring and summer and finally collected flowering plants in mid- August of 1987 immediately following the first sum- mer rains. It is a curious aside that T. S. Brandegee, who made exhaustive collections in the Sierra de la Laguna at the end of the last century, never encountered this distinctive shrub. A search of the herbarium at the University of California, where the major set of T. S. Brandegee's collections are deposited, was made to no avail. The short period of anthesis which occurs for 6 to 1 0 days following the inception of the erratic summer rains and the very local nature of this plants distribution best accounts for its absence from his collection. Daphnopsis lagunae is the only member of the Thyme- laeaceae known from Baja California (Wiggins 1980). Daphnopsis is the largest genus, with about 50 species, in the family in the New Madrono, Vol. 36, No. 4, pp. 266-270, 1989 1989] BREEDLOYE AND LEON: DAPHNOPSIS LACUNAE 267 World and has three areas with concentrations of species, Mexico, the Antilles and Brazil (Nevling, 1959). Daphnopsis lagunae Breedlove & Leon de la Luz, sp. nov. (Fig. 1).— Type: MEXICO. Baja California Sur: Municipio of La Paz, Sierra de la Laguna, Pine Oak forest along trail from La Laguna toElPicacho, 1900 m, 23°3rN, 110°02'W, 14 Aug 1987 (female plant), Leon de la Luz 2730 (holotype Cas; isotypes CIB, UC, MEXU). Frutex deciduus usque ad 1.5 m altus. Folia coriacea, 4-8 cm longa, 2-5 cm lata, margine revoluta. Inflorescentae pistillatae et stamineae portatae in pedunculi 1-5 mm longa cum 1-5 floribus. Flores staminales cum calycis tubus anguste obconicus, 6.5-7 mm longa, 1-1.5 mm lata basi, 3.5-4 mm lata orificii, extra tomentosus, intus glaber, petala 8, circa 0.3 mm longa, staminibus in planus 2, disco annulari. Flores pistillates cum calycis tubus 5-6 mm longa, 1.5-2 mm lata basi, 2-2.5 mm lata orificii. Deciduous, sparsely branched, dioecious shrubs to 1.5 m tall, young branches appressed tomentose. Leaves alternate, estipulate, sessile, coriaceous, 4-8 cm long, 2-5 cm wide, elliptic to lanceolate, glabrous to appressed tomentulose; apex acute to rounded; base cuneate to rounded; margins entire and revolute; venation pinnate with 8-10 ascending primary veins, secondary veins prominently reticulate and partially obscuring the primary veins, midvein prom- inent beneath, all veins reddish brown on the under surface, im- mersed and lightly differentiated on the upper surface. Inflorescence an umbel borne from leafy stems of recent growth, rachis appressed pubescent. Staminate inflorescence with the peduncle 8-12 mm long; staminate flowers 2-3(-5) per inflorescence; pedicel 0.5-1 mm long; calyx tube narrowly obconic, 6.5-7 mm long, 1-1.5 mm broad at base, 3.5-4 mm broad at orifice, appressed tomentose outside, gla- brous within; calyx lobes unequal, puberulent within, the outer lobes 2 mm long, 1.5 mm broad at base, the inner lobes 2 mm long, 1 mm broad at base, all acute; petals 8, papilliform, about 0.3 mm long, inserted above and to each side of the 4 alternisepalous sta- mens; stamens 8, inserted at 2 levels; antisepalous stamens inserted at the level of the orifice, exserted; alternisepalous stamens inserted 1 mm below the orifice, included; filaments 0.2 mm long; anthers oblong-ovate, 0.7-1 mm long, 0.3-0.5 mm broad; disc annular, irregularly adnate to the calyx tube base, undulate to irregularly lobed, glabrous, free marginally; pistillode ten pin shaped, 0.7-1.5 mm long, glabrous. Pistillate inflorescences with the peduncles 5- 12 mm long; pistillate flowers 1-4 per inflorescence; pedicels 4-6 mm long; calyx tube obconic to narrowly campanulate, 5-6 mm long, 1.5-2 mm broad at the base, 2-2.5 mm broad at orifice, ap- 268 MADRONO [Vol. 36 pressed tomentose outside, glabrous within; calyx lobes subequal, 1-1.5 mm long, 1-1.5 mm broad