ffs* 0 ’f-'-r'i' ■ ■J ■>' !ii: , \ ' t I ' f.^'; I VOLUME 61, NUMBER 1 JANUARY-MARCH 2014 MADRONO A WEST AMERICAN JOURNAL OE BOTANY Stage and Size Structure of Three Species of Oaks in Central Coastal California Ian S. Pearse, Sophie Griswold, Desirree Pizarro, and Walter D. Koenig ^ Chromosome Number and Reproductive Attributes for Erigeron A T (ASTERACEAE), A ClIFF-DwELLHN^ EnDEMIUOF Southeastern Arizona Richard D. Noyes and Pamela ...... .............. 9 Phytogenies and Chp.omosome EvoLUTiON.OF'FiMfcEZM (B OR agin ace ae; Hydrophylloideae) Inferred from Nuclear Riposomal and'ChloroplastaSequencI Genevieve K. WaldenylMura M. ‘^arnsoni Greg S Spicer, Frank W. Cipriono, and Robert Paiterson . . .'A A . .‘if. ;4Crt ............ A ..... . 16 /A-v y N Factors Determining the Establishment 'of Plant Zonation in a Southern Californian Riparian WooDLi#D,Ar> -^^JohnM Boland ''Zp. , 48 ■” I'fy , - i'. . '-'iUj' Identification and Taxonomic Stat6js of 'Cordylawbus jenuis subsp. RALLE5CENS (OrOBANCHACEAE) }p A Barbara L. Wilson, Richard E^Brainerd, Nick Ottin£,Brian J. Knaus, Julie Kierstead Nelson....... ........ 64 ' ^ Foliar Analyses of Conifers on Serpentine and Gabbro Soils in the Klamath Mountains Earl B. Alexander .................................................................................... 11 Ecology and Distribution of the Introduced Moss Campylopus INTROFLEXUS (DiCRANACEAE) IN WESTERN NORTH AMERICA Benjamin E. Carter ................................................................................. 82 Biology of the Geophyte, Triteleia ixioides Subsp. anilina (Themidaceae), in Coniferous Forests of Butte County, California Alfred Kannely and Robert A. Schlising ...................................... 87 PUBLISHED QUARTERLY BY THE CALIEORNIA BOTANICAL SOCIETY Madrono (ISSN 0024-9637) is published quarterly by the California Botanical Society, Inc., and is issued from the office of the Society, Herbaria, Life Sciences Building, University of California, Berkeley, CA 94720. Subscription information on inside back cover. Established 1916. Periodicals postage paid at Berkeley, CA, and additional mailing offices. Return requested. Postmaster: Send address changes to Madrono, Kim Kersh, Membership Chair, Uni- versity and Jepson Herbarium, University of California, Berkeley, CA 94720-2465. kersh@berkeley.edu. Editor — Matt Ritter Biological Sciences Department Cal Poly, San Luis Obispo 1 Grand Avenue San Luis Obispo, CA 93407 madronoeditor@gmail.com Editorial Assistant — Genevieve Walden Book Editor — Matt Ritter Noteworthy Collections Editor — David Keil Board of Editors Class of: 2014 — Brandon Pratt, California State University, Bakersfield, CA Tom Wendt, University of Texas, Austin, TX 2015 — Anna Jacobson, California State University, Bakersfield, CA CALIFORNIA BOTANICAL SOCIETY, INC. Oeficers for 2012-2013 President. V. Thomas Parker, Department of Biology, San Francisco State University, San Francisco, CA 94132, parker@sfsu.edu Eirst Vice President: Andrew Doran, University and Jepson Herbaria, University of California, Berkeley, CA 94720, andre wdoran @ berkeley.edu Second Vice President: Vacant. Recording Secretary: Michael Vasey, Department of Biology, San Francisco State University, San Francisco, CA 94132, mvasey@sfsu.edu Corresponding Secretary: Anna Larsen, Jepson Herbarium, University of California, Berkeley, CA 94720, secretary@calbotsoc.org Treasurer: Thomas Schweich, California Botanical Society, Jepson Herbarium, University of California, Berkeley, CA 94720, tomas@schweich.com The Council of the California Botanical Society comprises the officers listed above plus the immediate Past President, Dean Kelch, Jepson Herbarium, University of California, Berkeley, CA 94720, dkelch@berkeley. edu; the Membership Chair, Kim Kersh, University and Jepson Herbaria, University of California, Berkeley, CA 94720, kersh@berkeley.edu; the Editor of Madrono', and three elected Council Members: Staci Markos, University and Jepson Herbaria, University of California, Berkeley, CA 94720, smarkos@berkeley.edu. Graduate Student Representatives: Genevieve Walden, Department of Integrative Biology and Jepson Herbarium, University of California, Berkeley, CA 94720, gkwalden@gmail.com. Administrator: Lynn Yamashita, University of California, Berkeley, CA 94720, admin@calbotsoc.org. Webmaster: Ekaphan (Bier) Kraichak, University of California, Berkeley, CA 94720, ekraichak@gmail.com. @ This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Madrono, Vol. 61, No. 1, pp. 1-8, 2014 STAGE AND SIZE STRUCTURE OF THREE SPECIES OF OAKS IN CENTRAL COASTAL CALIFORNIA Ian S. Pearse Cornell Lab of Ornithology, 159 Sapsucker Woods Road, Ithaca, NY 14850 and Illinois Natural History Survey, 1816 S Oak St, Champaign, IL 61820 ianspearse@gmaiLcom Sophie Griswold and Desirree Pizarro Cornell Lab of Ornithology, 159 Sapsucker Woods Road, Ithaca, NY 14850 Walter D. Koenig Cornell Lab of Ornithology, 159 Sapsucker Woods Road, Ithaca, NY 14850 and Department of Neurobiology and Behavior, Cornell University, Ithaca, NY 14853 Abstract Oaks are foundational species in much of California, and many oak populations in the state may be in jeopardy due to a lack of recruitment of young trees. Despite considerable interest in this problem, there have been few comprehensive surveys of all stages of oak development. We surveyed all stages of three oaks: Quercus lobata, Q. douglasii, and Q. agrifoUa in a forest plot with mixed land-use in central coastal California. We found abundant seedlings of all oak species, but an apparent paucity of Q. lobata and Q. douglasii saplings. First year seedlings of all species were less abundant in parts of the study site with cattle grazing, but later-stage seedlings of Q. lobata and Q. douglasii were equally abundant across land-use types. Quercus agrifolia seedlings were associated with non-grazed areas; Quercus agrifolia late-stage seedlings in the grazed area were smaller and less abundant than in non- grazed areas. Quercus agrifolia seedlings of all stages tended to be clustered around conspecific mature trees. Quercus lobata late-stage seedlings, and to a lesser degree those of Q. douglasii, were often distant from any potential parent tree. These data indicate that young stages of the three species of oak have different spatial distributions and occur in different abundances at two sites with different grazing regimes. They are also consistent with a relative lack of regeneration in Q. lobata and Q. douglasii, although it remains to be determined that the small number of saplings of these species observed is insufficient to replace mortality of mature trees. This survey provides a baseline from which future resampling can assess the long-term demographic success of three Californian oak species. Key Words: regeneration, demography, Quercus. Californian oaks are a dominant component of at least a third of the forest and savanna ecosystems of the state (Pavlik et al. 1991). However, there has been recent concern for oak populations, as there is a perceived lack of regeneration of oak forests and a lack of young tree stages (Griffin 1976; Tyler et al. 2006). Studies have attributed the lack of oak regener- ation to various causal factors (Griffin 1971; Borchert et al. 1989; Callaway and Davis 1998; Gordon and Rice 2000), while others have questioned the severity of regeneration problem altogether (Tyler et al. 2006). The apparent lack of recruitment of two common California oaks, Quercus lobata Nee and Q. douglasii Hook. & Arn, (Fagaceae), have prompted numerous causal explanations for seedling mortality and lack of regeneration in these species, including cattle grazing (Hall et al. 1992; Tyler et al. 2008; Reiner and Craig 2011), competition from non-native grasses (Griffin 1971; Gordon and Rice 2000), altered herbivore populations (Griffin 1971; Borchert et al. 1989; Callaway and Davis 1998; Davis et al. 2011), and increased drought stress (Mahall et al. 2009). Each of these factors has undergone widespread anthropogenic change in the last two centuries, thus potentially explaining the apparently poor recruitment currently suffered by oak popula- tions (reviewed by Tyler et al. 2006). However, the relative importance of each of these factors may be site specific. For example, while grazing may decrease seedling abundance at some sites (Reiner and Craig 2011), even sites without livestock grazing may experience a lack of seedling recruitment (Griffin 1971). The lack of a single causal factor in altering oak populations calls for a more integrative approach in assessing oak population dynamics. A recent conceptual review (Tyler et al. 2006) and meta-analysis (Zavaleta et al. 2007) of California oak recruitment both suggested that two Californian oaks, valley oak {Q. lobata) and blue oak {Q. douglasii), suffer from declining populations. Both studies concluded that site- specific causes may affect recruitment of these 2 MADRONO [VoL 61 species. Moreover, they indicated that detailed demographic models are needed to address the question of whether low observed recruitment levels are actually insufficient to replace stands of long-lived adults. Likewise, a detailed demo- graphic analysis of oak populations could ask specifically which life stages most limit popula- tion growth (Tyler et al. 2006; Zavaleta et al. 2007; Davis et al. 2011). Meanwhile, studies have suggested that a third species of Californian oak, Q. agrifolia Nee, may not be undergoing similar low regeneration rates as its two sympatric congeners, even though it may be affected by some of the same recruitment challenges (Call- away and Davis 1998; Mahall et al. 2009). While studies have successfully documented multiple site specific factors likely affecting populations of Californian oaks, there is still a relative lack of the explicit demographic model- ing of oak populations needed to quantify the risk to oak populations and to assess which life stages of oaks are of the highest conservation value. The only current study that explicitly reconstructs Q. lobata demography uses an experimental approach, which is very valuable in determining causal factors that limit Q. lobata success (Davis et al. 2011). Observational studies that thoroughly survey oak populations that are both geographically referenced and include all life stages of the oaks complement experimental studies. Such surveys will be valuable because they can be easily resampled in order to directly assess vital rates and stage transitions. At the same time, comprehensive site surveys observe oak populations across all microhabitats at a site, which is difficult in manipulative studies. Here we present a baseline survey of the oaks on a portion of Hastings Natural History Reserve (HNHR) and the adjacent Oak Ridge Ranch (ORR) property in Monterey County of central coastal California. We surveyed the size, stage, and geographic location of 19,755 individuals of the three dominant oak species {Q. lobata, Q. douglasii, and Q. agrifolia) in a 52.6 hectare area encompassing forest, grassland, and savannah habitats. This survey also encompassed areas with strikingly different recent land use histo- ries, as HNHR has been protected from cattle grazing since 1937, but ORR continues to experience significant regular cattle grazing. We use this survey primarily to establish a baseline for parameterizing future demographic models of oak populations, and we make this baseline data available to the general scientific community. We analyzed the size structure and spatial distribution of early and late seedling stages of these oaks in order to ask to what degree habitat type and land use affect the size of late-stage seedlings and the proximity of seedlings to a potential parent. Methods Study Area and Natural History The survey was conducted on Haystack Hill and surrounding areas (36.385114N, — 121.561906 W) on Hastings Natural History Reservation (HNHR) and the adjacent Oak Ridge Ranch (ORR) (Fig. 1). The study area is typical of coast range oak habitat, is the site of several studies of oak populations (White 1966; Griffin 1971, 1976; Koenig and Knops 2007), and is central (within 70 km of 1/3 of the study sites) to the sites used in a recent meta-analysis of Californian oak popula- tions (Zavaleta et al. 2007). The study area has a Mediterranean climate typical of California’s coast range. In the year of the survey, the average rainfall was 475 mm, which was slightly less than the 30 year average for the site (540 mm - Hastings Weather Station http://www.hastingsreserve.org/ weather/Weather.html). The study area was rough- ly delimited by the contour of the hill. We excluded a large patch of dense chaparral that would have been within the study, as the area was too densely vegetated to use the same survey techniques as in other areas and contained few oak trees, with the exception of a small number of Q. agrifolia individuals at the bottom of dry washes that descended the hillside through chaparral. The study area was divided into three habitat types based on tree cover and groundcover type: forest, savanna, and grassland (Griffin 1990). ‘Forest’ had a complete canopy cover. Forest understory was dominated by poison oak {Tox- icodendron diver silobum [Torr. & A. Gray] Greene) and had little to no grass cover. One forest region had a substantial stand of madrone {Arbutus menziesii Pursh), whereas oaks domi- nated all others. ‘Savanna’ had a partial canopy cover and an understory of native and non-native grasses including Avena L. sp., Bromus hordea- ceus L., R diandrus Roth, Hordeum L. sp., Stipa pulchra Hitchc., and Air a caryophyllea L., as well as forbs such as Madia gracilis (Sm.) D. D. Keck, Asclepias eriocarpa Benth., Plagiobothrys Fisch. & C.A. Mey. sp., and Amsinckia menziesii (Lehm.) A. Nelson & J.F. Macbr. ‘Grasslands’ had sparse to no tree cover and were dominated by the same mix of grasses and forbs. Small remnants of perennial grasslands (sensu Griffin 1990) dominated by Stipa pulchra were included in this category. There was considerable variation in both slope and aspect of habitat types within the study area, but both study sites (HNHR and ORR) had both steep and relatively flat forests, grasslands, and savannahs. One difference be- tween the two sites was the presence of a small creek bed at the HNHR site that had a high abundance of Q. agrifoUa seedlings. The study site was divided approximately down the middle by a fence separating HNHR 2014] PEARSE ET AL.: SURVEY OF CENTRAL CALIFORNIAN OAK STAGES 3 Fig. 1. A map of the survey area (highlighted in light dashes and outlined with a dotted line). The survey is divided roughly in half by a property line (central E~W line), where the northern property (Oak Ridge Ranch) is grazed and the southern property (Hastings Natural History Reserve (HNHR) is ungrazed. In May-June 2012, we surveyed and recorded the location of all valley oak (Q. lobata, white dots), blue oak (Q. douglasii, grey dots), and coast live oak (Q. agrifolia, black dots) first year seedlings, late-stage seedlings, saplings, and trees within the demarcated study area. from ORR. The land use of the two areas can be summarized as follows. Prior to 1930, both sites had a history of mixed land use including ranching, timber harvest, and small-scale agricul- ture (Griffin 1990). HNHR was donated to the University of California in 1937 at which point all grazing and agriculture was stopped. At ORR, moderate year-round cattle grazing continues to present day, including during the active survey period, although no tillage or significant timber harvest has been recorded within the study area within recent history. While oak plantings and restoration efforts have been made in other portions of HNHR, none of these activities have occurred within the study area. The year preceding our survey had a moderate to good acorn crop for all three oak species. Based on the Koenig visual count method (Koenig et al. 1994a), the 2011 Q. lobata acorn crop was 15% greater than the 30-year mean Q. lobata crop (Pearse et al. 2014). Likewise, the 2011 Q. douglasii acorn crop was 63% greater than its 30-year mean, and the 2011 Q. agrifolia acorn crop was 29% greater than its 30-year mean. Survey Methods We surveyed all valley oak, blue oak, and coast live oak individuals within the study area in May through early July 2012. We marked 10 m wide transects across the study area and used a GPS to record the locality of each oak within the area to an accuracy of 5 m. For each individual, we recorded one of four growth stages. Seedlings were first year germinants from acorns. All had unbranched stems, and most were still connected to a partially buried acorn. Late-stage seedlings were over one year old, less than 1.5 meters in height, had branched or multiple stems, and lacked a connect- ed acorn. Saplings were greater than 1.5 meters in height, less than 10 cm DBH, and had not yet begun to produce fruit or flowers. Adult trees were over 10 cm DBH. We considered these classes to be biologically useful, as demographic parameters including growth rate, survivorship, and reproduc- tion likely vary between stages (e.g., Griffin 1971; Koenig and Knops 2007). For all late-stage seedlings, we also recorded height and canopy width and calculated seedling volume as height X widtffi. For all first-year seedlings, late-stage seedlings, and saplings, we also calculated Euclid- ean distance to the nearest conspecific adult tree. We tagged each adult tree and recorded which first year seedlings, late-stage seedlings, and saplings were within 20 m of its trunk. Statistical Analysis While the two landscape level factors - habitat and property, the latter corresponding to differ- 4 MADRONO [VoL 61 Table 1. The Density of Three Oak (Quercus) Species (Trees/Hectare) in Three Habitat Types Spanning the Ungrazed Hastings Natural History Preserve (HNHR) and Grazed Oak Ridge Ranch (ORR). All trees (>10 cm DBH), saplings (>1.5 m tall), late-stage seedlings (multiple stems, no attached acorn), and first-year seedlings (single stem, often acorn still attached) within the total 52.57 ha survey area were recorded resulting in 19,755 records of oak individuals. Forest Savannah Grassland HNHR ORR HNHR ORR HNHR ORR Area (ha) 8.21 8.89 8.23 7.73 10.83 8.65 Quercus lobata 1st year seedling 161.75 7.2 49.09 5.3 16.34 0.55 Late stage seedling 47.14 17.55 24.79 54.59 15.51 18.47 Sapling 0.24 0 0 0 0.09 0.09 Tree 17.54 10.01 13.49 6.47 2.4 0.37 Quercus douglasii 1st year seedling 18.39 176.04 84.93 34.67 3.23 0.09 Late stage seedling 29.6 194.94 94.9 60.28 5.26 2.4 Sapling 0.49 0 0 0 0.09 0 Tree 11.33 69.74 22.72 6.21 0.09 0.83 Quercus agrifolia 1st year seedling 455.18 23.17 57.84 1.29 61.68 0.28 Late stage seedling 209.38 76.83 24.18 8.8 6.74 3.42 Sapling 15.23 0.45 0 0 0.55 0 Tree 55.05 27.67 6.44 0.91 0.92 0 ent land use - are spatially non-independent, we used these data to build a case study of oak populations at a single, large site. Specifically, our goal was to explore variation in 1) the size of late- stage seedling oaks, as this size class may escape grazing at large sizes (either by being tall or very wide), 2) the distance of seedlings and late-stage seedling oaks from a potential (conspecific) parent tree, and 3) the composition of oak seedlings in the understory of individual adult trees. All statistics were calculated in R using package car (R Core Development Team 2008; Fox and Weisberg 2011). Our georeferenced dataset is made avail- able at the Dryad data depository http://dx.doi. org/10.5061/dryad.467g5 and on the authors’ websites http://www.nbb.cornell.edu/wkoenig; http://ianpearse.wordpress.com. Results Habitat and Land-use Associations We surveyed 52.6 ha and gathered data on a total 19,755 individuals of the three oak species across both properties (HNHR and ORR) and all habitats (Fig. 1, Table 1). The surveyed area of each habitat type at each property was similar, and we report the densities of oaks at each stage in each habitat type on the two properties (Table 1). Quercus lobata seedlings were more abundant at HNHR, the ungrazed site, than at ORR, the grazed site - 1909 seedlings versus 111 (Table 1). However, this difference did not carry into the late-stage seedling individuals, and there were slightly fewer late-stage seedling Q. lobata at HNHR than at ORR (759 versus 778). At HNHR both Q. lobata seedlings and late-stage seedling stages tended to be located in forested habitats, whereas at ORR, they were associated with savannah, although the distribution of overstory trees only partially reflected this trend. All growth stages of Q. douglasii were uncom- mon in grasslands at all sites (Table 1). There was a greater abundance of early growth stages of Q. douglasii at ORR than at HNHR (4059 versus 1966), reflecting the distribution of adult trees at the two sites (677 versus 281). There was a full- canopied Q. douglasii forest on ORR, and all growth stages of Q. douglasii at ORR were more closely associated with forest than at HNHR, where Q. douglasii was more abundant in open savannah habitat. Quercus agrifolia was associ- ated with full-canopied forest habitats at both sites and was particularly abundant in HNHR forests. Quercus agrifolia was also more abundant in HNHR grasslands and savannas than the same habitat types at ORR. There were strikingly fewer Q. agrifolia first year seedlings in all habitats on ORR than on HNHR (219 versus 4881), which would be consistent with an impact of grazing on Q. agrifolia seedling abundance. For Q. agrifolia, this trend persisted into late- stage seedlings, where there were fewer Q. agrifolia late-stage seedlings at ORR than at HNHR (788 versus 1991). There were few saplings of both Q. lobata and Q. douglasii in all habitats at both properties (four Q. lobata saplings and five Q. douglasii seedlings), consistent with perceived recruitment problems with these species. In contrast, Q. agrifolia saplings were reasonably common (n = 135), almost all of which were associated with forest sites on the ungrazed HNHR property (Table 1). Size of Late Stage Seedlings We measured the volume of late-stage seed- lings (excluding first-year seedlings) (Fig, 2). 2014] PEARSE ET AL.: SURVEY OF CENTRAL CALIFORNIAN OAK STAGES 5 C/5 05 g 0 0 (/) 0 05 S to I B o forest savannah forest savannah Q. agrifolia grassland grassland forest savannah grassland Fig. 2. The volume of late-stage seedlings of Q. lobata, Q. douglasii, and Q. agrifolia in three habitat types on ungrazed (HNHR, dark grey) and grazed (ORR, light grey) properties. Our designation of “late- stage seedling” included oaks that were less than 1.5 m in height, had multiple stems, and no attached acorn. This size of this stage may be particularly important to oak regeneration, as late-stage seedlings persist over many years and are susceptible to grazing, especially when small. Bars are means ± SE. Quercus lobata late-stage seedlings tended to be larger in more open habitats such as grasslands irrespective of whether the site was grazed or not (Fig. 2). Quercus lobata seedlings in savannahs were four times larger on ORR than on HNFIR (Fig. 2) and similar in size in other habitat types, a finding that is not consistent with cattle grazing being a factor limiting seedling growth of this species. Quercus douglasii late-stage seedlings showed similar size trends to Q. lobata, with larger late- stage seedlings in grasslands than in other habitat types (Fig. 2). Unlike Q. lobata, they were roughly the same size across land-use types in all habitats. Quercus agrifolia late-stage seedlings showed the strongest size association with land use type of the three species. In both grasslands and forests, Q. agrifolia late-stage seedlings at HNHR were 300-400% larger than those in the same habitats on ORR (Fig. 2). Quercus agrifolia late- stage seedlings in savannah habitat were smaller than in other habitats (Fig. 2). The variation in size of seedlings was greatest in grassland habitats. Spatial Distribution of Seedlings We measured the distance from each first year seedling and late stage seedling to its nearest mature conspecific tree (i.e., potential parent) (Fig. 3). Quercus lobata first-year seedlings aver- aged 12.4 meters from the nearest potential parent. Quercus lobata late stage seedlings were on average further from the nearest potential parent than seedlings of the other species, and this distance was even greater on ORR than on HNHR (15.6 m and 14.3 m respectively, Fig. 3). Quercus douglasii seedlings were generally closer to a potential parent than Q. lobata seedlings, likely reflecting the density of stands of these tree species (Fig. 3). Both Q. douglasii first-year and late-stage seedlings averaged eight m from the nearest potential parent on HNHR (Fig. 3). On ORR, first-year Q. douglasii seedlings averaged 4.7 m from the nearest potential parent and late- stage seedlings averaged seven m. Quercus agrifo- lia seedlings averaged 7.5 m from the nearest potential parent, a value that did not vary by either seedlings class (Fi 3855 =2.0, P = 0.16) or property (Fi^gss =0.002, P = 0.96, Fig. 3). We also measured the composition of the seedling oak community underneath mature trees of each of the three oak species and found that the understories of both Q. douglasii and Q. agrifolia were dominated by their conspecific seedlings (Fig. 4). In contrast, the understories of Q. lobata trees had a roughly even distribution of seedlings from all three oak species. Similarly, Q. lobata seedlings were the least abundant compo- nent of the understories of either Q. douglasii or Q. agrifolia. Discussion The size and age distributions of the three species of oaks surveyed in this study varied between properties with different land-use histo- ry, but these differences were fairly subtle. There were fewer valley oak {Q. lobata) first-year seedlings per adult tree on the grazed site (ORR) than on the ungrazed site (HNHR). 6 MADRONO [Vol. 61 Fig. 3. The distribution of the distance of Q. lobata, Q. douglasii, and Q. agrifolia first year seedlings (dashed line) and late-stage seedlings (solid line) to their nearest conspecific (i.e., potential parent) tree at the ungrazed site (HNHR) and grazed site (Oak Ridge Ranch). Valley oak late-stage seedlings and to a lesser degree first year seedlings are located at a greater distance from a potential parent than the other two species. However, the abundance of late-stage seedlings was similar at both sites (Table 1). Moreover, Q. lobata late-stage seedlings were on average larger at the grazed site, further suggesting that cattle grazing may affect the survival of first-year seedlings, but may have little or no effect on the long-term survival of established seedlings. Quer- cus douglasii populations had generally the same associations as Q. lobata with the two major land use patterns (Table 1, Fig. 2), except that there was a large number of first-year and late-stage Q. douglasii seedlings in a dense forest patch on ORR (Table 1). The larger size of late-stage seedlings at the grazed site suggests that cattle grazing does not limit (and may even promote) the growth of these seedlings at our study area. The land-use differences in this survey are consistent with effects of cattle grazing on Q. douglasii and Q. lobata observed in other studies. For example, experimental manipulation of grazing did not affect the survivorship of late- stage (2-year old) seedlings in a Sierra Nevada foothills population of Q. douglasii. (Hall et al. 1992), and grazing decreased the total abundance of naturally recruiting Q. douglasii seedlings (likely including a large portion of first-year seedlings) in only one out of four years (Reiner and Craig 2011). Livestock grazing may still be a significant mortality factor for Q. lobata and Q. douglasiv, however, our results suggest that other mortality factors are potentially more important even in ungrazed sites. Interestingly, our survey found fewer (Table 1) and smaller (Fig. 2) early-stage Q. agrifolia individuals on the grazed property (ORR) than on the ungrazed property (HNHR). Unlike in Q. douglasii, Q. agrifolia later-stage seedlings were both smaller and less abundant (with reference to mature trees) on ORR than on HNHR (Fig. 2, Table 1). Moreover, there were far more Q. agrifoUa saplings per mature tree on HNHR than on ORR (Table 1). As non-adult stages appear to be more represented at HNHR than at ORR, this suggests that Q. agrifolia population growth on HNHR may be higher than on the grazed adjacent property. Quercus agrifoUa pop- ulations are thought to be stable or even increasing in many regions with high forest or shrub cover, but are unable to colonize open habitat, presumably because shrubs provide a refuge from wildlife or livestock grazing (Call- away and Davis 1998). In the current survey, Q. agrifoUa was strongly associated with habitats with greater shrub and tree cover, and livestock grazing may have imposed a sufficient herbivore pressure to partially negate the protective effect of high vegetation cover for Q. agrifolia seedlings. The seedlings of each of the three oak species had unique spatial distributions with reference to their potential parental trees. The distribution of first-year seedlings likely reflects patterns of acorn fall and immediate dispersal. As acorn production is pulsed (i.e., varies dramatically from year to year) and synchronous between trees 2014] PEARSE ET AL.: SURVEY OF CENTRAL CALIFORNIAN OAK STAGES 7 Q. douglasii Q. agrifolia Fig. 4. The species identity of all immature oaks (first year and late-stage seedlings) within 20 m of a valley oak (Q. lobata = Q. lo), blue oak {Q. douglasii = Q. do), and coast live oak {Q. agrifolia = Q. ag). The understories of both blue oak and live oak were dominated by conspecifics (likely acorn fall from the focal tree), however the understory of valley oak had a roughly even representation of all three species. (Koenig et al. 1994b), the distribution of first- year seedlings may vary substantially from year to year. On the other hand, the late-stage seedlings of oaks in this habitat may be very long-lived (Koenig and Knops 2007), so their distribution is unlikely affected by temporal variation in acorn crop. Quercus lobata first year seedlings were clustered around potential paren- tal trees, likely reflecting patterns of acorn fall. Late-stage Q. lobata seedlings, however, were on average further from a potential parent, especial- ly on the grazed property (Fig. 3), This, in combination with few seedlings altogether at some sites, resulted in an underrepresentation of Q. lobata seedlings in the understory of Q. lobata mature trees (Fig. 4). Quercus douglasii seedlings were more abundant under their parent tree (Fig. 4), although late-stage seedlings tended to be located further from a potential parent than first-year seedlings (Fig. 3). Both first year and late-stage Q. agrifolia seedlings were strongly clustered around potential parental trees (Fig. 3), and Q. agrifolia seedlings were highly represented in the understories of Q. agrifolia trees (Fig. 4). These results suggest a hierarchy in the ability of oak seedlings to colonize open habitat. Quercus lobata was often found far from its parental species, consistent with previous studies showing that Q. lobata seedlings better tolerate dry habitats (Mahall et al. 2009) and deer browsing (Tyler et al. 2008) better than Q. agrifolia. Quercus douglasii has been shown to tolerate dry, exposed habitats (Griffin 1971), but it may be negatively affected by direct competi- tion from non-native grasses in open areas (Gordon and Rice 2000) and is associated with shrub cover (Callaway 1992). Quercus agrifolia has been shown to be strongly associated with shrubby or forested habitat, which provides protection from herbivores (Callaway and Davis 1998). The factors that limit oak populations are thought to be somewhat site-specific. This has led to both studies that attempt to find those factors that are important across many site (Zavaleta et al. 2007) and detailed studies that attempt to assess the causal factors limiting oak seedling success at individual sites (Davis et al. 2011). Our study will be useful in both of these goals. By establishing a baseline survey at a site, where the history of factors that could potentially affect oak populations (e.g., grazing, habitat affiliation, wildlife density) is well documented, our study can easily be combined with other surveys at different geographic locations in order to assess the generality of these factors. Likewise, by having a population of oaks where all stages are surveyed and georeferenced, this survey opens the possibility of resampling to determine the key demographic limitations at this representative site. Explicit demographic models of Californian oak populations are important to understand the sustainability of the state’s oak populations (Davis 2011). This study provides a thorough baseline estimate of populations of Q. lobata, Q. douglasii, and Q. agrifolia that can be used by future resampling and modeling efforts to esti- mate demographic transitions between life-histo- ry stages and thus the long-term dynamics of oak populations. For example, at the same site, a marked population of blue oak late-stage seed- lings has persisted with little growth, little mortality, and no transitions to saplings within 8 MADRONO the past 60 years (Koenig and Knops 2007). Moreover, there are currently no robust estimates of age-biased mortality for adult valley oaks, which is a critical parameter in estimating the necessary rate of replacement by seedlings (Tyler et ah 2006). By marking a large number of trees and seedlings, as well as the few observed saplings, resampling of this study may observe enough stage-transitions to accurately parame- terize demographic models of oak populations at a large, natural site. Acknowledgments We thank Vince Voegeli and the management of Oak Ridge Ranch for access to the study site. This study was supported by an NSF REU grants DEB- 102 14 17 and DEB-1212885 and NSF grant DEB 0816691 to WDK. Literature Cited Borchert, M. L, F. W. Davis, J. Michaelsen, and L. D. Oyler. 1989. Interactions of factors affecting seedling recruitment of blue oak {Quercus douglasii) in California. Ecology 70:389-404. Callaway, R. M. 1992. Effect of shrubs on recruit- ment of Quercus douglasii and Quercus lobata in California. Ecology 73:2118-2128. and F. W. Davis. 1998. Recruitment of Quercus agrifolia in central California: the impor- tance of shrub-dominated patches. Journal of Vegetation Science 9:647-656. Davis, F. W., C. M. Tyler, and B. E. Mahall. 2011. Consumer control of oak demography in a Mediterranean-climate savanna. Ecosphere 2:108, doi:10.1890/ESl 1-00187.1 Fox, J. and S. Weisberg. 2011. An R Companion to Applied Regression, Second Edition. Sage Publi- cations, Thousand Oaks, CA. Gordon, D. R. and K. J. Rice. 2000. Competitive suppression of Quercus douglasii (Fagaceae) seed- ling emergence and growth. American Journal of Botany 87:986-994. Griffin, J. R. 1971. Oak regeneration in upper Carmel Valley, California. Ecology 52:862-868. . 1976. Regeneration in Quercus lobata savannas, Santa Lucia Mountains, California. American Midland Naturalist 95:422-435. . 1990. Flora of Hastings Reservation, Carmel Valley, California. Third Edition. Regents of the University of California, Berkeley, CA. Hall, L. M., M. R. George, D. D. McCreary, and T. E. Adams. 1992. Effects of cattle grazing on [Vol. 61 blue oak seedling damage and survival. Journal of Range Management 45:503-506. Koenig, W. D. and J. M. H. Knops. 2007. Long-term growth and persistence of blue oak {Quercus douglasii) seedlings in a California oak savanna. Madrono 54:269-274. , , W. J. Carmen, M. T. Stanback, and R. L. Mumme. 1994a. Estimating acorn crops using visual surveys. Canadian Journal of Forest Research (Revue Canadienne De Recherche For- estiere) 24:2105-2112. , R. L. Mumme, W. J. Carmen, and M. T. Stanback. 1994b. Acorn production by oaks in central coastal California: variation within and among years. Ecology 75:99-109. Mahall, B. E., C. M. Tyler, E. S. Cole, and C. Mata. 2009. A comparative study of oak {Quercus, Fagaceae) seedling physiology during summer drought in southern California. American Journal of Botany 96:751-761. Pavlik, B. M., P. C. Muick, S. G. Johnson, and M. Popper. 1991. Oaks of California. Cachuma Press and California Oak Foundation, Los Olivos, CA. Pearse, I. S., W. D. Koenig, and J. M. H. Knops. 2014. Cues versus proximate drivers: testing the mechanism behind masting behavior. Oikos (in press). R Development Core Team. 2008. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Website http://www.R-project.org/ (ac- cessed Dec 2010). Reiner, R. and A. Craig. 2011. Conservation easements in California blue oak woodlands: testing the assumption of livestock grazing as a compatible use. Natural Areas Journal 31:408-413. Tyler, C. M., F. W. Davis, and B. E. Mahall. 2008. The relative importance of factors affecting age- specific seedling survival of two co-occurring oak species in southern California. Forest Ecology and Management 255:3063-3074. , B. Kuhn, and F. W. Davis. 2006. Demogra- phy and recruitment limitations of three oak species in California. Quarterly Review of Biology 81:127-152. White, K. L. 1966. Structure and composition of foothill woodland in central coastal California. Ecology 47:229-237. Zavaleta, E. S., K. B. Hulvey, and B. Fulfrost. 2007. Regional patterns of recruitment success and failure in two endemic California oaks. Diversity and Distributions 13:735-745. Madrono, Vol. 61, No. 1, pp. 9-15, 2014 CHROMOSOME NUMBER AND REPRODUCTIVE ATTRIBUTES FOR ERIGERON LEMMONII (ASTERACEAE), A CLIFF-DWELLING ENDEMIC OF SOUTHEASTERN ARIZONA Richard D. Noyes Department of Biology, University of Central Arkansas, Conway, AR 72035 rnoyes@uca.edu Pamela Bailey U.S. Army Engineer Research and Development Center, Vicksburg, MS 39180 Abstract Erigeron lemmonii A. Gray, restricted to Scheelite Canyon in the Huachuca Range, Arizona, has previously been proposed for federal listing as an endangered species but basic cytological and reproductive information has been wanting. The first chromosome count for the species is 2n = 18, which is the common diploid number in Erigeron. Analyses of eight plants from five disparate sites within the population show that pollen averages 85.8% staining in cotton blue in lactophenol. Microscopic observation using differential interference contrast optics shows that E. lemmonii combines about equally monosporic and bisporic megagametophyte development within a single capitulum. Despite variability in developmental route, the egg apparatus among mature megagametophytes appears to be nearly uniform in structure. In greenhouse culture, isolated plants fail to set seed indicating that plants are probably self-incompatible. Controlled crosses yield seed, but variation in seed set intimates the possible presence of genetic barriers within the population. Key Words: bispory, conservation, endemic, Erigeron, Erigeron lemmonii, gametogenesis, megaspo- rogenesis, monospory. Erigeron L. (Asteraceae) consists of about 390 species with 173 species documented for North America north of Mexico (Nesom 2006). In the United States, most species occur in the montane and arid West. General morphological uniformi- ty can make determining the identity of species difficult; most of the taxa are low perennials with simple or lobed, alternate, one-nerved leaves, and white to light-purple rays and yellow discs. Cronquist (1947) even concluded that indument was the most reliable character for species delimitation. Further complicating systematic elucidation in Erigeron is the occurrence of apomictic complexes that include local polyploid hybrid populations (microspecies) that reproduce asexually by seed. Despite these issues, a system- atic framework Erigeron is maturing based on a combination of morphological and molecular analyses (Noyes 2000; Nesom 2008). It is common for Erigeron species to be locally endemic and known from relatively few popula- tions in specialized habitats. For instance, in Arizona, out of 42 described Erigeron species, 13 have global conservation status ranks (www. NatureServe.org/explorer) of G1 (four taxa; critically imperiled), G2 (six taxa; imperiled), or G3 (three taxa; vulnerable). To effectively man- age such restricted and sensitive plant species, basic biosystematic data are essential. Such data may include assessments of chromosome number, breeding system (selling vs. self-incompatible). mode of reproduction (apomictic vs. sexual), and phylogenetic relationship. For many Erigeron species, these data are lacking. Erigeron lemmonii A. Gray is known only from Scheelite Canyon in the Huachuca Mountains of Cochise County, Arizona. It was described in 1883 based on a collection by John Gill Lemmon made the previous year (Gray 1883). It is classified as a member of Erigeron sect. Olygo- trichium Nutt. (Nesom 2008) and is a decumbent- ascending perennial forming clumps in crevices and on ledges of vertical limestone cliffs within the canyon. Plants produce relatively long, arching stems that give rise to solitary (or few) capitula on ascending branches. Erigeron lemmo- nii has a global rank of G1 (www.natureserve. org/explorer). Based in part on a report indicat- ing that it was known from only 108 individuals (Gori et al. 1990), it was proposed as a candidate for protection under the Endangered Species Act in 1993. Subsequent extensive census of suitable habitat in the Huachuca Mountains did not uncover new populations, but additional plants discovered within Scheelite Canyon brought the estimated total number of individuals to about 950 (Malusa 2006). In consideration of these new data and in determining that E. lemmonii was stable and unlikely to be extirpated, it was removed from the candidate list (U.S. Depart- ment of the Interior, Fish and Wildlife Service 2012). 10 MADRONO [Vol. 61 Fig. 1. Photographs of Erigeron lemmonii. A. In greenhouse culture at the University of Central Arkansas from rooted ramets collected in the field. B. On a cliff ledge in Scheelite Canyon, AZ. Capitula are approximately one cm diam. Distal leaves along flowering branches are typically entire (as viewed in A); more basally disposed leaves are commonly three- to five-lobed. Materials And Methods In the early June 2012, vegetative branches from nine plants of Erigeron lemmonii (Fig. 1) were collected and sent to the University of Central Arkansas, Conway. Plants were sampled from four sites in Scheelite Canyon: North Main Face, sample #1; Main Face, samples #2-6; Boulder, sample #7; Owl Canyon, samples #8-9 (Fig. 2). Upon arrival, bases of the stems were dipped in rooting hormone powder (Green Light Organic Rooting Hormone, Green Light Co., San Antonio, TX, greenlightco.com), placed in Fafard Professional Potting Mix (Conrad Fafard Inc., Agawam, MA, www.fafard.com/Products/) in four-inch pots, and provided with natural lighting and moderate watering regimen. As the plants came into flower in the greenhouse, their ability to make seed autonomously was assessed by inspecting shattered mature heads under a dissecting microscope using transmitted illumina- tion. In Erigeron, filled cypselae are opaque and light brown; empty ones are transparent. Her- barium vouchers of the North Main Face specimen (RDN #1687) were prepared and deposited at UCAC and ARIZ. Chromosome number and pollen stainability were determined for all plants. Root-tips were collected in the early morning and pretreated in 8-hydroxyquinoline for four hours and then fixed in 3:1 ethanol: acetic acid. Root-tips were then digested in 15% (~2 mol/L) HCl for 26 minutes at room temperature, rinsed in distilled water, macerated, stained with acetocarmine, squashed under a cover slip, and viewed at lOOOX using bright-field microscopy. To assess stainability, newly shed pollen was stained in Cotton Blue in lactophenol for four days and evaluated with bright field microscopy at 400 X (Stanley and Linskens 1974). Stainability was scored as percentage of darkly staining grains in a sample of 300 grains. Pollen size was estimated for the Main Face-4 accession by measuring on digital images a sample of 85 grains with diameters estimated from average mid-exine to mid-exine points. Reproductive development was evaluated for the North Main Face plant. Capitula at three stages of development were studied: stage one - pre-anthesis, to observe initial division of the megasporocyte; stage two - early anthesis, to view condition of mature female gametophytes; stage three - three days post anthesis, to detect evidence for autonomous embryo or endosperm formation. Capitula were fixed in FAA for two weeks and then dehydrated in 100% ethanol and cleared in methyl salicylate (Herr 1971). Ovaries (each bearing a single ovule) were dissected from the cleared florets and arrayed under a cover slip held in place with rubber cement. Cellular detail of ovules was observed at 600 X using differential interference contrast optics (D.LC). All microscopic observations were performed using an Olympus B54 microscope. Images were made with a CCD 8-bit digital camera, and measurements were made using AnalySIS (ver- sion 3.1) image-capturing software (Soft Imaging System, GmbH 1989-2001). The ability of plants to make seed by outcrossing was evaluated with controlled crosses among the plants in greenhouse culture. Four inter-site crosses were performed. Pollinations were made by removing a capitulum at anthesis from the pollen donor and thoroughly brushing it against newly emerging stigmas of the seed parent. Pollinations for a single cross were performed over five days to ensure pollen transfer to all florets as they opened in the capitulum of the seed parent. All florets of a capitulum opened within five days. Results and Discussion All nine ramets of Erigeron lemmonii that had been treated with rooting hormone developed nodal roots after about four weeks and developed aerial branches and flowers after about 10 weeks 2014] NOYES AND BAILEY: ERIGERON LEMMONII OF ARIZONA 11 Fig. 2. Map of Scheelite Canyon, Huachuca Mountains, Arizona, showing location of plants collected for study. 1: North Main Face; 2-6: Main Face; 7: Boulder; 8-9: Owl Canyon. (Fig. 1). Subsequent tests documented that ra- mets in standard soil without rooting hormone would stay green and appear healthy but would not develop roots. Ramets treated with rooting hormone but placed in potting soil-sand mixtures also rooted but more slowly than in potting soil alone. All rooted plants thrived in the green- houses of the University of Central Arkansas and did not require special watering or light treat- ment. Plants flowered in successive flushes about every three months with flowers emerging on new branches produced from near the base of the plant. Mature capitula inspected for each plant always consisted of empty ovaries; no filled cypselae were ever observed. We conclude that despite the restricted distribution and habitat of the species, E. lemmonii is easily cultured in greenhouse conditions and is not capable of making seed autonomously either by selling or apomixis. The evidence is consistent with the hypothesis that E. lemmonii possesses sporophyt- ic incompatibility, as has been described for other Asteraceae (Gerstel 1950; de Nettancourt 1977). Chromosome counts for all nine plants re- vealed 2n = 18, the first count for the species (Table 1, Fig. 3 A). This is the common diploid number for Erigeron and the presumed ancestral number for tribe Astereae (Brouillet et al. 2009). The chromosome complement is nearly uniform, consisting of approximately equal-length chro- mosomes about three pm long. B-chromosomes were not observed. Pollen produced by the eight plants had high percentage of staining grains (mean 85.8%; Table 1; Fig. 3B). The average measure of the diameters of 85 grains for the Main Face-4 sample was 16.5 pm (SD == 0.79) and the grains all had three evident pores. This grain size is only modestly greater than the mean value of 14.9 pm obtained for 36 diploid populations of E. strigosus Muhl. ex Willd. of eastern North America (Noyes and Allison 2005). For the capitula of North Main Face Erigeron lemmonii, we observed that each ovary consisted of a single standard unitegmic, tenuinucellate, anatropous ovule bearing a single megasporocyte (Fig. 4A). This condition is typical for Erigeron, though it has been reported that some species have multiple megasporocytes within a common nucellus that then compete for dominance in subsequent development (Marling 1951). We 12 MADRONO [Vol. 61 Table 1. Chromosome Numbers and Pollen StAINABILITY for ERIGERON LEMMONII, SCHEELITE Canyon, AZ. Chromosome numbers determined from acetocarmine squashes of root-tips. Pollen stainability reported as percentage of grains (out of 300) darkly and uniformly staining in cotton blue in lactophenol. Plant ID Chromosomes (2«) Pollen stained (%) 1. North Main Face 18 95.3 2. Main Face - 1 18 87.3 3. Main Face - 2 18 76.7 4. Main Face - 3 18 92.7 5. Main Face - 4 18 86.7 6. Main Face - 5 18 80.7 7. Boulder 18 71.7 8. Owl Canyon - 1 18 96 9. Owl Canyon - 2 18 84.7 Mean (SD) 85.8 (8.3) observed variation in the number and placement of cell walls separating the four products of meiosis. Five patterns were observed, three of which were approximately equal in frequency. Of 130 ovules, 42 (32.3%) were consistent with typical monosporic development, exhibiting four nuclei partitioned into separate spores by cell walls (Fig. 4B), and 46 (35.4%) exhibited a bisporic pattern with two cells each bearing two nuclei (Fig. 4C), which results when cell walls form between the two products of meiosis I but no walls form between the products of meiosis IL The third common type (33 ovules, 25.4%) appeared to be a blend of tetrasporic and bisporic development yielding three cells; the micropylar cell contained two nuclei and the two distal cells each contained a single nucleus (Fig. 4D). This pattern evidently results when a cell wall is deposited following meiosis I, but following meiosis II a wall is formed only between nuclei in the chalazal cell. The two low-frequency patterns observed are also interpreted to be developmental mixtures: a three-celled type (8 ovules, 6.2%) similar to the third common type above except that the two nucleate cell was chalazal rather than micropylar, and a type was observed only once (0.8%) that consisted of two cells, a uninucleate micropylar cell and a trinu- cleate chalazal cell. Tetraspory, i.e., the formation of a single coenospore containing all four products of meiosis, was not observed. In subsequent development, we observed most commonly (46 of 59 observations, 78.0%) the expansion and vacuolization of the chalazal spore (whether one- or two-nucleate) and compression and ultimate degeneration of the micropylar spore(s) (Fig. 4E, F, G). In the other 13 ovules (22.0%), we observed expansion of the micropy- lar or a median spore (Fig. 4H). In sum, early reproductive development for E. lemmonii is characterized by an equal mixture of mono- and bisporic types within a single capitulum and a diversity of intermediate forms. The mature megagametophytes yielded egg apparati that were highly regular in structure. Of 124 gametophytes observed, 108 (87.1%) consisted of a single domed egg cell, two wedge shaped synergids forming the micropylar termi- nus of the gametophyte, and two polar nuclei within the central cell usually directly adjacent to the egg cell (Fig. 41). Of the remaining ovules observed, ten (8.1%) either apparently lacked a gametophyte, or the gametophyte was collapsed into a dense mass. In six ovules (4,8%), gameto- phytic cells did not yield a recognizable egg apparatus. For the regular megagametophytes, the polar nuclei were fused into a common fusion nucleus and the degree of fusion varied. Of 62 ovules, 47 (75.8%) possessed a single fusion nucleus with a single large nucleolus, ten ovules Fig. 3. Chromosome complement and pollen for Erigeron lemmonii. A. Acetocarmine chromosome squash for Main Face-5 showing 2n = 18. B. Putatively viable pollen stained with cotton blue in lactophenol is dark; lightly and irregularly stained grains (arrows) are likely inviable. Scale bars = 20 pm. NOYES AND BAILEY; ERIGERON LEMMONII OF ARIZONA 13 2014] Fig. 4. Megagametophyte development for North Main Face Erigeron lemmonii. Abbreviations; a = antipodal cells, c = central cell, ch = chalazal region, e = egg cell, fn = fusion nucleus, m = micropyle, n = nucleus, nc = nucellus, nl = nucleolus. A. Megasporocyte within the ovule prior to meiotic division. B-D. Alternative spore arrangements resulting from meiosis. B. Four uninucleate megaspores indicative of monospory. C. Two binucleate megaspores indicative of bispory. D. Two uninucleate and one binucleate megaspore indicative of megasporo- genesis intermediate between monospory and bispory. E-H. Early megaspore gennination patterns prior to the first mitotic division. Arrows indicate nuclei within the selected developing spore. E. Enlargement of uninucleate chalazal spore, with compression of three micropylar spores. F. Enlargement of a binucleate chalazal spore, with compression of a micropylar binucleate spore. G. Enlargement of a uninucleate chalazal spore, with compression of micropylar uninucleate and binucleate spores. H. Enlargement of a micropylar binucleate spore, with early compression of two uninucleate chalazal spores. I. Mature megagametophyte showing egg, fusion nucleus, and antipodal cells and nuclei within the chalazal panhandle; synergids present but not visible in this view. Scale bars = 20 jam. (16.1%) had a fusion nucleus that was divided by a nuclear membrane into two compartments corresponding to the two polar nuclei, and five ovules (8.1%) contained a single fusion nucleus with two, distinct nucleoli. In contrast to the egg apparatus, the antipodals of the megagametophytes, residing within an elongate chalazal pan-handle (Fig. 41), were varied in number: from two (total nuclei within the gametophyte seven) to nine (total nuclei within the gametophyte 14). Eight-, nine-, and ten- nucleate gametophytes (with three, four, and five antipodal nuclei, respectively) accounted for 48 of 63 of the gametophytes (76.2%) inspected. In terms of reproductive development, Eriger- on as a whole is notable in exhibiting considerable diversity, unlike the majority of Asteraceae, which possess the classic monosporic Polygo- nww-type development (Harling 1951; Pullaiah 1984). Out of a total of 26 Erigeron taxa investigated (Harling 1951), 18 are tetrasporic (69.2%), five are bisporic (19.2%), and only three are monosporic (1 1.5%). There also appears to be considerable lability in development in the genus; five of the tetrasporic species yield occasional bisporic ovules, two of the bisporic species exhibit occasional monospory and tetraspory. In overall pattern, E. lemmonii appears to be most similar to 14 MADRONO [Vol. 61 Table 2. Seed from Controlled Crosses for Erigeron lemmonil Crosses performed for individual capitula selected for the seed parent with pollinations from the pollen parent over a five day period. Percentage cypselae resulting from crosses estimated as number of cypselae divided by the total number of florets (ray plus disc) in the capitulum. Cross (pollen parent X seed parent) # Cypselae / # Florets in capitulum Percentage cypselae formation 1. North Main Face X Boulder 39 / 98 39.8 2. North Main Face X Owl-2 3 / 76 3.9 3. North Main Face X Main Face-2 1 / 119 0.8 4. Owl-2 X Boulder 25 / 85 29.4 Mean (SD) 17.0 / 94.5 11.7 (19.1) E. giabelius Nutt., which is reported, within one individual, to possess equal proportions of monosporic and bisporic derived megagameto- phytes. Erigeron lemmonii differs in that, in addition to monosporic and bisporic gameto- phytes, it produces intermediate types and has one megasporocyte per ovule; E. giabelius pro- duces 2-11 megasporocytes per ovule (Marling 1951). Our data from four experimental crosses for Erigeron lemmonii show variation in percentage of seed produced per capitulum (0.8 to 39.8%, mean of 1 1.7 cypselae; Table 2). The North Main Face plant, even when used as a common pollen donor produced only one cypsela (0.8%) when crossed with Main Face-2 but 39 cypselae (39.8%) when crossed with Boulder. These data may indicate reproductive barriers in the popu- lation. Given that the E. lemmonii population is relatively small and evidently self-incompatible, it is possible that S-allele diversity has been reduced, which would limit opportunities for successful reproduction in the population (Busch and Schoen 2008). Our data at least show that plants are capable of producing seed; the design prevents further strong inference. Our observations shed light on reproduction in Erigeron lemmonii. First, it is surprisingly easy to cultivate vegetatively under standard greenhouse conditions. This means that if the population in Scheelite Canyon were to be threatened, ex situ propagation would likely not be problematic. Second, it is diploid {2n = 18), sexual, and self- incompatible. This means that it appears to be a distinct evolutionary lineage and is not an asexual polyploid microspecies. This potentially greatly simplifies the rationale and strategy in developing a conservation plan for it (Hey et al. 2003; Ennos et al. 2005). Third, our data show that E. lemmonii individuals exhibit a mixture of mono- sporic and bisporic female gametophyte develop- ment. Notwithstanding, this condition does not appear to limit its ability to produce regular appearing mature gametophytes. In the absence of further reproductive and phylogenetic data, it is uncertain if this condition is an isolated occurrence, or if it characterizes a group of related taxa. Acknowledgments We thank Kim A. Whitley, Northern Arizona University, for help collecting ramets, the University of Central Arkansas for the use of greenhouse space, and John Strother (University of California, Berkeley) and an anonymous reviewer for comments that improved the manuscript. Field collection funding was provided by the United States Army’s Environmental Quality and Installations Basic Research Program as part of a larger research project (Project 09-03). Literature Cited Brouillet, L., T. K. Lowrey, L. Urbatsch, V. Karaman-Castro, G. Sancho, S. Wagstaff, and J. C. Semple. 2009. Astereae. Pp. 449^90 in V. A. Funk, A. Susanna, T. F. Stuessy, and R. J. Bayer, (eds.), Systematics, evolution, and biogeog- raphy of Compositae. International Association for Plant Taxonomy, University of Vienna, Austria. Busch, J. W. and D. J. Schoen. 2008. The evolution of self-incompatibility when mates are limiting. Trends in Plant Science 13:128-136. Cronquist, a. 1947. Revision of the North American species of Erigeron, north of Mexico. Brittonia 6:121-302. De Nettancourt, D. 1977. Incompatibility in angio- sperms. Springer-Verlag, New York, NY. Ennos, R. A., G. C. French, and P. M. Hollings- worth. 2005. Conserving taxonomic complexity. Trends in Ecology and Evolution 20:164—168. Gerstel, D. U. 1950. Self-incompatibility studies in guayule. II. Inheritance. Genetics 35:482-506. Gori, D. F., P. L. Warren, and L. S. Anderson. 1990. Population studies of sensitive plants of the Huachuca, Patagonia, and Atascosa Mountains, Arizona. Unpublished report. Coronado National Forest, Tucson, AZ. Gray, A. 1883. Contributions to North American botany. I. Characters of new Compositae, with revisions of certain genera, and critical notes. Proceedings of the American Academy of Arts and Sciences 19:1-96. Marling, G. 1951. Embryological studies in the Compositae. Part III. Astereae. Acta Horti Ber- giani 16:73-120. Herr, J. M., Jr. 1971. A new clearing-squash technique for the study of ovule development in angiosperms. American Journal of Botany 58:785-790. Hey, j., R. S. Waples, M. L. Arnold, R. K. Butlin, and R. G. Harrison. 2003. Understanding and confronting species uncertainty in biology and 2014] NOYES AND BAILEY; ERIGERON LEMMONII OF ARIZONA 15 conservation. Trends in Ecology and Evolution 18:597-603. Malusa, J. 2006, Surveys and monitoring plots of the Lemmon's fleabane, Erigeron iemmonii, in the Huachuca Mts, Arizona. A report submitted to the United States Fish and Wildlife Service, the University of Arizona and the Arizona Department of Agriculture. Nesom, G. 2006. Erigeron. Pp. 256-348 in Flora of North America Editorial Committee (eds.), Flora of North America North of Mexico, Voi. 20: Magnoliophyta: Asteridae (in part): Asteraceae, part 2. Oxford University Press, New York, NY. . 2008. Classification of subtribe Conyzinae (Asteraceae: Astereae). Lundellia 11:8-38. Noyes, R. D. 2000. Biogeographical and evolutionary insights on Erigeron and allies (Asteraceae) from ITS sequence data. Plant Systematics and Evolu- tion 220:93-114. AND J. R. Allison. 2005. Cytology, ovule development and pollen quality in sexual Erigeron strigosus (Asteraceae). International Journal of Plant Sciences 166:49-59. PULLAIAH, T. 1984. Embryology of Compositae. International Bioscience Series 13. Today and Tomorrows Printers & Publishers, New Delhi, India. Stanley, R. G. and H. F. Linskens. 1974. Viability tests. Pp. 67-86 in R. G. Stanley and H. F. Linskens, (eds.). Pollen: biology, biochemistry, and management. Springer- Verlag, Berlin. U.S. Department Of The Interior, Fish And Wildlife Service. 2012. Endangered and threatened wildlife and plants; 12-Month finding for the Lemmon Fleabane, Endangered Status for the Acuna Cactus and the Fickeisen Plains Cactus and Designation of Critical Habtitat: Proposed Rule. Federal Register 77:60509- 60579. Madrono, Vol. 61, No. 1, pp. 16-47, 2014 PHYLOGENIES AND CHROMOSOME EVOLUTION OF PHACELIA (BORAGINACEAE: HYDROPHYLLOIDEAE) INFERRED FROM NUCLEAR RIBOSOMAL AND CHLOROPLAST SEQUENCE DATA Genevieve K. Walden Department of Integrative Biology, University of California, Berkeley, CA 94720 gk walden@gmaiL com Laura M. Garrison Department of Ecology and Evolutionary Biology, Brown University, Providence, RI 02912 Greg S. Spicer, Frank W. Cipriano, and Robert Patterson Department of Biology, San Francisco State University, 1600 Holloway Avenue, San Francisco, CA 94132 Abstract This project sampled throughout Phacelia using the internal transcribed spacer region (ITS-l, ITS- 2, and 5.8S gene) of nuclear ribosomal DNA (nrlTS) and the chloroplast DNA gene {ndhV) to infer phytogenies for nuclear and plastid partitions. Nuclear and plastid partitions were incongruent in our analyses. Phylogenetic analyses (maximum parsimony, maximum likelihood, and Bayesian inference) recovered gene tree topologies similar to previous molecular studies. We corroborate incongruence between nuclear and plastid topologies for placement of some problematic groups (e.g., Draperia, Romanzoffia and “core” Phacelia subg. Pulchellae, Phacelia sect. Baretiana). Combined analyses resulted in better resolution than separate analyses, and in a topology that favored the separate plastid topologies. Romanzojfia was sister to a monophyletic Phacelia in the combined analyses. Our results support combining incongruent partitions in a combined analysis to seek support for internal nodes. Maximum likelihood analyses were used to infer ancestral chromosome numbers and identify gains, losses, polyploid doubling, and whole genome duplication events from published chromosome counts in the genus. The predicted base number for the genus was x = 9, x = 11, or x = 12. Key Words: Boraginaceae, California flora, chromosome evolution, Hydrophylloideae, incongruence, molecular phylogenetics, Phacelia, Romanzoffia. Phacelia Juss. is the largest genus (207 spp.) in Hydrophylloideae (Boraginaceae). The majority of species (176 spp.) are distributed in western North America and an additional 16 species occur into Central America, with an amphitropi- cal disjunct group of nine species in southwestern South America. The center of diversity for the genus is the California Floristic Province (CFP); a third of described taxa occur within the CFP (ca. 70 spp., 40 spp. endemic) and ca. 90 spp. occur within the political boundaries of the state (Raven and Axelrod 1978; Patterson et al. 2012). In California, 33 taxa in Phacelia are ranked in the California Native Plant Society Rare and Endangered Plant Inventory (CNPS 2011). These include one of three federally endangered taxa {P. insularis Munz var. insularis) and one candidate taxon considered for federal protection {P. stellar is Brand) (U.S. Department of the Interior, Fish and Wildlife Service 1978, 1982, 1997, 2004, 2011) . Thus, Phacelia is one of ten largest genera and Boraginaceae one of ten largest flowering plant families occurring in the CFP and in California (Beard et al. 2000; Baldwin et al. 2012) . Phacelia, as the largest and most diverse genus in Hydrophylloideae, is often used as an example of the diversity of the California flora (Stebbins and Major 1965; Raven and Axelrod 1978; Ackerly 2009; Kraft et al. 2010). However, research regarding evolution and diversification in the genus and its significance in the California flora has been limited due to the lack of a well- resolved, broadly sampled molecular phylogeny with congruent nuclear and plastid partitions. Recent Molecular Studies Relationships of major lineages in Phacelia have been previously studied using molecular phylogenetic methods. Gilbert et al. (2005) combined thesis work in Phacelia sect. Euglypta S. Watson by Dempcy (1996) and in Phacelia sect. Miltitzia (A. de Candolle) J. T. Howell by Ganong (2002), along with sequences from dissertation work by Ferguson (1998), to publish an nrlTS partition of 51 taxa (84 accessions) in Phacelia. Gilbert et al. (2005) recovered Roman- zoffia Cham, sister to a monophyletic Phacelia and recovered a monophyletic Phacelia subg. Microgenetes (A. de Candolle) A. Gray that included a paraphyletic Phacelia sect. Euglypta and a paraphyletic Phacelia sect, Miltitzia. 2014] WALDEN ET AL.: PHYLOGENIES AND CHROMOSOME EVOLUTION OF PHACELIA 17 Hansen et aL (2009) published an nrlTS partition of 56 taxa (91 accessions) and a rpll6 intron partition of 22 taxa (37 accessions) for Phacelia sect. Gymnobytha (A, de Candolle) Benth. & Hook.f. and Phacelia sect. Whitlavia (Harv.) Benth. & Hook.f. Hansen et al. (2009) recovered Romanzoffia sister to a monophyletic Phacelia and recovered a paraphyletic Phacelia sect. Whitlavia in the separate nuclear partition. Phacelia was paraphyletic in the separate plastid partition. Both Phacelia sect. Euglypta and Phacelia subg. Pulchellae (Rydb.) Walden & Patt. were recovered as basal lineages with Romanzof- fia nested within the genus. The combined nrlTS and rpll6 intron analysis recovered Romanzoffia sister to a monophyletic Phacelia and monophy- letic infrageneric sections (Hansen et al. 2009). Ferguson (1998 [1999]) included 19 taxa (19 accessions) of Phacelia within a larger analysis of Hydrophylloideae for ndhF, recovering a paraphyletic Phacelia. Phacelia subg. Pulchellae was sister to a nested Romanzoffia and remaining sampled Phacelia. Collectively, researchers have published 124 accessions within Phacelia using the internal transcribed spacer region (ITS-1, ITS-2, and 5.8S gene) of nuclear ribosomal DNA (nrlTS), published 20 accessions for the chloroplast (cpDNA) gene ndh¥, and published 37 accessions for the chloroplast (cpDNA) marker rpll6 intron (Ferguson 1998 [1999]; Olmstead et al. 2000; Gilbert et al. 2005; Hansen et al. 2009; Glass and Levy 2011). We combined previously published nrlTS sequences with thesis work in Phacelia sect. Glandulosae (Rydb.) Walden & R. Patt. and Phacelia sect. Ramosissimae (Rydb) Walden & R. Patt. of Garrison (2007) in an expanded nuclear phylogeny to infer inter- and infraspecific evolu- tionary relationships in Phacelia. We combined previously published ndh¥ sequences with thesis work in Phacelia of Walden (2010) in an expanded cpDNA phylogeny to infer infrageneric relationships in the genus. Previous molecular phylogenetic studies in Phacelia and Hydrophylloideae have identified significant phylogenetic incongruence between nuclear and plastid partitions (Ferguson 1998; Moore and Jansen 2006; Hansen et al. 2009; Weeks et al. 2010; Nazaire and Hufford 2012; Taylor 2012). For an extensive discussion on incongruence between nuclear (nrlTS) and plas- tid {ndh¥) partitions in Phacelia and Hydrophyl- loideae see Ferguson (1998). When tests for homogeneity between partitions (e.g., incongru- ence length test [Farris et al. 1995]) reject the null hypothesis, a combined analysis is inappropriate and partitions are analyzed separately using the conditional combination approach (Bull et al. 1993; Huelsenbeck et al. 1996). Some researchers combine incongruent partitions using simulta- neous analyses, arguing that these combined analyses provide greater resolution than separate analyses of incongruent partitions (Nixon and Carpenter 2005). Our goal was to determine if increased sampling for a respective molecular marker (nrlTS, ndh¥) within Phacelia recovered similar gene tree topologies to previous studies for separate analyses of partitions, to compare nuclear and plastid partitions for character homogeneity and combinability (Cunningham 1997), and to determine if simultaneous analyses provided enhanced resolution for a reduced subset of samples for which both nrlTS and ndh¥ sequences were available. Chromosome Evolution Phacelia occupies a range of habitats and exhibits a variety of life history traits and ecological adaptations. Species differences have been traditionally based on morphological (e.g., seed shape and number) and cytological charac- ters (e.g., chromosome numbers), and less so on ecological factors (e.g., edaphic factors). These characters have been used to diagnose infra- and interspecific taxa, and to delimit infrageneric groups within the genus (for a review and current infrageneric classification in Phacelia, see Walden and Patterson [2012]). Both Constance’s (1963) and Gillett’s (1968) classifications were based largely upon chromosome numbers, drawing upon Constance’s extensive collaborations into chromosome number differences in Hydrophyl- laceae with Marion Cave (see Cave and Con- stance [1942, 1944, 1947, 1950, 1959]). Phacelia benefits from published chromosome counts for approximately two thirds of the genus, ranging from n = 5 {P. dubia [L.] Trek & Small, P. maculata Wood) ton = 33 (P. hastata Douglas ex Lehm. var. compacta [Greene ex Brand] Cron- quist, P. leptosepala Rydb.) (Cave and Constance 1947, 1950; Kruckeberg 1956; Kovanda 1978). Constance (1963); Heckard (1963), and Gillett (1968) hypothesized that n =\ \ was the ancestral condition for the genus and noted it was also the most common haploid count for extant taxa. Hypotheses proposed for the base number for the genus have not been tested in a broad phyloge- netic context. Previous studies considering evo- lution of chromosome numbers in a molecular context in Phacelia include mapping of chromo- some numbers to nrlTS tree topologies by Gilbert et al. (2005) and to ndh¥ sequence data using maximum parsimony by Walden (2010). Reconstructing ancestral states using a maximum parsimony approach without an explicit frame- work (e.g., biosystematic studies of chromosomal rearrangements across the genus) allows only for coding with a categorical character matrix using the unordered states assumption in Mesquite version 2.74 (Maddison and Maddison 2010), regardless of whether a transition represents an 18 MADRONO [VoL 61 increasing or decreasing dysploidy event or a doubling polyploid event (Mayrose et al. 2010). Although the maximum parsimony approach has real merit, the unordered states assumption option offers little resolution for this dataset at the present time. We were interested in determin- ing the ancestral base number for Phacelia to better understand patterns of chromosome evo- lution within infrageneric groups and within the genus using a maximum likelihood approach (Mayrose et al. 2010; Hallinan and Lindberg 2011a). chromEvol version 1.3 (Mayrose et al. 2010) and GDCN (Hallman and Lindberg 2011a) use explicit likelihood models of evolution to infer ancestral states for chromosome numbers at nodes in phylogenies from rooted ultrametric trees. These analyses offer the ability to test hypotheses for the base number for infrageneric groups and the genus using results from our expanded nuclear and plastid phylogenies. chro- mEvol v.1.3 and GDCN estimate probabilities of chromosome evolution events at nodes to explore patterns of gains, losses, polyploid doubling, and whole genome duplication (WGD) events within a known phylogeny (Mayrose et al. 2010; Hallinan and Lindberg 2011a). Saltational sped- ation has been an important factor in cladogen- esis in the California flora; we were interested in identifying any recent genome duplication events within Phacelia phylogenies using a maximum likelihood approach (Stebbins and Major 1965; Raven and Axelrod 1978; Wood et al. 2009; Hallinan and Lindberg 2011a). Materials and Methods Chromosome Numbers in Phacelia We reviewed the literature for published chromosome counts for taxa in Phacelia and outgroups sampled in this study (Table 1). Taxa are presented alphabetically for outgroups {Eu- ploca Nutt., Eriodictyon Benth., Draperia Torr,, Hesperochiron S. Watson, Howellanthus [Con- stance] Walden & R. Patt., Nama L., Romanzof- fia, Tricardia Torr. ex S. Watson) and ingroup taxa within Phacelia. Original names for pub- lished counts are noted for synonyms or where different from the current accepted name or specimen determination. Chromosome numbers for haploid {n) or diploid {2n) counts are given as published. We chose not to include references reporting unpublished counts or citing personal communications for these analyses. We note if a voucher specimen was not cited in the notes column of the table. No attempt was made to locate and examine all voucher specimens cited for each published count for this study. Refer- ences may include one or more counts for an individual taxon, we do not include summary numbers of the individual counts for each reference or a comprehensive list of voucher specimens and karyotype figures. Constance (1963) reported a count for Phacelia pauciflora S. Watson without reference to a voucher specimen. We included this count and corre- sponding voucher specimen examined at the University of California Herbarium (UC) in Table 1. Chromosome counts originally pub- lished as taxa in Phacelia but corrected or redetermined in later publication were excluded. Citations are listed chronologically within each taxon and a list of full references follows the table. Chromosome counts were not directly obtained from individuals or populations includ- ed in the direct sequence analyses for this study. This limitation may obscure cryptic diversity in sampled populations. This list should be consid- ered a working draft of chromosome numbers for Phacelia and Boraginaceae. A future comprehen- sive review of chromosome counts in Boragina- ceae is anticipated as a useful resource for workers in the family (G. K. Walden, unpub- lished manuscript). Taxon Sampling This study represents a joint publication of thesis work from Garrison (2007) and Walden (2010). The expanded nuclear partition (176- accession) included 89 taxa (42% genus) and the expanded plastid partition (126-accession) includ- ed 90 taxa (43% genus). Sampling within Phacelia for the expanded nuclear (176-accession) and plastid (126-accession) partitions included repre- sentatives from all subgenera and sections; the reduced (61 -accession) partition lacked a repre- sentative from Phacelia sect. Pachyphyllae Wal- den & R. Patt. Lor the nuclear partition 48 taxa were sampled from California, 44 taxa sampled from western North America, 5 taxa were sampled from Central America, and one taxon was sampled from South America. For the plastid partition 53 taxa were sampled from California, 37 taxa were sampled from western North America, five taxa were sampled from Central America, and one taxon was sampled from South America. Accessions of Romanzoffia were included to assess the relationship to and monophyly of Phacelia. We included accessions from Hydrophylloideae to briefly assess relation- ships between Phacelia and exemplar taxa and for purposes of chromosomal evolution. Euploca (Heliotropioideae: Boraginaceae) was included as the diploid outgroup to root the tree. Lield collections of fresh plant material were., preserved in silica gel for molecular work and voucher specimens were deposited in the Harry D. Thiers Herbarium at San Lrancisco State University (SLSU). Additional material was sequenced from banked molecular vouchers with herbarium vouchers received from the William L. 2014] WALDEN ET AL.: PHYLOGENIES AND CHROMOSOME EVOLUTION OF PHACELIA 19 Oc2 i I G S O o 3 a -n ^ o 'd Si ft d cd d o 0\ M m ^ m M 00 CN CN ^ CN ^ m <-H tsj m ^ ^ i ^ ^ o u 0 w g « M 0 I s &5 s g « il 0 <4) §• o o A 2 2 00 << =s =s »S "S s s & 5 5 A A r- ©> ON o r-- o>i ON r-- r-. ON e^ ©>■ O ^ c-> fN| CN O On r^ ^ ^ VN VN in ^ VI ^ m ^ VN VN 0\ ON CTn ON ON ON ON ON On ON ON ON d" ON On On On On On On On ON r— ! T— < 1—^ r— i F— ( !— ! »— < S-H <—«>—< f-H F— 1 On f-N 5—4 1—4 f-4 f-N 1^ 1— 1 s— i 8 § 8 8 8 8 8 8 ® w o O O Q 8 8 8 8 8 4) W W O O O O O O O fl d d d d d d d d d d d o d d d d d d d d d ^ 5 -M "M d S d d d d d B d B -2 cd cd si ^ Q\ m m m m w m Cfl M M C« c« S w M Cfl m M CO 60 Cfl CO > > > ^ 8 > > > -3 > > > > S > s> > d >■ > > > > > > d d cfl u u u u ^ ^ ^ u m u KUUUU^UUU^^UU d d c^ „ U U U U U w o% OO M u M •p 0 •S ■fc § 1 s D O A §3 8 -p p ctf oS •§11 e § ^ ill s^l A a5 :s H « M ^ ft ^ of o I 0^ S A 2 0 < •S I d o >. ^ 6^ W , H XO d © cd « . .2 m $=N ^ 0 - !-< « 0 -d I m d ^ o d - o fc A 2 0 < ]o o H ¥ f o 1 A S =S •fct S |.| C Sf. •3 1 K. .5 o ,o ^ o ft K -A ^ 5 o o s ilit -S fO »q »Q A A A A A A A A A A A •S 0 5^. W -S s s s R 1^ 2 N p p cd d . > > A u u 00 g " " "d •S -S ^ c c ^ ^ ^ p" s s P 1 13 S 20 MADRONO [Vol. 61 0) B cd C g _3).2 O ^ 3 3 a .'S ^ 5 S=i cd y cd ^ O G > y G y G M y y G y 0 y < S y y y a O O 2 c« X c,a ^ .Q a ■S 3 O 0 *3 K Q G O U § 1 ■-S g cs S ~s; as c 2 s .y X H -2 S y y 6<) S y y y y y y 'y y y y a,- a,' ca: a; a,- O S > O > 3 < « ^ Q b < •£ ^ X ^ ^ S s « s ^ ’S K O O S Cj O CJ d d d d CN ^ (N ^ O fNl d I -1 ^-n si c ^ "S u ^ H > • « s g ^ s cd w CD 0^0 >-< w •2 3 i« Co QJ m 2 y X ^ rK ^ •2 0 .g s; i=r ^ Co y 2 y 6- y 2 y. s| I S d -S a PQ as 5 y =3 s ^ y K 2 "2 •S -2 v2 2 ^ %S?E y y y ^ »y a.’ a^ a; a^ a; a.' oC a; a; OS in os ' OJ (U o o e fi cd cd os r-- os f\i ^ ^ in ^ os os Os os (U (U > ^ cd cd N U U 'O "O "d 'd deed cd cd cd cd y y y y > > > > cd cd cd cd u u u u fN fN d- d- os os y y y y d s cd cd CO CO a a O O U U -d -d a c cd cd y y > > cd cd U U d d ^ U 00 C« 00 CO G G OQ G G 0 0^00 U U 2 U U -d 'd G G cd d ^ 'S "g -d G G G d cd y y y y > > > > > cd cd O cd cd U U X U U o o m in (N cN ^ ^ — I rsi (N 00 ^ 00 < 00 y Q y d ^ O o -S H S 2 S ^ 'S y G . y G ^ ’5 0 S Q 0 cn y 0 cd fci jO-S O < . G H :k < « W S 5 s K ^ s dj ^ dj ^ ^ ^ y Q y y y s ^ ^ y ^ ^ ^ :a 05 g 6<) y ^ •S -S cd u o q; a,' a,- q, a, a ^ G 2 y W H 1^ “i A A y G y y u o Q s y y Haploid Diploid Original Taxon count («) count (2«) Reference publication name Notes chromEvol GDCN 2014] WALDEN ET AL.: PHYLOGENIES AND CHROMOSOME EVOLUTION OF PHACELIA 21 X o VO 2 s i cd '(J -M -(-> (n 00 M W 00 > fi fl fi VO :0 O O O 2 J u u u na -13 -a no fi fi fi fi p cd ^ > > > > o > Q ^ ^ *0 cd cd UOS JUUU >> s o 2 X m ^ I (N fN O ON O ON in in ^ in in ON On On ON ON OJ O > > > p p p p U U U U o o p ^ P OJ t« ^ P o O Pm U o ON m - § 'I "i ^ K K O ^ ^ >> p 0 $ 06 X 0 u j d 0 ^ E Q 6 u >. 0 |s^. ^ S s; <"§ S § S W O C) _ |« 5 p O !-i X — o ^ o p ^ O o Sh X X! P P P P J-i 5-1 ffl pq fv. woo P P U O (U ^ OWN? 5^ S-I > II .bo .bo > o § 8 K ^ 5-( SO § s ^ VJ S -s s -s 5 ^ ON ON P o X g X e p p p ^ p > % > p o ^ U Pm U 0^ o o p p VO VO ^ ^ g On On O -H ^ U bC iso ^ 5-1 5-i ^ o o ^ X -Q Poo o o o > p p P l-l 5-j U X X fN r- ■p- 'p- ON ON o o o o p p p p 4—* 4=J &0 DO p p O O u u X X P P P P > > P P u u o in On o O P P w P O U X p p o > P U o o p VO p p o gpu ^ § o o p > 5m 03 X U m ^ (N (N ^ CN fN . bO l-M 0 S Q ^ ^ Sp S p P Q O ^ I ^ r K ^ ts»*. ^ f n bopH bobobobcibf)bobobf)-^^ X 53 o r- in ON ON o o p p +-> un P O ON U oo 00 ^ VO 2) P 2 ^ p ^ o3 ii^ ^ ’3 -P P P CO U GO p F. howelUana N. D. Atwood 22 MADRONO [Voi. 61 I «3 •a 5 A o .a 3 a rs -2 ^ _a g Q Q o 3 s S ^ 'E fl cd p i o ^ o 1 XXI A ^ “O II ?3: g X « ^ 0'S ■§ S a . 1^. « w d d § 2 00 q; a. fN| ^r■ CNi ^ § 8 d d « o d 3 cfl w ca M W3 ■ ■ d ^ ^ o 8 § d d d o o d d o o UUUUU -a -d ’d -o d a a a fl cd ^ ^ ^ ^ ^ 0) cd ^ ^ ^ UUUUU u >> 2 S 0 I < I I o 0\ d' d 53 d » d o o a U !.»< U Q H ;§ •H •S A o S2 « ^ fl M U o\ d H -d d O'. « d g w ^ ^ cd g ^ ^ g ^ ^ > i §1 i > U £ N M £ U 0 I I •S .S S -S d “S H y &3 > I 8 -a 0 ^1 .EE is SH S CNi M m ^ ^ CN o m »-< ^ ^ CN M M m v O ^ ^ ^ T-< CNI fN ^ $ 4? ^ H U X X o u e < o o K a P r- ON ON On •4^ in IT) On ON On ON IF- 1 w—t 5—4 T— 1 « « « W o Q o o d d d d B E d B M M Cfl CO d d d d o 0 o o U U U U -d 'O -d d fl d fl e« d a d » w « « > > > > d cd d U U U U 5 d d o o d d > > a > ^ o g *r> ^ l:' 0\ ON d E U W _a O O n fe d d U O c^ ^ >4=J 4-^ ^ ^ U!3 d d d S H o o ca-oUU &§Cfi S 5-< d d d -2 u « 'd ^ ^ ^ Cd TO OUHUU > ^ m m m ON ON c^ 3 O fO ^ H :s • .d ^ -s ^ S "5 b« < I a ^•S -S •2 *0 ^ 1| i ^<1 i Q E 5 ^ d O' o S i a ? s' ^ 3 O 'g S' ^ § ffl £ 0 u e < © t o h s 2 1^ eh M r-' ON ON o o o o d d «43k:J ce m d d O O O S U U ON VN — ' ON 'd 2 S S § .2 » « W bO > Sd > otUOU *-i ^ — i CNI i^&i ^ ^ 2 Sll ■S^O to § s a A .s a^ a^ a; A On 0^ a: A 0^ a, ^ I ^ i •s i 5 5 a; Table 1 . Continued. 2014] WALDEN ET AL.: PHYLOGENIES AND CHROMOSOME EVOLUTION OF PH AC ELI A 23 a cd 2 ]3)-2 o .a 3 3 a .'2 fN o ^ Q § o S ^ cd 3 X 8 00 a^ ^ § I -S i ^ ^•S fU d ^ > cd 5^ 3 d U ^ 0 u OD flD dD O CD CD add cd cd cd CO 00 CO Add O O O UUU 'U 'd ’O d d d cd cd cd > > > cd cd cd UUU (N (N ^ — I ^ »-H (N (N CnI — I ^ — I — I (N ^ rsi Os Xo' GOO 0 8 d so cd Os g O MU ^'g CD d CD (U d > cd u (N -H CD Cd d " O CO PQ . CD (D 0) d '"O (U cd S c I 2 2 0 < cd u 0 < O ^ IT) Os Os (D OD O CD d d cd cd 00 CO d d o o U U -d X3 d d cd cd (D SD > > cd cd u u r- o ^ IT) Os OS OD (D O CD d d B B CO CO d d o o U U 'd -o d d cd cd (D OJ > > Cd cd u u o^ in os Os O CD O O d d cd cd cn CO oo ^ fi a o o ^ U U Oi CD XS 'd d d d cd cd cd +-> c^ (D flD > > cd cd UUU Ss o CD fN! d ^ ^ e e p o H U^ xs g d > ^ c^ O u 0^ (N a^ r- ^ in ^ Os Os Os Os fU O flD > > > cd cd cd cd u u u u -H fM On G « •S 5e S K ^ O G dj ?u SS K K cl: CL a; 0 G cd I “ S 3D A N E ^ d ^ -§^ G K K G ^ CJ ^ O O ^ u ^ O Cl 5«. CJ G G a, a, O < cd S-i o d d cd a 3 M d . W oi t a G G CL a, I d X ^ d > g 2 s io |o ><<■ %-Uq=^ t cd > O O d X G d Mr • G hd 'sf ^ in Os ids ON ON On ON aD CD CD O d d cd cd dD « OD 3D o o o o d d d d cd ^ cd cd OT CO d d o o _ _ _ _ UUUUUU CO CO CO CO d d d d o o o X X d d cd cd > > cd cd u u X X X X d d d d cd cd cd cd 3D (U 3D 3D > > > > cd cd cd cd u u u u 3D X I t m fN O ri cd cd cd *:sZ m 2 Q o ^ ^ ^ G H g “G C E g, Q ^ ^ o ^ ^ Ci, G G G 3D X U O X O 3D ^ CO Cd 8"^ *5 8^d o X 0 d ^ G < < ^ « -s 2 ^ 8“S S ^ S Ci, SL G, s a;a;aicL aia^ cLCLai^cL a:cLa;cLa:Q; c o a § ^ 2 ;§G ^•3 K ^ G a; a: ^ 3D g M O 3D X o d ^ 2 D ffl a a Table L Continued. 24 MADRONO [Voi. 61 O .a 3 3 a rs 3 3 "S Q g 3 s 'o ^ "E d cd d i o 00 ^ 04 ^ 3 3 ’^1 o^ u -O d NN 5_, Q 0 S w fs >, o ^ ^ ^ A 0 o o o u 'U 'd d d > > u u VO CT\ §■« d g d o d &'d Si d = S -S ? ^ ^ a 0 Q U U m « d > O Ki U U M m d a ^ 0\ ^ ^ d Crt d d 04 5^ 0) 'S .3 0) O Pm O M ^ ^ ^ ^ Ov o\ a\ Q\ O E fNl 04 VO 04 04 ^ W ^ u M W e d o o U U d d d eti > > d d u u 04 0\ > cd s_ 0 s A Cd < 2 «s ^ Q a jv) O vj ^ Co ^ A A m d ad O W u c^ NM II -S a ~d ^ ^ ^ t.« ■*« ^ C^ 6, O Q ss o 'S ^ VJ Q S a Q sg Q u w d g > > b 04 d . ^ 04 04 ^ 43 <^4 -d 2 g k « X X t =2 ?3 cd > d d o o H H 3 3 d d 04 04 ffl ffl u 2 •2 -2 A A 0 Ov f^ ov ov o ov o o A O in If in in in in in m Ov ov On ^ ov ^ ov ov ov ov ov ov ov ov A 44 0\ -d cd ■CH ,-H 04 04 04 P 04 Si 4) 04 04 04 04 44 O O w f^ o o O o O o o o O O d d d « d fl d d d d d d cd d d u d ffi d id d d d c^ d cd W M d d o o U U -o d d 04 04 > > d cd U U d 04 ffl 2 ^ < s ft, a, a, a, tt. AAA 'U 13 O o o SS II 0, CO >> d d o « IT) Cfl o d • in o! VO oo 00 ov cn oi 1 ov m I ^■vocn Ph ^ g- cd .-d .O • na “ S ^ O « E d d E S ^ d d d ’d ra ^ O O ^ ^ ffl ffl ^ « d d ^ .2 M d . o c« m d d o o d rU .3 M 32 S.S « I U ^ 2P ^ Cd ^ ■5.'g o-M-^ ^ 04 'g ^ .O ^ lii££| S d H ffiwggS’ I" i t2 v2 5 ^ '3 A 3 '-g < ^ ^ «4_< ^ 3 §• ° M ° C 4i §■ >- ST-a S2'2's a® « > g S 6 2 -g c c P M ■'^ i_3 K X 0) c« ■ a p > 2 ^ oJ' Si of 04 ^ d d d d I 8 8 8 8 d c^ c^ d c3 ^ Sd ^ ^ ^ 04 X >01 A» A» g 43 .g .g ,g 13 0) +J g s d a a oi P, o o o o o A U u u u u 2 'd 'U 'd 'U Q -M o a a ffi a S ,5 04 04 04 94 P (J ^ ^ ^ ^ .g .g .s .s 2 dv Cfl c« M 04 04 04 04 m ; ^ ^ ^ ^ 3 ! g S 0 S w : d d d d ^ P d fT) Q N « • A «J lASd 04 44 d K d 04 . 04 O 4) d ^ o S a d 0 g hu P &0 04 d 5-< A ^ 11-; 1 §2 g 04 S w) d ^ ^ o ^ 2 d cd g g ^ S .S O N P d ffl P A ^ ^ a d >.43 A « 'd »4 A bo d Jn d g; cd >>A d w 5-1 feH pH A > o A :0 m d a P w Q g A a 2 S s ^ Q o< 0\ . ^ VO a a ® Q p ■2jpg^ Pd 2014] WALDEN ET AL.: PHYLOGENIES AND CHROMOSOME EVOLUTION OF PHACELIA 25 Zhao, Z. 1996. Documented chromosome numbers 1996: 2. Miscellaneous USA. and Mexican species, mostly Asteraceae. SIDA: Contributions to Botany 17:259-263. AND B. L. Turner. 1993. Documented chromosome numbers 1993: 3. Miscellaneous USA. and Mexican species, mostly Asteraceae. SIDA: Contributions to Botany 15:649-653. 