at the base, puberulent within, apex acute; petals 8, papilliform; staminodia 8, papilliform, difficult to distinguish from the petals; pistil 5.5-6.5 mm long, the stigma capitate, exserted. Drupe ovoid, 8-1 1 mm long, apiculate, 1-4 ma- turing per inflorescence, calyx persistent. Seed smooth, brown, spherical, 4-6 mm across; hilum a small depressed black dot on the proximal end. Paratvpes MEXICO, Baja California Sur, type locality: 30 Oct 1985, Dommguez 37 (CAS, CIB) fruit; 12 Aug 1987, Leon de la Luz 2678 (CAS, CIB) male; Leon de la Luz 2679 (CAS, CIB) female; 14 Aug 1987, Leon de la Luz 2731 (CAS, CIB) female. Trail from La Burrera to La Laguna: 24 Sep 1930, M. E. Jones 27276 (CAS) fruit; 22 Oct 1977, Breedlove and Axelrod 43308 (CAS) fruit, Breedlove and Axelrod 43376 (CAS) fruit, 14 Aug 1987, Leon de la Luz 2677 (CAS, CIB) male; Leon de la Luz 2732 (CAS, CIB) male; 28 Aug 1987, Leon de la Luz 2817 (CAS, CIB) female. Cerro Verde: 12 Sep 1986, Leon de la Luz 2032 (CAS, CIB) fruit. 1989] BREEUhOVE AND LEON. DAPHNOPSIS LAGUNAE 269 4 cms. inode. c, branch in flower, d, branch in fruit. Floral tube drawings by Colleen Sudekum, habit and ovary drawn by J. L. Leon de la Luz. Distribution. Known from three populations in the Sierra de la Laguna. The largest population occurs near the summit of El Picacho (2000 to 2150 m) and extends south as scattered individuals along the ridge facing the Pacific Ocean to at least the point where the trail from La Burrera crosses to La Laguna where a second large popu- lation exists. A small population with only a few plants has been found near the summit of El Cerro Verde. Habitat. All of the populations occur on unstable, rocky, granitic soils in both shaded and exposed situations on steep slopes or on relatively flat ridge tops. The associated species include: Arbutus peninsularis Rose & Goldman, Arracacia brandegeii Britton & Rose, Bernardia lagunensis (M. E. Jones) Wheeler, Calliandra peninsularis Rose, Cyclanthera tamnoides Cogn., Helianthus similis (Brandegee) S. F. Blake, Lepechinia hastata (A. Gray) Epling, Mimosa xantii A. Gray, Mirabilis jalapa L., Nolina beldingii Brandegee, Pinus lagunae (Passini & Bailey) Passini, Quercus devia Goldman, Quercus tubercu- lata Liebm., Tagetes lacera Brandegee, Tephrosia canna Brandegee, Verbesina pustulata M. E. Jones, and others. 270 MADRONO [Vol. 36 Discussion Nevling (1959) in his excellent revision of the genus Daphnopsis does not delimit species groups as such and states that species that have the same level of development of petal type may or may not be related. Daphnopsis lagunae with eight petals and a few-flowered inflorescence seems morphologically distant from the species that is closest geographically, D. mexiae Nevling of Nayarit and Sinaloa, which has the petals connate into a faucal annulus and an umbellate inflorescence with up to 55 flowers. Of the seven species with eight petals, one from the Antilles, three from northern South America and three from Mexico, most are from wet forest locations and none share three or more of the key characters used by Nevling. Daphnop- sis lagunae appears to be unique within the genus. Its markedly deciduous habit, few-flowered inflorescence and eight petals clearly set it apart from other species. Acknowledgments The authors wish to express special thanks to Annetta Carter for her encouragement and advice in the preparation of this manuscript. Dr. Arturo Gomez-Pompa, Director of the UC MEXUS Program, provided support for Leon de la Luz to visit the herbaria of the University of California and the California Academy of Sciences to begin this study. This research has been partially funded by a grant from CONACYT-SPP, Mexico. Literature Cited Leon de la Luz, J. L. and R. Dominguez 1989. Rora of the Sierra de la Laguna, Baja California Sur, Mexico. Madrono 36:61-83. Nevling, L. I., Jr. 1959. A revision of the genus Daphnopsis. Ann. Missouri Bot. Card. 46:257-358. Wiggins, L L. 1980. Flora of Baja California. Stanford Univ. Press. 