26 MADROto [Vol. 61 Brown Center at the Missouri Botanic Garden (MO) and from living collections without her- barium vouchers received from the University of California, Berkeley Botanic Garden. University of California, Davis Center for Plant Diversity (DAV), San Diego Natural History Museum (SD), SFSU, and the University of Texas at Austin (TEX/LL) granted permission to destruc- tively sample herbarium specimens. Sample details, including voucher specimens, botanic garden accession numbers, herbarium accession numbers, and GenBank accession numbers for all sequences are included in Appendix 1. Updated identification of voucher specimens for accessions is noted parenthetically where these differ from previous studies. GenBank accessions and vouch- ers cited in previous studies excluded from these analyses are listed in Appendix 2. Gene Regions Sampled We examined the internal transcribed spacer region (ITS-1, ITS-2, and 5.8S gene) of nuclear ribosomal DNA sequences (nrlTS) for the nuclear partition (biparentally inherited charac- ter). nrlTS is a highly variable and rapidly evolving region, useful for inferring species level relationships (Baldwin et al. 1995; Alvarez and Wendel 2003). nrlTS is the molecular marker with the largest representative sampling of sequences published to date for the genus. For the plastid partition (maternally inherited char- acter) we examined the single copy gene ndh¥. The chloroplast NADH dehydrogenase gene ndh¥ codes for the F (ND5) subunit; for an expanded discussion of ndh¥ sequence evolution and phylogenetic utility see Kim and Jansen (1995). The ndh group of chloroplast genes is highly conserved across major plant taxa and has been used to assess and define infrageneric evolutionary relationships in plant families (Ney- land and Urbatsch 1996; Olmstead et al. 2000; Martin and Sabater 2010). ndh¥ also has a broad sampling of sequences published for Hydrophyl- loideae and Solanales (Olmstead and Sweere 1994; Bohs and Olmstead 1997; Ferguson 1998; Moore and Jansen 2006; Taylor 2012). DNA Isolation and PCR Amplification DNA isolation. Total genomic DNA was extracted from 0.020mg dry weight herbarium or silica-dried leaf material following Doyle and Dickson (1987) using Qiagee DNEasy Plant kits (Qiagen, Valencia, CA, USA) and grinding with liquid nitrogen for the homogenization step. For ndh¥, the protocol was modified to use homog- enization with acid-clean steel ball bearings in lieu of grinding with liquid nitrogen and substi- tution of NucleoSpin filters (Macherey-Nagel, Duren, Germany) for the filtration step. PCR amplification for nrlTS. Total genomic DNA extract was diluted 1 : 10 in ultra-pure H2O for best amplification in PCR (Mullis et al. 1987) for the internal transcribed spacer region (ITS-1, 5.8S, and ITS-2) of nuclear ribosomal DNA (nrlTS) with primers ITS-I (F) (Urbatsch et al. 2000) and ITS4 (R) (White et al. 1990). PCR was conducted with a final reaction volume of 50 pL, containing 24.8 pL ultra-pure H2O, 5 pL lOX GOLD buffer (containing 150mM Tris buffer at pH 8.0, 500 mM KCl), 25 pL 25 mM MgCl2, 8 pL 10 mM dNTPs, 2.0 pL each of forward and reverse primer at lOmM each, 2.0 pL 100% dimethyl sulfoxide (DMSO), 2.0 pL 1:10 dilution genomic DNA template, and 0.02 pL Gold TAQ polymerase or FastStart TAQ polymerase (Takara Bio Inc., Otsu, Shiga, Japan). PCRs reaction thermo-cycle profile had a 4 min at 94°C initial denaturation step, followed by 35 cycles of 30 sec at 94°C, 45 sec at 59°C, 45 sec 59°C ramping to 72°C at 0.5° increase per second, and terminated at 72°C for 5 min extension, followed by 5 min at 4°C to snap chill. PCR amplification for ndhF. Total genomic DNA extract was diluted 1:10 in ultra-pure H2O for amplification using overlapping primer sets 1F-1318R and 972F-2110R (Olmstead and Sweere 1994; Ferguson 1998 [1999]). PCR was conducted with a final reaction volume of 1 5 pL, containing 5.72 pL ultra-pure H2O, 1.5 pL lOX Exonuclease Taq DNA polymerase Mg2^ free buffer (ExTaq) (Takara Bio Inc., Otsu, Shiga, Japan), 1.5 pL 25 mM MgCl2, 1.2 pL 2.5 mM dNTPs, LO pL each of forward and reverse primer at lOmM, 1.0 pL bovine serum albumin (BSA), 2.0 pL 1:10 dilution genomic DNA template, and 0.08 pL ExTaq. PCRs reaction thermo-cycle profile had a 2 min at 96°C initial denaturation step, followed by 35 cycles of 30 sec at 94°C, 1 min at 61°C, 1 min at 72°C, terminated at 72°C for 5 min extension, followed by 5 min at 4°C to snap chill. This amplification protocol was followed for both sets of ndh¥ primer pairs. To verify amplification of nuclear and plastid PCR products a combined total of 3 pL of template PCR product and 1 pL of 6X dye was run on a 1.6% agarose gel (APEX agarose) in IX TBE buffer at 100 volts, with a standard 100 bp (wrITS) or Ikb {ndh¥) ladder to visually size fragments. The gel was stained in an ethidium bromide bath, rinsed in deionized water, and viewed under ultra-violet light. Gel photographs were taken for reference (not shown). Sequence polymorphism in the direct sequenced PCR product was not observed for either nrlTS or ndh¥ accessions. DNA Sequencing Cycle sequencing for nr ITS. Template DNA from PCR product was purified using MO BIO 2014] WALDEN ET AL.: PHYLOGENIES AND CHROMOSOME EVOLUTION OF PHACELIA 27 UltraClean PCR Clean-Up DNA Purification Kit (MO BIO Laboratories, Inc,, Solano Beach, CA). Cycle sequencing for nrlTS was conducted with ITS4, ITS-I, and internal primers ITS2 and ITS3 (White et ah 1990; Urbatsch et al. 2000) in a final reaction volume of 12 pL, containing 6.45 pL ultra-pure H2O, 0.5 pL BigDye (Applied Biosys- tems, Inc., Foster City, CA, USA), 2.0 pL 5X buffer, 0,75 pL primer, and 2.0 pL template DNA. Reaction parameters were an initial 30 sec at 95°C denaturation step, followed by 25 cycles of 10 sec at 94°C, 30 sec at 59°C, and terminated at 60°C for 4 min. Cycle sequencing for ndhF. PCR products were cleaned of excess nucleotides (dNTPs) and primers from the amplification reaction using 1 pL ExoSAP-IT (USB Corp., Cleveland, Ohio, USA) per 5 pL template, with an initial 37°C incubation for 30 min for digestion, followed by 80°C for 10 min to inactivate the enzymes. Cycle sequencing for ndh¥ was conducted with internal primers IF, 536F, 536R, 972F, 1318F, 1318R, and 2110R in a final reaction volume of 12 pL containing 6.45 pL ultra-pure H2O, 0.5 pL BigDye (Applied Biosystems Inc., Foster City, CA, USA), 2.0 pL 5X buffer, 0.75 pL primer, and 2.0 pL template DNA. Reaction parameters were an initial 30 sec at 96°C denaturation step, followed by 25 cycles of 10 sec at 96°C, 30 sec at 60°C, and terminated at 60°C for 4 min extension. Sequencing for nr ITS and ndhF. Cycle sequenc- ing products were precipitated using EDTA/ Sodium acetate in ethanol protocol, and then resuspended in 15 pL Hi-Di Formamide (Applied Biosystems, Inc., Foster City, CA, USA). Prod- ucts were denatured for 2 min at 95°C, followed by 5 min at 4°C to snap chill. Samples were loaded into a 96 well plate and spun down at low speed (700 rpm for 1 min). Sequencing was conducted using an ABI PRISM 377 Sequencer or ABI PRISM 3100 Sequencer (Applied Biosys- tems, Inc., Foster City, CA, USA). All molecular work was conducted in the SFSU Department of Biology Conservation Genetics Laboratory (now the SFSU Department of Biology Genomics/ Transcriptomics Analysis Core [GTAC]). Data Analysis Nucleotide sequences were edited and assem- bled using Sequencher 4.8 (Gene Codes Corpo- ration, Inc., Ann Arbor, MI, USA). Base calling was straightforward for both nrlTS and ndh¥. We used the conditional combination approach to determine if partitions should be analyzed separately or could be combined (Huelsenbeck et al. 1996). This study joined 124 previously published nrlTS sequences from GenBank with 52 new sequences from Garrison (2007) for a total of 176 nrlTS sequences (nrlTS expanded), and 28 previously published ndh¥ sequences from GenBank were joined with 98 new sequences from Walden (2010) for a total of 126 ndh¥ sequences {ndh¥ expanded). All sequences gener- ated for this study were deposited in GenBank (Appendix 1), Multiple sequences for each molecular marker were aligned in ClustalW2 (Larkin et al. 2007; Goujon et al. 2010) using default parameters and indels edited manually in MacClade v.4.8 OSX (Sinauer Associates Inc., Sunderland, Massachusetts, [Maddison and Maddison 2001]). We identified nrlTS and ndh¥ sequences derived from the same voucher speci- men for a reduced subset of 61 accessions. The reduced 61 -accession nrlTS (nrlTS reduced) and ndh¥ {ndh¥ reduced) partitions were concatenat- ed in Mesquite v.2.74 for a combined partition following the total evidence approach (nrlTS + ndh¥ 61 -accession). Maximum parsimony. The incongruence length difference test (ILD, as the partition homogeneity test) was implemented in PAUP* v. 4. Ob 10 (Swofford 2002) to detect conflicting signal and assess combinability in the nrlTS + ndh¥ 61- accession concatenated subset, uninformative characters excluded, using 1000 replicates, 100 random addition sequence replicates, multrees “ 1, and MAXTREES increased by 100 to a limit of 10,000. nrlTS (expanded, reduced) and ndh¥ (expand- ed, reduced) individual partitions and the com- bined nrlTS + ndh¥ 61 -accession partition were analyzed using the maximum parsimony criterion (MP) in PAUP* v.4.0bl0. MP phylogeny recon- struction was performed using a heuristic search of 1000 random addition sequence replicates with tree-bisection-reconnection (TBR) branch swap- ping algorithm, ACCTRAN, all characters un- ordered and weighted equally, gaps treated as missing data, MAXTREES increased by 100 to a limit of 100,000, and nchuck = 100. Nonpara- metric bootstrap analyses were performed using the starting strict consensus tree obtained via stepwise addition, using a heuristic search, including 100 random addition sequence repli- cates with 1000 bootstrap replicates (Felsenstein 1985). Model selection. jModeltest version 0.1.1 with PHYML was used to test 88 models of evolution for best fit for the nuclear and plastid partitions (Posada and Crandall 1998; Guindon and Gas- cuel 2003; Posada and Buckley 2004; Posada 2008). Calculations using the Akaike information criterion (AIC) and hierarchical likelihood ratio tests (hLRTs) via jModeltest selected GTR + G (General Time Reversible model of nucleotide substitution with the gamma [F] model of rate heterogeneity) as the best-fit model of evolution for the nuclear (expanded, reduced), plastid 28 MADRONO [VoL 61 (expanded, reduced), and combined nrlTS + ndh¥ 61 -accession partitions (-InL score: nrlTS expanded 8933.5854, nrlTS reduced 5998.5051, ndh¥ expanded 10029.4265, ndh¥ reduced 7996.9681, nrlTS + ndh¥ 14590.1426). Maximum likelihood. File formats were con- verted through the CIPRES portal (Miller et al. 2010) with NCLconverter version 2.1 (Lewis and Holder 2008) to a relaxed PHYLIP format for RAxML-HPC2 version 7.4.4 analysis on XSEDE (Extreme Science and Engineering Discovery Environment). The model of evolution for the nuclear (expanded, reduced), plastid (expanded, reduced), and combined nrlTS + ndh¥ 61- accession partitions was GTRGAMMA for the entire analysis, with 5000 rapid bootstrap replicates and best tree search, and gaps and undetermined values treated as missing data (Stamatakis 2006, Stamatakis et al. 2008). The combined analysis was partitioned by gene and unlinked in analysis (Brown and Lemmon 2007). Bayesian analysis. Analyses were initiated in MrBayes 3.1.2 (Huelsenbeck and Ronquist 2001; Ronquist and Huelsenbeck 2003) using XSEDE through the CIPRES portal (Miller et al. 2010) using default priors, random starting trees, four independent runs, number of generations for the nrlTS expanded partition was 20,00,000, all other partitions the number of generations were 10.000. 000, using four chains (three hot and one cold) sampled every 1000 generations (total samples per run: nrlTS expanded 20,001, all other partitions 10,001), with the general time reversible model with gamma-distributed rate variation across sites (GTR + G). The average standard deviation of split frequencies from each run was less than 0.01 at the end of the four runs (nrlTS expanded 0.008313, nrlTS reduced 0.006756, ndh¥ expanded 0.003820, ndh¥ reduced 0.001946, nrlTS + ndh¥ 0.001142), and PSRF (Potential Scale Reduction Factor) approached one for the 95% credibility interval for the expanded partitions and equaled one for the reduced and combined partitions (Geiman and Rubin 1992). The number of trees required to reach stationarity was determined using Tracer v.1.5 (Rambaut and Drummond 2009). Conver- gence of posterior probabilities of split frequen- cies of runs was assessed in AWTY (Are We There Yet?) using the between run compare diagnostic function; graphical plots did not reject convergence (p near one for burn-in at 10%) (Wilgenbusch et al. 2004; Nylander et al. 2008). Bum-in samples (nrlTS expanded 5000; all other partitions 2500) were discarded (samples included for analysis from each run: nrlTS expanded 15.001, all other partitions 7501), and runs were combined with posterior probabilities of nodes >95% strongly supported. Maximum likelihood ancestral chromosome number evolution analysis in chromEvol version 1.3. Sampled taxa were coded with haploid chromosome counts for analysis in chromEvol version 1.3 using the ML RAxML best tree from the expanded nuclear, expanded plastid, and combined nrlTS + ndh¥ analyses (Mayrose et al. 2010; Mayrose 2012). Diploid counts were divided by two and input as haploid if these were the only published counts available for a taxon (see Table 1 for the coding schema). Taxa without published chromosome counts were coded “X” and treated as a gap in the ML analysis. Polymorphic chromosome counts or taxa circumscribed as polyploid were both coded as polymorphic (e.g., 11_22). It was necessary to assign the root to the outgroup (Euploca) manually in the control file after an initial analysis to locate the appropriate node number using FigTree 1.3.1 (Rambaut 2009), and then rerun chromEvol v.1.3 analyses (Mayrose et al. 2010). All eight models were included as there was no a priori reason to exclude one or more of them. Accession JQ250033 is a collection from Chile received as a molecular voucher from MO and is likely Phaceiia artemisioides Griseb. (identified in Appendix 1 as P. aff. artemisioides). The voucher specimen was not examined for this study. All published chromosome counts for taxa from South America are n = ll (Covas and Schnack 1947; Cave and Constance 1959). Alternative coding and analysis in chromEvol v.i.3 using a missing count (X) for this accession inferred n = 1 1 as the ancestral condition at the node and did not change the best model chosen by AIC (alternate analyses not shown in results section). Maximum likelihood whole genome duplication event analysis in GDCN. The ML RaXML best tree was imported into Mesquite v.2.74 and chromosome counts for sampled taxa were coded as continuous characters for analysis in GDCN as specified in the manual (Hallinan and Lind- berg 2011a, b). Samples without published chromosome counts at sampled taxonomic rank (coded “X” in chromEvol v.L3 analyses) were recoded as required in GDCN (see Table 1 for coding schema). Taxa lacking chromosome counts were coded with published counts of the variety or species (e.g., P. insularis Munz var. continentis J. T. Howell, n = 10 [Cave and Constance 1947] was coded for P. insularis var. insularis). Polymorphic chromosome counts cod- ed in chromEvol were recoded with a single state in GDCN to match sampled population locations in this study and corresponding voucher speci- men localities for published chromosome counts. Where sampled populations and chromosome counts overlapped in geographic distribution the more common count was used (measured by 2014] WALDEN ET AL.: PHYLOGENIES AND CHROMOSOME EVOLUTION OF PHACELIA 29 number of published vouchers). If chromosome counts differed in number the count with a cited voucher specimen was chosen. For taxa where no chromosome count was available, the inferred chromosome number from chromEvol v.L3 analyses was used for an “informed” coding (Cusimano et al, 2012). WGD events were allowed (wgd = sto), chromosome duplications (k in GDCN) and losses (}i in GDCN) were equal (muset = lam), and the chromosome number for the root was determined by ML for each analysis (rootfit = ml). chromEvol v.L3 and GDCN use slightly different ways to symbolically represent and discuss gains, losses, and doubling chromosome events within lineages. The rate of chromosome gains is symbolized by S in chromEvol v.L3 and the rate of chromosome duplication is symbolized by X in GDCN. The rate of chromosome loss is symbolized by X in chromEvol v.1.3 and by p in GDCN. The rate of demipolyploidization is symbolized by p in chromEvol v.L3 and poly- ploidization is symbolized by p in chromEvol. In GDCN the rate of whole genome duplication is symbolized by S. We refer to the textual interpretation of the symbols used by each analysis and parenthetically indicate the symbol and corresponding analysis throughout this paper. Results Phylogenetic Analyses Partition homogeneity test results for the combined nrlTS + ndh¥ 61 -accession subset rejected the null hypothesis (P = 0.001), indicat- ing the nuclear and plastid partitions were significantly heterogeneous and should not be combined. Separate analyses. Separate analyses for the expanded and reduced partitions (nuclear, plas- tid) resulted in similar tree topologies for each respective molecular marker. For this reason we show results for the expanded individual parti- tions only (nrlTS expanded 176-accession Fig. 1, ndh¥ expanded 126-accession Fig. 2). The following groups (corresponding to the infrageneric classification of Walden and Patter- son [2012]) were supported as monophyletic in both the nuclear and plastid separate analyses: Fhaceiia subg. Microgenetes, Phacelia sect. Pha- ceiia, Phacelia sect. Eutoca (R. Br.) Benth. & Hook.f., Phacelia sect. Cosmantha (Nolle ex A. de Candolle) Benth. & Hook.f., Phacelia sect. Gymnobytha, and Phacelia sect. Ramosissimae. Phacelia sect. Eutoca and Phacelia sect. Cos- mantha were sister to each other in both the nuclear and plastid separate analyses, as were Phacelia ^cct. Gymnobytha and the group corre- sponding to Phacelia subsect. Whitiaviae (Harv.) G. W. Gillett (included within Phacelia sect. Whitlavia, subsections not labeled on figures). Although Phacelia subg. Microgenetes was sup- ported as monophyletic, neither the nuclear nor the plastid separate analyses (expanded and reduced) recovered a monophyletic Phacelia sect. Euglypta or Phacelia sect. Miltitzia. Incongruence occurred between the nuclear and plastid topologies for placement of Draperia relative to Tricar dia + Hesperochiron (-h Howel- ianthus sampled in the ndh¥ expanded partition), placement of Romanzoffia and “core” Phacelia subg. Puichellae, and placement of Phacelia sect. Baretiana Walden & R. Patt. (nested within Phacelia sect. Gianduiosae in the nuclear parti- tion). Phacelia subg. Phacelia was not supported as monophyletic in the nuclear partition due to recovery of a paraphyletic Phacelia sect. Whitia- via. Phacelia subsect. Phacelia and Phacelia subsect. Humiles Walden & R, Patt. were not supported as monophyletic in the nuclear parti- tion due to placement of P. breweri A. Gray sister to P. californica Cham. Phacelia was not supported as monophyletic in the plastid parti- tion (“core” Phacelia subg. Puichellae sister to Romanzoffia and remaining sampled Phacelia). nr ITS expanded partition. The expanded nrlTS partition included 176 sequences and contained 1973 total characters. 1612 (81.70%) characters were constant and 111 (5.62%) variable charac- ters were parsimony uninformative. The total number of parsimony informative characters was 250 (12.67%). nrlTS sequences identical to each other are as follows: FJ8 14643 to FJ8 14644, AY630311 to AY630312, FJ814633 to FJ814632 to FJ8 14634, FJ8 14651 to FJ8 14652. Identical nrlTS sequences, and ndh¥ sequences derived from the same genomics, are indicated in Appendix 1 . The total number of MP trees was 4400, with the best MP tree score = 1445. The phylogram of the 0.5 majority rule consensus tree with MP bootstrap values, ML bootstrap values, and BI posterior probabilities is shown in Figure L AY630269 (cited as P. bicolor Torr. ex S. Watson in Gilbert et al. [2005]) was recovered sister to FJ8 14624 (cited as P. fremontii Torr. in Hansen et al. [2009]). The voucher specimen for FJ8 14624 is the blue-throated form of P. fremontii, corresponding to some of the charac- ters described in the protologue for P. brannanii Kellogg, especially “lobes rounded, bright blue” (Kellogg 1877). The voucher specimen for AY630269 has faded corollas. These accessions are identified as P. aff brannanii in Appendix 1 and Figure 1. The name Phacelia brannanii is currently placed in synonymy with P. fremontii and the type specimen has not been located (CAS?). The type was collected “near Fresno” and was probably collected from populations in 30 MADRONO [Vol. 61 Phocelio sert. Borefiono celia subg. Phacelia Phatelia sett. Ramosissimae Phocelio sett. Whitlavia Phatelia sett. Gymnobylha Phatelia sett. Eutoca Phatelia sect. Cosmantha Phatelia sect. Euglypta Phatelia sect. Miltitzio Phocelio seel. Eyglypto celia sett. Miltitzio Ph celia subg. Microgenetes Phocelio sect. Euglypto Phocelio sect. Pothyphyllae Phacelio sett. Whiliovia celia subg. Phacelio phatelia sect. Phocelio Phacelio subg. Pulchellae Fig. 1. Phylogram of 0.5-majority rule consensus tree for the expanded nrlTS 176-accession partition from Bayesian phylogenetic analysis (scale bar = mean number of nucleotide substitutions per site). Support values are shown above branches (or at nodes if branches too short to show full text) for Maximum Parsimony bootstrap values / Maximum Likelihood bootstrap values / Bayesian posterior probabilities for nodes supported at >95% posterior probability. An asterisk (*) indicates clade with <75% support in 0.5 majority rule MP bootstrap tree or 0,5 majority rule ML bootstrap tree. A dagger (f) indicates a clade not resolved in 0.5 majority rule MP bootstrap tree or 0.5 majority rule ML bootstrap tree. Abbreviation: P. = Phacelia. Infrageneric groups labeled (subgenera, sections) following Walden and Patterson (2012). 2014] WALDEN ET AL.: PHYLOGENIES AND CHROMOSOME EVOLUTION OF PHACELIA 31 Phacelia sect. Glandulosae Phacelia sect. Ramosissimae Phacelia sect. Whillavio Pii|icelia subg. Phacelia Phacelia sect. Gymnobytha Phacelia sect. Phacelia Phocelia sect. Baretiona Phacelia sect. Eutoca X Phacelia sect. Cosmantha Phacelia subg. Pulchellae * Phacelia sett. Euglypta J Phacelia sect. Miltitzia Phacelia sect. Euglypta X Phacelia sect. Miltitzia I Phacelia celia su Phacelia sect. Euglypta Phhcelia subg. Microgenetes I Phacelia sect. Pachyphyllae J Phacelia subg. Pulchellae Fig. 2. Phylogram of 0.5-majority rule consensus tree for the expanded ndh¥ 126-accession partition from Bayesian phylogenetic analysis (scale bar = mean number of nucleotide substitutions per site). Support values are shown above branches (or at nodes if branches too short to show full text) for Maximum Parsimony bootstrap values / Maximum Likelihood bootstrap values / Bayesian posterior probabilities for nodes supported at >95% posterior probability. An asterisk (*) indicates clade with <75% support in 0.5 majority rule MP bootstrap tree or 0.5 majority rule ML bootstrap tree. A dagger (t) indicates a clade not resolved in 0.5 majority rule MP bootstrap tree or 0.5 majority rule ML bootstrap tree. Abbreviation: P. = Phacelia. Infrageneric groups labeled (subgenera, sections) following Walden and Patterson (2012). the western San Joaquin Valley. Both voucher specimens were made in the southern Sierra Nevada Mountains in Kern County. Howell (1946) and Dempcy (1996) previously identified the Owens Valley and the Mohave Desert as areas of intergradation between P. bicolor and P. fremontii populations. Thorough understanding requires sampling throughout the distribution of 32 MADRONO [Vol. 61 P. fremontii in the San Joaquin Valley, Trans- verse Ranges, and Tehachapi Mountains to determine range, characters, and relationships. Our results indicate that, at least for P. fremontii, there is need to reconsider the need for reconsid- eration of the names currently placed in synon- ymy with P. fremontii and possible recognition of P. brannanii as a variety. Taxonomic changes must be addressed in a future paper, pending determination of type material. Phacelia scariosa is a “problematic taxon” included in these analyses (Ferguson 1998). Phacelia sect. Baretiana was still supported as sister to Phacelia sect. Glandulosae when P. scariosa was pruned from the partition (results not shown). Phacelia crenulata Torr. ex S. Watson, as currently circumscribed, was not supported as monophyletic in this analysis. A clade of eleven accessions representing two of four infraspecific varieties sampled from Arizona and California was recovered with very little nrlTS variation. Accession JX233446 P. crenu- lata var. angustifolia N. D. Atwood was recov- ered sister to JX233435 P. coerulea Greene, a result consistent with prior work (Garrison 2007). Phacelia crenulata var. angustifolia should be treated outside of Phacelia crenulata, although no name is currently available. However, there is not adequate support at this time to propose taxonomic segregation due to incongruent nucle- ar and plastid partitions, lack of corresponding plastid sequences for Phacelia crenulata var. angustifolia, and the need for adequate sampling throughout the described varieties of Phacelia crenulata. Species and varietal segregations in Phacelia have been made based on morphology and distribution. Some of these segregations are in need of revision. For example, Phacelia distans sensu lato is paraphyletic and requires revision. A group of five accessions included plants corre- sponding to four names currently placed in synonymy. A clade of two accessions (JX2333454, JX233455) included plants sampled from Arizona that correspond to the expanded circumscription of Phacelia distans var. australis Brand (1913). The type for that name is from the “Greenhorn Range” of California (UC63348!) and additional sampling is needed to determine relationships for the original and expanded circumscriptions for this taxon and relationship with P. gentryi Constance (Brand 1913; Constance 1948). ndhF expanded partition. The ndh¥ dataset of 126 sequences contained 2547 total characters. 1956 (76.79%) characters were constant and 233 (9.14%) variable characters were parsimony uninformative. The total number of parsimony informative characters was 358 (14.05%). Best MP tree length was 1033. The phylogram of the 0.5 majority rule consensus tree with MP bootstrap values, ML bootstrap values, and BI posterior probabilities is shown in Figure 2. Phacelia subg. Pulchellae is not monophyletic as currently circumscribed. Paraphyletic Phacelia subg. Pulchellae consists of a “core” group sister to Romanzoffia and remaining sampled Phacelia, and a group consisting of Phacelia cookei, P. keckii, and P. suaveolens sister to Phacelia subg. Phacelia. Phacelia cookei, P. keckii, and P. suaveolens were not sampled for nrlTS sequences (Appendix 1) and could not be examined with a separate nuclear or combined analysis. This result would argue for a refined circumscription of “core” Phacelia subg. Pulchellae excluding P. cookei, P. suaveolens, and P. keckii. These taxa should be placed in an unresolved group with low support within Phacelia subg. Phacelia and the focus of future studies. Two accessions identified as P. patuliflora A. Gray in GenBank were placed in separate clades. Accession AF047781 was supported in every analysis (MP, ML, BI) within Phacelia sect. Cosmantha and sister to AF047777 P. hirsuta Nutt., results accordant with Ferguson (1998 [1999]), while accession AF 130 179 (Olmstead et al. 2000) was supported in every analysis (MP, ML, BI) within Phacelia sect. Glandulosae. The voucher specimen of AF047781 (2089 bp) was received by loan and identification confirmed, but the AF 130 179 (2223 bp) voucher specimen was not examined for this study. Placement of AF 130 179 within Phacelia sect. Glandulosae could be attributed to branch length attraction as the two sequences differ by some 200 base pairs, misidentiflcation of source material, se- quencing error, lineage sorting, or the paucity of representative accessions from Phacelia sect. Cosmantha (see chromosome evolution results section for additional discussion). Combined analysis. Acknowledging that com- bining the nuclear and plastid partitions was statistically inappropriate (as determined by ILD test result) and that the separate analyses corroborated incongruence between topologies reported in previous studies, we include results of the combined nrlTS + ndHF 61 -accession analyses (Fig. 3) in a departure from the conditional combination approach in favor of the total evidence approach. The combined nrlTS -f- ndh¥ analyses supported a monophyletic Phacelia {Romanzoffia sister to Phacelia). Phacelia consists of a basal lineage of Phacelia subg. Pulchellae sister to Phacelia subg. Microgenetes, which is sister to Phacelia subg. Phacelia. Phacelia sect. Euglypta and Phacelia sect. Miltitzia are not monophyletic (sections not labeled in figure) within Phacelia subg. Microgenetes and require taxonomic reconsideration. Phacelia sect. Eutoca is sister to Phacelia sect. Cosmantha, and these form the basal group in Phacelia subg. Phacelia. 33 2014] WALDEN ET AL.: PHYLOGENIES AND CHROMOSOME EVOLUTION OF PHACELIA la § ^ I ^ O g ^ w O. n M O m o S - O cfl a y 4m. 4:^ O « g s g ■S -o S “ o « ■3 'O w O « d 8 S ^ CO s o + d M erf g ^ "O o u ^ d € ^ a ^ o o o ^ 'S 8 2 d 5 (rf « > • c ^ U CO "^''8 a ”5 o o 4m d o a is S) « ^ d >. g 0, p c . «i m « d s £ II Parsimony bootstrap values / Maximum Likelihood bootstrap values / Bayesian posterior probabilities for nodes supported at >95% posterior probability. An asterisk (*) indicates clade with <75% support in 0.5 majority rule MP bootstrap tree or 0.5 majority rule ML bootstrap tree. A dagger (f) indicates a clade not resolved in 0.5 majority rule MP bootstrap tree or 0.5 majority rule ML bootstrap tree. Abbreviation: P. = Phacelia. Infrageneric groups labeled (subgenera, sections) following Walden and Patterson (2012). 34 MADRONO [Vol. 61 Fig. 4. RaxML phylogram of the single, best tree [-InL = 8926.106148, tree length = 2.993179] for the expanded nrlTS 176-accession partition from maximum likelihood analysis (scale bar = mean number of nucleotide substitutions per site). Tip labels are shown with GenBank accession numbers. Maximum likelihood inferred ancestral chromosome numbers are indicated at nodes for chromEvol analysis model Ml (k, 5, p, p = 0, AIC = 343.6) / GDCN analysis. Nodes and branches with expectations above 0.5 from chromEvol analysis for gains (A. in chromEvol) are indicated with © (total number of events 13.9872), for losses (5 in chromEvol) are indicated with 0 2014] WALDEN ET AL.: PHYLOGENIES AND CHROMOSOME EVOLUTION OF PHACELIA 35 Phaceiia sect. Bareiiana is sister to the group consisting of Phaceiia sect. Phaceiia, Phaceiia sect, Gymnobytha, and (a monophyletic) Phaceiia sect. Whitlavia. Phaceiia sect. Glandulosae is sister to Phaceiia sect. Ramosissimae. Combined nrlTS + ndhF 61 -accession partition. The nrlTS + ndhV dataset of 61 sequences contained 4520 total characters. 3711 (82.10%) characters were constant and 310 (6.85%) variable characters were parsimony uninformative. The total number of parsimony informative characters was 499 (11.03%), The total number of MP trees was 36, with the best MP tree score = 1772. The phylogram of the 0.5 majority rule consensus tree with MP bootstrap values, ML bootstrap values, and BI posterior probabilities is shown in Figure 3. Atwood (1975) proposed six complexes within Phaceiia sect. Glandulosae. The separate nuclear and plastid partitions do not support these complexes and the combined 61 -accession parti- tion does not include adequate representation to address the circumscription of the complexes. Chromosome Evolution Results from chromEvol v.1.3 and GDCN analyses for the nrlTS expanded partition are shown in Figure 4, for the ndh¥ expanded partition in Figure 5, and for the nrlTS + ndh¥ 61 -accession partition in Figure 6. The best supported chromEvol v.1.3 model for the ex- panded partitions was Ml and for the combined partitions the best supported model was M2 (see figure captions for details). The results of chromEvol v.1.3 and GDCN analyses for the inferred base number and genome doubling events are dependent on sampling and input tree topology. This study is limited in abilit)^ to resolve patterns of chromosomal evolution within infrageneric groups because there is a lack of multiple accessions for every taxon sampled across the range of documented diversity. The predicted base number for the genus was x = 9, x = 11, or X = 12. The predicted base number for Phaceiia excluding “core” Phaceiia subg. Pul- chellae (and excluding Romanzoffia for the ndh¥ expanded partition) was x=llorx==12. Phaceiia subg. Pulcheiiae contains descending dysploidy (n = 10, 11, 12) and annual polyploid taxa {n = 22, 24, not sampled). The predicted base number for “core” Phaceiia subg. Pulcheiiae was x=llorx=12. There is ascending and descending dysploidy within Phaceiia subg. Mi~ crogenetes {n = 11, 12, 13) and annual polyploid taxa {n = 23), discussed previously in Constance (1963), Heckard (1963), and Gilbert et al. (2005). The predicted base number for Phaceiia subg. Microgenetes was x = 1 1 or x — 12. Phaceiia sect. Pachyphyllae (« = 11) predicted base number was X == 11. The predicted base number for the paraphyietic Phaceiia sect. Euglypta and Phaceiia sect. Miltitzia {Phaceiia subg. Microgenetes ex- cluding Phaceiia sect. Pachyphyllae) was x = 12. Cytological characters will be helpful in a future revision of Phaceiia sect. Euglypta and Phaceiia sect. Miltitzia. There is descending dysploidy within Phaceiia sect. Cosmantha {n = 5, 6, 8, 9) and annual polyploid taxa {n = 14, not sampled). Sampling throughout this group is extremely limited for the nuclear and plastid partitions. The predicted base number for Phaceiia sect. Cosmantha was x = 9 as hypothesized by Constance (1949). The predicted base number for Phaceiia sect. Eutoca, Phaceiia sect. Baretiana, Phaceiia sect. Gymnobytha, and Phaceiia sect. Whitlavia was x = 1 1. Phaceiia sect. Phaceiia includes the annual Phaceiia subsect. Humiles (not labeled in figures) with ascending and descending dysploidy {n = 7, 8, 9, 10, 1 1) and the perennial polyploid complex Phaceiia subsect. Phaceiia {n — 11, 22, 33). The predicted base number for Phaceiia sect, Phaceiia was x = 1 1 . There is descending dysploidy within Phaceiia sect. Ramosissimae {n = 10, 11), although accessions of P. suffrutescens Parry {n = 10) w^ere not included in analysis. The predicted base number for Phaceiia sect. Ramosissimae was x == 11, There is ascending dysploidy in Phaceiia sect. Glandulosae {n “ 11, 12) and a putative annual polyploid taxon (AF 130 179). Our review of published chromosome counts for sampled taxa indicates wider variation than has previously has been noted for Phaceiia sect. Glandulosae (e.g., Phaceiia congesta, Phaceiia crenulata var. crenulata) (Atwood 1975; Walden and Patterson 2012). These counts may represent infraspecific cytotypes that should be the focus of future studies of cryptic diversity in the section. The predicted base number for Phaceiia sect. Glandu- losae was X = 11. Coding of AF 130 179 P. patulifiora with the published count of « = 9 for the taxon or with X did not change the inferred ancestral state of the nearest ancestral node {n = 11) in chromEvol v.L3 analysis (alternate anal- yses not shown in results section, see phylogenetic analysis results section for discussion regarding this accession). A loss event was inferred along the branch with coding of « == 9 (expectation 1 .89974). It is possible that, due to our coding schema, we did (total number of events 21 .5156), for polyploidizatiori (p in chromEvol) are indicated with ® (total number of events 4.24671). A whole genome duplication (WGD) event supported at >95% posterior probability in GDCN analysis is indicated with IE at nodes and branches (symbol shown at tip for WGD event along branch). 36 MADRONO [Vol. 61 0.0070 Fig. 5. RaxML phylogram of the single, best tree [-InL = 10029.632275, tree length = 0.533931] for the expanded ndh¥ 126-accession partition from maximum likelihood analysis (scale bar = mean number of nucleotide substitutions per site). Tip labels are shown with GenBank accession numbers. Maximum likelihood inferred ancestral chromosome numbers are indicated at nodes for chromEvol analysis model Ml (A,, 5, p, p = 0, AIC = 285.6) / GDCN analysis. Nodes and branches with expectations above 0.5 from chromEvol analysis for gains (A in chromEvol) are indicated with © (total number of events 7.04247), for losses (5 in chromEvol) are indicated with 2014] WALDEN ET AL.: PHYLOGENIES AND CHROMOSOME EVOLUTION OF PHACELIA 37 not capture a cryptic polyploidization event for this accession like that for P. ivesiana var. pediculoides. Results from chromEvol v. L3 and GDCN infer polyploidization and whole genome dupli- cation events occurring within Phacelia, although this is dependent upon sampling, coding of the chromosome counts at tips, choice of provided tree topology, and AIC selection of the best modeh All analyses identified a polyploidization event (for chromEvol v.L3) and a whole genome duplication (WGD) event (for GDCN) for Erio- dictyon and Phacelia ivesiana var. pediculoides. The plastid expanded analyses identified five WGD events (for GDCN) in Phacelia subsect Phacelia and the combined nrlTS + ndh¥ 61- accession analyses identified one polyploidization event (for chromEvol v.L3) and three WGD events (for GDCN). Tentatively, P. humilis {n = 11) appears to be the diploid ancestor for the perennial polyploid complex and P. breweri {n = 11) is the diploid ancestor for P. californica {n = 22). Future research requires sampling through- out Phacelia sect. Phacelia (North and South American members) and consideration of Heck- ard’s hypotheses of ploidy intergradations in the complex (Fig. 7 in Heckard [I960]). Discussion Phylogenetic Relationships This project sampled throughout Phacelia using the internal transcribed spacer region (ITS-1, ITS-2, and 5.8S gene) of nuclear ribo- somal DNA (nrlTS) and the chloroplast DNA gene {ndh¥) to infer expanded phytogenies for the nuclear and plastid partitions. An objective of this study was to determine if additional sampling within the genus for each nuclear and plastid molecular marker would recover similar topolo- gies as in previous analyses. We recovered similar nrlTS topologies and support values for this study as in Gilbert et al. (2005) and as in Hansen et al. (2009) (Fig. 2 in Gilbert et al. [2005], Fig. 2 in Hansen et al. [2009]). We recovered similar ndh¥ topologies and MP support values for this study as in Ferguson (1998) (Fig. 3 in Ferguson [1998 (1999)]), and note that we show only the MP support values in Figure 2 (MP tree not shown). The nuclear partition resulted in weaker support for deeper nodes along the backbone of trees when compared with the plastid partition. Overall, MP analyses for both the nuclear and plastid partitions (expanded, reduced) resulted in weaker support for deeper nodes along the backbone of trees and these were collapsed as polytomies and reported in the figure with a dagger or with an asterisk indicating low support. Within group relationships were well support- ed for both nuclear and plastid partitions in separate analyses, except for Phacelia sect. Glandulosae and for Phacelia subsect. Phacelia. nrlTS and ndh¥ were minimally variable for resolving relationships in these sections and indicate recent adaptive radiations in these sections. There was better resolution within groups where taxa were represented by multiple accessions rather than singletons. This sort of intensive sampling was balanced by practical concerns. Phacelia sect. Glandulosae includes taxa of conservation concern, some represented by extremely limited populations and few herbarium collections, including two federally endangered taxa (P. argillacea N. D. Atwood [not sampled] and P. formosula Osterh. {ndh¥ expanded only]). The gypsophilic and gypsovag taxa {P. gypso- genia 1. M. Johnst., P. marshall-johnstonii N. D. Atwood & Pinkava, P. palmer i Torr. ex S. Watson, P. vossii N. D. Atwood) were similarly unresolved in the section (Turner 2011). It is unfortunate that this study is inadequate to provide much needed answers regarding evolu- tionary relationships for these taxa, but indicates that an explicit sampling regime is required for future systematic studies. Phylogenetic incongruence between plastid and nuclear partitions was previously noted in mo- lecular studies in Phacelia (Ferguson 1998, 1998 [1999]; Gilbert et al. 2005; Hansen et al. 2009). The nuclear and plastid partitions were signifi- cantly incongruent in our analyses. We corrobo- rate incongruence between nuclear and plastid topologies for placement of some groups (e.g., Draperia, Romanzoffia and “core” Phacelia subg. Pulchellae, Phacelia sect. Baretiana). Long- branch attraction may be a factor in the incongruent results at the base of the tree for Draperia, Romanzoffia, and Phacelia, although this remains to be exhaustively tested (Bergsten 2005). The low support for groups for the nuclear partition may partially account for incongruence between the topologies. Different sampling be- tween nuclear and plastid partitions could partially account for incongruent topologies in the expanded analyses. For example, the expand- ed nuclear topology recovered a paraphyletic Phacelia subsect. Humiles (P. breweri sister to P. californica within Phacelia subsect. Phacelia). Phacelia breweri was not sampled for ndh¥. However, incongruence between the nuclear © (total number of events 36.2782), for polyploidization (p in chromEvol) are indicated with © (total number of events 2.42607). A whole genome duplication (WGD) event supported at >95% posterior probability in GDCN analysis is indicated with lx) at nodes and branches (symbol shown at tip for WGD event along branch). 38 MADRONO [Vol. 61 i^ssSssiP o'g8-2®'9^-8 ^ jrESat _ o siillipjf |2iS|2iE5S.I Ssli e|o2^-S"' igiS|ggiS2| -“ ©-^ O © ^ O O < C N N ©^’w^-^O ■?.S2 g g ® 8 0 6«-£|s-gi- X I- D Z uj uj S d \ O d .2 S ^ !h :S .2 U Cfi S ID QJ C W O • n !> P3 (rt s-g^2| s: s m ^ oj d 0) r =1 0 d gj > + 0 II e« bO ^ ^ O " ^ QJ Q Q.r o 'o ^ 't'S >2^ ''■o'S ss c -d ^ y ^ ^ B ^ •gSs^’S^.fr i -S - -S .s % ^ o e ^ ^ 6 id- o ■£ o OJ c« > cfl 13 Cl ^ o ^ O ^ •=s||s| M H ^ ” Si^=| - .S V o S Q o Dl ^ n 'tfi c« ^ ^ ^ O ^ jfj ^ c ^ c 2 50 g " s ^ P3 c > g ^ ^ 0^ d' 75 d . QJ 0) RS 0) ._ ^ <4_^ ^ ^ *43 O O d «« ii J s w s ■; ^ S8 gl if "-S -o Si'S c , S “ 11 g^ ||i*| s ^ gs®'® ID Dl LD ^ s ^ -d ^ Cfl p O Sh > CO W -S ID ^ d „ ^ ^ c Z 1 d d Qj U C s Ji § I 3 d :r d g ^ ^ d o ^ O 13 o ,_ S II S ^ ^ ‘5 !h g o u .a -b 2 '5 ^ :2 a ^2 a - 2 ^ I g o d H S ^ 0^322^2 d ^6 X ^ bv 2 S 2 d ^ LD box -3 _n •'^ 00 0^ o a ^ ^ to ^ Ai ’b a> ^ -3 ^q ''' X ^ O d CO CO "d s g-S^- N I d Z ^ ^ ^ d , o Pn g d CO ^ 5h d H d -rf CO gj CO g’'§ ec g X g a o| td ^ 2 (5 00 2014] WALDEN ET AL.: PHYLOGENIES AND CHROMOSOME EVOLUTION OF PHACELIA 39 (Fig. 1) and the plastid (Fig, 2) topologies remained for the 61 -accession partitions in separate analyses (results not shown), indicating that sampling of representatives using sequences from different specimens was not contributing to the incongruence between partitions. Incongruent topologies between the separate nuclear partition (expanded, reduced) separate and plastid partition (expanded, reduced) indi- cated that simultaneous analyses of the combined nrlTS + ndh¥ 61 -accession partition would be statistically inappropriate but potentially benefi- cial (Nixon and Carpenter 2005). Combined analyses resulted in better resolution than sepa- rate analyses, and in a topology that favored the separate plastid topologies (expanded, reduced). Where nuclear and plastid topologies were congruent in separate analyses, the combined nrlTS + ndhF 61 -accession topology was also congruent and internal nodes were recovered with support. Our results support combining incongruent partitions in a combined analysis to seek support for internal nodes. Previous molecular phylogenetic studies in Hydrophylloideae have identified significant phy- logenetic incongruence between nuclear and plastid partitions (Ferguson 1998; Hansen et al. 2009; Taylor 2012). Studies in Cordioideae (Weeks et al. 2010), Ehretioideae (Smith 2003; Moore and Jansen 2006), and broad sampling across Boraginaceae that included “problematic” subfamilies (Ferguson 1998 [1999]; Nazaire and Hufford 2012) have also identified significant phylogenetic incongruence between nuclear and plastid partitions. Broad sampling across Boragi- naceae with limited sampling from Cordioideae, Hydrophylloideae, and Ehretioideae and exclud- ing “problematic” taxa found nuclear and plastid partitions to be congruent (Cohen 2013). Studies focusing on Boraginoideae (Winkworth et al. 2002; Hilger et al. 2004; Hasenstab-Lehman and Simpson 2012) and Heliotropioideae (Luebert and Wen 2008) have found the nuclear and plastid partitions to be congruent for those subfamilies. It is clear that a comprehensive examination of these “problematic” but interesting lineages within the larger context of Boraginaceae is needed. Partic- ularly necessary is a comparative, statistical examination of partition homogeneity, incom- plete lineage sorting, and analysis of hidden support within datasets of Boraginaceae (Huel- senbeck and Bull 1996; Gatesy et al. 1999; Leigh et al. 2008; Sarkar et al. 2008; Simon et al. 2009). Chromosome Evolution Maximum likelihood analyses were used to infer ancestral chromosome numbers and identify gains, losses, polyploid doubling, and whole genome duplication events from published chro- mosome counts assembled from the literature for sampled taxa. Results of analyses from chrom- Evol V.L3 and GDCN are dependent on sam- pling and known tree topology (Mayrose et al. 2010; Hallinan and Lindberg 2011a). Our results demonstrate the utility of using both chromEvol v.1.3 and GDCN in combination to identify patterns of chromosome evolution in Phacelia. Constance (1963), Heckard (1963), and Gillett (1968) hypothesized that n = l \ was the ancestral condition for the genus and noted it was also the most common haploid count for extant taxa. Results for the separate nuclear, separate plastid, or combined dataset did not provide a consensus. The predicted base number for the genus was x = 9, X = 11, or X — 12. Total evidence approaches may provide better-resolved phylogenies for Phacelia and Hydrophylloideae, offering resolu- tion of inferred ancestral states for chromosome evolution, especially those that are shown as equivocal in areas of the tree with lower sampling. Conclusions This study contributes to our understanding of the evolutionary relationships of Phacelia, an entirely New World genus with a center of diversity in California. We investigated chromo- some evolution in an explicit molecular context using maximum likelihood models of evolution. Phacelia is an excellent group to study chromo- some evolution, as it is a large genus representa- tive of the California flora with a range of chromosome counts (Stebbins 1942; Stebbins and Major 1965; Stebbins 1971). Our study identifies patterns of gains, losses, and polyploid doubling events in lineages that likely contributed to the overall diversity in Phacelia and the California flora. Although an extensive dataset of chromo- some counts exists for the genus and the subfamily, approximately a third of the genus (ca. 70 taxa) lack published chromosome counts. Expanded sampling for each respective marker (nuclear nrlTS and plastid ndHF) and phyloge- netic analyses (maximum parsimony, maximum likelihood, and Bayesian inference) recovered similar topologies for separate and combined analyses as previous studies. The combined nrlTS + ndhF analyses supported Romanzoffia sister to a monophyletic Phacelia. Our results support combining incongruent partitions in a combined analysis to seek support for internal nodes. Results indicated recent adaptive radiations in Phacelia sect. Glandulosae and Phacelia sect. Phacelia. Future work is needed to understand and compare the rate of molecular and morpho- logical evolution in the genus. Acknowledgments We thank Ellen Dean, Dennis Desjardin, Bruce Baldwin, our graduate labs and cohorts, Jim Linnberg, and Trigger (service dog of GKW). We thank Brent 40 MADRONO Mishler, Dave Linnberg, Kip Will, and participants in IB200A/B for substantive discussions and critical comments on previous analyses and drafts. Diane Ferguson shared data matrices from her research and granted permission to use sequences. Debra Hansen shared genomics and primers. Bob Allen, John Dempcy, Richard Halse, Holly Forbes, Jim Shevock, and Jerry Tiehm shared specimen and locality data. Leigh Johnson, Staci Markos, Richard Olmstead, Tom Rosatti, and John Strother were generous with their time and advice. We are indebted to the kindness of botanists, curators, herbaria, botanic gardens, and permitting agencies for permission and access to populations, as well as gifts of material for destructive sampling. This research represents, in part, master’s theses by LMG (2007) and GKW (2010) submitted to SFSU. Funding was provided in part by the GAANN and ARCS Foundation Fellowships, and by the SFSU Department of Biology (COSE Advisory Board Schol- arship, Kenneth and Pamela Fong Scholarship, Hensill Fellowship), and by research grants from CNPS Santa Clara Chapter and CNPS East Bay Chapter to LMG. Funding was provided in part by NSF GK12, NSF TREE, NSF GRFP, and UC Berkeley Chancellor’s Fellowships, and by research grants from the California Native Plant Society (Graduate Student Grant), CNPS Bristlecone Chapter (Mary Dedecker Botanical Grant), CNPS Orange County Chapter (Charlie O’Neill Grad- uate Student Grant), Colorado Native Plant Society (John W. Marr Research Grant), Conservation Genet- ics Laboratory at SFSU, Nevada Native Plant Society (Margaret Williams Research Grant), Southern Cali- fornia Botanists (Susan Hobbs Field Research Grant), SFSU Department of Biology, and by the UC Valentine Eastern Sierra Reserve (Graduate Student Research Grant) to GKW. We thank Federico Luebert and one anonymous reviewer for their critical reviews to improve the manuscript. We thank the editor, Matt Ritter, for thoughtful comments and patience. Literature Cited Ackerly, D. D. 2009. Evolution, origin and age of lineages in the Californian and Mediterranean floras. Journal of Biogeography 36:1221-1233. Alvarez, I. and J. F. Wendel. 2003. Ribosomal ITS sequences and plant phylogenetic inference. Mo- lecular Phylogenetics and Evolution 29:417-434. Atwood, N. D. 1975. A revision of the Phaceiia Crenulatae group (Hydrophyllaceae) for North America. The Great Basin Naturalist 35:127-190. AWTY (Are We There Yet?): A system for graphical exploration of MCMC convergence in Bayesian phylogenetic inference. Website http://ceb.csit.fsu. edu/awty (accessed 05 August 2012). Baldwin, B. G., D. H. Goldman, D. J. Keil, R. Patterson, T. J. Rosatti, and D. H. Wilken, (EDS.). 2012. The Jepson Manual: vascular plants of California, 2nd ed. University of California Press, Berkeley, CA. , M. J. Sanderson, J. M. Porter, M. F. WojciECHOwsKi, C. S. Campbell, and M. J. Donoghue. 1995. The ITS region of nuclear ribosomal DNA: a valuable source of evidence on angiosperm phyiogeny. Annals of the Missouri Botanical Garden 82:247-277. Beard, J. S., A. R. Chapman, and P. Gioia. 2000. Species richness and endemism in the Western [VoL 61 Australian flora. Journal of Biogeography 27: 1257-1268. Bergsten, j. 2005. A review of long-branch attraction. Cladistics 21:163-193. Bohs, L. and R. G. Olmstead. 1997. Phylogenetic relationships in Solanum (Solanaceae) based on ndh¥ sequences. Systematic Botany 22:5-17. Brand, A. 1913. Hydrophyllaceae. Pp. 1-210 in A. Engler (ed.), Das Pflanzenreich IV, VoL 251 (Heft 59). Verlag von Wilhelm Engelmann, Leipzig, Brown, J. M. and A. R. Lemmon. 2007. The importance of data partitioning and the utility of Bayes factors in Bayesian phylogenetics. Systematic Biology 56:643-655. Bull, J., J. P. Huelsenbeck, C. W. Cunningham, D. L. SWOFFORD, AND P. J. Waddell. 1993. Partitioning and combining data in phylogenetic analysis. Systematic Biology 42:384-397. Cave, M. S. and L. Constance. 1942. Chromosome numbers in the Hydrophyllaceae. University of California Publications in Botany 18:205-216. . 1944. Chromosome numbers in the Hydro- phyllaceae, 11. University of California Publica- tions in Botany 18:293-298. . 1947. Chromosome numbers in the Hydro- phyllaceae, III. University of California Publica- tions in Botany 18:449-465. . 1950. Chromosome numbers in the Hydro- phyllaceae, IV. University of California Publica- tions in Botany 23:363-382. . 1959. Chromosome numbers in the Hydro- phyllaceae, V. University of California Publica- tions in Botany 30:233-258. chromEvol (v.L3). 2012. Website http://www.tau.ac. il/—itaymay/cp/chromEvol/index.htnil (accessed 01 May 2012). California Native Plant Society (CNPS). 2011. Inventory of Rare and Endangered Plants (online edition, v8-01a). California Native Plant Society. Sacramento, CA. Website http://www.cnps.org/ inventory (accessed 30 June 2013). Cohen, J. 1. 2013. A phylogenetic analysis of morpho- logical and molecular characters of Boraginaceae: evolutionary relationships, taxonomy, and patterns of character evolution, Cladistics. doi: 10.1111/ cla. 12036 (accessed 12 July 2013). Constance, L. 1948. A new species of Phaceiia from Sonora, Mexico. Madrono 9:255-256. . 1949. A revision of Phaceiia subgenus Cos- manthus (Hydrophyllaceae). Contributions to the Gray Herbarium 168:1M8. . 1963. Chromosome number and classification in Hydrophyllaceae. Brittonia 15:273-285. CovAS, G. and B. Schnack. 1947. Estudios cariolo- gicos en Antofitas. II Parte. Revista Argentina de agronoma 14:224-231. Cunningham, C. W. 1997. Can three incongruence tests predict when data should be combined? Molecular Biology and Evolution 14:733-740. Cusimano, N., a. Sousa, and S. S. Renner. 2012. Maximum likelihood inference implies a high, not a low, ancestral haploid chromosome number in Araceae, with a critique of the bias introduced by A’. Annals of Botany 109:681-692. Dempcy, J. M. 1996. Phylogenetic relationships in Phaceiia section Euglypta (Hydrophyllaceae). M.S. Thesis, Biology, San Francisco State University, San Francisco, CA. 2014] WALDEN ET AL.: PHYLOGENIES AND CHROMOSOME EVOLUTION OF PHACELIA 41 Doyle, J. J. and E. E. Dickson. 1987. Preservation of plant samples for DNA restriction endonuclease analysis. Taxon 36:715-722. Farris, J. S., M. Kallersjo, A. G. Kluge, and C. Bult. 1995. Constructing a significance test for incongruence. Systematic Biology 44:570-572. Felsenstein, j. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783-791. Ferguson, D. M. 1998. Phylogenetic systematics of Hydrophyllaceae. Ph.D. Dissertation, Biology, Harvard, Cambridge, MA. Ferguson, D. 1998 (1999). Phylogenetic analysis and relationships in Hydrophyllaceae based on ndh¥ sequence data. Systematic Botany 23:253-268. FigTree V. 1.3.1 2006-2009. Website http://tree.bio.ed. ac.uk/software/figtree/. (accessed 28 May 2009). Ganong, C. K. 2002. Phylogenetic analysis of Phacelia section Miltitzia. M.S. Thesis, Biology, San Fran- cisco State University, San Francisco, CA. Garrison, L. M. 2007. Phylogenetic relationships in Phacelia (Boraginaceae) inferred from nrlTS se- quence data. M.S. Thesis, Biology, San Francisco State University, San Francisco, CA. Gatesy, j., P. O’Grady, and R. H. Baker. 1999. Corroboration among data sets in simultaneous analysis: hidden support for phylogenetic relation- ships among higher level artiodactyl taxa. Cladis- tics 15:271-313. Gelman, a. and D. B. Rubin. 1992. Inference from iterative simulation using multiple sequences. Statistical Science 1:451-511. Gilbert, C., J. M. Dempcy, C. K. Ganong, R. Patterson, and G. S. Spicer. 2005. Phylogenetic relationships within Phacelia subgenus Phacelia (Hydrophyllaceae) inferred from nuclear rDNA ITS sequence data. Systematic Botany 30:627-634. Gillett, G. W. 1968. Systematic relationships in the Cosmanthus phacelias (Hydrophyllaceae). Brittonia 20:368-374. Glass, P. M. and F. Levy. 2011. Correspondence of morphology, phylogeny and reproductive barriers in Phacelia subgenus Cosmanthus (Hydrophylla- ceae). The Journal of the Torrey Botanical Society 138:341-352. Goujon, M., H. McWilliam, W. Li, F. Valentin, S. Squizzato, j. Paern, and R. Lopez. 2010. A new bioinformatics analysis tools framework at EMBL- EBI. Nucleic Acids Research 38:W695-W699. Guindon, S. and O. Gascuel. 2003. A simple, fast, and accurate algorithm to estimate large phyloge- nies by maximum likelihood. Systematic Biology 52:696-704. Hallinan, N. M. and D. R. Lindberg. 2011a. Comparative analysis of chromosome counts infers three paleopolyploidies in the mollusca. Genome Biology and Evolution 3:1150-1163. doi: 10.1093/ gbe/evr087 (accessed 25 April 2012). . 2011b. GDCN (Genome Duplication from Chromosome Numbers). Website http://code.google. com/p/gdcn/ (accessed 25 April 2012). Hansen, D. R., G. S. Spicer, and K. Patterson. 2009. Phylogenetic relationships between and within Phacelia sections Whitlavia and Gymno- bythus (Boraginaceae). Systematic Botany 34: 737-746. Hasenstab-Lehman, K. E. and M. G. Simpson. 2012. Cat’s eyes and popcorn flowers: phylogenetic systematics of the genus Cryptantha s.l. (Boragina- ceae). Systematic Botany 37:738-757. Heckard, L. R. 1960. Taxonomic studies in the Phacelia magellanica polyploid complex, with special reference to the California members. University of California Publications in Botany 32:1-126. . 1963. The Hydrophyllaceae. The Quarterly Review of Biology 38:117-123. Hilger, H. H., F. Selvi, A. Papini, and M. Bigazzi. 2004. Molecular systematics of Boraginaceae tribe Boragineae based on ITSl and trnL sequences, with special reference to Anchusa s.l. Ann Bot 94:201-212. Howell, J. T. 1946. A revision of Phacelia section Euglypta. American Midland Naturalist 36:381- 411. Huelsenbeck, j. P. and J. J. Bull. 1996. A likelihood ratio test to detect conflicting phylogenetic signal. Systematic Biology 45:92-98. , , and C. W. Cunningham. 1996. Combining data in phylogenetic analysis. Trends in Ecology & Evolution 11:152-158. and F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogeny. Bioinformatics 17:754-755. Kellogg, A. 1877. Notes and descriptions of some Californian plants. Proceedings of the California Academy of Sciences 7:89-94. Kim, K. j. and R. K. Jansen. 1995. ndh¥ sequence evolution and the major clades in the sunflower family. Proceedings of the National Academy of Sciences 92:10379-10383. Ko VANDA, M. 1978. Chromosome numbers of miscel- laneous United States dicotyledons. Rhodora 80:431-440. Kraft, N. J. B., B. G. Baldwin, and D. D. Ackerly. 2010. Range size, taxon age and hot- spots of neoendemism in the California flora. Diversity and Distributions 16:403-413. Kruckeberg, a. R. 1956. Notes on the Phacelia magellanica complex in the Pacific Northwest. Madrono 13:209-221. Larkin, M. A., G. Blackshields, N. P. Brown, R. Chenna, P. a. McGettigan, H. McWilliam, F. Valentin, I. M. Wallace, A. Wilm, R. Lopez, J. D. Thompson, T. J. Gibson, and D. G. Higgins. 2007. Clustal W and Clustal X version 2.0. Bioinformatics 23:2947-2948. Leigh, J. W., E. Susko, M. Baumgartner, and A. J. Roger. 2008. Testing congruence in phylogenomic analysis. Systematic Biology 57:104—115. Lewis, P. O. and M. T. Holder. 2008. NEXUS Class Library (version 2.1). Website http://sourceforge. net/projects/ncl/. Luebert, F. and j. Wen. 2008. Phylogenetic analysis and evolutionary diversification of Heliotropium sect. Cochranea (Heliotropiaceae) in the Atacama Desert. Systematic Botany 33:390^02. MacClade v.4.08. Analysis of phylogeny and charac- ter evolution. Sinauer Associates, Sunderland, MA Maddison, W. and D. Maddison. 2010. Mesquite: a modular system for evolutionary analysis. Version 2.74. Website http://mesquiteproject.org (accessed 3 October 2010). Martin, M. and B. Sabater. 2010. Plastid ndh genes in plant evolution. Plant Physiology and Biochem- istry 48:636-645. 42 MADRONO [VoL 61 Mayrose, L, M. S. Barker, and S. P. Otto. 2010. Probabilistic models of chromosome number evo- lution and the inference of polyploidy. Systematic Biology 59:132-144. Miller, M. A., W. Pfeiffer, and T. Schwartz. 2010. Creating the CIPRES Science Gateway for inference of large phylogenetic trees. Paper read at the Proceedings of the Gateway Computing Environments Workshop (GCE), 14 November 2010, New Orleans, LA. Moore, M. J. and R. K. Jansen. 2006. Molecular evidence for the age, origin, and evolutionary history of the American desert plant genus Tiquilia (Boraginaceae). Molecular Phylogenetics and Evo- lution 39:668-687. Mullis, K. B., F. Faloona, S. Scharf, R. Saiki, G. Horn, and H. A. Erlich. 1987. Specific enzy- matic amplification of DNA in vitro: the polymer- ase chain reaction. Cold Spring Harbor Symposia on Quantitative Biology 51:263-273. Nazaire, M. and L. Hufford. 2012. A broad phylogenetic analysis of Boraginaceae: implications for the relationships of Mertensia. Systematic Botany 37:758-783. Neyland, R. and L. E. Urbatsch. 1996, The ndh¥ chloroplast gene detected in all vascular plant divisions. Planta 200:273-277. Nixon, K. C. and J. M. Carpenter. 2005. On simultaneous analysis. Cladistics 12(3): 221-241. Nylander, j. a. A., J. C. Wilgenbusch, D. L. Warren, and D. L. Swofford. 2008. AWTY (Are We There Yet?): a system for graphical exploration of MCMC convergence in Bayesian phylogenetics. Bioinformatics 24:581-583. Olmstead, R. G., K.-J. Kim, R. K. Jansen, and S. J. Wagstaff. 2000. The phylogeny of the Asteridae sensu lato based on chloroplast ndh¥ gene sequences. Molecular Phylogenetics and Evolution 16:96-112. and j. a, Sweere. 1994. Combining data in phylogenetic systematics: an empirical approach using three molecular data sets in the Solanaceae. Systematic Biology 43:467N-81. Patterson, R., L. M. Garrison, and D. R. Hansen. 2012. Phacelia. Pp. 485-501 in B. G. Baldwin, D. H. Goldman, D. J. Keil, R. Patterson, T. J. Rosatti, and D. H. Wilken (eds.), The Jepson Manual: vascular plants of California, 2nd ed. University of California Press, Berkeley, CA. Posada, D. 2008. jModelTest: Phylogenetic model averaging. Molecular Biology and Evolution 25:1253-1256. and T. R. Buckley. 2004. Model selection and model averaging in phylogenetics: advantages of the AIC and Bayesian approaches over likelihood ratio tests. Systematic Biology 53:793-808. and K. A. Crandall. 1998. Modeltest: testing the model of DNA substitution. Modeltest 3.7. Bioinformatics 14:817-818. Raven, P. H. and D. Axelrod. 1978. Origin and relationships of the California flora. University of California Publications in Botany 72:1-134. Ronquist, F. and j. P. Huelsenbeck. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19:1572-1574. Sarkar, L N., M. G. Egan, G. Coruzzi, E. K. Lee, and R. Desalle. 2008. Automated simultaneous analysis phylogenetics (ASAP): an enabling tool for phlyogenomics. BMC Bioinformatics 9(1): 103. Simon, J., P. A. McLenachan, and P. J. Lockhart. 2009. A statistical approach for distinguishing hybridization and incom.plete lineage sorting. The American Naturalist 174:E54-E70. doi: 10.1086/ 600082. Smith, R. A. 2003. The systematics and molecular evolution of Lennoaceae. Ph.D. Dissertation, Biology, Vanderbilt University, Nashville, TN. Stamatakis, A. 2006. RAxML-VI-HPC: Maximum likelihood-based phylogenetic analyses with thou- sands of taxa and mixed models. Bioinformatics 22:2688-2690. , P. Hoover, and J. Rougemont. 2008. A fast bootstrapping algorithm for the RAxML web- servers. Systematic Biology 57:758-771. Stebbins, G. L., Jr. 1942. Polyploid complexes in relation to ecology and the history of floras. American Naturalist 76:36^5. . 1971. Chromosomal evolution in higher plants. Edward Arnold Ltd, London. and j. Major. 1965. Endemism and speciation in the California Flora. Ecological Monographs 35:1-35. PAUP* Phylogenetic analysis using parsimony (*and other methods), version 4,0 beta 10. Sinauer Associates Inc., Sunderland, MA, Taylor, S. 2012. Molecular systematics and the origin of gypsophily in Nama L. (Boraginaceae). Ph.D. Dissertation, Plant Biology, University of Texas at Austin, Austin, TX. Tracer v. 1.5.0. 2009. Website http://tree.bio.ed.ac.uk/ software/tracer/, (accessed 10 March 2010). Turner, B. L. 2011. A new gypsophilic Phacelia (Hydrophyllaceae) from Coahuila, Mexico. Phyto- logia 93(1): 88-93. U.S. Department of the Interior, Fish and Wildlife Service. 1978. Endangered and threat- ened wildlife and plants; determination of five plants as endangered species, final rule. Federal Register 43:448 10M48 12. . 1982. Endangered and threatened wildlife and plants; determination that Phacelia formosula is an endangered species, final rule. Federal Register 59:38540-38543. . 1997. Endangered and threatened wildlife and plants; final rule for 13 plant taxa from the northern Channel Islands, California. Federal Register 62:40954-40974, . 2004. Endangered and threatened wildlife and plants; review of species that are candidates or proposed for listing as endangered or threatened; annual notice of findings on resubmitted petitions; annual description of progress on listing actions; notices of reviev/; proposed rule. Federal Register 69:24876-24904. . 2011. Endangered and threatened wildlife and plants; determination of endangered status for Ipomopsis polyantha (Pagosa Skyrocket) and threatened status for Penstemon debilis (Parachute Beardtongue) and Phacelia submutica (DeBeque Phacelia); final rule. Federal Register 76:45054— 45075. Urbatsch, L. E., B. G. Baldwin, and M. J. Donoghue. 2000. Phylogeny of the coneflowers and relatives (Heliantheae: Asteraceae) based on nuclear rDNA internal transcribed spacer (ITS) sequences and chlorplast DNA restriction site data. Systematic Botany 25:539-565. 2014] WALDEN ET AL.: PHYLOGENIES AND CHROMOSOME EVOLUTION OF PHACELIA 43 Walden, G. K. 2010. Phylogeny of infrageneric relationships within Phacelia (Boraginaceae) in- ferred from chloroplast sequence data. M.S. Thesis, Biology, San Francisco State University, San Francisco, CA. AND R. Patterson. 2012. Nomenclature of subdivisions in Phacelia (Boraginaceae). Madrono 59:211-222. Weeks, A., K. E. Baird, and C. K. McMullen. 2010. Origin and evolution of endemic Galapagos Varronia species (Cordiaceae). Molecular Phyloge- netics and Evolution 57:948-954. White, T. J., T. Bruns, and J. Taylor. 1990. Amplification and direct sequencing of fungal ribo- somal RNA genes for phylogenetics. Pp. 315-322 in M. A. Innis, D. H. Gelfand, J. J. Sinisky, and T. J. White (eds.), PCR protocols: a guide to methods and applications. Academic Press, Inc., New York. WiNKWORTH, R. C., J. Grau, A. W. Robertson, and P. J. Lockhart. 2002. The origins and evolution of the genus Myosotis L. (Boraginaceae). Molecu- lar Phylogenetics and Evolution 24:180-193. Wood, T. E., N. Takebayashi, M. S. Barker, I. Mayrose, P. B. Greenspoon, and L. H. Riese- BERG. 2009. The frequency of polyploid speciation in vascular plants. Proceedings of the National Academy of Sciences 106:13875-13879. Appendix 1 List of taxa sampled in this study: named taxon (presented in alphabetical order, with bolded type and botanical authority given for first instance of taxon), locality of collection, date of collection or n.d., name of collector(s) and collection number, acronym of herbar- ium where voucher specimen is deposited and herbar- ium accession number (if available), GenBank accession numbers for nrlTS and ndh¥ (en dash [-] if information or sequence not available). Taxa names follow recent treatments in the second edition of The Jepson Manual (Baldwin et al. 2012) and treatments in preparation for FNANM, botanical authorities follow Authors of Plant Names edited by R. K. Brummitt and C. E. Powell (1992), and herbarium acronyms follow Index Herbar- iorum (http://sweetgum.nybg.org/ih/). Vouchers or se- quences originally published under different identifiers are indicated with an asterisk (*) and discussed parenthetically following the information string. Voucher specimens not examined for this study are indicated (n.v.). Eriodictyon californicum (Hook. & Arn.) Torr., USA, CA, Contra Costa Co., 27 Apr 1994, D. M. Ferguson 59 (GH00402724, n.v.), AF091159, AF047820; Euploca campestris (Griseb.) Diane & Hilger, -, n.d., Nee & Wen 53873 (US, n.v.), EF688856, EF688908; Draperia systyla (A. Gray) Torr., USA, CA, Tulare Co., 4-5 Sep 1979, R. Thorne 53719 (RSA341263, n.v.), AF091155, AF047770; HesperocMron pumilus (Griseb.) Porter, -, n.d., R. Olmstead and D. M. Ferguson 92 (GH, n.v.), AF091167, AF047783; Howellanthus dale- sianus (J. T. Howell) Walden & R. Patt., USA, CA, Siskiyou Co., 21 Jun 2005, R. Patterson and S. Santos 1982 (SFSU), -, JQ249933; Nama demissum A. Gray, USA, CA, San Bernardino Co., 12 April 1987, M. O. Bagiey and P. Athey 1932 (RSA395999, n.v.), AF091174, AF047767; Phacelia adenophora J. T. Howell, USA, CA, Lassen Co., 16 May 1993, J. Dempcy 114-2* (SFSU), AY630260, - (*cited as J. Dempcy 114 in Gilbert et al. 2005); P. adenophora, USA, NV, border of Storey Co. and Lyon Co., 1 1 May 1993, /. Dempcy 117-1* (SFSU), AY630261, - (*cited as J. Dempcy 117 m Gilbert et al. 2005, and a duplicate collection of J. Dempcy 117-2); P. adenophora, USA, NV, border of Storey Co. and Lyon Co., 1 1 May 1993, /. Dempcy 117-2* (SFSU), AY630262, - (*cited as J. Dempcy 117 in Gilbert et al. 2005, and a duplicate collection of J. Dempcy 117-1); P. adenophora, USA, NV, Washoe Co., 14 Jun 2008, A. Tiehm 15643 (SFSU), -, JQ249940; Phacelia affmis A. Gray, USA, NV, Nye Co., 10 Jun 1995, /. Dempcy 137-1 (SFSU), -, JQ249941; P. affinis, USA, NV, Nye Co., 10 Jun 1995, /. Dempcy 137-2 (SFSU), AY630625, -; P. affinis, USA, NV, Nye Co., 10 Jun 1995, J. Dempcy 138-1 (SFSU), AY630264, -; P. affinis, USA, AZ, Santa Cruz Co., 21 Mar 2005, L. M. Garrison 31 (SFSU), JX233424, JQ249942; P. affinis, USA, AZ, Mohave Co., F. McClintock 52-264 (CAS, n.v.), AY630263, -; Phacelia argentea A. Nelson & J. F. Macbr., USA, OR, n.d., D. M. Ferguson 82 (WTU, n.v.), AF091185, AF047810; P. argentea, USA, OR, Curry Co., wild collected for cultivation 26 Aug 1984, V. Stansell s.n., sourced from the Berry Botanic Garden, Portland, OR, 11 July 1986 (Berry BG SB84-103), living collection at UC Berkeley Botanic Garden (CA, Alameda Co.), UCBG collection for molecular research 24 Nov 2008, H. Forbes s.n. (UCBG 86.1064), -, JQ249944; P. argentea, USA, OR, Coos Co., 24 Jul 2008, G. K. Walden 81 (SFSU), FJ8 14625, JQ249943; Phacelia arizonica A. Gray, USA, AZ, Santa Cruz Co., 21 Mar 2005, L. M. Garrison 32 (SFSU), JX233426, -; Phacelia aff. artemisioides Griseb., CHILE, Antofagasta, El Loa Province, 1 1 Apr 1994, C M. Taylor and A. Pool 11569 (M0002977, n.v.), -, JQ250033; Phacelia bicolov Torr. ex S. Watson, USA, CA, Mono Co., 16 May 1993, J. Dempcy 89-2 (SFSU), AY630266, -; P. bicolor, USA, CA, Lassen Co., 16 May 1993, J. Dempcy 112-1 (SFSU), AY630267, -; P. bicolor, USA, NV, Humboldt Co., 13 May 1993, /. Dempcy 118-2 (SFSU), AY630268, P. bicolor, USA, CA, Mono Co., 18 Jun 1993, D. M. Ferguson 12 (GH), AF091186, AF047819; P. bicolor, USA, NV, Ormsby Co., 19 May 2008, A. Tiehm 15523 (SFSU), -, JQ249945; Phacelia bolanderi A, Gray, -, n.d., R. Olmstead 93-65 (WTU, n.v.), AF091 187, AF047762; P. bolanderi, USA, CA, Humboldt Co., near Weott, wild collected for cultivation s.d., W. Roderick s.n. , living collection at UC Berkeley Botanic Garden (CA, Alameda Co.), UCBG collection for molecular research 24 Nov 2008, H. Forbes s.n. (UCBG 61.0093), FJ814622, JQ249946; Phacelia bombycina Wooton & Standi., USA, AZ, Graham Co., 28 Mar 2004, L. M. Garrison 15 (SFSU), JX233427, -; P. bombycina, USA, AZ, Pima Co., 28 Mar 2004, L. M. Garrison 16 (SFSU), JX233434, -; P. bombycina, USA, AZ, Pima Co., 28 Mar 2004, L. M. Garrison 17 (SFSU), JX233428, P. bombycina, USA, AZ, Pima Co., 22 Mar 2005, L. M. Garrison 35 (SFSU), JX233429, -; P. bombycina, USA, AZ, Cochise Co., 25 Mar 2005, L. M. Garrison 43 (SFSU), JX233430, JQ249947; Phacelia brachyloba (Benth.) A. Gray, USA, CA, Santa Barbara Co., 19 Jun 1965, J. Ammirati 315 (SFSU08229), AY630271, -; P. brachyloba, USA, CA, Santa Barbara Co., 12 May 1994, J. Dempcy 123-1 (SFSU), AY630272, -; P. brachyloba, USA, CA, Santa Barbara Co., 12 May 1994, J. Dempcy 124-1 (SFSU), AY630273, -; P. brachyloba, USA, CA, San Diego Co., 28 May 2008, J. P. Rebman and M. Mulligan 15294 44 MADRONO [VoL 61 (SD 186946, n.v.), - JQ249949; R brachyloba, USA, CA, Orange Co., 26 Jun 2008, G. K Walden 68 (SFSU), JQ249948; Phacelia aff. brannamii Kellogg*, USA, CA, Kern Co., 17 May 1981, J. Shevock 8564 (CAS7 13422), AY630269, - (*cited as R bicolor in Gilbert et al. 2005); P. aff. brannanii*, Kem Co., CA, 26 Mar 2008, G. K. Walden 33 (SFSUX FJ814624, JQ249967 (*cited as P. fremontii Torr. in Hansen et al. 2009); Phacelia breweri A. Gray, USA, CA, Stanislaus Co., 23 Apr 1993, C. Condos 23 (SFSU), AY630274, -; Phacelia calif ornica Cham., USA, CA, San Francisco Co., 8 Apr 1991, M. Ely 40 (SFSU), AY630275, -; P. californica, USA, CA, Marin Co., wild collected for cultivation 12 Jun 1960, W. Roderick s.n., living collection at UC Berkeley Botanic Garden (CA, Alameda Co.), UCBG collection for molecular research 24 Nov 2008, H. Forbes s.n. (UCB60.0723), FJ814626, JQ249950; Phacelia calthifolia Brand, USA, CA, Inyo Co., 29 Mar 1970, D. Breedlove 17289A (RSA217251, n.v.), AY630276, -; P. calthifolia, USA, CA, Inyo Co., 30 Mar 1995, J. Dempcy 128C (SFSU), AY630278, -; P. calthifolia, USA, CA, Inyo Co., 8 Mar 1973, J. Thorne 42503 (RSA239876, n.v.), AY630277, Phace- lia campanularia A. Gray var. campamularia, USA, CA, San Bernardino Co., 29 Mar 2004, D. R. Hansen 16 (SFSU), FJ8 14643*, JQ249951 (*sequence identical to FJ8 14644); P. campanularia var. campanularia, USA, CA, San Bernardino Co., 22 Mar 2004, D. R. Hansen 22 (SFSU), FJ8 14644*, JQ249952 (*sequence identical to FJ8 14643); P. campanularia var. campanularia, USA, CA, San Diego Co., 2005, D. R. Hansen 45 (SFSU), FJ8 14640, -; P. campanularia A. Gray var. campanu- laria, USA, CA, San Bernardino Co., n.d., C. R. Richards 45 (SFSU), FJ8 14642, -; Phacelia campanu- laria A. Gray var. vasiformis (G. W. Gillett) Walden & R. Patt., USA, CA, Los Angeles Co., in cultivation at RSABG, voucher harvested 19 Apr 1994, D. M. Ferguson 56 (GH), AF091188, AF047786; P. campanu- laria A. Gray var. vasiformis, USA, CA, Riverside Co., 2005, D. R. Hansen 47 (SFSU^ FJ8 14646, -; P. campanularia A. Gray var. vasiformis, USA, CA, Riverside Co., 2005, D. R. Hansen 49 (SFSU), FJ8 14647, -; P. campanularia A. Gray var. vasiformis, USA, CA, San Bernardino Co., 25 Mar 2008, G. K. Walden 20 (SFSU), JQ249953; Phacelia capitata Kruckeb,, USA, OR, Douglas Co., 25 May 2008, R. R. Halse 7466 (SFSU), JQ249954; Phacelia cephalotes A. Gray, USA, UT, Washington Co., 11 May 2004, L. C. Higgins 25490 (DAVl 72082), -, JQ249935; Phacelia cicutaria Greene var. cicutaria, USA, CA, Kem Co., 22 Mar 2004, L. M. Garrison 02 (SFSU), JX233431, JQ249955; P. cicutaria var. cicutaria, USA, CA, Kem Co., 6 May 2005, L. M. Garrison 45 (SFSU), JX233432, -; Phacelia cicutaria Greene var. Mspida (A. Gray) J. T. Howell, USA, CA, Los Angeles Co., 6 May 2004, L. M Garrison 24 (SFSU), JX233433, -; P. cicutaria Greene var. hispida, USA, CA, San Diego Co., 3 Apr 2008, /. P. Rebman and M. Wall 14303 (SD 184369, n.v.), -, JQ249936; Phacelia coerulea Greene, USA, AZ, Co- chise Co., 21 Mar 2005, L. M. Garrison 30 (SFSU), JX233435, JQ249956; Phacelia congesta Hook., USA, TX, Starr Co., 23 Feb 1992, R. R. Halse 4436 (GH00288821), AF091189, AF047780; Phacelia cookei Constance & Heckard, USA, CA, Siskiyou Co., 29 May 2008, K. Schierenbeck s.n. (SFSU), - JQ249957; Phacelia corrugata A. Nelson, USA, AZ, Coconino Co., 10 May 2005, L. M. Garrison 53 (SFSU), JX233437, -; P. corrugata, USA, UT, Carbon Co., 12 May 2005, L. M. Garrison 62 (SFSU), JX233438, -; Phacelia cottamii'M. D. Atwood, USA, UT, Emery Co., 16 Jun 1999, TV. D. Atwood 24858 (DAY 1662 15), JQ249958; Phacelia crenulata Torr. ex S. Watson var. ambigua (M. E. Jones) J. F. Macbr., USA, CA, Imperial Co., 26 Mar 2004, L. M. Garrison 10 (SFSU), JX233440, P. crenulata var. ambigua, USA, CA, Imperial Co., 27 Mar 2004, L. M. Garrison 11 (SFSU), JX233441, -; P. crenulata var. ambigua, USA, AZ, Maricopa Co., 27 Mar 2004, L. M. Garrison 12 (SFSU), JX233442, -; P. crenulata var. ambigua, USA, AZ, Pima Co., 28 Mar 2004, L. M. Garrison 18 (SFSU), JX233443, -; P. crenulata var. ambigua, USA, CA, Riverside Co., 29 Mar 2004, L. M. Garrison 19 (SFSU), JX233444, P. crenulata var. ambigua, USA, CA, San Bernardino Co., 12 Apr 1992, K. Whitney 28 (SFSU), AY630279, Phacelia crenulata Torr. ex S. Watson var. angmtifolia N. D. Atwood, USA, AZ, Coconino Co., 10 May 2005, L. M. Garrison 55 (SFSU), JX233446, -; Phacelia crenulata Torr, ex S. Watson var. crenulata, USA, CA, Inyo Co., 23 Mar 2004, L. M. Garrison 05 (SFSU), JX233447, -; P. crenulata var. crenulata, USA, CA, Inyo Co., 24 Mar 2004, L. M. Garrison 06 (SFSU), JX233448, -; P. crenulata var. crenulata, USA, AZ, Pinal Co., 6 May 2005, L. M. Garrison 44 (SFSU), FJ8 14630, JQ249959; P. crenulata var. crenulata, USA, CA, San Bernardino Co., 6 May 2005, L. M. Garrison 46 (SFSU), JX233450, -; P. crenulata var. crenulata, USA, CA, InyO' Co., 8 May 2005, L. M. Garrison 48 (SFSU), JX233451, Phacelia cryptantha Greene, USA, CA, Riverside Co., 29 Mar 2004, L. M. Garrison 20 (SFSU), JX233453, -; Phacelia cumingii (Benth.) A. Gray, CHILE, L. Constance 3502 (CAS, n.v.), AY630282, -; P. cumingii, CHILE, Werde- man 1042 (CAS, n.v.), AY630283, -; Phacelia curvipes Torr. ex S. Watson, USA, CA, Mono Co., 10 Jun 2008, G K. Walden 55 (SFSU), JQ249961; Phacelia davMsonii A. Gray, USA, CA, Orange Co., 26 Jun 2008, G K. Walden 72 (SFSU), JQ249962; Phacelia demissa A. Gray, USA, NM, San Doval Co., 1 Jun 2004, N. D. Atwood and A. Clifford 30342 (DAVl 75809), JQ249963; Phacelia distans Benth. sensu lato*, USA, CA, Kem Co., 22 Mar 2004, L. M. Garrison 03 (SFSU), FJ8 14654, - (*cited as P. tanacetifolia in Hansen et al. 2009); P. distans s.L, USA, AZ, Pima Co., L. M. Garrison 14 (SFSU), JX233454, P. distans s.L, USA, CA, San Bernardino Co., 29 Mar 2004, L. M. Garrison 21 (SFSU), JX233478, JQ250026; P. distans s.L, USA, CA, Los Angeles Co., 6 May 2004, L. M. Garrison 25 (SFSU), JX233479, -; P. distans s.l., USA, AZ, Pima Co., 22 Mar 2005, L. M. Garrison 34 (SFSU), JX233455, P. distans s.l.*, USA, CA, San Diego Co., n.d., R. Peters 01 (SFSU), AY630280, - (*cited as P. crenulata var. minutifoUa in Gilbert et al. 2005); P. distans s.L„ USA, CA; Marin Co., 12 May 1991, P. Wharton 24 (SFSU), AY630284, -; Phacelia divaricata A. Gray, USA, CA, San Mateo Co., 23 Apr 1997, M. A. Hewlett 581 mah (SFSU3414), AY630285, P. divaricata, USA, CA, Colusa Co., 30 May 2008, G K. Walden 35 (SFSU), JQ249964; Phacelia douglasii (Benth.) Torr., USA, CA, Monterey Co., April 1969, L. S. Rose 69018 (SFSU), AY630286, -; Phacelia duMa (L.) Trel. & Small, n.d., direct submission to GenBank, P. Glass s.n. (TENN, n.v.), JN619425, -; Phacelia egena (Brand) J, T. Howell, USA, CA, Tehama Co., 30 Mar 1980, M. A. Showers 1679 (SFSU08308), AY630287, -; Phacettaeisemii Bran- degee, USA, CA, Sierra Co., 11 Jun 2008, D. Kruse- 2014] WALDEN ET AL.: PHYLOGENIES AND CHROMOSOME EVOLUTION OF PHACELIA 45 Pickier 21 (SFSU),-, JQ249965; Phacelia formosula Osterh., USA, CO, Jackson Co., 11 Aug 1973, N. D. Atwood and L. C. Higgins 5830 (DAV67851), JQ249966; Phacelia franklinii (R. Br.) A. Gray, CANADA, Saskatchewan, Saskatoon, 11 Jul 1973, V. L. Harms 20025 (GH00288820), -, AF047772; Phacelia fremontii Torn, USA, NV, Nye Co., 15 Apr 1992, J. Dempcy 101* (SFSU), AY630289, - (*cited as /. Dempcy 101-1 in Gilbert et al. 2005); P. fremontii^ USA, CA, San Bernardino Co., 17 Apr 1992, J. Dempcy 102* (SFSU), AY630288, - (*cited as J. Dempcy 102-2 in Gilbert et al. 2005); P. fremontii Torr., USA, CA, Inyo Co., 9 May 1992, J. Dempcy 105* (SFSU), AY630290, - (*cited as J. Dempcy 105-13 in Gilbert et al. 2005); Phacelia glabernma (Torr.) J. T. Howell, USA, NV, Pershing Co., n.d., A. Tiehm 11666 (OSC, n.v.), AY630291, -; Phacelia glandulifera Piper, USA, OR, Lake Co., 15 Jun 1993, /. Dempcy 119-2 (SFSU), -, JQ249968; P. glandulifera, USA, OR, 15 Jun 1993, J. Dempcy 119-3 (SFSU), AY630293, -; P. glandulifera, USA, OR, Harney Co., 16 Jun 1993, /. Dempcy 120* (SFSU), AY630294, - (*cited as J. Dempcy 120-2 in Gilbert et al. 2005); P. glandulifera, USA, OR, Harney Co., n.d., A. Tiehm 11063 (CAS, n.v.), AY630292, -; Phacelia glandulosa Nutt., USA, CO, Mineral Co., 1984, Weber and Randolph 17444 (TEX, n.v.), JX233456, -; Phacelia glechomifolia A. Gray, USA, AZ, Mohave Co., 12 May 2004, N. D. Atwood 30196 (D AVI 74936), -, JQ249969; Phacelia grandiflora (Benth.) A. Gray, USA, CA, Orange Co., 2008, B. Allen s.n. (SFSU), JQ249970; P. grandiflora, USA, CA, Los Angeles Co., 23 Jun 1993, D. M. Ferguson 26 (GH), AF091190, AF047818; P. grandiflo- ra, USA, CA, San Diego Co., 7 May 2004, D. R. Hansen 23 (SFSU), FJ8 14648, JQ249971; P. grandiflo- ra, USA, CA, Los Angeles Co., 8 May 2004, D. R. Hansen 29 (SFSU), FJ8 14649, JQ249972; Phacelia greenei J. T. Howell, USA, CA, Siskiyou Co., 21 Jun 2005, R. Patterson and S. Santos 1981 (SFSU), -, JQ249973; Phacelia grisea A. Gray, USA, CA, Mon- terey Co., 3 Jun 2009, G. K. Walden 151 (SFSU), -, JQ249974; Phacelia gymnoclada Torr. ex S. Watson, USA, NV, Washoe Co., 18 May 1993, J. Dempcy 115-1 (SFSU), AY630295, -; P. gymnoclada, USA, NV, Nye Co., 9 Jun 1995, J. Dempcy 136-1 (SFSU), AY630297, -; P. gymnoclada, USA, NV, Esmeraldo Co., n.d., Holmgren 11352 (CAS, n.v.), AY630296, -; P. gymno- clada, USA, OR, Malheur Co., 26 May 1995, P. F. Zika 12351 (GH, n.v.), AF091191, AF047793; Phacelia gypsogenia 1. M. Johnst., MEXICO, Nuevo Leon, 1992, Aramberri 00248757 (TEX, n.v.), JX233457, -; Phacelia hastata Douglas ex Lehm. var. compacta (Brand) Cronquist, USA, NV, Humboldt Co., 5 Jun 2008, A. Tiehm 15577 (SFSU), -, JQ249975; Phacelia heterophylla Pursh., USA, OR, Harney Co., 10 Jun 1996, D. M Ferguson 126 (GH), AF091 192, AF047805; Phacelia heterophylla Pursh var. virgata (Greene) R. D. Dorn, USA, CA, Modoc Co., 12 Jun 2008, F. Gauna s.n. (SFSU), -, JQ249976; P. heterophylla var. virgata, USA, OR, 2008, R. R. Halse 7464 (SFSU), -, JQ249978; P. heterophylla var. virgata, USA, NV, Douglas Co., 29 May 2008, A. Tiehm 15548 (SFSU), JQ249977; Phacelia hirsuta Nutt., USA, AR, Montgomery Co., 10 May 1991, D. F. Boufford, V. M. Bates and F. W. Wood 25539 (A), AF091193, AF047777; Phacelia howelliana N. D. Atwood, USA, UT, Grand Co., 11 May 2005, F. M. Garrison 59 (SFSU), JX233458, JQ249979; Phacelia hubbyi (J. F. Macbr.) L. M. Garrison, USA, CA, Ventura Co., 30 Mar 2004, L. M. Garrison 23 (SFSU), JQ249980; Phacelia humilis Torr. & A. Gray, USA, CA, Mono Co., Jun 1993, D. M. Ferguson 21 (GH), AF091194, AF047817; P. humilis, USA, CA, Sierra Co., 10 Jun 2008, D. Kruse-Pickler 20 (SFSU), -, JQ249981; P. humilis, USA, CA, Sierra Co., 4 Jul 1982, R. Patterson 1795 (SFSU 13836), AY630302, -; Phacelia hydrophyl- loides Torr. ex A. Gray, USA, CA, Sierra Co., Jun 2008, P. Hankamp s.n. (SFSU), -, JQ249982; P. hydrophylloides, USA, CA, Sierra Co., J. Shevock 5140 (CAS, n.v.), AY630304, -; Phacelia imbricata Greene var. imbricata, USA, CA, Lake Co., 9 Jun 1974, D. Toren 3582 (SFSU08372), AY630305, -; Phacelia infundibuliformis Torr., USA, TX, Presidio Co., 24 Mar 2005, L. M. Garrison 40 (SFSU), JX233436, -; Phacelia insularis Munz var. insularis, USA, CA, Santa Barbara Co., 2005, D. R. Hansen 55 (SFSU), FJ8 14627, JQ249937; Phacelia integrifolia Torr. var. integrifolia, USA, NM, Eddy Co., 24 Mar 2005, L. M. Garrison 37 (SFSU), JX233460, -; P. integrifolia var. integrifolia, USA, TX, Culberson Co., 24 Mar 2005, L. M. Garrison 38 (SFSU), JX233461, -; P. integrifolia var. integrifolia, USA, NM, Dona Ana Co., 25 Mar 2005, L. M. Garrison 42 (SFSU), JX233462, P. integrifolia var. integrifolia*, USA, TX, Mason Co., 2004, J. McDill 2004-20 (SFSU), JX233459, - (*cited as P. robusta in Garrison 2007); Phacelia integrifolia Torr. var. texana (J. W. Voss) N. D. Atwood, USA, TX, Reeves Co., n.d., - (TEX-00300216, n.v.), JX233463, -; Phacelia inyoensis (J. F. Macbr.) J. T. Howell, USA, CA, Inyo Co., 10 Apr 1995, M. De Decker 6444 (RSA627326, n.v.), AY630307, -; Phacelia inundata J. T. Howell, USA, CA, Lassen Co., 13 May 1993, B. Bartholomew, M. Gilbert, and L. Skog 6559 (CAS889055, n.v.), AY630306, -; P. inundata, USA, NV, Washoe Co., 2008, A. Tiehm 15553 (SFSU), -, JQ249983; Phacelia ivesiana Torr. var. pediculoides J. T. Howell, USA, CA, San Bernardino Co., 10 May 1978, B. A. Prigge and R. F. Thorne 2952* (RSA278650, n.v.), AY630308, - (*cited as Priggle 2952in Gilbert et al. 2005); P. ivesiana var. pediculoides, USA, CA, San Bernardino Co., Mar 2008, G. K. Walden 22 (SFSU), FJ8 14623, JQ249984; Phacelia keckii Munz & 1. M. Johnst., USA, CA, Orange Co., 26 Jun 2008, G. K. Walden 71 (SFSU), -, JQ249985; Phacelia laxiflora J. T. Howell, USA, AZ, Coconino Co., 23 May 1950, J. T. Howell 26440 (D AVI 9059), -, JQ249986; Phacelia leibergii Brand, USA, OR, Deschutes Co., 6 Jul 1991, J. Dempcy 100-2 (SFSU), AY630311*, - (*sequence identical to AY630312); P. leibergii, USA, OR, Deschutes Co., 16 Jun 1993, J. Dempcy 121* (SFSU), AY630312*, JQ249987 (*cited as J. Dempcy 121-2 in Gilbert et al 2005, *sequence identical to AY630311); Phacelia linearis (Pursh) Holz., USA, OR, Lake Co., 9 Jun 1996, D. M. Ferguson 123 (WTU, n.v.), AF091195, AF047806; P linearis, USA, CA, Modoc Co., 12 Jun 2008, F. Gauna s.n. (SFSU), FJ814629, JQ249988; P linearis, USA, CA, Siskiyou Co., 21 Jun 2005, R. Patterson and S. Santos 1978 (SFSU), -, JQ249989; Phacelia longipes Torr. ex A. Gray, USA, CA, Santa Barbara Co., 2005, D. Hansen 57 (SFSU), FJ8 14645, -; P longipes, USA, CA, Santa Barbara Co., 2005, D. Hansen 59 (SFSU), FJ8 14641, -; Phacelia lutea (Hook. & Arn.) J. T. Howell, USA, OR, Lake Co., n.d., D. M. Ferguson 122 (WTU, n.v.), AF091196, AF047807; Phacelia lutea (Hook. & Arn.) J. T. Howell var. lutea, USA, NV, Washoe Co., n.d., A. Tiehm 10617 (CAS, 46 MADRONO [Vol. 61 n.v.), AY630314, Phacelia lutea (Hook. & Am.) J. T. Howell var. calva Cronquist, USA, NV, Humboldt Co., n.d., A. Tiehm 12085 (CAS, n.v.), AY630313, Phacelia mammarillensis N. D. Atwood, USA, UT, Kane Co., 1 1 May 2005, L. M. Garrison 63 (SFSU), JX233464, Phacelia marshal-johnstonii N. D. Atwood & D. J. Pinkava, MEXICO, 1973, - (LL00248820, n.v.), JX233465, Phacelia minor (Harvey) Theil. ex F. Zimm., USA, CA, Los Angeles Co., 22 Jun 1993, D. M. Ferguson 24 (GH), AF091197, AF047802; P. minor, USA, CA, San Bernardino Co., 2004, D. R. Hansen 14 (SFSU), FJ8 14633*, JQ249990 (*identical sequence to FJ8 14632, FJ8 14634); P. minor, USA, CA, Los Angeles Co., im, D. R. Hansen 15 (SFSU), FJ8 14632*, JQ249991 (*identical sequence to FJ8 14633, FJ8 14634); Phacelia minor X Phacelia parryi (putative Fi hybrid, maternal and paternal identity unknown), USA, CA: Los Angeles Co., 2005, D. R. Hansen 30 (SFSU), FJ8 14634*, JQ249992 (*identical sequence to FJ814632, FJ814633); P. minor X P. parryi (putative Fi hybrid, maternal and paternal identity unknown), USA, CA: San Diego Co., 17 Apr 2003, J. P. Rebman and J. Gregory 8526 (SDl 59431, n.v.), FJ8 14631, -; Phacelia monoensis Halse, USA, NV, Lyon Co., 19 May 1993, J. Dempcy 116~H (SFSU), AY630315, - (duplicate collec- tion of Dempcy 116-2)', P. monoensis, USA, NV, Lyon Co., 19 May 1993, J. Dempcy 116-2* (SFSU), AY630316, - (*duplicate collection of Dempcy 116-1)', P. monoensis, USA, CA, Sierra Co., 10 Jun 2008, D. Kruse-Pickler 16 (SFSU), -, JQ249993; P. monoensis, USA, NV, A. Tiehm 15516 (SFSU), -, JQ249938; Phacelia mutabilis Greene, USA, CA, Mono Co., 11 Jun 2008, G. K. Walden 63 (SFSU), -, JQ249994; Phacelia nasMana Jeps., USA, CA, Kem Co., 19 Mar 2003, D. R. Hansen 04 (SFSU), FJ8 14637, JQ249995; P. nashiana, USA, CA, Kem Co., 13 Mar 2004, D. R. Hansen 13 (SFSU), FJ8 14638, JQ249996; P. nashiana, USA, CA, Kern Co., 2005, D. R. Hansen 41 (SFSU), FJ8 14639, -; P. nashiana, USA, CA, Kern Co., 26 Mar 2008, G. K. Walden 28 (SFSU), -, JQ249997; Phacelia neglecta M. E. Jones, USA, CA, Inyo Co., n.d., Castagnoli et al 124 (CAS, n.v.), AY630319, -; P. neglecta, USA, CA, San Bernardino Co., 2 Apr 1978, J. Hendrickson 16473 (RSA278404, n.v.), AY630318, -; P. neglecta, USA, CA, Riverside Co., 4 Apr 1992, A. C. Sanders and E. J. Lawlor 12090 (RSA554652, n.v.), AY630317, -; Phacelia nemoralis Greene var. nemoralis, USA, CA, Marin Co., n.d., H. Leschke s.n. (SESU), AY630320, -; Phacelia nemoralis Greene var. oregonen- sis (Heckard) Walden & R. Patt., USA, OR, Lane Co., 29 Jul 2008, R. R. Halse 7486 (SFSU), JQ249998; Phacelia neomexicana Thurb. ex Torr., USA, NM, 1997, -, (TEX00255260, n.v.), JX233466, -; Phacelia noven- millensis Munz, USA, CA, Kern Co., 26 May 1987, B. Enter, J. Shevock, and T Sholars 7005 (DA VI 12895), JQ249999; Phacelia pachyphylla A. Gray, USA, CA, San Bernardino Co., 2 Apr 1995, J. Dempcy 130-1* (SFSU), AY630323, - (*duplicate collection of Dempcy 130-2, 130-3)', P. pachyphylla, USA, CA, San Bernar- dino Co., 2 Apr 1995, J. Dempcy 130-2* (SFSU), AY630324, - (*duplicate collection of Dempcy 130-1, 130-3)', P. pachyphylla, USA, CA, San Bernardino Co., 2 Apr 1995, J. Dempcy 130-3* (SFSU), -, JQ250000 (*duplicate collection of Dempcy 130-1, 130-2)', P. pachyphylla, USA, CA, Kern Co., 26 May 1977, R. Gustavson 487 (RSA428130, n.v.), AY630321, -; P. pachyphylla, USA, CA, Kern Co., Sanders 227 (CAS, n.v.), AY630322, -; Phacelia palmeriToii. ex S. Watson, USA, UT, Washington Co., 9 May 2005, L. M. Garrison 51 (SFSU), JX233467, -; Phacelia parryi Torr., USA, CA, Orange Co., 2008, B. Allen s.n. (SFSU), JQ250001; P. parryi, USA, CA, San Diego Co., 8 Apr 1990, J. Dempcy 03 (SFSU), AY653742, -; P. parryi, USA, CA, San Diego Co., 7 May 2004, D. R. Hansen 24 (SFSU), FJ814635, -; P. parryi, USA, CA, San Diego Co., 7 May 2004, D. R. Hansen 26 (SFSU), FJ8 14636, JQ250002; P. parryi, USA, CA, San Diego Co., 03 Apr 2008, /. P. Rebman 14277 and M. IFa// (SDl 84376, n.v.), JQ250003; Phacelia patulijlora (Engelm. & A. Gray) A. Gray, USA, TX, Hidalgo Co., 21 Feb 1992, R. R. Halse 4425 (GH, n.v.), AF091 198,AF047781; P. patuli- flora, USA, TX, n.d., direct submission to GenBank (WTU, n.v.), -, AF 130 179; Phacelia pauciflora S. Watson, MEXICO, Baja California Norte, 26 Mar 2009, G. K Walden 128 (SFSU), -, JQ250005; P. pauciflora, MEXICO, Baja California Norte, 29 Mar 1989, G. L. Webster 26155 (DAV133615), -, JQ250004; Phacelia pedicellata A. Gray, USA, CA, Inyo Co., 25 Mar 2004, L. M. Garrison 07 (SFSU), JX233468, -; P. pedicellata, USA, AZ, Pima Co., 27 Mar 2004, L. M. Garrison 13 (SFSU), JX233469, -; P. pedicellata, MEXICO, Baja California Norte, 26 Mar 2009, G. K. Walden 127 (SFSU), -, JQ250006; Phacelia perityloides Coville, USA, CA, Inyo Co., 7 May 2005, L. M. Garrison 47 (SFSU), JX233470, JQ250007; Phacelia petrosa N. D. Atwood, F.J. Sm. & T. A. Knight, USA, AZ, Mohave Co., 27 Apr 2004, N. D. Atwood and L. C. Higgins 30142 (DAV175105), -, JQ250008; Phacelia popei Torr. & A. Gray, USA, NM, Eddy Co., 23 Mar 2005, L. M. Garrison 36 (SFSU), JX233471, JQ250009; P. popei, USA, TX, Culberson Co., 24 Mar 2005, L. M. Garrison 39 (SFSU), JX233472, -; Phacelia pringlei A. Gray, USA, CA, Siskiyou Co., 21 Jun 2005, R. Patterson and S. Santos 1980 (SFSU), -, JQ250010; Phacelia procera A. Gray, USA, CA, Sierra Co., Jun 2008, P. Hankamp s.n. (SFSU), JQ25001 1; P. procera, USA, CA, Shasta Co., 8 Jun 1969, H. Thiers 23458 (SFSU08424), JQ250012; P. procera, USA, CA, Lake Co., 20 Jul 1975, D. Toren 1986 (SFSU08423), AY630325, -; Phacelia pulchella A. Gray, USA, UT, Washington Co., 9 May 2005, L. M. Garrison 52 (SFSU), JX233473, JQ250013; Phacelia purpusii Bran- degee, USA, CA, El Dorado Co., 12 Jun 1990, Barron s.n. (DAV152217), -, JQ250014; Phacelia rafaelensis N. D. Atwood, USA, AZ, Coconino Co., 10 May 2005, L. M. Garrison 54 (SFSU), JX233474, JQ250015; Phacelia ramosissima Douglas ex Lehm., USA, CA, Mono Co., 18 Jun 1993, D. M. Ferguson 10 (GH), AF091199, AF047821; Phacelia ramosissima Douglas ex Lehm. var. austrolitoralis Munz, USA, CA, Santa Barbara Co., 24 Jun 2008, G. K. Walden 67 (SFSU), JQ250016; Phacelia ramosissima Douglas ex Lehm. var. eremophila (Greene) J. F. Macbr., USA, CA, Mono Co., n.d., H. D. Thiers 17121 (SFSU), AY630327, -; P. ramosissima var. eremophila, USA, CA, Mono Co., 11 Aug 2008, G. K. Walden 83 (SFSU), -, JQ250024; Phacelia ramosissima Douglas ex Lehm. var. latifolia (Torr.) Cronquist, USA, CA, Orange Co., 26 Jun 2008, G. K. Walden 76 (SFSU), -, JQ250017; Phacelia ramosissima Douglas ex Lehm. var. subglabra M. Peck, USA, CA, Inyo Co., 13 Jun ' 2009, G. K. Walden 200a (SFSU), -, JQ250018; Phacelia rotundifolia Torr. ex S. Watson, USA, CA, San Bernardino Co., 9 Apr 1993, C. Condos 09 (SFSU), AY630328, -; P. rotundifolia, USA, CA, Inyo Co., 18 Mar 1986, J. Morefield and McCarty 3274 (GH), AF091200, AF047779; P. rotundifolia, USA, CA, Inyo 2014] WALDEN ET AL.; PHYLOGENIES AND CHROMOSOME EVOLUTION OF PHACELIA 47 Co., 17 Apr 1973, B. Trowbridge 3180 (SFSU08433), JQ250019; Phacelia mpestris Greene, USA, TX, Jeff Davis Co., 31 Aug 1997, W. R. Carr 16928 (TEX00041271, n.v.), JX233475, Phacelia salina (A. Nelson) J. T. Howell, USA, WY: Sweetwater Co., n.d., B. E. Nelson 36344 (RM, n.v.), AY630329, Phacelia scariosa Brandegee, MEXICO, Baja California Sur, Sierra de Guadalupe, Mulege, Miguel Dominguez Leon 3274 (SDNHM, n.v.), JX233476, JQ250020; Phacelia scopulina (A. Nelson) J. T. Howell, USA, NV, Elko Co., n.d., A. Tiehm 10573 (OSC, n.v.), AY630330, Phacelia sericea (Graham) A. Gray, USA, UT, Grand Co., 22 Aug 1985, B. Franklin 2301 (GH), AF091201, AF047778; P. sericea, USA, CA, Modoc Co., n.d., M. A. Showers s.n. (SFSU), AY630331, — ; Phacelia splendens Eastwood, USA, UT, Grand Co., 1997, N. D. Atwood and S. Welsh 22060 (TEX, n.v.), JX233477, JQ250021; Phacelia stebMmsii Constance & Heckard, USA, CA, Placer Co., 8 Jul 1977, G. L. Stebbins 7761 (DAV79856), -, JQ250022; Phacelia stellaris Brand, USA, CA, San Diego Co., Mar 2008, C. Burrascano s.n. (SFSU), JQ250023; Phacelia suaveolens Greene, USA, CA, Santa Cruz Co., 10 Jul 2009, V. T. Parker s.n. (SFSU), -, JQ250025; Phacelia tetramera J. T. Howell, USA, NV, Humboldt Co., n.d., A. Tiehm 12133 (CAS, n.v.), AY630333, P. tetramera, USA, NV, Humboldt Co., 1 1 Jun 2008, A. Tiehm 15641 (SFSU), -, JQ250027; Phacelia thermalis Greene, USA, OR, Lake Co., 10 Jun 1996, D. M. Ferguson 125 (GH), AF091202, AF047795; Phacelia vallis-mortae J. W. Voss, USA, CA, Inyo Co., 23 Mar 2004, L. M. Garrison 04 (SFSU), -, JQ250028; Phacelia viscida (Benth.) Torr. var. albijlora A. Gray, USA, CA: Santa Barbara Co., 2005, D. R. Hansen 50 (SFSU), FJ8 14653, P. viscida var. albiflora, USA, CA, Santa Barbara Co., Channel Islands, 28 Mar 2005, D. R. Hansen 54 (SFSU), FJ8 14650, JQ250031; Phacelia viscida (Benth.) Torr. var. viscida, USA, CA, Ventura Co., 9 May 2004, D. R. Hansen 28 (SFSU), FJ8 14651*, JQ250029 (*sequence identical to FJ8 14652); P. viscida var. viscida, USA, CA, Santa Barbara Co., 2005, D. R. Hansen 56 (SFSU), FJ8 14652*, JQ250030 (*sequence identical to FJ8 14651); Phacelia vossii N. D. Atwood, MEXICO, Nuevo Leon, 1990, - (TEX00224027, n.v.), JX233480, JQ250032; Phacelia welsMi N. D. Atwood, USA, AZ, Coconino Co., 10 May 2005, L. M. Garrison 56 (SFSU), JX233481, -; Romanzoffia californica Greene, USA, CA, Romanzoffia californica, n.d., D. M. Ferguson 128 (GH, n.v.), AF091205, AF047804; Romanzoffia thompsonii Marttala, D. M. Ferguson 134 (GH, n.v.), AF091206, AF047784; Tricardia watsonii Torr. ex S. Watson, Inyo Co., 10 Apr 1986, J. Morefield & McCarty 3375 (GH, n.v.), AF091209, AF047775. Appendix 2 List of sequences excluded from analyses. Voucher specimens missing. The herbarium at SFSU was renovated as part of a seismic retrofit of Hensill Hall (2000-2005). The vascular plant collection was stored offsite during a portion of the renovation, prior to installation into the current facility (H.D. Thiers Herbarium, SFSU). Some voucher collections previ- ously cited in the literature have not been located in the current collections (including unmounted research material) at SFSU or located in personal collections of the researchers. These vouchers are presumed lost, missing, or destroyed, and GenBank sequences derived from those vouchers were excluded from this analysis and listed here: Phacelia adenophova J. T. Howell, USA, CA, Lassen Co., J. Dempcy 116 (SFSU), AY630259 (nrlTS); Phacelia holanderi A. Gray, USA, CA, Del Norte Co., C. Gilbert 54 (SFSU), AY630270 (nrlTS); Phacelia humilis Torr. & A. Gray, USA, CA, Sonoma Co., C. Gilbert 3 A (SFSU), AY630301 (nrlTS) {P. humilis as currently circumscribed is not known to occur in Sonoma Co., and without the voucher specimen it is impossible to speculate on the putative identity of this specimen); Phacelia hydrophylloides Torr. ex A. Gray, USA, CA, Sierra Co., J. Dempcy 126 (SFSU), AY630303 (nrlTS); P. ivesiana Torr., USA, CA, San Bernardino Co., J. Dempcy 8 (SFSU), AY630309 (nrlTS); Phacelia leibergii Brand, USA, OR, Deschutes Co., J. Dempcy 99-2 (SFSU), AY630310 (nrlTS) (although J. Dempcy 99-2 was not located, a duplicate sheet is deposited at SFSU as J. Dempcy 99-1). Voucher specimens destroyed. The following voucher specimens previously cited in studies were consumed entirely during genomic extraction. No duplicate voucher material is deposited at SFSU for corrobora- tion, and GenBank sequences derived from these vouchers were excluded from analyses and listed here: Phacelia rotundifolia Torr. ex S. Watson, USA, CA, Inyo Co., 2005, L. M. Garrison and D. Hansen ROT (SFSU), FJ8 14628 (nrlTS), FJ8 14681 {rpll6 intron). Romanzoffia californica, USA, CA, in private cultiva- tion in San Mateo Co., horticultural material from Annie’s Annuals (Richmond, Contra Costa Co., CA), R. Patterson s.n., no voucher, harvested for genomic extraction in 2008, FJ814619 (nrlTS), FJ814691 {rpll6 intron). Voucher specimens deposited at SFSU without label information. Voucher specimens previously cited in studies were located at SFSU as unmounted research material without label information. GenBank sequences derived from these vouchers were also excluded from these analyses and listed here: Phacelia hastata Douglas ex Lehm., USA, CA, Inyo Co., C. Gilbert 106 (SFSU), AY630298 (nrlTS); P. hastata Douglas ex Lehm., USA, CA, Mono Co., C. Gilbert 109 (SFSU), AY630299 (nrlTS); Phacelia heterophylla Pursh var. virgata (Greene) R. D. Dorn, USA, CA, Alpine Co., C. Gilbert 101 (SFSU), AY630300 (nrlTS); Phacelia ramosissima Douglas ex Lehm., USA, CA, Alpine Co., Gilbert 100 (SFSU), AY630326 (nrlTS); Phacelia vallis-mortae J. W. Voss, USA, CA, Inyo Co., C. Gilbert 108 (SFSU), AY630332 (nrlTS) (cited as P. tanacetifolia Benth. in Gilbert et al 2005). Madrono, Vol. 61, No. 1, pp. 48-63, 2014 FACTORS DETERMINING THE ESTABLISHMENT OF PLANT ZONATION IN A SOUTHERN CALIFORNIAN RIPARIAN WOODLAND John M. Boland Boland Ecological Services, 3504 Louisiana Street, San Diego, CA 92104 JohnBoland@sbcglobaLnet Abstract This paper describes plant zonation in a southern California riparian woodland and identifies the factors responsible for the zonation. Of the 25 common trees and shrubs in the Tijuana River Valley, three were numerically dominant: Baccharis salicifolia (Ruiz & Pav.) Pers., Salix lasiolepis Benth., and Salix gooddingii C. R. Ball, here referred to as BASA, SALA, and SAGO, respectively. Adults of these species displayed significant down-slope zonation, with BASA, SALA, and SAGO most abundant in the High, Intermediate, and Low zone, respectively. Among new recruits, SALA and SAGO seedlings displayed zonation similar to that of adults, indicating that the zonation of SALA and SAGO was established at the time of recruitment. In contrast, BASA seedlings were more broadly distributed than adults; they were abundant in all zones, particularly the Low zone where adults were rare, indicating that the zonation of BASA adults was established post recruitment. For SALA and SAGO, the timing of fruiting and timing of water-level decline were the factors producing adult zonation; the two species had nearly non-overlapping periods of seed production, and this led to zonation of their seedlings on the banks as water levels declined. The seedling zonation was then retained in the adults. Because factors affecting recruitment played an important role in their zonation, these two willow species provide a new example of the supply-side influencing community structure. For BASA, whose seedlings were widely distributed, zonation of adults was the result of poor seedling survivorship in the Low zone during the first winter and poor adult survivorship in the Intermediate zone later. Results of this study can help guide future riparian restoration projects in southern California. Based on the prolific natural recruitment and rapid development of dense, native-dominated stands, use of a natural restoration approach where possible is recommended instead of the more common horticultural approach. Key Words: Baccharis salicifolia, fruiting phenology, riparian woodland, Salix gooddingii, Salix lasiolepis, recruitment, supply-side ecology, zonation. Riparian habitats are rare in southern Califor- nia because many of the rivers have been channel- ized, and there has been extensive development on the floodplains. An estimated 95% of the original riparian community has been eliminated during the past 200 years (Faber et al. 1989). Only recently has the value of these communities been recognized; remnant riparian habitats are now being preserved and protected, and each year a considerable amount of funding is directed towards their restoration and enhancement. Because these rem- nant riparian communities have not been well studied (Faber et al. 1989), and because there is little local information, project managers have to rely for guidance on studies conducted outside southern California where different species domi- nate, e.g., the desert Southwest (e.g., Stromberg et al, 1996), the Sierra Nevada (e.g., McBride and Strahan 1984a, b), and elsewhere (e.g., Krasny et al. 1988; Niiyama 1990; Mahoney and Rood 1998). Many wetland plant communities are charac- terized by striking species-zonation patterns across elevational gradients (Keddy 2010). The study of these zonation patterns and the process- es producing them have often led to a deep understanding of the entire community, e.g., algal communities (Dayton 1971; Robles and Deshar- nais 2002), mangroves (Rabinowitz 1978; Sousa et al. 2007), and salt marshes (Pennings and Callaway 1992; Pennings and Bertness 2001; Pennings et al. 2005). There have been only a few studies of riparian community composition in southern California (Bendix 1994, 1999; Oneal and Rotenberry 2008), and none has examined zonation. A productive approach for determining the underlying causes of species distributions in general, and zonation in partic- ular, is to examine the two “sides” of commu- nity development — the supply side and the interaction side (e.g., Lewin 1986; Roughgarden et al. 1987; Underwood and Fairweather 1989; Grosberg and Levitan 1992; Schmitt and Hol- brook 1999). Supply side refers to factors such as seed production and dispersal that affect the supply of propagules to an area, and the interac- tion side refers to factors such as competition and predation that affect the survivorship of recruits in the community (Roughgarden 2009). The ap- proach, therefore, is to determine when the zonation pattern becomes established, and this requires the study of reproduction, seedling establishment, and community development. 