1025 pp. (Received 9 May 1989; revision accepted 28 Aug 1989.) ANNOUNCEMENT New Publication Goodrich, S. and E. Neese, Uinta Basin flora, USDA Forest Service— Intermountain Region, Ogden, UT, 1986, [iii], xvii, 320 pp., 1 foldout map, unillus., no ISBN, paperbound, price unknown. [Reproduced from single-spaced camera-ready copy. On ca. 1 660 specific and sub- specific taxa from a ca. 40,000 km^ area in Utah and Colorado. For review see J. Major, Fremontia 15(4):30.] MONARDELLA BENEOLENS (LAMIACEAE), A NEW SPECIES FROM THE CREST OF THE SOUTHERN SIERRA NEVADA, CALIFORNIA James R. Shevock Department of Botany, California Academy of Sciences, San Francisco, CA 941 18-4599 Barbara Ertter University and Jepson Herbarium, University of California, Berkeley, CA 94720 James D. Jokerst Department of Botany, California Academy of Sciences, San Francisco, CA 941 18-4599 Abstract Monardella beneolens, a new species from the crest of the southern Sierra Nevada in Inyo, Kern, and Tulare counties, California, is described and illustrated. The new species is ecologically and morphologically closest to M. cinerea of subalpine and alpine habitats in the San Gabriel Mountains, Los Angeles and San Bernardino counties, southern California. The subsessile, ovate, undulate-crisped leaves; abun- dant short-stalked glandular hairs on the leaves, bracts, and calyces; and longer spreading eglandular hairs throughout collectively distinguish M. beneolens from all other species of Monardella. The existence of this distinctive, sweet-smelling Monardella first came to the authors' attention in June 1986 during a botanical collecting expedition to Owens Peak. A subsequent search of Monar- della collections at CAS, DS, JEPS, RSA, UC, and UCSB revealed five collections of comparable material, labelled as M. odoratissima Benth. subsp. parvifolia (E. Greene) Epling. Among these was an undetermined vegetative specimen collected by C. A. Purpus in July 1896 on Olancha Peak that represents the first collection of this new species. It was later collected at Olancha Peak in 1950 by J. T. Howell and in 1975 by James Tatum, who was doing a flora of the peak for his master's thesis (Tatum 1979). Clare Hardham, a student of Mo- nardella and author of several endemic California species in this genus, annotated Howell's collections as an "undescribed species" in 1975. She was never able to study mature material and conse- quently did not formally describe the species. Monardella beneolens Shevock, Ertter, & Jokerst, sp. nov. (Fig. 1).— Type: USA, California, Kern Co., near the summit of Owens Madrono, Vol. 36, No. 4, pp. 271-279, 1989 272 MADRONO [Vol. 36 Fig. 1. Monardella beneolens.—A, habit; B, leaves with close-up of vestiture; C, outer leafy verticillaster bract; D, flower with glandular droplets at corolla tip; E, internal structure of corolla with enlargement of anther; F, nutlet with cross-section. Peak on granitic and metamorphic scree and bedrock, open mixed conifer forest, eastern slope below the Sierran crest, T26S R37E sect. 21, 8200 ft. (2500 m), 13 Jul 1986, Shevock, Bartel, and York 11727 (holotype: CAS; isotpyes: FSC, MO, NY, RSA, UC). 1989] SHEVOCK ET AL.: MONARDELLA BENEOLENS 273 Ab aliis Monardella combinatione habitu laxe rhizomatosa, glan- dulis abundis stipitatus trichomatibus villosis, foliis ovatis sessilibus crenatis undulatis distinguenda. Mat-forming, loosely rhizomatous perennial, vestiture a mixture of abundant fine glandular hairs ±0. 1 mm long and spreading non- glandular hairs ± 1 mm long. Stems erect, 0.7-3 dm high, sometimes branched above, villous, bark light brown below and green above. Leaves 5-13 pairs per stem, short petiolate to nearly sessile, those near the ground largest; blades ovate, densely hairy on both surfaces, margin undulate, ± crenate, 7-15 mm long and 2-10 mm wide. Verticillaster (head), solitary or occasionally with 1-2 additional heads in distinct whorls below, or in a panicle. Verticillaster bracts in three sets: outer set 2 pairs, 6-9 mm long, 4-7 mm wide, ovate, acute, scarious or the outer pair leaflike or leafy-tipped with green, undulate margins, scarious portions occasionally rose or lavender, closely subtending head or the outer pair l-5(-8) mm distant; middle set of bracts 2-3 pairs, lance-ovate, scarious, rose to lavender aging to straw-colored; inner set of bracts 0-3(-5), lanceolate or acicular, scarious. Calyx 6-8 mm long, sparsely to moderately hairy with short and long glandular hairs, 13-15 veined; lobes triangular-acute, margins glandular-ciliate with hairs to 0.5 mm long. Corolla 9-1 1 mm long, tube exserted from calyx % to % its total length, lavender to pale rose, sparsely pilose within and more densely so without; lobes tipped with gold colored glandular droplets, upper lip cleft its length; stamens unequal, with lower pair equal to or exceeding the corolla lobes and upper pair shorter than corolla lobes, filaments attached near or below the middle of the corolla; pistils shorter than the stamens, stigma branches < 1 mm long; nutlets 1 .5-2.0 mm long, triangular-ovate in cross section, light brown with black mottling. Chromosome number n=2\. Paratypes: USA, California, Kern Co., type locality, 2500 m, 1 1 Jun 1 986, Enter, Daniel, andBagley 6443 (UC); 8 Sep 1 987, Shevock andJokerst 11812 (CAS, RSA, UC). Tulare Co., Olancha Peak, W slope of summit at base of granitic boulders above foxtail pine forest, T19S R36E sect. 19 NEy4 SWy4, 11,600 ft (3530 m), 22 Jul 1950, Howell 27202 (CAS); Olancha Peak, Jul 1896, Purpus 1866 (UC). Inyo Co., Olancha Peak, southern saddle at timberline, T19S R36E sect. 19, 10 Sep 1986, 10,700 ft (3260 m), Shevock 11771 (CAS, FSC, MO, NY, RSA, UC); SE shoulder of Olancha Peak, 3402 m, 29 Jul 1975, Tatum 220 (UCSB); S ridge of Olancha Peak, 3353 m, no date, Tatum 502 (UCSB); Little Cottonwood Creek, 10,200 ft (3110 m), 12 Aug 1949, Howell 26254 (CAS). Distribution, habitat and phenology. Monardella beneolens grows on rocky granitic or metamorphic slopes in open mixed conifer and foxtail pine forests, from 2500 to 3530 m. Plants form extensive 274 MADRONO [Vol. 36 dense mats 1.5 m or more across where rooting substrate consists of rocky scree. Plants in bedrock crevices and ledges are smaller. Flowering extends from July to September, depending on elevation, snowpack duration, and patterns of seasonal aridity. Monardella beneolens is known from only three sites along the southern Sierra Nevada crest: Owens Peak in Kern County, Olancha Peak in Inyo/Tulare counties, and Cottonwood Creek in Inyo Coun- ty (Fig. 2). Although M. beneolens is not associated with a consistent group of species at these three populations, it occurs with a sur- prisingly high concentration of rare localized southern Sierra Nevada endemics. Noteworthy associates include Eriogonum wrightiiTomy ex Benth. var. olanchense (J. T. Howell) Rev. and Trifolium de- deckeme G. Gillett at Olancha Peak; while Erigeron aequifolius H. M. Hall, Eriogonum breedlovei (J. T. Howell) Rev. var. shevockii J. T. Howell, Haplopappus gilmanii S. F. Blake, Lomatium shevockii R. L. Hartman & Constance, and Raillardella muirii A. Gray occur at Owens Peak. Morphological variation and hybridization. Populations Monar- della beneolens from Olancha and Owens Peak are quite distinct. Olancha Peak material from above timberline, has generally shorter stems, denser villous and glandular hairs, and smaller crisped leaves. This may be a function of climate