2014] BOLAND: RIPARIAN ZONATION 49 The most common perennial species in the riparian communities of southern California — Salix L. spp. (willows), Populus L. spp. (cotton- woods) and Baccharis salicifoUa (Ruiz & Pav.) Pers, (mule fat) — reproduce in similar ways. They produce tiny, wind-dispersed seeds that are short- lived (Stella et al. 2006), have no dormancy requirement (Emery 1988), and germinate within hours of landing (Karrenberg et al. 2002, Boland unpublished data) in their recruitment safe sites (Harper 1977). These safe sites are places suitable for germination where the substrate is both moist and in the sun (Karrenberg et al. 2002; Seiwa et al. 2008), and they typically occur in a narrow band immediately above the water’s edge, re- ferred to as the capillary fringe (Mahoney and Rood 1998). The common species disperse seeds in phase with the seasonal retreat of floodwaters during spring and summer. Community develop- ment is rapid once seedlings are established (Faber et al. 1989) because members of the Salicaceae, willows in particular, are among the fastest-growing tree species (Karrenberg et al. 2002). This study describes zonation in a southern California riparian woodland and identifies the factors that produce the zonation. In particular, this paper: (1) describes the distribution of adults within the community; (2) describes the distribu- tion of seedlings in areas of new recruitment; (3) examines factors affecting recruitment — timing of fruiting, timing of water-level decline, and timing of seedling establishment — to determine their influence on adult zonation; (4) examines seedling survivorship and change in community structure over time to determine the influence of these post-recruitment factors on adult zonation; and (5) discusses how the findings can improve restoration projects. This study of patterns and processes in a riparian woodland is one of very few that simultaneously quantifies seed production, recruitment, and survivorship of co-occurring tree species. Study Site and Dominant Species The Tijuana River Valley (32°33.080'N, 117°4.971'W) in San Diego Co., California, is a coastal floodplain that covers 1457 ha at approx- imately sea-level at the end of a 448,000 ha watershed. The climate is Mediterranean, with most of the rain falling between November and April (Zedler et al. 1992). The Tijuana River is an intermittent stream; flows are strong during winter and spring but cease during summer, reducing the river to a few widely-spaced pools. The river was confined to a narrow, unarmored channel during the 20th century when the valley was used extensively for agriculture, and riparian forests were absent (Boland, personal observa- tion). Widespread flooding, particularly during the 1980s, expanded the channel and forced out much of the agriculture (Zedler et al. 1992). The largest flooding events occurred in 1980, 1993, and 2005, and the riparian woodlands that developed in the flooded sites were therefore 32, 19, and 7 years old, respectively, in 2012. The riparian woodlands in the valley are still consid- ered pioneer, because succession is slow in these communities and may take 50-70 or more years to complete (Faber et al. 1989). The woodlands are preserved within three adjoining parks: the County of San Diego’s Tijuana River Valley Regional Park, the Border Field State Park, and the U.S. Fish and Wildlife Service’s Tijuana Slough National Wildlife Refuge. They are relatively undisturbed and support numerous bird species, including the endangered Vireo bellii pusillus (least Bell’s vireo) and Empidonax traiUii extimus (southwestern willow flycatcher; U.S. Fish and Wildlife Service 1994, 2005). The riparian woodlands in the valley have not been studied previously, but the spread and dispersal of Arundo donax L., a non-native, invasive species, has received attention (Boland 2006, 2008). The three most abundant species in the riparian habitats of the Tijuana River Valley are Baccharis salicifoUa (Ruiz & Pav.) Pers., Salix lasiolepis Benth. (arroyo willow), and Salix gooddingii C. R. Ball (Goodding’s black willow), here referred to as BASA, SALA, and SAGO, respectively. These dioecious species are common throughout California and the Southwest (Bald- win et al. 2012). Methods Distribution of Adults Riparian woodland bisect. To illustrate the down-slope characteristics of the vegetation, a bisect {sensu Barbour et al. 1987) was made through a 19-year-old woodland in the center of the valley (Fig. 1). A transect line was laid perpendicular to the course of the river from the highest riparian shrub on the floodplain to the center of the river (120 m). Every perennial tree and shrub within a 5 m belt alongside the line was identified and measured so that a scale drawing could be made. In addition, percent canopy cover of each perennial species was estimated within a 10 m-wide belt transect alongside the line (within 10 X 10 m quadrats). Ground elevations along the transect were measured with a GPS unit (Trimble R8 Model 2 GNSS rover), using the water level in the nearest pool in the river bed as the zero datum. GPS data were post-processed using Trimble Geomatics Office and Trimble Business Center. Density. To document species composition and density within the riparian woodlands (19- 50 MADRONO [Vol. 61 Fig. 1. Location of study sites in the Tijuana River Valley. Sites were used for the woodland bisect (1), density and relative abundance of adults (1-12), seedling profile (A), seedling density and survivorship (A-C), recruitment pots (two P), fruiting phenology (three F), woodland development (A-C, 1^, 13-14), and BASA skeletons (1^). 32 years old), quadrat surveys were conducted at 12 sites chosen in a stratified-random manner (Fig. 1). At each site, the habitat was divided into three zones — High (outside the forest), Inter- mediate (the forest edge), and Low (inside the forest) — and a 10 X 10 m quadrat was randomly placed in each zone. In each quadrat, all perennial trees and shrubs over 2 m tall were identified and counted. For each zone, density of each species was calculated as the mean number of individuals per 100 m^ quadrat (n = 12). Relative abundance. A second set of surveys was conducted at the same 12 sites to calculate the relative abundance of species and to test the distributions of BASA, SALA, and SAGO. [Statistical tests could not be done on the quadrat-survey results because the number of individuals in each quadrat was low.] A line transect was started at the location of the quadrat in the above density survey and run parallel to the course of the river within each zone. Along each transect, the first 33 perennial tree and shrub individuals over 2 m tall were identified and counted. Transect lengths varied with plant density and were 50^00 m long. This method used a similar sampling effort in each zone to collect a large, representative sample. Colonial plants, such as Salix exigua Nutt, (narrow-leaved willow), were counted as one individual if they intersected <10 m of the transect, or as two individuals if they intersected 10—20 m. [None intersected >20 m.] For each zone, the relative abundance of each species was calculated as the percent of the total 396 individuals censused in the zone. For BASA, SALA, and SAGO, numbers of individuals in the three zones were compared using replicated G-tests of indepen- dence (Sokal and Rohlf 1995). Overlaps in the zonation of the species were calculated using the proportional similarity index (PSI) applied to the frequency distributions: PSI = S min {pi, qi), where pi and qi represent the proportion of species p and q in zone i (Zaret and Smith 1984). The above adult surveys quantified the distri- bution and abundances of all the perennial species; the other aspects of this paper focus on only the three dominant species: BASA, SALA, and SAGO. Distribution of Seedlings Recruitment sites. Three recruitment sites (also called nursery sites, e.g., Mahoney and Rood 1998) were found in the Tijuana River Valley in December 2009 and followed during 2010. All three had been recently cleared of vegetation, were large enough to include the full range of floodwater elevations, were inundated during floods, and were within 200 m of adult riparian vegetation. The sites were named New Channel (650 m^), Dirt Road (780 m^), and Dairy Mart (1849 m^; Fig. 1). New Channel was a natural, freshly-scoured meander channel, whereas the other two sites had been cleared by bulldozers. Each site was visited regularly during early 2010 to monitor water levels and to stake the lowering water’s edge. Each site was divided into three zones based on the water levels on particular dates. The division between the High and Intermediate zones was the water level in late February, the division between the Intermediate 2014] BOLAND: RIPARIAN ZONATION 51 and Low zones was the water level in late April, and the lower extent of the Low zone was the water level in mid-July. These dates were used because they corresponded to the fruiting periods of SALA and SAGO, which were being followed simultaneously as described below. Elevations of the zones were measured during summer with the GPS unit described above, using the water level in the nearest pool in the river bed as the zero datum. At the three sites, the High, Intermediate, and Low zones had mean elevation spans of: 0.66 (±0.08 SD) m, 0.50 (±0.02 SD) m, and 0.24 (±0.07 SD) m, respectively, for a total elevation range of 1.4 m. Seedling profile. To illustrate the down-slope distribution of the seedlings, a profile of the seedlings was made on a steep bank at New Channel during June 2010, A 6 m transect was laid perpendicular to the course of the river from the highest reach of the flood flows to the center of the river meander. BASA, SALA, and SAGO seedlings were identified to species, and their densities were measured within quadrats (20 X 20 cm) at 0.5 m intervals along the transect. Seedling density. To document the distribution and density of seedliriy,^ -M the three recruitment sites, seedlings were cerisused soon after they grew their first few leaves and could be identified to species. Because recruitment occurred first in the High and Intermediate zones as the water level declined, these zones were censused first (April to July 2010). The Low zones were censused later (June to July 2010). At each site, 1-3 transects were randomly placed within each zone parallel to the course of the river along elevation contours, and BASA, SALA, and SAGO seedlings were counted within quadrats (20 X 20 cm) at 1 m intervals along the transect. A total of 261 quadrats (86, 87, and 88 at the three sites) and 9040 seedlings were counted. For the purpose of this paper, the term recruits refers to these new seedlings (<5 months old). Overlaps in the zonation of the species were calculated using the PSI as above (Zaret and Smith 1984), and distributions of the species were tested using replicated G-tests of independence (Sokal and Rohlf 1995). Factors Affecting Recruitment Fruiting phenology. To determine temporal changes in seed production, fruiting of the three species was monitored weekly from December 29, 2009 through August 4, 2010 (n = 32 wk). Each week the same 12-15 adult females of each species were visited in three areas of the valley (Fig. 1). These plants had mean heights (±SD) of: 10.3 (±3.0) m for SAGO (n = 15); 6.5 (±1.3) m for SALA (n = 12); and 2.9 (±0.2) m for BASA (n = 12). Fruiting flowers (BASA) and fruiting catkins (SALA and SAGO) have a conspicuous fluffy appearance and, using the naked eye and binoculars, their percent cover on each plant was estimated from within 1 5 m of the plant. For consistency, the same person collected all of the fruiting data. Percent-cover data were arcsine transformed and averaged, giving the mean percent cover of fruiting flowers for each species on each survey date. Because the counts were of the mature, fluffy seeds that were ready for dispersal, the percent-cover data collected estimated both the abundance of fruiting and of seed production. Overlap between fruiting fre- quency distributions of the three species was calculated using the PSI as above (Zaret and Smith 1984). Similar visual estimations of percent cover are commonly used in field studies and have been shown to be accurate (e.g., Dethier et ai. 1993; Brakenhielm and Qinghong 1995). In this study, repeat estimates done on the same day indicated that the percent-cover method was sufficiently precise; the root mean square error of paired counts was ±4%, which compares favorably with a different fruiting abundance method used by Stella et al. (2006), who found the root mean square error of their paired counts to be ±10%. Timing of recruitment. To examine temporal changes in recruitment, artificial safe sites were made available for short periods during the fruiting months. Simple seed traps could not be used because the seeds of SALA and SAGO cannot be distinguished (Boland personal ob- servation). The artificial safe sites, or recruit- ment pots, were flower pots filled with moist sand. These provided suitable conditions for seeds that landed on them to germinate and grow to a size at which they could be identified to species. The pots were one-gallon, plastic flower pots (16.5 cm tall, 15 cm diameter, 0.018 surface area) filled with clean sediment from the nearby Goat Canyon sedimentation basin. This sediment has been identified as 53% sand, 40% clay, and 7% silt, and is classified as sandy clay (Nautilus Environmental, unpub- lished data). Each pot was placed alone in a plastic basin (14.2 liter; 35 X 31 X 15 cm) that was filled with water so that the surface of the sediment in the pot was kept moist through capillary action. Ten recruitment pots (with their basins) were put out each month from January to August 2010 (n = 8 monthly sets of pots). They were placed close to the river, 1 m apart, in two open areas within the riparian habitat; five pots were in the riparian forest near Hollister Bridge, and five were in the riparian shrub community approxi- mately 800 m to the west (Fig. 1). The pots were left in the field for two weeks, during which time wind-borne seeds landed on the sediment surface, 52 MADRONO [VoL 61 germinated, and began to grow. After two weeks of exposure, the pots were taken to a sunny location outside the valley where they were kept moist in basins of water. The pots were covered with a fine netting to exclude any new seeds, and seedlings were grown to a size that allowed species identification, approximately 1 cm tall. Seedlings in each pot were counted after 4-10 wk (winter seedlings needed longer to grow than summer seedlings). Each month, one control pot was prepared and treated like the others, but was not exposed in the field. No seedlings grew in the control pots, showing that the sediment contained no viable seeds. For analysis, the five pots at each site were treated as subsamples, and the average number of seedlings per pot was calculated each month as the average of the two sites (n = 2 sites). For each species, the strength of the link between fruiting percent cover and the number of seedlings in the pots was quantified using linear regressions. Arcsine-transformed fruiting percent-cover data were averaged for the two-week period that the recruitment pots were in the field, and that average was run against the average number of seedlings per pot during the same period. Correlation coefficients were tested for significance using the t-test (Sokal and Rohlf 1995). Water-level change and the predicted distribu- tion of seedlings. To test whether only the timing of fruiting and the timing of water-level change could account for the observed zonation of seedlings, the two factors were combined into a Seedling Distribution Prediction Index (SDPI). Fruiting data for 2009-2010 were from the section above. Water-level data for 2009-2010 were obtained from the International Boundary and Water Commission for station #1000, which is on the Tijuana River approximately 2 km upstream from the study sites. Daily flows (cubic m per second) were converted to daily elevations (height in m above stage) using the rating table for the station. Precipitation data for the same period were obtained from the Tijuana River National Estuary Research Re- serve's weather station, which is approximately 4 km downstream from the study sites. To determine the SDPI for each species, the total range of water levels during 2009-2010 was divided into 1 cm intervals, and the 1 cm intervals were populated with virtual seeds by putting the fruiting percent cover for each day (n = 32 d) into that day’s water level. The virtual seeds were then summed for each 1 cm interval. The resulting SDPI for each species showed the predicted distribution of seedlings against height on the river bank, based on only fruiting and water levels. Overlaps in the zonation of species in the SDPIs were calculated using the PSI as above (Zaret and Smith 1984). Post-recruitment Factors Seedling survivorship. The 2010 seedling cohort was re-censused twice to determine survivorship over the first summer and the first winter. To measure survivorship over the first summer, seedlings at the three recruitment sites were re- censused in late summer (August to September 2010). The time between the initial census and this second census averaged 68 days (±11.5 SD) for each zone at the three sites (n = 9). At each site, 1-3 horizontal transects were placed in each zone, and BASA, SALA, and SAGO seedlings were counted within quadrats (20 X 20 cm) at 1 m intervals along each transect. A total of 218 quadrats (86, 46, and 86 at the three sites) and 4279 seedlings were counted. Transects and quadrats were in approximately the same posi- tions as those in the initial census described above. To measure survivorship over the first winter, the seedlings (now yearlings) were censused again during summer 2011 (May to August). The time between the second (late summer) and this third census averaged 309 days (±16.4 SD) for each zone at the three sites (n = 9). At each site, 1-3 horizontal transects were placed in each zone, and yearlings were counted within quadrats (20 X 20 cm) at 1 m intervals along each transect. Transects and quadrats were in approximately the same positions as those in the initial census. A total of 275 quadrats (90, 92, and 93 at the three sites) and 396 yearlings were counted. These yearlings (approximately 1.5 years old) were easily distinguished from any new 2011 recruits by their larger size; yearlings were —1,5 m tall, whereas new recruits were —5 cm tall. Riparian woodland development. To understand changes in community structure during woodland development, four characteristics were measured within stands of different ages. The characteris- tics were plant density, canopy height, canopy percent cover, and light level. The stand ages were 0.5 years (n = 3); 1.5 years (n = 3); 7 years (n = 2); and 19 years (n = 4; Fig. 1). Plant densities of BASA, SALA, and SAGO were obtained from the seedling, yearling, and adult surveys described above. Additional densities were obtained in two 7-year-old stands. In these stands, 1 X 1 m quadrats were placed at 3 m intervals along 30 m transects parallel to the river flow (n = 1 1 quadrats in each zone at each site). Maximum canopy heights were measured with a meter stick, stiff meter tape, or laser distance measurer (Bosch DLR130K). In the 0.5- and 1.5- year-old stands, heights were taken every meter along transects parallel to the river flow (n = 18- 32 measurements in each zone at each site). In the 7“ and 19-year-old stands, heights were taken every 10 m along 30 m transects parallel to the river flow (n = 3 measurements in each zone at 2014] BOLAND; RIPARIAN ZONATION 53 [ BASA HIGH I INTERMEDIATE" | LOW -20 Oj 10 20 30 40 50 60 70| 80 S0| 100 110 120| Fig. 2. A bisect through a typical 19-year-old riparian woodland in the Tijuana River Valley. The bisect ends in the center of the river bed, and the opposite bank is a mirror image of the one shown. Canopy percent cover is shown above. Scales for elevation (left), tree height (right), and distance (bottom) are in meters. The X indicates an adult BASA that was almost dead. each site). Percent canopy cover measurements were taken 1.15 m above ground level using a spherical densiometer (Forest Densiometers). At all sites, measurements were taken at 3 m intervals along 30 m transects parallel to the river flow (n = 10 measurements in each zone). Light levels were measured at 1.15 m above the ground, during midday (11:30 a.m.-l:30 p.m.) on cloudless days using a light meter (Extech Instruments Model 401025). At each site, light was measured first outside the riparian zone in bright sunshine (ambient), then within each zone, and finally outside the riparian zone a second time. At all sites, measurements were taken at 3 m intervals along 30 m transects parallel to the river flow (n == 10 measurements in each zone). Light levels in each zone were averaged, and are presented as the percent of the ambient light level. BASA skeletons. Because dead BASA were conspicuous in the Intermediate zones of the forests, the height of BASA skeletons and the canopy overhead were documented. At each of the four 19-year-old stands, a transect line was randomly placed parallel to the river in the Intermediate zone. Of the first 50 BASA individ- uals along the transect (30-50 m in length), the tallest 10 BASA skeletons were flagged. The height of each flagged skeleton was recorded, along with the height of the willow canopy directly overhead at that spot. Results Distribution of Adults The bisect at one site showed that the riparian woodlands in the Tijuana River Valley consisted of a tall forest in the riverbed with shorter shrubs on the terraces above (Fig. 2). It also showed that three species — BASA, SALA, and SAGO — were dominant and exhibited a distinct down-slope zonation pattern. Each species had its greatest density and canopy cover in a different zone: BASA in the High zone on the upper terraces; SALA in the Intermediate zone at the edge of the forest; and SAGO in the Low zone within the forest in the riverbed. The more extensive surveys of adults in the 1 9- to 32-year-old sites reinforced this view of a 54 MADRONO [Vol. 61 Table 1. Density and Relative Abundance of the Perennial Species Within the Adult Riparian Woodlands of the Tijuana River Valley. Density is the average number of individuals per quadrat (100 m^) within each zone (n = 12 sites). Relative abundance is the percent of total individuals along line transects (n = 396 individuals in each zone). Nomenclature follows The Jepson Manual (Baldwin et al. 2012). N = native; ex = non- native/exotic; INT. = Intermediate zone. Species name Common name Origin DENSITY HIGH INT. LOW RELATIVE ABUNDANCE HIGH INT. LOW Baccharis salicifolia (BASA) Mule fat N 3.0 0.6 0 68% 10% 3% Salix lasiolepis (SALA) Arroyo willow N 0 3.4 1.3 3% 78% 15% Salix gooddingii (SAGO) Goodding’s black willow N 0 0.8 5.9 0.8% 8% 76% Salix exigua Narrow-leaved willow N 0.3 0.7 0.1 9% 1% 0 Ricinus communis Castor bean ex 0.2 0.6 0.3 4% 1% 2% Baccharis sarothroides Broom baccharis N 0.2 0 0 4% 0 0 Myoporum laetum Myoporum ex 0 0.1 0 2% 0.3% 0 Arundo donax Giant reed ex 0 0 0.1 2% 1% 2% Tamarix ramosissima Tamarisk ex 0.1 0 0.3 1% 0 0.8% Isocoma menziesii Coastal goldenbush N 0 0 0 1% 0 0 Nicotiana glauca Tree tobacco ex 0.3 0 0 1% 0 0 Atriplex lentiformis Big saltbush N 0 0 0 0.8% 0 0 Hymenoclea monogyra Desert fragrance N 0.4 0 0 0.8% 0 0 Acacia cy clops Cyclops acacia ex 0 0 0 0.5% 0 0 Eucalyptus globulus Blue gum ex 0 0 0.1 0.3% 0 0.3% Schinus terebinthifolius Brazilian pepper tree ex 0 0 0 0.3% 0 0.3% Salix laevigata Red willow N 0 0 0 0.3% 1% 0 Tamarix aphylla Athel ex 0 0 0 0.3% 0 0 Artemisia douglasiana Mugwort N 0 0 0 0.3% 0 0 Sambucus nigra subsp. canadensis Blue elderberry N 0 0 0 0.3% 0 0 Malosma lamina Laurel sumac N 0 0 0 0.3% 0 0 Populus fremontii Fremont cottonwood N 0 0 0 0 0 0.3% Morus alba White mulberry ex 0 0 0 0 0 0.3% Foeniculum vulgare Fennel ex 0.3 0 0 0 0 0 Schinus molle Peruvian pepper tree ex 0.1 0 0 0 0 0 Total number of individuals 57 66 97 396 396 396 Mean number of individuals per quadrat 4.8 5.5 8.1 SE 0.9 1.3 1.1 riparian community dominated by three species that exhibit down-slope zonation. Quadrat sur- veys at 12 sites showed that BAS A, SALA, and SAGO had peak densities in different zones: BASA in the High zone (mean of 3 individuals per 100 m^), SALA in the Intermediate zone (3.4 individuals per 100 m^), and SAGO in the Low zone (5.9 individuals per 100 m^. Table 1). Line- surveys at the same 12 sites showed that, of the 25 perennial shrub and tree species present, BASA, SALA, and SAGO accounted for a high percent- age of the individuals overall (88%), and each had its greatest relative abundance in a different zone: BASA accounted for 68% of individuals in the High zone; SALA for 78% in the Intermedi- ate zone; and SAGO for 76% in the Low zone (Table 1). Zonation of BASA, SALA, and SAGO was seen at all 12 survey sites, and the pattern was significant within each site (G-test, df = 4, P < 0.001) and for the valley as a whole (pooled data, G-test, n = 1041 individuals, G = 1254, df “ 4, P < 0.001). There was relatively little overlap between the adult distributions of the three species: on average, BASA and SALA had 17% overlap (±0.3% SE); BASA and SAGO had 7% overlap (±0.3% SE); and SALA and SAGO had 22% overlap (±0.2% SE; n = 12). Distribution of Seedlings The seedling profile at one site showed that new BASA, SALA, and SAGO recruits estab- lished in high numbers, up to 3375 seedlings per m^ (Fig. 3). BASA seedlings were the most broadly distributed and occurred in all zones, with greatest densities in the Intermediate and Low zones. SALA and SAGO seedlings were more narrowly distributed and non-overlap- ping, with SALA higher on the bank than SAGO. The more extensive surveys at the three recruitment sites also showed that new recruits established in high numbers — up to an average of 1523 seedlings per m^ (Low zone at New Channel) — and exhibited the same distribution patterns (Table 2). At all three sites, BASA seedlings had relatively high densities in all zones, SALA seedlings occurred almost exclu- sively in the High and Intermediate zones, and SAGO seedlings occurred almost exclusively in the Low zone. 2014] BOLAND: RIPARIAN ZONATION 55 BASA — 0 500 /m 2 1,000 /m 2 -- 1,500 /m 2 — 2,000 /m 2 2,500 /m 2 HIGH INTERMEDIATE LOW Fig. 3. A profile of the seedlings at a recruitment site in the Tijuana River Valley. Seedling densities are shown as number per m^. The profile was done on a steep bank, and the transect was 6 m long. Among seedlings, therefore, SALA and SAGO exhibited dear zonation but BASA did not. Zonation of SALA and SAGO seedlings was significant within each site (G-test, df = 2, P < 0.001) and in the pooled total for the three sites (G-test, n = 341 1 individuals, G = 3755, df = 2, P < 0.001). The distributions of SALA and Table 2. Seedling Density in the High, Intermediate, and Low Zones at the Three Recruitment Sites in Early Summer 2010. Numbers are means per m^; n = number of quadrats. BASA = Baccharis salicifolia, SALA = Salix lasiolepis, and SAGO = Salix goodingii. Recruitment site HIGH INT. LOW NEW CHANNEL BASA 407 840 1287 SALA 330 518 1 SAGO 0 48 236 TOT 737 1406 1523 SE 136 140 157 n 21 30 35 DIRT ROAD BASA 539 1078 694 SALA 780 368 1 SAGO 0 13 396 TOT 1319 1460 1091 SE 230 377 175 n 33 15 39 DAIRY MART BASA 11 10 27 SALA 49 23 0 SAGO 4 7 110 TOT 64 40 137 SE 12 8 23 n 27 31 30 SAGO seedlings were so distinct that they overlapped by only 3%, 9%, and 17% at the three sites. In contrast, the distribution of BASA seedlings overlapped extensively with both SALA (55%, 44%, and 49%) and SAGO (33%, 65%, and 68%) in the three sites. SALA and SAGO seedlings exhibited zonation similar to that observed among adults (SALA higher on the bank than SAGO, with little overlap in their distributions). In contrast, BASA seedlings were far more broadly distributed than adults, and they were most dense in the lower zones where BASA adults were virtually absent. Factors Affecting Recruitment Fruiting phenology. The three species had different fruiting curves (Fig. 4). BASA produced seeds during the entire study period (December 2009 to August 2010), and 95% of its seeds were produced in the six months between January 6 to July 14. In contrast, SALA and SAGO produced seeds for short periods during spring. SALA fruited first with a peak in mid-March, and 95% of its seeds were produced in the 10 weeks from February 24 to April 28. SAGO fruited later with a peak in mid-May, and 95% of its seeds were produced in the 13 weeks from April 21 to July 14. Peak seed production in SALA and SAGO was nine weeks apart, and there was only a 7.4% overlap in their seed production curves. Timing of recruitment. Seedlings of the three species were abundant in the recruitment pots, with total mean densities up to 136 seedlings per pot (7667 seedlings per m^). The three species had different recruitment curves (Fig. 5). BASA 56 MADRONO [VoL 61 Fig. 4. Fruiting periods of BASA, SALA, and SAGO recruited over a broad period from January through June. In contrast, SALA and SAGO recruited for shorter periods. SALA recruited from February to April, with a peak in March. SAGO recruited from April to June, with a peak in May. Peak recruitment in SALA and SAGO during 2010. Data are means ±1 SD. were eight weeks apart, and there was only a 1% overlap in their recruitment curves. For all three species, the number of seedlings in the pots during a given month was positively correlated with the intensity of fruiting of the adults during that month. The correlations were Fig. 5. Density of BASA, SALA, and SAGO seedlings in recruitment pots during 2010. Error bars are ±1 SE (n = 2 sites). Density per pot can be converted to density per m^ by multiplying by 55.6. 2014] BOLAND: RIPARIAN ZONATION 57 Fig. 6. Daily precipitation and river height in the Tijuana River Valley for December 28, 2009 to August 5, 2010. Data cover the period during which fruiting and recruitment were studied. relatively weak and not significant for BAS A (r^ = 0.496; P = 0.05). The correlations were strong and significant for SALA (r^ = 0.976; P < 0.001) and SAGO (r^ = 0.913; P < 0.001) and for the total community (r^ = 0.535; P < 0.05). These results indicate that there was close coupling between the timing of fruiting and the timing of recruitment. Water-level change and the predicted distribu- tion of seedlings. Water levels in the Tijuana River Valley fluctuated with rainfall during the study period (Fig. 6). They reached their highest levels (1.4 m) during January when rainfall was heavy and dropped to zero during May. The Seedling Distribution Prediction Index, based on fruiting and water-levels, predicted that: (1) SALA and SAGO seedlings would have nearly non-overlap- ping distributions (only 5% overlap); (2) SALA seedlings would be higher on the bank than SAGO seedlings; and (3) BASA seedlings would occur within all zones and overlap extensively with SALA and SAGO (overlaps of 48% and 36%, respectively). These predictions were all consistent with the observed seedling distribu- tions. Post-recruitment Factors Seedling survivorship. During the first summer, seedlings of all three species survived well in all zones, with the exception of SALA seedlings in the High zone (Table 3); summer survivorship for SALA in the High was only 3%, in contrast to 41% for BASA, which was also abundant in that zone. Many young SALA seedlings (<10 cm tall) were standing dead in the High zone during summer 2010, and they appeared to have died from desiccation. The result was that SALA densities in the High zone declined from an initial average of 386 seedlings per to an average of only 12 individuals per at the end of summer. During the first winter, seedlings of all three species survived well in most zones except for BASA seedlings in the Low zone (Table 3); winter survivorship of BASA in the Low was only 0.2%, in contrast to 19% for SAGO, which was also abundant in that zone. Many BASA yearlings (1.1-1 .9 m tali) were standing dead in the Low zone during early spring 2011. They were in pools of shallow water and appeared to have died from sustained anoxic conditions around their roots. In addition, many BASA yearlings could not be found and appeared to have been washed away by the winter flows. The result was that BASA densities in the Low zone declined from an initial average of 608 seedlings per in late summer to an average of only 1 individual per by the end of the first winter. Overall, despite many deaths during the first year, yearlings of the three species had total average densities of 22-59 per and formed dense, green thickets 1. 0-2.5 m tall in all zones. The yearlings also displayed a zonation that was 58 MADRONO [Vol. 61 Table 3. Seedling Density and Survivorship in the High, Intermediate, and Low Zones. 3A. Seedling density censused at three times; numbers are means per m^ (n = 3 sites). 3B. Seedling survivorship (%) calculated from the densities in A. An ‘x’ indicates that initial densities were too low to measure survivorship (<2 individuals per m-). BASA = Baccharis salicifoUa, SALA = Salix lasiolepis, and SAGO = Saiix gooddingii. HIGH INT. LOW A. DENSITY 1. Early summer 2010 BASA 319 643 669 SALA 386 303 1 SAGO 1 23 247 2. Late summer 2010 BASA 130 490 608 SALA 12 226 1 SAGO 1 8 106 3. Early summer 201 1 BASA 21 16 1 SALA 3 41 1 SAGO 0 2 20 B. SURVIVORSHIP After first summer (A1 to A2) BASA 41% 76% 91% SALA 3% 75% X SAGO X 33% 43% After first winter (A2 to A3) BASA 16% 3% 0.20% SALA 28% 18% X SAGO X 21% 19% similar to the adult zonation pattern, except that BASA yearlings were still abundant in the Intermediate zone. Riparian woodland development. To determine how and when BASA was essentially eliminated from the Intermediate zone, several aspects of the developing community were examined (Table 4). Plant densities decreased with age in all three zones. Densities were high at the time of recruitment (707-969 individuals per m^ on average) and were four orders of magnitude lower in the 19-year-old stands (0.06-0.11 indi- viduals per m^ on average). Canopy heights increased with age in all three zones. The greatest canopy heights were in the Low zone of 19-year- old forests, where trees reached an average height of approximately 12 m. Canopy cover increased with age in the Intermediate and Low zones where it was almost 100% in 7- and 19-year-old stands; it remained 4% or less in the High zone. Light levels decreased with age in the Intermedi- ate and Low zones, to less than 15% in 7- and 19- year-old stands; they remained 100% in the High zone. Overall, the community structure changed rapidly with age. In the Intermediate and Low zones, a dense forest grew up with a tall and dense canopy that greatly decreased the light levels within. BASA skeletons. In the Intermediate zone of 19-year-old woodlands, the maximum height of dead BASA individuals averaged 3.3 m (±0.6 SD; n = 40), whereas the height of the overhead SALA canopy averaged 8.3 m (±1.1 SD; n = 40). The presence of full-grown but dead BASA indicates that conditions in the Intermediate zone were initially favorable for growth, but condi- tions deteriorated some time after BASA had reached adult size. Discussion This study of a riparian woodland in southern California demonstrated that BASA, SALA, and SAGO were numerically and structurally domi- nant, and that they displayed a significant down- slope zonation pattern not previously described. Investigation of seed production, seedling recruit- ment, seedling survivorship, and community development identified the factors most respon- sible for the adult zonation patterns, and showed that the factors for SALA and SAGO were different than those for BASA. Factors Affecting Recruitment The zonation of SALA and SAGO was established at the time of recruitment, and the factors most responsible were the timing of fruiting and the timing of water-level changes. Seeds of these species germinated in the moist sediment just above the water line, and the two species established in sequence on the bank as the water level dropped. SALA established higher on the bank because its seeds were dispersed earlier when water levels were higher; SAGO established lower because its seeds were dispersed later when water levels were lower. The fruiting periods of SALA and SAGO were separate enough, and the water-level decline was steady enough, to result in distinct zonation of the seedlings. This zonation pattern observed among seedlings was retained as the stand aged and observed among adults (with only minor post-recruitment modification). Because Table 4. Characteristics of Riparian Woodlands in the Tijuana River Valley. Numbers are means (and SE) for density, canopy height, canopy cover, and light level in the High, Intermediate (INT), and Low zones of stands of four ages. 2014] BOLAND: RIPARIAN ZONATION 59 ^ o ^ ^ d d d t-- WON ^ q fN ^ ^ W WON ^ ^ in ^ q ^ ^ -Hfs^^ddoooo 04 ^ w wo ^ in ^ q P ^ ^ dr-oddoooo ON ^ 00 fN ^ q q q fNj ^ ^ ONodoNdoo— iONr4 ^ w WON q I NO cn F-^*d'ddr^f^‘ooo' W w On w ^ w q ^O) q ^ ^ ^LosG'^doooo in r4 ^ ^ doN^ddoo'oo' VO NO W wo ^ On ^ NO ^ ^ q q q q ^ ^ ONO^d^O^O^ en 00 ^ ^ ^ c-dddd'^t^oo w w wo ^ ^ ^ •^^ONddoooo CM w w wo ^ o m O NO cn cn o d d o o o o wo g c H 2 W > (U a ed w o X u < >- >< £ hS q o < 0 < H X 2 00 Q u u factors that affect recruitment play a primary role in their adult distributions, SALA and SAGO pro- vide a new example of the supply-side influencing zonation in a community (Grosberg and Levitan 1992; Sousa et al. 2007). The combination of fruiting period and water- level change as factors influencing seedling distribution has been predicted elsewhere for other species, usually with the proviso that sediment grain size and moisture content of the soil also play important roles (McBride and Strahan 1984b; Niiyama 1990; Van Splunder et al. 1995; Mahoney and Rood 1998). However, the recruitment pots used in this study eliminated these physical factors as alternative explanations for the observed zonation of seedlings, because the pots always contained the same sediment grain size and soil-moisture content. In addition, the Seedling Distribution Prediction Index based only on fruiting and water-levels predicted the observed zonation of the seedlings. It is unusual for fruiting period to play an important role in the structure of any community (Levine and Murrell 2003; Morisette et al. 2009), but the willow-dominated, riparian forest in this study has three characteristics that allow fruiting period to play such a role. First, events from fruiting to seedling recruitment occurred quickly with no time lags. It takes just minutes for wind- blown seeds to disperse to recruitment sites and, when the seeds land in a suitable area, they germinate within a few hours (Emery 1988; Young and Young 1992; Karrenberg et al. 2002, Boland, unpublished data). Fruiting period is unlikely to play an important role in commu- nity structure if species have persistent seed banks, longer seed-dormancy requirements, or slower dispersal mechanisms. Second, there was a relatively steady shift of recruitment safe sites down the bank as water levels declined. This allowed currently fruiting species to establish in sequence as water levels declined. Fruiting period is unlikely to play an important role in commu- nity structure when water-level declines are not orderly. For example. Van Splunder et al. (1995), working on the River Waal in the Netherlands, noted the sequential seed production of four Salicaceae species, but found no clear patterns in the distribution of seedlings because water levels rose and fell several times during seed produc- tion. Third, zonation patterns established in the first year were not disrupted by recruitment in later years. Once seedlings established in a disturbed site, they developed into a dense stand of even-aged adults; recruitment in later years, with different water levels and zones of recruit- ment, was so unsuccessful that it did not modify the zonation patterns of the stand (Boland unpublished data). Awareness of the importance of fruiting period to the distribution of riparian species should improve our general understand- 60 MADRONO [Vol. 61 ing of the role of phenology in species distribu- tions (Chuine 2010). It has been hypothesized that the early flowering of some wind-pollinated, deciduous trees has evolved because pollination rates are higher when the leaves are not fully developed, and the flowers are more exposed to breezes (Willson 1983). The findings of this study suggest that, for plants like SALA and SAGO, it is more likely that the timing of flowering and fruiting has evolved in response to selective pressure on the placement of seeds at an elevation on the banks where adults are most successful. Post-recruitment Factors The adult zonation pattern of BASA is primarily the result of post-recruitment factors affecting survivorship. BASA’s long fruiting period translated into a broad vertical distribu- tion of seedlings on the river bank, and post- recruitment mortality narrowed its distribution such that adults were abundant in only the High zone. This study identified two occasions in the development of the community when BASA deaths affected its zonation. During the first winter, nearly all of the BASA that had recruited to the Low zone died. Their deaths were due, in part, to their inability to withstand high flows and, in part, to their inability to survive several months in standing water. BASA, unlike the willows, is not noted for its ability to tolerate fast flows or anoxic conditions associated with standing water (Karrenberg et al. 2002). Then, during the next several years as the forests developed around them, many BASA adults in the Intermediate zone died. As shown by the presence of adult BASA skeletons under the dense willow canopy, BASA adults initially grew well in this zone until they were shaded and outcompeted by the taller-growing willows. Together, these results show that physiological tolerances and interspecific competition are the post-recruitment factors most responsible for the adult zonation of BASA. Because post-recruit- ment factors play the primary role, BASA is a new example of the interaction side (Rough- garden 2009) determining zonation in a commu- nity. BASA also appears to be an example of a competitively inferior plant that is forced to inhabit a more stressful zone (Grime 1979; Pennings and Bertness 2001). This study has shown that the positions of the zones and the main zonation patterns observed among adults are established by the end of the first year. At that time, the Low is dominated by SAGO, the Intermediate by SALA (and BASA), and the High by BASA. During later develop- ment the configuration of the bank may be altered through erosion or sedimentation, but the zones stay in essentially the same place. After 20 or 30 yr the cohort of initial recruits has developed into even-aged stands in each zone. The overall goal of studying patterns in this riparian woodland was to gain an understanding of the processes important in shaping the community. The study found that several factors acted consecutively to produce the species’ zonation; fruiting period and water levels acted first to produce seedling patterns, which, in SALA and SAGO, were retained to adulthood. Then physiological tolerances and interspecific competition acted to modify the seedling pattern, especially in BASA. The simultaneous study of seed production, recruitment, seedling survivor- ship, and woodland development showed where and when factors were important for each of the dominant species in the community. Applications in Riparian Restoration Projects The results of this study provide much-needed information to help managers plan and conduct riparian restoration projects in southern Califor- nia, The two basic approaches to riparian restoration are horticultural restoration (Griggs 2009) and natural restoration (Briggs 1996). In southern California, horticultural restoration is typically used; nursery-grown container plants are planted in low densities, and farming practices, such as irrigation and weed control, are used to sustain the plants for the required maintenance period, usually five years (Griggs 2009). This is the appropriate method in sites that are not inundated by winter floods and at sites that are distant from a natural seed source. However, horticultural restoration can be costly and is not always successful; often there are extensive deaths when the irrigation is discontin- ued, and often the installed assemblage does not resemble a natural riparian community (Boland personal observation). This study can provide some guidance for horticultural restoration projects regarding spe- cies relative abundances, spatial arrangements, and densities. First, because this study found that only a few species are dominant among the many present, horticultural projects should use a species palette that approximates these unequal natural relative abundances. Second, because this study showed a down-slope zonation of the dominant species, horticultural projects should plant species clumped in the appropriate zones as opposed to random, haphazard, or other mixed arrangements. Finally, because this study showed that 19-32-year-old natural riparian woodlands are relatively dense, horticultural projects should plant at high densities. If, for example, one assumes a 95% annual survivorship rate over 19 years for installed plants, then initial planting densities would need to be 2. 6-4.4 X greater than typically used at present (—200 per acre, or —500 2014] BOLAND: RIPARIAN ZONATION 61 per ha; River Partners 2007) to equal natural, 19- year-old densities. By making the relative abun- dances, spatial arrangements, and densities more closely resemble those observed in natural riparian communities, restoration projects that use a horticultural approach are more likely to be successful. The alternative to horticultural restoration is natural restoration. In natural restoration, a site is prepared (usually cleared and graded), and revegetation is allowed to proceed naturally with little or no human intervention (Briggs 1996). This is an appropriate method in sites that are inundated by floods during winter and which have natural seed sources nearby. If properly timed and carried out, this method can be more effective and less costly than horticultural resto- ration. Unfortunately, natural restoration is rarely used in southern California at present. This study provides the empirical foundation needed for managers to use the natural restora- tion approach. In particular, knowing that fruiting period and recruitment are so closely linked means that one can predict when the dominant riparian species will recruit. Also, knowing that recruitment safe sites and water level are so closely linked means that one can predict where the recruitment will occur. In natural restoration, this time-and-place predict- ability is of critical importance because success relies on water-level decline coinciding with peak seed production (Mahoney and Rood 1998). If a project involves the breaching of a berm to flood a restoration site, the water-level decline can be timed to ensure the recruitment of desired species to the site. In rangeland sites, natural restoration (called natural recovery in rangeland literature) has been criticized for slow development, dominance by undesirable plants, unpredictable results, and excessive herbivore damage (Whisenant 1999). Clearly, these problems have not occurred in the natural development of riparian sites of the Tijuana River Valley, where development was rapid, native plants were dominant, and there was little evidence of herbivore damage. In addition, the riparian woodlands that have developed naturally at these sites are of high quality and support many species, including endangered bird species. Natural restoration may be more successful in riparian habitats because, unlike rangelands, riparian communities are well adapted to frequent, extensive distur- bances and have the ability to regenerate quickly after a disturbance (Faber et al. 1989; Sher et al. 2002). Ideally, natural restoration will become more common in southern California in the future because it has several characteristics that make it superior to horticultural restoration. First, it produces a community with a high density (and cover) of seedlings and adults. If adult riparian trees are nearby (Friedman et al. 1995), recruit- ment is likely to be on the order of 40-1500 seedlings per m^ as observed in this study. Over time, naturally recruited stands will remain denser (and have greater cover) than horticultur- ally restored stands. In the Tijuana River Valley, densities within 7-year-old stands were 26 X greater than — and densities within 19-year-old stands were approximately double — the initial planting density of a typical horticultural project (—200 individuals per acre, or —500 per ha; River Partners 2007). Second, natural restoration pro- duces a community with the appropriate spatial distribution of species, including the down-slope zonation of dominant species. In horticultural restoration projects, species are often mixed in space resulting in poor survivorship of individu- als planted at less than optimal elevations, especially once irrigation has been discontinued. Third, natural restoration results in a community in which local plant species have appropriate sex ratios and genetic diversity (Briggs 1996; Landis et al. 2003). Some horticultural practices, such as the use of multiple cuttings from only a few source individuals or use of non-local stock, can result in unnatural sex ratios or unnatural genetic diversity among the installed plants. Finally, natural restoration is likely to be considerably less expensive than horticultural restoration. It requires no container plants or workers to plant them, no irrigation systems or workers to install and maintain them, and no weed control or post- recruitment maintenance. Because of the substantial benefits of natural restoration of riparian habitats, resource agencies with the authority to approve restoration plans should require that natural restoration be at- tempted before horticultural restoration at sites where such an approach would be appropriate. Greater emphasis on natural restoration will require a shift in the way restoration projects are planned, approved, and conducted, but likely would lead to decreased costs, increased quality, and improved long-term success of restoration projects. Acknowledgments I thank Lisa Ordonez for assistance in the field; Michelle Cordrey at the Tijuana River National Estuarine Research Reserve for the GPS measure- ments; Jeff Crooks at Tijuana River National Estua- rine Research Reserve for rainfall data; Steve Smullen at the International Boundary and Water Commission for river-flow data; and Deborah Woodward, Tito Marchant, and two anonymous reviewers for helpful comments on an early draft of this manuscript. I also thank the U.S. Fish and Wildlife Service, California State Parks, and the San Diego Co. Parks and Recreation Department for preserving the Tijuana River Valley woodlands and making them available for study. 62 MADRONO [Vol. 61 Literature Cited Baldwin, B. G., D. H. Goldman, D. J. Keil, T. J. Rosatti, and D. H. Wilken, (eds.). 2012. The Jepson manual: vascular plants of California, 2nd ed. University of California Press, Berkeley, CA. Barbour, M. G., J. H. Burk, and W. D. Pitts. 1987. Terrestrial plant ecology, 2nd ed. Benjamin/Curnmings Publishing Co., Menlo Park, CA. Bendix, j. 1994. Among-site variation in riparian vegetation of the southern California Transverse Ranges. American Midland Naturalist 132:136- 151. . 1999. Stream power influence on southern Californian riparian vegetation. Journal of Vege- tation Science 10:243-252. Boland, J. M. 2006. The importance of layering in the rapid spread of Arundo donax (giant reed). Madrono 53:303-312. . 2008. The roles of floods and bulldozers in the break-up and dispersal of Arundo donax (giant reed). Madrono 55:216-222. BrAkenhielm, S. and L. Qinghong. 1995. Compar- ison of field methods in vegetation monitoring. Water, Air and Soil Pollution 79:75-87. Briggs, M. K. 1996. Riparian ecosystem recovery in arid lands. University of Arizona Press, Tucson, AZ. Chuine, I. 2010. Why does phenology drive species distribution? Philosophical Transactions of the Royal Society B 365:3149-3160. Dayton, P. K. 1971. Competition, disturbance, and community organization: the provision and subse- quent utilization of space in a rocky intertidal community. Ecological Monographs 41:351-389. Dethier, M. N., E. S. Graham, S. Cohen, and L. M. Tear. 1993. Visual versus random-point percent cover estimations: “objective” is not always better. Marine Ecology Progress Series 96:93-100. Emery, D. E. 1988. Seed propagation of native California plants. Santa Barbara Botanic Garden, Santa Barbara, CA. Faber, P. A., E. Keller, A. Sands, and B. M. Massey. 1989, The ecology of riparian habitats of the southern California coastal region: a commu- nity profile. U.S. Fish and Wildlife Service Biological Report 85 (7.27). Friedman, J. M., M. L. Scott, and W. M. Lewis. 1995. Restoration of riparian forest using irriga- tion, artificial disturbance, and natural seedfall. Environmental Management 19:547-557. Griggs, F. T. 2009. California Riparian Habitat Restoration Handbook. Published by River Part- ners. Website http://riverpartners.org/documents/ Restoration_Handbook_Final_Dec09.pdf (ac- cessed 28 March 2012). Grime, J. P. 1979. Plant strategies and vegetation processes. J. Wiley & Sons, Chichester, UK. Grosberg, R. K. and D. R. Levitan. 1992. For adults only? Supply-side ecology and the history of larval biology. Trends in Ecology and Evolution 7:130-133. Harper, J. L. 1977. Population biology of plants. Academic Press, New York, NY. Karrenberg, S., P. j. Edwards, and J. Kollmann. 2002. The life history of Salicaceae living in the active zone of floodplains. Freshwater Biology 47:733-748. Keddy, P. a. 2010. Wetland Ecology: principles and Conservation. Cambridge University Press, Cam- bridge, MA. Krasny, M. E., K. a. Vogt, and J. C. Zasada. 1988. Establishment of four Salicaceae species by seed and vegetative reproduction on river bars in interior Alaska. Holarctic Ecology 11:210-219. Landis, T. D., D. R. Dreesen, and R. K. Dumroese. 2003. Sex and the single Salix: considerations for riparian restoration. Native Plants Journal 4:110- 117. Levine, J. M. and D. Murrell. 2003. Community- level consequences of seed dispersal patterns. Annual review of Ecology and Systematics 34: 549-574. Lewin, R. 1986. Supply-side ecology. Science 234:25- 27. Mahoney, J. M. and S. B. Rood. 1998. Streamflow requirements for cottonwood seedling recruitment — an integrative model. Wetlands 18:634—645. Morisette, j. T., a. D. Richardson, A. K. Knapp, J. 1. Fisher, E. A. Graham, J. Abatzoglou, B. E. Wilson, D. D. Breshears, G. M. Henebry, J. M. Hanes, and L. Liang. 2009. Tracking the rhythm of the seasons in the face of global change: phenological research in the 21st century. Frontiers in Ecology and the Environment 7:253-260. McBride, J. R. and J. Strahan. 1984a. Fluvial processes and woodland succession along Dry Creek, Sonoma County, California. Pp. 110-119 in R. E. Warner and K. M. Hendrix (eds.), California riparian systems: ecology, conservation, and pro- ductive management. University of California Press, Berkeley, CA. and . 1984b. Establishment and survival of woody riparian species on gravel bars of an intermittent stream. American Midland Naturalist 112:235-245. Niiyama, K. 1990. The role of seed dispersal and seedling traits in colonization and coexistence of Salix species in a seasonally flooded habitat. Ecological Research 5:317-331. Oneal, a. S. and j. T. Rotenberry. 2008. Riparian plant composition in an urbanizing landscape in southern California, USA. Landscape Ecology 23:553-567. Pennings, S. C. and M. D. Bertness. 2001. Salt marsh communities. Pp. 289-316 in M. E. Hay (ed.), Marine community ecology. Sinauer Associ- ates, Sunderland, MA. , M. Bestor-Grant, and M. D. Bertness. 2005. Plant zonation in low-latitude salt marshes: I disentangling the roles of flooding, salinity, and competition. Journal of Ecology 93:159-167. AND R. M. Callaway. 1992. Salt marsh plant zonation: the relative importance of competition ; and physical factors. Ecology 73:681-690. j Rabinowitz, D. 1978. Early growth of mangrove seedlings in Panama and an hypothesis concerning the relationship of dispersal and zonation. Journal ! of Biogeography 5:1 13-133. } River Partners. 2007. Riparian restoration plan for the Drumheller Slough Unit, Sacramento River Nation- al Wildlife Refuge, Glenn County, California. Prepared for U.S. Fish and Wildlife Service, Bureau of Reclamation, and Wildlife Conservation Board, i Website: http://www.riverpartners.org/documents/ 2014] BOLAND: RIPARIAN ZONATION 63 DrumhellerSloughFinalRestorationPlanRev 1 .pdf (accessed 28 March 2012). Robles, C. D. and R. A. Desharnais. 2002. History and current development of a paradigm of predation in rocky intertidal communities. Ecology 83:1521-1537. Roughgarden, J. 2009. Is there a general theory of community ecology? Biology and Philosophy 24: 521-529. , S. D. Gaines, and S. W. Pacala. 1987. Supply side ecology: the role of physical transport processes. Pp. 491-581 in J. H. Gee and P. S. Giller (eds.), Organization of communities past and present. Blackwell, London, UK. Schmitt, R. J. and S. J. Holbrook. 1999. Settlement and recruitment of three damselfish species: larval delivery and competition for shelter space. Oeco- logia 118:76-86. Seiwa, K., M. Tozawa, N. Ueno, M. Kimura, M. Yamazaki, and K. Matuyama. 2008. Roles of cottony hairs in directed seed dispersal in riparian willows. Plant Ecology 198:27-35. Sher, a. a., D. L. Marshall, and J. P. Taylor. 2002. Establishment patterns of native Populus and Salix in the presence of invasive nonnative Tamarix. Ecological Applications 12:760-772. SOKAL, R. R. and F. j. Rohlf. 1995. Biometry. W. H. Freeman and Company, New York, NY. Sousa, W. P., P. G. Kennedy, B. J. Mitchell, and B. M. Ordonez. 2007. Supply-side ecology in mangroves: do propagules dispersal and seedling establishment explain forest structure? Ecological Monographs 77:53-76. Stella, J. C., J. J. Battles, B. K. Orr, and J. R. McBride. 2006. Synchrony of seed dispersal, hydrology and local climate in a semi-arid river reach in California. Ecosystems 9:1200-1214. Stromberg, j. C., R. Tiller, and B. D. Richter. 1996. Effects of groundwater decline on riparian vegetation of semiarid regions: the San Pedro River, Arizona, USA. Ecological Applications 6:113-131. Underwood, A. J. and P. G. Fairweather. 1989. Supply-side ecology and benthic marine assem- blages. Trends in Ecology and Evolution 4:16- 20. U.S. Fish and Wildlife Service. 1994. Endangered and threatened wildlife and plants: designation of critical habitat for the least Bell’s vireo. Final Rule. Federal Register 59:4845^867. . 2005. Designation of critical habitat for the Southwestern Willow Flycatcher {Empidonax trail- lii extimus): Final Rule. Federal Register 70: 60886-61009. Van Splunder, L, H. Coops, L. Voesenek, and C. Blom. 1995. Establishment of alluvial forest species in floodplains: the role of dispersal timing, germination characteristics and water level fluctu- ations. Acta Botanica Neerlandica 44:269-278. Whisenant, S. G. 1999. Repairing damaged wildlands: a process-oriented, landscape-scale approach. Cambridge University Press, Cambridge, UK. Willson, M. F. 1983. Plant reproductive ecology. John Wiley and Sons, New York, NY. Young, J. A. and C. G. Young. 1992. Seeds of woody plants in North America. Timber Press, Portland, OR. Zaret, T. M. and E. P. Smith. 1984. On measuring niches and not measuring them. Pp. 127-137 in T. M. Zaret (ed.). Evolutionary ecology of neo- tropical freshwater fishes. Dr. W. Junk Publishers, The Hague, The Netherlands. Zedler, j. B., C. S. Nordby, and B. E. Kus. 1992. The ecology of Tijuana Estuary, California: a National Estuarine Research Reserve. NOAA Office of Coastal Resource Management, Sanctu- aries and Reserves Division, Washington, DC. Madrono, VoL 61, No. 1, pp. 64—76, 2014 IDENTIFICATION AND TAXONOMIC STATUS OF CORDYLANTHUS TENUIS SUBSP. PALLESCENS (OROBANCHACEAE) Barbara L. Wilson^ Richard E. Brainerd, and Nick Otting Carex Working Group, 1377 NW Alta Vista Drive, Corvallis, OR 97330 Brian J. Knaus Pacific Northwest Research Station, USDA Forest Service, 3200 Jefferson Way, Corvallis, OR 97331 Julie Kierstead Nelson Shasta-Trinity National Forest, 3644 Avtech Parkway, Redding, CA 96002 Abstract Cordylanthus tenuis subsp. pallescens. Pallid Bird’s-beak, is a rare plant of the Mount Shasta area of northern California. Recent reports of dozens of populations outside its limited expected range, observations of plants with morphology intermediate to other subspecies of C. tenuis, and populations that seemed to include individuals of more than one subspecies, raised questions about its rarity and taxonomic validity. Examining populations in the field suggested that many reported populations were misidentified because they were based on a single trait, often foliage color. The name C. t. subsp, pallescens should be restricted to populations in which all or most plants have the combination of traits expected of this taxon, including yellow-green foliage, four to eight flowers per cluster, and short, mostly non-glandular calyx hairs. In the context of other variation recognized at the subspecies level in C. tenuis, recognizing C. t. subsp. pallescens taxonomically is a reasonable choice, despite its very limited range. Key Words: Cordylanthus tenuis subsp. pallescens, intraspecific identification, Orobanchaceae, plants with legal protection, rare plants. Cordylanthus tenuis A. Gray is a widespread, variable species of the California Floristic Prov- ince. Like its congeners, it is an annual hemi- parasite that blooms in mid- to late summer. Host plants of Cordylanthus Nutt, ex Benth. species are usually trees and shrubs, though herbaceous plants can also be parasitized (Chuang and Heckard 1971). The inflorescence is complex. Each flower is partly hidden between the spathe-like calyx and a similar inner floral bract, and subtended by entire or three-parted (trifid) leaf-like outer bracts (Chuang and Heckard 1976). Morphological variation within C. tenuis has led to the recognition of six intergrading subspe- cies (Chuang and Heckard 1986). Five of them had previously been recognized at the species level. Two of the subspecies, C t. subsp. pallescens (Pennell) T. 1. Chuang & Heckard and C t. subsp. capillar is (Pennell) T. I. Chuang & Heckard, are rare California endemics on CNPS List IB. 2 (rare, threatened, or endan- gered), the former with state rank Sl.l (critically imperiled; CNPS 2011) and the latter federally listed as Endangered (U.S. Department of the Interior, Fish and Wildlife Service 1995). Cordy- lanthus tenuis subsp. pallescens was described as endemic to the Mount Shasta region of Siskiyou ‘ bwilson@peak.org County, California, from a very small area near the southern base of Black Butte. Two other subspecies also occur in the Mount Shasta region: C t. subsp. tenuis and C. t. subsp. viscidus (Howell) T. 1. Chuang & Heckard. Efforts to protect C. t. subsp. pallescens became confused when approximately 40 popu- lations were reported from the area south of Dunsmuir to the north side of Mount Shasta and east into the Eddy Mountains (CNPS 2011). These populations were found to be morpholog- ically diverse, potentially confounding attempts to identify populations with legal status. An additional population was reported from Lake County (vouchered by Isle 1704, at CHSC). These reports raised questions. Are all the reported C. t. subsp. pallescens populations correctly identified? Is C. t. subsp. pallescens distinct enough to consider it a valid taxon? Is C. t. subsp. pallescens too widespread and common to protect as a rare plant? What are its true geographic range and substrate fidelity? To address these questions, putative populations of C t. subsp. pallescens were visited in the field in 2010 and herbarium specimens were examined. Methods Twenty-four Cordylanthus tenuis populations were sampled at peak flowering time, 2-10 ii Table 1. Populations of Cordylanthus Tenuis Sampled in 2010. n = population size, determined by counting plants in the smaller populations and estimated in the larger populations, rpt = reported as C. tenuis subsp. pallescens (CNPS 2011). 2014] WILSON ET AL.: CORDYLANTHUS TENUIS SUBSP. PALLESCENS 65 0^ I S •a ^ s s I <0 (U g §1 S I Co •s g -s .£2 2 ^ -a .'<1 ^ ^ Qu) 'ii ^ 'ij ^ Co Co .'o Co .Co om^o^ooo^ooooooooooomfNo o^icn^^^OMotnoooooooM^oooasooo ■^m'^cNinooOL- oov^ONOOinroicn^m-— imo rn mcNL'^oo i> ^ 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 o O o O o O O O o o o O O o O O O O O o O O >> Oi b M M 13 M 13 M 13 13 ‘-3 22 13 13 13 M 13 13 CO Cd c« CO CO Vi Vi _c« Vi ^Vi .Vi .Vi .Vi .Vi _w Vi .Vi _co .2 .Vi .c^ .Vi GO m CO CO CO CO CO CO CO CO w io CO CO io io CO CO io io CO a IT) o ol 00 NO i> o o o o o ON O' m O o in o o o m 00 m lO >— ( NO r—t I— H ON O' NO NO m in o 1—1. NO m o > o ON NO 00 CN in m r- 04 •r—i o NO ON •tT -— ( NO o m tT m m m m ro in CN (N m CO m 'It 04 m 'si- (D oo m CM i> ON NO 00 NO 04 m 04 oo in m NO m NO 00 'd NO r- oo CN oo CN ON ON m O' f—i ON CN in m 00 m ON NO m m IT-- NO O < m o r-- ON 04 NO in -—1 00 04 in m o o O' NO m m m rn ■^. CN CN ON CO rn CO O^ ro 'sf cn m cn rn < 'W) (N (N oi (N (N CN CN CN CN ri ri a CN 04 oi a a (N CN 04 04 CN CN (N (N CN CN n CN CN CM (N CN CN CN 04 CN CN CN 04 CN 04 CN 04 04 CN o 7 T T T T 7 T T T T T T T T T T T T T T T 0) NO r-- m ro m in 00 ■p" O' CN CN NO ON CN ON m x r- NO NO m oo I— 1 NO 00 in oo s CN NO O' O' m ON »— i 7j- ON O p o IT) ( o r- in CN n o NO o CN 00 O' O' O' CN NO NO L' Vd 00 00 NO O On o 00 NO ON in NO o ON CN O' in ON 04 m "a ro ro CN rn < < < 04 04 04 04 04 CN 'f CN cn m < < < cd a L At a i~4 a a L a a L L a a a ON , i a -4 rL H-1 m p- NT •P- &D bO bObObObOMbObObO&ObObObO&ObObOibObDbObObO&i}^ ;3 <<<<<<<<<<<<<<<<<<<<a5asONON’--i»--i fi •2 a S cd ^ O oj X) fi •3 m u .s o Os m o o 2 >7 bOco S .M 5 W T3 d S o p^ O P^ o a u, O o •aO o s 73 p^ ^ W.2> m .g ^ afe GO Z X .-g W W ^ o GO W a a :5; cd a g .tfSS Cd ^ g a o 0)0^ ^ w -rt ■o S 9 c5 C P^ d ^ S cd cd Q 00 5 a g* 3-3^ ^ « S a2 s c '2 c §0C u 1j M ^ (U S >^G0 a^^ -2 S o « « X ^ ^ 0(Nm'!d-ir)Nor^oooNOcNim*ONor^oooNO -HCNm^NOON— i--(fSi(N3 flowers/cluster 10.95608 2 0.004178 0.010753 Bracts entire 9.712409 2 0.00778 0.014004 Callus tips present 3.241385 2 0.197762 0.197762 Tips expanded 7.124084 2 0.028381 0.03649 Calyx maroon 12.0913 2 0.002368 0.010753 No long hairs 10.68701 2 0.004779 0.010753 Glands sparse 8.690403 2 0.012969 0.019453 Short hairs dense 5.982017 2 0.050237 0.056516 70 MADRONO [Vol. 61 Fig. 3. Non-metric multidimensional scaling axes 1 (x axis) and 2 (y axis) for Cordylanthus tenuis. Black = C. t. subsp. pallescens, white = C t. subsp. tenuis, and light gray = C. t. subsp. viscidus. X = ambiguous population. populations overlapped some individuals from populations assigned to other taxa, and plants from the Weed populations (site 28) were scattered (results not shown). Because of concern that the scoring of foliage color in some individuals was influenced by the color of most plants in the population, this NMS was repeated omitting that trait from analysis (Fig. 4). With Table 4. Factor Loadings of Traits on Axes for NMS OF Cordylanthus Tenuis Individuals, Omitting Foliage Color. Axis 1 Axis 2 Flowers/cluster (maximum) 0.14671 0.15412 Bracts trifid 0.06409 0.37117 Bracts entire -0.06017 -0.22096 Bract tips callused 0.06596 0.16508 Bract tips expanded 0.1275 0.36158 Calyx maroon -0.26302 -0.25126 Long calyx hairs present -0.27545 -0.16508 Short calyx hairs (density) 0.19569 0.06919 Calyx glands (density) -0.19351 -0.11496 foliage color omitted, individuals from the core and Weed populations clustered together, though they were not strongly differentiated from the large cluster of plants from other populations, identified as C t. subsp. tenuis, C. t. subsp. viscidus, and intermediates. Axis 1 reflected mainly variation in the calyx (color, and density of hairs and glands) while axis 2 reflected mainly variation in bracts (shape and tips) and calyx color (Table 4). Cordylanthus tenuis subsp. pallescens popula- tions were found along roads (sites one and 25) and in openings in shrub-dominated habitats (sites two and nine) on excessively drained volcanic soils. They were not observed in microsites that -1.5 -1 -0.5 0 0.5 1 1.5 # C.t. pallescens • C.t. pallescens (Weed) □ C.t. tenuis 0 C.t. tenuis (Hotium) ^ C.t. viscidus A C. t. tenuis/viscidus o C.t. tenuis/pallescens Axis 1 Fig. 4. Non-metric multidimensional scaling for Cordylanthus tenuis individuals omitting foliage color. 2014] WILSON ET AL.: CORDYLANTHUS TENUIS SUBSP. PALLESCENS 71 had thick layers of pine needles, even when they grew in adjacent microsites that had bare mineral soil They were not found in fully shaded sites= Observed populations of other C tenuis subspe- cies occurred in similar habitats, though the range of substrates was greater and included serpentine substrates. Nearly every mature C. tenuis flower in the core area and in some of the nearby sites had a hole on the side near the base, where an insect had chewed through the inner bract, calyx, corolla, and ovary wall. Discussion Validity of Cordylanthus tenuis subsp. pallescens as a Taxon The core Cordylanthus tenuis subsp. pallescens populations (sites one, two, nine, and 25) appear to form a geographically coherent metapopula- tion around Black Butte with outlying popula- tions in nearby Weed. These plants probably number in the tens of thousands, growing on a single substrate and characterized by a distinctive combination of morphological traits. They look strikingly different from typical C t. subsp. tenuis and C t. subsp. viscidus, and were originally described as a species, C pallescens (Pennell 1947). Because some morphologically intermedi- ate plants occur, this taxon was reclassified as a subspecies of C. tenuis (Chuang and Heckard 1986). The reported existence of large numbers of plants with intermediate morphology led to the concern that recognizing C t. subsp. pallescens taxonomically might be inappropriate (L. W. Heckard, personal communication). Defining the concept of the species has challenged taxonomists for decades. Many solu- tions have been proposed (e.g., Mayr 1963; Sokal and Crovello 1970; Wiley 1978; Nixon and Wheeler 1990; Baum and Shaw 1995; Coyne and Orr 2009) but resolution has not been achieved, in part because the need to name species reflects both the pattern of biodiversity and the human need for neat, mutually exclusive names that match the way human minds classify organisms (Hey 2001; Yoon 2009). Criteria for naming subspecies and varieties are even more vague than criteria for naming species (Haig et al. 2005). Intraspecific names may be applied to incipient species that are morpholog- ically very similar but geographically isolated, or to virtually any grouping within a species that seems identifiable and is of sufficient interest. Intraspecific groupings differing by a single trait presumably caused by a single gene, such as hairiness (e.g., Elymus glaucus [Wilson et aL 2000]), awn length, or flower color are usually not recognized taxonomically. Cordylanthus tenuis subsp. pallescens is identified by at least four dusters of traits: foliage color (which may result from gland density), number of flowers per duster, the density of the various classes of calyx hairs (long non-glandular, short non-glandular, and short glandular hairs), and bract shape (tip traits and perhaps overall trifid/entire shape; Fig. IB). Although these traits occur together in the core C. t. subsp. pallescens populations, they occur separately in nearby populations (Fig. 1 A-C). Therefore, the name C t. subsp. pallescens cannot be dismissed as the inappropriate labeling of a single gene. Arguments can be made for three taxonomic solutions to the C t. subsp. pallescens problem. Option 1: Synonymize C t. subsp. pallescens with one of the other subspecies. It is similar to the other subspecies. Even its most distinctive traits occur at least occasionally in C t. subsp. tenuis, C. t. subsp. viscidus, or both. Clusters of four or more flowers, though otherwise rare in C tenuis, occur in northern populations of C t. subsp. viscidus and in C t. subsp. barbatus T. I. Chuang & Heckard of the southern Sierra Nevada (Chuang and Heckard 1986). Foliage color seems diagnostic in C t. subsp. pallescens, but green plants occur occasionally in popula- tions of the other two C tenuis subspecies in the Mount Shasta area and predominate in the rare C t. subsp. capiiiaris of Sonoma County, California (Chuang and Heckard 1986). Color is known to vary within some related taxa, such as the rare Chloropyron maritimum (Nutt, ex Benth.) A. A. Heller subsp. paiustre (Behr) Tank & J. M. Egger, which is usually brown although individ- uals and even entire small populations may be green (Tom Kaye, Institute of Applied Ecology, Corvallis, Oregon, personal communication). The problem with synonymizing C /. subsp. pallescens with one of the other C. tenuis subspecies is choosing the subspecies with which to synonymize it. The core C /. subsp. pallescens cannot easily be included in either C t. subsp. tenuis or C t. subsp. viscidus because C t. subsp. pallescens combines the trifid bracts of C. t. subsp. viscidus with the short hairs and sparse glands of C. t. subsp. tenuis, together with yellow- green coloration that is rare in the other subspecies (Fig. 2). Option 2: Combine C t. subsp. pallescens, C. t. subsp. tenuis, and C L subsp. viscidus. Combin- ing these three subspecies into a single wide- spread, variable taxon would not only simplify C tenuis taxonomy by removing taxonomic recog- nition from a small series of odd populations, but also solve two other problems in C. tenuis classification. In the Mount Shasta area, some C tenuis plants exhibit traits in combinations not typical of any subspecies, making subspecific identification difficult or impossible (Fig. ID; Rhonda Posey, personal communication). Mor- phologically intermediate plants are also common 72 MADRONO [Vol. 61 in the western foothills of the Sierra Nevada (Lawrence Janeway, personal communication). Merging the three subspecies would also do away with the anomalous range of C t. subsp. tenuis, widespread in the Sierra Nevada and disjunct in the Trinity Mountains (Chuang and Heckard 1986). However, this change could destabilize the other C. tenuis subtaxa including the geographi- cally isolated and federally listed endangered C. t. subsp. capillaris, because although subspecific taxa can be subjective groupings and the degree of difference between them may vary from species to species, one prefers the subspecific taxa within one species to be equivalently distinct. It would be best to study all six C. tenuis subspecies before making such a change. Option 3: Recognize C t. subsp. pallescens as an intraspecific taxon. Subspecific taxa are expected to be a bit fuzzy, to blend in with the relatives at the edges of their range. Individual C t. subsp. pallescens traits, such as yellow-green coloration and large flower clusters, are rare in other C tenuis taxa. The combination of traits characteristic of C. t. subsp. pallescens is rare in populations of the other subspecies even near Mount Shasta and absent or nearly so elsewhere. Cordylanthus tenuis subsp. pallescens seems to be at least as distinctive as any other C. tenuis subspecies currently recognized (Chuang and Heckard 1986). The plants in and near the core C t. subsp. pallescens populations appear mor- phologically similar (this study; Robin Fallscheer, California Department of Fish and Game, personal communication; Robert Hawkins, Tim- ber Products Company, personal communica- tion). Many reports of C t. subsp. pallescens or intermediate populations (CNPS 2011) result from misidentified yellow-green plants of C t. subsp. tenuis or C. t. subsp. viscidus. The only problem with recognizing C t. subsp. pallescens seems to be that it has such a tiny geographic range that the area of intergradation with other subspecies is larger than its core range. The pattern of variation that led Chuang and Heckard (1986) to recognize three subspecies in the Mount Shasta area is real (Fig. 1), even though morphologically intermediate plants do occur. Pending a broader study that includes the three C tenuis subspecies that grow outside the geographic region covered in this study, we recommend recognizing C t. subsp. pallescens at the subspecific level. Identification and Management of C t. subsp. pallescens As interpreted in this study, Cordylanthus tenuis subsp. pallescens is a rare taxon. Rarity can be measured in three basic ways (Rabinowitz 1981). Cordylanthus t. subsp. pallescens clearly has two of the three; it has a very limited range and lives in a limited habitat type. Its total population size may be in the tens of thousands, not extremely small but not large for an annual plant. Cordylanthus tenuis subsp. pallescens is more geographically restricted than some earlier re- ports suggest, extending from just south of Black Butte to the town of Weed in the Southern Cascade Range of Siskiyou County. Much of its range lies on private land managed for timber production (Robin Fallscheer, California Depart- ment of Fish and Game, personal communica- tion). Silvicultural practices applied in this area in recent years may have created favorable condi- tions for this plant, at least for short periods of time, by opening the canopy and disturbing the duff layer. The number of C. t. subsp. pallescens populations appears to have increased over the last twenty years because plants appear in logging roads, along road edges, and on skid roads where they were not previously observed. Once estab- lished, these populations increase exponentially at least for a few years, and sometimes expand into adjacent regenerating clearcuts (Robert Hawkins, Timber Products Company, personal communication) . Cordylanthus tenuis populations commonly grow along roadsides, presumably from seeds moved by road maintenance equipment. It is possible that C t. subsp. pallescens was more geographically isolated and more neatly set off morphologically from other C tenuis populations before extensive road construction allowed the three subspecies present in the Mount Shasta area to expand their ranges and establish contact with each other. Nearly all the mature C. t. subsp. pallescens flowers observed and most of the flowers in other nearby C tenuis populations had been chewed near the base, through the inner bract, calyx, corolla, and ovary wall, perhaps by an insect eating the ovules. The effect of this predation on seed set could not be determined during this study, because the plants were at anthesis. Thinking in terms of populations rather than individuals is important when attempting to find and manage C. t. subsp. pallescens populations. In a C. t. subsp. pallescens population, all or most of the individuals have all these traits: yellow- green foliage and calyx, clusters of four or more flowers, and sparse calyx glands. Their non- glandular calyx hairs consist of a carpet of very short hairs and very few longer hairs that are perhaps two or three times as long as the short hairs. In most populations, some or all of the outer bracts are trifid. The plants of site 28 in the town of Weed seemed to be C t. subsp. pallescens except that they had entire bracts. Bract shape may be less consistent than its use in classifying Cordylanthus species suggests (Chuang and Heckard 1975). 2014] WILSON ET AL.: CORDYLANTHUS TENUIS SUBSP. PALLESCENS 73 Classifying the Weed population as C t. subsp. pallescens is reasonable. Possession of only one C t. subsp. pallescens trait does not make a plant C. t. subsp. pallescens. Thus, a green C. tenuis is not necessarily C t. subsp. pallescens. Similarly, the presence of one individual that might be classified as C t. subsp. pallescens does not make the population C subsp. pallescens or a priority for protection. If most of the plants in the population seem to be C t. subsp. tenuis, C. t. subsp. viscidus, or interme- diate between those two, the population should not be classified as C. t. subsp. pallescens. In such a population, the combination of genes produc- ing the single C. t. subsp. pallescens-like individ- ual would likely be broken up in the next generation, and might reappear again in later generations. The genes producing the more distinctive C t. subsp. pallescens traits seem to be moving out into other C. tenuis populations, probably creating novel combinations that will affect the evolution of the species. This may be an interesting process to watch, but we do not see a reason for humans to try to control it by protecting populations that produce only occa- sional plants that might be identified as C. t. subsp. pallescens. Protecting the populations that are clearly C /. subsp. pallescens (the core populations in this study, plus other populations on nearby Forest Service and private lands) should help assure the continued survival of this taxon. Acknowledgments We thank Rhonda Posey of the Shasta-Trinity National Forest for information and maps. We thank Robin Fallscheer, California Fish and Game, and Robert Hawkins, Timber Products Company, for providing useful information. Tom Kaye, Institute for Applied Ecology, helped with the statistics. Maps were produced by Rachel Schwindt, Institute for Applied Ecology. We thank Richard Halse of Oregon State University Department of Botany and Plant Pathology for managing herbarium loans. This work v/as funded by Shasta-Trinity National Forest. Literature Cited Baum, D. and K. L. Shaw. 1995. Genealogical perspectives on the species problem, Pp. 289- 202 in P. C. Hoch and A. G. Stephenson (eds.), Experimental approaches to plant systematics. Monographs in Systematic Botany from the Mis- souri Botanical Garden, Volume 53, St. Louis, MO. Benjamini, Y. and Y. Hochberg. 1995. Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society, Series B (Methodological) 57:289-300. California Native Plant Society (CNPS). 2011. Inventory of Rare and Endangered Plants (online edition, v8-01a). California Native Plant Society, Sacramento, CA. Website http://www.cnps.org/ inventory (accessed 04 March 2011). Chuang, T. I. AND L. R. Heckard. 1971, Observa- tions on root-parasitism in Cordylanthus (Scrophu- lariaceae). American Journal of Botany 58: 218-228. . 1975. Re-evaluation of bract morphology in taxonomy of Cordylanthus (Scrophulariaceae). Madrono 23:169-236. •. 1976. Morphology, evolution, and taxonomic significance of the inflorescence in Cordylanthus (Scrophulariaceae). American Journal of Botany 63:272-282. . 1986. Systematics and evolution of Cordy- lanthus (Scrophulariaceae-Pediculariaceae) (includ- ing the taxonomy of subgenus Cordylanthus). Systematic Botany Monographs 10:1-105. . 1993. Cordylanthus. Pp, 1027-1030 in J. C. Hickman (ed.), The Jepson Manual: higher plants of California. University of California Press, Berkeley, CA. Coyne, J. A. and H. A. Orr. 2004. Speciation. Sinauer Associates, Sunderland, MA. Haig, S. M., E. A. Beever, S. M. Chambers, H. M. Draheim, B. D, Dugger, S. Dunham, E. Elliot-Smith, j. B. Fontaine, D. C. Kesler, B. J. Knaus, I. F. Lopes, P. Loschl, T. D. Mullins, and L. M. Sheffield. 2006. Taxonom- ic considerations in listing subspecies under the U.S. Endangered Species Act. Conservation Biol- ogy 6:1584-1594. Hey, Jody. 2001. Genes, categories, and species: the evolutionary and cognitive causes of the species problem. Oxford University Press, Oxford. Mayr, E. 1963. Animal species and evolution. Belknap Press of Harvard University Press, Cambridge, MA. McCune, B. and M. j. Mefford. 2006. PC-ORD, multivariate analysis of ecological data, version 5.10. MjM Software, Gleneden Beach, OR. Nixon, K. C. and Q. D. Wheeler. 1990. An amplification of the phylogenetic species concept. Cladistics 6:211-223. Oksanen, j., F. G. Blanchet, R. Kindt, P. Legendre, R. B. O’Hara, G. L. Simpson, P. Solymos, M. Henry, H. Stevens, and H. Wagner. 2011. vegan: Community Ecology Pack- age. R package version 1.17-8. Website http:// CRAN.R~project.org/package=vegan (accessed 6 March 2011). Pennell, F. W. 1947. Some hitherto undescribed Scrophulariaceae of the Pacific States. Proceedings of the Academy of Natural Sciences of Philadel- phia 99:155-199. R Development Core Team. 2010. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Website http://www.R-project.org/ (ac- cessed 6 March 2011). Rabinowitz, D. 1981. Seven forms of rarity. Pp. 205- 217 in H. Synge (ed.), The biological aspects of rare plant conservation: proceedings of an international conference held at King’s College Cambridge, England, 14-19 July 1980. Wiley Press, New York, NY. SOKAL, R. R. AND T. Crovello. 1970. The biological species concept; a critical evaluation. American Naturalist 104:127-153. 74 MADRONO [VoL 61 Tank, D. C., J. M. Egger, and R. G. Olmstead. 2009. Phylogenetic classification of subtribe Cas- tilleiinae (Orobanchaceae). Systematic Botany 34(1):182-197. Thiers, B. (continuously updated). Index Herbar- iorum: a global directory of public herbaria and associated staff. New York Botanical Garden’s Virtual Herbarium. Website http://sweetgum.nybg. org/ih/ (accessed 20 March 2011). U.S. Department of the Interior, Fish and Wildlife Service. 1995. Endangered and threat- ened wildlife and plants; determination of endan- gered status for ten plants and threatened status for two plants from serpentine habitats in the San Francisco Bay Region of California. Federal Register 60:6671-6685. Venables, W. N. and B. D. Ripley. 2002. Modern applied statistics with S. Fourth Edition. Springer, New York, NY. Wiley, E. O. 1978. The evolutionary species concept reconsidered. Systematic Zoology 27:17-26. Wilson, B. L., J. Kitzmiller, W. Rolle, and V. D. Hipkins. 2000. Isozyme variation and its environ- mental correlates in Elymus glaucus from the California Floristic Province. Canadian Journal of Botany 79:139-153. Yoon, C. K. 2009. Naming nature: the clash between instinct and science. W.W. Norton & Company, New York, NY. Appendix 1 Specimens examined. Herbarium acronyms follow Thiers (continually updated) except that STNF = Shasta-Trinity National Forest. The Wilson and Brainerd specimens cited were collected for this study and recently distributed to the herbaria cited. Cordylanthus tenuis subsp. pallescens. USA. Califor- nia. Siskiyou Co.: just s Weed Elementary School (s side of Hilltop Dr, less than 0.1 mi e of S. Davis Ave.), Weed NW Quadrangle, 3520 ft, 1 Aug 1990, Burk DB17 (JEPS); near sw base Black Butte (along hwy 5), 3500 ft, 21 Jul 1975, Chuang and Chuang 7546 (JEPS); sw of Black Butte, nw of Mt Shasta City, 41.33°N 122.35°W, 3500 ft, 21 Aug 1969, Chuang and Heckard 6741 (CHSC, JEPS); w Mt. Shasta City (se of Black Butte, along garbage disposal road just off hwy 5 [old 99]), 3850 ft, 3 Jul 1968, Heckard and Chuang 1972 (JEPS); ca. 2 mi ne Mt. Shasta City (between old hwy 99 and new freeway), 15 Aug 1971, Heckard and Chuang 2774 (JEPS); 4 mi nw of City of Mount Shasta (sw of Black Butte, along Garbage Disposal Road just off Highway 5 [old 99]), Garbage Disposal Rd, 3850 ft, 12 Aug 1968, Heckard et al. 2122 (JEPS); ca. 4 mi nw Mt. Shasta City, s of Black Butte, 0.1 mi from road parallelling old US Hwy 99, along Pony Trail Dr, near Shasta Abbey (Zen Mission), 3900 ft, 4 Aug 1972, Heckard 3150 (JEPS); n Mt. Shasta City (just sw of Black Butte, corner of old Hwy 99 and Pony Tail Dr.), 3000 ft, 16 Aug 1979, Heckard 5157 (JEPS); near Miner’s Peak ( = Black Butte, w side of Mt. Shasta), Upton Plateau, 1 Jul 1894, Jepson 21345 (JEPS); N of Mt. Shasta City, Siskiyou Co., 19 Jul 1940, Pennell 26184 (UC); Shasta- Trinity National Forests, se of County Dump site, 41.35°N 122.34°W, 1200 m, 2 Aug 1994, Ulloa~Cruz 42 (CHSC); road to trailhead for Black Butte Trail; study site 1, 4375 ft, 2 Aug 2010, Wilson and Brainerd 16186 (DAV, OSC, STNF, UC); s of Black Butte, e of the dump, on Shasta-Trinity National Forest property; UTM zone lOT, 555437 E, 4577898 N, study site 2, 3 Aug 2010, Wilson and Brainerd 16187 (CHSC, OSC, STNF, UC); Summit Drive just s of junction with Sunrise Drive, n of Abrams Rd, w edge of Interstate Highway 5, s of Black Butte; study site 9, 41.3471 5°N 122.34889°W, 3855 ft, 4 Aug 2010, Wilson and Brainerd 16201 (OSC, STNF); road to trailhead for Black Butte Trail; study site 25, 41.37442°N 122.33354°W, 4441 ft, 9 Aug 2010, Wilson and Brainerd 16280 (CHSC, DAV, OSC, UC); town of Weed; above the asphalt play- ground at the elementary school, below Hillside Drive, e of the parking lot; study site 28, 41.42691°N 122.37964°W, 3635 ft, 9 Aug 2010, Wilson and Brainerd 16290B (CHSC, STNF, UC). Cordylanthus tenuis subsp. pallescensiviscidus. USA. California. Siskiyou Co.: Weed, 3200 ft, 26 Aug 1914, Heller 11723 (OSC, UC). Cordylanthus tenuis subsp. tenuis. USA. California. Eldorado Co.: at lower end of Tamarack Trail, Glen Alpine Canyon, just sw of Leaf Lake, 7000 ft, 16 Aug 1952, Bacigalupi 930 (OSC). Lassen Co.: on road from Janesville to Thompson Peak, Sierra Nevada, Diamond Range, 5700 ft, 31 Jul 1973, J. T. Howell and True 50019 (JEPS, OSC). Plumas Co.: SW of Bucks Lake (NE of Frenchman Hill), 5600 ft, 28 Jul 1994, Ahart 7529 (JEPS). Siskiyou Co.: 2.2 mi n Scott Mt. summit (on Hwy 3, at road junction), 4500 ft, 26 Jul 1987, Ertter 7325 (OSC, UC); Wagon Creek on E side of Mt Eddy, ca 4 airmiles W. of Mt Shasta City, 5800 ft, 14 Jul 1990, Ertter 9348 (UC); e side of Mt. Eddy, 4500 ft, 28 Aug 1914, Heller 11744 (OSC, UC); Shasta-Trinity National Forest, along road 40N26 w of Lake Siskiyou; T32N R6W S25; study site 3, 41.28563°N 122.37418°W, 3932 ft, 3 Aug 2010, Wilson and Brainerd 16190 (OSC); Shasta-Trinity National Forest, along road 40N26 w of Lake Siskiyou; T32N R6W S25; study site 3, 41.28563°N 122.37418°W, 3932 ft, 3 Aug 2010, Wilson and Brainerd 16191 (OSC); Forest Service Rd 41N26, the Eddy Creek Rd, 2.8 mile above the junction with Old Stagecoach Rd; study site 11, 41.39273°N 122.47898°W, 4200 ft, 4 Aug 2010, Wilson and Brainerd 16205 (CHSC, STNF); Forest Service Rd 41N26, the Eddy Creek Rd, 2.8 mile above the junction with Old Stagecoach Rd; study site 11, 41.39273°N 122.47898°W, 4200 ft, 4 Aug 2010, Wilson and Brainerd 16206 (OSC, UC); Old Stagecoach Rd, about 0.5 mile n of Forest Service Rd 41N26, the Eddy Creek Rd; study site 12, 41.41280°N 122.43398°W, 3267 ft, 4 Aug 2010, Wilson and Brainerd 16210 (CHSC, DAV, OSC, UC); Ney Springs Rd e of Lake Siskiyou; intermittent along road starting 0.6 miles e of Rd 20M020 and extending at least 0.6 mile; study site 19, 4L26872°N 122.31984°W, 3060 ft, 6 Aug 2010, Wilson and Brainerd 16233 (OSC); on road 26 between Lake Siskiyou and Gumboot Lake, at a junction about 3/4 mile uphill of Rd 40N43; study site 23, 41.2271 UN 122.45552°W, 4953 ft, 8 Aug 2010, Wilson and Brainerd 16255 (CHSC, DAV, STNF, UC); Shasta-Trinity National Forest; road A 10 between town of Mount Shasta and Panther Meadow; study site 26, 41.35249°N 122.30864°W, 441 1 ft, 9 Aug 2010, Wilson and Brainerd 16281 (CHSC, DAV, STNF, UC); Shasta-Trinity National Forest near Hotlum, along dirt road paralleling the railroad track, between Rd 42N48 and the trestle, n of Mount Shasta; study site 27, 4L49153°N 122.30346°W, 3900 ft, 9 Aug 2010, Wilson and Brainerd 16290 (STNF); Stewart Springs Rd 1.5 miles from the Gazelle Rd; study site 29, 2014] WILSON ET AL.: CORDYLANTHUS TENUIS SUBSP. PALLESCENS 75 41. 4360 TN 1 22.4668 rw, 9 Aug 2010, Wilson and Brainerd 16291 (CHSC, DAY, STNF, UC); Forest Service Rd 41N26, the Eddy Creek Rd, 0.6 mile above the junction with Old Stagecoach Rd; study site 10, 41.40562°N 122.44258°W, 3411 ft, 4 Aug 2010, Wilson and Brainerd 1 6202 B (OSC); Shasta-Trinity National forest, on Stewart Springs Rd (Forest Service Rd 17) 5.3 miles from the Gazelle Rd, study site 30, 41.42462°N 122.51649°W, 4281 ft, 9 Aug 2010, Wilson and Brainerd 16291 B (CHSC, DAY, UC). unknown Cou Sierra Nevada Mountains, 19 Sep 1882, Pringle s.n. (OSC). Cordylamthus tennis subsp. tenuisipaliescens interme- diate. USA, California. Siskiyou Co.i Shasta-Trinity National Forest; n of Mount Shasta itself, near the w end of Forest Service Rd 43N18, study site 13, 41.49203°N 122.20539°W, 5400 ft, 4 Aug 2010, Wilson and Brainerd 16214 (STNF); Shasta-Trinity National Forest; n of Mount Shasta itself, near the w end of Forest Service Rd 43N18, study site 13, 41.49203°N 122.20539°W, 5400 ft, 4 Aug 2010, Wilson and Brainerd 16216 (OSC). Cordylanthus tenuis subsp, tenuhlviscidus, USA. California. Butte Co,: Forbestown Rd, 39.49799°N 12L32413°W, 2647 ft, 7 Aug 2010, Wilson and Brainerd 16244 (CHSC, DAY, OSC, UC), Siskiyou Co.: Shasta- Trinity National Forest, along Rd 40N43 w of Lake Siskiyou; at first creek crossing w of junction with road 26; study site 4, 41.28133°N 122.36802°W, 3488 ft, 3 Aug 2010, Wilson and Brainerd 16192 (OSC); Duns- muir, road 1M036 (railroad Park Rd) across form the railroad museum, study site 15, 41.18648°N 122.29558°W, 2310 ft, 5 Aug 2010, Wilson and Brainerd 16221 (CHSC, DAY, OSC, STNF, UC), Cordylanthus tenuis subsp. viscMus. USA. California. Butte Co.: 2.2 mi sw of and below Nimshew (along Humbug Rd on the eastern canyonside of Butte Creek), Butte Creek, 2000 ft, 26 Jul 1957, Bacigalupi and Whisler 6038 (JEPS); Butte Meadows, 4500 ft, 26 Jul 1957, Bacigalupi and Whisler 6045 (JEPS); 0.2 mi below junction of Ridge Rd on Humbug Rd (near Nimshew [nw of Magalia]), 2100 ft, 20 Aug 1969, Chuang and Heckard 6738 (JEPS); Jonesville, 1650 m, 29 Jul 1929, Copeland 469 (OSC, UC); Nimshew (site), nw of Magalia, 2800 ft, 15 Aug 1971, Heckard and Chuang 2766 (JEPS); South entrance Manzanita Street (to Stirling City), 3500 ft, 15 Aug 1971, Heckard and Chuang 2767 (JEPS); 1.5 mi w Butte Meadows (at Thatcher Ridge Rd junction), 4200 ft, ^ 15 Aug 1971, Heckard and Chuang 2768 (JEPS); Little Summit, 4800 ft, 22 Jul 1914, Heller 11586 (UC). Del Norte Co,: divide between Smith River and south fork of Smith River, French Hill, 17 Aug 1927, Applegate 5244 (UC); Old Gasquet Toll Rd, Danger Point Ridge top Danger Point, 1425 ft, 19 Jun 1975, Barker 957b (JEPS); French Hill Rd, 3.3 miles up from junction with Hwy 99 w of Gasquet; T17N R2E S31, 1700 ft, 28 Jul 1983, Chambers 5100 (OSC); up the Illinois River beyond Takilma, 24 Aug 1929, Henderson 11530 (OSC); State line n of Monumental, 3100 ft, 9 Jul 1940, Tracy 16683 (JEPS, UC); Summit Valley Mountain, 3500 ft, 4 Jul 1938, van Deventer 130 (JEPS). Eldorado Co.; 1.7 mi ne Georgetown, 2800 ft, 20 Aug 1969, Chmmg and Heckard 6732 (JEPS). Humboldt Co.: Grouse Creek, 1 Aug 1888, Chesnut s.n. (UC); Pine Point Ridge sw Mt. Lassie Lookout, North Coast Ranges, 14 Jul 1949, Hoffman 3063 (JEPS, UC); Koby Ranch, on road between Willow Creek (on Three Creeks), Koby Ranch, 2000 ft, 20 Jul 1924, Tracy 6735 (JEPS, UC); Salmon Summit, 5800 ft, 30 Jul 1935, Tracy 14354 (UC); South Fork of Trinity River (mountain slopes w of its mouth), 1000 ft, 14 Aug 1938, Tracy 16118 (JEPS, UC); Grouse Mountain, 5000 ft, 16 Aug 1939, Tracy 16418 (JEPS, UC); ridge top 1 mi w Mud Springs, Trinity Summit, 3800 ft, 24 Aug 1947, Tracy 17884 (UC); Mud Springs Trinity Summit, 4400 ft, 4 Aug 1949, Tracy 18411 (UC); Mud Springs Trinity Summit, 4400 ft, 4 Aug 1949, Tracy 18412 (UC); Mud Springs Trinity Summit, 4400 ft, 4 Aug 1949, Tracy 18413 (UC). Lake Co.: Crooked Tree Ridge Snow Mt, (nw side of mt,), 5400 ft, 7 Jul 1982, Heckard and Hickman 5965 (JEPS); Mendocino National Forest, North Coast Ranges, Rice Fork Watershed, On Mason Trail to Poges Peak on block spur just n of dear cut, 3120 ft, 28 Jul 1999, Isle 1407 (CHSC). Mendocino Co.: just ne Ham Pass (ca 12 air mi ne of Covelo), Pine Ridge, 4900 ft, 5 Sep 1975, Heckard and Ashton 4115 (JEPS); Modoc Co,: Big Valley Mts., n.d., Nutting s.n. (UC). Plumas Co.: SW of Bucks Lake (NE of Frenchman Hill), 5600 ft, 28 Jul 1994, Ahart 7529 (JEPS); e and above Squirrel Creek (ne corner of jet 25N42 and 401, ca 3 air mi n of Spring Garden on Hwy 70), Spring Garden quad, 4500 ft, 21 Jul 1981, M A Taylor 4081 (JEPS). Shasta Co,: Lassens Peak, 1 Aug 1896, Austin 411 (UC); Pinewood Bear Valley Mts., 6000 ft, Jun-Aug 1893, Baker s.n. (UC); e of junction with Inwood- Whitmore Rd (along route 44), 3000 ft, 21 Jul 1975, Chuang and Chuang 7543 (JEPS); 4.5 mi n Pollard Flat (Highway 5), 21 Aug 1969, Chuang and Heckard 6740 (JEPS); along Route 44 0,4 mi e of juction with Inwood- Whitmore Rd Route 44, 2900 ft, 15 Aug 1971, Heckard and Chuang 2771 (JEPS); 5 mi se Siskiyou County Line (Hwy 89 ca 1 mi se Dana Rd [A- 19] jet), 1340 m, 15 Jul 1990, Heckard and Chuang 6765 (JEPS); A mi w Dana Rd Junction (along U.S. Highway 89), 4000 ft, 3 Aug 1971, Heckard and Rubtzoff 2700 (JEPS); 0.2 mi w of junction with McArthur Rd (along highway 89), 4000 ft, 13 Aug 1968, Heckard et al 2125 (JEPS); 2 mi e of junction with road to Inv/ood (along Highway 44, around Northern Pines Motel), 3200 ft, 14 Aug 1968, Heckard et al 2129A (JEPS); along trail to Castle Dome near junction of trail to Indian Springs (Castle Crags State Park), Castle Crags State Park, 4700 ft, 14 Aug 1967, Heckard 1733 (JEPS); Upper Fall River Valley, 3400 ft, 11 Aug 1914, Jepson 5773 (JEPS); n Lamoine (along Shotgun Creek), Shotgun Creek, 1700-2000 ft, 19 Jul 1940, Pennell 26189 (JEPS, UC); 3 mi s Hat Creek (Route 89), 3800 ft, 20 Jul 1940, Pennell 26208 (UC); 5.6 mi n Pollard Flat Guard Station (on U.S. Highway 99), 5 Sep 1956, Raven 10472 (JEPS); Lassen National Forest; Red Mountain, along road 34N19, 40.77773°N 121.53062°W, 4706 ft, 6 Aug 2010, Wilson and Brainerd 16235 (CHSC, DAY, OSC, STNF, UC). Siskiyou Co.: 1/2 mi w Hamburg (along hwy following Klamath River), 1500 ft, 12 Aug 1954, Bacigalupi et al 4792 (JEPS); on road from Callahan to Carrville (1.2 mi above [and s of] junction with Gazelle road, lower part of grade over Scott Mt.), 3600 ft, 13 Aug 1954, Bacigalupi et ai. 4795 (JEPS); 0.3 mi below Gumboot Creek crossing (both sides of rd 40N26), The Eddy’s, 5600 ft, 18 Aug 1990, Burk s.n. (JEPS); w Dunsmuir (n side of road 40N26, ca. 1.8 mi below South Fork Sacramento River Bridge), Shasta National Forest, 6000 ft, 3 Nov 1990, Burk s.n. (JEPS); near Shackelford Creek, Quartz Valley, 11 Aug 1908, Butler 25 (UC); s side Dunsmuir, 2500 ft, 21 Jul 1975, Chuang and 76 MADRONO [VoL 61 Chuang 7545 (JEPS); 1/2 mi w Hamburg (on Hwy 96 above the Klamath River), 1400 ft, 21 Aug 1969, Chuang and Heckard 6742 (JEPS); 27 mi n Happy Camp (Elder Mt Rd junction, on road to O'Brien), 2700 ft, 21 Aug 1969, Chuang and Heckard 6743 (JEPS); Shasta Springs, 2 Sep 1917, Eastwood 6664 (UC); n side Mt. Shasta (along Bolam logging road at e-w fork), 4800 ft, 2 Jul 1968, Heckard and Chuang 1971 (JEPS); 1 mi s of downtown Dunsmuir (below freeway), 2700 ft, 15 Aug 1971, Heckard and Chuang 2773 (JEPS); just a mile over s of downtown Dunsmuir (along Dunsmuir Ave), 2700 ft, 13 Aug 1968, Heckard et al 2123 (JEPS); 13 1/2 mi w Shasta County Line (5 1/2 mi w of Medicine Lake junction, along Hwy 89), 3500 ft, 13 Aug 1968, Heckard et at 2124 (JEPS); 1 5 air mi e Montague (Ball Mt. Rd, nw of Goosenest Mt), 3800 ft, 18 Jul 1978, Heckard et al. 4873 (JEPS); n side Mt. Shasta (along Bolam logging road), 5000 ft, 5 Jul 1967, Heckard 1603 (JEPS); 5.3 mi s Callahan (3 mibelow road summit, along road up to Scott Mt.), 4400 ft, 28 Aug 1971, Heckard 2784 (JEPS); southern edge Dunsmuir (near first Dunsmuir freeway exit), 2700 ft, 4 Aug 1972, Heckard 3149 (JEPS); sw Weed (s of Edgewood, road below Stewart Springs), 3200 ft, 7 Sep 1978, Heckard 4973 (JEPS); s Edgewood (road above Stewart Springs, on Deadfall Meadow-Mt. Eddy Rd.), 4500 ft, 7 Sep 1978, Heckard 4974 (JEPS); 1 mi s of downtown Dunsmuir (below Hwy 5), 2700 ft, 28 Aug 1979, Heckard 5160 (JEPS); w Sisson, Mt. Eddy, 6500 ft, 16 Jul 1918, Heller 13034 (UC); 17.2 mi n of Klamath River, western Cedar Camp Rd, 11 Jul 1950, Hoffman 3720 (JEPS); Area several miles w of Weed, 4L43°N- 122.39°W, 2 Sep 1978, Jokerst 152 (CHSC); 3 mi se Scott Mountain Lodge (ridge between Mill and Mule Creeks), Scott Mts., 1340 m, 8 Aug 1938, Keck 4863 (UC); 5 mi s of summit of road from Happy Camp CA to Waldo, OR, 16 Jul 1950, Mason 14064 (JEPS, UC); North Fork of Sacramento River, 4500 ft, 5 Sep 1956, Raven 10469 (JEPS); mouth of river Northern Coast Ranges, Salmon River, 800 ft, 14 Aug 1920, Tracy 5351 (UC); Dunsmuir; Panorama Drive n of Interstate Highway 5, 4L2008rN 122.27902°W, 2540 ft, 5 Aug 2010, Wilson and Brainerd 16220 (CHSC, DAY, OSC, STNF, UC); Shasta-Trinity National Forest, about 10 miles e of McCloud, at junction of Highway 80 and Forest Service Rd 41N06, 41.2008rN 12L27902°W, 2540 ft, 5 Aug 2010, Wilson and Brainerd 16222 (CHSC, DAY, STNF, UC); North Shore Rd, n side of Lake Siskiyou, 1.2 miles w of cross road; study site 14, 4L29067°N 122.35 152°W, 3270 ft, 6 Aug 2010, Wilson and Brainerd 16231 (OSC); at sign at border of Shasta- Trinity National Forest on Highway 89 e of McCloud; study site 20, 4L20736°N 12L7824°W, 4449 ft, 6 Aug 2010, Wilson and Brainerd 16234 (CHSC, DAY, OSC, UC). Tehama Co.: Pine Creek, ca. 8 miles w of Cottonwood; along Benson Rd, 600 ft, 10 Jun 1997, D. W. Taylor 16049 (JEPS); 1 mi e Lyonsville road junction (on route 36, w of Mineral), 4000 ft, 15 Aug 1971, Heckard and Chuang 2769 (JEPS); ca 0.5 mi from Regan Meadow (along road from Regan Meadow to Brushy Mountain), Shasta Trinity National Forest, 4600 ft, 27 Jul 1979, Nelson et al. 5047 (JEPS). Trinity Co.: on bank above trail to Chloride Mine, Trinity Mountains, Dedrick, 3000 ft, 3 Aug 1948, Alexander and Kellogg 5385 (UC); along road to Stuart Gap 0. 1 mi se jet with Rd. 30 (Mad-Wildwood Rd.), Devils Camp, 1160 m, 28 Jul 1988, Dean 155 (UC); Boulder Creek Basin, along trail ca. 0.25 mile from Canyon Creek, 41°N, 123°W, 5200 ft, 8 Aug 1970, Ferlatte and Howard 275 (OSC); ca 0.25 mi from Canyon Creek (along trail). Trinity Alps, Boulder Cr. Basin, 4 UN 123°W, 5200 ft, 8 Aug 1970, Ferlatte and Howard 1275 (JEPS); about 2 mi se Wildwood (on Peariut-Beegum rd), 3500 ft, 10 Sep 1971, Heckard et al. 2797 (JEPS); 2.7 mi e Hayfork post office, 2400 ft, 10 Sep 1971, Heckard et al. 2800 (JEPS); n Trinity Center (1/2 mi s of junction of road to Castella), 2900 ft, 28 Aug 1971, Heckard 2785 (JEPS); about 4 mi below (s) road summit over Scott Mt., 3800 ft, 28 Aug 1971, Heckard 2784 A (JEPS); about 2 mi s road Junction to Deadfall Lakes (along Scott Mt. -Trinity Center road [about 8 mi s of Scott Mt. road summit]), 3800 ft, 28 Aug 1971, Heckard 2784B (JEPS); Red Mountain Trail, 5000 ft, 20 Aug 1953, Pollard (JEPS); 2 mi e Burnt Ranch, Spellenberg property, 1600 ft, 13 Aug 1965, Spellenberg 1225 (UC); sw Wildv/ood (crest of ridge at head of gulch), Muldoon Gulch, 4200 ft, 19 Aug 1972, Stebbins (JEPS). uiikiiowii Co.: Head of Rush Creek, 5600 ft, 20 Jul 1914, Yates 537 (UC). USA. Oregon. Curry Co,: Sourdough Trail in Lemmingsworth Gulch area; area is reached by For. Serv. roads #3907 and #4014, 42.0159°N-124.0033°W, 15 Jul 1979, Hess s.n. (OSC). Jackson Co,: along Lewis Rd on the ne end of Lost Creek Lake, one mile sw of State Hwy. 62, 42.7009°N-122.6084°W, 555 m, 29 Jul 1981, Hake 2433 (ORE, OSC); along Lewis Rd on the nw end of Lost Creek Lake, one mile sw of State Hwy. 62, 42.7009°N-122.6084°W, 555 m, 29 Jul 1981, Halse 2433 (OSC); about 1 mile w of Prospect, 42,751 1°N- 122.5076°W, 27 Aug 1971, Heckard 2779 (OSC); Along Crater Lake Rd, 14 miles ne of Trail (2 miles w of Laurelhurst State Park), 2100 ft, 4 Aug 1972, Heckard 3151 (OSC); part way up Ashland Butte, s. Oregon, 8 Jul 1886, Henderson 784 (OSC); Ashland, 6 May 1887, Henderson s.n. (ORE, OSC); along Ashland-KJamath Falls road 25 mi. E of Ashland, 22 Aug 1916, Peck 241 3 (WILLU); about 1/3 mile w of confluence of Steve Fork and Sturgis Fork, 42.074rN 123.2196°W, 18 Jul 1991, Rolle 464 (OSC); Butte Falls Resource Area (Medford District Bureau of Land Management), Bieber Wasson. BLM land accessed by gravel road off of Salt Creek Rd. Within BLM fuels treatment unit BW304, 42.4694°N- 1 22.57 rw, 19 Jun 2004, Sikes 130 (OSC); no locality, 13 Jul 1967, Waring 742 (OSC). Josephine Co.: Wimer Rd, 3. 6 mi. SW of O’Brien, at Rock Ck. bridge, 28 Jul 1983, Chambers 5110 (OSC); 1.8 miles sw of O’Brien on Oregon Mountain Rd, 21 Aug 1969, Chuang and Heckard 6744 (OSC); Elder Creek Trail, Siskiyou Forest, 22 Aug 1919, Ingram 1034 (OSC); Illinois Valley, Rough and Ready Creek, 26 Jun 1990, Kagan 6269001 (OSC); Grants Pass, 6 Jul 1909, Peck 2407 (WILLU); mountains near Cal. line, sw of Waldo, 4 Aug 1913, Peck 2408 (WILLU); near Ore-Cal line along Grants Pass-Crescent City Rd, 2 Jul 1918, Peck 8100 (WILLU); Takiima, 8 Jul 1918, Peck 8423 (WILLU). Lane Co.: Young Rk. trail 3685. ca. 60 mi. se of Eugene; ca. 20 mi. se of Hills Crk. Res. Approx. 1/6 mi. from where the trail crosses Spur Rd. #435, 43.5207°N- 122.4005°W, 30 Jul 2005, Harvey s.n. (OSC). Madrono, Vol. 61, No. 1, pp. 77-81, 2014 FOLIAR ANALYSES OF CONIFERS ON SERPENTINE AND GABBRO SOILS IN THE KLAMATH MOUNTAINS Earl B. Alexander Soils and Geoecology, 106 Leland Lane, Pittsburg, CA 94565 alexgeoeco@gmaiLcom Abstract Soils, timber site data, and the foliage of mature conifer trees were sampled at 16 Klamath Mountains sites with serpentine soils and four sites with gabbro soils. The basal areas of trees at the gabbro sites were greater than the basal areas at the serpentine sites. Tree height growth, based on old growth site curves, was significantly greater on the gabbro soils than on the serpentine soils. The main soil difference was about five times greater exchangeable Mg from the serpentine soils than from the gabbro soils. The higher Mg in serpentine soils was not reflected in higher concentrations of Mg in the foliage of yellow pine (Pinus spp.), Douglas-fir (Pseudotsuga menziesii [Mirb.] Franco), or incense cedar {Calocedrus decurrens [Torr.] Florins) trees on serpentine soils, but Ca was significantly higher in the Douglas-fir tree foliage on gabbro soils. The high Mg in the serpentine soils may interfere with the utilization of Ca by Douglas-fir trees causing their foliage to have lower Ca concentrations than in the foliage of Douglas-fir trees on gabbro soils. For each of the three tree species (yellow pine, Douglas- fir, and incense cedar), foliar N, P, and K concentrations were not appreciably different on serpentine versus gabbro soils. Although soil N and P were not determined, organic matter concentrations did not differ between serpentine and gabbro soils. Key Words: Foliar analyses, conifers, gabbro, serpentine, Klamath Mountains, soil chemistry, tree growth. Serpentine and gabbro soils support many kinds of plants that do not grow on other kinds of soils. Many plant species are unique habitants of gabbro soils, not growing on serpentine soils or any other kinds of soils (Oberbauer 1993; Wilson et al. 2010). Most speculation about the special characteristics of gabbro soils that allow unique plants to grow on them has concentrated on the chemistry of the soils (Dayton 1966; Hunter and Horenstein 1992; Alexander 2011; Burge and Manos 2011). The definitive relationships of unique plants to the gabbro soils on which they grow are elusive; the reasons that some plants grow only on gabbro soils are unknown. Foliar analyses might provide some clues. The elemental compositions of plant leaves depend upon many factors, including the availabilities of the elements from the soils which their roots occupy. Because plant leaf compositions commonly reflect the chemical compositions of soils in which they are growing, leaf analyses from the same kinds of plants on gabbro and serpentine soils can be compared to learn what elemental concentration differences in the plants are most representative of differences in the soils that might be responsible for growth and survival differences (Turner et al. 1978). Although there are foliar analyses of plants growing on serpentine soils (e.g., Alexander et al. 2007), there is a dearth of comparable data for plants growing on gabbro soils. In a previous investigation of serpentine soils that included four soils on gabbro (Alexander et al. 1989), leaves were taken from conifers for foliar analyses. Data from the foliar analyses have not been published; they are a source of information that can be utilized to compare differences in the utilization of some plant nutrient elements by different conifer species on serpentine and gabbro soils. Serpentine and gabbro soils, and leaves from trees on the soils, were sampled on unmanaged forested land over the Trinity ultra- mafic body, or Trinity ophiolite, in the Klamath Mountains of California. Although the original interest was timber management (Alexander et al. 1989), data from analyses of the soils and conifer foliage are useful for investigating responses of the conifers to the contents of plant nutrient elements in the soils. The soils data, timber site, and basal area data were reported in Alexander et al. (1989), but the gabbro soils were not differentiated from the serpentine soils, and no foliar analyses were reported for the trees. The current objective is examination of elemental chemical analyses from conifer foliage on 16 serpentine soils and on four gabbro soils of the Trinity ophiolite to learn how differences in plant nutrient element concentrations in the foliage relate to chemical differences in the soils. Area of Investigation The Trinity ophiolite is a large area of predominantly serpentinized peridotite and gab- bro in the eastern part of the Klamath Mountains (Fig. 1). Soils and conifer foliage were sampled at 78 MADRONO [Vol. 61 San Francisco Ecosystem Plot 62 miles 100 km Fig. 1 . Locations of the soil and foliage sampling sites in the Klamath Mountains of California. Serpentine ecosystem sites are represented by open circles. Gabbro ecosystem sites are represented by closed circles. altitudes from 700 to 1920 m on serpentinized peridotite and from 660 to 1700 m on gabbro. The mean annual precipitation, mainly winter rain and snow, ranges from 750 to 1400 mm on the serpentine and from 1250 to 1750 mm on the gabbro soils. The soils were mostly cool (mesic soil temperature regime), moderately deep In- ceptisols and Alfisols (Fig. 2), with some Molli- sols on the serpentinized peridotite. They are in loamy-skeletal, fine-loamy, and clayey-skeletal families (Soil Survey Staff 1999). The overstory was predominantly yellow pine {P. jeffreyi Balf. on serpentine and P. ponder osa P. Lawson & C. Lawson on gabbro soils), Douglas-fir {Pseudo- tsuga menziesii [Mirb.] Franco), and incense cedar {Calocedrus decurrens [Torr.] Florins). Huckle- berry oak (Quercus vaccinifolia Kellogg) was the most common shrub; California coffeeberry {Fragula californica Eschsch. subsp. occidentalis [Howell ex Greene] Kartesz and Gandhi) was common on both serpentine and gabbro soils, even though it is supposedly a serpentine endemic (Safford et al. 2005). Grasses, mainly fescues {Festuca idahoensis Elmer and F. californica Vasey) and wheatgrass {Elymus spicatus [Pursh] Gould) were common on the serpentine soils with the more open overstories and less shrub cover. Methods Soils were sampled from the 0-10 and 10-30 cm depths at three locations on 0.1 acre-ft (0.04 ha) plots; samples from each depth were mixed, air- dried, and passed through a sieve to obtain fine- earth (particles <2 mm) for laboratory analyses. Ages and heights of representative trees were measured and basal areas were measured for all trees on each plot. Current year leaves were sampled from the lower south sides of mature yellow pine, Douglas-fir, and incense cedar trees; three subsamples from each tree group were combined for analyses. Weight loss at 450°C (LOI, loss-on-ignition) from air dry soil was chosen as an indicator of soil organic matter. Exchangeable Ca, Mg, and K were extracted with neutral ammonium acetate and Mn, Fe, and Ni were extracted in Na citrate- dithionite solution (Alexander et al. 1989); the quantities of all cations were recorded by atomic Fig. 2. A moderately deep soil on gabbro that has intruded the Trinity ophiolite. Graduations on the tape are each 10 cm, and the white mark at the bottom is at 75 cm. 2014] ALEXANDER: FOLIAR ANALYSES OF CONIFERS ON SERPENTINE AND GABBRO 79 absorption (AA) spectroscopy. Data for samples from the 0-10 and 10-30 cm depths were combined in 1:2 ratios for the current data analyses. Differences between means for each element in serpentine and in gabbro soils were compared by an unpaired test (Snedecor and Cochran 1967). Current year foliage was dried and pulverized, N was ascertained by the micro-Kjehldahl method, and other elements were determined following perchloric acid digestion of pulverized foliage (Zinke and Stangenberger 1979). Phos- phorus was ascertained by molybdenum blue colorimetry and Ca, Mg, K, and Mn by A A spectroscopy. For each conifer group and foliar element, differences between sample means for serpentine and for gabbro soils were compared by an unpaired t-test (Snedecor and Cochran 1967). Results and Discussion Soil and timber site data. Basal areas of the trees, and standard deviations, were 41.4 ± lie m^/ha on serpentine and 65.0 ±9.1 m^/ha on gabbro soils. A timber site index (TSI) that is based on the presumed heights of trees at 300 yr (Dunning 1942) and reported by Alexander et al. (1989) was significantly higher for trees on four sites with gabbro soils than for the trees on the 16 sites with serpentine soils (Table 1, unpaired Ntest, a < 0.01). The trees measured on the serpentine and gabbro soils were mostly yellow pine, some Douglas-fir, and few white fir {Abies concolor [Gordon & Glend.] LindL ex Hildebr. var. concolor) (C. Adamson, USD A Forest Service, 1989, now retired, personal communication) . Exchangeable Mg was much greater from the serpentine soils than from the gabbro soils, but there were no significant differences for ex- changeable Ca or exchangeable K. Although gabbro soils had more dithionite-citrate extract- able Cr, Mn, and Fe, only the Cr was signifi- cantly greater than from serpentine soils. The amounts of Cr reported in Table 1 are not total amounts, and toxic Cr(VI) was not differentiated from nontoxic Cr(III), which is much, more comm.on in unpolluted soils (Adriano 2001). It is unlikely that any of the first transition elements were toxic to the conifers. Foliage analyses. The Dougias-fir and yellow pine tree data can be compared to analyses for foliage from 82 Dougias-fir and 78 ponderosa pine trees collected throughout the conifer forests of California, Oregon, and Washington (Zinke and Stangenberger 1979). Zinke and Stangenber- ger (1979) produced distribution curves for chemical element concentrations in the Douglas- fir and ponderosa pine tree foliage. Based on those curves, means of data from trees on the 16 cd u e, w ^ c« .S ^ Cfl .."S ^ u S o 5 ^ o 0) .S o 'd o « 2 T ^ s .S S 2 « d p « M O d a « w a S o g O •a <4-4 N- ^ CO K S pq o o V O OT 0) n s ^ s ^ < s 0 g e a H S I- g a V. . o 2 w .2 ^ A 6 > •< flj Z 'O A -d • d d w ^ n H I a E3 e ^ W) VO 00 o o m fN| N A r-4 m d d C\1 00 d A r- VO d r-- OV /-N fS! d o ^ VO ^ w ”1 ° S M ^ 18 df. 14.8** 0.64 0.04 4.39** 0.97 2.11* 1.87 1.07 0.64 0.84 80 MADRONO [VoL 61 Table 2. Comparisons of Elemental Concentrations in First Year Yellow Pine (Ponderosa and Jeffrey Pines) and Douglas-fir Needles from the Soils on th^ Trinity Ophiolite in the Klamath Mountains. Numbers in parentheses are percentiles based on foliar analyses from 78 ponderosa pine and 82 Douglas-fir trees in a variety of habitats (Zinke and Stangenberger 1979). Significant differences at the 95% (a < 0.05) or 99% (a < 0.01) levels of confidence are indicated by one (*) or two (**) asterisks. Tree and substrate n N(g/kg [ppt]) P(g/kg [ppt]) Ca (g/Mg [ppm]) Mg (g/Mg [ppm]) K (g/Mg [ppm]) Fe (g/Mg [ppm]) Mn (g/Mg [ppm]) Yellow pine Serpentine 13 10.6(48) 1.44(67) 1.23(36) 1.52(95) 6.64(58) * 37(24) 72(37) Gabbro 4 10.2(39) 1.40(64) 1.49(41) 1.64(98) 5.96(45) 42(29) 182(77) Douglas-fir Serpentine 13 9.0(18) ** 1.53(63) 2.02(13) 1.94(84) 8.15(79) 76(8) 117(23) Gabbro 4 9.4(22) 1.24(42) 3.99(61) 2.53(95) 7.09(63) 57(1) 505(93) Incense cedar Serpentine 13 9.5 1.17 8.18 3.88 5.88 76 34 Gabbro 3 9.0 0.96 8.57 3.15 6.88 83 72 serpentine and four nonserpentine soils (gabbro) in the Klamath Mountains investigation were given percentile ratings, which are in parentheses in Table 2, Foliar Douglas-fir N was very low on serpentine soils (18th percentile, indicating that only 17% of the trees in the 82-tree sample had lower foliar N contents), and Douglas-fir foliar N was low on gabbro soils (22nd percentile); but N was only marginally low (percentile <50) in yellow pine foliage on both kinds of soils. Magnesium was very high in foliage from all of the yellow pine and Douglas-fir trees and Ca was relatively low in the foliage of Douglas-fir trees on serpentine soils. The lower Ca in foliage of Douglas-fir trees on serpentine soils, compared to the trees on gabbro soils, is reflected in Figure 3, where the Douglas-fir foliage Ca concentrations for three trees on gabbro soils are shown nearer to the Ca apex (higher Ca) than the Douglas-fir foliage Ca contents of trees on serpentine soils. Higher Mn concentrations in the foliage of trees on gabbro soils than in the foliage of trees on serpentine soils (Table 2) was related positive- ly to higher exchangeable Ca/Mg ratios for the gabbro soils (a < 0.01) and to higher timber site index on them (Alexander et al. 1989), rather than to differences in the contents of Mn extractable from soils. Exchangeable Ca was about the same from both serpentine soils (64 mmol+/kg mean) and gabbro soils (65 mmol+/kg mean), but exchangeable Mg was significantly greater (ot < 0.01) from serpentine soils (218 mmol-h/kg mean) than in gabbro soils (41 mmol+/kg mean); consequently, the Ca/Mg ratios were greater in gabbro soils (Alexander et al. 1989). The significantly lower Ca in Douglas-fir foliage from trees on serpentine soils, compared to trees on gabbro (Table 2) may be a result of significantly higher Mg in serpentine soils inhibiting the uptake of Ca from the soils. Foliar Mg, and especially Ca, were considerably higher in the incense cedar foliage than in the yellow pine and Douglas-fir trees, but Zinke and Stangenberger (1979) had no probability distri- butions for foliage from incense cedar trees. Conclusions The growth of coniferous trees is considerably greater on gabbro soils than on serpentine soils. High Mg in serpentine soils appears to be the main limitation of tree growth of serpentine soils. The high Mg may inhibit the utilization of Ca, which is evident in the higher concentration of Ca in the foliage of Douglas-fir trees on gabbro than in foliage of Douglas-fir trees on serpentine soils. Most of the first transition elements from Cr to K Fig. 3. The proportions of Ca, Mg, and K ions in Douglas-fir, yellow pine, and incense cedar foliage from trees on the Trinity ophiolite of the Klamath Moun- tains. Data for the trees on serpentine soils are represented by closed symbols (cross hatched circles, triangles, and squares). Data for the trees on gabbro soils are represented by open symbols (circles, triangles, and squares). Apices of the triangle represent 100% K (top), Ca (left), and Mg (right). 2014] ALEXANDER: FOLIAR ANALYSES OF CONIFERS ON SERPENTINE AND GABBRO 81 Ni, which are sometimes toxic to plants, had higher concentrations in dithionite-citrate ex- tracts from gabbro soils than from serpentine soils, but there was no evidence that any of these elements were toxic to the trees. None of the comparisons of the N, P, or K concentrations in yellow pine, Douglas-fir, or incense cedar foliage for trees on serpentine and gabbro soils revealed any significant differences between soils. If the serpentine soils contain less N and available plant P than the gabbro soils, the differences are not reflected in the foliar analyses. The significantly greater Ca in Douglas-fir tree foliage on gabbro soils than on serpentine soils is a clear distinction related to the differences between the different soils. Although the foliar Ca differences between the soils were not signif- icant for yellow pine and incense cedar trees, the results for Douglas-fir trees indicate that foliar analyses of plants can reveal interesting clues about the effects of gabbro soils on the plants. Acknowledgments C. “Bud” Adamson, formerly a Shasta-Trinity National Forest botanist, established the plots and recorded the features of each one. James Bertenshaw did the foliar analyses in the laboratory of Paul Zinke at the University of California, Berkeley. Literature Cited Adriano, D. C. 2001. Trace elements in the terrestrial environment: biogeochemistry, bioavailability, and risks of metals. Second edition. Springer, New York, NY. Alexander, E. B. 2011. Gabbro soils and plant distributions on them. Madrono 58:113-122. , C. Adamson, P. J. Zinke, and R. C. Graham. 1989. Soils and conifer forest produc- tivity on serpentinized peridotite of the Trinity ophiolite, California. Soil Science 148:412^23. , R. G. Coleman, T. Keeler-Wolf, and S. Harrison. 2007. Serpentine geoecology of western North America: geology, soils, and vegetation. Oxford University Press, New York, NY. Burge, D. O. and P. S. Manos. 2011. Edaphic ecology and genetics of the gabbro-endemic shrub Ceano- thus roderickii (Rhamnaceae). Madrono 58:1-21. Dayton, B. R. 1966. The relationship of vegetation to Iredell and other Piedmont soils in Granville County, North Carolina. Journal of the Elisha Mitchell Scientific Society 82:108-118. Dunning, D. 1942. A site classification for the mixed- conifer selection forests of the Sierra Nevada. Research Note 28. U.S. Department of Agricul- ture, Forest Service, California Forest and Range Experiment Station, Berkeley, CA. Hunter, J. C. and J. E. Horenstein. 1992. The vegetation of the Pine Hill area (California) and its relation to substratum. Pp. 197-206 in A. J. M. Baker, J. Proctor, and R. D. Reeves (eds.). The vegetation of ultramafic (serpentine) soils: proceed- ings of the first international conference on serpentine ecology, University of California, Davis, 19-22 June 1991, Intercept, Andover, NH. Oberbauer, T. a. 1993. Soils and plants of limited distribution in the Peninsular Ranges. Fremontia 21:3-7. Safford, H. D., j. H. Viers, and S. P. Harrison. 2005. Serpentine endemism in the California flora: a database of serpentine affinity. Madrono 52: 222-257. Snedecor, G. and W. Cochran. 1967. Statistical Methods. Iowa State University Press, Ames, lA. Soil Survey Staff. 1999, Soil taxonomy: a basic system of soil classification for making and interpreting soil surveys. Second edition. Agricul- ture handbook 436. U.S. Department of Agricul- ture, Natural Resources Conservation Service. U.S. Government Printing Office, Washington, DC. Turner, J., S. F. Dice, D. W. Cole, and S. P. Gessel. 1978. Variation of nutrients in forest tree foliage - a review. Internal Report No. 167. College of Forest Resources, University of Washington, Seattle, WA. Wilson, J. L., D. R. Ayres, S. Steinmaus, and M. Baad. 2010. Vegetation and flora of a biodiversity hotspot: Pine Hill, El Dorado County, California, USA. Madrono 56:246-278. Zinke, P. J. and A. Stangenberger. 1979. Ponder- osa pine and Douglas-fir foliage analyses arrayed in probability distributions. Pp. 221-225 in S. P. Gessel, P. M. Kenady, and W. A. Atkinson (eds.). Proceedings of the Forest Fertilization Conference, September 25-27, 1979, Alderbrook Inn, Union, WA (Seattle, WA), University of Washington, College of Forest Resources, Institute of Forest Resources Contribution No. 40. Madrono, Vol. 61, No. 1, pp. 82-86, 2014 ECOLOGY AND DISTRIBUTION OF THE INTRODUCED MOSS CAMPYLOPUS INTROFLEXUS (DICRANACEAE) IN WESTERN NORTH AMERICA Benjamin E. Carter Department of Biology, Duke University, Durham, NC 27708 benj amin . carter @duke . edu Abstract Campylopus introflexus (Hedw.) Brid. is a moss native to the southern hemisphere and well documented as an invasive species throughout Europe. The species was first collected in North America in 1967 in Del Norte Co., California, and now occurs primarily in coastal areas from Santa Barbara Co. to southern British Columbia. Herbarium specimens were assembled and verified to document the distribution of the species in western North America. Collection dates of specimens indicate a rapid invasion along the coast north and south from the first record, with a slower establishment of inland populations. The species is most prevalent in coastal, foggy areas. Many of the populations are associated with anthropogenic environments, but several populations appear to be invading relatively undisturbed sites. These are mostly characterized by poor soils along the coast, with the largest populations associated with Bishop Pine forests and stabilized sand dunes. A brief summary of ecological literature documenting the European invasion of C. introflexus suggests that the species, especially in the presence of disturbance, has the potential to negatively impact native bryophyte, lichen, and vascular plant species. Key Words: Bryophyte, British Columbia, California, introduced species, invasive species, Oregon, Washington. The spread of invasive species is widely regarded as one of the primary threats to global biodiversity (Wilcove et al. 1998). While invasive vascular plants are the subject of active research, and millions of dollars in eradication efforts, the threats of more poorly known plants can go underappreciated because of a lack of basic documentation of their distribution and ecology. Life history traits associated with invasive plant species are common among the bryophytes (hornworts, liverworts and mosses). Many bryo- phyte species are fast growing, excellent dispers- ers that rely heavily on asexual reproduction (During 1979). Despite this, very little informa- tion on bryophyte invasions exists in the litera- ture (Soderstrom 1992). These gaps in knowledge include the basic parameters of invasions - for example, rates of spread and influence on native biota - that inform resource prioritization for the control of invasive vascular plant species. One of the few cases of a bryophyte invasion that has been well studied is the European invasion of the southern hemisphere moss Cam- pylopus introflexus (Hedw.) Brid. Introduced to Great Britain in 1941 (Richards 1963), it spread across the entire continent in less than fifty years, and is now widespread and locally common throughout Europe (Hassel and Soderstrom 2005; Klinck 2009). Over the last several decades, the species has been the source of ecological studies examining ecosystem impacts, habitat requirements and potential control methods throughout its European distribution as workers there move toward understanding the potential impacts of the species (reviewed by Klinck 2009). The earliest collection of C introflexus in North America was in northern California in 1967 (D. Norris 8167 [CAS]). Since then it has spread north to Vancouver and south to the Channel Islands. The purpose of this study is to document the spread in western North America to date, to discuss the apparent ecological preferences of the species in western North America based on herbarium label data and field observations, and to briefly summarize the relevant European literature on the ecology and environmental impacts of the species. Materials and Methods All collections of Campylopus introflexus from western North America at CAS, DUKE, UC, and UBC were examined, as well as collections from the personal herbaria of several bryologists. Field observations of California populations were made sporadically from 2007 to 2012. These included several populations in the Mendocino/ Fort Bragg area, Mendocino Co. (near the putative introduction site), and populations at Salt Point State Park, Sonoma Co., the Monterey peninsula, Monterey Co., and Santa Cruz Island, Santa Barbara Co. Results and Discussion Global Distribution and Ecology Campylopus introflexus has a broad native distribution in the southern hemisphere between the latitudes of 22 and 66 (Gradstein and Sipman 2014] CARTER: DISTRIBUTION AND ECOLOGY OF CAMPYLOPUS 83 1978). Common in Australia, New Zealand, South Africa, and southern South America, it is also present on many of the islands in the south Atlantic and Indian Oceans (Gradstein and Sipman 1978). The first verified record of the species from outside its native range is from Great Britain in 1941 (Richards 1963). From there it spread quickly across Europe and the current distribution spans from Iceland to Italy and from Estonia to Hungary (summarized by Klinck 2009). In Great Britain, where the spread was particularly well tracked, the species demon- strated the lag-period followed by rapid dispersal and colonization that is typical of invasive species (Hassel and Sdderstrom 2005). From 1941 to about 1960, the geographic distribution remained very limited, but this was followed by exponential growth until saturation around 1980 (Hassel and Sdderstrom 2005). It is unknown how the species was transported to Europe and whether there were single or multiple introductions. In some cases, molecular data can be useful in identifying the geographic origin of introduced bryophytes (Carter 2010), however molecular investigations in Campylopus using ITS and the chloroplast atpB-rbcL spacer have so far demonstrated insufficient variation to ascertain the origin of European and North American populations of C. introflexus (Stech and Dohrman 2004, Stech and Wagner 2005). In Europe, where it is well studied, C introflexus grows in a wide variety of habitats but only becomes dominant in sandy, acidic, soils (van der Meulen et al. 1987). In these environ- ments, it can form extensive, dense turfs several centimeters thick. Some of the worst invasions in northern Europe are in stabilized dune commu- nities where C. introflexus threatens fragile lichen dominated communities (Ketner-Oostra and Sy- kora 2008). In these systems, important factors for establishment, persistence and/or competitive advantage of C. introflexus include elevated soil C and N (Sparrius and Kooijman 2011), high soil organic matter (Hasse 2007; Daniels et al. 2008), and physical disturbance (Hasse 2007; Daniels et al. 2008). Under a high disturbance regime induced by ungulates, C. introflexus was docu- mented overtaking a lichen dominated grassland and forming mono-dominant stands (Biermann and Daniels 1997), however a follow-up study ten years after removal of the ungulates demonstrat- ed some recovery of the native lichen community (Daniels et al. 2008). Campylopus introflexus has also been shown to increase in frequency with an associated decrease in frequency of the moss Polytrichum piliferum (Hasse 2007), suggesting that it has the capacity to outcompete other moss species. Woody species can also be impacted by the presence of large turfs of C. introflexus through reduction of germination and establish- ment due to the inability of seedling roots to penetrate through the tall turfs (Equihua and Usher 1993). Distribution and Ecology in North America In North America, Campylopus introflexus has been documented from California, Oregon, Washington, and southern British Columbia. Vouchers from the Aleutian Islands in Alaska and the Queen Charlotte Islands in British Columbia {Schofield 125265 [UBC, DUKE] and Schofield 33675 [DUKE], respectively) were misidentified specimens of other Campylopus species. The first specimen documenting Campy- lopus introflexus in western North America was collected by D. Norris in 1967 near Gasquet, Del Norte Co., California (Appendix 1). This collec- tion precedes earlier reports citing 1971 as the first record (Frahm 1980; O’Brien 1999). At that time Campylopus in North America was poorly understood taxonomically, especially regarding separation of C. introflexus from C pilifer, which is native to the southeastern U.S. and Mexico. After circumscriptions were clarified, it was realized that C introflexus in western North America constitutes an introduced population (Gradstein and Sipman 1978; Frahm 1980, 2007). The majority of known populations of C introflexus in North America occur near the Pacific coastline within range of the marine influence (Fig. 1). For more than 20 years after the first voucher was collected, collections were mostly restricted to a narrow band of coastline within approximately 200 km north and south of the initial collection. A noteworthy exception is the population from Lassen Volcanic National Park first documented by Showers in 1975 (Showers 1978, as C. atrovirens De Not.). That population has been the source of some confusion. The original collection, Showers 1909, is from Boiling Springs Lake, Plumas Co. not Lassen Co. as reported by Frahm (1980). Although Frahm (1980) cited the Showers 1909 specimen (currently housed at UC) as a voucher of C. introflexus in California, the specimen was annotated by him, dated Novem- ber 1979, as C. pilfer . The specimen matches Frahm’s (1980) morphological description of C introflexus and the annotation on the specimen was apparently an error. O’Brien (1999) men- tioned the discrepancy, but assumed that the literature report was erroneous (as opposed to the annotation). The population has recently been re-visited and a specimen collected (M. Hutten, personal communication). The population, which is clearly C introflexus based on current inter- pretations of the morphological circumscriptions, is still established at Boiling Springs Lake and has expanded to the nearby Terminal Geyser (Appendix 1). 84 MADRONO [VoL 61 Fig. 1. Distribution of Campylopus introflexus in North America. By the year 2000, C. introflexus had been documented along the coast from southern British Columbia to the Monterey Bay region, and in the last decade populations have been documented as far south as Santa Cruz Island and inland to Plumas National Forest. Through- out much of its distribution, and most notably in southern populations, C. introflexus is often associated with closed cone coniferous forests, especially Bishop Pine (Pinus muricata D. Don) forests (e.g., Christy Pines on Santa Cruz Island, Santa Barbara Co., CA; Del Monte Forest on the Monterey Peninsula, Monterey Co., CA; Salt Point State Park, Sonoma Co., CA; Pygmy forests of Mendocino Co., CA). Northern occur- rences in Oregon and British Colombia are typically associated with nutrient poor sand dune soils, waterlogged soils of bogs and fens, or ruderal areas including roadsides, rooftops, and landscaped areas. The inland populations in Lassen Volcanic National Park are associated with hydrothermal areas (Showers 1978), while the other inland populations from Yuba Co. and Lake Co. are associated with highly disturbed areas (D. Toren, personal communication). Collections from disturbed environments consti- tute a large proportion of the records throughout the distribution, but several populations appear to be well established in otherwise undisturbed vegetation. These include some of the popula- tions in the pygmy forests around Fort Bragg and Mendocino, as well as the Santa Cruz Island population. Populations at the Del Monte Forest and Salt Point are mostly found in the immediate vicinity of hiking trails and unpaved roads. Clones can be robust, forming dense monospe- cific turfs to five centimeters thick or more and, in ideal conditions, several square meters or more in extent. Although no quantitative ecological data are available, these extensive patches appear to be outcompeting other bryophyte and lichen species. Campylopus introflexus is reportedly dioicous (i.e,, it has separate male and female gameto- phytes; Frahm 2007), however sporophytes have been observed in one of the northernmost populations at Burns Bog in southwestern British Columbia (Taylor 1997), from the southernmost population on Santa Cruz Island, and from the inland Boiling Springs Lake population. Given that the presence of sporophytes in a population implies colonization of at least two individuals (one male, one female), the presence of sporo- phytes throughout the distribution is noteworthy. Conclusions Campylopus introflexus has a well-documented history of invasion in Europe, and efforts there now focus on understanding the life history of the species and its impact on native ecosystems. In western North America much less is known due to the lack of adequate historical collections. With that important caveat, the existing collec- tions are consistent with the expected pattern for a rapidly expanding distribution. While nothing is known about the potential of the species to impact the native flora in western North Amer- ica, negative impacts in similar European ecosys- tems are well documented and should serve as a warning. This study establishes baseline information on the currently known distribution and qualitative ecological preferences of the species. Future work should include efforts to increase awareness of botanists to introduced bryophytes, and imple- mentation of quantitative ecological studies to better understand the ecosystem impacts and potential future distribution of C. introflexus. This is especially important in closed cone coniferous forests, and stabilized dune commu- nities, both of which exist across relatively small areas and are rich in sensitive endemic vascular plants. Acknowledgments I thank the curators of CAS, DUKE, UBC, and UC for access to their specimens. I am indebted to M. Hutten, J. Harpel, D. Toren, and D. Wagner for bringing to my attention several populations and for 2014] CARTER: DISTRIBUTION AND ECOLOGY OF CAMPYLOPUS 85 providing valuable insights based on their extensive fieldwork and collections. Literature Cited Biermann, R. and F. J. A. Daniels. 1997. Changes in a lichen-rich dry sand grassland vegetation with special reference to lichen synusiae and Campylopus introflexus. Phytocoenologia 27:257-273. Carter, B. E. 2010. The taxonomic status of the Tasmanian endemic moss Scleropodium australe. Bryologist 1 13(4):775-780. Daniels, F. J. A., A. Minarski, and O. Lepping. 2008. Long-term changes in the pattern of a Corynephorion grassland in the inland of the Netherlands. Annali di Botanica 8:9-19. During, H. J. 1979. Life strategies of bryophytes: a preliminary review. Lindbergia 5:2-18. Equihua, M. and M. B. Usher. 1993. Impact of carpets of the invasive moss Campylopus introflexus on Calluna vulgaris regeneration. Journal of Ecology 81:359-365. Frahm, j. 1980. The genus Campylopus in North America north of Mexico. Bryologist 83:570-588. . 2007. Campylopus. Pp. 366-376 in Flora of North America Editorial Committee, eds. 2007, Flora of North America North of Mexico, Vol 27: Bryophyta: Mosses, part 1. Oxford University Press, New York, NY. Gradstein, S. R. and H. J. M. Sipman. 1978. Taxonomy and world distribution of Campylopus introflexus and C. pilifer. (=C. polytrichoides): a new synthesis. Bryologist 81:114—121. Hasse, T. 2007. Campylopus introflexus invasion in a dune grassland: succession, disturbance, and rele- vance of existing plant invader concepts. Herzogia 20:305-315. Hassel, K. and L. Soderstrom. 2005. The expansion of the alien mosses Orthodontium linear e and Campylopus introflexus in Britain and continental Europe. Journal of the Hattori Botanical Labora- tory 97:183-193. Ketner-Oostra, R. and K. V. Sykora. 2008. Vegeta- tion change in a lichen-rich inland drift sand area in the Netherlands. Phytocoenologia 38:267-286. Klinck, j. 2009, The alien invasive species Campylopus introflexus - a threat to the Danish coastal dune system. M.S. Thesis, Copenhagen University, Copenhagen, Denmark. O’Brien, T. J. 1999. Noteworthy collections: Califor- nia. Madrono 46:113-114. Richards, P. W. 1963. Campylopus introflexus (Hedw.) Brid. and C. polytrichoides De Not. in the British isles: a preliminary account. Transactions of the British Bryological Society 4:404^17. Showers, D. W. 1978. A moss flora of Lassen Volcanic National Park, California. M.A. Thesis, San Francisco State University, San Francisco, CA. Soderstrom, L. 1992. Invasion and range expansion and contractions of bryophytes. Pp. 131-158 in J. Bates and A. Farmer, (eds.), Bryophytes and lichens in a changing environment. Oxford Univer- sity Press, Oxford. Sparrius, L. B. and a. M. Kooijman. 2011. Invasiveness of Campylopus introflexus in drift sands depends on nitrogen deposition and soil organic matter. Applied Vegetation Science 14:221-229. Stech, M. and j. Dohrman. 2004. Molecular relationships and biogeography of two Gondwa- nan Campylopus species, C. pilifer and C. intro- flexus (Dicranaceae). Monographs in Systematic Botany, Missouri Botanic Garden 98:415^31. and D. Wagner. 2005. Molecular relation- ships, biogeography, and evolution of Gondwanan Campylopus species (Dicranaceae, Bryopsida). Tax- on 54(2), 377-382. Taylor, T. 1997. Campylopus introflexus- moss intro- duced in British Columbia. Botanical Electronic News no. 162. Website http://www.ou.edu/cas/ botany-micro/ben/ben 162.html (accessed 15 Sep- tember 2011). VAN DER MEULEN, F., H. VAN DER HAGEN, AND B. Kruijsen. 1987. Campylopus introflexus: invasion of a moss in Dutch coastal dunes. Proceedings of the Koninklijke Nederlandse Akademie van We- tenschappen, Series C 90:73-80. WiLCOVE, D. S., D. Rothstein, j. Dubow, a. Phillips, and E. Losos. 1998. Quantifying threats to imperiled species in the United States. Bioscience 48:607-615. Appendix 1 Voucher Specimens, with Locality, Year of Collection, and Herbarium CANADA. B.C.: Vancouver, 2010, Joya 609 (UBC); Texada Island, 2002, Sadler s.n. (UBC); Burns Bog, 1997, Schofield 107660 (DUKE, UBC); Burns Bog, 1997, Schofield 107661 (DUKE, UBC); Burns Bog, 1997, Schofield 107665 (UBC); Lulu Island, Richmond Nature Park, 2002, Schofield 120055 (DUKE, UBC); Lulu Island, W of Richmond Nature Park, 2006, Schofield 124572 (DUKE, UBC); Burns Bog, 1994, Taylor s.n. (UBC); Burns Bog, 1996, Taylor 96-2 (UBC); South Burnaby, near Byrne Rd, 1996, Taylor 96-9 (UBC); Lulu Island, W of Richmond Nature Park, 2002, Taylor s.n. (UBC). USA. California. Del Norte Co.: Hwy 199 near Stoney Creek, 1971, Chapman 101 (UC); Darlingtonia bog 2 mi N of Gasquet, 1967, Norris 8167 (CAS); Gasquet, 1971, Santana 267 (UC). Humboldt Co,: Little River, near Airport, 1988, Mishler 3762 (Duke); Seashore 3 mi N of Manila, 1975, Montalvo s.n. (UC); Logging Rd S of Lumbar Hills, 1984, Nelson 7616 (UC); Humboldt State University Campus, 1979, Norris 53101 (UC); Bayside Golf Course, 1979, Norris 53102 (UC); Prairie Creek Redwoods State Park, 1980, Norris 56821 (UC); Greenwood Heights Rd ca 3.5 mi SE of Old Areata Rd, 1981, Norris 58319 (UC); Freshwater Creek 2 mi SW of Kneeland, 1983, Norris 68275 (UC); Areata City Forest 1 mi from Fickle Hill Rd, 1983, Norris 68751 (UC); Prairie Creek Redwoods State Park, 1984, Norris 71732 (UC); Hwy 101 Rest stop 5 mi N of Trinidad 1987, Silver 780 (UC); Sunny Brae, near Areata, 1980, Spjut 6319 (UC). Lake Co.: Sulfur Bank Mine, near Clearlake, 2000, Toren 8027 (D. Toren, personal herbarium); Sulfur Bank Mine, near Clear- lake, 2005, Toren 9383a (conf. B. Allen) (D. Toren, personal herbarium); Sulfur Bank Mine, near Clear- lake, 2005, Toren 9383b (conf. B. Allen) (D, Toren, personal herbarium). Marin Co.: Limantour Beach, 1999, Obrien 3353 (UC). Mendocino Co.: Pygmy Forest 1.2 mi E of Hwy 1 along Gibney Ln, 1998, Becking 86 MADRONO [Vol. 61 98091118 (UC); Pygmy Forest 1.2 mi E of Hwy 1 along Gibney Ln, 1998, Becking 98091119 (UC); Logging Road 5 air miles E of Mendocino, 2007, Carter 2337 (UC); Pygmy forest east of Mendocino, 2007, Carter 2340 (UC); Mackerricher State Park 2007, Carter 2253 (UC); Pygmy forest near Airport SE of Mendocino 2007, Carter 2393({JC); Pygmy Forest 2 mi from Hwy 101 along Rd 409, 1983, Norris 68762 (UC); Pygmy forest 1 mi E of Hwy 1 along Airport Rd, 1987, Norris 73153 (UC); Pygmy forest 6 mi E of Hwy 101 along Little Lake Rd, 1994, Norris 82496 (UC); Pygmy forest 6 mi. E of Hwy 101 along Little Lake Rd 1994, Norris 82502 (UC); Pygmy forest near Fort Bragg Airport, 2002, Norris 103725 (UC); Pygmy forest near Fort Bragg Airport, 2002, Norris 103731 (UC); Jackson State Forest, 3 mi E of Fort Bragg, 1971, Selva s.n. (UC); Pygmy forest near Albion, 2002, Shevock 21904 (CAS, UC); Pygmy forest near Albion, 2002, Shevock 21912 (CAS, UC); Pygmy forest at Sholar’s Bog 2004, Shevock 24650 (CAS, UC); Fen area at Big River Laguna, 2004, Shevock 24676 (CAS, UC); Pygmy forest along Simpson Lane, 1976, Showers 3258 (UC); Mitchell Creek Rd 1 mi S of Simpson Lane, 1976, Showers and Toren s.n. (UC). Monterey Co.* Del Monte Forest, 2011, Carter 5942 (UC); Del Monte Forest, 2008, Kellman and Ladder 5921 (CAS); Pebble Beach, 2001, Yadon s.n. (UC). Plumas Co.: Lassen Volcanic National Park, Boiling Springs Lake, 2012, Hutten s.n. (M. Hutten, personal herbarium); Lassen Volcanic National Park, Terminal Geyser, 12 Oct 2012, Hutten s.n. (M. Hutten, personal herbarium); Boiling Springs Lake, 1979, Showers 1909 (UC). San Francisco Co.: Mt. Davidson, 2000, Toren 7760 (CAS). San Mateo Co.: San Pedro Valley County Park, Montara Mt. Trail, 2009, Shevock 32699 (CAS). Santa Barbara Co.: Santa Cruz Island, Christy Pines, 2011, Carter 5330 (UC); Santa Cruz Island 1 mi W of Mt. Diablo, 2011, Carter 5954 (UC); Santa Cruz Island, Sierra Blanca Ridge, 2012, Carter 6686 (UC); Santa Cruz Island, Los Sauces Trail, 2002, Robertson 2154 (CAS). Santa Clara Co.: Geng Rd, Palo Alto, 2004, Kellman 3634 (CAS). Sonoma Co.: Salt Point State Park, 2012, Carter 6530 (UC). Yuba Co.: Along Oregon Hill Rd. near Green- ville, 5 mi SE of Challenge, 2010, Toren 9735 (CAS). USA. Oregon. Coos Co.: Muddy Lake, 8 km N of Langlois, 1981, Christy 3172 (DUKE, UBC). Curry Co.: Floras Lake State Park, 1981, Christy 3206 (DUKE, UBC); Floras Lake State Park, 1981, Christy 3207 (conf. Frahm) (UBC); 8 mi. north of Port Orford, 2006, B. Shaw 602 (DUKE); Lane Co.: Siuslaw NF, Oregon Dunes Recreation Area, 1993, Christy 8329 (UBC); Siuslaw NF, Oregon Dunes Recreation Area, 1993, Christy 8337 (UBC); Siuslaw NF, Sutton Recreation Area, 1999, Christy 9206 (UBC); Eugene, Lane Community College Campus, 1995, Dorris 1288 (UC); Florence, 2006, B. Shaw 593 (DUKE); Florence, 2006, B. Shaw 594 (DUKE); Florence, 2006, B. Shaw 595 (DUKE); Florence, 2006, B. Shaw 596 (DUKE). USA. Washington. Grays Harbor Co.: Hwy 109 near Moclips, Township 21N, Range IIW, Section S29, 2001, Hutten 4952 (conf. Harpel 16 May 2002) (M. Hutten, personal herbarium). Madrono, Vol. 61, No. 1, pp. 87-104, 2014 BIOLOGY OF THE GEOPHYTE, TRITELEIA IXIOIDES SUBSP. ANILINA (THEMIDACEAE), IN CONIFEROUS FORESTS OF BUTTE COUNTY, CALIFORNIA Alfred Kannely 3558 Paseo Avenue, Live Oak, CA 95953 Alfred@ofir.DK Robert A. Schlising Department of Biological Sciences, California State University, Chico, CA 95929“0515 Abstract The life cycle of the widespread, cormous geophyte, Triteleia ixioides (W. T. Aiton) Greene subsp. anilina (Greene) L. W. Lenz (Themidaceae) was studied in the field over several years in coniferous forest sites in Butte County, CA. All sites, ranging from 1383 to 1774 meters in elevation, have a Mediterranean climate and loamy soils above Tuscan mudflow substrate. Flowering scape height, corm weight, and corm depth varied significantly. Old corms decreased in dry weight throughout the winter, as shoots, roots, and new corms grew. By spring, each old corm had produced a leaf and a flowering scape, was depleted of stored food, and had a new “replacement corm” (without roots) already developed on top of it. Fruit set varied from 38.5% to 89.1%, with significant differences among populations sampled in different years and sites. Seed set was low (overall mean 38.4%), but varied significantly among years and sites, ranging from 26.4% to 56.3%. Plants are partially self- compatible. Soil moisture was not limiting to seed set, nor was pollen vectors, except on cold days. The rriost important pollinator was a bee fly {Bombylius facialis, Bombyliidae), although native bees provided some pollination at higher sites. Ten-minute observation periods, during bee fly activity, showed up to 15 visits to plants in a square meter. Percent seed germination was high (88-100%), with germination after the first rains and seedling growth continuing during the cold months. Seedlings produced single leaves up to 60.4 mm long, and corms averaged 1.6 ± 0.1 mm wide at one site. Contractile roots, produced lateral to the primary roots, averaged 16.13 ± 1.00 cm long. Shrinking and wrinkling at the upper part of the contractile root pulled the newly formed corm deeper in the soil. Study of contractile roots in Triteleia and other geophytes may help illustrate adaptation to summer drought in Mediterranean California. Key Words: Bombylius facialis, contractile root, corm, geophyte, low seed set, scape-wasting, Triteleia ixioides subsp. anilina. The geophyte is a very common growth form among perennial herbaceous plants in the region of California with Mediterranean climate. The term “geophyte” was defined by Raunkiaer (1934) for plants that die down yearly to an underground bulb, corm, or tuber at the end of one season and then renew growth from this organ when favorable growth conditions return. Although geophytes occur in many regions with different climates, Raunkiaer and others have long described them as especially adaptive in regions with Mediterranean and other summer- dry climates. In addition to drought avoidance, Pate and Dixon (1982) and Proches et al. (2005) concluded that a belowground “food storage function” was also of central importance in defining geophytes in the areas they studied. In California a few recent studies on geophytes have focused on floral adaptations and pollina- tors (e.g., Dilley et al. 2000; Patterson and Givnish 2003). Several other studies have focused on belowground growth of the corm or bulb in California plants (Rimbach 1902; Smith 1930; Jemstedt 1984; Han 2001), some of which have dealt with the life cycle in relation to the summer drought of the Mediterranean climate (e.g., Tyler and Borchert 2002; Schlising and Chamberlain 2006). However, and importantly, in some accounts of California geophytes (Rundel 1996; Parsons 2000; Schlising and Chamberlain 2006) the authors noted there is a lack of basic information available on the field biology and complete life history for most members of the rich geophyte component in Mediterranean California. Further investigation of individual species, and how they might differ from each other, will contribute to better understanding of the geo- phyte life form as a “strategy” in coping with predictable extremes inherent in Mediterranean climate. Such understanding may aid in long- term conservation of this important element of California’s native flora. Although not one of the seven taxa of Triteleia listed in the California Native Plant Society’s Inventory of Rare and Endangered Plants of California (CNPS 2013), we chose to study Triteleia ixioides (W. T. Aiton) Greene (Themi- 88 MADRONO [Vol. 61 daceae). It is an abundant and conspicuous geophytic component of open coniferous forests and meadow edges in our area (northern Sierra Nevada and southern Cascade Range). This species occurs from southwest Oregon to the Klamath and Cascade Ranges, the Sierra Ne- vada, and the Coast Ranges of California. Within this large geographic range, six subspecies are recognized (Hoover 1941; Lenz 1975, 1976; Fires and Keator 2012). Very little information has been published on this species. Han (2001) investigated weights of corms in relation to scape and flower production when grown in cultivation and in the field. Berg (2003) described ovules and the embryology for T. ixioides flowers, but did not specify which subspecies was examined. Kannely (2005) report- ed the pollen: ovule ratio for T. ixioides subsp. anilina (Greene) L. W. Lenz, from one of the populations observed in detail in the present study. The most recent morphological description (Fires and Keator 2012) for the subspecies we studied, Triteleia ixioides subsp. anilina, indicates that its spheroid corm produces one or two linear leaves 10-25 cm long and a flowering scape up to 30 cm tall when the leaves are drying. The umbel- like inflorescence has two to many yellow flowers with funnel-shaped tube and spreading lobes. The general habitat (Fires and Keator 2012) is conifer forest edges at 600-3000 m in elevation. Since T. ixioides subsp. anilina occurs in the Klamath and Cascade Ranges, the Coast Ranges, and the Sierra Nevada (Fires and Keator 2012), our study in northern Butte County thus represents this subspecies near the center of its overall range. Apparently no field studies have been done that describe the field biology of T. ixioides subsp. anilina. Our study was done to provide basic information illustrating the phenol- ogy and life history of this common plant as it exists in forest habitats in northern Butte County. Main goals were to 1) provide information on size of vegetative parts of the plants in relation to growth during different times of the year, 2) document flower visitors that might serve as pollinators, 3) determine the degree of reproduc- tive success plants had in producing seeds, and 4) to investigate seed germination and seedling growth in relation to Mediterranean climate. Study Area and Regional Climate To incorporate potential variation in the study plants in northeastern Butte County three main study sites were used. These extended over about 25 km of coniferous forest habitat, and ranged from 1383 to 1774 m in elevation. Bedrock is Tuscan Formation volcanic mudflow, here rep- resenting the southernmost extent of the Cascade Range. Soils in major portions of all study sites are rocky, and consist of sandy loam or loam, as determined with composite samples sent to A. and L. Analytical Laboratories, Inc., Memphis, TN. Current names for the specific soil types are not available, since the recent Soil Survey of the Butte County Area did not include this forested region. All three study sites have a Mediterranean climate, with a shorter winter at the lowest site than at the other two. The lowest site has a weather-recording station nearby, approximately two kilometers southwest of Butte Meadows, at 1467 m elevation, for which 12 years of climatic measurements were available (www.wrcc.dri.edu/ cgi-bin/rawMAIN.pl?caCCRR accessed 12 April 2012). Climate data from this source were used to approximate conditions for all three sites, as no other climate data were available. The lowest study site, “Ridge,” is near a USGS benchmark at 1383 m elevation, on Carpenter Ridge, about 6.9 km southwest of Butte Mead- ows at 40°03'19.9" latitude, 121°35'40.1" longi- tude. Here, Triteleia plants grow mainly on the flat ridge top, and receive at least some shade every day during the growing season. Main forest trees are Pinus ponderosa Lawson & C. Lawson, P. lambertiana Douglas, Pseudotsuga menziesii (Mirb.) Franco, Calocedrus decurrens (Torr.) Florin, Abies concolor (Gordon & Glend.) Lindl. ex Hildebr. and Quercus kelloggii Newb.; shrubs include Ceanothus integerrimus Hook. & Arn., C. prostratus Benth. and Arctostaphylos patula Greene. Herbaceous associates include Penste- mon neotericus D. D. Keck, Senecio aronicoides DC., Calochortus coeruleus (Kellogg) S. Watson, Allium campanulatum S. Watson, Kelloggia ga- lioides Torr., Cordylanthus tenuis A. Gray, Mimulus torreyi A. Gray, Microsteris gracilis (Douglas ex Hook.) Greene, and Stipa stillmanii Bol. Average annual precipitation (2000 through 2011) is 1211 mm from September through August (the Mediterranean rainfall year) with a low of 443 mm in the 2007-2008 study-year, and a high of 2160 mm in 2005-2006. Soil texture is sandy loam (71.3% sand, 21.1% silt and 7.5% clay) and is rockier than at the two other study sites. The second site, “Meadow,” is located 6.6 km east of Jonesville, and 0.32 km south of the road along Scotts John Creek, on U.S. Forest Service road 26N27, at 40°06'21.0" latitude, 121°25'35.4" longitude, and 1746 m elevation. Triteleia plants grow in bands between dense forest trees and a wet, sodden, three-lobed meadow. Here they receive shade from trees during much of the year. Trees are mostly Abies concolor and Pinus jeffreyi Balf.; shrubs include Symphoricarpos mollis Nutt, and Ribes roezlii Regel. Herbaceous associates are Sidalcea glaucescens Greene, Hack- elia californica (A. Gray) 1. M. Johnst., Allium campanulatum, Mimulus torreyi, Dicentra uniflora 2014] KANNELY AND SCHLISING; BIOLOGY OF THE GEOPHYTE, TRITELEIA IXIOIDES 89 Kellogg, and grasses such as Elymus elymoides (Raf.) Swezey subsp. calif ornicus (J. G. Sm.) Barkworth, and E. glaucus Buckley. This site, 363 m higher than Ridge, has snow persisting longer. In all years of the study, snow on the local roads prevented spring observations until late May or early June. Soil texture is loam (50.8% sand, 35.4% silt and 13.7% clay). The third (less intensively used) study site, “Creek,” occurs 10.7 km east of Jonesville on Humbug Summit Road, adjacent to the California Department of Fish and Wildlife’s Butte Creek House Ecological Reserve, at 40°05'16.8" latitude, 121°24'59.2" longitude, at 1774 m elevation. Trite- leia occurs in openings and among sparse stands of Pinus contorta Douglas ex Loudon subsp. murrayana (Balf.) Critchf. and Pinus jeffreyi; the Triteleia plants receive at least some shade daily during the growing season. Associates include Acmispon americanus (Nutt.) Rydb., Allium cam- panulatum, Mimulus torreyi, and Elymus ely- moides subsp. californicus. Being at the same approximate elevation, the climate here is similar to that of the Meadow site, as is the soil texture (loam, with 48.8% sand, 39.4% silt, 11.6% clay). Methods Sampling Protocol Populations were sampled similarly in most years from 2004 to 2009 from baselines (not starting at the exact same points), subjectively placed in areas with abundant Triteleia plants. Most parameters were measured on plants selected randomly along transects every 1-2 m, that were up to three m long and at right angles to the baselines. Additional transects were used to do destructive sampling, including collecting of corms and flower visitors, and to do experimental manipulations of flowers. Destruction of corms was limited to only several seasons so as to not significantly change the density of the plants in populations being studied. Specific features chosen for measurement were those that would characterize both vegetative growth and repro- duction, and could be easily measured or assessed in the field. Vegetative Features Scape height was measured to the nearest 0.5 cm, from soil surface to base of the single umbellate inflorescence when plants were in fruit. Depth to the base of the corm was measured from the soil surface when plants were in fruit. To assess characteristics of the corms, harvested corms had their coats, shoots, and roots gently removed before corm diameters (width and height) were measured to the nearest 0.5 cm. Since corms were slightly flattened spheres. volume was derived using the formula for an oblate sphere (Vobiate — 4/3Tca^b). Corms were dried at about 37°C for one month to determine dry weight. Phenology and growth of new corms, and old corm replacement were studied in the field at Ridge, starting in September 2005 and ending in June 2006. Thirty plants in fruit were chosen randomly in September to have their corms harvested and measured. The four nearest plants (at the major compass directions from the harvested plant) were marked with wooden skewers in the soil, for comparative corm harvesting during the following months to show the nature of changes in the corms during the winter. In Meadow in 2007 plants with different degrees of symptoms of a “scape- wasting syn- drome” were recorded in a 50 m^ belt transect within the population that had been used for measurements and counts of various vegetative and reproductive features in preceding years. Reproductive Features Flowers. Flowers per plant were determined at time of fruiting, by counting the number of fruits and then adding the number of fruitless pedicels present (pedicels remain when flowers abscise). In order to characterize the mature adult portion of the population, only plants with two or more flowers were sampled. Pollen: ovule ratio was determined using the sonication method (Kan- nely 2005). To determine if there was autonomous self- pollination, inflorescences at Ridge with flower buds ready to open were “bagged” using 15 by 20 cm lens paper (Fisher Scientific) closed at the base of the inflorescence with lightweight (Tot- 50) staples. Space was left in the resulting “bag” for flowers to open completely; the “bag” was also secured to a wooden skewer pushed into the ground several cm from the scape. Entire inflorescences were harvested in five weeks. In 2007 this experiment was done with 32 plants, each paired with the closest control plant to the east, which was marked but not bagged. Nectar was sampled at Creek in several flowers that had been bagged overnight, by using 10 pi micropipettes and a Bellingham and Stanley pocket refractometer reading 0-50% sucrose equivalents. Quantification of insects visiting Triteleia flowers was done at Ridge in two seasons and at Meadow in one, by completing numerous 10- minute observation periods. One-meter square areas with six to 75 flowering plants were chosen subjectively to prevent spatial overlap. A stan- dardized page with a large square representing a square meter of ground had all flowering plants quickly mapped, with the number open flowers written on each. The plot was watched for 10 min, and visits were recorded when an insect probed a 90 MADRONO [Vol. 61 flower with its tongue. When an insect returned to probe a flower that it had probed even seconds earlier in the 10-minute period, this was recorded as another visit. A thermometer in the shade of the observer was read after 10 min. Thus, a record was made of flowers per plant, insect visits per plant, and how many plants were visited in sequence by a particular insect. Twelve specimens of the fly Bombylius facialis Cresson, were collected at Ridge on the morning of 29 June 2010 and were examined for pollen on their bodies. Over three years 180 plots were watched (at Ridge, 70 in 2005, 40 in 2007 and 30 in 2008; at Meadow, 40 in 2005). In the three years a total of 2582 plants and 4281 flowers were observed. Isolated plants lying outside of the main transects of the study areas were not included. Fruits and seeds. To assess the “success” of fruit and seed formation, number of fruits (and percent fruit set) per plant was detemiined just prior to dehiscence in July. Two mature fruits were collected, from opposite sides of each inflorescence, from plants along the main tran- sects. These were put into separate #1 coin envelopes (Swinton Avenue Trading, Boca Ra- ton, FL). In the laboratory each fruit was dissected to count mature seeds and ovules that had failed to become seeds, to determine ovules per ovary, seeds per fruit, and percent seed set. Seeds were counted as mature if they were black and had normal subspheric shape and size. Seed weight, viability, and germination, and early seedling growth. Features of the seeds, including weight and viability, were determined using 200 black seeds collected from dehisced fruits, at Ridge and Meadow on 6 September 2005, and at Creek on 9 August 2008. In 2005 two Petri dishes had 50 seeds each from Ridge placed on filter paper, flooded with distilled water and kept in the dark for 45 hours. Seeds were cut in half, and placed with the cut surfaces facing downward in dishes of filter paper soaked with tetrazolium chloride. After 24 hours seeds were examined for staining of the embryos as indica- tion of viability (Bradbeer 1988). Timing of seed germination under field condi- tions was determined by planting sets of seeds collected from open fruits at the end of the season before fall rains began (Ridge, Meadow: 6 September 2005; Creek: 5 October 2008). Lots of 50 seeds each were spread along 20-cm strips of v-folded, non-inked newsprint (to permit finding seeds later), and covered with 8-10 mm of local soil. Rows were covered with squares of hardware cloth to help prevent disturbance by animals; seeds here received only natural precip- itation. Rows of seeds and their supporting newsprint were recovered at intervals during the wet season and examined in the lab. Data Analysis One-Way ANOVAs and Tukey-Kramer com- parisons were used on most plant parameters measured to determine if location and year were factors. When data did not meet assumptions for normality and equality of variances Kruskal- Wallis and Wilcoxen tests were compared with parametric tests. However, since we had reason- ably large sample sizes, and the probabilities obtained from the non-parametric tests were nearly identical to the P-values obtained from the One-way ANOVA and Tukey-Kramer com- parisons, we have reported the latter. Two- sample t-tests were used to compare means for corm diameters, seed weights, and length of seedling leaves. Most data were analyzed using Minitab 16 (Minitab Inc., State College, PA) and JMP PRO 10 (SAS Institute Inc., Cary, NC). Results Vegetative Features General Morphology. Means for flowering scape heights were fairly uniform in the region, but scapes were significantly taller (P < 0.001) at the higher elevation site (Meadow) each year except 2007 (Fig. 1). The aboveground growth of scapes began approximately one month later in the spring at Meadow than at Ridge. In 2007, an unidentified “scape- wasting syn- drome” appeared on many of the plants inter- spersed among apparently healthy plants at Meadow. Of 169 scapes observed in a represen- tative section of the population, only 12% of the plants appeared healthy, showing no apparent symptoms when observed on 22 June. The remaining plants showed symptoms, including whole scape upright, but with flowers dead (79%), or scape bent in the middle and with the inflorescence dying (9%). None of the plants with symptoms produced flowers that year. Several plants were noted with similar symptoms at Creek on 13 June 2007, but were not seen in any other year. Triteleia plants have also been found parasitized by a dodder (probably Cuscuta californica Hook. & Arn.) - one plant at Meadow in July 2006 and many plants at Creek in July 2008. Corm depths (soil surface to base of corm) were significantly greater by three or more cm at Meadow than at Ridge both years measured (P < 0.001; Table 1). Corm volume did not vary significantly throughout the region, but corm dry weight was higher at Meadow (P = 0.01) (Table 1). Small cormlets, vegetative side shoots pro- duced by mature corms during spring growth, were rarely seen on corms dug at Ridge, but occurred more often at Meadow. At Meadow in 2014] KANNELY AND SCHLISING: BIOLOGY OF THE GEOPHYTE, TRITELEIA I XI OWES 91 16 14 S- 12 § 10 e £ & M 8 6 I 2 0 2004 2005 2007 2008 Fig. 1. Flowering scape height in Triteleia ixioides subsp. anilina at two sites in Butte County. One-Way ANOVA with P == 0.0001 was followed by Tukey- Kramer comparisons, showing that scapes at Meadow were significantly taller than at Ridge, in each year except in 2007. Scape heights did not differ among the four years at Meadow except that they were significantly taller in 2008 than in 2007. Scapes at Ridge were shorter in 2008 than 2005. N was 30 in all samples, except 40 at Meadow in 2008, and 39 in 2007 and 41 in 2008 at Ridge. 2004, of 30 adult corms dug, 22 had between one and four cormlets. Cormlet dry weight (X ± SE) was 0.0256 ± 0.003 g (n = 45), compared with 0,259 ± 0.025 in the adult corms (Table 1). Annual corm replacement. Corms that reached full size (Fig. 2a) at the time of fruit and seed maturation in June or July remained largely inactive until the fall/winter rains commenced. All of five corms dug 20 October 2005 at Ridge showed signs of renewed shoot (but not root) growth after 30.5 mm precipitation in the preceding weeks; shoot length above the flat upper surfaces of the corm was only 2-5 mm. Corms sampled on 5 November 2005 (n — 25) after an additional 70.6 mm of rainfall, and again on 10 February 2006 (n = 24) showed an increase in shoot length and a decrease in corm weight (Fig. 3). Shoot length (cm) increased 357% (from 0.54 ± 0.027 to 2.47 ± 0.149), while corm weight (g) decreased 60% (0.208 ± 0.041 to 0.084 ± 0.012). In February these corms had numerous delicate roots which broke readily, several reaching up to 3.2 cm long. Heavy snowpack prevented sampling again until 5 May, by which time each old corm had produced a leaf (sheathing the shoot) that reached the soil surface. The single dark green leaf on the new shoot (n — 28) reached 5.88 ± 0.38 cm long, as measured from the ground surface. In addition, a new corm was obvious at the top of each old corm (Fig. 2b, top arrow). The weight of the old corms in May averaged less than in February but was still higher than the weight of new corms (0.035 ± 0.005 g vs. 0.009 ± 0.001 g; n = 23; P = 0.001). In May the growing scape (n = 28), visible Table 1 . Features (Mean ± SE) of Corms in Triteleia ixioides subsp. anilina at Two Sites in Butte County. Sample size is shown in parentheses (n). A t-test was used to compare corm volumes. One-Way ANOVAs were used for other parameters, with Tukey-Kramer comparisons used for means of corm depth and dry weight. In each column, means with the same superscript letters do not differ significantly. Site Year Diameter (cm) Depth, from soil surface to base of corm (cm) Volume (cm^) Dry weight (mg) Ridge 2005 0.98 ± 0.03" 5.03 ± 0.29" 0.39 ± 0.06" 0.168 ± 0.013" (n) (30) (30) (30) (31) Meadow 2004 1.07 ± 0.04" 8.22 ± 0.6*’ 0.259 ± 0.025*’ (n) (30) (30) (-) (30) 2005 1.01 ± 0.03" 9.31 ± 0.50*’ 0.41 ± 0.04" 0.236 ± 0.026"*’ (n) (30) (30) (30) (29) 92 MADRONO [VoL 61 Fig. 2. Features in the life cycle of Triteleia ixioides subsp. anilina. a. Rootless new conn (about 14 mm in diameter) in July, showing remains of earlier rooted corm. b. Multiple generations of corms, showing the youngest as a bulge (top arrow) at top of functional, rooted corm, and old and decaying remains of two earlier generations (bottom two arrows), c. Concurrent shoot and (white) new corm. development in early June, growing from the functional, rooted old corm below, d. Bombylius facialis (head about four mm wide) probing a flower of Triteleia. e. Ventral view of Bombylius, showing Triteleia pollen at base of tongue, f. Seedling, showing primary and contractile (arrow) roots. Seed about three mm long. g. Seedling, near end of first year’s growth, with small corm at the top of large contractile root. Gridlines at five mm. within the sheathing leaf, averaged 2.86 ± 035 cm long. When this study to characterize corm replace- ment ended on 8 June (Fig. 2c), the new corm (n = 29) was considerably larger and heavier (X = 0. 