since Olancha Peak populations occur nearly 770 meters higher in elevation than those on Owens Peak. Below timberline south of Olancha Peak, possible hybridization with M. linoides A. Gray has been observed. The apparent hybrids are taller than M. beneolens, with narrower, less crisped leaves, and bracts exceeding the calyces with parallel veins that branch from the midrib near its base. In addition, plants from Shevock 11762 (along the Pacific Crest Trail S of Olancha Peak) with petiolate leaves, fascicles of undeveloped leaves in the axils, longer intemodes, and softly pubescent calyces may be hybrids between M. beneolens and M. odoratissima, a relative of M. linoides. Given the propensity for hybridization in Monardella, ecological rather than genetic factors appear to be most responsible for maintaining what distinctions exist among sympatric species (Epling 1925; Hardham 1966; Hardham and Bartel in press). Owens Peak was selected for the type locality because plants there display the least hybridization of the M. beneolens populations known. This contrasts with Clare Hardham's (pers. comm.) belief that the Olancha Peak material is the least hybridized and therefore more suitable as the type. In our understanding, plants from Owens Peak have traits that could be interpreted as ancestral to the Olancha Peak population while the converse seems less tenable. In addition to exhibiting the 1989] ^YlEWOCYiET XU: MONARDELLA BENEOLENS 275 II8*»00' W Fig. 2. Distribution of Monardella beneolens along the crest of the southern Sierra Nevada in Inyo, Kern, and Tulare counties, California. fewest traits of other locally occurring species, plants from Owens Peak more closely resemble other geographically restricted Monar- della from southern California and Arizona. It is likely that M. beneolens was historically more widespread and the isolated pop- ulations we observe today have since diverged as a result of hy- bridization and other forces. Relationships. Unlike many species Monardella, which are often not well delimited, M. beneolens is so distinct in its combination of openly rhizomatous habit, ovate crenate-undulate leaves, and mixed short glandular and spreading villous vestiture, that determining its relationships is challenging. The existence of putative hybrids with sympatric Monardella could lead one to conclude that M. beneolens is closely related to M. linoides and M. odoratissima. These latter MADRONO [Vol. o o fli I ^ 00 3 C (u c o O 05 B 0-6 • ^ C3 <- ;/5 4> a ^ ri 1/3 C3 03 c3 13 1/3 43 ^ ■i ^ ^ O ^ o Vh 4> o eg 03 (U 3 O 1 a; > on J3 03 ^ ^ - 03 O 00 03 o 03 C3 I dJ ^ O t/3 O O i3 C/3 TD eg 00 > 03 O « o S3 O S3 O O o >^ O 03 3 >» 00 ^ 6 03 O O t/5 ■(-> c 03 6 03 o (U c 03 ^ 03 o a c/3 52 o o 13 S-i H 3 3 o o +1 ■(-> O X) 0, o 3> 00T3 S3 S3 ^ B 2 13 3 c "ob I S3 O S3 >'^ 3 a t/3 ^ t/3 53 o --^ l-H CO C 43 3 J3 3) lU u 3 o 00 1/3 3 2 C/3 73 1/3 00 3 3 'S- CX (U 03 t/3 •r" Cd 03 > X ^ o >. o 73 3 S3 " 03 00 03 3 o a O c/3 o 03 03 c/3 > 03 (U C E 43 O 03 O, cd 43 CAl (U 73 03 (U a 03 43 c/3 > ■*-> 0) u S3 :3 • "—I ' S3 o 03 C X c/0 > H c/3 (U 43 ■t-> 03 OJ .52 tJi 'c/3 03 03 flJ 3 3 1^ U U 1989] SHEVOCK ET AL.: MONARDELLA BENEOLENS 3 i ■*-> s "o o o 00 +1 C3 (U 1/1 o C3 'l/3 o o O (U (=1 "ob 1/3 ^3 §^ i| O 73 P C O u c| O C/3 C 'ob X ■C O cd 03 u — 1J C " c 00- 3 S « Si 00 ■(-< 'c 03 O. u O 03 O 1/3 1/3 OJ (U 43 > o c3 T3 0) J3 o C 03 )-( Xi U( (U ^ « § Id c J3 o -t- ;3 T3 03 a 03 > 03 > 03 03" o 03 03 T3 t; C/3 (U ■(-< o c :3 in "o O o o 00 c o c 73 C3 > 0) 1/3 o 73 a t/3 ;> 5/3 'c3 43 o c o B cd C 03 C/3 I c/3 .-H 1/3 00 o o V3 „ ■c3 6 O " t3.i3 C r3 o 43 1/3 00 O 5^ O 03 03 73 C C C ^ ^ 3 03 00 O "33 C3 ""^■^ =>o c 2 C 5 -2 ob I 1/3 .^•^ *-H f . . O O c/3 „ ■c3 6 o ■* 03 -P c« (U O 03 s o N X ' 4«i 03 U 00 c/3 G C/3 C/3 (U (U 42 X) O O c o c/3 (U o 03 c 42 03 i3