140 g) than the depleted old corm (X = 0.014 g). The plants had flower buds on lengthening scapes, now at 7.68 ± 0.64 cm, but had leaves at final length (1 1.79 ± 0.49 cm). The new corms did not have any roots. By early summer old corms were shrunken and dried. In many plants there were also the stilholder dried remains of a corm from earlier years visible (Fig. 2b, bottom two arrows). Corms were heavily predated by pocket gophers {Thermomys bottae [Eydoux & Gervais, 1836]) most years at Meadow and Creek, but not at Ridge. Judging by tracks, black-tailed deer were responsible for many nipped scapes and leaves at all three sites. No insect damage was detected on vegetative parts. Flowers and Fruits Flower numbers and floral behavior. Numbers of flowers per plant from samples ranged from two to 27 (Table 2), with the overall mean for the 335 plants in the entire study being 8.2 ± 0.3. 2014] KANNELY AND SCHLISING: BIOLOGY OF THE GEOPHYTE, TRITELEIA IXIOIDES 93 0.35 0.3 (A +1 X m E ’S >» hm ■O E 0.25 0.2 0.15 0.1 0.05 0 Nov Feb May Jun Month of corm samples, from 2005 to 2006 Fig. 3. Corm dry weights in grams (X ± SE) for Triteleia ixioides subsp. anilina at Carpenter Ridge, from samples of neighboring individuals measured from 5 November 2005 to 8 June 2006. Old corms, producing the roots, leaf and flowering stem were depleted as aboveground growth continued during 2006. New corms, without roots, formed on top of the old, spent corms and permitted the plant to persist in a dormant state through the summer drought. One-way ANOVA (P = 0.0001) and Tukey-Kramer comparisons indicated that weights for old corms, November through May, were all significantly different from each other (P < 0.0001). New corm weight in June was significantly higher than in May (P < 0.0001). Sample sizes for the four harvests of corms were 25, 24, 28, and 29. Flower number per plant was statistically higher at Meadow than at Ridge in 2005 and 2009, the two years when both sites were sampled (Ta- ble 2). The highest mean number of flowers per plant for all years was 11.2 ± 0.9 at Meadow in 2009. Flowers opened by mid morning and remained open for 2-3 days. Anthers dehisced and stigmas appeared receptive at anthesis. Ovule number per ovary ranged from six to 22, depending on site and year (Table 3), but overall, mean number of ovules per ovary was generally consistent, with a mean value for all years of 12.9 ± 0.1 (n = 592). Mean ovules per ovary at Meadow was not significantly different in any of the five study years (Table 3). The pollen : ovule ratio from 24 flower buds sampled at Meadow 25 June 2004 was 905: 1 ± 70 (Kannely 2005), with 12,951 ± 1022 SE pollen grains per flower. At Ridge in 2007, 97% of the plants (n = 32) with unbagged inflorescences set fruits, compared with (19%) that were bagged. Number of fruits per plant averaged 2.6 ± 0.24 in unbagged plants and 0.2 ± 0.09 in bagged plants. Nectar production occurred simultaneously with anther dehiscence. This caused nectar samples to be easily contaminated with pollen and not suitable for a test. Nectar samples had to be combined to get a sufficient quantity to be measurable. Nectar from several flowers at Creek on 22 July 2006 read 35% sucrose equivalents. Flower-visitors. Of the 4281 flowers observed in plots, 789 (18.4%) were visited by insects. A bee fly, Bombylius facialis (Cresson 1919) (Bombylii- dae) (Fig. 2d, e) was the most frequent flower visitor; 721 (91.6%) of the visits were by this species. Other visitors included bumblebees (Api- dae), solitary bees (including Megachilidae), syrphid flies (Syrphidae), an additional species of bee fly (Bombyliidae), and butterflies (Lycae- nidae and Nymphalidae) (Appendix 1). Further quantification of insect visits to flowers is presented here only for B. facialis. Numbers of Bombylius individuals that entered a single plot in 10 min ranged from zero to eight. Percentages of the plants in a single plot visited by at least one fly ranged from zero to 100%, with up to a maximum of 14 separate plants visited in sequence by a single fly that entered a plot, in 10 minutes (Fig. 4). Flies were usually scarce by mid-afternoon, probably due to depletion of the nectar in the flowers and/or very high air temperatures (e.g., 29-30°C). Flies were occa- sionally observed probing flowers of Mimulus torreyi or Calyptridium monospermum Greene, with flowers only about half as high off the ground as the Triteleia flowers, but nearly all the flies were seen only on Triteleia flowers. Early in the season (e.g., at Ridge on 14-20 June 2005) there were few sightings of Bombylius 94 MADRONO [Vol. 61 Table 2. Comparisons of the Mean ± SE (Range) for Number of Flowers per Plant, Fruits per Plant, and % Fruit Set in Triteleia ixioides Subsp. anilina at Three Sites in 2004-2009. A One-Way ANOVA indicating P = 0.000 for each parameter, was followed by Tukey-Kramer comparisons. In each column, means with the same superscript letters do not differ significantly. Year Site n Flowers per plant Fruits per plant % fruit set 2004 Meadow 26 7.2 ± 4.5 ± O.S"^’ 65.0 ± 4.5"*= (2-18) (0-13) (0-100) 2005 Ridge 26 5.0 ± 0.4^^ 3.6 ± 0.3"‘’ 73.8 ± 3.3"*=“* (2-11) (2-7) (40-100) Meadow 30 9.8 ± l.P*^ 8.0 ± 0.9"^*" 84.0 ± 3.3"^* (2-26) (2-22) (30-100) 2006 Ridge 31 8.8 ± 0.8‘‘*^^ 4.9 ± 0.4"'=*' 58.8 ± 3.5*= (3-27) (2-12) (18.2-100) 2007 Ridge 32 7.1 ± 0.7'"‘^‘' 5.0 ± 0.5"*="^ 72.2 ± 3.6"*=^* (2-19) (0-16) (0-100) Meadow 28 8.2 ± 6.0 ± 0.7"^*"*' 71.9 ± 4.0"*=^ (3-21) (2-19) (22.2-100) 2008 Ridge 33 5.4 ± 0.5‘^‘' 2.2 ± 0.3*= 38.5 ± 4.8" (2-18) (0-7) (0-100) Meadow 32 8.7 ± 1.0^'^‘^^ 7.7 ± 0.9"‘*"*' 89.1 ± 2.3" (2-22) (2-21) (50-100) 2009 Ridge 32 7.3 ± 0.5"*"^ 4.8 ± 0.4"*=*" 66.6 ± 3.4"*= (3-16) (2-13) (33.3-100) Meadow 30 11.2 ± 0.9^ 9.1 ± 0.9" 79.1 ± 3.5""^ (4-21) (2-20) (33.3-100) Creek 35 10.5 ± 0.9"" 8.6 ± 0.8"^* 79.9 ± 2.5""^* (3-26) (2-22) (43-100) or other visitors); plot data from these dates are not illustrated. Most other days at both Ridge and Meadow had obvious active visitation by the flies. Percent of both plants and flowers visited by Bombylius facialis for these days is summarized by 1-hour periods at Ridge (Fig. 5) and at Meadow (Fig. 6). Additional data (including non-Bombylius visitors) for all plots covered in Figures 4-6 are shown in Appendix 1. All bee flies collected had masses of pollen grains at the top of their non-retractable tongues (Fig. 2e), or elsewhere on the lower part of the head — positions that could contact a Triteleia stigma while a fly probed for nectar. Features of fruits. Mean number of fruits per plant with one or more seeds for the entire study was 5.9 ± 0.2, but varied (Table 2) from zero to 22. In 2005, 2008, and 2009, three of the four years that Meadow and Ridge could be compared for fruits per plant. Meadow had statistically higher fruit set. Mean % fruit set per plant was 71.1 ± 1.3 in the 335 plants examined in this study. Percent fruit set per plant ranged from zero to 100, but 2008 was the only year where Meadow had a significantly higher percent of the flowers form- ing fruits than did Ridge (Table 2). In recording phenological features of flowering and fruiting plants, we found that by early July at Ridge and late July at Meadow and Creek, scapes and leaves were dead; fruits were mature and beginning to dehisce. As the summer progressed the fruit valves spread widely apart, permitting seeds to drop to the ground or to be flung out when scapes were moved by the wind or brushed by an animal. A few seeds remained in some fruits until fall. Dead scapes that did not get blown away or moved by animals stood into the winter if the snow was not deep, but under the weight of snow scapes were often pressed to the ground and were not visible by spring. Numbers of mature seeds in the 592 fruits (from 296 plants counted in the six years of this study) ranged from 0-19 (Table 3), with an overall mean of 5.0 ± 0.1 seeds per fruit. Undeveloped ovules and seeds counted in sam- pled fruits indicated that reproductive success in terms of percent seed set, was low, with yearly means ranging from 26.4 to 56.3 (Table 3). The mean seed set combining all sites and years in this study was 38.4% ± 0.9%. No indication of predation on maturing seeds was found. Seeds and Seedlings Seed weight and viability. In 2005, Triteleia mean seed weight was higher (X ± SE) at Ridge (1.60 ± 0.03 mg) than at Meadow (1.24 ± 0.02 mg) (P = < 0.001; n = 200). Seeds collected 9 August 2008 at Creek averaged 1.23 ± 0.05 mg 2014] KANNELY AND SCHLISING: BIOLOGY OF THE GEOPHYTE, TRITELEIA I XI OWES 95 Table 3. Year and Site Comparisons of the Mean ± SE (Range) Number of Ovules per Ovary, Seeds PER Fruit and % Seed Set in Triteleia ixioides subsp. anilina at Three Sites in 2004-2009. A One-Way ANOVA indicating P = 0.000 for each parameter, was followed by Tukey-Kramer comparisons. In each column, means with the same superscript letters do not differ significantly. Year Site n Ovules per ovary Seeds per fruit % seed set 2004 Meadow 24 13.3 ± 0.5"" 4.5 ± 0.5"" 33.3 ± 3.1"" (11-18) (2-10) (11.5M5.8) 2005 Ridge 26 11.9 ± 0.5^ 6.9 ± 0.6" 56.3 ± 3.4" (6-17) (2-14) (22.7-91.3) Meadow 28 12.1 ± 0.5“ 4.5 ± 0.4"" 32.9 ± 2.7"" (7-16) (0-8) (0-62.5) 2006 Ridge 31 11.2 ± 0.5‘' 4.7 ± 03"" 42.3 ± 2.9"^" (8-16) (1-9) (12.5-85.0) 2007 Ridge 32 13.0 ± 0.5"“ 6.8 ± 0.4" 50.8 ± 2.2"^" (9-18) (3-13) (26.3-80.0) Meadow 28 13.1 ± 0.5"“ 6.0 ± 0.7" 26.4 ± 2.6" (^18) (2-19) (8.3-58.6) 2008 Meadow 32 12.4 ± 0.3“ 3.6 ± 0.2" 28.9 ± 1.5"" (9-16) (2-6) (13M4) 2009 Ridge 32 14.5 ± 0.5"" 6.8 ± 0.5" 46.1 ± 3.0"^ (11-22) (2-14) (18.2-86.7) Meadow 30 11.4 ± 0.4“ 3.4 ± 0.3" 30.0 ± 2.6" (7-16) (1-8) (5.6-64.3) Creek 32 16.0 ± 0.4" 5.5 ± 0.3"" 34.2 ± 1.5"" (12-22) (3-9) (18.8-53.0) (n = 200). All embryos of the 100 Ridge seeds tested with tetrazolium chloride in 2005 stained bright red, implying high viability of seeds. Seed germination and seedling growth. Recov- cries from batches of 50 seeds on newsprint at all study sites indicated that 88 to 100% of the seeds germinated. Weather extremes and poor accessi- bility caused results to be spotty and incomplete for parts of this study (Table 4). Time of germination and earliest seedling growth was documented only at Ridge in 2005- 2006, since plantings at Meadow in 2005 (and at Meadow and Creek in 2008) were not accessible until late the following springs due to heavy snows. At Ridge, sprouts on seeds measured on 10 February (Table 4) showed that even at 1383 m elevation, Triteleia ixioides seeds had germinated and that seedlings were growing during the coldest period of the year. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Nymber of Triteleia plants visited by a fly entering a plot Fig. 4. Number of Triteleia plants visited in sequence by 85 individual Bombylius flies entering square-meter plots. Occurrences are combined from both sites and all dates shown in Figs. 5 and 6. 96 MADRONO [VoL 61 [lOa.m. 11a.m. 1p.m. 2p.m.j 30 May 2007 [10a.m. 1p.m. 2p.m.] [10a.m. 11a.m. 1p.m. 2p.m.] [11a.m. 1p.m.] 8 June 2007 11 June 2008 15 June 2008 Fig. 5. Percent of Triteleia plants and flowers in plots at Ridge visited at least one time by Bombylius facialis in five consecutive 10-minute observation periods on dates when the flies were active. Note that dates for 2007 and 2008 are both included. The first 10-minute period started on the hour listed, but the other four periods in the hour started at slightly different times, as new plots were set up. Numbers of plants and flowers are shown in Appendix 1. As sprouts grew, the tip of the cotyledon stayed in the seed, but the base of the cotyledon attached to the minute seedling axis often penetrated the newsprint. Later stages of growth in the spring months were sometimes inhibited by the newsprint, and only some, or portions of, seedlings could be reliably and accurately mea- sured. For example, length of the first “true” leaf (arising on the shoot near the base of the cotyledon) was reliably measured in several instances, but length of seedling roots could be assessed well only one time (5 May 2006, in Table 4). On 5 May at Ridge all seedlings’ green first leaves were visible above the soil surface. Several of the seedlings at Ridge on 8 June 2009 showed initiation of the corm. There was a short region, up to 1.5 mm wide (wider than the base of the axis with the first leaf) that indicated early growth of the corm. More complete data were obtained at Creek on 28 June 2009. Here all seedlings were apparently near the end of first- year growth, with single leaves projecting above- ground and drying. First-year plants of Triteleia ixioides at Creek produced corms with mean 60 I 30 ^ 20 c 0) i* a 10 0 ii mean % plants visited M mean % flowers visited 11a.m., 15 July 12p.m., 15 July 11a.m., 16 July 1p.m., 16 July Fig. 6. Percent of Triteleia plants and flowers in plots at Meadow visited at least one time by Bombylius facialis in five consecutive 10-minute observation periods on dates when the flies were active in 2005. The first 10-minute period started on the hour listed, but the other four periods in the hour started at slightly different times, as new plots were set up. Data are shown as means from simultaneous observations by two people each hour. Although fewer plots were studied at Meadow than at Ridge (Fig. 5), these data illustrate the importance of Bombylius facialis at this higher elevation site. Numbers of plants and flowers are shown in Appendix 1. 2014] KANNELY AND SCHLISING: BIOLOGY OF THE GEOPHYTE, TRITELEIA IXIOIDES 97 Table 4. Selected Features of Triteleia Seed Germination and Seedlings, from Experimental Plantings. All length measurements are in mm, as X ± SE (n). The n (number of samples per planting) for % seeds germinated, lengths of sprout, first leaf, and contractile root may be fewer than the 50 seeds planted, due to loss or breakage. Mean length of first leaf of seedling on 8 June 2006 (from seeds planted 16 September 2005) was significantly longer at Ridge than at Meadow (P < 0.0001). For seedlings collected at Creek on 28 June 2009 leaves and roots were too desiccated or broken to provide accurate measurements, but mean diameter of first-year corms on these seedlings was 1.64 ± 0,075 mm. *Contraction wrinkling was seen in the proximal part of 28 of the 41 roots measured; primary roots were deteriorated, **A small swelling above top of root indicated start of corm formation in 45 of the 50 seedlings. Study site Date seeds planted Date of % seeds seedling collection germinated (n) Length of “sprout” from seed coat (n) Length of first leaf, from point of cotyledon attachment to the seed, to tip (n) Length of contractile root from base of cotyledon (n) Ridge 16 Sep 2005 10 Feb 2006 94 22.11 ± 2,17 _ _ (47) (47) (4 (4 5 May 2006 100 14.08 ± 0.83 16.13 ± 1.00* (48) (4 (40) (41) 8 June 2006 100 _ 48.4 ± 1.69" (48) (4 (43) (4 Meadow 16 Sep 2005 8 June 2006 88 _ 15.7 ± 0.68" _ (44) (4 (42) (4 Creek 5 Oct 2008 8 June 2009 100 _ 60.36 ± 1.51 9.94 ± 0.68** (50) (4 (50) (50) 28 June 2009 100 _ (39) (4 (4 (4 diameters of 1.64 (±0.08 SE) mm before the young plants (n = 39) went dormant in the summer. Other plantings at Ridge and Meadow provided counts of seeds that germinated (Ta- ble 4), indicating 88 to 100% germination. Seedlings were observed to have a contractile root. It appeared lateral to the primary root (Fig, 2f), at the base of the stern, where the corm forms. Figure 2g shows a large specimen, with the contractile root wider than the developing corm. Discussion Growth of Plants During Different Times of the Year Measurements of mature scape heights in different sites and years indicate there is variation in this aspect of growth (Fig. 1), as might be expected due to somewhat different microhabi- tats. Mature heights were reached just two to three weeks after the scape first appeared aboveground in May or June, and this occurred about a month earlier at the lowest elevation site (Ridge). We have also documented a “wasting syndrome” appearing in the scapes that heavily impaired reproduction in the Meadow popula- tion in 2007, but was not seen here in other years. Similar symptoms were reported for two popu- lations of Triteleia iaxa Benth. in grassland habitats in the northern Sacramento Valley in 1999 and 2001 (Schlising and Chamberlain 2006). The cause of the symptoms or name of the disease has not been determined for either species, but these field observations illustrate a potential stress factor for T. ixioides. Our field studies have shown that there is considerable subterranean growth in T. ixioides under natural conditions during the Mediterra- nean winter. Even with the low winter tempera- tures at 1383 meters elevation (or higher), considerable growth of the new corm occurs (Fig, 3). This represents an important variation in strategy when compared to herbaceous peren- nials that are winter-dormant. We also illustrate how cormous geophytes like T. ixioides have plant bodies that last only a single year, since corms (Figs. 2-3) as well as aboveground parts are totally replaced each year. As the old rooted corm is expended producing a new shoot and flowers, a new rootless corm is being created before spring flowering is complete. This shows that provisioning for the survival of the individual plant body precedes reproduction by seed. New, rootless corms spend the late summer in a dormant state and begin to develop roots with the fall rains. Thus it appears that moisture availability may have a larger influence than temperature on breaking dormancy, a subject which may warrant further study. Bee Flies as Pollinators of Triteleia Flowers Our fieldwork provides a detailed example of a single species of bee fly as the primary pollinator 98 MADRONO [VoL 61 of Triteleia ixioides in Butte County. Bombylius facialis has been listed earlier, but without detail, as a pollinator of the vernal pool plant Pogogyne abramsii J. T. Howell in San Diego County (Schiller et al. 2000). Other published information on bee flies using California plants mentions the genus Bombylius, but does not provide numerical information on the fly visits (e.g., Grant and Grant 1965) or does not identify the specific Bombylius studied (e.g., Schmitt 1983). Our information comes from close examination in 1- m^ plots over 110 separate 10-minute periods spread over different hours of the day (Figs. 4-6, Appendix 1). Thus we have obtained concrete, numerical documentation for a specific bee fly on a specific plant. Adult flies in the Bombyliidae are thought to be pollinators while they collect nectar (Grant and Grant 1965; Kastinger and Weber 2001). Interestingly, Panov (2007) found that bombyliid flies (including several species of long-tongued Bombylius in Austria), also ingested large amounts of pollen through their tongues during short flower visits. A few studies in North America suggest that bombyliids may take in pollen as well as nectar, as from Hedyotis L. in New York (Grimaldi 1988) and Commelina L. and Tradescantia L. in Florida (Deyrup 1988). In Triteleia, nectaries are located in the three septa that separate the chambers of the ovary (see Vogel 1998 on the closely-related T. lugens Greene). Nectar secreted in the septa oozes out and drains through shallow channels to the base of the flower. While using pipettes to extract nectar, pollen was often found mixed in with it. This suggests that Bombylius facialis may have opportunity to ingest pollen in the nectar of Triteleia flowers. When these flies probe flowers deeply with their extended tongues, abundant loose pollen accumulates on the base of the tongue and lower face (Fig. 2e); this pollen may brush onto stigmas during the probing, causing pollination. Our data show that in many cases a single fly may visit four to seven Triteleia plants (and less commonly up to 13 or 14 plants) in one square meter in just a 10-minute period (Fig. 4). Thus, they have great potential to significantly spread pollen among plants in a given population. Bombylius facialis was seen as the most frequent flower-visitor beginning when ambient temperature was at least 20°C. This species of fly is considered common in the lower elevations of this region (Neal Evenhuis, personal communi- cation). We have also observed this species at populations of T. ixioides subsp. anilina outside the study areas. However, we have evidence that along with B. facialis, bees and butterflies may also have importance as pollen vectors at the higher elevation sites such as Meadow (Appendix 1) and Creek. Additional quantitative field study is needed to document variation in pollen vectors for T. ixioides and for other species of Triteleia that have wide geographic and/or elevational ranges. Flower color in taxa of Triteleia ranges from yellow to white, lavender and blue. The larger- and blue-flowered T. laxa, for example, observed at low elevations in the northern Sacramento Valley, was not visited by bombyliid flies, but rather by large butterflies and by bees (Cham- berlain and Schlising 2008). Reproductive Success, Based on Production of Fruits and Seeds We have shown that reproductive success (based on fruit-set and seed set) varies between years and sites in our region (Tables 2 and 3). Definitive causes for this variation are not known, but could include pollen vector limita- tions and weather. Qualitative observations made during the flowering period suggested that soil water was not lacking. Our observations also show that pollen vectors are abundant, but only during warm weather. The short flowering period of Triteleia, along with temperature limitations of pollinators, suggests that vector activity does not always coincide favorably with flower receptivity to maximize seed set. The presence of undeveloped white ovules along with mature black seeds in all 592 fruits examined shows consistent seed set below max- imum potential. It is possible that low seed set (overall 38.4%) may be due to chromosomal anomalies, including polyploidy (as suggested by Lenz 1975). Lenz noted that for most populations of T. ixioides subsp. anilina he examined, plants had two sets of chromosomes {2n — 14); but he found one population in Sierra County where 2n = 42. Furthermore, he noted that at meiosis chromosomes in the polyploids form rings or chains, and after division, daughter nuclei do not develop further. The relationship between low seed set and polyploidy warrants further investi- gation. Despite low overall seed set, nearly all fruits we sampled set some seed. Our bagging tests indicated that there is a degree of self- compatibility at the Ridge population, but we did not do tests using self-pollen. Although bagged plants were exposed only to self-pollen, it was not applied by hand. Seed Germination and Seedling Growth in Relation to Mediterranean Climate Seeds germinated at our lowest elevation site sometime after the first fall rains rather than in the spring, a phenomenon typical of many plants in Mediterranean areas. Seedling growth contin- ued throughout the winter (Table 4) despite low air temperatures occurring at 1383 m elevation 2014] KANNELY AND SCHLISING: BIOLOGY OF THE GEOPHYTE, TRITELEIA IXIOIDES 99 (Ridge). Even with seasonal access difficulties, comparison of seedling morphology indicated that winter seed germination and seedling growth were similar at all three study sites. Harvesting batches of these seedlings at several times during their first year, permitted us to document the presence of contractile roots, which have also been described for Triteleia hyacinthina (Lindl.) Greene (Smith 1930, referred to as Brodiaea lactea (Lindl.) S. Watson; Putz 1992) and T. laxa (Schlising and Chamberlain 2006). Since contrac- tile root morphology and behavior is important in the context of plant growth under Mediterra- nean conditions, our documentation for T. ixioides warrants additional discussion of the scant literature here. More than a century ago (Rimbach 1902) briefly described contractile roots as “subterra- nean organs” in several California “lilies” (e.g., ScoUopus Torr., Trillium L., Zigadenus Michx.) noting that such roots may function in pulling, to different degrees, young bulbs of seedlings downward in the soil from where they were first produced. Jernstedt (1984) described contractile roots in detail for Chlorogalum pomeridianum (DC.) Kunth, a large and widespread geophyte of Mediterranean California. She illustrated how the proximal (top) portion shrinks due to differential collapse and elongation in tiers of cells. The continued downward growth of the distal, growing portion of the root helps pull the bulb downward into the “channel” left by the shrinking root. Piitz (1996) also described the “pulling force” of contractile roots in some detail. Our outdoor plantings of Triteleia ixioides demonstrated fast-growing contractile roots lat- eral to the primary roots (Fig, 2f), but did not permit us to measure the extent that the shrinking of the roots moved the young seedlings’ corms deeper in the soil. We found first-year contractile roots were significantly longer at Ridge than at Meadow (Table 4), perhaps because it was warm longer in the day, permitting more growth at the lower elevation. The authors mentioned above who focused on California plants, described these roots in the context of the Mediterranean life cycle, and the fast “planting” of the seedling corm or bulb before the summer drought. Additional numerical informa- tion (Schlising, unpublished data) demonstrates contractile roots in seedlings, for species of Brodiaea Sm., Calochortus Pursh, Dichelostemma Kunth, and Odontostomum Torr. in northern California. Contractile roots on seedlings were emphasized by Rimbach (1902), but he also noted that contractile roots can be formed again on older plants in the years after the seedling stage. Putz (1996) discussed the varying morphologies of contractile roots on both seedlings and older plants, and he noted that contractile roots are widespread geographically and occur in diverse habitats. In addition he illustrated this growth behavior as widespread in flowering plants in general - beyond the petaloid monocots that have been discussed here. An unusual example of eudicots with contractile roots in California was illustrated for Jepsonia heterandra Eastw. (Saxi- fragaceae) by Ornduff (1969). Despite earlier work, contractile roots remain a poorly known plant feature. In the recent book “Seedling Ecology and Evolution” (Leek et al. 2008) such roots are not described as a feature of Mediterranean plants. Observations from our field study suggest that the “fast planting” by contrac- tile roots in seedlings of Triteleia ixioides subsp. anilina may be a critical part of their adaptation to early summer drought in Mediterranean Califor- nia. This may apply especially to the movement of the first-year corm to a somewhat safer, less- desiccating depth for over-summering during the dry season. More study is needed on contractile root ecology in general, especially as part of the life cycle of California geophytes. Our field study has provided much information on life history and phenology for this species in nature. This is one of the two species of Triteleia for which much field information has been collected, with T. laxa being the other (Schlising and Chamberlain 2006; Chamberlain and Schlis- ing 2008). It is hoped that our work provides a basis for continued study of not only this species but some of the rare Triteleia taxa, is a starting point for comparative studies on variations in strategy, and assists in conservation of Califor- nia’s geophytes. Acknowledgments The authors thank the many people who helped measure or collect plant parts and record data in the field with us during the eight years of this study. We thank Jody Ryker and Scott Chamberlain for help with statistical analyses. We appreciate the identification of the major flower visitor by Neal Evenhuis, Bishop Museum. We are also grateful for helpful comments on our manuscript by John Dittes and by anonymous reviewers. Literature Cited Berg, R. Y. 2003. Development of ovule, embryo sac, and endosperm in Triteleia (Themidaceae) relative to taxonomy. American Journal of Botany 90: 937-948. Bradbeer, J. W. 1988. Seed dormancy and germina- tion. Chapman and Hall, New York, NY. California Native Plant Society (CNPS). 2013. Inventory of Rare and Endangered Plants (online edition, v8-02). California Native Plant Society, Sacramento, CA. Website http://www.cnps.org/ inventory (accessed July 2013). Chamberlain, S. A. and R. A. Schlising. 2008. Role of honey bees (Hymenoptera: Apidae) in the pollination biology of a California native plant, Triteleia laxa (Asparagales: Themidaceae). Envi- ronmental Entomology 37:808-816. 100 MADRONO [Vol. 61 Deyrup, M. a. 1988. Pollen-feeding in Poecilognathus punctipennis (Diptera; Bombyliidae). Florida Ento- mologist 71:597-605. Dilley, J., P. Wilson, and M. R. Mesler. 2000. The radiation of Calochortus: generalist flowers moving through a mosaic of potential pollinators. Oikos 89:209-222. Grant, V. and K. Grant. 1965. Flower pollination in the phlox family. Columbia University Press, New York, NY. Grimaldi, D. 1988. Bee-flies and bluets: Bombylius (Diptera: Bombyliidae) flower-constant on the distylous specids, Hedyotis caerula (Rubiaceae), and the manner of foraging. Journal of Natural History 22:1-10. Han, S. S. 2001. Flowering of three species of Brodiaea in relation to bulb size and source. Scientia Horticulturae 91:349-355. Hoover, R. F. 1941. A systematic study of Triteleia. American Midland Naturalist 25:73-100. Jernstedt, J. a. 1984. Seedling growth and root contraction in the soap plant, Chlorogalum pomer- idianum (Liliaceae). American Journal of Botany 71:69-75. Kannely, a. 2005. Preparation and quantification of entomophilous pollen using sonication and an area-counting technique. Madrono 52:267-269. Kastinger, C. and a. Weber. 2001. Bee-flies {Bombylius spp., Bombyliidae, Diptera) and the pollination of flowers. Flora 196:3-25. Leck, M. a., V. T. Parker, and R. L. Simpson. 2008. Seedling ecology and evolution. Cambridge Uni- versity Press, New York, NY. Lenz, L. W. 1975. A biosystematic study of Triteleia (Liliaceae). I. Revision of the species in section Calliprora. Aliso 8:221-258. . 1976. A biosystematic study of Triteleia (Lilia- ceae). II. Chromosome numbers and karyotypes of the species of section Calliprora. Aliso 8:353-377. Ornduff, R. 1969. Ecology, morphology, and systematics of Jepsonia (Saxifragaceae). Brittonia 21:286-298. Panov, A. A. 2007. Sex-related diet specificity in Bombylius major and some other Bombyliidae (Diptera). Entomological Review 87:812-821. Parsons, R. F. 2000. Monocotyledonous geophytes: comparisons of California with Victoria, Australia. Australian Journal of Botany 48:39-43. Pate, J. S. and K. W. Dixon. 1982. Tuberous, cormous and bulbous plants: biology of an adaptive strategy in Western Australia. University of Western Aus- tralia Press, Perth, Western Australia. Patterson, R. and T. J. Givnish. 2003. Geographic cohesion, chromosomal evolution, parallel adap- tive radiations, and consequent floral adaptations in Calochortus (Calochortaceae): evidence from a cpDNA phylogeny. New Phytologist 161:253-264. Pires, j. C. and G. Keator. 2012. Triteleia. Pp. 1512- 1514 in B. G. Baldwin, D. H. Goldman, D. J. Keil, R. Patterson, T. J. Rosatti, and D. H. Wilken (eds.), The Jepson Manual: vascular plants of California, Second Edition. University of Califor- nia Press, Berkeley, CA. Proches, S., R. M. Cowling, and D. R. du Preez. 2005. Patterns of geophyte diversity and storage organ size in the winter-rainfall region of southern Africa. Diversity and Distributions 11:101-109. PuTZ, N. 1996. Development and function of contrac- tile roots. Pp. 859-874 in Y. Waisel, A. Eshel, and U. Kafkafi (eds.). Plant roots: the hidden half Marcel Dekkar, Inc., New York, NY. Raunkiaer, C. 1934. The life forms of plants and statistical plant geography. Oxford University Press, London, England. Rimbach, a. 1902. Physiological observations on the subterranean organs of some Californian Liliaceae. Botanical Gazette 33:401-M20. Rundel, P. W. 1996. Monocotyledonous geophytes in the California flora. Madrono 43:355-368. Schiller, J. R., P. H. Zedler, and C. H. Black. 2000. The effect of density-dependent insect visits, flowering phenology, and plant size on seed set of the endangered vernal pool plant Pogogyne abram- sii (Lamiaceae) in natural compared to created pools. Wetlands 20:386-386. SCHLISING, R. A. AND S. A. CHAMBERLAIN. 2006. Biology of the geophytic lily, Triteleia laxa (Themidaceae), in grasslands of the northern Sacramento Valley. Madrono 53:321-341. Schmitt, J. 1983. Density-dependent pollinator forag- ing. Flowering phenology, and temporal pollen dispersal patterns in Linanthus bicolor. Evolution 37:1247-1257. Smith, F. H. 1930. The corm and contractile roots of Brodiaea lactea. American Journal of Botany 17:916-927. Tyler, C. and M. Borchert. 2002. Reproduction and growth of the chaparral geophyte, Zigadenus fremontii (Liliaceae), in relation to fire. Plant Ecology 165:11-20. Vogel, S. 1998. Remarkable nectaries: structure, ecology, organophyletic perspectives. III. Nectar ducts. Flora 193:113-131. Appendix 1. Number of Open Flowers on Plants of Triteleia ixioides Subsp. anilina, within Different One Meter^ Plots, and Quantification of Visits TO Flowers by Bombylius facialis during io-minute Periods at Ridge in 2007 and 2008 and at Meadow in 2005. Totals are sums for each of the five 10- minute periods done during the hour indicated. Visits by flies are listed separately for plot, plant, and flower. Percents (rounded) are given in parentheses. The last column lists other insects observed on Triteleia during the period (B = bumblebee, b = solitary bee, sy = syrphid fly, by == other bombyliid fly, lep = butterfly). 101 2014] KANNELY AND SCHLISING; BIOLOGY OF THE GEOPHYTE, TRITELEIA IXIOIDES c« C« +-> D O > (U > M o S 53 60 O- s ^ O o H C g « s 6^ X O C d 60 a 60 00 d ‘C ^ ^ u Cl, e cd w d s-i a S 01 ‘S ^ > O ^ > B c 3 s d ^ & 0) H S 00 5 c2 d m o 2 £ ‘-S -2 o ^ X m ir> ^ fn m ^ n(Nm\i“OChr^r-in'^ooi>-om^mmr^<=-nm(NrsioO’^0'^fONoir)in'-H,-tO(N,— icNOOOOO in'^NONnoooominocO'^'>--iNOir)’^ooirN<--^Nnir>NOmorNSvor'iONmNn r^rMO'“ifNfNOOoo»n(N^NoooNooooooNm-HOONO S d o o g d o o ri 1 1:00 a.m. Appendix 1 . Continued. 102 MADRONO [Vol. 61 +-* 0) o > w g o •S ^ u, 60 I -I o 'I #1 cd ao H G C (U cd g -Q .S X) g G 5 w (/) S S 0) ^ o G S fl QO U > ^ (U (U C/D 00 s « ^ a o -o 'S s S >> ® a G c s ^ 3 cd fi SBt S a S:s > 60 w f") w 'C .5 s (D a Uh ^ ^ i-i a I a c +-> c i-{ S G t-4 2 o s i G ■£ ^ a£ (D Wi fi 2 « '— ^ on '■G (U VI o .a > o o G I ^ ^00 ^ O ^ I to 'ovo'o fo^r-ONiT) r- VO ^os o 04 00 ON VO 00 inoo'^oor^ONONt^ r^fNOOomvO'— 'Or^oo-^r-vo^-iooin'rj-roooNOOmooo o o o m ^ VO VO VO O 04 O ^ 00 m On ’-H VO 00 /— N O /~s /— N O ooomvooNO'-i lONoooo-— 'f-- o^-iON^ovooNmr^fNooom-H^ON oor^r^ON-^fMocoNino— 'OOmo O^M^O'-^C'irvi'-H^ooo^'-i’-HfNirNioom '^r4fM«N'^fn0m00'^0(N ff^fN|^^000'-^r4^'— ifNOOVO'^'^fNfOr^OO'^OinOOeNOfN ovo^nvor^O(N'-HON^OMO'--^'^'-ivoooornvomvnoNO‘noNOOir)ONONOfnvo <^(Nr4r4'-'m^ ir)^mm'^‘n»nm»-iONONt^ONOr40'^oo'^ooN‘ONin^ONONr^r^— ^r-inooocN s a o o s a o o S a o o (N Appendix 1 . Continued. 2014] KANNELY AND SCHLISING; BIOLOGY OF THE GEOPHYTE, TRITELEIA IXIOIDES 103 (U ^ O ■S ^ W) ^ .S o -2 > f 1 « -o 2 >> g ■> « c Cfl W ^ ^ fl ^ O O O H a >v >> X) g -d § 3 s:2 •;< > cd w s fl ’a, ^ ^ ’2 ^ ^ ^ a ^ d >, o ^ ^ X H OJ c t Sst Ss > feO ^ fi <>-s .2 'C -§ X IS 3c2 m « > X d .d O u X X d 2 « g- c. O o X a jj X x" a a >.Xx x''x X X X X m ^cn (N r- fsj p cn 00 rn tn o m rn o© (Nmocn'<;f-oo>r)OOOOTO inr^X»r)rsirs|fnir)'^‘r>Cs!fn xooooxo^oo^'m'^x»-lmootr^xo^of^^^r)r--o^^ mcNOOX'^'^mooin— iini>ooininX'?t-^^^<-OOxr-OOX'^^X»-'OXenXOfNONXXOX ni^OrNifNmf\|(NOOCNi^rN|>— 'OOrs|cN^fN»-i^^'— ifNCN'-i o — i ^ m — I o f<-)^0»-ifn'^mmoOfn^(N^OOfNfn(Nm— I— irs|^ C/D O •S tH W) u _g 6 ’I #1 'd S A, cd (U H G ^ >v ^ 0) cd fi >. O XI d “d § ^ S .s •s > cd w ^ i "a X (U , O ! d « , d cr (U • ^ T3 ^ ^ ^ d 'S d >^ d -5 ^ C r^ ^ d 5 s ° o -S X C CJ B'^ ^ d d d S o-g ‘S ^ ^ d ^ > bO d Oi d ,0 d ^-1 a ^ n o X ^ t3 0^ ■S| c/^ C S « d (u E ^ ^ o d « o dl^' f’B S o w X m 00 oo ^ ro CN xxr--(NOr^ON^— lO fO'd-^fNOXXX'-nO ^n-d-'^CNior-r-o CNl — < -H ^ O fN O fN rs| (N — I o '^xio'^oooxmorn ^^fN— I— ^CN—iCsI-HrNl Tt r4 — I m . u S&o X d o o O d • • 3 m in height only Bird 1+ 53 209 344 8 of height (P - 1.2 X 10-'^ n = 2300). For %Green and total number of branches, greater branching tended to show lower frequencies in the higher %Green categories (P < LO X 10“^°°, n = 12,202), and the results remain significant when isolating only plants 4-5.9 m in height (P = 0.03, n - 2605). By height, %Green had an unusual (and significant) pattern. The two higher classes %Green 3 and 4, clearly had greater frequencies of plants at lower heights (i.e., browning increases with height), but in the low %Green categories 1 and 2, the distribution was more even and any trends were less apparent (Fig. 2). Girdling Girdling is not independent of epidermal browning (as described above). Girdling and damage are not independent, as there is more girdling on damaged plants and non-girdled plants have a relatively lower frequency of damage (P == 1.4 X 10“^^ n = 11,297; height controlled: P - 0.00012, n - 2301). Girdling increases with height, but girdling is independent of nurse species, total number of branches and bird cavities. As observed below, topping may be linked to girdling. Damage As noted above, %Green and girdling are statistically related to damage. Not surprisingly, damage increases with plant height, as well as with bird cavities (which are not a component of damage) (damage and bird cavities: P = 2.2 X 10“^^ n = 11,328; height controlled: P = 1.4 X 10“®, n — 2309). Damage is independent of total number of branches when height is controlled for (in the 4-5.9 m tall subset). A total of 114 plants were identified as having some sort of fire damage and 345 plants as having gun damage. Seven plants were identified as having lightning damage, while nearly 10,000 of the plants were identified as having no damage. Further, 257 plants were identified as being topped (that is, the top portion of the main stem was absent). Of these, 178 were identified as being 75- 100% green, and 22 had no bird cavities. Seventy- five percent (191/254) however, were identified as being girdled. By comparison 5504/12,205 (45%) plants were girdled in the entire dataset. Some types of damage are provided in Table 2. Bird Cavities In addition to the relationships outlined above, bird cavities and nurse species is marginally insignificant (P = 0.053, n = 1931), though it should be noted that sample sizes are small as cacti with sampled nurses are primarily the smallest plants, while birds tend to establish on taller plants. In the control subset of plants 1- 1.9 m in height, only one bird cavity was recorded among the three main nurse species. Chi-square results show that plant height (P = 1.6 X 10“^^^, n = 6042 for plants 2 m and taller) and number of branches (height-controlled: P — 7.1 X 10“’°, n = 2614) are both significantly related to bird cavities. Correlation results to determine which of these two variables (plant height or number of branches) is the better predictor of bird cavities, particularly considering the very strong relationship between plant height and branching, show that plant height is a superior predictor of cavities (P = 5.7 X 10“’^^, n — 6042) than number of branches is in predicting cavities (P = 2.2 X 10“^^, n = 6042). Interestingly, total number of branches had a stronger relationship with bird cavities (P — 2,2 X 10“°^) than number of primary branches (P = 7.0 X 10“^^, n = 6042), and secondary (P = 3.8 X 10“^^, n = 6042) (and tertiary) branches had weaker correlations still, though all were signif- icant, likely due in part to the large sample size. Branches, Nurse Plants, and Height The distribution of branches on Carnegiea by height is shown in Fig. 3. The strongest correla- 120 MADRONO [VoL 61 ToGreen 1 %Green 2 Fig. 2. Camegiea plant frequencies by different height classes in each epidermal browning class. %Green represents the amount of the plant that does NOT have epidermal browning and classifications are: %Green 1: <50% green; %Green 2: 50-75% green; %Green 3: 75-99%; %Green 4: 100% green. Height classes are 0: plants 0- 0.9 m in height; 1: 1.0-1. 9 m; 2: 2,0-2. 9 m, etc. to 8+ which includes all plants 8.0 m in height and taller. For these charts, some classes were merged for ease of interpretation. For example, most plants with no epidermal browning (%Green 4) are 0-0.9 m in height. tion is between height and number of primary branches (P < 1.0 X 10“^“, n = 6042, F = 2580.8), followed by height and total number of branches (P < 1.0 X 10“^“, n = 6042, F = 2327.7), with secondary (P = 5.2 X 10“", n = 6042) branches (only 14 plants have tertiary branches) exhibiting a weaker relationship with height. Table 2. Frequencies of Various Classifications of Damage (Fire, Gun, Lightning, Topped (Decapitated), Tree, and No Damage [None]) and Other Variables Including Girdling (Absent, Present) and Epidermal Browning %Green Categories (1: <50% Green; 2: 50-75% Green; 3: 75-99%; 4: 100% Green, le., no Epidermal Browning). Girdling Epidermal browning A P %Green 1 %Green 2 % Green 3 %Green 4 Fire ' 66 44 35 44 31 0 Gun 118 214 9 96 221 6 Lightning 0 7 0 4 3 0 Topped 63 191 20 59 175 3 Tree 185 450 14 71 539 10 None 6009 3949 87 489 6037 3340 2014] DANZER AND DREZNER: CARNEGIEA DAMAGE 121 8+ 7.0- 7.9 6.0- 6.9 5.0- 5.9 4.0A.9 3.0- 3.9 2.0- 2.9 ■ None H Primary □ Secondary & Tertiary 0 10 20 30 40 50 60 70 80 90 100 Fig. 3. Distribution of branches by height starting at two m. None represents plants with no branches. Primary represents the frequency of plants (not number of branches) that have only primary branches (those emanating from the main stem). Secondary and Tertiary represents the number of plants that have at least one branch emanating off of a branch. Each height bar is scaled to 100% of plants in that height class, so the branch classes are depicted as a proportion rather than raw frequency. The height classes are in meters (e.g., the shortest class includes plants between 2. 0-2. 9 m in height). 8+ includes all plants 8.0 m in height and taller. Finally, there were 5220 plants between 0-1.0 m in height. Of these, 50.3% are nursed by Ambrosia deltoidea, 17.7% by Larrea tridentata, and 10.8% Cercidium microphyllum. A wide variety of other species make up the remainder of the nurse identifications. Discussion Epidermal Browning Epidermal browning is associated with scaling and barking as well as loss of spines (Evans et al. 1992) on a variety of columnar cactus species (Evans and Fehling 1994; Evans et al. 1994b). It is far more common on the south-facing side of Carnegiea plants (Evans et al. 1992), and in South American cacti, epidermal browning is predom- inantly on the northern equator-facing sides of plants (Evans et al. 1994b) and browning is lessened on those plants shaded on their sun- facing side by larger nurse plants (Evans et al. 1994a). The pattern is widespread over large intra-continental areas (Evans et al. 1992). Epidermal browning is not related to trace metal pollution (Kolberg and Lajtha 1997). Mounting evidence has linked browning to solar radiation (Evans et al. 2001), and they suggest more specifically that UV-B radiation causes browning, though other work does not confirm this (Lajtha et al. 1997). At Saguaro National Park, epidermal browning has been linked with premature death of these cacti (Evans et al. 2005). Other factors that may contribute to epidermal browning should be considered, such as tissue heating (Drezner 2011) above lethal temperatures (McAuliffe 1996). Regardless, epidermal brown- ing may predict future longevity of plants and may result in premature mortality (Evans et al. 2005). Epidermal browning is found across numerous taxa, including across many genera of columnar cacti throughout the Americas (Evans and Fehling 1994; Evans et al. 1994b; Evans 2005) as well as several succulent South African Euphorbia species (Evans and Abela 2011). The association between height and browning has been recognized (Duriscoe and Graban 1992), where browning increases with Carnegiea height. Of plants with greater epidermal brown- ing (categories 1 and 2 of %Green), the frequen- cies are rather evenly distributed (fluctuating from 5-15% in category 1 plants across the one meter height classes to nine meters, and for category 2, from 5-18%). That is, plants with greater epidermal browning are fairly evenly distributed across the height spectrum. Converse- ly, of individuals with little to no browning, the 0-0.9 m height class has 28% of all category 3 %Green individuals, and 82% of all category 4 individuals. Thus, individuals that have little epidermal browning are dominantly smaller plants, but plants with high amounts of epider- mal browning span the entire height range relatively evenly. There were significantly more bird cavities associated with plants with a lower %Green classification (height-controlled). Whether this is due to avian preferences for plants, for example, with fewer spines (associated with epidermal browning), or whether bird cavities may act to hasten browning is uncertain. Plants with greater girdling also appear to be less green, particularly apparent in the significantly low frequency of plants that are in the highest %Green class and also girdled. Plants without recorded natural and anthropogenic damage are more frequent in the higher %Green classes. Even branching appears to be related to %Green, where browning 122 MADRONO [Voi„ 61 increases with branches. While solar radiation is surely the primary cause of epidermal browning, perhaps these other factors weaken the plant or hasten the browning process. Girdling Greater girdling is observed on damaged plants (damages not associated with rodent action). It is notable that girdling increases with height. Taller plants may be more girdled simply due to greater time periods over which there was opportunity for small mammals to cause the girdling. It is also possible that several variables are confounded, particularly in relation to height. Spine loss increases with epidermal browning and thus presumably with height (Duriscoe and Graban 1992), which may increase the attractiveness of taller plants to small mammals for girdling. Epidermal browning is related to girdling; whether this relationship is direct (girdling hastens browning or browning draws rodents for girdling), or confounded by another variable confounded by height is uncertain though plants with no epidermal browning are disproportion- ately ungirdled while girdled plants are under- represented in the highest percent green class. This supports past observations that epidermal browning may hasten death and provides some indication of the overall health of the plant (Evans et ah 2003; 2005). It is possible that epidermal browning may be hastened by damage. Also, as Carnegiea plants get taller, they generally get thicker (Drezner 2003a), so greater available plant material may draw rodents, increasing girdling on taller plants. However, the chi-square results for the height-controlled subset of the sample maintain the significant relationship between girdling and epidermal browning (%Green) and the relationship between girdling and damage (though bird cavities are indepen- dent of girdling). Perhaps plants are weakened by girdling making them more prone to at least some types of damage (Steenbergh and Lowe 1977, 1983); damage classifications were primarily tree, topped, gun, and fire. The data (Table 2) show that while fire frequency data are similarly distributed to the no damage data, girdling is disproportionately more common for topping, tree and gun damage. Damaged plants may also be compromised making them more attractive to rodents for girdling. Also, small mammals may seek out plants with greater epidermal browning and spine loss. Rogers (1985) suggests that Carnegiea individuals with spine loss may be at greater risk of contact with cattle. Despite the possibility that girdling may hasten mortality by increasing susceptibility to wind- throw and freezing, this is relatively rare (Steen- bergh and Lowe 1977, 1983). Nurse species, and thus architecture (trees versus the open Larrea shrub nurse), is also independent of rodent girdling of cacti. Damage Since plants get taller with age, it is not surprising that damage increases with height, as there is more time over which damage can occur. In fact, the damage classifications are inherently height related. Potential damage from a tree or topping increase with height, as does the likeli- hood of being struck by lightning as the tallest individuals are disproportionately struck (Steen- bergh 1972), or being used as a target for gunfire. Estimates suggest that lightning kills, “much less than 0.1% of the population per year” (Steen- bergh 1972; Steenbergh and Lowe 1983, pp. 137). Our dataset has only seven lightning strike identifications that represent 0.06% of the plants in total, in strong agreement with past observa- tions (Steenbergh 1972). Our value is not adjusted per year, however Steenbergh (1972) suggests decomposition is rapid in such cases. Decapita- tion was documented in 2% of the sample. Of those 2%, nearly 70% had less than 25% epidermal browning. Girdling, however, was documented in 75% of decapitated plants, compared to the dataset average of 45% girdled plants. Damage increases with bird cavities (also when controlling for height). In the northern Sonoran Desert, Carnegiea decapitation (= topping) may occur where freezing occurred at a bird cavity (Steenbergh and Lowe 1977). Bird cavities thus increase a plant’s susceptibility to topping (Steenbergh and Lowe 1983; McAuliffe and Hendricks 1988), but other damage types are less clear in their possible association with bird cavities. For example, perhaps some gunfire targets birds rather than the plants, or that birds may utilize bullet entry points for nest excava- tion. When isolating plants that are 5. 0-6. 9 m in height only, the observed number of plants v/ith bird cavities and with gun damage (versus no damage) is significantly higher than expected (data not shown). The impact of bird cavities naturally on Carnegiea gigantea populations is relatively small (Steenbergh and Lowe 1983). Branches however, is independent of damage. The presence of combustible material or a substantial ground covering of grass or other vegetation that encourages fire burn and spread is strongly related to Carnegiea gigantea fire dam- age and mortality, while Carnegiea in more barren or rocky areas appear to suffer far less (Steenbergh and Lowe 1977). Mortality may exceed 70% in burned areas (Rogers 1985). Smaller plants are particularly susceptible and may be underrepresented in estimates (Rogers 1985). Our dataset shows 110 plants (Table 2) with fire damage, representing less than 1% of the 2014] DANZER AND DREZNER: CARNEGIEA DAMAGE 123 population. Not surprisingly, buffelgrass {Penni- setum ciliare (L.) Link), a major contributor to combustible ground cover and fire damage in the Sonoran Desert [Schiermeier 2005]) is not well established at this site. Ground cover at this site is rather sparse except during unusually wet winters when a generous cover of annuals becomes established. However, as an active military training site, the potential for ignition is ever present. Bird Cavities Bird cavities are significantly related to both plant height and to total number of branches (which are themselves strongly related [Drezner 2003b]). Korol and Hutto (1984) found a positive association between bird cavities and branching, and large, old, often heavily branched Carnegiea may support 50 woodpecker cavities (Steenbergh and Lowe 1983). Correlation indicates that total number of branches is a stronger predictor of bird cavities than number of primary branches, further supporting the overall importance of branches for presence of avian nesting cavities. However, our correlation results show that number of bird cavities is better predicted by plant height than by number of branches. While branching increases available ‘real estate’ for bird cavities (and is a significant predictor), height appears to be the superior predictor. On the highly branched Pachycereus pringlei (S. Watson) Britton & Rose (cardon), bird cavities are oriented in directions with fewer branches in order to maximize visibility (Zwartjes and Nor- dell 1998), leading to the possibility that branch- es, or excessive branching, may result in some bird avoidance of such highly branched cacti. Further, branches act as a windbreak, which is detrimental for some species such as the Gilded Flicker {Colaptes chrysoides [Malherbe 1852]); they excavate nesting sites that face into the prevailing wind direction as this aids in cooling (Zwartjes and Nordell 1998). Gila woodpeckers {Melanerpes uropygialis [Baird 1854]) show a preference for cavities that are higher above the ground which may be related to greater predation risk near the ground, difficulties associated with excavating smaller plants, or even greater thermal stress in nests in cavities that are closer to the ground (Korol and Hutto 1984). It is also possible, however, that there is greater opportu- nity for bird cavities to accumulate in taller plants (McAuliffe and Hendricks 1988). The relationship between bird cavities and nurse species, although marginally insignificant, is decidedly inconclusive as the vast majority of bird cavities are in taller plants, while nurse species were recorded primarily for shorter plants. In the 1.0-1. 9 meter height-controlled subset, only one bird cavity was recorded across the three nurse species analyzed in this study (of n — 236 Carnegiea associated with these nurse plants). Conclusions Our study sheds light on many previously unknown factors related to the Sonoran Desert keystone species, Carnegiea gigantea and its ecology, including: (1) Plants with large amounts of epidermal browning are evenly represented across height classes, fluctuating from about 5- 1 5% up to plants nine meters in height. However, plants with little to no sign of epidermal browning have very high representation in the smallest height class (0-0.9 m). (2) Epidermal browning is associated with the presence of bird cavities, girdling, natural and anthropogenic damage, and branching. Whether these individ- ually hasten browning or are the result of it (for example perhaps small mammals select plants with browning for girdling due to decreased presence of spines) is unclear. (3) Girdling is associated with greater natural and anthropogen- ic damage, including decapitation; plants that are damaged may be more attractive to rodents for girdling. Girdling is independent of presence of bird cavities and nurse architecture. (4) Lightning strike evidence was documented in far below 1% of the sampled plants, and fired-damaged plants also made-up less than 1% of the sampled plants. Decapitation made-up about 2% of the sample and most of those had little epidermal browning, but nearly 75% were girdled, compared to 45% girdling in the entire dataset. (5) Damage tends to increase with plant height. Even when height is controlled for, damage (including gunshot wounds) is associated with presence of bird cavities. Birds may select plants with entry wounds for cavity excavation, or perhaps some gunfire targets birds rather than plants. (6) Bird cavities are more strongly predicted by cactus height than number of branches, likely due to some disadvantages associated with branching (e.g., visibility, reduced exposure to winds that aid in cooling), as well as benefits of establishing higher on the plant (thus, necessarily, selecting taller plants for excavation sites) such as an elevated position above the ground and structural issues that may be associated with excavating smaller plants. Columnar cacti are often keystone species (Fleming 2002), yet columnar cacti face a variety of pressures on their long-term survival. Epider- mal browning is found across numerous taxa throughout the Western Hemisphere and is associated with early mortality (Evans and Fehling 1994; Evans et al. 1994b; Evans 2005). We show that it is also linked to presence of various types of damage in Carnegiea, including bird holes, girdling, and physical damage to the plant (e.g., gunshot wounds, proximal tree, fire). 124 MADRONO In addition to the widespread (geographically and taxonomically) occurrence of epidermal browning, some columnar cactus species suffer from flat top decay syndrome and fatal bleach- ing, as well as human incursion and destruction (Bashan et ah 1995), It is imperative that we develop a global-scale understanding of the dynamics of these species and the factors that impact their survival and mortality. Acknowledgments We thank the Arizona Army National Guard who own the data and granted us permission to analyze and publish our results from the dataset, and the Arizona Department of Game and Fish (AZGF) for providing useful background information and answering field- work and other related questions. Literature Cited Abou-Haidar, F. 1989. Influence of livestock grazing on Saguaro seedling establishment. M.S. Thesis, Arizona State University, Tempe, AZ. Bashan, Y., G. Toldeo, and G. Holguin. 1995. Flat top decay syndrome of the giant cardon cactus {Pachycereus pringiei): description and distribution in Baja California Sur, Mexico. Canadian Journal of Botany 73:683-692. Carleton, a. M. 1986. Synoptic-dynamic character of “bursts” and “breaks” in the south-west U.S. summer precipitation singularity. Journal of Cli- matology 6:605-623. . 1987. Summer circulation climate of the American Southwest, 1945-1984. Annals of the Association of American Geographers 77:619-634. Danzer, S. and T. D. Drezner. 2010. Demographics of more than 12,000 individuals of a keystone species in the northern Sonoran Desert since the mid- 1800’s. International Journal of Plant Sciences 171:538-546. Drezner, T. D. 2003a. Revisiting Bergmann’s Rule for saguaros (Carnegiea gigantea [Engelm.] Britt, and Rose): stem diameter patterns over space. Journal of Biogeography 30:353-359. . 2003b. A test of the relationship between seasonal rainfall and saguaro cacti branching patterns. Ecography 26:393^04. . 2004a. Saguaro recruitment over their Amer- ican range: a separation and comparison of summer temperature and rainfall. Journal of Arid Environments 56:509-524. . 2004b. Saguaros and their nurses in the Sonoran Desert: a review. Desert Plants 20:3-10. . 2011. Cactus surface temperatures are impact- ed by seasonality, spines, and height on plant. Environmental and Experimental Botany 74:17-21. and R. C. Balling Jr. 2008. Regeneration cycles of the keystone species Carnegiea gigantea are linked to worldwide volcanism. Journal of Vegetation Science 19:587-596. Duriscoe, D. M. and S. L. Graban. 1992. Epidermal browning and population dynamics of giant saguaros in long-term monitoring plots. Pp. 237- 258 in C. P. Stone and E. S. Bellantoni (eds.), Proceedings of the Symposium on Research in Saguaro National Monument, 23-24 January 1991. National Park Service, Rincon Institute, and [Vol. 61 Southwest Parks and Monuments Association, Tucson, AZ. Evans, L. S. 2005. Stem surface injuries of Neobux- haumia tetetzo and Neobuxbaumia mezcalaensis of the Tehuacan Valley of central Mexico. Journal of the Torrey Botanical Society 132:33-37. , K. A. Howard, and E. J. Stolze. 1992. Epidermal browning of saguaro cacti {Carnegiea gigantea): is it new or related to direction? Environmental and Experimental Botany 32:357- 363. AND B. J. Fehling. 1994. Surficial injuries of several long-lived columnar cacti of the Sonoran Desert, Mexico. Environmental and Experimental Botany 34:19-23. , V. A. Cantarella, K. W. Stolte, and K. H. Thompson. 1994a. Epidermal browning of sagua- ro cacti {Carnegiea gigantea): surface and internal characteristics associated with browning. Environ- mental and Experimental Botany 34:9-17. , C. McKenna, R. Ginocchio, G. Montene- gro, AND R. Kiesling. 1994b. Surficial injuries of several cacti of South America. Environmental and Experimental Botany 34:285-292. , V. Sahi, and S. Ghersini. 1995. Epidermal browning of saguaro cacti {Carnegiea gigantea): relative health and rates of surficial injuries of a population. Environmental and Experimental Bot- any 35:557-562. , J. H. Sullivan, and M. Lim. 2001. Initial effects of UV-B radiation on stem surfaces of Stenocereus thurberi (organ pipe cacti). Environ- mental and Experimental Botany 46:181-187. , M. ZUGERMAYR, AND A. J. YOUNG. 2003. Changes in surface injuries and mortality rates of Saguaro {Carnegiea gigantea) cacti over a twelve- year period. Journal of the Torrey Botanical Society 130:238-243. , A. J. B. Young, and S. J. Harnett. 2005. Changes in scale and bark stem surface injuries and mortality rates of a saguaro cacti {Carnegiea gigantea, Cactaceae) population in Tucson Moun- tain Park. Canadian Journal of Botany 83:31 1-319. and N. Abela. 2011. Stem surface injuries of 20 species of succulent Euphorbia (Euphorbiaceae) from South Africa. Environmental and Experimen- tal Botany 74:205-215. Fleming, T. H. 2002. Pollination biology of four species of Sonoran Desert columnar cacti. Pp. 207- 224 in T. H. Fleming and A. Valiente-Banuet (eds.). Columnar cacti and their mutualists: evolution, ecology, and conservation. University of Arizona Press, Tucson, AZ. Goad, M. S. and R. W. Mannan. 1987. Nest site selection by Elf Owls in Saguaro National Monu- ment, Arizona. Condor 89:659-662. Hutto, R. L., J. R. McAuliffe, and L. Hogan. 1986. Distributional associates of the saguaro {Carnegiea gigantea). The Southwestern Naturalist 31:469-476. INOUYE, R. S., N. J. Huntly, and D. W. Inouye. 1981. Non-random orientation of Gila Woodpeck- er nest entrances in saguaro cacti. Condor 83: 88-89. Kerpez, T. a. and N. S. Smith. 1990a. Competition between European Starlings and native woodpeck- ers for nest cavities in saguaros. Auk 107:367-375. 2014] DANZER AND DREZNER: CARNEGIEA DAMAGE 125 . 1990b. Nest-site selection and nest-cavity characteristics of Gila Woodpeckers and Northern Flickers. Condor 92:193-198. Kolberg, K. and K. Lajtha. 1997. Population dynamics and trace metal biogeochemistry of the saguaro cactus {Carnegiea gigantea) in the Sonoran Desert. Ecoscience 4:223-231. Korol, J. J. and R. L. Hutto. 1984. Factors affecting nest site location in Gila Woodpeckers. Condor 86:73-78. Lajtha, K., K. Kolberg, and J. Getz. 1997. Ecophysiology of the saguaro cactus {Carnegiea gigantea) in the Saguaro National Monument: relationship to symptoms of decline. Journal of Arid Environments 36:579-590. McAuliffe, j. R. 1996. Saguaro cactus dynamics. Pp. 96-131 in W. L. Halvorson and G. E. Davis (eds.), Science and Ecosystem Management in the National Parks. University of Arizona Press, Tucson, AZ. AND P. Hendricks. 1988. Determinants of the vertical distributions of woodpecker nest cavities in the sahuaro cactus. Condor 90:791-801. Niering, W. a. AND R. H. Whittaker. 1965. The saguaro problem and grazing in southwestern Na- tional Monuments. National Parks Magazine 39:4-9. , , AND C. H. Lowe. 1963. The saguaro: a population in relation to environment. Science 142:15-23. Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43:223-225. Rogers, G. F. 1985. Mortality of burned Cereus giganteus. Ecology 66:630-632. SCHIERMEIER, Q. 2005. Pall hangs over desert’s future as alien weeds fuel wildfires. Nature 435:724. Schmidt-Nielson, K. S. 1964. Desert animals: phys- iological problems of heat and water. Clarendon Press, Oxford. Sheppard, P. R., A. C. Comrie, G. D. Packin, K. Angersbach, AND M. K. Hughes. 2002. The climate of the US Southwest. Climate Research 21:219-238. Steenbergh, W. F. 1972. Lightning-caused destruc- tion in a desert plant community. Southwestern Naturalist 16:419^29. AND C. H. Lowe. 1977. Ecology of the Saguaro: 11. Reproduction, germination, establish- ment, growth, and survival of the young plant. Department of the Interior, National Park Service Scientific Monograph Series No. 8., U.S. Govern- ment Printing Office, Washington, D.C. . 1983. Ecology of the Saguaro: III. Growth and demography. Department of the Interior, National Park Service Scientific Monograph Series No. 17, U.S. Government Printing Office, Washington, D.C. Stevens, J. and D. A. Falk. 2009. Can buffelgrass invasions be controlled in the American Southwest? Using invasion ecology theory to understand buffelgrass success and develop comprehensive restoration and management. Ecological Restora- tion 27:417^27. Zwartjes, P. W. and S. E. Nordell. 1998. Patterns of cavity-entrance orientation by gilded flickers {Colaptes chrysoides) in cardon cactus. Auk 115: 119-126. Madrono, VoL 61, No. 1, pp. 126-136, 2014 CHOLLA MORTALITY AND EXTREME DROUGHT IN THE SONORAN DESERT Edward G. Bobich Biological Sciences Department, California State Polytechnic University, Pomona, CA 91768 egbobich@csupomona.edu Nick L. Wallace and Keely L. Sartori Department of Biology, Whittier College, Whittier, CA 90608 Abstract At the beginning of the 21st Century, deserts and woodlands of the southwestern United States were in the midst of a drought that contributed to massive die-offs for many plant species. In the northwestern Sonoran Desert, the drought contributed to extreme desiccation and mortality of cholla (Cylindropuntia) cacti. At low elevations (210-290 m), over half of the Cylindropuntia bigelovii, over 17% of the C. echinocarpa, 6% of the C ramosissima, and less than 2% of the C. ganderi were dead by winter 2004/2005. At a higher elevation (—820 m), 26% of C. bigelovii and 14% of C ganderi were dead by winter 2005/2006. The lower mortality for C. bigelovii at the higher elevation site was likely due to milder temperatures and greater precipitation; the greater mortality for C. ganderi at the same site may be the result of freezing or frost damage to tissues during the drought. Differences in mortality among species at both elevations were attributed to several factors, including: intra- and interspecific competition, reproductive strategy, and shoot architecture. Based on this study and previous research, it is believed that C. bigelovii experiences periodic extensive die-offs in response to extreme droughts, whereas populations of the other three species apparently change little during such droughts. Key Words: Chollas, Cylindropuntia, desert, drought, mortality. Populations of desert perennials are dynamic, with plant cover, mortality, and recruitment tied to several factors, especially total yearly or seasonal rainfall (Martin and Turner 1977; Goldberg and Turner 1986). Extreme climate events are believed to have a disproportionate effect on a variety of processes within organisms of a population (Gutschick and BassiriRad 2003). For populations of desert perennials to experi- ence measurable changes, events such as drought must be prolonged or especially severe compared to the conditions they usually experience (Gutschick and BassiriRad 2003). Such a drought occurred throughout much of the southwestern United States around the turn of the 21st Century, with 2002 being the driest year on record for much of the region (NOAA, NCDC website, Breshears et al. 2005). Massive die-offs of many desert perennials occurred as a result of this drought (Miriti et al. 2007; McAuliffe and Hamerlynck 2010), as well as Pinus edulis Engelm. (Breshears et al. 2005). There is evidence that, in addition to low precipitation, increasing ambient temperatures (Breshears et al. 2005), differences in soil composition (McAuliffe and Hamerlynck 2010), and plant age (Miriti et al. 2007) affected mortality during this drought. Measured increases in minimum winter temper- atures, as well as predicted increases in annual mean temperatures and a predicted drying of the entire region, indicate that droughts in the Desert Southwest will likely increase in frequency and be more extreme in the future (Weiss and Overpeck 2005; Seager et al. 2007). Desert cacti, especially barrel and columnar forms, are able to survive long-term drought because they have voluminous amounts of water- storage tissues (Gibson and Nobel 1986). They also tolerate some of the highest tissue temper- atures of any group of vascular plants and can acclimate to higher temperatures than they typically experience (Smith et al. 1984). For cacti, the relative tolerance to high temperatures and drought is linked with morphology. Chollas (genus Cylindropuntia [Engelm.] F.M. Knuth of North America) have stems with relatively small volumes of water-storage tissues and high surface area to volume ratios (Gibson and Nobel 1986), characteristics that can lead to lower relative tissue temperatures in the field (Nobel 2009), which may be one reason why this growth form is successful in warm deserts of North America. However, because chollas likely experience lower stem tissue temperatures in the field, they have lower maximum viable tissue temperatures than barrel and columnar cacti (Smith et al. 1984). Thus, during extreme long-term drought with elevated temperatures, such as that experienced in the drought at the turn of the century (Breshears et al. 2005) and those predicted for the future (IPCC 2007), chollas may experience greater mortality than barrel and columnar cacti not 2014] BOBICH ET AL.: CHOLLA MORTALITY AND EXTREME DROUGHT 127 only because they store less water than these growth forms, but because their tissues may not be able to tolerate increases in temperature. The severe drought at the end of the 20th and the beginning of the 21st centuries appeared to lead to extensive mortality for several cholla species in the northwestern Sonoran Desert. In early fall 2004, nearly every Cylindropuntia bigelovii (Engelm.) F. M. Keuth, Cylindropuntia echinocarpa (Engelm. & J. M. Bigelow) F. M. Knuth, Cylindropuntia ganderi (C. B, Wolf) Rebman & Pinkava, and Cylindropuntia ramo- sissima (Engelm,) F. M, Knuth at low elevations were extremely desiccated or dead and most living plants had not flowered or produced new stems for two years (M. Fisher personal com- munication), In the 1960’s, C bigelovii experi- enced a die-off at the same location (McDo- nough 1965). For the recent drought, C bigelovii clearly experienced extensive mortality. Mortality for the other species was difficult to assess initially because cacti can experience extreme desiccation and survive (Gibson and Nobel 1986), Thus, the purpose of the study was to determine how the morphology of each species and the apparent competition they experience were related to their relative mortality as a result of the drought. Because of the length and severity of the drought, it was hypothesized that species survival should depend on the ability to store water, which is related to stem volume and the amount of water storage tissue in the stems (Nobel 2009). Based on these criteria and previous research on the four species in Deep Canyon (Bobich and Nobel 2001), C. ramosissima should have experienced the greatest percent mortality, because its stems had the smallest volume and smallest amount of water storage tissue, follov/ed in order by C echinocarpa, C. bigelovii, and C ganderi. It was also hypothesized that shorter species with more basal branches should have experienced lower mortality because they would experience lower water stress due to gravity and be able to sacrifice some of their canopy and survive, compared with taller species with fewer basal branches. In addition, species that experience the greatest competition for resources should experience the highest mortality. The effects of elevation on plant mortality were also assessed by observing C bigelovii and C ganderi at a higher elevation where they were hypothesized to have experi- enced lower mortality, because temperature decreases and rainfall increases as elevation increases. To test these hypotheses, morpholog- ical measurements, distances to nearest neigh- bors, and descriptions of patterns of shoot death were recorded. In addition, climate data were analyzed to compare the conditions prior to this study and those leading up to the previously reported die-off at the same site (McDonough 1965), and to compare the climate at the two sites in this study. Methods Field Observations Research was performed in the flood plain (210-290 m; 33°39'N, 116°22'W) and at Agave Hill (820 m; 33°38'N, 116°24'W) within the University of California, Riverside Philip L. Boyd Deep Canyon Desert Research Center (Deep Canyon) in Palm Desert, California. The following cholla species were studied in the flood plain: Cylindropuntia bigelovii, C. echinocarpa, C. ganderi, and C ramosissima. Only C bigelovii and C ganderi occurred at Agave Hill. In the flood plain, 10 plants occurring approximately 238-250 m in elevation were not included in the study because they could not be positively identified to species. At both sites, chollas were sampled in 10 m X 10 m quadrats. In the flood plain, four quadrats were placed roughly west to east every 0.2 km along a total distance of 2.8 km south from the fence at the north boundary of the Research Center. At Agave Hill, three quadrats were positioned roughly west to east on two north- and two south-facing slopes, for a total of 12 quadrats. Only chollas whose bases were in the quadrat boundaries were sampled. Dead chollas were included in the study if they possessed an intact epidermis over approximately 75% of their shoots, which would indicate that they died during the drought. Chollas in the flood plain were determined to be living or dead during the winter and early spring of 2004/2005 because the site received 269 mm of rain from October 2004 through February 2005 (average precipitation from July 1-Jurie 30 for the site was 140 mm), allowing living plants to rehydrate and produce both new shoots and flowers. Chollas were studied at Agave Hill during winter 2005/2006 when all living plants possessed turgid shoots. The following morphological measurements were recorded for individual plants: height, lengths of the major and minor axes of the canopy, and number of basal branches, which were defined as major branches originating within 0.2 m of the plant base. Height measure- ments for some of the dead C bigelovii were approximated by measuring the height of the upright portion of the shoot and the length of the nodding or detached upper portion of the shoot lying near the base of the plant. Canopy spread was calculated by taking the square root of the product of the major and minor axes. Canopy axes were not measured for dead C. bigelovii because the majority of the lateral branches had detached; the dead individuals of the other three species appeared to have retained most of their 128 MADRONO [Vol„ 61 stems, allowing for canopy measurements. Ob- servations of shoot death, including detached dead shoots, retained dead shoots, and the apparent direction of the death of the shoot (acropetal or basipetal), were made in an effort to relate morphology to mortality. Distance to the nearest cholla was recorded because chollas demonstrate a high-level of intraspecific and interspecific competition (Cody 1986a, b). Dis- tance to the nearest shrub was also recorded because most of the shrubs at each site, represented primarily by Ambrosia dumosa (A. Gray) W.W. Payne, Enceiia farinosa A. Gray ex Torr., and Larrea tridentata (Sesse & Moc. ex DC.) Coville in the flood plain and Ambrosia dumosa, Bahiopsis parishii (Greene) E.E. Schill. & Panero, Bebbia juncea (Benth.) Greene, Enceiia farinosa, Eriogonum fascicuiatum Benth., and Larrea tridentata at Agave Hill, likely had a significant portion of their roots within the rooting zone of the chollas (Cody 1986b; Rundel and Nobel 1991; Nobel 1997; E. Bobich personal observations). Climate Data Precipitation and mean yearly temperature in Palm Springs were determined for the 10 seasons (July 1-June 30; 1954/1955-1963/1964) prior to the publication of the die-off of C. bigelovii in the early 1960’s (McDonough 1965) and for the 10 seasons (1994/1995-2003/2004) leading up to this study. Climate data for Palm Springs were downloaded from the Desert Research Institute’s Western Regional Climate Center website (wrcc.dri.edu). Because climate data for Agave Hill was not recorded until 1973, precipitation and mean yearly temperature for seasons 1973/ 1974-2003/2004 and the 10 seasons leading up to the drought (1994/1995-2003/2004) were com- pared between the flood plain and Agave Hill. In addition, the number of days with low temper- atures <0°C and days with low temperatures 0- 2°C, which can lead to radiation frost or frost pockets and freezing damage to cacti, were recorded for the 10 seasons leading up to the drought for both the flood plain and Agave Hill. Data Analysis There were typically few individuals of each species in each quadrat, so the data for each species was pooled for both the flood plain and Agave Hill. Percent mortality was calculated for each species at both sites. Morphological data and distances to the nearest cholla and nearest shrub were not distributed normally for any of the species and the variances among species were rarely equal. Because most of the data could not be transformed to allow for parametric compar- isons, comparisons of morphological data and distances to the nearest cholla and nearest shrub among the species in the flood plain were performed using Kruskal- Wallis one way AN- OVA. Comparisons between C bigelovii and C ganderi at Agave Hill were performed using Mann-Whitney U tests. Comparisons for C. bigelovii and C ganderi between the flood plain and Agave Hill and between living and dead plants for each species within each site were also performed using Mann-Whitney U tests. For C ganderi in the flood plain, living and dead individuals could not be compared because only one had died. Climate data for Palm Springs were compared between the 1954/1955-1963/1964 and 1994/1995-2003/2004 seasons using Student’s t- test. Precipitation and temperature were com- pared between the flood plain and Agave Hill using paired t-tests, with data grouped by year. Data are presented as means ±1SE. Results Distribution, Die-back, and Mortality of Chollas Each cholla species in the flood plain had a different elevational distribution. The only spe- cies that occurred throughout the entire flood plain was Cylindropuntia ramosissima. Cylindro- puntia echinocarpa had the lowest elevational range and was not sampled in quadrats above 244 m. The lowest elevation at which C bigelovii and C ganderi were sampled in the flood plain was 250 m, with C ganderi apparently replacing C. echinocarpa; C. bigelovii was especially com- mon on rockier soil. At Agave Hill (—820 m), C bigelovii was sampled only on south-facing slopes, which is where they primarily occurred, whereas C ganderi occurred in quadrats on both north- and south-facing slopes. In late fall 2004, chollas in the Deep Canyon flood plain were extremely dehydrated and most had experienced substantial die-back, although the patterns of shoot die-back differed among species (Fig. 1). Most of the dead C bigelovii had blackened shoots with black spines; living stems were usually bright green with yellow spines (Figs. lA and 2). The shoots of C. bigelovii appeared to have died acropetally, the evidence being that living plants, as well as some of those that had collapsed, retained living stems on approximately the top or distal third of their shoots (Fig. 2). The shoots of many dead or dying C bigelovii failed near the base and fell over or appeared to have collapsed (Fig. 2), but most plants that died had either a nodding shoot, with approximately the top third of the shoot bent downward, or the top third or quarter of the shoot had completely detached from the rest of the plant (Fig. lA), For C echinocarpa (Fig. IB), dead stems occurred throughout the shoot canopy; in some cases, entire major branches 2014] BOBICH ET AL.: CHOLLA MORTALITY AND EXTREME DROUGHT 129 Fig. 1. Representatives of A) Cylindropuntia bigelovii, B) C. echinocarpa, C) C. ganderi, and D) C. ramosissima during the fall of 2004 in the flood plain of the Philip L. Boyd Deep Canyon Desert Research Center, Palm Desert, CA. Arrows indicate nodding shoots in A) and dead branches in B), C), and D). had died. There were almost always dead small stems near the base of each C echinocarpa. Cylindropuntia ganderi and C. ramosissima clearly experienced basipetal dieback, which often re- sulted in the death of entire major branches (Fig. 1C, D); however, whereas the dead terminal stems of C. ganderi were almost always attached to the plant, copious numbers of dead stems occurred at the base of all but the smallest C. ramosissima. The shoots of many of the largest C ramosissima apparently split at the base, appear- ing initially as separate plants, especially those individuals whose bases were buried in mounds of sandy alluvium. After the rainfall in the winter of 2004/2005, many plants that initially appeared to have been dead, especially certain C. ramo- sissima, rehydrated and developed new stems. The cholla species that experienced the highest mortality at both sites in the Boyd Reserve was C. bigelovii (Table 1). Over half of the sampled C bigeiovii were dead in the floodplain, whereas just over one-fourth were dead at Agave Hill (Table 1). Of the other three species that occurred in the flood plain, only C. echinocarpa experienced over 10% mortality (Table 1). For C. ganderi in the flood plain, only one sampled individual had died, whereas 14% died at Agave Hill (Table 1). Almost 6% of the sampled C ramosissima were dead in the flood plain by the spring of 2005 (Table 1). Morphology and Proximity to Other Chollas and Shrubs In the flood plain, C ganderi was on average 75% as tall as the other three cholla species, but at Agave Hill it was statistically the same height as C bigelovii (Table 2). Cylindropuntia bigelovii had the smallest canopy spread of the four species, with a canopy 40% as broad as those of C echinocarpa and C ganderi, and 24% of that of C. ramosissima in the flood plain. At Agave Hill, the spread of C bigelovii was less than 50% of that of C. ganderi (Table 2). Cylindropuntia bigelovii also had the fewest basal branches, with almost all plants having a single trunk at both sites. In the flood plain, C ramosissima had the most basal branches, followed in order by C 130 MADRONO [Vol. 61 Fig. 2. A) Two healthy and two collapsed dying individuals of Cylindropuntia bigelovii and B) a view of the collapsed dying individuals showing dead bases, leading to the collapse, and living stems on the apices of the shoots. The plants were located in the flood plain of the Philip L. Boyd Deep Canyon Desert Research Center, Palm Desert, CA. gander i and C. echinocarpa (Table 2). At Agave Hill, C ganderi had three times as many basal branches as did C. bigelovii (Table 2). In general, chollas were smaller at Agave Hill than in the flood plain. The heights of C. bigelovii and C. ganderi at Agave Hill averaged 71% of those in the flood plain (Table 2). Canopy spread and the number of basal branches did not differ between C. bigelovii in the flood plain and those at Agave Hill. For C. ganderi at Agave Hill, canopy spread was 75% smaller and individuals averaged one less basal branch than did those in the flood plain (Table 2). In the flood plain, C bigelovii individuals were nearly 1.0 m closer to other chollas, usually other C bigelovii, than were plants of the other three species, all of which were similar distances from other chollas (Table 2). At Agave Hill, C bigelovii was 0.7 m closer to other chollas than was C ganderi (Table 2). Cylindropuntia bigelovii individuals were also 1.0 m closer to shrubs than were the other three species in the flood plain, whereas at Agave Hill C. bigelovii and C. ganderi were similar distances from the nearest shrubs (Table 2). Overall, chollas at Agave Hill were much closer to their nearest neighboring chollas and shrubs than were similar species in the flood plain (Table 2). Dead individuals did not differ significantly from living individuals for any of the species in the flood plain or at Agave Hill in terms of morphology (height, spread, and number of basal branches), or proximity to other chollas or shrubs (Table 2). At Agave Hill, dead C. ganderi nearly differed in height from living individuals; how- ever, no other comparisons yielded P < 0.12 (Table 2). Climate For Palm Springs, the average precipitation for the 10 seasons (July 1-June 30) prior to the publication of McDonough (1965) was statisti- cally the same as it was for the 10 seasons leading up to this study (Table 3), whereas the average yearly temperature for the 10 seasons leading up to this study was almost two degrees higher than it was from 1954/55-1963/64 (Table 3). From 1973/74 to 2003/04, the climate in the flood plain was warmer and drier than that of Agave Hill (Table 4). In fact, for every season, except 1986/ 87 and 1989/90, Agave Hill received more precipitation than did the flood plain, with the average seasonal precipitation in the flood plain 80% of that at Agave Hill. During that same period, the average daily temperature in the flood plain was over 3°C greater than that at Agave Hill (Table 4). Over the ten seasons leading up to this study, the flood plain again received approx- imately 80% of the precipitation at Agave Hill. During the same period, both the flood plain and Agave Hill received an average of 69% of the Table 1. Percent Mortality for Cylindropuntia Bigelovii, C. echinocarpa, C. ganderi, and C. RAMOSISSIMA IN THE FLOOD PLAIN AND AT AGAVE HILL IN THE PHILIP L. BOYD DEEP CANYON DESERT Research Center, Palm Desert, CA. Values in parentheses indicate the total number of individuals sampled at each location. C. bigelovii C. echinocarpa C. ganderi C. ramosissima Flood plain 55.6 (275) 17.5 (63) 1.6 (61) 5.9 (85) Agave Hill 25.8 (128) — 14.0 (57) — Table 2. Morphological Characteristics, Distance to the Nearest Cholla, and Distance to the Nearest Shrub for all of the Sampled Individuals, and the Living and the Dead Individuals of Cylindropuntia Bigelovii, C. echinocarpa, C. ganderi, and C ramosissima in the Flood Plain and of C. bigelovii and C. ganderi at Agave Hill in the Philip L. Boyd Deep Canyon Desert Research Center, Palm Desert, CA. For ail of the 2014] BOBICH ET AL.: CHOLLA MORTALITY AND EXTREME DROUGHT 131 "TS O TS TS C cd £d w w w ^ a.S C § “ -O I ,, ■S -o S s s M ^ W czi ' ‘ 'o :s .. ■' " 6 Z 2 I S is ^ O S) ^ gU C« J-l ^ cd a o ^ ^ ^ cd 'g ^ d S ™ g -S Q »■< c „ . cd « d .2 M c« o 2 ^ .2 g « a ‘S g-ScQ - ^ ^ 'Id M c« i2 2 3 o 3 8^ g “.>D ° 2-sl &| X d G S X g ^ g:d > - 2 « i-S ^ d X 5 d 0 o o 2 X S w cd cn b •? -2 & ^ S S d Cd C« bp .O .O w ^ “ d ■5 ^ ^ ^ c/D y > (U d) cd ^ ^ « -s § E D .S X >. >>6 '2 |S3S| d IS S ^ o '■5 ^ ^ 3 S ■K.e iS a & d d .> o I ^ ^ X ^ ^ X) I - 2 cd 8 O d ^ ^ O cd cn cd X) oo , cDNX O— i^rimcNirsirNi fNfNroooo*^^’-^ odddddddoNddEdddddd +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 ooi>-0'd‘'^cnmfnx ininr^ininin(Nri.-H cN (N fN d d d -=< L ^ < X < cd bi} < 132 MADRONO [VoL 61 Table 3. Mean Precipitation and Temperature for Palm Springs for the 1954/1955-1963/1964 Seasons AND THE 1994/1995-2003/2004 Rainfall Seasons. Temperatures were not included for 1958/59 and 1963/64 due to insufficient data. Values with different letters within a row indicate a significant difference (P < 0.05) for Student's t-test. Data are means ± ISE. 1954/1955-1963/1964 1994/1995-2003/2004 95.0 ± 14.6a 85.0 ± 22.4a 22.1 ± 0.2a 24.0 ± 0.3b Precipitation (mm) Temperature (°C) average seasonal rainfall they received from 1973/ 1974 to 2003/2004 (Table 4). The two seasons with the lowest precipitation at both sites were 1998/1999, with 49.3 rnm received in the flood plain and 68.8 mm at Agave Hill, and 2001/2002, with 57.4 mm received in the flood plain and 67.4 mm at Agave Hill; the calendar year with the lowest precipitation was 2002, with the flood plain receiving 26.9 mm and Agave Hill 37.8 mm. The average temperature for the 10 seasons leading up to this study was over 4°C greater in the flood plain than at Agave Hill (Table 4). The low temperatures in the flood plain were never <2°C during the 10 seasons prior to the study. At Agave Hill, there were three freezing days, occurring in December 1998 (0.0°C), April 1999 (0.0°C), and January 2002 (-0.5°C). Each freezing day was preceded by and/or followed by days with low temperatures 0-2°C. Overall, there were 23 d with low temperatures 0-2°C at Agave Hill, with all of them occurring during the 1997/ 98-2001/02 and 2003-2004 seasons. Discussion The cholla species that experienced the greatest mortality in Deep Canyon during the most recent drought was C. bigelovii, with populations decreasing over 50% in the flood plain and over 25% at Agave Hill, McDonough (1965) reported that C bigelovii experienced greater than 50% mortality in Deep Canyon approximately 40 yr earlier, but did not attribute the mortality to any one factor. Populations of C. bigelovii may be stable for several decades and individual plants can live over 60 yr (Hastings and Turner 1965). Based on the results of this study and McDonough (1965), and the climate leading up to both studies, it can be assumed that, at least in the northwestern Sonoran Desert, the popula- tions of C bigelovii are dynamic and experience repeated die-offs that likely result from extreme drought. The greater susceptibility of C. bigelovii to drought compared with the other cholla species is related to its morphology, reproductive mode, and low genetic diversity. The shoots of C bigelovii almost always had one basal branch (trunk), which is typical for the species (McDo- nough 1965; Turner et ah 1995). Thus, if the trunk fails, that individual essentially dies. Unlike the other three species, all of which depend on sexual reproduction to regenerate populations, C bigelovii rarely produces viable seeds (Pinkava et al. 1985) and almost solely propagates vegetatively via the successful rooting of detached stems (Benson 1982; Nobel et al. 1986; Bobich and Nobel 2001). Thus, C. bigelovii likely experiences acropetal shoot death to allow individuals to keep their young stems, which are their propagules, alive to regenerate their popu- lations. For the other three species that depend on sexual reproduction, survival of adults is important because successful seedling recruitment of cacti in deserts is rare (Gibson and Nobel 1986). Finally, C bigelovii appears to be sterile and completely dependent on vegetative repro- duction because it is triploid (Pinkava and Parfltt 1982; Baldwin et al, 2012) throughout much of its range. As a result, the genetic diversity of C. bigelovii is likely very low, which could make certain clones especially susceptible to adverse conditions like extreme drought, resulting in relatively high mortality for this species. Because C. bigelovii reproduces vegetatively and its propagules are large (Bobich and Nobel 2001), its populations tend to be aggregated (McDonough 1965), which should lead to in- creased intraspecific competition. Cylindropuntia acanthocarpa, C. echinocarpa and C. ramosissima, tend to be randomly or regularly spaced and avoid neighboring conspecifics and/or congeners Table 4. Mean Precipitation and Temperature for the 1973/1974-2003/2004 and 1994/1995-2003/2004 Seasons in the Flood Plain and Agave Hill in the Philip L. Boyd Deep Canyon Research Center. Yearly rainfall and mean temperature were compared between sites using paired t-tests; values with different letters within a row indicate a significant difference (P < 0.05) between the sites. Data are means ± ISE. 1 973/1974-2003/2004 1 994/ 1 995-2003/2004 Flood plain Agave hill Flood plain Agave hill Precipitation (mm) Temperature (°C) 156 ± 17a 23.8 ± 0.2a 193 ± 20b 20.6 ± 0.2b 108 ± 24a 24.5 ± 0.3a 134 ± 26b 20.0 ± 0.5b 2014] BOBICH ET AL.: CHOLLA MORTALITY AND EXTREME DROUGHT 133 (Cody 1986a, b). The aggregated distributions of C. bigelovii are likely to exist for long periods between extreme droughts, after which extensive mortality drives populations towards a more random distribution, as hypothesized by McDon- ough (1965). Cylindropuntia bigelovii were also closer to shrubs than the other cholla species in the flood plain, reflecting the fact that the species occurs in areas with relatively high plant ground cover (E. Bobich, unpublished results). Thus, interspecific competition may have also contrib- uted to the relatively high mortality of this species in the flood plain. The species with the lowest percent mortality in the flood plain was C gander i, which was also the shortest of the cholla species; C. ganderi also had a similar number of basal branches to those of the most prolifically branched species, C ramo- sissima. Like C. ramosissima, C. ganderi experi- enced basipetal shoot death, reflecting the ten- dency for adults to persist because they likely reproduce solely from seed (Bobich and Nobel 2001). Cylindropuntia ganderi did appear to lose fewer stems during the drought than the other species in the study, which is likely because it has extremely fibrous wood in its terminal stem junctions, whereas the other three species com- pletely lacked fibers in those regions (Bobich and Nobel 2001). The resistance of C ganderi to mortality in the flood plain may be because it has the greatest stem capacitance based on stem volume of the four species studied (Bobich and Nobel 2001), and its stems are very tolerant of high temperatures (Nobel and Bobich 2002). The effect of how its relatively short height might have contributed to its high survivorship is unclear, because the difference in height between C. ganderi and the other cholla species would have a small effect on water potential, yet the relatively low stature and surface area/volume of its shoots could have resulted in it depleting less soil water than the other species in the flood plain. Elevation affected mortality of C. bigelovii and C ganderi differently. Typically, C ganderi occurs at higher elevations than does C. bigelovii (Baldwin et al. 2012), which may be related to their relative abilities to establish in colder microclimates. In fact, at Agave Hill C bigelovii occurs almost exclusively on south-facing slopes, which experience the highest local temperatures, whereas C. ganderi occurs on slopes of all aspects. At Agave Hill, mortality for C bigelovii was half of what it was in the flood plain, supporting the hypothesis that the lower daily high temperatures and greater rainfall at Agave Hill should lead to reduced mortality; however, mortality of C ganderi was over seven times greater at Agave Hill than in the flood plain. Although produc- tivity appears to increase with increases in elevation for C. ganderi, the amount of tissue damage for this species also increases with increasing elevation (Nobel and Bobich 2002). One reason for the increased tissue damage at higher elevations compared with lower elevations may be because C. ganderi stems are less tolerant of low than high temperatures, suffering 50% tissue death at — 4.9°C and 69.5°C after temper- ature acclimation, respectively (Nobel and Bob- ich 2002); these values are — 7.3°C (Nobel 1982) and 59.0°C (Didden-Zopfy and Nobel 1982) for C. bigelovii, respectively (Nobel 1988). Further- more, C ganderi has a relatively open canopy that results in lower shoot temperatures than other cacti (Nobel et al. 1991) and has an apparent increased need for nurse plants at high elevations (Nobel and Bobich 2002), likely for protection from freezing temperatures; nurse plants would not have supplied much cover during the drought because of dieback and relative leaflessness. Considering that there were three freezing episodes at Agave Hill in the seven seasons prior to the study, with one being in January 2002, the driest year on record, it is possible that drought combined with susceptibil- ity to freezing and frost damage led to greater mortality of C. ganderi at Agave Hill than in the flood plain. In this study, C ramosissima had the smallest stems in terms of mass and diameter (Bobich and Nobel 2001), making individual stems more susceptible to dehydration (Nobel 1988) than those of the other three species due to the lack of stem capacitance. Although there was a consider- able amount of standing dead and detached dead stems near individuals, mortality was relatively low (6%) for C ramosissima, which may be due to several factors, including the relatively high number of basal branches and the large canopy spread of the individuals. Different branches emanating from the bases of C ramosissima appeared to respond to the drought differently, with some branches dying completely, while others survived and produced new stems and flowers after the winter of 2004/2005, indicating that this species experiences branch sacrifice during drought (Rood et al. 2000). The shoots of larger C ramosissima appeared to split at the base, possibly allowing each resulting section to act independently of each other, as occurs for desert shrubs (Schenk 1999). Thus, individual C. ramosissima may increase their chances of long- term survival by compartmentalizing their shoots and stem redundancy. Of the four cholla species in this study, C ramosissima was the only one that apparently experienced shoot splitting. The low mortality of C ramosissima in the flood plain of Deep Canyon is in contrast with the response of the species to the drought in Joshua Tree National Park, where over 50% of the adults died between 1999 and 2004 (Miriti et al. 2007). The difference in mortality of this 134 MADRONO [Vol. 61 species between the sites may be related to the fact that the drought was especially severe in Joshua Tree National Park, resulting in massive die-offs of A. dumosa and other subshrubs as well as the highest mortality observed for L. tridentata in the Mojave or Sonoran Deserts after the drought (McAuliffe and Hamerlynck 2010). Furthermore, the site where C ramosissima was studied (1006 m) was near the elevational limit of the species (1100 m; Baldwin et al. 2012) and, as such, the individuals at that site could have been more susceptible to extreme changes in climate. Cylindropuntia echinocarpa was the only spe- cies other than C bigelovii to experience over 10% mortality in the flood plain. The stems of C echinocarpa have a small capacitance compared to those of C. bigelovii and C. ganderi (Bobich and Nobel 2001), explaining why there were dead stems at the bases of individuals. Cylindropuntia echinocarpa had an intermediate canopy spread and number of basal branches compared to the other three species, meaning that although they might be able to compartmentalize their shoot mortality, the loss of one basal branch would result in a loss of 40% of their canopy compared to less than 25% for C. ganderi and C. ramosissima. Thus, it is likely that the intermedi- ate mortality of C. echinocarpa during the drought is related to its morphology. Dead individuals did not differ from living individuals in morphology or their proximity to other chollas or shrubs for any of the cholla species. Juvenile desert shrubs tend to experience greater mortality than adults and their mortality is not related to their relative distribution within a population (Wright and Howe 1987). Further, seedlings and young juvenile cacti are extremely susceptible to desiccation (Gibson and Nobel 1986). Miriti et al. (2007), who separated plants into juveniles and adults for perennials in Joshua Tree National Park, found that juvenile C. ramosissima experienced greater percent mortal- ity than did adults between 1999 and 2004. Because the drought was prolonged and there was likely low recruitment during that period, it is probable that the cholla populations in Deep Canyon had few individuals younger than five- years-old at the time of this study. Thus, even if there was selection against younger individuals in Deep Canyon during the drought, it might not have been detectable in this study. Most climate models for the Sonoran Desert indicate the region will be warmer and drier, with droughts being longer and more severe (Weiss and Overpeck 2005; Seager et al. 2007), which is somewhat supported by the 2°C increase in mean annual temperature in Palm Springs from the 1950’s and 1960’s to the most recent drought. Climate has already been cited as the main factor in the upward change in elevational ranges of some of the most prominent perennial plants in the Deep Canyon drainage (Kelly and Goulden 2008). How these predicted changes in climate will affect cholla populations is unclear because mortality was low for the sexually reproducing species during this drought, which included the driest year on record, and because each species appears to have different adaptations to deal with prolonged drought. The sexually reproducing species, C echinocarpa, C. ganderi, and C ramosissima, may experience greater changes than C bigelovii in elevational range over time because seedling recruitment should be lower in the predicted future climates, whereas C. bigelovii stems have a large capacitance and can produce roots readily when exposed to moist soil (Bobich and Nobel 2001). Thus, although C bigelovii individuals clearly experienced the greatest mor- tality due to the most recent major drought compared with sexually reproducing species, it is not indicative of their relative future success in the northwestern Sonoran Desert. Conclusions Cholla species responses to the recent severe drought in the northwestern Sonoran Desert were related the primary mode of reproduction, which was related, in part, to their morphological characteristics, and proximity to other competing plants. Mortality was highest in the flood plain at Agave Hill for C. bigelovii, which is the only cholla species that relied completely on vegetative reproduction. Because their youngest stems are their propagules, C bigelovii plants sacrificed their single trunk, essentially killing the individ- uals, to keep their youngest stems alive. Cylin- dropuntia bigelovii also experiences the greatest intraspecific competition of the four species because its reproductive mode leads to closer groupings of individuals. Individuals of the other three species appeared to sacrifice different parts of their shoots for plant survival. Young stems died throughout the canopy for C. echinocarpa, with plants rarely losing entire basal branches, whereas terminal stems and entire major branch- es died on C ganderi and C ramosissima, both of which had several major branches; C. ganderi retained its stems and responded as a single plant, whereas all C. ramosissima shed their youngest stems and the largest individuals responded to the drought as several independent units. Mortality for C. bigelovii was lower at Agave Hill, which is cooler and receives more rain than does the flood plain, whereas mortality for C. ganderi increased with increases in elevation, possibly reflecting a susceptibility to low temperature tissue damage during the extreme drought. Acknowledgments We thank Allan Muth and Mark Fisher for allowing us to perform the research at the Philip L. Boyd Deep 2014] BOBICH ET AL.: CHOLLA MORTALITY AND EXTREME DROUGHT 135 Canyon Desert Research Center, for sharing their observations of the drought, and for providing the climate data for the flood plain and Agave Hill. We also thank Gretchen B. North for her insightful comments in the preparation of the manuscript. Literature Cited Baldwin, B. G.D. H. Goldman, D. J. Keil, R. Patterson, and T. J. Rosatti, (eds.) 2012. The Jepson Manual: vascular plants of California, 2nd ed. University of California Press, Berkeley, CA. Benson, L. 1982. The cacti of the United States and Canada. Stanford University Press, Stanford, CA. Bobich, E. G. and P. S. Nobel. 2001. Vegetative reproduction as related to biomechanics, morphol- ogy, and anatomy of four cholla cactus species in the Sonoran Desert. Annals of Botany 87:485^93. Breshears, D. D., N. S. Cobb, P. M. Rich, K. P. Price, C. D. Allen, R. G. Balice, W. H. Romme, j. H. Kastens, M. L. Floyd, J. Belnap, J. J. Anderson, O. B. Myers, and C. W. Meyer. 2005. Regional vegetation die-off in response to global-change-type drought. Proceedings of the National Academy of Sciences of the United States of America 102:15144-15148. Cody, M. L. 1986a. Structural niches in plant communities. Pp. 381^05 in J. Diamond and T. J. Case (eds.) Community ecology. Harper & Row, New York, NY. . 1986b. Spacing patterns in Mojave Desert plant communities: near-neighbor analyses. Journal of Arid Environments 11:1 99-2 1 7 . Didden-Zopfy, B. and P. S. Nobel. 1982. High- temperature tolerance and heat acclimation of Opuntia bigelovii. Oecologia 52:176-180. DRI, WRCC. Western US Historical Summaries. Website http.7/www. wrcc.dri.edu/Climsum.html (accessed 07 September 2011). Gibson, A. C. and P. S. Nobel. 1986. The cactus primer. Harvard University Press, Cambridge, MA. Goldberg, D. E. and R. M. Turner. 1986. Vegetation change and plant demography in permanent plots in the Sonoran Desert. Ecology 67:695-712. Gutschick, V. P. and H. BassiriRad. 2003. Extreme events as shaping physiology, ecology, and evolu- tion of plants; toward a unified definition and evaluation of their consequences. New Phytologist 160:21^2. Hastings, J. R. and R. M. Turner. 1965. The changing mile: an ecological study of vegetation change with time in the lower mile of an arid and semiarid region. University of Arizona Press, Tucson, AZ. Intergovernmental Panel on Climate Change (IPCC). 2007. Climate change 2007: The physical science basis, summary for policy makers. Contri- bution of Working Group I to the Fourth Assess- ment Report of the Intergovernmental Panel on Climate Change. IPCC Secretariat, Geneva, Swit- zerland. Kelly, A. E. and M. L. Goulden. 2008. Rapid shifts in plant distribution with recent climate change. Proceedings of the National Academy of Sciences of the United States of America 105:11823-11826. Martin, S. C. and R. M. Turner. 1977. Vegetation change in the Sonoran Desert region, Arizona and Sonora. Journal of the Arizona Academy of Science 12:59-69. McAuliffe, j. R. and E. P. Hamerlynck. 2010. Perennial plant mortality in the Sonoran and Mojave deserts in response to severe, multi-year drought. Journal of Arid Environments 74:885-896. McDonough, W. T. 1965. Pattern changes associated with the decline of a species in a desert habitat. Vegetatio 13:97-101. Miriti, M. N., S. Rodriguez-BuriticA, S. J. Wright, and H. F. Howe. 2007. Episodic death across species of desert shrubs. Ecology 88:32-36. NOAA, NCDC. 2003. Climate of 2002 Annual Review; Drought: National Paleoclimatic (Pre-Instrumen- tal) Perspective. Website http://www.ncdc.noaa. gov/oa/climate/research/2002/ann/paleo-drought. html (accessed 03 December 2011). Nobel, P. S. 1982. Low-temperature tolerance and cold hardening of cacti. Ecology 63:1650-1656. . 1988. Environmental biology of agaves and cacti. Cambridge University Press, New York, NY. . 1997. Root distribution and seasonal produc- tion in the northwestern Sonoran Desert for a C3 subshrub, a C4 bunchgrass, and a CAM leaf succulent. American Journal of Botany 84:949-955. — . 2009. Physicochemical and environmental plant physiology. Fourth Edition, Elsevier/ Aca- demic Press, Burlington, MA. and E. G. Bobich. 2002. Plant frequency, stem and root characteristics, and CO2 uptake for Opuntia acanthocarpa: elevational correlates in the north- western Sonoran Desert. Oecologia 130:165-172. , M. E. Loik, and R. W. Meyer. 1991. Microhabitat and diel tissue acidity changes for two sympatric cactus species differing in growth habit. Journal of Ecology 79:167-182. , G. N. Geller, S. C. Kee, and A. D. Zimmerman. 1986. Temperatures and thermal tolerances for cacti exposed to high temperatures near the soil surface. Plant, Cell & Environment 9:279-287. PiNKAVA, D. J. AND B. D. Parfitt. 1982. Chromo- some numbers in some cacti of western North America. IV. Bulletin of the Torrey Botanical 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. System- atic Botany 10:471-483. Rood, S. B., S. Patino, K. Coombs, and M. T. Tyree. 2000. Branch sacrifice: cavitation-associat- ed drought adaptation of riparian cottonwoods. Trees 14:248-257. Rundel, P. W. and P. S. Nobel. 1991. Structure and function in desert root systems. Pp. 349-378 in D. Atkinson (ed.). Plant root growth: an ecological perspective. Special publication of the British Ecological Society, no. 10. Blackwell Scientific Publications, Oxford, United Kingdom. Schenk, H. J. 1999. Clonal splitting in desert shrubs. Plant Ecology 141:41-52. Seager, R., M. Ting, I. Held, Y. Kushnir, J. Lu, G. Vecchi, H. Huang, N. Harnik, A. Leetmaa, N. Lau, C. Li, j. Velez, and N. Naik. 2007. Model projections of an imminent transition to a more arid climate in southwestern North America. Science 316:1181-1184. 136 MADRONO [Vol. 61 Smith, S. D., B. Didden-Zopfy, and P. S. Nobel. 1984. High-temperature responses of North Amer- ican cacti. Ecology 65:643-651. Turner, R. M., J. E. Bowers, and T. L. Burgess. 1995. Sonoran Desert plants: an ecological atlas. The University of Arizona Press, Tucson, AZ. Weiss, J. L. and J. T. Overpeck. 2005. Is the Sonoran Desert losing its cool? Global Change Biology ! 11:2065-2077. Wright, S. J. and H. F. Howe. 1987. Pattern and mortality in Colorado Desert Plants. Oecologia 73:543-552. Madrono, VoL 61, No. 1, pp. 137-143, 2014 SCOULERIA SISKIYOUENSIS (SCOULERIACEAE), A NEW RHEOPHYTIC MOSS ENDEMIC TO SOUTHERN OREGON, USA James R. Shevock California Academy of Sciences, Department of Botany, 55 Music Concourse Dr., Golden Gate Park, San Francisco, CA 94118-4503 j shevock@calacademy . org Daniel H. Norris University Herbarium, 1001 Valley Life Sciences Building, University of California, Berkeley, CA 94720-2465 Abstract Scouleria siskiyouensis Shevock & D. H. Norris, a new species restricted to rivers and streams in southwestern Oregon is described and illustrated. This species appears related to the widespread S'. aquatica Hook, in Drummond of western North America but is distinguished by a combination of features including lamina bistratose except for a few cells at immediate leaf margin unistratose, leaves lanceolate to ovate-lanceolate and with leaf apex acute to acuminate. Key Words: Aquatic mosses, rheophytes, Scouleria aquatica, Scouleria marginata. Scouleria Hook, is a very distinctive genus of rheophytic mosses. Described by Hooker in 1829, the dark coloration and growth habit resem- blance to certain rheophytic Racomitrium Brid. s.l. {Codriophorus P. Beauv. and Bucklandiella Roiv.) and Schistidium Bruch & Schimp. species probably contributed to its initial placement by early bryologists in the Grimmiaceae. The presence of systylious capsules is also similar to species of Schistidium. Scouleria is immediately recognized in the field by its habitat (aquatic in rapidly flowing streams and rivers); by its size and coloration (drying nearly black with leaves mostly about five mm or more broad); and by its capsule morphology (nearly sessile, large and globose with very thick capsule walls and presence of a columella). As a rheophytic genus, Scouleria has one of the most unusual distributions of bryophytes: west- ern North America, North Asia (Siberia and the Russian Far East), and southern South America (Churchill 2007). In North America, Scouleria primarily occurs along the Pacific Slope from Alaska to California with scattered disjunct populations inward to Montana but is not found in the eastern half of the continent. Only two species of Scouleria have been recognized for North America. These two species are readily recognized in the field by a hand-lens examina- tion of the leaf border. In Scouleria aquatica the border is merely thickened but flat caused by either enlarged cells or by the development of a bistratose layer at the margin. However, in Scouleria marginata E. Britton, the margin appears inflated and rounded due to thickened multistratose cells along the margin forming a more prominent border, especially seen on older leaves. These two species rarely grow sympatri- cally. Scouleria marginata prefers larger river systems and generally lower elevations whereas S. aquatica has a much broader range of ecological attributes occurring from near sea level to the headwaters of tributary streams. For over 150 yr Scouleria remained in the Grimmiaceae but the largely morphological work conducted by Churchill (1985) convincingly elevated Scouleria to family status. Recent molecular evidence also supports the recognition of this family of rheophytic mosses (Tsubota et al. 2003; Carter et al. 20 14). The other genus attributed to this family, Tridontium Hook, f., is monospecific, and is endemic to Australia and New Zealand. Scouleria is a genus that is easily recognized in the field. It is restricted to rapidly flowing unpolluted rivers and streams where plants are seasonally submerged then exposed on boulders and rock walls as rivers decrease in flow. When hydrated, Scouleria is dark green in color but upon drying becomes nearly black. Scouleria generally occurs in large patches and when in a dry state these blackened Scouleria populations can be recognized on large boulders and rock walls from a distance of several meters. On a worldwide basis all members of Scouleria seem to be the ultimate rheophyte, annually receiving severe scouring during peak flows when sediment yields are increased, deep submersion, an extended period in white water rapids, followed by a period of desiccation during which they are generally exposed in full sun during the hottest months of the year. All species of Scouleria have a prominent morphologic feature that exhibits an unusual structure for a moss leaf. In addition to the usual 138 MADRONO [VoL 61 feature of alar cells and median laminal cells (isodiametric and thick-walled in Scouieria) there is an area of several rows of elongate and glistening white-walled cells which is seen as an intramarginal limbidium. This intramarginal lim- bidium seems to be unique to the genus but its prominence varies greatly from leaf to leaf even on a single stem. Some leaves may have only a few cells between the alar and median cells while other leaves may have an intramarginal limbi- dium that comprises most of the area of the leaf. The variability in the prominence of the intra- marginal limbidium does not seem to be related to species differences within the genus, and it seems also unrelated to apparent habitat differ- ences. This limbidium although often is the dominant visual feature of some leaves apparent- ly has no taxonomic importance beyond merely signaling the generic identification. An additional generic character for Scouieria is the frequency of rhizoids on the abaxiai base of the costa. While the genus is easy to determine, the recognition of species has been more difficult. The actual number of species based solely on morphological features has ranged from three to five species, and part of this recognition was based on geography rather than a good set of morphological characters. Like many rheophytes, the aquatic growth form creates a large array of morphological character variations generally useful for identification purposes such as leaf shape, leaf size and leaf border. Most keys developed to separate Scouieria species have relied on cross sections of leaves to describe the patterns and locations of various levels of bistratosity. Crosby et ai. (2000) treat the genus as comprised of three species; two species in the Northern Hemisphere and one in the Southern Hemisphere. However, two additional species, S. rschewinii Lindb. & Arnell described in 1890 and S. puicherrima Broth, described in 1916 from Siberia and the Russian Far East were subse- quently reduced to synonymy within the geo- graphically widespread S. aquatica (Churchill 1985), As late as 1970, sporophytes were un- known for these two Russian taxa (Savic- Ljubickaja and Smirnova 1970). However, Churchill (2007) acknowledges that much varia- tion exists within the S', aquatica complex and other taxonomic arrangements at variety or species level may be warranted through future molecular study. While sporophytic characters can be highly informative between members within a genus, in Scouieria capsules are generally infrequently produced, are of short duration and look basically the same in shape and color. Even when capsules are present on herbarium speci- mens they are rather fragile and peristome teeth or their remnants are easily lost or detached. The only distinction in capsule morphology described in the genus to date is the lack of peristome teeth in S. marginata, a species endemic to the Pacific Slope of North America from British Columbia to California with isolated populations in Idaho and Montana (Lawton 1971; Christy et al 1982; Churchill 1985, 2007; Norris and Shevock 2004a). Kurbatova (1998) states that the seta is longer in S. aquatica var. puicherrima, but we have not seen Russian Scouieria with capsules for comparison to North American taxa. Fieldwork contributing toward a bryoflora of California (Norris and Shevock 2004a) led to the collection of many bryophytes, especially from the northern extension of the California Floristic Province within southern Oregon. This area is well-known as one of the biodiversity hot spots for vascular plants and contains a high level of endemism (Norris 1997). The area of the Kaimiopsis Wilderness on the Siskiyou portion of the Rogue River-Siskiyou National Forest is exceptionally important as an evolutionary refu~ gium. The same area is also rich in bryophyte diversity. Following examination of Scouieria collections from that general region it was determined that an undescribed species was likely at hand based on vegetative leaves that are strikingly more lanceolate and acute at apex. Under a compound microscope, a cross section displayed far greater development of bistratosity across the lamina than seen previously in the genus. However, describing a new Scouieria from North America was deemed to require a critical examination of S', aquatica throughout its range including plants from the Russian Federation. Due to the view of extreme plasticity displayed among morphological features of rheophytes in general and wide variation reported within the S. aquatica complex in particular, this issue was not pursed further at that time although this entity was referenced in the Scouieria key as "species A’ (Norris and Shevock 2004b). Eventually a molec- ular study was initiated to determine the affinities of Scouieria within its geographical range. That study (Carter et al. 2014) confirmed that the Russian species are worthy of species rank, S. aquatica is restricted to North America, and that the southwestern Oregon plants represent a species new to science and is described herein. Taxonomic Treatment Scouieria sisMyouensis Shevock & D, H. Norris, sp. nov. (Figs. 1-2). — TYPE: USA, Oregon, Douglas Co., Coos Bay District, Bureau of Land Management, Middle Fork Coquille River along highway 42 west of Roseberg at milepost 26, former site of Bear Creek Recre- ation Area, T30S, R9W, section 9, 42°58'9.2"N, 123°45'57.8''W, 850 ft, 19 Mar 2005, Shevock and Keiiman 26365 (holotype: CAS; isotypes: H, HYO, KRAM, MHA, MO, NY, OSC, UBC, VBGI, UC). 2014] SHEVOCK AND NORRIS: NEW RHEOPHYTIC MOSS 139 Fig, 1. Scouieria siskiyouensis Shevock & D. H. Norris. 1„ Fertile plant, wet. 2. Portion of stem transverse section. 3-6. Leaves. 7-14. Leaf transverse sections. (All from Shevock 26365, isotype, KRAM). Scale bars: a - 100 jam (7- 14); b - 1 mm (3-6); c - 100 pm (2); d - 0.5 cm (1). 140 MADRONO [VoL 61 Fig 2. Scouleria siskiyouensis Shevock & D. H. Norris. L Fertile plant, dry. 2-3. Leaf apices. 4. Middeaf cells. 5. Middeaf cells at margin. 6. Basal angular cells. 7. Basal juxtacostal cells. 8-10. Perichaetial leaves. 11. Calyptra. 12. Operculate capsule with calyptra, wet. 13. Operculate capsule, wet. 14-15. Young systylious capsuks, dry. 16. Old capsules with destroyed peristome, dry. 17. Exothecial cells at base of urn. 18. Mid-urn exothecial cells. 19. Exothecial cells at orifice and portion of peristome. 20-24. Spores. (All from Shevock 26365, isotype, KRAM). Scale bars: a — 1 mm (1, 1 1—16); b — 1 mm (8—10); c — 100 pm (17—24); d — 100 pm (2—7). 2014] SHEVOCK AND NORRIS: NEW RHEOPHYTIC MOSS 141 Plants aquatic, usually in rapid water of streams and rivers, often rather closely branched, mostly to 8 cm long. Plants dark green to brownish-green in younger portions, glossy black in older portions, frequently with most of the leaf laminae removed by water scouring, leaving only the costa. Leaves to 5 mm long, 1.5-3, 5: 1, not decurrent. Leaf apex acute to acuminate with apex not cucullate. Leaves broadly keeled above midleaf. Leaf margins serrulate to serrate to finely dentate, usually plane. Leaves mostly bistratose, but with margins unistratose for several cells along the border. Laminal cells of median limb isodiametric, about 10 pm or nearly so, smooth. Laminal cells of median portions of the sheathing base to 12 pm broad, short- rectangular, mostly 2-3: 1, thick-walled with lumen; wall ratio 1-2: 1, mostly not pitted. Costa on adaxial side having 6-8 cell rows of rectangu- lar epidermal cells, filling more than 20% of leaf, tapering very little to the subpercurrent apex, with rhizoids cloaking its abaxial basal 1/5 or more. A broad and almost white to pale infra- marginal limbidium usually present, often re- stricted to proximal one quarter of leaf, but sometimes extending to near leaf apex with this variability in extent shown even on leaves of the same clone. Axillary hairs 5-6 cells long to 100 pm with 2 short but concolorous basal cells, cylindric and of constant diameter thick- walled (lumen: wall ratio 1.5-3: 1). Rhizoids red-brown with somewhat warty surfaces, to 35 pm in diameter at insertion, mostly on oldest parts of stem, and also on the adaxial face of the costa, mostly below mid-leaf. Plants presumed dioicous, perigonia not seen; perichaetial leaves generally reduced in length and width compared to vegetative leaves. Seta mostly less than 1 mm long straight and smooth emerging from a short vaginulum with a heavy cover of paraphyses; calyptra covering half or less of maturing capsule, cucullate, smooth, naked. Capsule systylious, mostly 2.5-3 mm long, nearly globose when young and operculate but shrinking and becoming ring-like after dehiscence of apicu- late operculum, then becoming 2-3 times broader than tall with the longitudinally ribbed columella occupying the central axis of the shortened capsule and with the spores surrounding that columella. Median exothecial cells to 15 pm wide, 1-2.5: 1 rather thick-walled, arranged in regular longitudi- nal rows. Peristome of 32 fragile, short, triangular teeth, to 400 pm long, reddish when young, aging brown, inserted 1-2 cells below mouth of capsule, smooth with prominent dorsal trabeculae, blunt at apices, reflexed when dry, slightly incurved when wet. Teeth cover only a very small portion of the capsule mouth. Spores spherical, generally light brown, to 40 pm in diameter, ornamented with low anastamosing ridges. Paratypes: USA, OREGON. Coos Co.: South Fork Coquille River along forest road 33 at Elk Creek Falls, Rogue River-Siskiyou National Forest, 165 ft, 21 Mar 2005, Shevock 26421 (CAS, CONC, H, KRAM, LE, MO, NY, UBC, UC); Same location, 14 Jun 2005, Wagner ml697 (CAS, OSC); Myrtlewood Grove Campground, 375 ft, 21 Mar 2005, Shevock 26438 (CAS, MO, NY, UC). Curry Co.: Elk River Road about 17 mi E of highway 101, Rogue River-Siskiyou Na- tional Forest, 8 Mar 1972, Norris 21977 (UC); Elk River Road about 17 mi E of highway 101, Rogue River-Siskiyou National Forest, 8 Mar 1972, Norris 21979 (UC); about 10 mi SE of highway 101 east of Port Orford, 300 m, 25 Jan 1995, Norris 84649 (CAS, H, LE, UC); Elk River Road 1.4 mi E of fish hatchery, 530 ft, 1 Mar 2013, Shevock 41901 (CAS, CONN, DUKE, E, F, H, KRAM, LE, MHA, MO, NY, OSC, UBC, UC, US); Bear Creek at confluence with Elk River 2.2 mi E of fish hatchery, 550 ft, Shevock 41906 (BOL, CAS, COLO, CONC, E, H, KUN, MO, NY, OSC, UC); Bear Creek at confluence with Elk River 10 mi E of fish hatchery, 850 ft, Shevock 41910 (CAS, CONC, CONN, H, HO, KRAM, L, LE, MHA, MO, NY, OSC, TNS, UBC, UC, US); Redwood State Park about 8 mi E of Brookings, along Chetco River, 3-9 Sep 1950, Koch 3245 (UC); Redwood State Park about 8 mi E of Brookings, along Chetco River, 3-9 Sep 1950, Koch 3270 (UC); Redwood State Park about 8 mi E of Brookings, along Chetco River, 3-9 Sep 1950, Koch 3290 (UC); Chetco River, Siskiyou National Forest, T39S, R12W, SI 3, 400 ft, 17 Oct 2002, Jones 3954 (OSC); South Fork Chetco River just above confluence with Chetco River, at milepost 8, Rogue River- Siskiyou National Forest, 155 ft, 19 Jan 2013, Shevock and Lambio 41752 (BOL, CAS, COLO, DUKE, E, F, H, HO, KRAM, L, LE, MHA, MO, NY, OSC, UBC, UC, US); Winchuck River at forest boundary, 0.5 mi west of Winchuck Campground, Rogue River-Siskiyou National Forest, 175 ft, 19 Jan 2013, Shevock and Lambio 41759 (CAS, CONC, H, KRAM, LE, MHA, MO, NY, OSC, UBC, UC). Josephine Co.: South Fork Taylor Creek along forest road 25 near Tin Can Campground, Rogue River-Siskiyou Na- tional Forest, 1000 ft, 28 Feb 2013, Shevock and Loring 41881 (CAS). Taxonomic Relationships Scouleria siskiyouensis has gone undetected due to its similarity to other Scouleria species in southern Oregon. The plants form robust colo- nies similar in appearance to populations of both S. aquatica and S. marginata. These taxa, however, are readily separated. When sporo- phytes are present, S. siskiyouensis has 32 fragile peristome teeth reflexed on the mouth of the capsule while peristome teeth are absent in S. marginata. Gametophytically the leaves in S. 142 MADRONO siskiyouensis are primarily bistratose across the median region but cells adjacent to the border are unistratose. Scouleria marginata on the other hand is unistratose across the lamina (rarely with bistratose streaks) but the leaf margins are multistratose and appear considerably thickened called ‘pseudocostae’ by Churchill (1985). Scou- leria aquatica resembles S. siskiyouensis but it is generally smaller in stature with a more rounded leaf apex. Sporophytes appear to be identical between these two species although they are exceedingly more common in S. siskiyouensis compared to S. aquatica. A leaf cross-section is the most reliable method of species recognition between these related taxa. Scouleria aquatica can be distinguished from S. marginata by the marginal cells of the vegetative leaves that are in only one layer or occasionally with bistratose streaks and are somewhat enlarged and thin- walled compared to the 4+ layers of stereid like cells in S. marginata. This feature of having multistratose borders can be observed with a hand-lens in S. marginata. Scouleria aquatica has peristome teeth, which differ from the eperisto- mate S. marginata. Based on a recent molecular study (Carter et al. 2014), Scouleria siskiyouensis is sister to S. aquatica. Habitat and Ecology Scouleria siskiyouensis is restricted to fast flowing, unpolluted rivers and streams with large boulders. Plants seem to require seasonal sub- mergence, an extended period in the splash zone, then a period of complete desiccation. Several other rheophytic mosses can be associated with Scouleria siskiyouensis. Scleropodium obtusifolium (Mitt.) Kindb. and to a lesser degree Codrio- phorus aciculare (Hedw.) P. Beauv. and Schisti- dium rivulare (Brid.) Podp., are the most common associates. Scouleria is a genus that requires seasonal submersion, cold, clean water, and periods of desiccation during the dry summer months. Changes in such hydrologic function could rapidly cause the extirpation of this species from those locations. Scouleria can also be very localized even in river and stream systems where it occurs. Plants are found on large boulders or walls of bedrock micro-sites where they cannot be displaced during floods or periods of peak flows. Populations can occupy up to a meter wide band in the water column depending on river flow. Generally, Scouleria occurs just below the high water zone. A period of time in the splash zone among white-water rapids appears to be critical for the establishment of sporophytes. Being both dioicous and a rheophyte may account for limited success of sporophyte production. As an aquatic moss, fertilization has to occur while the plant is hydrated and free water is readily available and plants need to stay hydrated (in the splash zone) [VoL 61 i while sporophytes are maturing. However, among North American Scouleria, S. siskiyouen- sis produces sporophytes considerably more frequently with mature capsules appearing during the winter season and sporophytes also appear to develop in series as water levels fluctuate between winter storms. Sporophytes of Scouleria sis- kiyouensis have been observed in all of the populations documented to date. In S. aquatica and S', marginata, sporophytes are generally produced later in the year depending on eleva- tion, however, based on our field observations both species produce sporophytes more frequent- ly in Oregon than in adjacent California. Among rheophytic mosses the peristome teeth in Scouleria are rather unusual. The role of the peristome is traditionally understood as a means of facilitating spore dissemination by opening of the capsule mouth at times when environmental conditions are favorable for spore release. Peri- stomate bryophytes differ in the patterns of reflexing and inflexing generally based upon ambient humidity. In Scouleria, the peristome seems to have neither function. Peristome teeth in Scouleria are so small and fragile that they are unlikely to carry their unusually large spores toward the capsule mouth to aid in dispersal. The mouth of the capsule is exceedingly broad in relation to the tiny ring of peristome teeth, and therefore, these teeth provide no means of restriction of the capsule mouth. The release of spores in Scouleria actually occurs when the systylious globose capsule shrinks longitudinally from the operculum and columella like a donut causing the bulk of the spores to be forced out by this compression action. The remaining spores lodged in the base of the capsule are disseminated as the capsule wall disintegrates. Additional spores can also are removed from the opened capsules in the event water levels rise again after capsule dehiscence. The peristome teeth in Scouleria are but mere ornamentations and seem to be a useless structure, normally reflexed along the capsule wall. Of the six species of Scouleria worldwide, peristome teeth have been lost in only S. marginata. Distribution Populations of Scouleria siskiyouensis are currently restricted to southwestern Oregon, primarily in the Siskiyou portion of the Rogue River-Siskiyou National Forest within the Coast Range Ecoregion. Population elevations range from 150 to 1000 feet. This species is a very narrow rheophytic endemic with the distance between the northern and southern occurrences being only 68 air miles. We anticipate additional populations of S. siskiyouensis within this range will be discovered as more bedrock river and stream habitats are surveyed along the western 2014] SHEVOCK AND NORRIS: NEW RHEOPHYTIC MOSS 143 and northern boundaries of the Siskiyou portion of the Rogue River-Siskiyou National Forest and adjacent BLM Coos Bay District lands. Although the southernmost Oregon occurrence along the Winchuck River is just 1.75 air miles north of the California/Oregoe border, we are of the opinion that locating a California occurrence may be limited. We were unable to locate S. siskiyouensis during field sampling in the adjacent Smith River watershed in California although both N. aqua- tica and S. marginata are present but as small, isolated populations. The riparian zone within the Smith is not optimum habitat for the new Scouleria taxon due to the dominance of serpen- tine geology resulting in less riparian vegetation along river and stream corridors. Conservation Implications The majority of the known occurrences of Scouleria siskiyouensis occur on public lands administered by either the USD A Forest Service or the USDI Bureau of Land Management, thereby offering greater opportunities for long- term conservation. In addition, many rivers and perennial streams have added layers of protection by law and regulation to conserve riparian values and anadromous fisheries, especially those occur- rences along Congressionally designated Wild and Scenic Rivers. Dams, however, could be fatal to this species since river hydrology and ecology would be significantly altered. Key to Oregon Scouleria 1. Plants with leaves lanceolate to ovate lanceo- late, leaf apex acute to acuminate; transverse section of leaves primarily bistratose except for a few unistratose cells at immediate leaf margin ................ Scouleria siskiyouensis T. Plants with leaves mainly ovate, leaf apex almost consistently obtuse to bluntly-rounded; transverse section of leaves unistratose or with occasional bistratose streaks or with thicken- ings (multistratose layers) along the immediate margin 2. Leaf margins unistratose or occasionally with bistratose streaks or larger cells along margin but without areas of greater thickness; capsule with 32 reddish short peristome teeth .......... Scouleria aquatica 2'. Leaf margins prominently thickened throughout, somewhat cartilaginous with 4-6 layers of cells in the immediate margin; capsule lacking peristome teeth . ................ Scouleria marginata Acknowledgments We thank Halina Bednarek-Ochyra for producing the wonderful illustration plates for this remarkable plant. Colleagues David Wagner and Scot Loring provided data on Oregon Scouleria populations and specimens for which we are most grateful. A loan of Oregon Scouleria from OSC was very helpful in determining species ranges and is appreciated. We also thank the various land management agencies for collecting permits to develop baseline data on California and Oregon bryophytes, primarily the Rogue River-Siskiyou and Six Rivers National Forests. Comments provided by David Wagner and two anonymous reviewers enhanced the final version. Literature Cited Carter, B., S. Nosratinia, and J. R. Shevock. 2014. A revisitation of species circumscriptions and evolutionary relationships in Scouleria (Scouleria- ceae). Systematic Botany, 39:4-9. Christy, J. A., J. H. Lyford, and D. H. Wagner. 1982. Checklist of Oregon mosses. The Bryologist 85:22-36. Crosby, M., R. E. Magill, B. Allen, and S. He. 2000. A checklist of the mosses. Missouri Botanical Garden, St. Louis. Churchill, S. 1985. The systematics and biogeogra- phy of Scouleria Hook. (Musci: Scouleriaceae). Lindbergia 11:59-71. . 2007. Scouleriaceae. Pp. 311-313 in Flora of North America Editorial Committee (eds.). Flora of North America North of Mexico, Vol 27: Bryophyta: Mosses, part 1. Oxford University Press, New York, NY. Kurbatova, L. E. 1998. De genere Scouleria Hook, in Rossia notula. Novosti sistematiki nizsikh rastenii 32:162-169. [In Russian.] Lawton, E. 1971. Moss flora of the Pacific Northwest. Hattori Botanical Laboratory, Nichinan, Miya- zaki, Japan. Norris, D. H. 1997. The Oregon-California border: important in bryogeography. Journal of the Hattori Botanical Laboratory 82:185-189. and j. R. Shevock. 2004a. Contributions toward a bryoflora of California I: a specimen- based catalogue of mosses. Madrono 51:1-131. . 2004b. Contributions toward a bryoflora of California II: a key to the mosses. Madrono 51:133-269. Savic-Ljubickaja, L. L and Z. N. Smirnova. 1970. Handbook of mosses of the U.S.S.R. The acro- carpous mosses. The Academy of Sciences of the USSR. The Komarov Botanical Institute, Lenin- grad, Russia. Tsubota, H., Y. Ageno, B. Estebanez, T. Yamaguchi, and H. Deguchi. 2003. Molecular phylogeny of the Grimmiales (Musci) based on chloroplast rbcL sequences. Hikobia 14:55-70. Madrono, VoL 61, No. 1, pp. 144-145, 2014 REVIEW The Drunken Botanist. The Plants That Create the World’s Great Drinks. By Amy Stewart. 2013. Al- gonquin Books of Chapel Hill, Chapel Hill, NC. 381 pp. ISBN 97816 16200466. Price, about $20.00, hardcover. Anyone who has ever taken or taught an Economic Botany course, or who simply has been intrigued by natural ingredients in foods, cannot help but be baffled by the tremendous array of botanical products that go into various alcoholic beverages. Grapes, grains, junipers, agaves, nearly everyone knows a bit about the role these plants play in various libations. But this list does not come anywhere close to the array of plant products that appear in the bottles of a well- stocked liquor store. Amy Stewart has written a super-detailed compendium of the plants that “create the world’s great drinks,” and not a few of some of the world’s more obscure drinks. The treatments of plant species vary in their depth and detail, but one would be hard pressed to find a plant that Stewart omitted. In her introduction, appropriately titled “Aperitif,” she states that it would be “impossible to describe every plant that has ever flavored and alcoholic beverage.” Perhaps, but she must be very close to that goal. Around 160 plants from nearly eighty families are treated in the book, plus the necessary fungi and bacteria, and even a few invertebrates. The Drunken Botanist is organized naturally into three parts, beginning with discussions of the processes, natural and human-invented, that account for alcoholic solutions. A brief note at the start of the book titled “About the Recipes” is dedicated to instructing the reader on the proper proportions and size of a cocktail: “do get into the habit of mixing one small, civilized drink at a time” and “please get rid of your jumbo- sized cocktail glasses.” This advice is well placed, as most of the recipes in the book are fairly simple with quality ingredients, but high in alcohol content. Next there is a long catalogue of herbs and spices, flowers, trees, fruits, and seeds; who, for example, would have expected entries on genera such as Croton L., Drosera L., Adiantum L., Araucaria Juss., or Arbutus L. in a book on alcoholic potations? Some plants warrant only part of a page; then there is the six-page account of the labyrinthine saga of angostura bitters. Part three deals with gardening and growing your own plants to use in preparing cocktails (“Gardeners are the ultimate mixolo- gists”). Bay Area dwellers with tiny yards will love her cocktail garden planted in a seven foot wide side yard at her Humboldt County home (see her blog on Gardenrant.com for photos). The book is further punctuated with sidebars titled “Bugs in Booze;” who knew that, up to 2006, the shocking red color of Campari was due to cochineal insects. There are brief discussions on whether it’s mescal or mezcal, whisky or whiskey. There are instruc- tions on how best to grow your own plants to be used in drinks. And there are, of course, recipes for fifty cocktails, and several others for syrups, infusions, and garnishes. A lot in this book will be familiar to the Madrono readership, devoted botanists that we are. In the section on Agave L., taxonomists will appreciate the sidebar titled “The Lumpers, the Splitters, and Howard Scott Gentry,” where Stewart gives a brief but detailed history on the taxonomy of the genus. Gentry believed that the Agave species should be split based on floral characters. Stewart points out the difficulty in using this character due to the fact that one might have to wait as long as thirty years to actually see a specimen in bloom! But there are still likely to be items new to readers: why does gin cause allergic reactions; why precisely does adding water make absinthe cloudy; and which plant contributes to speedy recovery in sexually ex- hausted male rats. The book is attractively designed and Stewart is a delightful raconteur (she is also the author of Wicked Plants: The Weed That Killed Lincoln’s Mother, and Other Botanical Atrocities [2009], and Flower Confidential: The Good, the Bad, and the Beautiful [2008], both published by Algonquin Books of Chapel Hill). She is equally interested in the botany, the history, and the chemistry of plants, and is adept in ferreting out the most obscure and mysterious facts (for example: in its time, George Washington’s whiskey distillery was one of the largest in the country). Stewart has researched her subject very thoroughly. Her taxonomy is up to date (fide Angiosperm Phylog- eny Group) and she even points out recent taxonomic reassignments. The text is refreshingly devoid of large numbers of botanical errors and misstatements that botanists so readily catch in popular writings and cookbooks; we did notice that, on a page discussing European centaury, she 2014] BOOK REVIEW 145 has a drawing of Centaur ea L., not Centaurium Hill, and there are a few sentences that begin with an abbreviated first letter of the genus name, rather than the spelled-out name; botanists don’t do that. The Drunken Botanist is a fun and easy read, easily affordable, and would make an excellent gift for botanists and bartenders alike. And how much fun it would be to offer a course in which this book could be used as the text. It is easy to become addicted to all things Amy Stewart, especially when a good drink is involved. If you like The Drunken Botanist, be sure to visit her website at www.amystewart.com to see, for example, high-resolution photos of many of the cocktails listed in the book. How about a Herbarium Cocktail? Indeed, there is a recipe for it! — Sheryl Creer and Robert Patterson', Department of Biology, San Francisco State University, San Francisco, CA 94132. ^patters@sfsu.edu. Madrono, VoL 61, No. 1, pp. 146-147, 2014 REVIEW Intermountain Flora: Vas- cular Plants of the Inter- mountain West, U.S.A. Subclass Magnoliidae- Caryophyllidae, Vol. 2 A. By N. H. Holmgren, P. K. Holmgren, J. L. Re- veal, AND Collabora- tors. 2012. The New York Botanical Garden Press, Bronx, New York. 742 pp. ISBN 978^0=-89327=- 520-4, Price $150.00, hard- cover. This work constitutes the finale of the Intermountain Flora series, begun in 1972 and encompassing six volumes total, with volumes 2 and 3 each published in two parts. The series covers the plants of an extensive region between the Sierra Nevada and Rocky Mountains, encompassing the entire Great Basin plus the Wasatch Mountains, Uinta Mountains, and Colorado Plateau. This range includes central and northern Nevada, extreme eastern/northeast- ern California, southeastern Oregon, southern Idaho, extreme southwestern Wyoming, all of Utah, and extreme northwestern Arizona. From its inception the Intermountain Flora series has used Cronquist’s (e.g., 1981) classifica- tion, with this last treatment (Volume 2, Part A) covering his subclasses Magnoliidae, Hamameli- dae, and Caryophyllidae. However, because Cron- quist’s system has been largely superseded by others derived from molecular phylogenetic anal- yses (notably the Angiosperm Phylogeny Classifi- cation [APG], with the most recent version being APG III 2009), some discrepancies with these recent systems are apparent. In the Intermountain Flora, the Magnoliids include not only the Mageoliaceae and Saururaceae, but also both the Nymphaeales/Nymphaeaceae (now accepted as one of the first lineages of angiosperms and basal to the Magnoliids), and the Ranunculales (now accepted as basal Eudicots). In subclass Hamamelidae of the Intermountain Flora, the three orders treated — Fagales, Hamamelidales, and Urticales — have family compositions mostly corresponding to APG III, but their higher-level classification is quite different in the latter, indicative of the fact that wind pollination evolved independently among many of these taxa. The third subclass treated in the Intermountain Flora, the Caryophyllidae (Order Caryophyllales), agrees essentially entirely with that of APG III. Despite some differences in classification of families within subclasses and some orders, the actual delimita- tion of families follows that of APG III or other recent studies. For example, the Montiaceae and Sarcobataceae are treated as families separate from other caryophyllids. Aside from these obvious differences in higher-level groupings, this last treatment of the Intermountain Flora adheres to very high stan- dards. A practical key of the families, without reference to suprafamilial groups, precedes taxo- nomic treatments. Subclasses, orders, and families are each described, with summarizing features, differences in classification from other systems, and up-to-date references cited. Keys to genera and species are very well constructed. Listed in the treatments are number of total species within families and genera, etymology of scientific names, chromosome numbers, habitat informa- tion, elevation, geographic range, notes on cultivation, close relatives, and notes on classifi- cation, including recent phylogenetic studies. What I particularly like about the Intermoun- tain Flora series is the complete listing of synonyms, both heterotypic and homotypic, for each taxon, these including author(s), publica- tion, type-collection data (collector, number, date of collection, and habitat), and citation of holotype, isotype, and/or lectotype specimen(s). Common names are given for every taxon, granted that this is a source of contention among professional taxonomists. Species descriptions are particularly complete, with descriptive termi- nology precise and specialized (not “watered down”), the latter a joy for professional bota- nists. Line drawings are of extremely high quality and illustrate the whole plant plus close-ups of flowers and/or fruits, with emphasis on diagnostic features. Illustrations are presented for every taxon in the book, eliminating a common source of frustration experienced with other floras. Finally, the text formatting is concise and efficient, with artistic and pleasing use of font type, font size, and paragraph and line spacing. In summary, this last contribution of the Intermountain Flora follows the same scholarly standards of earlier volumes and incorporates up- to-date information from current research pa- pers. However, this book is more for the professional botanist, as opposed to the layper- son. Although the higher-level grouping of families is sometimes antiquated and the geo- graphic scope of the Intermountain Flora volumes is somewhat limited, the quality of taxonomic treatments makes this book exceptionally worthy as a reference. Botanists will find the series useful not just for keys and taxonomic treatments of INTERMOUNTAIN FLORA Vascular Plants of the Intermountain West, U.S.A. 2014] BOOK REVIEW 147 plants occurring in this geographic region, but also for the detailed information presented for specific taxonomic groups. These are essential references to be consulted on the plants of western North America. — Michael G. Simpson, Department of Biology, San Diego State Univereity, San Diego, CA 92182, msimpson@ mail.sdsu.edu. Literature Cited 1. Angiosperm Phytogeny Group (APG III). 2009. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG III. Botanical Journal of the Linnean Society 161:105-121. 2. Cronquist a. 1981. An integrated system of classification of flowering plants. Columbia Univer- sity Press, New York, NY. Madrono, Vol. 61, No. 1, p. 148, 2014 NOTEWORTHY COLLECTION CALIFORNIA Taxus brevifolia Nutt. (TAXACEAE). — Tuo- lumne Co., Mi Wok Ranger District, Stanislaus National Forest, along Basin Creek below the crossing of Forest Road 1N04, UTM NAD 1983 zone ION 4208683 751056, T2N R17E Sec31 NENE, the portion of the stand on Forest Service land has been mapped as 8.9 acres and has approximately 200 trees, the largest tree measured was 12" dbh., most of the larger trees had multiple stems, the stand also extends to section 30 SE and, on private land, into section 32 NWNW, old growth mixed conifer forest on a north aspect at 3560 ft elev, growing on a granitic, highly weathered Holland soil derived from Mesozoic granitic rocks, 18 March 2009, Jonathan Day s.n. (UC1965998, UC1965999). This stand was first noted by Hans Bayer in 1992 and was mapped by Kathy Aldrich around 1995. Previous knowledge. Prior to this collection, T. brevifolia was thought to reach its southern limit in the Sierra at Calaveras Big Trees State Park in Calaveras County (Rundel 1968). Rundel ruled out the reports from Yosemite and Tulare Counties including an 1874 specimen labeled Yosemite Valley (Mariposa County, Lemmon s.n.). This species was not included in an illustrated flora of Yosemite National Park (Botti 2001) and no other recent reports have surfaced. As follow-up on the Lemmon collection, Jonathan Day, Kate Day, and Margaret Willits checked two of the most likely areas in Yosemite Valley for this species on 13 March 2009. One additional likely area was not surveyed due to rockslide danger. We did not find T. brevifolia. There is a possibility of T. brevifolia occurring in the Merced Canyon below Yosemite. Brandegee (1891, pp. 160) reported “The yew {Taxus brevifolia) grows near the water in the canon of the Merced.” and “Neither this tree [Torreya Am.] nor the last [Taxus L.] quite reaches the valley.” This information was not considered by Rundel (1968). Significance. This is the first collection from Tuo- lumne County. It is also the southernmost extant stand known in the Sierra. This collection is approximately 34 km SSE of the North Grove, Calaveras Big Trees, the previous limit. This species is on the watchlist for the Stanislaus National Forest and would be addressed and consid- ered in the NEPA (National Environmental Policy Act) process. — Margaret L. Willits*, Jonathan Day, Mi Wok Ranger District, P.O. Box 100, Mi Wuk Village, CA 95346. 'mwillits@fs.fed.us; Heppner Ranger Dis- trict, 117 South Main Street, Heppner, OR 97836. Acknowledgments Jennie Haas, Alison Colwell, Stanley Scher, Dieter Wilkin, and Matt Ritter provided helpful suggestions and editing. A huge effort of many people stopped the Rim Fire before it reached this stand. Jonathan and Kate Day provided careful surveying and good company on cold, snowy north aspects. Literature Cited Botti, S. J. 2001. An illustrated flora of Yosemite National Park. Yosemite Association, El Portal, CA. Brandegee, K. 1891. The flora of Yo Semite. Zoe 2:155-167. Rundel, P. W. 1968. The southern limit of Taxus brevifolia in the Sierra Nevada. Madrono 19:300. Madrono, VoL 61, No. 1, pp. 149-150, 2014 NOTEWORTHY COLLECTION New Mexico Hexalectris COLEMAN II (Catling) A. H. Kenn. & L. E. Watson (ORCHIDACEAE).— Hidalgo Co., Coronado National Forest, Peloncillo Mountains, Skeleton Canyon 7.5' USGS quad, 12A 0685596mE, 3486984mN (NAD83), 31.5028°N, 109.0457°W (WGS84), 1620 m (5315 ft) elevation, Cottonwood Canyon, 390 m east of Arizona-New Mexico state boundary, 1 30 m east of Geronimo Trail, 31 May 2012, M. A. Cloud- Hughes 027, with M. A. Baker and R. A. Coleman (UNM 126990). Single individual with five flowers (one flower collected) in sandy soil between boulders on eastern side of Cottonwood Canyon in duff under Quercus grisea Liebm./g. arizonica Sarg. canopy with Agave palmeri Engelm., Arctostaphylos pungens Kunth, Elymus ely- moides (Raf.) Swezey, Ericameria laricifolia (A. Gray) Shinners, Juniperus deppeana Steud., Muhlenbergia emersleyi Vasey, Nolina microcarpa S. Watson, Q. emoryi Torr., Rhus glabra L., and R. trilobata Nutt. Previous knowledge. Formerly considered endemic to the mountains of southeastern Arizona, this species was originally collected by Toolin and Reichenbacher in 1981 in Baboquivari Canyon, Baboquivari Mountains, Pima Co., Arizona but was identified as Hexalectris spicata (Walter) Barnhart (ARIZ252881) (Southwestern Environmental Information Network [SEINet] 2013). In 1986, McLaughlin collected the species from McCleary Canyon, Santa Rita Mountains, Pima Co., Arizona, and these specimens were also initially identified as H. spicata (ARIZ271012) (SEINet 2013). In 1996, orchid specialist Ron Coleman photographed plants at McLaughlin’s McCleary Canyon collection site and, along with Canadian orchid expert Paul Catling, identified the plants as Hexalectris revoluta Correll (Coleman 1999). Coleman located the species at additional sites in Pima Co., Arizona, and in 2004, Catling determined that these specimens were suffi- ciently different from other Hexalectris revoluta speci- mens to separate them as Hexalectris revoluta var. colemanii Catling (Catling 2004), This taxon was subsequently raised to species status, Hexalectris colemanii, after a phylogenetic analysis of the genus Hexalectris by Kennedy and Watson (2010). Hexalectris colemanii has not been found in the Baboquivari Mountains since its original collection in 1 98 1 , in spite of repeated surveys by Coleman (Coleman 2002). Toolin and Reichenbacher found only 12 shoots of the species in 1981, and it is most likely this population has been extirpated by grazing. As part of a sensitive plant survey for the Coronado National Forest in 2003, Baker confirmed all earlier Santa Rita Mountains localities of Hexalectrix colemanii and found an additional population further south in Sawmill Canyon (Baker 2003). In 2010 through 2012, WestLand Resources, Inc. conducted surveys for Hexalectris colemanii as part of the permitting process for a proposed copper mine in the Santa Rita Mountains. These surveys resulted in the discovery of small populations of Hexalectris colemanii in four additional canyons within the Santa Rita Mountains, four canyons in the Big Dragoon Moun- tains, two canyons in the Whetstone Mountains, one canyon in the Chiracahua Mountains, and one canyon in the Peloncillo Mountains (WestLand Resources, Inc. 2012). All of these recently discovered localities are in Arizona. In 2012, Southwestern Botanical Research was contracted through the U.S. Fish and Wildlife Service and the Arizona Department of Agriculture to conduct surveys for Hexalectris colemanii. After conferring with WestLand to avoid duplication of effort with their large survey teams, we focused our surveys in the outlying Atascosa, Pajarito, Mule, northern Chiracahua, and Peloncillo Mountains. We found three shoots of Hexalectris colemanii in the Peloncillo Mountains of Arizona: two in Cottonwood Canyon, which were almost certainly from the same rhizome, and one in Miller Canyon. The final shoot of Hexalectris colemannii found on our survey was in Cottonwood Canyon approximately 400 m over the border in New Mexico. Significance. This is the first record for Hexalectris colemanii in New Mexico. Because only one individual was found, Hexalectris colemanii can reasonably be considered the rarest orchid in New Mexico. Few Hexalectris species have been reported in the most recent floras of New Mexico (Martin and Hutchins 1980; Allred et al. 2012). A search of the SEINet herbarium specimen database shows one specimen (TEUI5647) collected approximately 25 km northwest of our specimen and identified as Hexalectris c.f. revoluta (SEINet 2013). The authors and Coleman have examined this specimen and are in agreement that, based on floral and fruit characters, it is Hexalectris arizonica (S. Watson) A. H. Kennedy & L. E. Watson. Although the Coleman, WestLand Resources, Inc., and Southwestern Botanical Research surveys have significantly extended the known range of Hexalectris colemanii, this orchid should still be considered a very rare species. The largest number of individual shoots in any one population in a single year is approximately 140. Because this orchid is rhizotomous, the number of shoots likely represents a much smaller number of genetically distinct individuals. Furthermore, Cole- man’s monitoring efforts in McCleary Canyon have shown that the number of Hexalectris colemanii shoots in a population is extremely variable, with nearly every “good” year (e.g., 40^5 shoots in 2001) followed by a crash to zero or near zero the following year (WestLand Resources, Inc. 2012). Major threats to Hexalectris colemanii include fire, grazing and trampling by livestock, mining, and threats to pollinator populations. As an obligate myco- heterotroph, Hexalectris colemanii is also vulnerable to any threat to its host plants, their ectomycorrhizae, or the litter layer. This orchid has so far been found only in “sky-island” habitats, which are inherently limited. Hexalectris colemanii currently has no federal status, in spite of its limited range, very small known populations, and multiple threats. Efforts should continue to be made to protect known populations and locate new ones to ensure the long-term survival of this orchid. 150 MADRONO [Vol. 61 — Michelle A. Cloud-Hughes^ and Marc A. Baker^. ‘Desert Solitaire Botany and Ecological Resto- ration, San Diego, CA 92103. mcloudhiighes@gmaiL com. ^Southwest Botanical Research, Chino Valley, AZ 86323. Literature Cited Allred, K. W. 2012. Flora Neomexicana Volume I: Annotated checklist, 2nd edition. Lulu, Raleigh, N.C. Baker, M. A. 2003. Sensitive plant survey for the Coronado National Forest. Project No. 2003-01- FLORA: in fulfillment of Coronado National Forest Contract No. 43-8197-3-0038. Final Report, 19 December 2003. Southwestern Botanical Re- search, Chino Valley, AZ. Unpublished report. Catling, P. M. 2004. A synopsis of the genus H exale ctr is in the United States and a new variety of Hexalectris revoiuta. The Native Orchid Con- ference Journal 1:5-25. Coleman, R. A. 2002. The wild orchids of Arizona and New Mexico. Comstock Publishing Associates, a division of Cornell University Press, Ithaca, New York. Pp. 98-102. . 1999. Hexalectris revoiuta in Arizona. North American Native Orchid Journal 5:312-315. Kennedy, A. H. and L. E. Watson. 2010, Species delimitations and phylogenetic relationships with the fully mycO“heterotrophic Hexalectris (Orchida- ceae). Systematic Botany 35:64-76. Martin, W. C. and C. R. Hutchins. 1980. A flora of New Mexico, Volume 1. J. Cramer, Vaduz. Southwest Environmental Information Net= WORK, SEINet. 2013. Website http//:swbiodiversity. org/seinet/index.php (accessed April- August 2013). Westland Resources, Inc. 2012, Biology and life history of Coleman’s Coralroot {Hexalectris cole- manii) and surveys for Hexalectris in 2011. West- Land Resources, Inc., Tucson, AZ. . • ' .. Mr' .. „:. . ;'v •■ ~ : Htiit- iU •> i. • , f r ',nt.'tth ■> .-r- re I !', ■•■ > / 4 ^ I J ♦'!'• M ' -■> -•• f,.-. '■:,'i V '' .if. •Tff' \ t: i ; .U ' •*** ‘.4, »■'' .. >! ' . 'iji '^' ' > f '• f ; ■' i H ' tn' . uf M • ’ ,•^ \\7,<:- ■ A nttU ’.'^'"•4 •■■'' ;| ‘f ‘ .. -I ■ ^ V ■>'*>•*?»», j«?t...- -V, Tv-". T] ^ ‘j -f '»,K -K, A. »»■ «^-s» i,. '''«* '■■Vl ■n»‘J23Wr®'»* • ♦>■ ■ 'iwi -, - ''vi -t-A, JilifU-*' ^101^ .;. •■ u •'•'^#** / ■ ■':•'.■■ .w tuTA/i'fi^.;^ Nrijtli H ' < it,,tr'*.t> ' ^ t .'■''* * ; f... ’H i >: ’•v.‘COVP^ .i-rKD^ J' j .\ <_$*ii ir ^ 1, Vl'tvs N^t. v./i- ■ .'■ vvu it' j.i 4. ‘ ■-f (ii .-j- #* • '■•■ ■' . '■' «a >'-^^'s 'im^' ■ .'v ■ ® ' ■JU'-'f-t i ■■^(.■.H*^'-^^^, .I-' . ':ii 7. '• „i«Vv , ■ lit .7^ , ^ # - “.til >4 ■ffi ■■ i-."'ti(i >-.'v3!3^j' : * I -4 • ^ :-Wjt.^ ',- :: (Vu. ' C • ■ . li ■'f.%^'.- . * ■ W; ;:..®: Mak ^ '>' ’* 74- *w'' ^ ■3r4> • (gw's. '1 Subscriptions — Membership The California Botanical Society has several membership types (individuals ($40 per year; family $45 per year; emeritus $32 per year; students $27 per year for a maximum of 7 years). Late fees may be assessed. Beginning in 2011, rates will increase by $5 for all membership types except life memberships, for which rates will increase by $100, and student memberships, which will not show a rate increase. Members of the Society receive Madrono free. Institutional subscriptions to Madrono are available ($75). Membership is based on a calendar year only. Life memberships are $850. Applications for membership (including dues), orders for subscriptions, and renewal payments should be sent to the Membership Chair. Requests and rates for back issues, changes of address, and undelivered copies of Madrono should be sent to the Corresponding Secretary. Information for Contributors Submit manuscripts to http://www.edmgr.com/madrono/. If you are submitting a manuscript for the first time to the Madrono PeerTrack website, you must first register by clicking "register now" and following the instructions. Manuscripts by authors having outstanding page charges will not be sent for review. Manuscripts may be submitted in English or Spanish. English-language manuscripts dealing with taxa or topics of Latin America and Spanish-language manuscripts must have a Spanish Resumen and an English Abstract. For all articles and short items (NOTES, NOTEWORTHY COLLECTIONS, POINTS OF VIEW, etc.), follow the format used in recent issues for the type of item submitted. Detailed information for contributors, including instructions for manuscript preparation, is available at the California Botanical Society website, www.calbotsoc. org; follow the link for Madrono. Authors are encouraged to include the names, addresses, and e-mail addresses of two to four potential reviewers with their submitted manuscript. All California Botanical Society members current in the volume year that their contributions are published are allowed five free pages per volume year. Additional pages will be charged at the rate of $40 per page. Joint authors may apply all or a portion of their respective five-page allotments to a jointly-published article. Partial pages will be charged as full. 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 color plates and other complex matter (including illustrations, charts, maps, photographs) will be charged at cost. Author’s changes after typesetting @ $4.50 per line will be charged to authors. Page charges are important in maintaining Madrono as a viable publication, and timely payment of charges is appreciated. 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. Geomorphic Landforms and Plant Immunity Structure and Dominance in the Central Desert Region of Baja California, Mexico., m Philip W^Rundel, M. Rasoul Skarifi, Erik T. Niisen, Gail A. Baker, Ross..: L^irginia, and Leila M. Rei^tionships Between Epidermal Browning, Girdling, Damage, and Bird Cavitie^in a Military Restricted Database of 12,000+ Plants of the. Keystone ~Carnegiea gigantea in the Northern Sonoran Desert ' ’ Shelley Danzer and Taiy-Dawn Drezner.,.,... ........ ................................... . V A"'- A .Cholla Mori 1(1 A} - iTALiTY AND Extreme Drought in the Sonoran Desert Edward G. Bobich, Mck L. Wallace andsKeely_^rSarmrL NEW SPECIES -C/: ScOULERIA SISKIYOUENSIS (ScOULERIACEAE), A Nw ^EOPHYTIC Moss ’.j - Endemic Southern Oregon, James R. sfievock and DanielH. r Norris 13 BOOK REVIEWS / ^ fJ y. ^y;L ,V The Drunken Botanist ^ .. - ‘3- / ^Sheryl Creer and'Robert PattersJr^.... ................. .................................... 14 ' / /.I "I ' , A . Intepaiountain Flora yVAScuLAR Plants of the Intermountain West, U.S.A. / / V' ^ Michael G. Simpson fi... ................................................. lA NOTEWORTHY COLLECTIONS California... New Mexico 't: : 'V El VOLUME 61, NUMBER 2 APRIL-JUNE 2014 MADRONO A WEST AMERICAN JOURNAL OF BOTANY Serotiny in California Oaks Walter D. Koenig, Eric L. Walters, Ian S. Pearse, William J. Carmen, and Johannes M. H. Knops 151 New Records for Rare and Under-Collected Aquatic Vascular Plants of Yellowstone National Park ^ C. Eric Hellquist, C. Barre Hellquist, and Jennifer J. Whi^fie Stomata Size in Relation to Ploiby-^Leveljw^^^ Hawthorns ( Cra ta e W5,-RasACE ae) ^ ' ' Brechann V McGoey, Morphology and DiscocACjm Root Gpfeih Fire Effectjstn a M Joh^A^ Chris^^(Cjn^id McCain, S^qi r/q0 \i' Me >■— ■/ -JiiKSim-.. .... ;m^V^®^ENTRALL|j^p6^ Oregon irah ^cj^reenmfimdyennifer V|i.3^!!'k.'L 201 How Does Simulated Gopher Disturbance Affect wL Establishment OF HoLCUS LAI^US L. (PoaIjEAEIhN CkLFORNI^^^fTAL PrAIRIE? Meredith A. Thomsen and Carla MTD'Jmtonio 3^^/.... 218 Z A New Species of Triteleia (Themidaceae) from the Southern Sierra Nevada Ed Kentner and Kim Steiner..... 227 Cylindropuntia chuckwallensis (Cactaceae), a New Species from Riverside and Imperial Counties, California Marc A. Baker and Michelle A. Cloud-Hughes 231 Annotated Checklist of the Vascular Plants of Santa Cruz County, California Jenn Yost 244 PUBLISHED QUARTERLY BY THE CALIFORNIA BOTANICAL SOCIETY Madrono (ISSN 0024-9637) is published quarterly by the California Botanical Society, Inc., and is issued from the office of the Society, Herbaria, Life Sciences Building, University of California, Berkeley, CA 94720. Subscription information on inside back cover. Established 1916. Periodicals postage paid at Berkeley, C A, and additional mailing offices. Return requested. Postmaster: Send address changes to Madrono, Kim Kersh, Membership Chair, Uni- versity and Jepson Herbarium, University of California, Berkeley, CA 94720-2465. kersh@berkeley.edu. Editor — Matt Ritter Biological Sciences Department Cal Poly, San Luis Obispo 1 Grand Avenue San Luis Obispo, CA 93407 madronoeditor@gmail.com Editorial Assistant — Genevieve Walden Book Editor — Matt Ritter Noteworthy Collections Editor — David Keil Board of Editors Class of: 2014 — Brandon Pratt, California State University, Bakersfield, CA Tom Wendt, University of Texas, Austin, TX 2015 — Anna Jacobson, California State University, Bakersfield, CA CALIFORNIA BOTANICAL SOCIETY, INC. Ofeicers for 2012-2013 President: V. Thomas Parker, Department of Biology, San Francisco State University, San Francisco, CA 94132, parker@sfsu.edu Eirst Vice President: Andrew Doran, University and Jepson Herbaria, University of California, Berkeley, CA 94720, andrewdoran@berkeley.edu Second Vice President: Vacant. Recording Secretary: Michael Vasey, Department of Biology, San Francisco State University, San Francisco, CA 94132, mvasey@sfsu.edu Corresponding Secretary: Anna Larsen, Jepson Herbarium, University of California, Berkeley, CA 94720, secretary@calbotsoc.org Treasurer: Thomas Schweich, California Botanical Society, Jepson Herbarium, University of California, Berkeley, CA 94720, tomas@schweich.com The Council of the California Botanical Society comprises the officers listed above plus the immediate Past President, Dean Kelch, Jepson Herbarium, University of California, Berkeley, CA 94720, dkelch@berkeley. edu; the Membership Chair, Kim Kersh, University and Jepson Herbaria, University of California, Berkeley, CA 94720, kersh@berkeley.edu; the FJ/Yor of Madrono; and three elected Council Members: Staci Markos, University and Jepson Herbaria, University of California, Berkeley, CA 94720, smarkos@berkeley.edu. Graduate Student Representatives: Genevieve Walden, Department of Integrative Biology and Jepson Herbarium, University of California, Berkeley, CA 94720, gkwalden@gmail.com. Administrator: Lynn Yamashita, University of California, Berkeley, CA 94720, admin@calbotsoc.org. Webmaster: Ekaphan (Bier) Kraichak, University of California, Berkeley, CA 94720, ekraichak@gmail.com. @ This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Madrono, VoL 61, No. 2, pp. 151-158, 2014 SEROTINY IN CALIFORNIA OAKS Walter D. Koenig Cornell Lab of Ornithology, 159 Sapsucker Woods Road, Ithaca, NY 14850, and Department of Neurobiology and Behavior, Cornell University, Ithaca, NY 14853 wdk4@co rnell . edu Eric L. Walters Department of Biological Sciences, Old Dominion University, Norfolk, VA 23529 Ian S. Pearse Cornell Lab of Ornithology, 159 Sapsucker Woods Road, Ithaca, NY 14850 William J. Carmen 145 Eldridge Ave., Mill Valley, CA 94941 Johannes M. H. Knops School of Biological Sciences, University of Nebraska, Lincoln, NE 68588 Abstract Although prolonged seed retention, or serotiny, is believed to be an adaptation to highly variable environments such as the Mediterranean regions of California, no prior study has systematically investigated the prevalence of seed retention among California oaks (family Fagaceae), the dominant woody taxon in California foothill woodlands. We quantified the extent to which acorns were retained into and through the winter and spring within the canopy of five species of California oaks at Hastings Reservation, Monterey County. Significant serotiny was found in coast live oak {Quercus agrifolia) and, to a lesser extent, valley oak (Q. iobata), but was absent in blue oak (Q. douglasii), canyon live oak {Q. chrysoiepis), and California black oak (Q. kelloggii). In both species where serotiny was observed, seed retention was primarily predicted by the size of the focal tree’s acorn crop. In addition, serotiny in coast live oaks was more prevalent in dry years and when the overall acorn crop of coast live oaks was large. We found no evidence that acorn fall in these species is triggered by a specific environmental event. Prolonged seed retention in California oaks renders acorns available in the canopy to wildlife throughout the winter in some years with potentially significant effects extending beyond those of acorn abundance per se. Key Words: Acorns, acorn retention, oaks, serotiny. Serotiny is an adaptation of plants to retain their seeds for an extended period of time, in some cases for years or decades, often, although not necessarily, releasing them in response to some specific environmental trigger (Lament 1991). A common and widespread trigger is fire, and there has been extensive study of species that release their seeds in response to fire or intense heat (Lament et al. 1991; Bond and Van Wilgen 1996). The factors known to facilitate seed release in serotinous species are varied, however, and include plant or branch death, solar radiation, and dryness (Cowling and Lament 1985; Lament et al. 1991; Nathan et al. 1999). Serotiny has been suggested to be particularly common in environ- ments subject to high spatial and temporal variation in water availability (Evanari et al. 1982; Nathan et al. 1999). Mediterranean climates are notable for their high temporal variability in rainfall, frequency and intensity of extreme events, and potential sensitivity to climate change (Sanchez et al. 2004; Xoplaki et al. 2004). Thus, to the extent that serotiny is important in environments that are highly variable in water availability, it would be surprising if oaks (genus Quercus L.) — the dominant tree taxon in California woodland- savanna (Griffin and Critchfield 1972; Allen-Diaz et al. 2007) covering some 5 X 10® ha of hardwood rangeland in the state (Standiford 2002) — failed to exhibit some degree of serotiny, with potentially important consequences to the diverse community of wildlife depending on acorns for food (Pavlik et al. 1991). As part of a more comprehensive investiga- tion of acorn production in California oaks, we quantified acorn retention over a 10-year period by five species of oaks at Hastings Natural History Reservation, Monterey Coun- ty, central coastal California. Here we report on the degree of serotiny observed in these populations and explore correlations between seed retention and environmental factors with the goal of understanding the extent and potential significance of this phenomenon to California oak woodlands. MADRONO 152 [Vol. 61 Table 1. Taxa and Sample Sizes for Seed Retention at Hastings Reservation, 2002-2011 Seasons. Common name Scientific name Comments Elevational range (m) N trees, 2002-2005 N trees, 2006-2011 Valley oak Q. lobata White oak (sect. Quercus), deciduous, matures acorns in one year 476-875 16 20 Blue oak Q. douglasU White oak (sect. Quercus), deciduous, matures acorns in one year 538-847 15 24 Coast live oak Q. agrifolia Red oak (sect. Lobatae), evergreen, matures acorns in one year 484-875 17 17 California black oak Q. kelloggii Red oak (sect. Lobatae), deciduous, matures acorns in two years 800-850 — 7 Canyon live oak Q. chrysolepis Intermediate oak (sect. Protobalanus), evergreen, matures acorns in two years 800-875 8 Methods We quantified acorn production and retention on marked individuals of five species of oaks at Hastings Reservation at monthly intervals when they were present on trees between September 2002 and April 2012. Acorns of all five species generally mature in late September or October and typically fall off the trees in October and November. Acorns were counted using the visual survey method of Koenig et al. (1994a) and involved a subset of the trees whose patterns of acorn production are described and analyzed elsewhere (Koenig et al. 1994b; Koenig et al. 1996). Trees were individually marked and included valley {Quercus lobata Nee), blue {Q. douglasU Hook. & Arn.), and coast live {Q. agrifolia Nee) oaks between the 2002-2003 and 2005-2006 seasons. Starting in autumn 2006 the survey was expanded to include additional valley and blue oaks as well as a small number of canyon live {Q. chrysolepis Liebm.) and Califor- nia black {Q. kelloggii Newb.) oaks (Table 1). During initial surveys each September, two observers using binoculars counted as many acorns that they could during 15-second counts. Counts were added (N30) and //^-transformed (///(N30 + 1) = LN30) to reduce the correlation between the mean and the variance. Subsequent surveys were conducted similarly, except that only intact acorns still remaining on the trees were counted (i.e., empty caps still attached to a branch were ignored). Although values were In- transformed for analysis, we graph the untrans- formed data. In order to quantify the length of time acorns were retained on the trees, we determined, for each tree, the month in which its LN30 value first dropped below two (that is, < seven acorns were counted during the 30-second sampling period), an arbitrary cut-off corresponding to a significant number of acorns still remaining on the trees. Thus, trees for which < seven acorns were counted during the initial survey were excluded from analyses of the factors influencing serotiny, since acorn retention in such trees could not be meaningfully measured. Concordance across years among individuals was calculated using Kendall’s coefficient of concordance (Sokal and Rohlf 1981). We tested a total of eight variables (nine for valley oak) for their potential effects on the degree of serotiny using linear mixed models. On an annual basis, we looked at the initial size of the focal tree’s acorn crop (/^-transformed) (LN30), the mean ///-transformed acorn crop size of all trees of the target species surveyed, mean average daily temperature, and total rainfall, the last two variables measured from 1 November to 31 March, the main period during which acorns were retained by trees in years when serotiny occurred. On an individual tree basis, we tested for the effects of tree size, water availability, soil nitrogen availability, soil phosphorus availability, and spring leaf phenology during the prior spring. Spring leaf phenology was available for valley oaks only as part of a study of the relationship between this character and acorn production in this species. Weather was measured at Hastings Reserva- tion headquarters, within 3.5 km of all trees. Tree size was measured by the diameter at breast height (dbh). Water availability was estimated by predawn xylem water potential using a pressure bomb during September 1991 and 1994-1998 (Knops and Koenig 1994). Differences in xylem water potential among trees have been shown to be concordant among years (Knops and Koenig 2000); that is, despite annual variation in overall water availability, trees that are relatively wet because they have good access to ground water are relatively wet in both wet and dry years. Thus, we used xylem water potential values from 1991, when all trees were measured. Nitrogen (N) and phosphorus (P) availability was estimated using four ion-exchange resin bags placed under each tree (at a depth of 5-10 cm) between October 1992 and April 1993 (Knops and Koenig 2014] KOENIG ET AL.: SEROTINY IN CALIFORNIA OAKS 153 1997). Bags were then analyzed for NO3, NH4, and PO4 to estimate available nitrogen and phosphorus levels (soil nitrogen and soil phos- phorus). Spring leaf phenology of valley oaks was quantified by means of weekly surveys for budburst and flowering activity conducted be- tween 2003 and 201 1 (Koenig et al. 2012). For the analyses conducted here, we used the date on which budburst occurred (defined as >5% of the tree having leafed out and turned green). Analyses were conducted using the fime’ function in the R library ‘nlme’ (Pinheiro et al. 2013) in which the number of months at least seven acorns were counted on the tree during surveys (than is, for which LN30 > 2) was the dependent variable and ‘tree’ was included as a random factor. We compared a set of 11 (10 for coast live oak) candidate models as listed in Table 2. Most of these included the focal tree’s initial acorn crop along with one or two of the other variables; also included were the null model (intercept only), the full model, and a model with only the overall mean acorn crop of all conspe- cific trees counted each year (TV = 85 [valley oak]; N ^ 62 [coast live oak]). Models were ranked within species using the Akaike information criterion corrected for sam- ple size (AICc) with the ‘AICctab’ function in the R package ‘bbmle’ (Bolker 2012). We then model-averaged coefficients (mean ± standard error) from analyses of models with competing support (i.e., AAIQ < 10). All statistical analyses were conducted in R2.15.1 (R Development Core Team 2012). Results Interspecific and Inter-year Variability Seed retention differed considerably among the five species (Fig. 1). In general, acorns, which mature in October and November, had largely fallen or been removed by seed predators by December, one to two months later. The most notable exception to this was among coast live oaks, which frequently retained acorns two additional months until February and in some cases into (and occasionally beyond) March and April. Similar seed retention was also observed more rarely among valley oaks. During the 10 years of the study, we documented nine individuals (seven coast live oaks [11% of the individuals surveyed] in two different years and two valley oaks [2% of the individuals surveyed] in three different years) that achieved counts of seven acorns or more (LN30 > 2) in April, five to six months following normal acorn maturation. Two of these individuals (one of each species) retained acorns in their canopies into April in two different years (Table 3). Although systematic surveys were not conducted after April, in two cases coast live oaks surveyed in July yielded counts of at least seven acorns; both these were during the unusually large crop year of 2011- 2012. Trees that retained acorns were typically trees that produced very large acorn crops initially, but this was not always the case (Table 3). Overall, annual variation in the degree to which trees retained acorns was modest for valley oaks (Fig. 2a) and considerable for coast live oaks (Fig. 2b). There was significant concordance in seed retention across trees for both species (Kendall coefficient of concordance, valley oak: X^i5 ~ 47.2, P < 0.001; coast live oak: x^is = 43.2, P < 0.001). Virtually all acorns of the other three species had disappeared from trees by December of all years (Fig. 1). Because significant serotiny was absent in three of the species, we only performed mixed- effects models for valley oak and coast live oak. For valley oaks, four models had non-trivial support involving four variables: the initial size of the focal tree’s acorn crop, overall mean acorn crop size, mean winter temperature, and xylem water potential. Only the initial size of the focal tree’s acorn crop, however, had a model- averaged estimate whose 95% confidence inter- val did not overlap zero (Table 2). For coast live oak, two models involving three variables garnered non-trivial support, including the size of the focal tree’s acorn crop, the overall mean acorn crop size, and winter rainfall. Although the estimates of all three model-averaged variables had 95% confidence intervals that did not overlap zero, the standardized effect sizes indicated that the size of the focal tree’s acorn crop was by far the most important predictor of serotiny (Table 2). Thus, serotiny in both species was most strongly associated with the initial acorn crop size of the focal tree (Fig. 3). In addition, coast live oaks were more serotinous when the overall mean acorn crop of conspecifics was large and during dry winters. Discussion Acorns are a major food resource for Califor- nia wildlife, which at our study site include a range of species spanning birds such as band- tailed pigeons {Patagioenas fasciata [Say, 1823]) (Fry and Vaughn 1977), western scrub-jays {Aphelocoma calif ornica [Vigors, 1839]) (Carmen 2004; Koenig et al. 2009) and acorn woodpeckers {Melanerpesformicivorus [Swainson, 1827]) (Koenig and Mumme 1987; Koenig et al. 2008) to mam- mals including mice (Peromyscus spp. [Gloger, 1841]) (Merritt 1974), wood rats {Neotoma fuscipes [Baird, 1858]), ground squirrels {Spermo- philus beecheyi [Richardson, 1829]), mule deer {Odocoileus hemionus [Rafinesque, 1817]), and 154 MADRONO o ^ e ^ o §1 u o 3 a « 60 •S .s c« b ftOGO . hi 0) u ^ O cd u $1 'O ^■H S-S . fi J cd X ^ c+' « S ^ S w 43 *-> uo cd (U S’ 'o'H o cd h 'V ^ fi 'd ^ Q oj GO cd d ^ 'S •S > -o ^ ll .1 s 'S ^ 2 o S o > eO C4_i > (N z H . ^ O ^ ^ cd W O ^ >. tL, 'U o=i UJ *-i S o Q l! V. 60 II S'! •C o a Orl OO cd > cd 52 r;:3 cd O rd CO d cd g s^ .S M ^ & X ^ O &i D rH i-i X) H 'O .s c ^2 2 c ^ 2 S 3 2 ^ S w 2 -SI m d C-'-^OOOOOOOO o^qqoqqoooo dddddddddd V V V V V qqqqqqqq’-^q dNd'-^rnd-^Nr^d X I I I X I I I I X X X X X X X X X X X X X X xxxxxxxx ^^lm'^»r(vor^ooo^ 2 oj 3. ^ o '-='^^lfO•^sr)VO^'00O^^ [Vol. 61 2014] KOENIG ET AL.: SEROTINY IN CALIFORNIA OAKS 155 Fig. 1. Mean number of acorns counted during surveys of trees of the five species of oaks by month averaged across all years. feral pigs (Sus scrofa [Linnaeus, 1758]) (Pavlik et al. 1991). Although most of the mammals typically forage on acorns that have fallen from trees, many of the birds frequently, and in some cases primarily, forage on acorns present in the canopy. Such species would potentially benefit by having access to trees that hold onto their acorns through the winter and into the next spring, even if such serotiny was relatively rare. We report here on the finding that acorns of two species of California oaks are weakly serotinous, being retained in the canopy for six and rarely up to nine months post-maturity. Serotiny has not been previously recognized or quantified in California oaks, with the exception of data on two individuals of each of four species (including coast live oak, which were found to retain their acorns significantly longer than the other species) measured during a single year by Carmen (2004). Given that wildlife in our study typically begin eating immature acorns as early as August, this means that in some years acorns are present in the canopy of a small number of trees for virtually the entire year. Although the proportion Fig. 2. Mean number of acorns counted during surveys of trees by year for (a) valley oak and (b) coast live oak. Although valley oak generally lost their acorns by December, this was not the case in 2004, Coast live oak was the most variable, retaining a large fraction of its crop into the next spring in 201 1-2012 but otherwise generally dropping acorns by February. of trees with substantial number of acorns retained through the winter is small, such trees may potentially attract and benefit wildlife from a relatively wide area around individual trees. Thus acorns, although usually considered an ephemeral ‘pulsed’ resource (Ostfeld and Keesing 2000), are in some cases available to canopy- foraging California birds throughout the year, both through storage of mature acorns by some species but also through the retention of acorns on some trees. Serotiny was most prevalent among coast live oaks, but was found occasionally in valley oaks as well. The latter is particularly surprising given that valley oaks are deciduous. Indeed, tree 161, which exhibited an unprecedented degree of serotiny in 2004-2005 (Table 2), retained its acorns, but not its leaves, throughout this particular winter (W. Koenig personal observa- tion). We did not observe significant serotiny in Table 3. Cases of Trees Retaining a Substantial Number of Acorns (LN30 > 2) into April, Six Months after Normal Acorn Maturation. Species Tree number Crop year Acorns counted in September Acorns counted in April Valley oak 161 2004 156 93 Valley oak 161 2011 118 9 Valley oak 201 2007 108 9 Coast live oak 162 2011 59 22 Coast live oak 166 2011 64 20 Coast live oak 176 2011 62 10 Coast live oak 187 2011 42 12 Coast live oak 194 2011 24 7 Coast live oak 195 2006 85 23 Coast live oak 195 2011 83 67 Coast live oak 202 2011 132 37 156 MADROto [VoL 61 Jun _ • • • •A • • •A'' Apr * # A A A mm • A © -a • •• • •• • • • m • • c >, Feb _ m m mm m m--'' m A m _c - • wm m ••■■'•A# A«ni ? ffi Dec _ A /MIA A A m - A A ^ A A Oct _ • A - A" valley oak - Am A coast live oak 1 1 0 25 1 1 1 1 50 75 100 125 1 1 1 150 175 200 Initial acorn crop Fig. 3. The serotiny index (first month that seven or fewer acorns were counted) plotted again the initial size of the tree's acorn crop for valley and coast live oak. Only included are trees for which the initial number of acorns counted was at least seven; A = 1 84 (valley oak) and TV = 169 (coast live oak). In both cases, the acorns of trees with larger initial crops were present in the canopy for longer than trees with smaller acorn crops. any year among blue oak, canyon live oak, or California black oak. Among the species where serotiny was present we found that acorn retention was concordant among individuals across years; that is, trees that held onto their acorns relatively longer in one year did so in other years as well, compared to other trees in the population. To a large extent this is most likely a side effect of the strong positive relationship between seed retention and individual productivity, which varies consider- ably among individuals (Koenig et al. 1994b). Because of this relationship, trees that are consistently good acorn producers will inevitably exhibit greater serotiny than relatively unproduc- tive trees. What drives the differences in seed retention we observed? We suggest three hypotheses. First, interspecific differences correlated to some extent with differences in the water relationships of mature trees. Specifically, the two weakly serotinous species, coast live oak and valley oak, are drought-intolerant species that regularly tap into the water table, while blue oak and California black oak, which do not exhibit serotiny, are relatively drought-tolerant species whose root system often does not tap into easily available ground water (Knops and Koenig 1994). Countering this pattern is canyon live oak, another drought-intolerant species but one that did not retain acorns beyond maturity in November. Furthermore, it is unclear how or why drought tolerance of mature trees might be related to seed retention. Second, and perhaps more likely, is that interspecific differences in serotiny are related to differences in acorn morphology, which differ considerably among species and have important consequences for water loss and desiccation resistance of acorns (Xia et al. 2012). Snow (1991), for example, found that harvested coast live oak acorns desiccated faster than Engelmann oak acorns {Q. engelmannii), a result that he related to differences in distributions of these species at the Santa Rosa Plateau in Riverside County, California. Seed retention in coast live oaks might thus serve to reduce dehydration and extend the lifespan of acorns during dry winters, a hypothesis supported by the significant negative relationship between serotiny in this species and winter rainfall. Studies of the morphology and water relations of California acorns are needed in order to critically test this possibility. Third, acorn retention may facilitate dispersal by scatter-hoarding birds, which would have a longer time to harvest and disperse acorns before they fall to the ground and are consumed by rodents or deer. The bird species most likely to be involved is the western scrub-jay, which is well- known to harvest and store large numbers of acorns (Grinnell 1936; Carmen 2004; Koenig et al. 2009); the other major avian acorn harvester in this habitat, the acorn woopdecker, stores acorns in granaries (Koenig and Mumme 1987) and is a relatively inefficient acorn dispersal agent. There are no data to support the hypothesis that either of these species preferentially harvest or store either coast live oak or valley oak acorns (Koenig and Benedict 2002; Koenig et al. 2008), and thus this hypothesis cannot explain the interspecific differences in serotiny we observed. This hypothesis may, however, play an important role in selecting for serotiny in general among these species, since individuals trees could poten- tially gain considerable benefits by retaining acorns in the canopy beyond the time they are available from other trees, thereby enhancing the harvesting and dispersal of their seeds. Clearly the drivers of interspecific differences in seed retention in oaks deserve further study. Conclusion Intraspecifically, we found that serotiny in both coast live oak and valley oak is strongly related to a tree’s initial crop size. Beyond this, we detected no significant factor predicting serotiny in valley oak while among coast live oak serotiny was more pronounced during dry winters and when the overall coast live oak acorn crop was large. Thus, there appears to be no simple driver of differences in seed retention among years or within species in California oaks beyond a relationship in coast live oak with winter rainfall. Other potentially important factors remain to be examined; for example, it is possible that differences in the frequency or strength of winter storms (which is potentially correlated with winter rainfall) drives annual differences in 2014] KOENIG ET AL.: SEROTINY IN CALIFORNIA OAKS 157 serotiny. Similarly, the possibility that seed retention has coevolved with avian seed dispers- ers remains a strong possibility. There is much to be learned by further study of the causes and consequences of seed retention in California oaks, as well as other California taxa. Acknowledgments We thank the reviewers for their comments and Vince Voegeli and the Museum of Vertebrate Zoology, University of California, Berkeley, for logistical support. Numerous field assistants helped with the winter acorn surveys. Financial support for the project has come from the National Science Foundation (DEB-0816691 and IOS-09 18944) and the University of California’s Inte- grated Hardwood Range Management Program. Literature Cited Allen-Diaz, B., R. B. Standiford, and R. D. Jackson. 2007. Oak woodlands and forests. Pp. 313-338 in M. G. Barbour, T. Keeler-Wolf, and A. A. Schoen- herr, (eds.), Terrestrial vegetation of California, 3rd ed. University of California Press, Berkeley, CA. Bolker, B. 2012. bbmle: tools for general maximum likelihood estimation. R package version 1.0. 5. 2. Website http://cran.r-project.org/web/packages/ bbmle/index.html (accessed 1 September 2012). Bond, W. J. and B. W. Van Wilgen. 1996. Fire and plants. Chapman and Hall, London, U.K. Carmen, W. J. 2004. Noncooperative breeding in the California scrub-jay. Studies in Avian Biology 28:1-100. Cowling, R. M. and B. B. Lamont. 1985. Variation in serotiny of three Banksia species along a climatic gradient. Australian Journal of Ecology 10:345-350. Evanari, M., L. Shanan, and N. Tadmor. 1982. The Negev: the challenge of a desert, 2nd ed. Harvard University Press, Cambridge, MA. Fry, M. E. and C. E. Vaughn. 1977. Acorn selection by band-tailed pigeons. California Fish and Game 63:59-60. Griffin, J. R. and W. B. Critchfield. 1972. The distribution of forest trees in California. Research paper PSW-82. USDA Forest Service, Pacific Southwest Research Station, Berkeley, CA. Grinnell, j. 1936. Up-hill planters. Condor 38:80-82. Knops, j. M. H. and W. D. Koenig. 1994. Water use strategies of five sympatric species of Quercus in central coastal California. Madrono 41:290-301. . 1997. Site fertility and leaf nutrients of sympatric evergreen and deciduous species of Quercus in central coastal California. Plant Ecology 130:121-131. . 2000. Annual variation in xylem water potential in California oaks. Madrono 47:106-108. Koenig, W. D. and L. S. Benedict. 2002. Size, insect parasitism, and energetic value of acorns stored by acorn woodpeckers. Condor 104:539-547. , K. A. Funk, T. S. Kraft, W. J. Carmen, B. C. Barringer, and J. M. H. Knops. 2012. Stabilizing selection for within-season flowering phenology confirms pollen limitation in a wind- pollinated tree. Journal of Ecology 100:758-763. , J. M. H. Knops, W. J. Carmen, M. T. Stanback, and R. L. Mumme. 1994a. Estimating acorn crops using visual surveys. Canadian Journal of Forest Research 24:2105-2112. . 1996. Acorn production by oaks in central coastal California: influence of weather at three levels. Canadian Journal of Forest Research 26:1677-1683. , A. H. Krakauer, W. B. Monahan, J. Haydock, j. M. H. Knops, and W. J. Carmen. 2009. Mast-producing trees and the geographical ecology of western scrub-jays. Ecography 32:561-570. , J. P. McEntee, and E. L. Walters. 2008. Acorn harvesting by acorn woodpeckers: annual variation and comparison with genetic estimates. Evolutionary Ecology Research 10:811-822. AND R. L. Mumme. 1987. Population ecology of the cooperatively breeding acorn woodpecker. Princeton University Press, Princeton, NJ. , , W. J. Carmen, and M. T. Stanback. 1994b. Acorn production by oaks in central coastal California: variation within and among years. Ecology 75:99-109. Lamont, B. B. 1991. Canopy seed storage and release — what’s in a name? Oikos 60:266-268. , D. C. Le Maitre, R. M. Cowling, and N. J. Enright. 1991. Canopy seed storage in woody plants. Botanical Review 57:277-317. Merritt, J. F. 1974. Factors influencing the local distribution of Peromyscus californicus in northern California. Journal of Mammalogy 55:102-114. Nathan, R., U. N. Safriel, I. Noy-Meir, and G. Schiller. 1999. Seed release without fire in Pinus halepensis, a Mediterranean serotinous wind-dis- persed tree. Journal of Ecology 87:659-669. OSTFELD, R. S. AND F. Keesing. 2000. Pulsed resources and community dynamics of consumers in terrestrial ecosystems. Trends in Ecology and Evolution 15:232-237. Pavlik, B. M., P. C. Muick, S. Johnson, and M. Popper. 1991. Oaks of California. Cachuma Press, Los Olivos, CA. PiNHEiRO, J., D. Bates, S. DebRoy, D. Sarkar, and THE R Development Core Team. 2013. nlme: linear and nonlinear mixed effects models. R package version 3.1-109. Website http://cran.r- project.org/web/packages/nlme/index.html (ac- cessed 10 June 2012). R Development Core Team. 2012. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN3-90005 1-07-0, Website http://www. R-project.org (accessed 25 July 2012). SAnchez, E., C. Gallardo, M. A. Gaertner, A. Arribas, and M. Castro. 2004. Future climate extreme events in the Mediterranean simulated by a regional climate model: a first approach. Global and Planetary Change 44:163-180. Snow, G. E. 1991. Germination characteristics of Engelmann oak, and coast live oak from the Santa Rosa Plateau, Riverside County, California. Pp. 360-365 in R. B. Standiford (tech, coor.). Proceedings of the symposium on oak woodlands and hardwood rangeland management. USDA Pacific Southwest Research Station General Tech- nical Report PSW-126. SOKAL, R. R. and F. j. Rohlf. 1981. Biometry. W. H. Freeman, San Francisco, CA. Standiford, R. B. 2002. California’s oak woodlands. Pp. 280-303 in W. J. McShea and W. M. Healy, (eds.), Oak forest ecosystems. Johns Hopkins University Press, Baltimore, MD. 158 MADRONO [VoL 61 XiA, K., M. L Daws, W. Stuppy, Z.-K. Zhou, and H. W. Pritchard. 2012. Rates of water loss and uptake in recalcitrant fruits of Quercm species are determined by pericarp anatomy. PLoS One 7(10): e47368. doi:10.1371/joumaLpone.0047368 XoPLAKi, E., J. F. GonzAlez-Rouco, J. Luter^ BACKER, AND H. WANNER. 2004. Wet season Mediterranean precipitation variability: influence of large-scale dynamics and trends. Climate Dy- namics 23:63-78. Madrono, VoL 61, No. 2, pp. 159-176, 2014 NEW RECORDS FOR RARE AND UNDER-COLLECTED AQUATIC VASCULAR PLANTS OF YELLOWSTONE NATIONAL PARK C. Eric Hellquist Department of Biological Sciences, State University of New York Oswego, Oswego, NY 13216 eric.hellquist@oswego.edu C. Barre Hellquist Department of Biology, Massachusetts College of Liberal Arts, North Adams, MA 01247 Jennifer J. Whipple National Park Service, Yellowstone National Park, WY 82190 Abstract During 2008 and 2010, we conducted a floristic inventory of the aquatic vascular plants of Yellowstone National Park (YNP). Here, we report noteworthy records found during a survey of over 224 collecting locations within YNP. We documented 26 noteworthy species from 151 collections. These records represent rare, under-collected or newly located species in YNP, Montana, and Wyoming. New records include members of the Araceae (inch Lemnaceae), Haloragaceae, Hydrocharitaceae, Isoetaceae, Lentibulariaceae, Potamogetonaceae, and Typhaceae (incl. Spargania- ceae). During our fieldwork, we located two species new to Montana {Myriophyllum quitense and Potamogeton strictifoUus) and three species new to Wyoming {Isoetes echinospora, Najas flexilis, and Utricularia ochroleuca). We also found a single location for P. obtusifolius, a species known historically from Wyoming. In addition, we located P. zosteriformis, a species that had not been collected recently in Wyoming. We also found several species that were either new to the flora of YNP (Cailitriche heterophylla, N. flexilis, Potamogeton obtusifolius, P. zosteriformis) or species previously collected in YNP, but were misidentified in herbarium records {Lemna gibba, P. strictifoUus, Stuckenia vaginata, and U. ochroleuca). Our work also has greatly augmented the known collections of additional rare species in the Greater Yellowstone Ecosystem such as Lemna valdiviana, M. verticillatum, P. amplifolius, P. friesii, P. robbinsii, and U. minor. In addition, we found new sites for infrequently encountered or easily overlooked aquatic species including Ceratophyllum demersum (Ceratophylla- ceae), Elatine rubella (Elatinaceae), Elodea nuttallii (Hydrocharitaceae), Ruppia cirrhosa (Ruppiaceae), and Subularia aquatica (Brassicaceae). Key Words: Aquatic vascular plants, Myriophyllum, Potamogeton, Potamogeton foUosus subsp. fibrillosus, Utricularia, Yellowstone National Park. Yellowstone National Park (YNP) covers 8987 km^ and has been shaped by geological forces dominated by uplift of mountains, volca- nism, and glaciation (Marston and Anderson 1991). Much of the landscape of YNP is characterized by aquatic habitats. YNP contains the headwaters of the Missouri-Mississippi and Snake-Columbia watersheds (Marston and An- derson 1991). Approximately 5% (ca. 46,000 ha) of the surface area of YNP is covered by water (Varley and Schullery 1998). Much of this water is distributed across 150 lakes, with Yellowstone Lake being the largest high elevation lake in North America (Varley and Schullery 1998). In addition to the wealth of lakes in YNP, over 500 streams with a total length of over 4000 km drain the watersheds of YNP (Varley and Schullery 1998). Along with the abundant aquatic eco- systems of YNP, the park also has extensive wetlands. Many YNP wetlands are hydro- thermally influenced, thus creating a com- plex chemical environment that affects the abundance and distribution of wetland and aquatic vegetation. The wide variety of aquatic and semi-aquatic habitats distributed across the geologically complex landscape of YNP provides an ideal setting to examine patterns of aquatic plant distribution, abundance, and habitat char- acteristics. There are approximately 1150 native vascular plant species in Yellowstone National Park (Young 2007) and approximately 2082 in the Greater Yellowstone Ecosystem (GYE) (Evert 2010). Of these species, a proportionally small (ca. 9% YNP, ca. 5% GYE), but ecologically impor- tant number of these plants are aquatic. Hydro- phytes grow in or on water and may be anchored to inundated or saturated sediments for some period of the growing season (Tiner 2012). Hydrophyte biomass may be entirely submerged or may be a combination of submerged, floating, and emergent. Hydrophytes are classified as obligate indicators of aquatic or wetland condi- tions (Lichvar et al. 2012). 160 MADRONO [Vol. 61 Ecologically, hydrophytes are important food for waterfowl (Fassett 1957) and provide habitat for fish that is used for nesting, egg attachment, and refuge (Sculthorpe 1967; Mulligan 1969). In addition, aquatic plants oxygenate water, reduce currents, shade and cool littoral sediments, promote sedimentation and retention, add or- ganic matter to sediments, and provide habitat heterogeneity and substrate for aquatic verte- brates, invertebrates, and periphyton (Mulligan 1969). Despite their importance to aquatic ecosystems, aquatic plants are often under- represented in herbarium collections and floras. Yellowstone National Park has a long history of plant exploration (e.g.. Tweedy 1886; Rydberg 1900), but aquatic plants have been under- collected. Species collected in this study have probably existed for many years in YNP and are not new introductions. Many taxa occur in neighboring regions or states and thus YNP records are likely filling gaps within ranges that are more contiguous than previously realized. In addition, several are small species that are easily overlooked. Below, we provide comments on overlooked species or noteworthy additions to the aquatic vascular flora of YNP, the GYE, as well as Montana, Wyoming, and Idaho. Methods In 2008 and 2010, we conducted an inventory of the aquatic vascular flora of YNP to thoroughly document species occurrences, while also linking patterns in water chemistry to the abundance and distribution of aquatic vascular plants. The project also entailed use of the Yellowstone National Park Herbarium (YELLO) and the Rocky Mountain Herbarium (RM) to locate previous collections and to annotate specimens as needed. During early July to mid- August, we sampled YNP ponds, lakes, rivers, and wetlands focused specifically on aquatic plant diversity. Surveys took place via shoreline wading or by canoe. Over the two summers, we surveyed more than 224 YNP collection sites. Specimens for herbarium collections were col- lected by hand or by grapple. We collected all aquatic vascular species that we could locate at each site. Water chemistry parameters at sam- pling locations (pH, alkalinity, conductivity, nitrates, phosphorus, and trace metals) were measured and will be presented elsewhere (Hell- quist et ak, unpublished data). Unless otherwise noted, species determinations were made by C. B. Hellquist and C. E. Hellquist. Collections of the Lentibulariaceae were anno- tated by Garrett E. Crow (University of New Hampshire, Durham, NH) and Barry Rice (The Nature Conservancy, Davis, CA). James Hickey (University of Miami, Oxford, OH) annotated the Isoetaceae. Donald Les (University of Con- necticut; DNA analysis of Elodea Michx.) and Zdenek Kaplan and Judith Fehrer (Academy of Sciences of the Czech Republic; DNA analysis of Potamogeton L, and Stuckenia Bdrner) deter- mined problematic collections with DNA analy- sis. Collections confirmed with DNA fingerprint- ing are noted with “DNA” following the date in the collection citations. Nomenclature follows Haynes (2000a) for the Hydrocharitaceae, Taylor et al. (1993) for the Isoetaceae, Haynes and Hellquist (2000) for the Potamogetonaceae, Kaul (2000) for the Sparganiaceae, Landolt (2000) for the Lemnaceae, Les (1997) for the Ceratophylla- ceae, and Crow and Hellquist (2000) for the Brassicaceae, Elatinaceae, Haloragaceae, Lenti- bulariacae, and Plantaginaceae. Global and Subnational Ranks are cited from NatureServe (2012). State Heritage Ranks and species of concern recognition are cited from the Wyoming Natural Diversity Database (Heidel 2012). Results and Discussion We located a variety of noteworthy records for aquatic plants that were either new locations of state species of concern, new records for Wyo- ming or Montana, new records for YNP, under- collected species, or a combination of these categories. We documented 102 aquatic vascular plant species and hybrids, and collected over 2200 dried specimens of aquatic or other wetland plants (Hellquist et ak, unpublished data). These specimens are deposited in the Yellowstone National Park Herbarium (YELLO), Yellow- stone Heritage and Research Center, Gardiner, MT. Below, we summarize notable members of the YNP aquatic vascular flora that are either listed as species of concern by their respective states or are uncommon or under-collected members of the aquatic flora of YNP. Araceae (Including Lemnaceae) Lemna gibha L. Lemna gibha (swollen duckweed) is found in quiet, eutrophic habitats (Landolt 2000) and has distinctive inflated fronds that can be confirmed in cross section (Dorn 1984), Lemna gibha is found in Eurasia, South America, northern Mexico, and Africa. In North America, its range is most contiguous from Texas to California and Nevada. Scattered stations are located in Illinois, Nebraska, and the GYE (Landolt 2000). Dorn (1984, 2001) states that L. gibha is reported from Montana and Wyoming, respectively, but no location information is provided. Evert (2010) doubts that L. gibha is native to GYE. Tweedy (1886) noted that L. gibha is rare in YNP and cites its presence in ponds near the head of Broad Creek. In 1979, L. gibha was collected in a 2014] HELLQUIST ET AL.: UNCOMMON AQUATIC PLANTS OF YELLOWSTONE NP 161 marshy pond along the Specimen Ridge Trail between Soda Butte Creek and the Lamar River {T. Caprio 706 [YELLO]). In addition, Kesonie and Hartman (2011) report two collections from Grand Teton National Park and John D. Rock- efeller Parkway. We found L. gibba in two quiet ponds in the Mammoth-Gardiner vicinity of YNP. USA. MONTANA. YELLOWSTONE NA- TIONAL PARK. Park Co,: Ice Lake, “Landslide Lake,” at Gardiner Basin, first lake upstream along Landslide Creek, ca. 1.0 mile from road, elev. 5487 ft, 45°01.993'N, - 1 10°45.023^W, 9 July 2007, C E. Hellquist and C. B. Hellquist 16-08\ Sixth pond, the largest along Beaver Pond Trail Loop at the northeast corner of the loop, starting at the west end from Mammoth Hot Springs, elev. 6518 ft, 44°59.950'N, ~110°43.117'W, 26 July 2010, C E. Hellquist and C. B. Hellquist 830-10. Lemna valdiviana Philippi Lemna valdiviana (small duckweed) is a wide- spread duckweed found in South America, Central America, North America, and the West Indies (Landolt 2000). In western North Amer- ica, L. valdiviana is found in several isolated sites, with most locations concentrated along the Atlantic coast, the lower Mississippi Valley, and the Gulf of Mexico (Landolt 2000). Fronds of L. valdiviana tend to be somewhat elongate and have a midvein that ends just short of the frond apex at least three quarters from the node. Although not mapped by Landolt (2000) from the GYE it is recorded from Park and Teton Counties in Wyoming (Dorn 2001; Evert 2010). This species is of concern in Wyoming (G5/S1). Evert (2010, p. 401) describes the species as “rare” and typical of quiet waters including warm springs between 3800-8200 ft in elevation. In YNP, Lemna valdiviana has been collected in the Bechler District in the Robinson Creek thermal area from a warm spring discharge (/. J. Whipple 3488 [YELLO]). Our single collection was in a small thermally influenced seep on the west bank of the Firehole River near Ojo Caliente Spring. USA. WYOMING. YELLOWSTONE NA- TIONAL PARK. Teton Co.: Spring seep above Firehole River north of Ojo Caliente Spring, elev. 7203 ft, 44°33.927'N, - 1 10°50.654'W, 19 July 2010, C E. Hellquist and C. B. Hellquist 756-10. Brassicaceae Subularia aquatica L. Subularia L. (awlwort) is one of five genera of the mustard family that have aquatic or semi- aquatic species. There are only two species of Subularia worldwide, with S. aquatica the only North American species (Al-Shehbaz 2010). Awlwort is found entirely submerged in shallow waters or emergent in littoral zones of freshwater or saline wetlands (Al-Shehbaz 2010). Subularia is an annual with a basal rosette of tapered, quill- like leaves that can be inconspicuous in the field. Subularia is found widely distributed primarily in northern states and provinces in North America (Al-Shehbaz 2010). In the western United States, Subularia is found in Wyoming, Utah, Colorado, Idaho, Washington, and California (Al-Shehbaz 2010). In YNP, Subularia was found in a small pond at the outlet of Yellowstone Lake (Tweedy 1886) and from an 1873 collection at Lake (Rydberg 1900). Mulligan and Calder (1964) map a Subularia collection from Yellowstone Lake and from sites in Fremont and Sublette Counties, Wyoming. Evert (2010, p. 238) stated that Subularia is “rare, seldom seen or collected” in the region. Subularia has been collected in YNP from the Bechler District (“Boundary Lake #2”), which is probably located along the west boundary of YNP near Boundary Creek {B. Gresswell s.n., det. by D. Despain [YELLO 4467]). Additional YNP Subularia sites include Scaup Lake (/. J. Whipple 2946 [YELLO]) and Plover Point, Yellowstone Lake {J. J. Whipple 5753 [YELLO]). USA. WYOMING. YELLOWSTONE NA- TIONAL PARK. Park Co,: Pond on the west side of Grebe Lake Trail, ca. 1.25 miles north of Norris-Canyon Rd, elev. 8140 ft, 44°43.708'N, - 1 10°32.099'W, 30 July 2010, C. E. Hellquist and C. B. Hellquist 883-10. Teton Co.: Lewis River, north of Lewis Lake, downstream of fast water, elev. 7782 ft, 44°19.804'N, 1 10°38.749'W, 9 Aug 2009, C E. Hellquist and C. B. Hellquist 524-08; Mallard Lake east of Old Faithful, elev. 8054 ft, 44°28.555'N, - 1 10°46.480'W, 14 Aug 2010, C E. Hellquist, C. B. Hellquist, and X. Vu 1094-10; Marshy open water at the southwest corner of the Southeast Arm of Yellowstone Lake, north of Trail Creek, elev. 7732 ft, ca. 44°17.7074'N, -110M2.9316'W, 17 Aug 2010, C. E. Hellquist, C. B. Hellquist, and H. Anderson 1126-10. Ceratophyllaceae Ceratophyllum demersum L. Ceratophyllum demersum (Hornwort) is a floating species that lacks roots and rarely produces fiowers and fruit. Ceratophyllum L. is widely disjunct from eastern North America to Washington, Oregon, and interior British Co- lumbia (Les 1997). In YNP, C. demersum is found in the quiet waters of lakes and ponds primarily in the Northern Range. Ceratophyllum demersum is one of the most widespread aquatic plants in North America (Les 1997), but we found it in only six sites in YNP. Evert (2010, p. 267) notes 162 MADRONO [Vol. 61 that C demersum is found in a variety of habitats but “seldom seen or collected in our area”. Interestingly, Tweedy (1886) states that Cerato- phyllum is often seen in “great quantities” (p. 19) and is “frequent in sluggish streams and lake sloughs throughout up to 8500 ft alt” (p. 62), However, Rydberg (1900) only cites one collec- tion of Ceratophyllum from YNP at Broad Creek. Herbarium specimens of C. demersum have been collected at Lost Lake {H. S. Conard 1237 [YELLO]), Cygnet Lakes {B. Gresswel! s.n. [YELLO]), Fern Lake (USFWS s.n. [YELLO 5655]), eastern White Lake (USFWS s.n. [YELLO 5656]), and an unnamed lake asso- ciated with a tributary of Rescue Creek (Gr ess- well s.n. [YELLO 4603]). This collection (YELLO 4603) is a mixed collection with Myriophyllum verticillatum. USA. MONTANA. YELLOWSTONE NA- TIONAL PARK. Park Co.: Landslide Lake (“Ice Lake”), at Gardiner Basin, first lake upstream along Landslide Creek, ca. 1.0 mile from road, elev. 5487 ft, 45°01.993'N, -110°45.023'W, 9 July 2007, C. E. Hellquist and C. B. Hellquist 15a-08; Slide Lake, Gardiner Basin along Old Gardiner Rd, elev. 5723 ft, 45m30rN, -110°42.00rw, 9 July 2008, C. E. Hellquist and C. B. Hellquist 5-08; Lower Rainbow Lake at Gardiner Basin, south of Old Gardiner Rd, elev. 5876 ft, 45°0L444'N, -110°44.62rW, 29 Aug 2008, C E. Hellquist and C. B. Hellquist 293-08. WYOMING. YEL- LOWSTONE NATIONAL PARK. Park Co.: Lost Lake, west of Roosevelt Lodge. Deep lake fringed with Nuphar. elev. 7041 ft, 44°54.560'N, -110°25.665'W, 11 July 2008, C E. Hellquist and C. B. Hellquist 34-08; Blacktail Pond east of Mammoth, Blacktail Deer Plateau, Mammoth- Tower Rd, elev. 6643 ft, 44°57.264'N, -110°36.215'W, 25 July 2008, C E. Hellquist and C. B. Hellquist 224-08. Teton Co.: Old marina, Yellowstone Lake at Grant Village, elev. 7732 ft, ca. 44°23.607'N, - 1 10°33.234'W, 20 Aug 2010, C E Hellquist and M. T. Hellquist 1141-10. Elatinaceae Elatine rubella Rydb. Elatine L. (waterwort) was found in two sites in YNP, Elatine can grow interspersed with other small, submerged species in shallow water or emergent on exposed mud. By rooting at its nodes it can form small patches or mats. In YNP, we found Elatine in marshy, littoral habitats. Confusion regarding the taxonomy of E. americana (Pursh) Arn,, E. brachysperma A. Gray, E. rubella, E. triandra Schkuhr, and E. triandra var. brachysperma (A. Gray) Fassett is reflected in the literature. Elatine rubella is the oldest recognized name for this group of Elatine in North America (Holmgren et al. 2005). Sometimes referred to as E. triandra (Evert 2010), Dorn (2001) stated that E. rubella was the only species in Wyoming. Tweedy (1886) cited collections of E. americana and E. triandra from Yellowstone Lake. In 1900, Rydberg recognized what had previously been referred to as E. americana, as E. rubella. Elatine rubella was considered a new species found along Yellow- stone Lake in volcanic sands. Rydberg acknowl- edged that E. rubella was very similar to the previously described E. americana and that it may be closely related. Today, E. americana is recognized as having an eastern distribution with its westernmost stations in Missouri and Okla- homa (Crow and Hellquist 2000). Elatine rubella and E. brachysperma are both rare in the GYE (Evert 2010). Elatine brachy- sperma is known from a single 1933 collection in Hebgen Lake in the GYE (Evert 2010) but is not known from YNP or Wyoming. Elatine rubella and E. brachysperma can be distinguished based on the amount of seed pits on the surface of the seeds. Elatine rubella has seeds with 16-25 pits per row while E. brachysperma has seeds with 9- 15 pits per row (Crow and Hellquist 2000). USA. MONTANA. YELLOWSTONE NA- TIONAL PARK. Gallatin Co.: Divide Lake along east side of Gallatin Rd, north of West Yellowstone, elev. 7252 ft, 44°54.64UN, -111°03.061'W, 6 Aug 2008, C. E. Hellquist and C. B. Hellquist 447-08. WYOMING. YEL- LOWSTONE NATIONAL PARK. Teton Co.: Marshy open water at the southwest corner of the Southeast Arm of Yellowstone Lake, north of Trail Creek, elev. 7732 ft, ca. 44U7.7074'N, -110°12.9316'W, 17 Aug 2010, C E. Hellquist, C. B. Hellquist and H. Anderson 1124; Marshy open water at the southwest corner of the Southeast Arm of Yellowstone Lake, north of Trail Creek, elev. 7732 ft, ca. 44°17.7074'N, -110°12.9316'W, 17 Aug 2010, C E. Hellquist, C. B. Hellquist and H. Anderson 1125. Haloragaceae Myriophyllum quitense Kunth Myriophyllum quitense (Andean water-milfoil) initially appears similar to the widespread and common M. sibiricum Kom. It has bright green foliage that appears stiff and rough when held out of water. When fertile, M. quitense produces large floral bracts that easily distinguish it from M. sibiricum, which produces minute bracts (see key in Ceska et al. 1986). Myriophyllum quitense has an unusual disjunct geographic distribution with populations in South America and North America. In North America, M. quitense is found in Mexico, the western United States (Arizona, Montana, Oregon, Washington, and Wyoming) 2014] HELLQUIST ET AL.: UNCOMMON AQUATIC PLANTS OF YELLOWSTONE NP 163 and eastern Canada (Prince Edward Island, New Brunswick; Ceska et al. 1986; McAlpine et al. 2007; Lesica 2012). Myriophyllum qui tense has been considered both native (Ceska et al. 1986; McAlpine et al. 2007) and introduced (Aiken 1981; Couch and Nelson 1988) in North America. Myriophyllum quitense is a species of concern in Wyoming (G47/S1) and has been recorded in YNP since the late 1 800’s with the apparent first record collected in 1885 by F. Tweedy from Yellowstone Lake (Couch and Nelson 1988). Another early record was an 1899 collection by A. and E. Nelson {A. Nelson 6791 [RM]) collected from the '‘Madison River 3 miles from R.S.”. In Nelson’s notes, M. quitense was described as being abundant in slow moving water of rivers and bays. “R.S.” is likely the Riverside Station that was located 3-4 miles west of Seven Mile Bridge, but it is unclear in what direction from Riverside Station the collection was made. The site is likely in Wyoming, but it could also have been Montana. Records of M. quitense from the early 1900’s were collected by Jepson from the Firehole River and by M. E. Jones from Swan Lake (Couch and Nelson 1988). Myriophyllum quitense was collected in 1953 in the Madison River west of Madison Junction {Porter 6379 [RM]). In 1965, M. quitense was collected in the Firehole River in the Lower Geyser Basin {Hitchcock 23891 [RM]). In the 1990’s, additional collections of M. quitense were made in the Firehole River upstream of the Ojo Caliente Bridge (/. J. Whipple 4136 [YELLO]) and in Nez Perce Creek just upstream of the bridge on the Grand Loop Rd (/. J. Whipple 4200 [YELLO]). The Firehole River and Nez Perce sites were still thriving populations in 2010. We found M. quitense in the Falls, Firehole, Madison, Snake, and Lewis Rivers as well as Nez Perce and Fairy Creeks. The most extensive stands of M. quitense in YNP were in cold, shallow currents of the Firehole and Madison Rivers as well as Nez Perce Creek. Along the Firehole River and Nez Perce Creek, we found M. quitense in areas with hydrothermal influence. In a few instances, we found M. quitense in slow moving waters such as within Lewis River. The records of M. quitense in the Madison River just inside the YNP boundary, and just outside of YNP at Baker’s Hole, are the first records for the species in Montana (S. Mincemoyer, personal communication) . Rydberg (1900) noted that Myriophyllum spicatum L., an introduced species, was present in the Madison River in Montana based on 1886 collections by Tweedy. We have not seen these specimens, but based on the present occurrence of M. quitense and M. sibiricum in the Madison River, it is unlikely these collections were M. spicatum. Myriophyllum sibiricum is the most abundant water-milfoil in YNP and we believe the Tweedy collections are likely either M. sibiricum or M. quitense. Locating any aquatic invasive species was a focus of our work. We looked thoroughly for M. spicatum and we did not find it in YNP, although the range of M. spicatum is approaching the GYE. USA. MONTANA. YELLOWSTONE NA- TIONAL PARK. Gallatin Co.: Madison River north of West Yellowstone, east of USFS Baker’s Hole Campground along west boundary of the park, elev. 6591 ft, 44°42.21 UN, - 1 1 r05.719'W, 5 Aug 2008, C. E. Hellquist and C. B. Hellquist 434-08; Madison River on north side of West Yellowstone-Madison Jet Rd, just east of park entrance, elev. 6665 ft, 44°40.013'N, -lir04.075'W, 6 Aug 2008, C. E. Hellquist and C. B. Hellquist 457-08. WYOMING. YEL- LOWSTONE NATIONAL PARK. Park Co.: Washed ashore on north end of Yellowstone Lake, east of Pelican Creek and east of Fish- ing Bridge, elev. 7736 ft, 44°33.161'N, -110°21.108'W, 11 Aug 2010, C. E. Hellquist and C. B. Hellquist 1034-10; Yellowstone River between Buffalo Ford and Nez Perce Ford, north of Fishing Bridge, elev. 7708 ft, ca. 44°37.()93'N, -110°25.096'W, 12 Aug 2010, C. E. Hellquist and C. B. Hellquist 1058-10; Yellowstone River at “Grizzly Curve” Hayden Valley, Canyon- Lake Rd, elev. 7680 ft, ca. 44°39.256'N, ~-110°27.809'W, 12 Aug 2010, C E. Hellquist and C. B. Hellquist 1068-10; Yellowstone River at Alum Creek, Hayden Valley, Canyon-Lake Rd, elev. 7664 ft, ca. 44°40.552'N, - 1 10°29.193', 14 Aug 2010, C. E. Hellquist and C. B. Hellquist 1080-10. Teton Co.: Small backwater of Snake River at Snake River picnic area at trailhead at the South Boundary Ranger Station, elev. 6917 ft, 44°08.173'N, -110°39.933'W, 28 July 2008, C. E. Hellquist and C. B. Hellquist 261-08; Lewis River backwater on east side of main channel at confluence with Snake River, elev. 6871 ft, 44°08.173'N, -110°39.933^W, 28 July 2008, C E. Hellquist and C. B. Hellquist 277-08; Fire- hole River south of Old Faithful and Upper Geyser Basin, elev. 7433 ft, 44°27.388'N, ~110°49.257'W, 30 July 2008, C. E. Hellquist, C. B. Hellquist and J. J. Whipple 312-08; Firehole River, Upper Geyser Basin below Old Faithful Inn, south of bridge to Geyser Hill, elev. 7390 ft, ca. 44°27.708'N, -110°49.58rW, 30 July 2008, C E. Hellquist, C. B. Hellquist and J. J. Whipple 316-08; Firehole River north of Upper Geyser Basin from Artemisia Geyser to Biscuit Basin, elev. 7284 ft, 44°28.865'N, - 1 10°51.057'W, C. E. Hellquist, C. B. Hellquist and J. J. Whipple 326- 08; Nez Perce Creek, warmed by thermal activity, east of Madison-Old Faithful Rd, elev. 7146 ft, 44°34.377'N, - 1 10°49.106'W, 31 July 2008, C. E. Hellquist and C. B. Hellquist 333-08; Firehole River, warmed by thermal activity upstream, at Fountain Flats Drive, west of Ojo Caliente Hot 164 MADRONO [VoL 61 Springs along Madison-Old Faithful Rd; elev. 7199 ft, 44°34d46'N, ~ 1 10°50.065'W, 31 July 2008, C E. Heilquist and C B. Hellquist 345-08; Nez Perce Creek between Nez Perce Picnic Area and confluence with Firehole River, elev. 7191 ft, 44°34A82'N, - 1 10M9.906^W, 31 July 2008, C E. Hellquist and C. B. Hellquist 350-08; Firehole River at confluence with Nez Perce Creek just north of Nez Perce Picnic Area, elev. 7161 ft, 44°34.86rN, - 1 10°50.00rW, 1 Aug 2008, C E. Hellquist and C. B. Hellquist 358-08; Firehole River on east side of Ojo Caliente Bridge and Ojo Caliente Spring, Fountain Flats, elev. 7183 ft, 44°33.768'N, - 1 10°50312'W, 4 Aug 2008, C F. Hellquist and C. B. Hellquist 405-08; Firehole River south of Firehole Picnic Area along Madison Jct.-Old Faithful Rd, elev. 7159 ft, 44°35.625^N, - 1 10°49.974^W, 4 Aug 2008, C E. Hellquist and C. B. Hellquist 409-08; Madison River south of West Yellowstone-Madison Jet Rd at Harlequin Lake Trailhead, elev. 6810 ft, 44°38.290'N, - i 10°53306'W, 5 Aug 2008, C E. Hellquist and C. B. Hellquist 425-08; Lewis River, north of Lewis Lake, downstream of fast water, elev. 7782 ft, 44°19.804'N, - 1 10°38.749'W, 9 Aug 2009, C. E. Hellquist and C. B. Hellquist 526- 08; Lewis River ca. 0.5 mile south of Lewis Falls along Grant - South Entrance Rd, elev. 7759 ft, 44N5.586'N, - 1 10°38.382'W, 11 Aug 2009, C E. Hellquist and C. B. Hellquist 569-08; Madison River backwater, east of Seven Mile Bridge along Madison-West Yellowstone Rd, elev. 6770 ft, ca. 44°39.622^N, - 1 10°57.770'W, 14 Aug 2008, C E. Hellquist and C. B. Hellquist 594-08; Firehole River north of Ojo Caliente Springs, elev. 7203 ft, 44°33.927'N, - 1 10°50.654'W, 19 July 2010, C E. Hellquist and C. B. Hellquist 752-10; Fairy Creek at confluence with Firehole River near Ojo Caliente Springs, elev, 7204 ft, 44°33,755'N, -110°50.473'W, 19 July 2010, C E. Hellquist and C. B. Hellquist 754-10; Shore of Shoshone Lake near DeLacy Creek, fragments along shore, elev. 7791 ft, 44°24.587'N, - 1 10°41.7895'W, 24 July 2010, C. E. Hellquist and C. B. Hellquist 781- 10; West shore of Beula Lake at trails end, elev. 7411 ft, 44°09.2429'N, - 1 10°46.0304'W, 28 July 2010, C E. Hellquist and C. B. Hellquist 837-10; Falls River at South Boundary Trail ford, Bechler District, elev. 6225 ft, 44°08.2irN, -llim53r: 6 Aug 2010, C E. Hellquist and C. B, Hellquist 971-10; Southeast shore of Lewis Lake, south of public launch site, elev, 7752 ft, 44°16.786'N, - 1 10°37.903'W, 9 Aug 2010,. C E. Hellquist and C. B. Hellquist 1006-10. Myriophyiium verticiiiatum L. Of the three Myriophyiium species in YNP ( M. quitense, M. sibiricum, and M. verticiiiatum), M. verticiiiatum (whorl-leaf watermilfoil) is the least common. Myriophyiium verticiiiatum differs from M. quitense and M. sibiricum by producing larger numbers of leaflets on closely spaced internodes along the stem. When removed from the water, the leaves fall on each other giving the plant a soft, plush appearance. Later in the season, M. verticiiiatum forms pectinate bracts along its inflorescence and large club-shaped turions along the stem. We found M. verticiiiatum at seven sites throughout the park, typically in standing water. Myriophyiium verticiiiatum is a species of concern in Wyoming (G5/S1). In YNP, specimens have been collected from Robinson Creek in Idaho (/. J. Whipple 3468 [YELLO]), Harlequin Lake, {Gresswell s.n. 6223 [YELLO]), Little Robinson Lake {Gresswell s.n. 4616 [YELLO]), from a small lake on Mt. Everts {Gresswell s.n. 4603 [YELLO]), and from ponds east of Twin Buttes {E. F. Evert 39598 [RM]), Myriophyiium verticiiiatum was recorded from 'lakes, head of Broad Creek” in YNP (Tweedy 1886, p. 38). In addition, a specimen was collected at a small pond near Norris Junction {Porter 6376 [RM]). USA. MONTANA. YELLOWSTONE NA- TIONAL PARK. Park Co.: First pond along Beaver Pond Trail Loop, starting at the west end from Mammoth Hot Springs, elev. 6841 ft, 44°59.661'N, - 1 10°43.175'W, 26 July 2010, C E. Hellquist and C. B. Hellquist 805-10. WYO- MING. YELLOWSTONE NATIONAL PARK. Park Co.: Oxbow wetlands along Slough Creek near the junction of Slough Creek Campground and Tower - Lamar Rd, 44°55.597'N, -110°19.249'W, 13 July 2008, C E. Hellquist and C. B. Hellquist 53-08; Glen Creek, east of parking lot for Bunsen Peak Trail, Swan Lake Flat, elev. 7296 ft, 44°55.923'N, --110°43.596'W, 15 July 2008, C. E. Hellquist, C. B. Hellquist and S. Gunther 81-08, 13 Aug 2008, C E. Hellquist and C. B. Hellquist 586-08; Gibbon River west of Norris Picnic Area, elev. 7520 ft, 44°43.773'N, --110°41.365'W, 16 July 2008, C E. Hellquist and C. B. Hellquist 95-08; Small pond south of Swan Lake, Swan Lake Flats, elev. 7288 ft, 44°54.830'N, - 1 10°43.498'W, 22 July 2008, C E. Hellquist and C. B. Hellquist 161-08; Triglochin marsh on west side of Norris - Madison Rd at the south end of Elk Park, elev. 7526 ft, 44°42.956'N, - 1 10M3.572'W, 23 July 2008, C E. Hellquist and C. B. Hellquist 202-08; North Twin Lake along Mammoth - Norris Rd, elev. 7574 ft, 44°46.5irN, - 1 10°44.180'W, 7 Aug 2008, C E. Hellquist and C. B. Hellquist 464-08. Teton Co.: Harlequin Lake north of West Yellowstone-Madison Jet Rd, elev. 6896 ft, 44°38.67rN, - 1 10°53.575'W, 5 Aug 2008, C E Hellquist and C. B. Hellquist 417-08; Southwest- ern-most pond north of trail to Nez Perce Patrol cabin, north of Nez Perce Creek, Madison-Old Faithful Rd, pond mostly filled with grasses and sedges, elev. 7220 ft, 44°34.632'N, -110°48.828'W, 17 July 2010, C E. Hellquist 2014] HELLQUIST ET AL.: UNCOMMON AQUATIC PLANTS OF YELLOWSTONE NP 165 and C. B. Hellquist 704-10; Middle pond north of trail to Nez Perce Patrol cabin, north of Nez Perce Creek, Madison-Old Faithful Rd, pond mostly filled with grasses and sedges, elev. 7224 ft, 44°34.754'N, - 1 10°48.741 'W, 17 July 2010, C E. Hellquist and C. B. Hellquist 720-10; DeLacy Creek, at confluence with Shoshone Lake, elev. 7791 ft, 44°24.5878'N, - 1 10°4L7895'W, 24 July 2010, C. E. Hellquist and C. B. Hellquist 799-10; South end of large pond at jet of Bechler Meadows Trail and Boundary Creek Trail, Bechler District, elev. 6475 ft, 44° 1 0.033 'N, -111°0L958'W, 5 Aug 2010, C. E. Hellquist and C. B. Hellquist 964-10. Hydrocharitaceae Elodea nuttallii (Planch.) H. St. John Elodea nuttallii (waterweed) was found in sites in the central and southern regions of YNP associated with Yellowstone Lake as well as the Lewis, Snake, and Yellowstone Rivers. Previous- ly, only E. canadensis Michx. was recognized within YNP (Evert 2010). Morphologically, our collections resembled E. canadensis, however DNA fingerprinting revealed the collections to be E. nuttallii (D, Les, personal communication). Elodea nuttallii is more prevalent in eastern North America and widely scattered in the west. Elodea nuttallii was last collected in the GYE in 1949 in Hebgen Lake, Gallatin CO, MT (Evert 2010). Ongoing molecular work on Elodea by D. H. Les is calling into question the determination of Elodea species based solely on vegetative morphology. Fertile material is recommended for accurate identification of Elodea due to the vegetative plasticity of taxa (D. H. Les, personal communication) . USA. WYOMING. YELLOWSTONE NA- TIONAL PARK. Teton Co.: East shore of Snake River, backwater in meadow along South Boundary trail, adjacent to hot spring channel, elev. 6887 ft, 44°08.068'N, - 1 10°39.759'W, 28 July 2008, C E. Hellquist and C. B. Hellquist 265- 08; Narrow cut-off channel of the Snake River along the South Boundary Trail, at the South Boundary Station, elev. 6917 ft, 44°08.173'N, -1 10°39,759'W, 28 July 2008, C E. Hellquist and C. B. Hellquist 271-08; Lewis River backwater on east side of main channel at confluence with Snake River, elev. 6871 ft, 44°08.537'N, -110°39.933'W, 28 July 2008, C E. Hellquist and C. B. Hellquist 276-08; Lewis Channel, north of Lewis Lake, downstream of fast water, elev. 7782 ft, 44°19.804'N, - 1 10°38.749'W, C E. Hellquist and C. B. Hellquist 516-08; Embayment at northeast end of Lewis Lake, elev. 7825 ft, 44°19.013'N, -110°36.460'W, 12 Aug 2008, C E. Hellquist and C. B. Hellquist 557-08; East shore of Snake River, backwater in meadow along South Boundary trail, adjacent to hot spring channel (same as 265-08), elev. 6887 ft, 44°08.068'N, - 1 10°39.759'W, 28 July 2008, DNA, C. E. Hellquist and C. B. Hellquist 852- 10; Yellowstone Lake at Bridge Bay bridge, floating along the shore, elev. 7768 ft, N°44°3L894'N, - 1 10°26.178'9 Aug 2010, C E. Hellquist and C. B. Hellquist 999-10; Southeast shore of Lewis Lake, south of public launch site, elev. 7752 ft, 44°16.786'N, - 1 10°37.903'W, 9 Aug 2010, C. E. Hellquist and C. B. Hellquist 1009-10; Marshy open water at the southwest corner of the Southeast Arm of Yellowstone Lake, north of Trail Creek, elev. 7732 ft, ca. 44°17.707'N, -110°12.93'W, 17 Aug 2010, DNA, C E. Hellquist, C. B. Hellquist, and H. Anderson 1130-10; Old marina, Yellowstone Lake at Grant Village, elev. 7732 ft, ca. 44°23.607'N, -110°33.234'W, 18 Aug 2010, DNA, C E. Hellquist and M. T. Hellquist 1142-10; Little Thumb Creek backwater at Yellowstone Lake, ca. 1 .0 mile north of West Thumb Geyser Basin, elev. 7854 ft, 44°26.246'N, - 1 10°34.770'W, 22 Aug 2010, DNA, C. B. Hellquist and M. T. Hellquist 1148-10. Park Co.: Eastern shore of Yellowstone River north of Fishing Bridge, elev. 7739 ft, 44°34.00rN, - 1 10°22.652'W, 8 Aug 2010, C E. Hellquist, C. B. Hellquist and X. Vu 996-10; Washed ashore on north end of Yellow- stone Lake, east of Pelican Creek and east of Fishing Bridge, elev. 7736 ft, 44°33.161'N, -110°2L108'W, 11 Aug 2010, C E. Hellquist and C. B. Hellquist 1037-10; Yellowstone River between Buffalo Ford and Nez Perce Ford, north of Fishing Bridge, elev. 7708 ft, 44°37.093'N, -110°25.096'W, 12 Aug 2010, DNA, C E. Hellquist and C. B. Hellquist 1053-10; Yellow- stone River at “Grizzly Curve” Hayden Valley, Canyon-Lake Rd, elev. 7680 ft, 44°39.256'N, - 1 10°27.809'W, 12 Aug 2010, C E. Hellquist and C. B. Hellquist 1071-10. Najas jlexilis (Willd.) Rostk. & W. L. E. Schmidt Najas jlexilis (water-nymph) is an annual with subopposite, long tapered, serrulate leaves that appear whorled. Najas Jlexilis was found in Wine- gar Lake along the southern boundary of YNP. This site is the first record for Najas Jlexilis in YNP and in Wyoming. The species is found predom- inantly in eastern North America with a more or less contiguous range from Newfoundland south to New Jersey, Ohio, and Missouri and west to Minnesota (Haynes 2000a). In western North America, the range of N Jlexilis is interrupted with records from Saskatchewan, Alberta, British Columbia, Washington, Oregon, California, Ida- ho, Montana, and Utah (Haynes 2000). USA. WYOMING. YELLOWSTONE NA- TIONAL PARK. Teton Co.: Eastern end of Winegar Lake, south of South Boundary Trail, 166 MADRONO [VoL 61 Bechler District, elev. 6452 ft, 44°08.234'N, -110°57.483'W, 6 Aug 2010, C E. Hellquist and C. B. Hellquist 976-10. ISOETACEAE Isoetes echinospora Durieu Only Isoetes occidentalis L. F. Hend. and I. bolanderi Engelm. were previously known to occur in the GYE (Evert 2010) and Wyoming (Dorn 2001). Isoetes echinospora (spiny-spored quillwort) is a new species of quillwort for Wyoming. We found /. echinospora in one locality along the south boundary of YNP. Isoetes echinospora has spores with distinct spines that easily distinguish it from the two other species of Isoetes in YNP. This circumboreal quillwort is a species of eastern and northern affinity in North America (Taylor et al. 1993). Its range extends from Newfoundland to Alaska, across the northeastern United States and Great Lakes region. In the west, /. echinospora is found in the northern prairie provinces, south into Idaho, western Montana, and along the Pacific coast to northern California. The YNP locality is a regionally isolated location between the pan- handle of Idaho and additional isolated localities in northern Utah and central Colorado (Taylor et al. 1993). USA. WYOMING. YELLOWSTONE NA~ TIONAL PARK. Teton Co.: Eastern end of Winegar Lake, south of South Boundary Trail, Bechler District, elevation 6452 ft, 44°08.234'N, -110°57.483'W, 6 Aug 2010, C. E. and C. B. Hellquist 1154-10. Isoetes occidentalis L. F. Hend. Isoetes occidentalis (western quillwort) was found in five sites in YNP. It is a species of western affinity found from southern British Columbia south to central California and east into Washington, Idaho, Montana, Colorado, Utah, and northwest Wyoming (Taylor et al. 1993) and is a species of concern in Wyoming (G4G5/S1). Isoetes occidentalis was found in Yellowstone Lake and sites further south. Robust I. occidentalis with stout leaves were found washed up on the northern shoreline of Yellow- stone Lake at the inlet of Pelican Creek. Evert (2010) states that /. occidentalis is known only from YNP in Wyoming and that it is found at elevations between 7500-7800 ft, although we did find it in the Lewis River at 6871 ft. Isoetes bolanderi and /. occidentalis can be differentiated by leaf structure. Isoetes bolanderi has leaves that taper to an acute tip while I. occidentalis has leaves that taper gradually to the tip (Taylor et al. 1993). USA. WYOMING. YELLOWSTONE NA- TIONAL PARK. Park Co.: Washed ashore on north end of Yellowstone Lake, east of Pelican Creek and east of Fishing Bridge, elev. 7736 ft, 44°33.16rN, -110°21.108'W, 11 Aug 2010, C. E. Hellquist and C. B. Hellquist 1025-10. Teton Co.: Lewis River backwater on east side of main channel at confluence with Snake River, elev. 6871 ft, 44°08.537'N, - 1 10°39.822'W, 28 July 2008, C E. Hellquist and C. B. Hellquist 279-08; Lewis River ca. 0.5 mile south of Lewis Falls along Grant-South Entrance Rd, elev. 7759 ft, 44°15.586'N, - 1 10°38.382'W, 11 Aug 2008, C. E. Hellquist and C. B. Hellquist 566-08; East end of Winegar Lake, south of South Boundary Trail, Bechler District, elev. 6452 ft, 44°08.234'N, -110°57.483'W, 6 Aug 2010, C. E. Hellquist and C. B. Hellquist 979-10; Southeast shore of Lewis Lake, south of public launch site, elev. 7752 ft, 44°16.786'N, - 1 10°37.903'W, 9 Aug 2010, C. E. Hellquist and C. B. Hellquist 1008-10. Lentibulariaceae One result of our work in YNP has been the realization that the bladderworts of the region need substantial taxonomic attention. Based on our work and the critical assistance of B. Rice and G. Crow, the bladderworts in YNP consist of four species: U. intermedia Hayne, U. minor L., U. ochroleuca Hartm., and Utricularia vulgaris L. subsp. macrorhiza (LeConte) Clausen. Although we collected fertile material whenev- er possible, vegetative stands of Utricularia were abundant. Differences between the four YNP Utricularia species can be judged based on floral characteristics related to the pedicel and the relative sizes of the corolla lip and spur (Thor 1988). The vegetative features of the species are also variable, and potential differences are subtle. We found that small vegetative U. vulgaris subsp. macrorhiza was easily confused with U. minor. U. ochroleuca also was easily misidentified as U. minor. Thor (1988) states that bristles on the teeth of leaf segments as well as the relative spur size compared to the lower lip of the corolla distinguishes U. minor from the other three YNP Utricularia. In the absence of flowers, the angles of quadrifid glands in traps are critical for species determination (Thor 1988). Due to the intergrading floral and vegetative characteristics of YNP Utricularia, the best characteristics to determine U. ochroleuca, U. minor, and U. intermedia are the arrangement of internal glands within the bladders (G. E. Crow, personal communication) . Utricularia intermedia Hayne Utricularia intermedia (flatleaf bladderwort) is a species of concern in Wyoming (G5/S1), which 2014] HELLQUIST ET AL.: UNCOMMON AQUATIC PLANTS OF YELLOWSTONE NP 167 is quite similar in appearance to U. ochroieuca. The earliest Wyoming collection of U. intermedia was in 1977, a specimen annotated by G. Crow during research for the treatment of the Lenti- bulariaceae for the Flora of North America, collected in the Gros Ventre Range {Lichvar 1290 [RM]). Utricuiaria intermedia is known from the Elk National Refuge, near Jackson Hole, WY. Utricuiaria intermedia is the rarest bladderwork in YNP, It is known from a Sphagnum-^QdgQ dominated peatland at the south end of Beula Lake along the southern border of YNP {E. F. Evert 41455 [YELLO]). In 2010, we collected U ochroieuca from Beula Lake as well as U. intermedia in what we presume is the same boggy area along the Beula Lake outlet stream where Evert collected. The overlap in habitat preferenc- es of these species adds to the identification difficulties that cannot be readily discerned in the field (see U. ochroieuca below). Another specimen from the northwest corner of Beula Lake appears to be a mixed collection of U. intermedia and U. ochroieuca {J. M. Lemly and D. J. Cooper 1412 [T'ELLO]). USA. WYOMING. YELLOWSTONE NA- TIONAL PARK. Teton Co.: Shallow shoreline of stream at southwestern corner of Beula Lake, elev. 7411 ft, 44°09.2429'N, - 1 10°46.0304^W, 28 July 2010, C E. Hellquist and C. B. Heliquist 846- 10; Narrow stream (1 ft wide) at southwestern corner of Beula Lake at trails end, elev. 7411 ft, 44°09.2429'N, -110°46.0304'W, 28 July 2010, C E. Hellquist and C B. Hellquist 847U0. Utricuiaria minor L. Utricuiaria minor (lesser bladderwort) is listed as a Plant Species of Potential Concern (G5/S3). Previous records for U. minor are known from approximately 30 extant and two historic records in Wyoming (B. Heidel, personal communica- tion). Previous collections of U. minor in YNP include YELLO records from thermal wetlands west of the Grand Loop Road in the Old Faithful area {Anderson 13 IB) and from a pond on the floating mat of Robinson Lake {Anderson and Harpel 150B), We found that in the field, small, sterile, more gracile individuals of U. vulgaris were easily confused with U. minor. We also learned that sterile specimens of what often appeared to be U. minor were in fact U. ochroieuca when quadrifid glands were examined (G. E. Crow personal communication; see U. ochroieuca below). USA. IDAHO. YELLOWSTONE NATION- AL PARK. Fremont Co.: Robinson Lake, Bechler District, lake a peat mat surrounded by a band of water dominated by Nuphar. elev. 6560 ft, 44°09.930'N, -111 04.21 3' W, 5 Aug 2010, C E. Hellquist and C. B. Hellquist 941-10. WYOMING. YELLOWSTONE NATIONAL PARK. Park Co.: Warm stream on the east side of the Norris-Madison Jet. highway near culvert at the south end of Elk Park. elev. 7526 ft, 44°42.956'N, - 1 10°43.592'W, 23 July 2008, C E. Hellquist and C. B. Hellquist 204-08; “Second Quadrant Pond” due west of Swan Lake and power line, pond covered by sedges and grasses, elev. 7620 ft, 44°55.303'N, - 1 10°45.790'W, 20 July 2010, C. E. Hellquist and C B. Hellquist 762- 10; “Third Quadrant Pond”, due west of Swan Lake and power line, pond covered by sedges and grasses, elev. 7674 ft, 44°46.096'N, -110°46.096'W, 20 July 2010, C E. Hellquist and C. B. Hellquist 766-10; “Fourth Quadrant Pond” due west of Swan Lake and power line, only small area of standing water, elev. 7679 ft, 44°55.567^N, - 1 10°46.127'W, 20 July 2010, C E. Hellquist and C. B. Hellquist 769-10. Teton Co.: Embayment at northeastern end of Lewis Lake. Elev. 7825 ft, 44°19.013'N, 1 10°36.420'W, 11 Aug 2008. C E. Hellquist and C. B. Hellquist 604-08. Utricuiaria ochroieuca Hartm, Utricuiaria ochroieuca (yellowish-white biad- derwort) was documented from several new locations in this study, and is also known from several specimens annotated in 2013 by G. E. Crow. The collections from this study and recently annotated records add a new species to the flora of Wyoming and YNP. Utricuiaria ochroieuca is a circumboreal bladderwort with yellowish-white flowers. It is found in Canada from the Northwest Territories south to Manitoba and east to Nova Scotia (Taylor 1989). In the United States, this uncommon bladderwort has been reported from nine states including New York, Minnesota, Colorado, Alaska, Washington, Oregon, and California (NatureServe 2012). Taylor (1989) also includes Ohio in its range. The nearest record of U. ochroieuca to YNP is located in Colorado (Taylor 1989; USDA 2012). The YNP records fill a geographic gap in the distribution of U. ochroieuca in the western United States. Utricuiaria ochroieuca primarily occurs in shallow waters associated with wetlands such as bogs and marshes (Taylor 1989). In YNP, we typically found U. ochroieuca in quiet, sheltered water such as small gaps in peaty turf extending into open water and gently flowing streams. These are also the habitats where we found all other Utricuiaria species as well. However, almost all of our collection sites for U. ochroieuca were found in the vicinity of or in water connected to hydrothermal features such as in the Gibbon and Lower Geyser Basins. Water chemistry analyses may help distinguish chemical preferences of Utricuiaria species in YNP (Hellquist et ah, unpublished data). 168 MADRONO [Vol. 61 Based on the inconspicuous appearance of U. ochroleuca, it can be mistaken for the more common U. intermedia (Taylor 1989). In our experiences, U. ochroleuca can be easily mistaken for U. minor if the quadrifid glands in the bladders are not microscopically examined (see below). It seems likely that U. ochroleuca may be more widespread, but easily overlooked within its range. For example, two YNP records of U. ochroleuca were found during research for the treatment of the Lentibulariaceae for the Flora of North America (G. Crow, personal communica- tion). One specimen deposited at Rocky Moun- tain Herbarium (RM) was collected in 1991 northwest of Norris Jet. by E. F. Evert (21763), and originally identified as U. minor. A second RM specimen, collected in 1901 by M. E. Jones, was identified only as Utricularia with the location cited as “Yellowstone National Park” (G. Crow, personal communication). In addition, two 2012 collections by B. Heidel from the Upper Green River in Wyoming, also have been identified as U. ochroleuca by G. E. Crow. In 2013, G. E. Crow annotated additional YELLO specimens of U. ochroleuca that were either originally determined as U. minor or Utricularia sp. from the Gibbon River (Lemly and Cooper 196), Tangled Creek (J. J. Whipple 3112), Tuff Cliff fen (/. J. Whipple 4326), a geothermally influenced wetland adjacent to the Shoshone Geyser Basin (/. J. Whipple 6345), in creeks and thermal areas at Three River Junction (Whipple and Anderson 6768, 6778, 6788), and in a geothermally influenced wetland near Hot Lake ( Whipple, Fertig, and Welp 4950). Differentiating U. intermedia, U. minor, and U. ochroleuca can be problematic, especially in the field. Distinguishing characteristics of the species are provided by Taylor (1989) based on previous work by Gliick (1906). Characteristics observed with magnification that we found helpful to tentatively identify U. intermedia were acute leaf segment tips, curved setiform appendages above the trap mouth, and the traps almost always completely segregated from green foliage. These characters contrast with U. ochroleuca, which has more obtuse leaf segment tips, long, straight setiform appendages above the trap mouth, and the sparsely scattered traps among green leaves. However, the most convincing characteristics require microscopic examination of the arrange- ment, length, and arm angle of the internal quadrifid glands of bladders (Thor 1988; Taylor 1989, G. Crow, personal communication). A thorough assessment of the abundance of U. ochroleuca in the Rocky Mountain region is needed. USA. WYOMING. YELLOWSTONE NA- TIONAL PARK. Park Co.j Small thermal seep east of main channel of the Gibbon River west of Norris Picnic Area, elev. 7699 ft, 44°43.659'N, -110°41.336'W, 5 Aug 2008, C. E. Hellquist and C. B. Hellquist 9F08; Small muddy pool east of Gibbon River at Gibbon Meadows north of Artists Paint Pots, Norris-Madison Jet. Rd, elev. 7376 ft, ca. 44°42.138'N, - 1 10°44.84FW, 1 Aug 2008, C. E. Hellquist and C. B. Hellquist 370-08; North Twin Lake along Mammoth- Norris Rd, elev. 7574 ft, 44°46.57rN, -110M4.180'W, 7 Aug 2008, C. E. Hellquist and C. B. Hellquist 465-08. Teton Co.: Fire- hole River south of Old Faithful and Upper Geyser Basin, elev. 7433 ft, 44°27.388'N, -110°49.257'W, 30 July 2008, C. E. Hellquist, C. B. Hellquist and J. Whipple 307-08'; Gibbon River upstream from Gibbon Falls north of Beryl Spring along Norris-Madison highway, elev. 7147 ft, 44°40.922'N, -110°44.657'W, 1 Aug 2008, C E. and C. B. Hellquist 390-08’, Embay- ment at northeast end of Lewis Lake, elev. 7825 ft, 44°19.013'N, - 1 10°36.460'W, 11 Aug 2008, C. E. Hellquist and C. B. Hellquist 560-08; Thermally-warmed stream flowing out of Senti- nel Meadows at intersection of Sentinel Meadows Trail, elev. 7242 ft, 44°33.804'N, - 1 10°50.974'W, 19 July 2010, C. E. Hellquist and C. B. Hellquist 741-10; Thermal seep west of Flat Cone Spring, Sentinel Meadows, elev. 7240 ft, 44°33.857'N, -110°51.692'W, 19 July 2010, C. E. Hellquist and C. B. Hellquist 742-10; Pond immediately along- side Norris-Mammoth Rd, north of Norris and south of Nymph Lake, pond thermally influ- enced, elev. 7540 ft, 44°45.027'N, -110°43.000, 29 July 2010, C. E. Hellquist and C. B. Hellquist 862-10; North shore of Nymph Lake, Norris- Mammoth Rd, north of Norris, lake thermally influenced, elev. 7492 ft, 44°45.0175'N, -110°43.590'W. 29 July 2010, C. E. Hellquist and C. B. Hellquist 868-10; South shore of Lewis Lake at outlet, in thermal seep, elev. 7752 ft, 44°16.595'N, - 1 10°38.164'W, 9 Aug 2010, C E. Hellquist and C. B. Hellquist 1003-10, 1014-10. Plantaginaceae Callitriche heterophylla Pursh Callitriche heterophylla (water-starwort) was found at two sites within YNP. These two records were the first for YNP. Evert (2010) cites an additional record from Teton Co, WY. In Wyoming, C. heterophylla is widely distributed, but is the least common species of Callitriche (Dorn 2001). Callitriche heterophylla could be mistaken for C. verna L. (syn. C. palustris L.). Dimensions of the fruit are diagnostic to distinguish the two species. In C. heterophylla, the fruit is as long as wide while in C. verna the fruit is longer than wide (Crow and Hellquist 2000). USA. WYOMING. YELLOWSTONE NA- TIONAL PARK. Park Co.: Dried pond above 2014] HELLQUIST ET AL.: UNCOMMON AQUATIC PLANTS OF YELLOWSTONE NP 169 Crystal Creek, springs to the southeast, dried buffalo wallow, elev. 6839 ft, 44°53.946'N, -110°17.886'W, 14 July 2010, C E, Hellquist and C. B. Hellquist 680-10', Yellowstone River between Buffalo Ford and Nez Perce Ford, north of Fishing Bridge, elev. 7708 ft, 44°37.093'N, -110°25.096'W, 12 Aug 2010, C E, Hellquist and C. B. Hellquist 1051-10. POTAMOGETONACEAE Potamogeton amplifolius Tuckerman Although abundant in eastern and midwestern North America, Potamogeton amplifolius (big- leaved pondweed) populations are scattered in western North America (Haynes and Hellquist 2000). Potamogeton amplifolius is a striking species with large, long-petiolate (up to 23 cm) floating leaves and broad, arcuate submersed leaves (up to 13 cm long) that have stout, persistent stipules (up to 12 cm; Haynes and Hellquist 2000). We found four YNP locations for P. amplifo- lius in deep water of lakes. Evert (2010, p. 546) states that the species is rare in the GYE and is found in “a few” YNP lakes in addition to Fremont County and Park County, WY. Evert (2010) does not note where in YNP P. amplifolius is found, and no records from YNP prior to our inventory are deposited in YELLO or RM. In Wyoming, P. amplifolius is a Wyoming species of concern (G5/S1) with three extant and two historic locations. USA. WYOMING. YELLOWSTONE NA- TIONAL PARK. Teton Co.: Goose Lake along east side of Fairy Falls Trail between Fountain Flats and Midway Geyser Basin, elev. 7224 ft, 44°32.540'N, ~ 1 10°50.624'W, 4 Aug 2008, C E. Hellquist and C. B. Hellquist 400-08', Lewis River north of Lewis Lake, downstream of fast water, elev. 7782 ft, 44°19.804^N, - 1 10°38.749'W, 9 Aug 2008, C. E. Hellquist and C. B. Hellquist 509- 08', Lewis Lake in ca. 6 ft. of water along northeast shore, 44°18.5irN, - 1 10°37.179'W. 9 Aug 2008, C. E. Hellquist and C. B. Hellquist 603-08', Shore of Shoshone Lake near DeLacy Creek, fragment along shore, elev. 7791 ft, 44°24.5878'N, - 1 10°4L7895'W, 24 July 2010, C. E. Hellquist and C. B. Hellquist 780-10', Floating on southeast shore of Lewis Lake, south of public launch site, elev. 7752 ft, 44°16.786'N, -110°37.903'W, 9 Aug 2010, C E. Hellquist and C B. Hellquist 1010-10. Potamogeton folio sus subsp. fibrillosus (Fernald) R. R. Haynes and Hellq. Potamogeton foliosus subsp. fibrillosus (fibrous pondweed) has been long reported from YNP (Haynes and Reveal 1973) and was a taxon of significant interest for our survey. The range of Potamogeton foliosus subsp. fibrillosus includes California, Washington, Idaho, and Utah (Hell- quist et al. 2012). Potamogeton foliosus subsp. fibrillosus is often reported from waters influ- enced by geothermal activity, but it is not exclusively found in geothermal hydrology (Haynes and Reveal 1973). This perplexing western North American taxon has been recognized at a variety of levels. Initially recognized as P. fibrillosus Fern. (Fernald 1932), its classification was later revised to P. foliosus var. fibrillosus (Fern.) Haynes and Reveal (Haynes and Reveal 1973). More recently, it has been classified as P. foliosus subsp. fibrillosus (Haynes and Hellquist 2002; Hellquist et al. 2012). Despite a focused effort to find P. foliosus subsp. fibrillosus in YNP, we were unable to find any conclusive populations of this taxon. In YNP, we collected what we believed to be P. foliosus subsp. fibrillosus at several sites, all influenced by hydrothermal conditions (sites in Fountain Flats, as well as along the Firehole River and Nez Perce Creek). To confirm our initial determinations, we sent five YNP collec- tions we believed were P. foliosus subsp. fibrillo- sus to Z. Kaplan and J. Fehrer for genetic analysis (Hellquist et al., unpublished data). DNA fingerprinting of our morphologically variable specimens that we suspected were P. foliosus subsp. fibrillosus revealed some of our collections were P. pusillus, P. foliosus, or a hybrid between P. pusillus and P. foliosus (Z. Kaplan, personal communication). Our inability to locate this taxon, in addition to the mixture of collections we thought were P. foliosus subsp. fibrillosus prior to genetic analysis was surprising. We believe that the morphological diversity within P. foliosus subsp. fibrillosus (e.g., Haynes and Reveal 1973) may be the result of specimens of this taxon being of hybrid origin. Our fieldwork in YNP and our experiences with hybrid Potamogeton make us increasingly skep- tical of the taxonomic validity of P. foliosus subsp. fibrillosus. Potamogeton friesii Rupr. Potamogeton friesii (Fries’ pondweed) is a species restricted to highly alkaline waters throughout its range (Hellquist 1980). Potamoge- ton friesii is one of the narrow-leaved pusilloid species that can be difficult to identify. It is often confused with the more common P. pusillus L. or the equally uncommon P. strictifoUus. Potamo- geton strictifolius has a bold vein-like margin on its leaves that is absent P. friesii. The leaf tips of P. friesii are mucronate, while P. strictifolius usually has an acute leaf-tip. Both species have fibrous stipules that are obvious when plants are 170 MADRONO [VoL 61 dried. Potam.ogeton friesii is most easily distin- guished by its distinct turions that consist of leaves organized in a firm, fan-shaped arrange- ment with the inner leaves at right angles to the outer leaves. In YNP, P. friesii was most often found in small, alkaline ponds in the northern range. This species is uncommon in YNP and Wyoming (G4/ SI). In Wyoming, P. friesii is presently known from four extant and three historic sites (B. Heidel, personal communication) as well as two sites within YNP (Swan Lake, [s.m USFWS, YELLO 5653]; eastern portion of Tern Lake, [5.«. USFWS, YELLO 5654]). In 2008 and 2010, eleven unrecorded YNP populations were locat- ed, six in the Gardiner Basin vicinity of Montana. USA. MONTANA. YELLOWSTONE NA- TIONAL PARK. Park Co.: Slide Lake, Gardiner Basin along Old Gardiner Rd, elev. 5723 ft, 45°00.30rN, -110°42.00rW, 9 July 2008, C E. Hellquist and C. B. Hellquist 10-08; Landslide Lake (‘Tee Lake”) at Gardiner Basin, first lake upstream along Landslide Creek ca. 1.0 mile from Stevens Creek Rd, elev. 5487 ft, 45°01.993'N, -110°45.023'W. 9 July 2008, C E. Hellquist and C. B. Hellquist 13-08; Lower Rainbow Lake at Gardiner Basin, south of Old Gardiner Rd, elev. 5876 ft, 45°0L444'N, -110°44.621'W, 29 Aug 2008, C E. Hellquist and C. B. Hellquist 290-08; Fourth pond off Beaver Pond Trail Loop to the south- west, starting at the west end from Mammoth Hot Springs, elev. 6519 ft, 44°59.802'N, -110°43.418'W, 26 July 2010, C E. Hellquist and C. B. Hellquist 819-10; Sixth pond, the largest along Beaver Pond Trail Loop at the northeast corner of the loop, starting at the west end from Mammoth Hot Springs, elev. 6518 ft, 44°59.950'N, - 1 10°43.1 17'W, 26 July 2010, C E, Hellquist and C. B. Hellquist 824-10; Small pond adjacent and below old beaver dam at sixth pond, the largest along Beaver Pond Trail Loop at the northeast corner of the loop, starting at the west end from Mammoth Hot Springs, elev. 6518 ft, 44°59.950'N, - 1 10°43.1 17'W, 26 July 2010, C E. Hellquist and C. B. Hellquist 825-10, WYOMING. YELLOWSTONE NATIONAL PARK. Park Co.: Buck Lake northeast of Trout Lake, Lamar Valley, elev. 6948 ft, 44°54.269'N, - 1 10°07.600^W, 10 July 2008, C E, Hellquist and C. B. Hellquist 28-08; Swan Lake, Swan Lake Flat along west side of highway, elev. 7288 ft, 44°55.168'N, - 1 10°44.124'W, 15 July 2008, C E. Hellquist, C. B. Hellquist, and S. Gunther 86-08, 23 July 2008, C E, Hellquist and C B, Hellquist 1 73-08, Blacktail Pond east of Mammoth, Black- tail Deer Plateau, Mammoth-Tower highway, elev. 6643 ft, 44°57.264'N, - 1 10°36.215'W, 25 July 2008, C E. Hellquist and C. B. Hellquist 221- 08; Floating Island Lake on south side of Mammoth-Tower highway, elev. 6600 ft, 44°56.493'N, - 1 10°27.064'W, 25 July 2008, C E, Hellquist and C. B. Hellquist 230-08; “Foster Lake” west of Soda Butte, north of Lamar River trailhead along the Tower-Northeast Rd, elev. 6723 ft, 44°52.344'N, -110°10.076'W, 14 Aug 2008, C E. Hellquist and C. B. Hellquist 580-08. Teton Co.: marshy open water at the southwest corner of the Southeast Arm of Yellowstone Lake, at mouth of Trail Creek, elev. 7732 ft, ca. 44°17.7074'N, - 1 10°12.9316'W, 17 Aug 2010, C E. Hellquist, C. B. Hellquist and H. Anderson, 1117-10. Potamogeton obtusifolius Mert. & W. D. J. Koch Potamogeton obtusifolius (bluntleaf pondweed) is a new species for the YNP flora and was listed as a historical species for Wyoming prior to our work (Heidel 2012). We found P. obtusifolius at single location (Cascade Lake) near Canyon Junction. This species is noted for having especially wide leaves (5-7 mm) for a linear- leaved pondweed (Haynes and Hellquist 2000). The leaves often appear two-ranked, are flaccid, and have distinct, broadly rounded leaf tips (Haynes and Hellquist 2000). The species can reproduce prolifically from winter buds. Potamo- geton obtusifolius is found primarily in northeast- ern North America and in the northern Great Lakes region. Isolated occurrences are found in western North America in far northern Canada, the Prairie Provinces, Alaska, British Columbia, Washington, Montana, and Wyoming (Haynes and Hellquist 2000). USA. WYOMING. YELLOWSTONE NA- TIONAL PARK. Park Co.: Cascade Lake, northeast of Canyon Jet., elev. 8059 ft, 44°44.996'N, - 1 10°3L514'W, 2 Aug 2010, C E. Hellquist and C. B. Hellquist 895-10. Potamogeton praelongus Wulfen Potamogeton praelongus (white-stem pond- weed) is similar to the common P. richardsonii (Ar. Benn.) Rydb., but has a distinct white zigzag stem and leaves that are less perfoliate than P. richardsonii. Potamogeton praelongus typically has keeled leaf tips that split when pressed (Crow and Hellquist 2000). This species, like P. ampli- folius, is characteristic of deep waters of lakes and is one of the earliest species to flower and fruit during the growing season. Potamogeton prae- longus is a species of concern in Wyoming (G5/ S1S2) and is known from eleven Wyoming sites, four within YNP (B. Heidel, personal communi- cation). One site is from a pond near the Mammoth Hot Springs hotel that no longer exists. Other known YNP sites include a pond near Mt. Everts {Gresswell s.n. [YELLO]) and Heart Lake (/. J. Whipple 6550 [YELLO]). Our 2014] HELLQUIST ET AL.: UNCOMMON AQUATIC PLANTS OF YELLOWSTONE NP 171 YNP collections include two sites from Montana and six from Wyoming. Some of our collections of P. praelongus were somewhat puzzling morphologically. Although morphologically consistent in most characters, our YNP collections did not always show the usually reliable trait of keeled leaf tips in our collections. We suspect that some of our collec- tions may be hybrids. The genetic composition of these YNP collections is currently being investi- gated (Hellquist et al., unpublished data). USA. MONTANA. YELLOWSTONE NA- TIONAL PARK. Park Co.: Slide Lake, Gardiner Basin along Old Gardiner Rd, elev. 5723 ft, 45m30rN, -110°42.00rw, 9 July 2008, C. E. Hellquist and C. B. Hellquist 6-08; Upper Rainbow Lake at Gardiner Basin, south of Old Gardiner Rd, elev. 5883 ft, 45°01.338'N, ^110°44.463'W, 29 July 2008, C E. Hellquist and C. B, Hellquist 296-08. WYOMING. YEL- LOWSTONE NATIONAL PARK. Park Co.: Fragment floating on shore of Grebe Lake north of Norris-Canyon Rd, elev. 8023 ft, 44°45.0373'N, - 1 10°33.2234'W, 30 July 2010, C. E. Hellquist and C. B. Hellquist 870-10; Floating along shore of Cascade Lake, northeast of Canyon Jet., elev. 8059 ft, 44°44.996'N, -110°3L514'W, 2 Aug 2010, C E. Hellquist and C. B. Hellquist 907-10. Teton Co.: North end of Riddle Lake, elev. 7913 ft, 44°26.713'N, -110°33.026'W, 27 July 2008, C E. Hellquist and C. B. Hellquist 245-08; Bridge Bay, Yellow- stone Lake, floating on surface, elev. 7769 ft, 44°3i.884'N, - 1 10°26.178^W, 10 Aug 2008, C E. Hellquist and C. B. Hellquist 530-08; West shore of Beula Lake at trails end, fragment floating along shore, elev. 741 1 ft, 44°09.2429'N, - 1 10°46.0304'W, 28 July 2010, C E. Hellquist and C. B. Hellquist 835-10; Deep water at the southwest corner of the Southeast Arm of Yellowstone Lake, north of Trail Creek, elev. 7732 ft, ca. 44°17.7074'N, -110°12.9316'W, 17 Aug 2010, C. E. Hellquist, C. B. Hellquist and H. Anderson 1118-10. Potamogeton robbinsii Oakes Potamogeton robbinsii (Robbins’ pondweed) is a rare plant in YNP and Wyoming (G5/S1). Prior to 2008, only three records were documented for P. robbinsii in YNP. Previously P. robbinsii was collected from Winegar Lake {R. Gresswell s.n. [YELLO]), Heart Lake (C Rich- ardson s.n. [GH, RM]), and from the shore of Heart Lake (H Anderson et al. 038 [YELLO 12070]). In 2010, we found new locations in Grebe Lake, Winegar Lake, and several sites around Yellowstone Lake. Potamogeton robbinsii, like P. obtusifolius, has the most contiguous part of its range in the northeast into the Great Lakes region with disjunct populations in the far north, the Rocky Mountains, and in the Pacific states and provinces (Haynes and Hellquist 2000). Potamogeton robbin- sii is easily recognizable with its stiff, two-ranked, sessile leaves that are serrulate with auriculate bases. Potamogeton robbinsii rarely blooms. It is unusual as the only pondweed with branched inflorescences (Haynes and Hellquist 2000). Al- though the species does not have true turions, P. robbinsii easily breaks apart into fragments that produce prolific adventitious roots. These frag- ments can be found washing on shore, lying flat on submerged sediments, and can contribute to the formation of dense stands in deep water. USA. WYOMING. YELLOWSTONE NA- TIONAL PARK. Park Co.: Grebe Lake north of Norris-Canyon Rd, elev. 8023 ft, 44°45.0373'N, -110°33.2234'W, 30 July 2010, C. E. Hellquist and C. B. Hellquist 869-10. Teton Co.: Eastern end of Winegar Lake, south of South Boundary Trail, Bechler District, elev. 6452 ft, 44°08.234'N, -110°57.483'W, 6 Aug 2010, C E. Hellquist and C. B. Hellquist 975-10; Yellowstone Lake at Bridge Bay bridge, floating along shore, elev. 7768 ft, 44°31.894'N, - 1 10°26.178'W, 9 Aug 2010, C. E. Hellquist and C. B. Hellquist 1000-10; Floating in marshy open water at the southwest corner of the Southeast Arm of Yellowstone Lake, north of Trail Creek, elev. 7732 ft, ca. 44°17.7074'N, - 1 10n2.9316'W, 17 Aug 2010, C. E. Hellquist, C. B. Hellquist, and H. Anderson 1129-10; Washed ashore on north end of Yellow- stone Lake, east of Pelican Creek and east of Fishing Bridge, elev. 7736 ft, 44°33.16UN, -110°2L108'W, 11 Aug 2010, C E. Hellquist and C. B. Hellquist 1035-10, 1036-10; Old marina, Yellowstone Lake at Grant Village, elev. 7732 ft, ca. 44°23.607'N, - 1 10°33.234'W, 20 Aug 2010, C. E. Hellquist and M. T. Hellquist 1139-10. Potamogeton strictifolius A. Benn. We found Potamogeton strictifolius (narrow- leaf pondweed) in four small alkaline ponds in the far northern portions of YNP within Mon- tana. Like P. friesii, P. strictifolius is usually found with species associated with alkaline water (Hellquist 1980). There appear to be no verified records of this species from Montana, even though it is thought to occur in the state (S. Mincemoyer, personal communication). Potamo- geton strictifolius is a species of concern in Wyoming (G5/S1?). It is known from several sites in the state (B. Heidel, personal communi- cation) including a specimen from the Firehole River (Porter 6385; RM). Two collections in YELLO {Gresswell 4601, 4740) from a pond on Mt. Everts and Big Trumpeter Lake are also P. strictifolius. Previously, P. strictifolius was found in the Yellowstone River (/. J. Whipple and M. Hektner 6593 [YELLO]) and Madison Rivers. We could not relocate P. strictifolius at the 172 MADRONO [VoL 61 Yellowstone River location. Potamogeton stricti- folius was cited from the upper reaches of the Madison River (Wright and Mills 1967), but we have not located any populations in that area or seen any herbarium records from the Madison River vicinity. Some collections that we initially believed were P. strictifolius in the field were in fact P. pusillus as shown by genetic analysis (Z, Kaplan, personal communication). Based on the abundance of P. pusillus in YNP rivers, and our own YNP collections that showed that P. pusillus can take on a P. strictifolius-likQ aspect in YNP rivers, we believe that Wright and Mills (1967) were probably observing P. pusillus in the Madison River. Collections of pusilloid Potamo- geton species from rivers that appear at first glance to be P. strictifolius in the GYE should be identified with caution. USA. MONTANA. YELLOWSTONE NA- TIONAL PARK. Park Co.: Slide Lake, Gardiner Basin along Old Gardiner Rd, elev. 5723 ft, 45°00.30rN, -110°42.00rW, 9 July 2008, C. E. Hellquist and C. B. Hellquist 287-08; Landslide Creek Lake (‘Tee Lake”) in Gardiner Basin, first lake upstream along Landslide Creek ca. 1.0 mile from Old Gardiner Rd, elev. 5487 ft, 45°0L993'N, - 1 10°45.023'W, 9 July 2008, C. E. Hellquist and C. B. Hellquist 13a-08; Lower Rainbow Lake at Gardiner Basin, south of Old Gardiner Rd, elev. 5876 ft, 45°01.444'N, -110°44.62rW, 29 July 2008, C E. Hellquist and C. B. Hellquist 291-08; Upper Rainbow Lake at Gardiner Basin south of Old Gardiner Rd, elev. 5883 ft, 45°0L338'N, - 1 10°44.463'W, 29 July 2008, C E. Hellquist and C. B. Hellquist 299- 08. Potamogeton zosteriformis Fernald Potamogeton zosteriformis (flat-stemmed pondweed) is a linear-leaved species that has three distinct veins, but upon closer observation with magnification many smaller veins are visible. The numerous fine veins (15-35) are most conspicuous when specimens are dried. The stem is broadly flattened or winged and may be as wide as the leaves (Haynes and Hellquist 2000). Potamogeton zosteriformis is often found growing from large winter buds as was observed at Sylvan Lake and in Bridge Bay (Yellowstone Lake). Potamogeton zosteriformis is uncommon in the northwest portion of its range and is a species of concern in Wyoming (G5/S1). Potamogeton zosteriformis was known historically from a single site in Grand Teton National Park (Solheim 4067 [RM]). Our survey includes four new Wyoming YNP populations located in lakes or slow moving water. USA. WYOMING. YELLOWSTONE NA- TIONAL PARK. Park Co.: Sylvan Lake west of Sylvan Pass along East Entrance Rd, elev. 8461 ft. 44°28.677'N, - 1 10°09.477'W, 24 July 2008, C E. Hellquist and C. B. Hellquist 209-08; 8 Aug 2008 C. E. Hellquist and C. B. Hellquist 507-08. Teton Co.: Lewis River north of Lewis Lake, down- stream of fast water, elev. 7782 ft, 44°19.804'N, -110°38.749'W, 9 Aug 2008, C E. Hellquist and C. B. Hellquist 510-08; Bridge Bay, Yellow- stone Lake, elev. 7768 ft, 44°31.894'N, -110°26.178'W, 10 Aug 2008, C E. Hellquist and C. B. Hellquist 531-08; Lewis Lake, floating along shore at north end of lake, elev. 7825 ft, 44°19.013'N, -110°36.460'W, 12 Aug 2008, C E. Hellquist and C. B. Hellquist 604-08. Stuckenia vaginata (Turez.) Holub The genus Stuckenia is well represented in YNP. Both Y filiformis and S. pectinata (L.) Borner are widespread. Stuckenia vaginata (sheathed pondweed) is identified by its large size and its coarse, multiple-branched morphol- ogy. It is distinct from the closely related S'. pectinata by its green fruit, large inflated stipules, and blunt-tipped leaves versus flesh-colored fruit, tight stipules and acute-tipped leaves of S. pectinata. Stuckenia vaginata is the largest mem- ber of the genus in North America and is most abundant in the western United States and Canada. In Wyoming, it is not tracked because of its abundance in the southern counties of the state (B. Heidel, personal communication). In Montana, it is known from ten western counties (Lesica 2012). An additional population of Y vaginata was located just north of YNP in, Dailey Lake in Park County, MT (C. E. Hellquist 1176). Evert (2010) states that there is a single collection of S. vaginata from YNP collected in Geode Lake {Gresswell s.n. [YELLO]). However, this specimen is a misidentifled collection of Ruppia cirrhosa (see below). In 2010, we surveyed Geode Lake and did not find S. vaginata, but found abundant S. pectinata and R. cirrhosa. During herbarium research for this project, we found a previously misidentifled specimen of S. vaginata in YELLO collections from Foster Lake in the Lamar Valley and in 2008 located a large population of S. vaginata there. A second YNP site for S. vaginata was found at Slide Lake, located in the northwestern corner of YNP. At Slide Lake, S. vaginata was found with other species indicative of alkaline water including Potamogeton friesii and P. strictifolius. Two large populations also were found in separate basins in the Beaver Ponds to the northwest of Mammoth Hot Springs. USA. MONTANA. YELLOWSTONE NA- TIONAL PARK. Park Co.: Slide Lake, Gardiner Basin along Old Gardiner Rd, elev. 5723 ft, 45°00.30LN, - 1 10°42,00rW, 9 July 2008, DNA, C E. Hellquist and C. B. Hellquist 7-08; 29 July 2008, DNA, C E. Hellquist and C. B. Hellquist 2014] HELLQUIST ET AL.: UNCOMMON AQUATIC PLANTS OF YELLOWSTONE NP 173 288-08; Fourth pond off Beaver Pond Trail Loop to the southwest, starting at the west end from Mammoth Hot Springs, elev. 6519 ft, 44°59.802'N, - 1 10°43.418'W, 26 July 2010, C E. Hellquist and C. B. Hellquist 817-10; Sixth pond, the largest along Beaver Pond Trail Loop at the northeast corner of the loop, starting at the west end from Mammoth Hot Springs, elev. 6518 ft, 44°59.950'N, - 1 10°43.1 17'W, 26 July 2010, C. E, Hellquist and C. B. Hellquist 823-10. WYOMING. YELLOWSTONE NATIONAL PARK. Park Co.: “Foster Lake” west of Soda Butte, north of Lamar River trailhead along Tower-Northeast Rd, elev. 6723 ft, 44°52.344'N, ■~110°10.076'W, 14 Aug 2008, C. E. Hellquist and C. B. Hellquist 579-08. Teton Co.: “Little Lower Basin Lake”, Lower Geyser Basin, west of Firehole Lake Drive, Madison-Old Faithful Rd, elev. 7243 ft, 44°32.23rN, -110°49.872'10W, 17 July 2010, C. E. Hellquist and C. B. Hellquist 730-10; Washed ashore on north end of Yellow- stone Lake, east of Pelican Creek and east of Fishing Bridge, elev. 7736 ft, 44°33.161'N, --110°2L108'W, 11 Aug 2010, C E. Hellquist and C. B. Hellquist 1038-10. Ruppiaceae Ruppia cirrhosa (Petagna) Grande Sometimes mistaken at a distance for a Potamogeton, Ruppia (widgeon-grass) can be distinguished by its growth along sediments where rooting occurs at nodes. The inflorescences are initially sheathed by leaves and later extended on distinctly coiled peduncles. Ruppia cirrhosa is found widely scattered from the Midwest of the United States to the Pacific Coast from Alaska to California (Haynes 2000b). Described as “seldom seen or collected” in the GYE (Evert 2010, p. 608), R. cirrhosa was found primarily in sites concentrated in the Upper and Lower Geyser Basins. Ruppia was found in water associated with hydrothermal features where it was the only or one of only a few aquatic vascular plants present. In the Upper Geyser Basin, Ruppia has been seen in bloom in November (J. J. Whipple, personal observation). Tweedy (1886, p. 66) describes Ruppia as “common” in geothermal areas of YNP and sometimes found in 32°C water. Ruppia cirrhosa is found in waters with high concentrations of calcium (Crow and Hellquist 2000; Haynes 2000b) and sulfur (Haynes 2000b), conditions that are consistent with our findings (Hellquist, unpublished data). USA. WYOMING. YELLOWSTONE NA- TIONAL PARK. Teton Co.: Geode Lake north of the Mammoth-Tower Rd, near Black Canyon of the Yellowstone River, elev. 6003 ft, 44°58.604^N, - 1 10°29.32rw, 22 July 2010, C E. Hellquist and C. B. Hellquist 776-10; Pond between Geode Lake and Black Canyon of the Yellowstone, pond has abundant purple Sulfur bacteria, Mammoth-Tower Rd. elev. 5831 ft, 44°58.736'N, - 1 lp°28.964'W, 22 July 2010, C E. Hellquist and C. B. Hellquist 777-10; West Sentinel wetland west of West Sentinel between Daisy Geyser and Grand Loop Rd, Upper Geyser Basin, elev. 7352 ft, 44°28.508'N, -110°50.813'W, 30 July 2008, C E. Hellquist, C. B. Hellquist and J. J. Whipple 329-08; Small pond to the east of the Fairy Falls Trail south of Ojo Caliente Springs and Firehole River, elev. 7194 ft, 44°33.778'N, - 1 10°50.23UW, 4 Aug 2008, C. E. Hellquist and C. B. Hellquist 394-08; Lewis Lake, floating along shore at north end of lake, elev. 7825 ft, 44°19.013'N, - 1 15°36.460'W, 12 Aug 2008, C E. Hellquist and C. B. Hellquist 551-08; Fountain Flats, pool near cone west of bike trail south of Ojo Caliente Spring, elev. 7164 ft, 44°33.478'N, -- 1 10°50.509'W, 10 July 2010, C E. Hellquist and C. B. Hellquist 640-10; Fountain Flats, meandering tributary stream of Fairy Creek, stream has minor thermal influence, 44°33.465'N, - 1 10°50.786'W, 10 July 2010, C. E. Hellquist and C. B. Hellquist 649-10; Fountain Flats, Lower Geyser Basin, Feather Lake east of Goose Lake, bottom of sandy gravel, elev. 7293 ft, 44°32.675'N, ~ 1 10°50.270'W, 11 July 2010, C E. Hellquist and C. B. Hellquist 651-10; Lower Geyser Basin, River Group thermal area wetlands, small shallow wetland pools, gaseous bubbling present, ca. four pools with S. pectinata, old spring with travertine, coffee- colored water partially surrounded by Eleocharis and Schoenoplectus, elev. 7227 ft, 44°33.859'N, --110°49.988'W, 12 July 2010, C. E. Hellquist and C. B. Hellquist 660-10; Fountain Flats on east side of bike trail west of Firehole River “pond 3”, elev. 7161 ft, 44°33.3785'N, - 1 10°50.293'W, 12 July 2010, C. E. Hellquist and C. B. Hellquist 667- 10; Lower Basin Lake, Lower Geyser Basin, west of Firehole Lake Drive, Madison-Old Faithful Rd, elev. 7283 ft, 44°32.140'N, - 1 10°49.390'W, 17 July 2010, C. E. Hellquist and C. B. Hellquist 426-10; Warm seep along former White Creek, west of Firehole Lake Drive, Madison-Old Faithful Rd., elev. 7312 ft, 44°32.166'N, ”110°49.194'W, 17 July 2010, C E. Hellquist and C. B. Hellquist 732-10. Typhaceae (Including Sparganiaceae) Sparganium natans L. Sparganium natans (small bur-reed) is the smallest of the floating-leaved species of bur-reed known from the GYE. Sparganium natans was often found in shallow, geothermally influenced waters or in isolated, often stagnant pools in drying wetlands. In YNP, the foliage often has a distinct bright yellow green coloration. It is 174 MADRONO [Vol. 61 especially abundant in the Obsidian Creek drainage between Norris and Mammoth. In Wyoming, Sparganium natans was recently con- sidered a species of concern, but has been collected often in YNP over the years. Sparga- nium natans was reported from YNP in collec- tions by Rydberg and Bessey (1897) from Shoshone Lake and collections by F. Tweedy (1884) in the Gibbon River (Rydberg 1900). Sparganium natans was recently reported from sites including Nymph Lake, Obsidian Creek, North Twin Lake, and South Twin Lake (J. J. Whipple, unpublished data) and was found at these sites as well as others listed below. USA. IDAHO. YELLOWSTONE NATION- AL PARK. Fremont Co.: “Pond” southeast of Bechler Ranger Station, only open water at east end, remainder covered by Schoenoplectus acutus, elev. 6355 ft, 44°08.510'N, - 1 1 r03.343'W, 4 Aug 2010, C E. Hellquist and C. B. Hellquist 922- 10', Robinson Lake, Bechler District, lake a peat mat surrounded by a band of water dominated by Nuphar, elev. 6569 ft, 44°09.930'N, -lir04.213'W, 5 Aug 2010, C E. Hellquist and C. B. Hellquist 944-10. WYOMING. YEL- LOWSTONE NATIONAL PARK. Park Co.: Channeled wetlands and Obsidian Creek by Beaver Lake in waters warmed by geother- mal activity, elev. 7402 ft, 44°49.083'N, -110°43.870^W, 18 July 2008, C E. Hellquist and C. B. Hellquist 133-08', Obsidian Creek just west of the crossing of the Mammoth-Norris Rd at Obsidian Cliff, elev. 7460 ft, 44°49.418'N, -110°43.748'W, 18 July 2008, C E. Hellquist and C. B. Hellquist 141-08', Obsidian Creek be- tween entrance to Indian Creek Campground and Mammoth-Norris Rd. elev. 7444 ft, 44°52.947'N, - 1 10°44.072'W, 18 July 2008, C. E. Hellquist and C. B. Hellquist 147-08’, North Twin Lake along Mammoth-Norris Rd, elev. 7574 ft, 44°46.5irN, - 1 10°44.180'W, 7 Aug 2008, C. E. Hellquist and C. B. Hellquist 462-08', South Twin Lake along Madison - Norris Rd, elev. 7571 ft, 44°46.275'N, - 1 10°44.005'W, 7 Aug 2008, C E. Hellquist and C. B. Hellquist 469- 08', Stream flowing into Nymph Lake, north of Norris along the Norris-Mammoth Rd, stream thermally influenced, elev. 7492 ft, 44°45.201'N, - 1 10°43.613'W, 29 July 2010, C. E. Hellquist and C. B. Hellquist 858-10; “Third Quadrant Pond”, due west of Swan Lake and power line, pond covered by sedges and grasses, elev. 7674 ft, 44°55.44LN, - 1 10°46.096'W, 20 July 2010, C E. Hellquist and C. B. Hellquist 764-10; Pond immediately alongside Norris-Mammoth Hwy., north of Norris and south of Nymph Lake, pond thermally influenced, elev. 7540 ft, 44°45.027'N, - 1 10°43.000'W, 29 July 2010, C E. Hellquist and C. B. Hellquist 860-10; Small Nuphar pond on north side of Norris-Canyon Hwy., elev. 8103 ft, 44°43.077'N, - 1 10°32.390'W, 2 Aug 2010, C. E. Hellquist and C. B. Hellquist 911-10. Teton Co.: Partially dried oxbow of the Gibbon River at Madison Jet., elev. 6793 ft, 44°38.534'N, -110°5L469'W, 1 Aug 2008, C E. Hellquist and C. B. Hellquist 386-08; Harlequin Lake north of West Yellowstone-Madison Jet Rd, elev. 6896 ft, 44°38.598'N, - 1 10°53.575'W, C. E. Hellquist and C. B. Hellquist 419-08; Southwest- ern-most pond north of trail to Nez Perce Patrol Cabin, north of Nez Perce Creek, Madison-Old Faithful Rd., pond mostly filled with grasses and sedges, elev. 7220 ft, 44°34.632'N, -110°48.828'W, 17 July 2010, C. E. Hellquist and C. B. Hellquist 708-10; South, small pond at jet of Bechler Meadows Trail and Boundary Creek Trail, elev. 6475 ft, 44°10.033'N, -lir0L958'W, 5 Aug 2010, C E. Hellquist and C. B. Hellquist 953-10; South end of large pond at jet of Bechler Meadows Trail and Boundary Creek Trail, Bechler District, elev. 6475 ft, 44°10.033'N, -lir0L958'W, 5 Aug 2010, C. E Hellquist and C. B. Hellquist 966-10. Conclusions The complex geology and corresponding hy- drology of YNP provides diverse aquatic habitat for uncommon and unusual aquatic vascular plants of the Rocky Mountain region. With additional reconnaissance into the interior of YNP, we anticipate more records for these species will be found. Forthcoming papers will provide a complete species inventory of the aquatic flora of YNP based on continuing fieldwork and will examine patterns of aquatic plant distribution in YNP as related to water chemistry. Our work also illustrates that field determinations of select aquatic taxa can some- times be problematic due to morphological subtlety or potential hybridization. In an upcom- ing paper, we will describe the abundance and distribution of several hybrid Potamogeton and Stuckenia taxa found in YNP that we have confirmed with genetic analysis. The prevalence of hybrid taxa within Potamogeton and Stuckenia as well as the morphological similarities found within some Potamogeton and Utricularia high- lights the need for further taxonomic studies of aquatic plants. Our results provide a basis for re- evaluating rarity among aquatic vascular plants of Yellowstone National Park, the GYE, and the northern Rocky Mountain region as a whole. Acknowledgments We are very grateful to the Yellowstone Park Foundation for funding our research and for the assistance of Molly Pickle. We greatly appreciate the logistical support, assistance, and encouragement of the personnel of Yellowstone National Park. Mary Hektner (former YNP Supervisory Vegetation Specialist) pro- vided essential support for our work throughout the duration of the project. YNP botanist Heidi Anderson 2014] HELLQUIST ET AL.: UNCOMMON AQUATIC PLANTS OF YELLOWSTONE NP 175 provided valuable field and herbarium assistance. Christie Hendrix and Stacey Gunther in the YNP Research Office provided critical logistical support during our work regarding our research permit and related procedures. We appreciate the assistance of Garrett Crow, Judith Fehrer, R. James Hickey, Zdenek Kaplan, Donald Les, and Barry Rice for annotating specimens noted in the methods. Thanks also to Scott Mincemoyer (Montana Natural Heritage Program) for his assistance with Montana records. We are very grateful to Bonnie Heidel (Wyoming Natural Diversity Database) for answering a variety of questions about Wyoming aquatic plant records and for providing valuable comments on the manuscript. At the Univer- sity of Michigan Biological Station (UMBS), Mike Grant provided water chemistry analysis and UMBS provided logistical support during manuscript prepara- tion. CEH and CBH also thank Marion Hellquist for assistance editing the manuscript and for her patience and support during our field seasons. We thank our reviewers for their thoughtful edits and suggestions that helped us improve the manuscript. This research was conducted under the guidelines of YNP Research Permit 5734 issued to CEH and CBH. Literature Cited Aiken, S. G. 1981. A conspectus of Myriophyllum (Haloragaceae) in North America. Brittonia 33:57-69. Al-Shehbaz, I. A. 2010. Brassicaceae Burnett. Pp. 224- 747 in Flora of North America Editorial Commit- tee (eds.), Flora of North America North of Mexico, Vol. 7: Magnoliophyta: Salicaceae to Brassicaceae. Oxford University Press, New York, NY. Ceska, O., a. Ceska, and P. D. Warrington. 1986. Myriophyllum quitense and Myriophyllum ussur- iense (Haloragaceae) in British Columbia, Canada. Brittonia 38:73-81. Couch, R. and E. Nelson. 1988. Myriophyllum quitense (Haloragaceae) in the United States. Brittonia 40:85-88. Crow, G. E. and C. B. Hellquist. 2000. Aquatic and wetland plants of northeastern North America, Volume 1. Pteridophytes, gymnosperms, and an- giosperms: dicotyledons. University of Wisconsin Press, Madison, WI. Dorn, R. D. 1984. Vascular plants of Montana. Mountain West Publishing, Cheyenne, WY. . 2001. Vascular plants of Wyoming. 3rd ed. Mountain West Publishing, Cheyenne, WY. Evert, E. F. 2010. Vascular plants of the Greater Yellowstone Area: Annotated catalog and atlas. Published by Erwin F. Evert, Park Ridge, IL. Fassett, N. C. 1957. A manual of aquatic plants. With revision appendix by E. C. Ogden. University of Wisconsin Press, Madison, WI. Fernald, M. L. 1932. The linear-leaved North American species of Potamogeton section Axillares. Memoirs of the American Academy of Arts and Sciences 17:1-183. Gluck, H. 1906. Biologische und morphologische Untersuchungen iiber Wasser- und Sumpfge- wachse. Vol. 2, Untersuchungen fiber die mitte- leuropaischen Utricularia-AitQn, fiber die Turio- nenbildung bei Wasserpflanzen, sowie fiber Ceratophyllum. Verlag von Gustav Fischer, Jena, Germany. Haynes, R. R. 2000a. Najadaceae Jussieu. Pp. 77-83 in Flora of North America Editorial Committee (eds.). Flora of North America North of Mexico. Vol. 22: Magnoliophyta: Alismatidae, Arecidae, Commelinidae (in part), and Zingiberidae. Oxford University Press, New York, NY. . 2000b. Ruppiaceae Hutchinson. Pp. 75-76 in Flora of North America Editorial Committee (eds.). Flora of North America North of Mexico. Vol. 22: Magnoliophyta: Alismatidae, Arecidae, Commelinidae (in part), and Zingiberidae. Oxford University Press, New York, NY. AND C. B. Hellquist. 2000. Potamogetona- ceae Dumortier. Pp. 47-74 in Flora of North America Editorial Committee (eds.), Flora of North America North of Mexico. Vol. 22: Magno- liophyta: Alismatidae, Arecidae, Commelinidae (in part), and Zingiberidae. Oxford University Press, New York, NY. AND J. L. Reveal. 1973. A re-evaluation of Potamogeton fibrillosus Fern. (Potamogetonaceae) Rhodora 75:75-77. Heidel, B. 2012. 2012 plant species of potential concern list. April 7, 2012. Wyoming Natural Diversity Database, University of Wyoming, Lar- amie, WY. Website www.uwyo.edu/wyndd/species- of-concem/piants/index.html (accessed July 2012). Hellquist, C. B. 1980. Correlation of alkalinity and the distribution of Potamogeton in New England. Rhodora 82:331-344. , R. F. Thorne and R. R. Haynes. 2012. Potamogeton foliosus subsp. fibrillosus (Fernald) R. R. Haynes and Hellq. in Jepson Flora Project (eds.), Jepson eFlora. Website http;//ucjeps.berke- ley.eciu/cgi-bin/get_IJM.pl?tid= 80561 (accessed 22 Aug 2012). Holmgren, N. H., P. K. Holmgren, and A. Cronquist. 2005. Intermountain flora: vascular plants of the Intermountain West, USA. Volume 2, Part B. Subclass Dilleniidae. New York Botanical Garden, Bronx, NY. Kaul, R. B. 2000. Sparganiaceae Rudolphi. Pp. 270- 277 in Flora of North America Editorial Commit- tee (eds.). Flora of North America North of Mexico. Vol. 22: Magnoliophyta: Alismatidae, Arecidae, Commelinidae (in part), and Zingiber- idae. Oxford University Press, New York, NY. Kesonie, D. and R. L. Hartman. 2011. A floristic inventory of Grand Teton National Park, Pinyon Peak Highlands, and vicinity, Wyoming, USA. Journal of the Botanical Research Institute of Texas 5:357-388. Landolt, E. 2000. Lemnaceae Gray. Pp. 143-153 in Flora of North America Editorial Committee (eds.), Flora of North America North of Mexico. Vol. 22: Magnoliophyta: Alismatidae, Arecidae, Commelinidae (in part), and Zingiberidae. Oxford University Press, New York, NY. Les, D. H. 1997. Certatophyllaceae Gray. Pp. 81-84 in Flora of North America Editorial Committee (eds.). Flora of North America North of Mexico. Vol. 3: Magnoliophyta: Magnoliidae and Hama- melidae. Oxford University Press, New York, NY. , N. M. Murray, and N. P. Tippery. 2009. Systematics of two imperiled pondweeds {Potamo- geton vaseyi, P. gemmiparus) and taxonomic 176 MADRONO [Vol. 61 ramifications for subsection Pusilli (Potamogeto- naceae). Systematic Botany 34:643-651. Lesica, P. 2012, Manual of Montana vascular plants. Botanical Research Institute of Texas, Fort Worth, TX. Lichvar, R. W., N. C. Melvin, M. L. Butterwick, AND W. N. Kirchner. 2012. National wetland plant list indicator rating definitions. U.S. Army Corps of Engineers, Engineer Research and Devel- opment Center (ERDC), Cold Regions Research and Engineering Laboratory (CRREL), ERDC/ CRREL Technical Note 12-1, Hanover, NH. Marston, R. a. and J. E. Anderson. 1991. Watersheds and vegetation of the Greater Yellow- stone Ecosystem. Conservation Biology 5:338-346. McAlpine, D. F., G. Bishop, O. Ceska, M. L. Moody, and A. Ceska. 2007. Andean water- milfoil, Myriophyllum quitense (Haloragaceae), in • the Saint John River Estuary System, New Bruns- wick, Canada. Rhodora 109:101-107. Mulligan, H. F. 1969. Management of aquatic vascular plants and algae, Pp. 464^82 in G. A. Rohlich, A. M. Beeton, D. C. Chandler, W. T. Edmundson, and A. D. Hasler (eds.). Eutrophica- tion: causes, consequences, correctives; proceedings of a symposium, 11-15 June 1967, University of Wisconsin, Madison. National Academy of Sci- ences, Washington, D.C. Mulligan, G. A. and J. A. Calder. 1964. The genus Subularia (Cruciferae). Rhodora 66:127-135. Natureserve. 2012. NatureServe Explorer: an online encyclopedia of life (web application). Version 7.1. NatureServe, Arlington, VA. Website http:// www.natureserve.org/explorer. (accessed 19 July 2012). Rydberg, P. A. 1900. Catalogue of the Flora of Montana and the Yellowstone National Park. Memoirs of the New York Botanical Garden, Vol. 1 , Press of the New Era Print. Co., New York, NY. SCULTHORPE, C. D. 1967. The biology of aquatic vascular plants. Edward Arnold, London, U.K. Taylor, P. 1989. The genus Utricularia: A taxonomic monograph. Kew Bulletin Additional Series XIV. Her Majesty’s Stationery Office, London, U.K. Taylor, W. C., N. T. Luebke, D. M. Britton, R. J. Hickey, and D. F. Brunton. 1993. Isoetaceae Reichenbach. Pp. 64—75 in Flora of North America Editorial Committee (eds.), Vol. 2: Pteridophytes and Gymnosperms. Oxford University Press, New York, NY. Thor, G. 1988. The genus Utricularia in Nordic countries, with special emphasis on U. stygia and U. ochroleuca. Nordic Journal of Botany 8:213-225. Tiner, R. W. 2012. Defining hydrophytes for wetland identification and delineation. U.S. Army Corps of Engineers, Engineer Research and Development Center (ERDC), Cold Regions Research and Engineering Laboratory (CRREL), ERDC/ CRREL CR Technical Note 12-1, Hanover, NH. Tweedy, F. 1886. Flora of the Yellowstone National Park. Published by F. Tweedy, Washington, D.C. United States Department of Agriculture, Natural Resources Conservation Service (USDA, NRCS). The Plants Database. Website http://www.plants.usda.gov/ (accessed January 2012). Varley, j. D. and P. Schullery. 1998. Yellowstone Fishes: ecology, history, and angling in the park. Stackpole Books, Mechanicsburg, PA. Wright, J. C. and L K. Mills. 1967. Productivity studies on the Madison River, Yellowstone Na- tional Park. Limnology and Oceanography 12:568-577. Young, L. (ed.) 2007, Yellowstone Resources and Issues 2007. Division of Interpretation, Yellow- stone National Park, Mammoth Hot Springs, WY. Madrono, Vol. 61, No. 2, pp. 177-193, 2014 STOMATA SIZE IN RELATION TO PLOIDY LEVEL IN NORTH AMERICAN HAWTHORNS {CRATAEGUS, ROSACEAE) Brechann V. McGoey Department of Ecology and Evolutionary Biology, University of Toronto, Toronto, ON, Canada M5S 3B2 brechann.mcgoey@utoronto.ca Kelvin Chau Canadian Food Inspection Agency, 1124 Finch Ave. W, Unit 2, Toronto, ON, Canada M3J 2E2 Timothy A. Dickinson Green Plant Herbarium (TRT), Department of Natural History, Royal Ontario Museum, 100 Queen's Park, Toronto, ON, Canada M5S 2C6, and Department of Ecology and Evolutionary Biology, University of Toronto, Toronto, ON, Canada MSS 3B2 Abstract The impacts of ploidy level changes on plant physiology and ecology present interesting avenues of research, and many questions remain unanswered. Here, we examine the connections between cytotype, taxon, stomata characteristics, and environmental variables in black-fruited hawthorns of the Pacific Northwest {Crataegus ser. Douglasianae\ Maleae, Amygdaloideae, Rosaceae). We explore the extent to which stomatal measurements can be used to predict ploidy level and how differences in ploidy level and stomata characteristics relate to geographic distributions. We sampled trees from across the geographic ranges of the putative sister taxa Crataegus suksdorfii (Sarg.) Kruschke (diploids and autotriploids) and C. douglasii Lindl. (tetraploids). We found that stomata differed between the two species, with tetraploid C. douglasii having larger average stomata sizes than diploid and triploid C suksdorfii We also obtained climatological and elevation data for the sites at which these samples were collected, and examined the associations between taxon, ploidy level, stomatal size and density, elevation, and environmental parameters. Our analyses indicate positive associations between stomatal size and latitude, and between ploidy level and elevation. Negative associations were found between temperature and precipitation variables and both ploidy level and stomatal size, particularly for the fall and winter quarters. There appeared to be no significant association between stomatal density and any of the environmental variables. Tetraploid C. douglasii occupied a wider range of environmental conditions than did either the diploids or the autotriploids. Key Words: Agamospermy, climate, Crataegus douglasii, Crataegus gaylussacia, Crataegus suksdorfii, geographical parthenogenesis, polyploidy, stomata. Determining the causal factors behind species distributions is a fundamental area of study in ecology (Krebs 2001). The realized range of an individual species can be affected by abiotic factors such as climate, biotic interactions such as competition, or a more complex interplay of environmental factors (Angert and Schemske 2005). Species ranges vary dramatically in area, and even closely related species can have very divergent range sizes (Brown et al. 1996). For some close relatives, differences in cytotype may be the underlying cause of range size differences; polyploidy in one species could lead to an ability to tolerate a greater or different range of conditions, and therefore facilitate a wider or novel distribution, compared to a diploid conge- ner (Levin 1983; Krebs 2001; Parisod et al. 2010; Ramsey 2011). A polyploidization event may present opportunities in terms of range expansion, including a possible increase in genetic diversity, and a decrease in barriers to self-fertilization (Angert and Schemske 2005, Lowry and Lester 2006). In addition, polyploids have been found to have a number of advantages over their diploid relatives, including larger seed sizes, more vigor- ous seedlings, increased resistance to pathogens and pests, and higher tolerances for poor soils and drought conditions (Levin 1983; Brown et al. 1996). These differences can translate into diver- gent ecological tolerances and geographic distri- butions (Lewis 1980; Lowry and Lester 2006; Ramsey 2011; McIntyre 2012). Although differ- ences between cytotypes in range sizes and environmental factors are not universal (Martin and Husband 2009), some rapid divergence after polyploidization is theoretically necessary for the new polyploid population to avoid direct compe- tition with its progenitor, thereby allowing for its persistence (Coyne and Orr 2004; McIntyre 2012). However, the connections between polyploidy and observed ecological differentiation remain poorly understood (Li et al. 1996; Ramsey 2011). 178 MADRONO [VoL 61 Polyploidy is frequently associated with gameto- phytic apomixis, notably in Ranunculus L. species (Horandl 2008) and in many Asteraceae, Poaceae, and Rosaceae (Nogier 1984). In these cases the wider range of apomicts may be related not only to polyploidy but also to the ability to set seed in the absence of other conspecific individuals (Horandl 2006; Lo et ah 2013). Investigations of the distribution and fre- quency of polyploidy typically depend on the availability of tissue from which chromosome counts can be made or, increasingly, from which stainable, intact nuclei can be extracted and passed through a flow cytometer. Neither of these approaches offers much chance of success with existing specim.ens from museum collections. Taxonomic recognition of major morphological differences between ploidy levels may be sufficient in some cases, as in the contrast between diploid hawthorns, all of which have 20 stamens per flower, while almost all tetraploid hawthorns have 10 stamens per flower (Talent and Dickinson 2005). Unfortu- nately, this pattern is confounded by intraspe- cific variation in ploidy level (e.g., Crataegus suksdorfii [Sarg.] Kruschke [Dickinson et al. 2008]), and morphological variation among closely related tetraploids (e.g., C. crus-galti L. sensu lato [Dickinson and Phipps 1986]). In addition, flowering traits may not always be visible on herbarium specimens. Considering these issues, researchers may seek proxies for direct measurements of chromosome number or nuclear DNA content. Cell size is one such proxy, since polyploids typically have larger cells than diploids (Stebbins 1950; Levin 1983; Otto 2007; Beaulieu et al. 2008). In plants, the size of pollen grains and of stomatal guard cell pairs may both be accessible with herbarium material and work on a variety of taxa, including Crataegus (Marshall 1978) and other Rosaceae (Joly and Bruneau 2007), has shown how these traits can be used to predict ploidy level in herbarium specimens (Buechler 2000; Joachimiak and Grabowska- Joachimiak 2000; Saitonstall et al. 2007; Chen et al. 2009). Relatively simple methods for light microscopy may have the potential to reveal the ploidy level of long dead individuals from which herbarium specimens of historical interest were collected. However, in hawthorns the variation in pollen grain size within ploidy levels is so great that, while overall size increases with nuclear DNA content mea- sured in leaves, no prediction from pollen measurements to ploidy level is possible; only stainability differentiates the pollen of diploids (highly stainable) from that of polyploids (Dick- inson, unpublished data). Importantly, features such as stomatal size and density have been recognized as varying not only with ploidy level, but also with environmental factors, such as temperature and water availabil- ity (Wang and Clarke 1993). The links between cytotype, stomata, and environmental variables have important implications for the realized ranges of plant species, and could help explain the differences in the breadth and characteristics of habitats occupied by congeners. Here, we investigate the way in which ploidy level, stomatal dimensions, and stomatal density vary with each other, and in relation to geographic distribution and the environment, as indexed by latitude, elevation, and climate parameters. We ask the following questions: (1) does size or some other readily measurable characteristic of cells with a fixed ontogenetic trajectory like stomata vary with ploidy level in such a way that this parameter can be used to predict ploidy level?, (2) does size or some other characteristic of stomata vary between taxa even when ploidy level is kept constant?, and finally (3) is cell size or other morphological variation between ploidy levels and taxa associated in any way with differences in geographic distribution? Materials and Methods Crataegus L. is a genus of approximately 200 species of woody plants, part of a mostly fleshy-fruited clade (tribe Maleae Small) within an expanded subfamily Amygdaloideae Am. (Rosaceae) (Potter et al. 2007; McNeill et al. 2012). Crataegus is taxonomically complex be- cause, while the subgeneric classification of the genus into taxonomic sections and series (Phipps et al. 1990; Phipps and O’Kennon 2002) is broadly supported by morphology, not all subgeneric groups have been shown to be monophyletic (Lo et al. 2007, 2009). Moreover, species circumscriptions often inadequately re- fleet the occurrence of hybridization, gameto- phytic apomixis, and polyploidy. Black-fruited Crataegus series Dougiasianae (Loud.) Rehder is one example of an apparently monophyletic group (Lo et al. 2007, 2009). This group is widespread in the Pacific Northwest of the United States and Canada and includes two well-known species, Crataegus suksdorfii (Sarg.) Kruschke and C. douglasii Lindl. Crataegus suksdorfii is restricted to the mesic conifer forest region of the Pacific Northwest (Brunsfeld and Johnson 1990), while C douglasii can grow in more xeric areas and so has a much wider distribution, extending from the eastern slopes of the Cascades to the Rocky Mountains, with disjunct populations east of the continental divide (Hargrove and Luxmoore 1998; Dickinson et al. 2008). Diploid, triploid, tetraploid, and pentaploid C suksdorfii are now known (Dickinson et al. 2008; Lo et al. 2010a, 2013; Goughian 2012), although only one of these cytotypes has, as yet, been given 2014] MCGOEY ET AL.: PLOIDY LEVEL AND STOMATA SIZE IN CRATAEGUS 179 taxonomic recognition. Crataegus gaylussacia A. Heller was described from Sonoma County, California, and has been shown to be triploid (Coughlan 2012, Dickinson et al. unpublished data), and to correspond more or less to the morphology of autotriploid individuals found elsewhere that would otherwise be referred to C. suksdorfii (Lo et al. 2010a). Crataegus gaylussacia has priority over the name C suksdorfii if these names are applied to all 20-stamen, black-fruited hawthorns of the Pacific Northwest without regard to ploidy level (Phipps 2012; VASCAN 2013). In view of the variation in ploidy level documented for these plants (Talent and Dickinson 2005; Lo et al. 2013), we will retain the older usage of these names here (i.e., as referring to different biological entities). A publication clarifying the use of these names is currently in preparation (T. A. Dickinson, unpublished manuscript). In contrast to C suksdorfii, individuals of C douglasii are almost exclusively tetraploid (Talent and Dickinson 2005; Lo et al. 2013), and distinctive morphotypes that have received taxo- nomic recognition (Phipps and O’Kennon 1998, 2002) appear to be of hybrid origin (Zarrei et al. 2012). Introgression from C douglasii appears to be responsible for the allopolyploid cytotypes of C. suksdorfii (Lo et al. 2010a). The wider geographic distribution of tetraploid C. douglasii is associated with an apparently almost complete reliance on asexual seed production (Talent and Dickinson 2007; Coughlan 2012; Lo et al. 2013), as is also the case in polyploid C. suksdorfii (Dickinson et al. 1996; Lo et al. 2013). Diploid C suksdorfii, on the other hand, appears to produce exclusively biparental, sexual seeds (Coughlan 2012; Lo et al. 2013). Sampling Sampling for this study draws on a large collection of mainly Pacific Northwest and Ontario specimens of Crataegus series Douglasia- nae, assembled and georeferenced for studies of morphological, breeding system, and ploidy level variation, that are housed for the most part in the Green Plant Herbarium (TRT) of the Royal Ontario Museum (Table 1). Many, but not all of these, are vouchers for studies of DNA sequence variation in the group (Lo et al. 2007, 2009; Coughlan 2012; Zarrei et al. 2012). To more fully sample the continental United States range of C douglasii, permission was obtained to remove leaf material from two borrowed specimens from Michigan (Table 1). Specimens were selected that had sufficient mature leaves to permit removal of some for destructive sampling. In the geographic study, specimen selection had the further cri- terion that only C. douglasii and C. suksdorfii (both diploids and autotriploids) would be compared, unconfounded by being combined as in allopolyploid C suksdorfii (Lo et al. 2009, 2010a). Locations of the individuals and popula- tions studied were mapped using SimpleMappr (Shorthouse 2010). Ploidy level determination. Most specimens studied are vouchers for flow cytometric ploidy level determinations (Talent and Dickinson 2005). However, because C. douglasii sensu lato has been found to be almost exclusively tetra- ploid (Talent and Dickinson 2005, Talent and Dickinson unpublished data), we have included some specimens of this taxon for which flow cytometric or chromosome count data are unavailable, and for our purposes have assumed that they are tetraploid (Table 1). Stomatal Size - Population Level Comparison Specimens for this study were selected to compare population samples of diploid (two populations, n ^ 1 and n = 10, Table 1), autotriploid, allotriploid, and allotetraploid (one population each, all n — 10) C. suksdorfii with two samples (each n = 10) of tetraploid C douglasii, each of the latter sympatric with one of the allopolyploid C. suksdorfii populations Table 1 (Dickinson et al. 1996; Lo et al. 2009, 2010a). Two leaves, randomly chosen from both short and long shoots were removed from each specimen, and two approximately 1 cm X 1 cm segments were excised from each leaf. For each leaf a random number from one to six was obtained and used to select a sampling position on the leaf (Fig. 1). Leaf segments were softened for 30-40 min in 70% ethanol, rinsed in tap water for 15-20 min, and then submerged overnight in 100% Drano® (S. C. Johnson & Son, Inc., Racine, WI) to clear the leaf segments (Buechler 2000). The following day, the samples were rinsed in water for 15-20 min, mounted in water on a microscope slide, and examined under Nomarski differential interference contrast at 800 X magni- fication. Using MorphoSys v. 1 .29 (Meacham and Duncan 1991), stomatal outlines were captured in order to calculate their length, width, area, and perimeter. In each leaf segment, five stomates were measured in five different quadrants of the leaf fragment: upper left, upper right, bottom left, bottom right, and middle, giving a total of 25 stomates measured per segment. To avoid count- ing a stomate more than once, the same counting sequence was used consistently (i.e., after the first stomate, go up, skip one, measure the following one; next, go left, skip one, count the following one; then up again and so on). Stomatal Size and Density - Geographic Study Specimens for this study were selected to compare diploid and autotriploid C suksdorfii 180 MADRONO [VoL 61 w ^ o w W3 m Z ■ ■ « o ^ > CM o o cd Si . Q W S 3 6^ OT ►-J n 1-^ m « " Q 8 a ^ Z ^ S A o H « . , <.a > ^ ^ .a W o o ® -o [L d W O P H< Z H 2 H m i§ ^ CM z Q OJ ^ w ^RJ J H « W P ^ w 6 I g §.§P o ^ = u W o « «y C ^ ^ « . !l ® w ^ « P « > P -O >> .. w "2 d d ’o o d^.S „ c^ w p VO S go op I zSh< Q W'O^.- ^ >H ^ L, H H _ < a? ^ > R. g OT 3 < Q 5 o j Q P z z>^ 9 w H < D o p g ^ < Q 94 2 o > 8 z ^ w bj t 5 m ^ Q ^ < .s H 'S dj S 6 3 o K 'S W 0) ^ 'O s .S o fi cfl .5 c« a ^ 8 § cd M ’o a ’B '^ S H O M ap b ’2 W d o f- fa ti o d o ^ p a o p o U o « ^ > s 'd ^ .a gj ^ S S o ^ s s w 2 d s 8p .-d X P Q M W 'O d £ ^ a "a p 3 p. 'S g o .2 o d > 3 a o a w • &a 3§ XX XXX m m m rs| d d Vo o VO O O o o O r^ fN I I g IT) \0 VO VO »0 >40 S R 0 O S, S, 1 I O O s s iD CN fN S Q> n-i S ^ 'y ^ ^ 2v P p X. ^ 1^ -o a d ^ .. o a g S'S d d o — , s o 3 > “ V U I ^ fl S 3 I ^ I - i s g -g ’d g . o d c^ d 3 m ’Ei O 'rt 05 X O d O oa B d ed ^~» I f-i G ^ §s o 2 S eO 2^3 Eu o « 3 8 30 I'S .3 'm 3 ■d < ca o s a I gj m M a i -i cSd d a^ a ” 3 -I o l> M ^ ^ ^x r^ S K >>^ H H V Vo Vo H H P P 2v > t\ Qo H H p b5 N N rx 0© H H P ao o « .s 8 50 CO d o- fX O MX CO VO o m VO OD d 0© Ov d cn o\ Ol o o M cx i~=4 cx d m VO VO CX fO ON VO cn Ov o o d m 5— i o\ d « m cx m ov j>. vd d p d d vd vd vd d vd OI cx cx d !— 1 ?— M »«N T—l ¥— ( 1—) ?— i ?— N cx f-=< ¥—1 ¥— I 5-»l ¥—( «\ »- •r—i ov d 4' d d es ov s r- CX 00 ON Ov VO m d VO fX ov o I— 1 ex ro rn o\ ov d rn MX ov d ■d d m d d tn i 5 vi d vd I I N d p ^ K o < I UN s :«l 6^ 60 Q . 3 a MO O < H o o m 2 I' o> m 0 a o ^ +-^ d- S ^ o w W) y .2 o 'S > 3 a o a W 2 ^ • isD 3 § X X JO ^ n K, pp (V) K A O e °9 Op Go o 00 a o o G N OC fN kj r\j in a ^ ^ fn fn fn > ^ H H H H tjq bq '-'<>00 0 fT) pn fV) > i i i * ^ ^ 'o k^ tq kj 5 a| ^ e ^x-g S .’g o S^i-s g:a - cd O ^ cd-S o o m o^ os VO Os VO o O m C OS r^ m lO VO in VO o -— ( «n 1— 1 in q q os \o o as as os 00 o Os m CN o m q q q so ^ q q q K fk fk tk in *n in rn -O cd ^ u '8-B ^ I " S g-c ^PQ Q Q G ^ Ph o « H S Ph 3 g O S rt O o .ft -*-> Q .rt +-' H-l .3 hJ fi cd .1 § :§ § I u D ..X bi) w s Co § O .S 2 ^ nJ 3 . o ^ S faL ® w O 00 t; c*H -Ts 3 a o O PQ Q r- Qo s 3 'm ^ S ^ •|X I ^ §0 o ^ 60 3 o Q pi u 182 MADRONO [Vol. ^ .a o 5 ft O cd bX) O M) O cd > 3 a o a W fN fN-i 'VO lo H H ki Oo fN oo a . g“- CiO 'o Oo ^ 00 3 ^ S 3 « ^ 'O N .s N OJ s . on Oo so Oo On On cd > On 00 P Oo ^ ’'O N .s N N ^ . tq rxj ix ^ ‘o *o 2: ^ ^ 1; k] ki Lq ’'xi fv-) No n 2 ^ ^ 2 'o o ao^^cdcdOj. o ^ s-i 'd ^.1 'S s go B § CD .'d ^ <1^ o j I S g o o d o ^ G NO cd rx ON CM 00 00 (N 00 —< 00 m ON O' 00 m' (N rn Cvi o 04 04 oi (N (N CN r— ( O) 04 04 ^—1 1— ( »— ( »— 1 rs! i-H Y— ( r—l >— ( r2' r-" rx-" Md r-- NO m MO ro 50) NO (N l> cn MO 7 00 00 ^ 2 O M h oj B M fi 'd op eN u bi) w S g Table 1. Continued. Lat. N, Vouchers Vouchers State Taxon County Site Long. W Elev. (m) Site notes (population study) (geographic study) Pioidy Wyoming C. douglasii Teton (29) Targhee 43.277, 110.795 1782 Snake River Phipps 7459 2014] MCGOEY ET AL.: PLOIDY LEVEL AND STOMATA SIZE IN CRATAEGUS 183 •S 'E o o a a ■§ o JD 00 On G o.> o H d CQ Q s s I .^1 . d 2 i ^ 2 ^ U O 2 la with tetraploid C douglasii across their continen- tal United States ranges (Fig. 2, Dickinson et al. 2008). Leaf samples were examined from 15 C. suksdorfii, including eight diploids and seven autotriploids, and 17 tetraploid C. douglasii trees (Table 1). The methods were the same as for the population-level study above, with the following exceptions. Softening the leaves was deemed unnecessary, and leaves were submerged in Drano® for 72 hr. Since the results of the previous study demonstrated that the sampling location on the leaf (Fig. 1) affected stomata size significantly (see Results), we restricted this study to leaf sections in the middle right of the leaf (section four in Fig. 1). Within each 1 cm^ leaf section we examined an areole in each of four different, spatially separated regions (areoles are areas bounded by quarternary veins on all sides). We captured images of an areole at different focal depths using an Infinity 1 digital camera and the Infinity Analyze program supplied with it (Lu- mQuera Corporation, Ottawa, ON). The images were stacked using Combine Z image software (Hadley 2010), and the stacked images were then imported back to the Infinity Analyze program. This was used to delimit individual areoles, determine their area, and count all the stomata found within an areole. We also used the program to measure the length and width of each stomata (Fig. 3). All measurements were automatically stored in an M.S. Access^^ data- base (Microsoft, Redmond, WA). Environmental data. Latitude and longitude for each collection site (Table 1) were used to obtain climate data (1971-2000 monthly normals for precipitation, and maximum and minimum tem- perature) using the online PRISM Data Explorer (PRISM Climate Group 2012). The geographic coordinates were also used to ascertain the ele- vation at each site using Google Earth (Version 7.1.2.2041 [software] available from http://www. google.com/earth/). Google Earth was also employed to estimate the proximity of sampling sites to bodies of water. The analyses described below used seasonal (quarterly) averages of the monthly normal values (Ql, winter, December- February; Q2, spring, March-May; Q3, summer, June-August; and Q4, fall, September-Novem- ber). Some analyses also included the average minimum and maximum temperatures for the year, and total normal precipitation. Analyses of just the quarterly averages also included site elevation and latitude. Statistical analyses. Statistical analyses were performed using R (R Development Core Team 2004). Normality of the size measurements was checked using qq plots and the Shapiro-Wilks test. Measurements found to be lognormally distributed were log-transformed. When analyz- 184 MADRONO [Vol. 61 Fig. 1. Diagram of a Crataegus leaf divided into six sections. For the population-level study, the sampled section was chosen randomly while, for the geographic study, the mid-right section (4) was always used. In all cases the abaxial surface of the leaf was examined as shown here. ing the stomata size data, two different methods were used to avoid the pseudoreplication problem of having many stomata measurements from the same leaves and trees. First, the mean values for each tree were calculated, and then compared. In the second analysis, we used a linear mixed effects model to determine the effect of ploidy and taxonomic group while accounting for the replication within leaves and trees. To analyze the relationship between stomatal traits, ploidy level, and climate variables we conducted a MANOVA and principal component analysis. We used the fourth-corner function in R from the ade4 package (Legendre et al. 1997, Dray and Legendre 2008) in order to infer relationships between our morphological data (stomata size and density; nuclear DNA content) and the environmental features of the sites referred to above. Results While examining our C. douglasii and C. suksdorfii leaves, we made basic observations on the leaf surface characteristics (Fig. 3). The outlines of epidermal cells were wavy and cells surrounding the stomata appeared anomocytic. This is consistent with Metcalfe and Chalke’s (1979) description of Rosaceae epidermis, but not with observations of other Crataegus species studied by Ganeva et al. (2009). Fig. 2. Map showing sites where tetraploid C douglasii (closed squares), diploid C. suksdorfii (open circles), and triploid C. gaylussacia and C. suksdorfii (open triangles) were collected for the geographic study (Table 1). Sites used in the population study are circled. 2014] MCGOEY ET AL.: PLOIDY LEVEL AND STOMATA SIZE IN CRATAEGUS 185 Fig. 3. Panels show Crataegus douglasii stomatal measurements: (A) Stomata width (i) and length (ii) were measured as the distances between the midpoints; (B) Crataegus douglasii stomata shown under 80 X magnification. Stomata density was estimated by counting the number of stomata within an areole, and dividing by the total area of the areole. Stomatal Size - Population-Level Study Variation in stomatal size. The stomata ranged in length from 21 pm to 54 pm and their total areas ranged from 264 pm^ to 1486 pm^. As expected, stomata from tetraploid individuals were the largest (mean length 36 pm, compared to the overall mean of 34 pm). These results are consistent with analyses on the other size measures (width, perimeter, and area). Diploids had slightly larger average sizes than triploids, which were the smallest of the three ploidy groups. This is surprising, since triploids did have higher DNA content then diploids, which is expected to be correlated with cell size. We tested for such a relationship between DNA content and stomata size within each ploidy level, but there were no significant correlations (all P > 0.05). The mean values for each leaf were determined and found to be normally distributed. Average stomata characteristics were consistently different between ploidy levels (all P values <0.001, see Fig. 4 for length), but not in a proportional manner (see below). Using an ANOVA to assess which factors affected stomatal characteristics, we found that both ploidy level and section of the leaf (Fig. 1) were significant (ploidy level and stomata length, P < 0.001; ploidy level and area, P < 0.001; leaf section and stomata length P < 0.001; leaf section and stomata area P < 0.001). The mixed effects model of all the individual stomatal measurements was consistent with the analysis using size means. When examining the length of stomata, we found that ploidy was a significant factor (P < 0.001), and that the location measured on the leaf was as well (P < 0.001). In addition there was a significant interaction between ploidy level and measured location (P = 0.0071). The same was true when examining stomata area, with both ploidy level (P < 0.001) and leaf section (P <0.001) being significant factors, the interaction term was also significant (P = 0.0124). Within tetraploids (C. douglasii, Ax C. suksdorfii; Fig. 4; Table 1), species was not a significant factor for stomata size (all P values >0.05). Stomatal Size - Geographic Study Variation in stomatal size. Stomata ranged in length from 15 pm to 57 pm, and the widths ranged from 9.8 pm to 38 pm. As was true in the population-level study, tetraploid C. douglasii had the largest stomata, and autotriploid C. suksdorfii had the smallest, although they were not significantly different from their diploid conspecifics (Fig. 5). Again, two analyses were conducted to avoid pseudoreplication. Tree means were used to examine the effect of ploidy on stomata length and width. Ploidy level was a significant factor for both length and width. The second analysis, using the full data set while accounting for pseudorep- lication within leaves and trees by including them as random factors, confirmed these findings, as the linear mixed effect models showed that ploidy was a significant factor for both stomata length (P < 0.001) and width (P < 0.001). Variation in stomatal density. Stomatal density ranged from 88 stomates/mm^ to 323 stomates/ 186 MADRONO [VoL 61 "T" 2x 1 i i 3x auto 3x ailo 4x 4x C. suksdorfii C. gaylussacia C. suksdorfii C. suksdorfii C. douglasii and C. suksdorfii Fig. 4. Box plot of size (length) of stomata for Crataegus taxa in the population study. The dark middle lines show the median values, and two ends of the box show the 1st and 3rd quartiles. The whiskers show the interquartile range multiplied by 1.5. All values outside this range (outliers) are shown as open circles. Both ploidy level and section of the leaf (Fig. 1) were significant (ploidy level P < 0.0001; leaf section P < 0.0001). mm^ (Fig. 5A). Note that the densities are reported here in stomates/mm^ for ease of interpretation, but the areoles were 0.16 mm^ on average. The C. douglasii had an average stomata density of 166, while the diploid and triploid C suksdorfii had average densities of 169 and 214 respectively. When all data were included in a simple ANOVA model, there was a significant difference between C. douglasii and C suksdorfii (P = 0.0215), but this difference was driven by the higher density in triploids, so that when tree and leaf were included as random effects, the species difference was no longer significant (P = 0.21). Likewise, an ANOVA on tree means showed that there was not a significant difference in stomata density between ploidy levels (P = 0.16) or species (P = 0.24). Environmental correlates. Comparisons be- tween sites occupied by tetraploid C douglasii and those occupied by C. suksdorfii (both diploids and autotriploids) showed the former to be cooler and more xeric (Figs. 6, 7). It is important to note that the environmental data were atmospheric, and may not be a precise reflection of water relations on the ground at each site. Nevertheless, except for the two Michigan sites for which hydrological conditions are unknown, each of the samples studied here are on flood plains, adjacent lakeshores, or on slopes down which water moves at or below the surface in response to local atmospheric condi- tions (Table 1). The results of the MANOVA analysis demonstrated a significant relationship between three climatological variables (annual precipitation, minimum and maximum tempera- ture) and the Crataegus species at a site (P = 0.0058). Fourth-corner analyses demonstrated a significant relationship between several environ- mental factors and some but not all of the features of the hawthorn stomata (Fig. 8). Stomatal length and width covaried significantly with latitude, precipitation, and temperature, although the direction of the relationship de- pended on the quarter of the year examined (Fig. 8). Ploidy level was positively associated with elevation and negatively associated with precipitation and temperature (Fig. 8). There were no significant relationships between stoma- tal density and any of the environmental factors (Fig. 8). 2014] MCGOEY ET AL.: PLOIDY LEVEL AND STOMATA SIZE IN CRATAEGUS 187 Fig. 5. Boxplot of stomata characters of diploid (2x) and autotriploid (3x) C suksdorfii and tetraploid (4x) C douglasii in the geographic study: (A) Density; (B) Length (pm); (C) Width (pm). The dark middle lines show the median values, and two ends of the box show the 1st and 3rd quartiles. The whiskers show the interquartile range multiplied by 1.5. All values outside this range (outliers) are shown as dark circles. Predicting Ploidy Level from Stomatal Measurements Data from the population-level study and the geographical one were pooled in order to depict the predictive relationship between nuclear DNA content and stomatal size. Plotting nuclear DNA content against stomatal length (Fig. 9) demonstrates the way in which varia- tion at the levels of leaf, individual, and sampling site vitiates any attempt to predict ploidy level from stomatal size: the ranges in size for each ploidy level overlap so much that any given size except the very smallest or the Fig. 6. Boxplot of climate summaries comparing the C. douglasii and C. suksdorfii sites (diploid and autotriploids pooled) used in the geographic study. Precipitation is the annual total (mm), maximum and minimum temperatures are annual averages. The dark middle lines show the median values, and two ends of the box show the 1st and 3rd quartiles. The whiskers show the interquartile range multiplied by 1.5. All values outside this range (outliers) are shown as dark circles. largest might correspond any one of the three ploidy levels studied here (Fig. 9). Similar relationships were observed by Marshall (1978) in Manitoba Crataegus chrysocarpa and C. succulenta. Although Marshall was probably incorrect in his inference that the smaller stomata of C. succulenta meant that this taxon is diploid (Ontario specimens have been shown to be triploid; Talent and Dickinson 2005), their size relative to those of C. chrysocarpa (tetraploid; Talent and Dickinson 2005) demonstrates the same trend we have observed. 188 MADRONO [Vol. 61 CD CM O CM I I Fig, 7. Principal component analysis of quarterly climate data of the Crataegus sites used in the geographic study (numbered as in Table 1). The four quarters are winter (Ql), spring (Q2), summer (Q3), and fall (Q4). Left panel: convex hulls enclose sites grouped by taxon and ploidy level present at each one. Right panel: PCA biplot with vectors representing contributions of quarterly climate variables. Discussion The taxonomic complexity of the genus Cra- taegus has long been recognized (Camp 1942), as has the connection between this complexity and the occurrence of gametophytic apomixis, hy- bridization, and polyploidy (Dickinson and Phipps 1985, 1986; Lo et aL 2010b), This connection makes knowing the ploidy level and breeding system critical in order to make clear taxonomic judgments regarding Crataegus spe- cies. Determining ploidy level from chromosome counts is difficult because the tissues in which meiotic metaphases can be accumulated are available for less than a week, annually, at any given locality. Flow cytometry is an extremely valuable tool when studying Crataegus, but this method works mainly on living tissue, making it nearly useless especially with older herbarium material (e.g., type specimens). The only excep- tion to this is the ability to obtain flow cytometric data from seeds (Table 1; Talent and Dickinson 2007). This approach has been shown to work on decades-old seeds, provided they have been stored in the cold (Talent personal communica- tion). The ability to use a phenotypic trait like stomata size as an indicator of ploidy level would greatly simplify the study of this complicated genus. Unfortunately, based on our results, variation in this trait is not consistently related to ploidy level. Although the largest stomata in our study were from the tetraploid C. douglasii trees, stomata from autotriploid C suksdorfii individuals were generally smaller than those of diploids (Fig. 5). In addition, there was wide variation in size within a ploidy level, and too much overlap between ploidy levels to make a comparison of average stomata size useful for assessing ploidy (Fig. 9). The relationship between genome and cell size is weaker in trees than in other plants (Beaulieu et al. 2008). The lack of a difference between C. suksdorfii cytotypes in stomata size may be a reflection of this. Stomata size is not solely controlled by genome size; environmental or physiological factors may select for smaller cell size, or specifically for smaller stomata. Note that we deliberately excluded allopolyploid C suks- dorfii from the geographic comparison in order to focus on the contrast between diploid and autotriploid C suksdorfii and tetraploid C douglasii (Fig. 5). There seemed to be more of a taxonomic grouping with respect to stomata size than a difference between ploidy levels (Fig. 5). The difference in stomata size between C douglasii and C suksdorfii could prove useful in determin- ing the identity of previously collected specimens. This might be of use in the case of sterile specimens, where per flower stamen numbers are unavailable. Other factors affect stomata size, and the relationship between environmental 2014] MCGOEY ET AL.: PLOIDY LEVEL AND STOMATA SIZE IN CRATAEGUS 189 0 c o X 03 c o X CO c o X as c o X CO c o "O 0 E E 0 E "e 0 E "e 0 E E 0 > h“ h" CL h- 1“ CL h— h— CL H- K a. _J 0 T— cu CM CM CO CO CO LU 5 5 o a o o O O a o o O Fig. 8. Associations between Crataegus stomatal characteristics and ploidy level, and the geographic and climatic variables at the sites from which they were collected, as calculated by a fourth-corner analysis (Legendre et al. 1997). Significant positive associations are shown in black, significant negative associations are in grey and non- significant associations are in white. Latitude values were collated from the herbarium specimens used in the geographic study. Elevation values were gathered from Google Earth. Monthy averages for maximum temperature, minimum temperature, and precipitation were acquired using the PRISM database and used to calculate quarterly values. The four quarters are winter (Ql), spring (Q2), summer (Q3), and fall (Q4). characteristics and stomata traits illustrate such an example. For C douglasii and C. suksdorfiU spring and summer precipitation averages are positively correlated with stomata size (Figs. 7, 8). There are a limited number of triploid sites, and there is much environmental heterogeneity between those in California (presumptively auto- triploid C. gaylussacia) and those in Oregon (autotriploid C suksdorfii). However, these sites are representative of the known, and quite limited, distribution of C. gaylussacia on the one hand, and of autotriploid C. suksdorfii on the other. Despite large strides in polyploidy research, we still know very little about the ecological implications of polyploidy (Soltis et al. 2010; Manzaneda et al. 2012). Past researchers have hypothesized that genome duplication confers advantages that lead to an increased or novel range for the polyploid species. However, little empirical work has been done to test this idea (Soltis et al. 2010). Stebbins addressed the uncertainty surrounding the effect of polyploidy on species ranges decades ago and suggested that clarification could be gained by carefully studying the morphological traits and habitats of increas- ing numbers of species groups (Stebbins 1950). In the years since, many different taxonomic groups have been examined. This research has demon- strated that polyploids often tolerate more extreme conditions than their diploid relatives (Lewis 1980), and one ecological difference that does seem common across many taxa is that plants with higher ploidy levels are found in more xeric conditions than their diploid relatives (Levin 2002). In our study group, tetraploid C. douglasii is found in more xeric conditions (in terms of mean annual precipitation) than diploid and triploid C. suksdorfii (Fig. 6). This is consistent with studies on other plants, where taxa with higher ploidy levels grow in drier conditions than their lower ploidy-level relatives, including in Achillea L. (Ramsey 2011), Chamer- ion (Raf.) Raf. ex Holub (Maherali et al. 2009), 190 MADRONO [Vol. 61 Fig. 9. Nuclear DNA content and stomatal length plotted to show why stomatal size cannot be used as a predictor of ploidy level given the variation between leaves, individuals, and sites using data from the combined population level (leaf region 4 only) and geographic samples. Tetraploid C. douglasii, filled squares. Open symbols, C suksdorfii: circles, diploid; triangles, triploid. and Brachypodium P. Beauv. (Manzaneda et al. 2012). Sites at which C. douglasii occurs are also, on the average, colder than those occupied by C suksdorfii (Fig. 6). When the climate data are broken down into quarterly values, we see that while there is some overlap between the environ- ments in which diploid C. suksdorfii and tetra- ploid C douglasii occur, the tetraploid occupies a much wider range of environments than does the diploid (Fig. 7). Combined with the variation in breeding systems, this pattern of apomictic polyploids distributed over a wider range of environments than closely related sexual diploids (hence exhibiting a larger geographic range; Fig. 2) is what is referred to as geographical parthenogenesis (Horandl 2006; Lo et al. 2013). As seen in our study, a higher ploidy level tends to lead to larger cells and stomata. In many taxa this increase in stomata size is accompanied by a significant decrease in stomata density. Never- theless, by using the fourth-corner method to infer the association between nuclear DNA content and features of stomata, and quarterly climate parameters and elevation, we have shown that stomata density does not vary significantly in relation to any of the environmental factors we examined (Fig. 8). In other groups it does not seem advantageous to have larger stomata in more xeric conditions, but in many cases, the decreased density more than compensates, so that the total pore space on a leaf is lower in higher cytotypes (Levin 2002). Polyploidy may confer other advantages on plants that enable them to tolerate dry environments. Polyploids tend to grow more slowly than diploid congenerics, which could be an advantage when resources are scarce (Grant 1981; Levin 1983; Deng et al. 2011). In addition, some polyploids have in- creased leaf water content and transpiration efficiency (Chen and Tang 1945; Tal and Gardi 1976). In Crataegus, allopolyploid C. suksdorfii occurs in drier areas, where no diploid popula- tions occur. Perhaps this is due to the contribu- tion from the C. douglasii genomes in these individuals. It is also important to consider effects separate from cytotype; genetic divergence driven by natural selection will occur between lineages after a polyploidization event. More research on the magnitude and mechanism for increased tolerance to dry environments in polyploids is needed. As noted above, environmental descriptors like “drier” or “wetter,” not to mention the precipi- tation data used in our analyses, relate to atmospheric conditions. In fact all of our sites except possibly Hillsboro, Washington, and the two Michigan ones, are within a kilometer (usually much less) of at least seasonally running or standing water, and the topography of many of these sites is such that ground water likely flows through the root zone most or all of the year (Table 1; e.g., Curtis 1986). Hawthorns are frequently found in riparian zones and this fact may help to explain the lack of correlation between stomatal size and growing season pre- cipitation (Fig. 8). Figure 8 also documents the way in which diploid C suksdorfii and autotri- ploid C. gaylussacia are restricted to a region of Mediterranean climate (high winter precipitation, summer drought) where stomatal size is negative- ly correlated with Q2 temperatures, and positively correlated with Q3 precipitation. Conclusions The interplay between cytotype, physiology and ecology is complex. Our study demonstrates that neither stomatal density nor stomatal size can be used to determine ploidy level in the Crataegus series Douglasianae, However, stomata size may be useful in differentiating between C. douglasii and C suksdorfii. The two species occur in distinctive environments, and some climato- logical (precipitation, temperature) and geo- graphic traits (latitude) are significantly related to stomata size. Polyploidy is a unique mecha- nism of speciation in its immediacy, and its propensity to cause speciation in sympatry. The ecological implications of polyploidization con- tinue to be an important avenue of research. Acknowledgments We are greatly indebted to Eugenia Lo and especially Nadia Talent for sharing flow cytometry data with us; 2014] MCGOEY ET AL.: PLOIDY LEVEL AND STOMATA SIZE IN CRATAEGUS 191 also to Nadia Talent for advice on nomenclature. We also thank Rebecca Dotterer, Rhoda Love, and Peter Zika for plant collections and collecting site informa- tion, and Annabel Drover, Jenny Bull, Cheying Ng, and the many Green Plant Herbarium volunteers and 2011 summer staff for helping to organize and database the vouchers for this study. The late Steve Brunsfeld suggested several of our Idaho study sites to TAD. We are grateful to Doreen Smith, Walter Earle, Margaret Graham, Bill Jensen, Gene Cooley, Rocky Thompson, John Herrick, Ralph Johnson, and Marcia Johnson, for information about and access to Califor- nia hawthorn collecting sites. The Siskiyou Field Institute & Deer Creek Center, Selma, Oregon, kindly permitted us to collect on their property. Data on pollen size and stainability were collected by undergraduate project students Yi-Min Chun, Jason Lee, Jennifer Shiller, and Jurgena Tusha. Adam Dickinson did the initial GIS comparison of climate variables at C. dougiasii and C suksdorfii collecting sites. Many thanks to the Royal Botanical Gardens Science Department, especially Natalie Iwanycki, for their support of this project. Financial support from a Natural Sciences and Engineering Research Council of Canada Undergrad- uate Student Research Award to KC and Discovery Grant A3430 to TAD, the Department of Ecology & Evolutionary Biology of the University of Toronto, and the Royal Ontario Museum is gratefully acknowledged. Literature Cited Angert, A. L. AND D. W. SCHEMSKE. 2005. The evolution of species’ distributions: reciprocal trans- plants across the elevation ranges of Mirnulus cardinalis and M. lewisii. Evolution 59:1671-1684. Beaulieu, J. M., I. J. Leitch, S. Patel, A. Pendharkar, and C. a. Knight. 2008. Genome size is a strong predictor of cell size and stomatal density in angiosperms. New Phytologist 179: 975-986. Brown, J. H., G. C. Stevens, and D. M. Kaufman. 1996. The geographic range: size, shape, boundar- ies, and internal structure. Annual Review of Ecology and Systematics 27:597-623. Brunsfeld, S. J. and F. D. Johnson. 1990. Cytolog- ical, morphological, ecological and phenological support for specific status of Crataegus suksdorfii (Rosaceae). Madrono 37:274-282. Buechler, W. K. 2000. Estimating polyploidy levels using cell size in Salix leaves. American Journal of Botany 87:238. Camp, W, 1942. The Crataegus problem. Castanea 7:51-55. Chen, G., W.-B. Sun, and H. Sun. 2009. Morpho- logical characteristics of leaf epidermis and size variation of leaf, flower and fruit in different ploidy levels in Buddleja macrostachya (Buddlejaceae). Journal of Systematics and Evolution 47:231-236. Chen, S.-L. and P. S. Tang. 1945. Studies on colchicine-induced autotetraploid barley. III. Phys- iological studies. American Journal of Botany 32:177-179. Coughlan, j, 2012. Mechanisms for range size and distribution variation in the polyploid complex black-fruited hawthorn {Crataegus series Dougla- sianae): biogeographic implications for the mainte- nance of cytotype diversity. M.Sc. thesis, Univer- sity of Toronto, Canada. Coyne, J. A. and H. A. Orr. 2004. Speciation. Sinauer Associates, Sunderland, MA. Curtis, A. B. 1986. Upper Elk Meadows Research Natural Area. USD A, Forest Service, Pacific North- west Forest and Range Experiment Station 1-20. Deng, B., W. Du, C. Liu, W. Sun, S. Tian, and H. Dong. 2011. Antioxidant response to drought, cold and nutrient stress in two ploidy levels of tobacco plants: low resource requirement confers polytolerance in polyploids? Plant Growth Regu- lation 66:37-47. Dickinson, T., S. Belaoussoff, R. Love, and M. Muniyamma. 1996. North American black-fruited hawthorns: I. Variation in floral construction, breeding system correlates, and their possible evolutionary significance in Crataegus sect. Dou- glasii Loudon. Folia Geobotanica et Phytotaxono- mica 31:355-371. Dickinson, T. A. and J. Phipps. 1985. Studies in Crataegus L. (Rosaceae: Maloideae). XIIL Degree and pattern of phenotypic variation in Crataegus sect. Crus-galli in Ontario. Systematic Botany 10:322-337. and . 1986. Studies in Crataegus (Rosaceae: Maloideae) XIV. The breeding system of Crataegus crus-galli sensu lato in Ontario. American Journal of Botany 73:116-130. , E. Y. Y. Lo, N. Talent, and R. M. Love. 2008. Black-fruited hawthorns of western North America — one or more agamic complexes? Botany 86:846-865. Dray, S. and P. Legendre. 2008. Testing the species traits-environment relationships: the fourth-corner problem revisited. Ecology 89:3400-3412. Ganeva, T., K. Uzunova, and D. Koleva. 2009. Comparative leaf epidermis investigation in species of genus Crataegus L. (Rosaceae) from Bulgaria. Feddes Repertoriiim 120:169-184. Grant, V. 1981. Plant Speciation, 2nd ed. Columbia University Press, New York, NY. Hadley, A. 2010. Combine Z. http://www.hadieyweb. pwp.blueyonder.co.uk/CZP/News.htm Hargrove, W. and R. Luxmoore (eds.). 1998. A new high-resolution national map of vegetation eco- regions produced empirically using multivariate spatial clustering. Oak Ridge National Laboratory. Website http://www.geobabble.org/~hnw/esri98/ (accessed August 2011). Horandl, E. 2006. The complex causality of geo- graphical parthenogenesis. New Phytologist 171:525-538. . 2008. Evolutionary implications of self-com- patibility and reproductive fitness in the apomictic Ranunculus auricomus polyploid complex (Ranun- culaceae). International Journal of Plant Sciences 169:1219-1228. Joachimiak, a. and a. Grabowska-Joachimiak. 2000. Stomatal cell length and ploidy level in four taxa belonging to the Phleum sect. Phleum. Acta Biologica Cracoviensia Series Botanica 42:103-107. JOLY, S. AND A. Bruneau. 2007. Delimiting species boundaries in Rosa sect. Cinnamomeae (Rosaceae) in eastern North America. Systematic Botany 32:819-836. Krebs, C. 2001. Ecology: the experimental analysis of distribution and abundance. Benjamin Cummings, San Francisco, CA. 192 MADRONO [VoL 61 Legendre, P., R. Galzin, and M. L. Harmelin- Vivien. 1997. Relating behavior to habitat: solu- tions to the fourth-corner problem. Ecology 78:547-562. Levin, D. A. 1983. Polyploidy and novelty in flowering plants. American Naturalist 122:1-25. . 2002. The role of chromosomal change in plant evolution. Oxford University Press, New York, NY. Lewis, W. H. 1980. Polyploidy in species populations. Pp. 103-144 in W. H. Lewis (ed.), Polyploidy: biological relevance. Proceedings of the Interna- tional Conference on Polyploidy, Biological Rele- vance, held at Washington University, St. Louis, Missouri, May 24—27, 1979. Plenum Publishing Corp., New York, NY. Li, W., G. Berlyn, and P. Ashton. 1996. Polyploids and their structural and physiological characteris- tics relative to water deficit in Betida papyrifera (Betulaceae). American Journal of Botany 83: 15-20. Lo, E. Y. Y., S. Stefanovic, and T. A. Dickinson. 2007. Molecular reappraisal of relationships be- tween Crataegus and Mespihis (Rosaceae, Pyr- eae) — Two genera or one? Systematic Botany 32:59^616. , , AND . 2009. Population genetic structure of diploid sexual and polyploid apomictic hawthorns {Crataegus; Rosaceae) in the Pacific Northwest. Molecular Ecology 18:1145-1160. , , AND . 2010a. Reconstructing reticulation history in a phylogenetic framework and the potential of allopatric speciation driven by polyploidy in an agamic complex in Crataegus (Rosaceae). Evolution 64:3593-3608. , , AND . 2013. Geographical parthenogenesis in Pacific Northwest hawthorns {Crataegus; Rosaceae). Botany 91:107-116. , , K. Ritland, and T. a. Dickinson. 2010b. Fine-scale comparisons of genetic variabil- ity in seed families of asexually and sexually reproducing Crataegus (Hawthorn; Rosaceae). American Journal of Botany 97:1014-1024. Love, R. M. and M. Feigen. 1978. Interspecific hybridization between native and naturalized Crataegus (Rosaceae) in western Oregon. Madrono 25:211-217. Lowry, E. and S. E. Lester. 2006. The biogeography of plant reproduction: potential determinants of species’ range sizes. Journal of Biogeography 33: 1975-1982. Maherali, H., a. E. Walden, and B. C. Husband. 2009. Genome duplication and the evolution of physiological responses to water stress. New Phytologist 184:721-731. Manzaneda, a., P. Rey, J. Bastida, C. Weiss- Lehman, E. Raskin, and T. Mitchell-Olds. 2012. Environmental aridity is associated with cytotype segregation and polyploidy occurrence in Brachypodium distachyon (Poaceae). New Phytolo- gist 193:797-805. Marshall, H. H. 1978. Identification of Crataegus species native to Manitoba. The Canadian Field- Naturalist 92:321-326. Martin, S. L. and B. C. Husband. 2009. Influence of phylogeny and ploidy on species ranges of North American angiosperms. Journal of Ecology 97: 913-922. McIntyre, P. J. 2012. Polyploidy associated with altered and broader ecological niches in the Claytonia perfoliata (Portulacaceae) species com- plex. American Journal of Botany 99:655-662. McNeill, J., F. R. Barrie, W. R. Buck, V. Demoulin, W. Greuter, D. L. Hawksworth, P. S. Herendeen, S. Knapp, K. Marhold, J. Prado, W. F. Prud’Homme Van Reine, G. F. Smith, J. H. WiERSEMA, AND N. J. TuRLAND. 2012. Internation- al Code of Nomenclature for algae, fungi, and plants (Melbourne Code) adopted by the Eighteenth International Botanical Congress Melbourne, Aus- tralia, July 2011. Regnum Vegetabile 154. Koeltz Scientific Books, Koenigstein, Germany. Meacham, C. AND T. Duncan. 1991. MorphoSys: an automated morphometric system. Version 1.29. University of California, Berkeley, Regents of the University of California, Berkeley, CA. Metcalfe, C. and L. Chalk. 1979. Anatomy of the dicotyledons, 2nd ed. Clarendon Press, Oxford, U.K. Nogler, G. a. 1984. Gametophytic apomixis. Pp. 475- 518 in B. M. Johri (ed.). Embryology of angio- sperms. Springer- Verlag, Berlin. Otto, S. P. 2007. The evolutionary consequences of polyploidy. Cell 131:452^62. Parisod, C., R. Holderegger, and C. Brochmann. 2010. Evolutionary consequences of autopolyploi- dy. New Phytologist 186:5-17. Phipps, J. 2012. Crataegus, in Jepson Flora Project (eds.), Jepson eFlora. Website http://ucjeps.berkeley. edu/cgi-bin/get_IJM.pl?tid=91319 (accessed on 9 June 2013). AND R. O’Kennon. 1998. Three new species of Crataegus (Rosaceae) from western North Amer- ica: C. okennonii, C. okanagenensis, and C. phippsii. Sida 18:169-191. AND . 2002. New taxa of Crataegus (Rosaceae) from the northern Okanagan-south- western Shuswap diversity center. Sida 31:355-371. Phipps, J. B., K. R. Robertson, P. G. Smith, and J. R. Rohrer. 1990. A checklist of the subfamily Maloideae (Rosaceae). Canadian Journal of Bota- ny 68:2209-2269. Potter, D., T. Eriksson, R. C. Evans, S. Oh, J. E. E. Smedmark, D. R. Morgan, M. Kerr, K. R. Robertson, M. Arsenault, T. A. Dickinson, AND C. S. Campbell. 2007. Phylogeny and classification of Rosaceae. Plant Systematics and Evolution 266:5^3. PRISM Climate Group. 2012. PRISM Data Explor- er. Oregon State University. Website http://www. prism.oregonstate.edu/ (accessed August 2011). R Development Core Team. 2004. R: A language and environment for statistical computing. R Foundation for Statistical Computing. Website http://www.r-project.org/ (accessed 15 April 2011). Ramsey, J. 2011. Polyploidy and ecological adaptation in wild yarrow. Proceedings of the National Academy of Sciences of the United States of America 108:7096-7101. Saltonstall, K., K. Glennon, a. Burnett, R. B. Hunter, and K. L. Hunter. 2007. Comparison of morphological variation indicative of ploidy level in Phragmites australis (Poaceae) from eastern North America. Rhodora 109:415-429. Shorthouse, D. P. 2010. SimpleMappr, an online tool to produce publication-quality point maps. Web- 2014] MCGOEY ET AL.: PLOIDY LEVEL AND STOMATA SIZE IN CRATAEGUS 193 site http://www.simplemappr.net (accessed on 9 June 2013). Soltis, D. E., R. J. A. Buggs, J. J. Doyle, and P. S. Soltis. 2010. What we still don’t know about polyploidy. Taxon 59:1387-1403. Stebbins, G. L. 1950. Variation and evolution in plants. Columbia University Press, New York, NY. Tal, M. and I. Gardi. 1976. Physiology of polyploid plants: water balance in autotetraploid and diploid tomato under low and high salinity. Physiologia Plantarum 38:257-261. Talent, N. and T. A. Dickinson. 2005. Polyploidy in Crataegus and Mespilus (Rosaceae, Maloideae): evolutionary inferences from flow cytometry of nuclear DNA amounts. Canadian Journal of Botany 83:1268-1304. and . 2007. Endosperm formation in aposporous Crataegus (Rosaceae, Spiraeoideae, tribe Pyreae): parallels to Ranunculaceae and Poaceae. New Phytologist 173:231-249. VASCAN. 2013. Crataegus gaylussacia A. Heller. Database of Vascular Plants of Canada (VAS- CAN). Website http://data.canadensys.net/vascan/ taxon/20019 (accessed on 7 June 2013). Wang, H. and J. Clarke. 1993. Genotypic, intra- plant, and environmental variation in stomatal frequency and size in wheat. Canadian Journal of Plant Science 73:671-678. Zarrei, M., N. Talent, J. Coughlan, S. Stefanovic, and T. a. Dickinson. 2012. Nuclear ribosomal ITS2 DNA sequences illuminate hybridization in a diploid- polyploid agamic complex of Crataegus (Rosaceae). Presented at Botany 2012, Columbus, OH. Madrono, VoL 61, No. 2, pp. 194-200, 2014 MORPHOLOGY AND DEVELOPMENT OF SUNKEN TERMINAL CEPHALIUM IN DISCOCACTUS (CACTACEAE) Root Gorelick Department of Biology, School of Mathematics and Statistics, and Institute of Interdisciplinary Studies, Carleton University, 1125 Colonel By Drive, Ottawa, Ontario, Canada, KIS 5B6 Root.Gorelick@carleton.ca Abstract Growth of photosynthetic portions of Discocactus shoots is seemingly not suppressed by cephalium formation, as vascular traces are prominent between the vascular cylinder and the circular juncture of cephalium and juvenile growth. Neither of these morphological traits has previously been documented in the Cactaceae. I therefore provide a pair of hypotheses consistent with these morphological traits and suggest ways to test these hypotheses, although do not test them myself. One hypothesis for these morphological traits is that the shoot vascular cylinder bifurcates when the cephalium first forms, with one cylindrical ‘branch’ of the vascular cambium terminating in the shoot apical meristem and the other concentric cylindrical ‘branch’ of the vascular cambium terminating in a circular meristem at the base of the terminal cephalium. A second hypothesis is that subapical development occurs very slowly surrounding a depressed shoot apical meristem in Discocactus. Vegetative portions of the shoot mature sufficiently slowly that the cephalium can be well formed even while juvenile areoles and photosynthetic internodes continue to grow and develop for several years after production by the juvenile phase of the shoot apical meristem. Key Words: Cephalium, Coieocephalocereiis, Discocactus, Espostoa, Melocactus, Podostemaceae, shoot apical meristem, vascular cambium. Several species of the family Cactaceae possess highly modified reproductive structures known as terminal cephalia. In these species, a single shoot apical meristem begins life as a normal- looking photosynthetic cactus, but eventually undergoes a juvenile/adult transition to form a non-photosynthetic apical portion of the shoot, called a terminal cephalium (the adjective ‘terminal’ is a misnomer and should probably be called an ‘apical cephalium’ because the cephalium grows indefinitely from the shoot apical meristem). The cephalium contains con- tiguous axillary buds that produce huge num- bers of hairs, bristles and spines, each of which are highly modified leaves. Terminal cephalia occur in two different clades of cacti, Pachycer- einae in the Core Cactoideae I and the Browningieae-Cereeae-Trichocereeae clade in the Core Cactoideae II (Hernandez-Hernandez et al. 2011) (although some botanists call the reproductive structures in Pachycereinae ‘pseu- docephalia’; see the discussion section regarding Cephalocereus Pfeiff. and see hypothesis two regarding Pachycereus Britton & Rose). In several taxa {Stephcmocereus leucostele A. Berger, Arrojctcloa Britton & Rose, and Cephalocereus apicicephalium E. Y. Dawson), the shoot apical meristem later reverts to producing the photo- synthetic juvenile morphology, thereby forming a series of ring-like terminal cephalia. However, in Melocactus Link & Otto, Discocactus Pfeiff, and Pachycereus militaris (Audot) D. R. Hunt, the conversion of the shoot apical meristem from juvenile to adult form is permanent: Once a terminal cephalium is formed, it continues to grow and produce flowers until the shoot or portion of the shoot dies (in Pachycereus militaris the portion of the shoot containing all of the cephalium and small part of the underlying photosynthetic tissue abscises and dies, but the remainder of the vegetative shoot below usually lives for many more years, forming new vegeta- tive and reproductive shoots from axillary branches; Mauseth et al. 2005). In a detailed review of anatomy of highly modified shoots in the Cactaceae, Mauseth (pp. 910, 2006) asserted that for Melocactus, Discocactus, and Pachycereus militaris, “The cephalium becomes longer every year, every year the juvenile portion merely becomes older - and it is the only photosynthetic tissue the plant has. Because the shoot is produced by one single SAM (shoot apical meristem) and does not branch, no new photosynthetic cortex can be added, so the ratio of photosynthetic tissue to heterotrophic tissue decreases every year.” If true, this would be especially problematic for Melocactus and Dis- cocactus because their terminal cephalia appear to completely lack chlorenchyma and stomata and because photosynthetic portions of their shoots only branch when seriously injured. In this paper, I show an exception to the claim of no new photosynthetic growth after cephalium formation for several species of Discocactus. I 2014] GORELICK: HYPOTHESES FOR SUNKEN MARGINS OF DISCOCACTUS CEPHALIA 195 Fig. la, b. Discocactus heptacanthus Britton & Rose subsp. catingicola (Buining & Brederoo) N.P. Taylor & Zappi (synonym D. catingicola Buining & Brederoo) near Porto Novo, Bahia, Brazil showing new photo- synthetic growth at the base of the terminal cephalium. The cephalium in Figure lb is about four to five times the diameter of the cephalium in Figure la. further discuss two hypotheses to explain this observation. Figure la shows a mature shoot of Discocactus heptacanthus Britton & Rose subsp. catingicola (Buining & Brederoo) N.P. Taylor & Zappi (synonym D. catingicola Buining & Brederoo) with a cephalium that is probably a few years old, but also with new photosynthetic tissue and cortex, including new axillary buds (areoles, with spines) arising from just below the cephalium. The brown color of the spines surrounding the base of the cephalium shows that these are probably new leaves on newly formed axillary buds. The beige colored trichomes in these axillary buds also indicate that this is new growth. Other plants of this subspecies {D. heptacanthus subsp. catingicola), with even larger cephalia, show this new growth after cephalium formation (Fig. lb). Buining' s (1980) monograph on Discocactus shows two taxa growing new vegetative tissue after cephalium formation, D. catingicola (pp. 179) and D. griseus Buining & Brederoo (pp. 175), both of which are usually considered synonyms of D. heptacanthus. New photosyn- Fig. 2. Discocactus placentiformis (synonyms D. latispinus Buining & Brederoo, D. crystallophilus Diers 6 Esteves) near Monjolos, Minas Gerais, Brazil with new vegetative spines surrounding the cephalium at the 7 o'clock position. thetic tissue also appears to be produced from the base of the cephalium in at least two other taxa, D. placentiformis K. Schum. (synonyms D. latispinus Buining & Brederoo, D. crystallophilus Diers & Esteves) and D. bahiensis Britton & Rose subsp. gracilis P.J. Braun & Esteves (Figs. 2^). The genus Discocactus was named for the accentuated disc-shaped form of the photosyn- thetic portion of the shoot. Discocactus photo- synthetic stems are proportionally much wider than those of Melocactus or Pachycereus mili- taris. This also explains how Discocactus shoots grow cephalia at fairly small sizes, but large diameter photosynthetic stems are often noted in the field. Disc-shaped photosynthetic juvenile stems are ubiquitous in the genus Discocactus, not just in D. heptacanthus, D. placentiformis, and D. bahiensis. Several of the ground-hugging (virtually geophytic) members of the genus, such as D. bahiensis subsp. gracilis (Fig. 4) and D. horstii Buining & Brederoo would probably perish without this ability to grow new photo- synthetic tissue as the older chlorenchyma gets trampled. Widening of the photosynthetic portion of the shoot after initial cephalium formation largely debunks the notion that newly formed spines at the base of the cephalium are the product of indeterminate growth of vegetative (juvenile) areoles. For example, Espostoa blossfeldiorum (Werdermann) Buxbaum grows a basal skirt of spines for many years after seedling germination. Cephalia too can have indeterminate growth of areoles, as seen in Espostoa lanata (Kunth) Britton & Rose and Coleocephalocereus goebelia- nus (Vaupel) Buining, which can both flower and grow axillary branches from areoles that are several years old (Go relick and Machado 2012; Gorelick 2014). Discocactus species, however, do more than grow new nodes (areoles) below the cephalium. They also grow new internodes. 196 MADRONO [Vol. 61 Fig. 3. Discocactus placentiformis in cultivation (dis- played as D. crystaUophihis). Not only are new axillary buds being produced at the base of the cephalium, but it appears that new chlorenchyma is also being produced, at least compared with the older reddish epidermis (photo credit: Geoff Stein). apparently with new chlorenchyma, and grow a substantially wider shoot. A radial section of a mature cultivated specimen of Discocactus zehntneri Britton & Rose subsp. araneispimis (Burning & Brederoo) P.J. Braun showed a remarkably sunken cephalium and vascular traces arising from the vascular cylinder to the epidermal juncture between juvenile growth and cephalium (Fig. 5). This is consistent with new photosynthetic growth and new juvenile areoles formed after the cephalium arose, but in two very different possible ways. Hypothesis 1 : Bifurcation of Vascular Cambium Forming Concentric Vascular Cambia I hypothesize bifurcation of the vascular cambium to form a second concentric cambium that temiinates as a circular meristem at the base of the cephalium. Such a circular meristem would cause the photosynthetic base of the shoot to grow wider over time, after cephalium formation. This circular meristem would make Discocactus photosynthetic stems proportionally much wider than those of Melocactus and Pachycereus militaris. This would also explain how Discocac- tus shoots grow cephalia at fairly small sizes, but large diameter photosynthetic shoots are often noted in the field. Given how short terminal cephalia are in all species of Discocactus, especially compared with the closely related genus Melocactus, a circular meristem in Dis- cocactus would probably only contribute new tissue to the photosynthetic (juvenile) portion of the shoot, while supplying little or no new tissue to the cephalium. Fig. 4. Discocactus bahiensis subsp. gracilis, near San Rafael, Bahia, Brazil with new vegetative spines surrounding the cephalium near the 12 o’clock position. Assuming this first hypothesis, the vascular cambium bifurcated from the primary vascular cambium at around the time the cephalium first formed, with one trace of the vascular cylinder (in radial section) extending to the shoot apical meristem and the other trace extending to the junction of the cephalium and vegetative tissue. Because the terminal end of this meristem is circular (Fig. 5e), not a point (as with the shoot apical meristem), tissues derived from this circular meristem would not form a closed vascular cylinder, but rather an open funnel- shaped structure. This can be easily envisioned by rotating the vascular traces to the cephalium about a vertical axis going through the center of the pith, with the widest part of the conical frustum being the circular meristem (Fig. 5e). This could not form a closed vascular cylinder because, if it did, the tissues would completely enclose the cephalium, precluding pollinators and seed dispersers from reaching the cephalium. The only situation possibly analogous with what I report here in Discocactus may occur in the aquatic family Podostemaceae, which have bizarre morphology (Eckardt and Baum 2010). For instance, Indotristicha ramosissima (Wight) P. Royen plants have a cup-like structure composed of parts that may be stems and/or leaves surrounding the terminal inflorescences (Rutishauser and Huber 1991) (it is not obvious what constitutes stems or leaves in the Podoste- maceae). Morphologies are often unusual in submerged aquatic angiosperms, at least when compared with their terrestrial relatives. Succulent plants are well known for adding extra layers of mitotically dividing cambial cells (Robert et al. 2011), some of which are unifacial, even including reverse cambia (Terrazas et al. 2011; Gorelick in press). But with these other taxa, concentric cambia arise from ground tissue (usually cortex), not from bifurcation of existing cambia. Making this first hypothesis less likely. 2014] GORELICK: HYPOTHESES FOR SUNKEN MARGINS OF DISCOCACTUS CEPHALIA 197 Fig. 5. Radial section of Discocactus zehntneri subsp. araneispinus in cultivation (—10 cm diameter). 5a. One section. 5b. The same section as Fig. 5a, but with lines drawn over vascular cylinder (black) and vascular traces to the cephalium (red). 5c. Facing section. 5d. The same facing section as Fig. 5c, but with lines drawn over vascular cylinder (black) and vascular traces to the cephalium (red). 5e. If a circular meristem exists at the base of the cephalium (hypothesis 1), then that meristem would be depicted by the dotted red circle. concentric vascular cambia have never been documented in the Cactaceae. Hypothesis 2: Slow Development of Juvenile Axillary Buds from Depressed Shoot Apical Meristem Even After Cephalium Formation The second hypothesis is that axillary buds in the juvenile (photosynthetic) stage can take several years to fully develop. The shoot apical meristem may produce nascent juvenile-stage axillary buds and photosynthetic internodes before initial cephalium formation, but these juvenile-stage axillary buds and photosynthetic internodes may not fully develop until long after the cephalium has formed. Discocactus is typical of the Cactaceae, especially the subfamily Cactoideae, in having a depressed shoot apical meristem, one that lies vertically below tissues that have recently developed from 198 MADRONO [Vol. 61 it. The resulting sunken apex is thought to protect mitotically actively dividing cells from the exigen- cies of desert life. However, if juvenile-stage development occurs as slowly as hypothesized above, then new development of areoles and photosynthetic internodes will look as though a circular meristem has developed at the base of the cephalium. Mature and almost mature areoles (axillary buds) occupy the rim of a crater that surrounds the sunken shoot apical meristem. Slowly developing juvenile tissues exit between this rim and the shoot apical meristem. However, with this hypothesis, development of reproductive tissue in the cephalium - which arises from the same shoot apical meristem - is faster than that of later developing portions of the juvenile tissues. The cephalium and shoot apical meristem eventually overtop the slowly developing late-formed juvenile tissues. This would explain the sunken look of the lateral edges of the cephalium in radial section (Fig. 6). Mauseth (personal communication; 20 August 2013) analogizes this to formation of the caldera of a strato-volcano, such as Mount St. Helens. Inside the main caldera, subsequent magma domes can form, which can and often do overtop the rim of the caldera. The deeply sunken lateral edges of the cepha- lium (Fig. 5) seem to be unique to Discocactus. While the phenomenon is not obvious from a casual look at a plant, but only visible with sectioning the shoot, a sunken cephalium has never been noted in radial sections of Melocactus nor Pachycereus {Backebergid) militaris (although I would say that Pachycereus militaris has a pseudocephalium because reproductive parts of its shoot contain stomata and parenchyma, but lack a narrow cork-laden cortex). I have never seen nor made radial sections of Arrojadoa nor Stephanocereus leucostele to see whether their terminal cephalia are sunken. This hypothesis of slow development of late- forming juvenile (photosynthetic) tissues is also consistent with the prominent vascular trace in Figure 5 from the vascular cylinder to the epidermal juncture between juvenile tissue and cephalium (red lines in Fig. 5b, d). Especially while the cephalium is young, this juncture should be densely packed with undeveloped axillary buds. It may not even be obvious without detailed microscope examination what consti- tutes individual axillary buds in this slowly developing annular mass. Each axillary bud is a short shoot and therefore will have a vascular trace going to it. A sufficiently dense mass of such axillary buds should have what appears to be a dense mass of vascular traces going to them, which may look like a single large vascular trace in the radial section (Fig. 5). These nascent developing axillary buds form a circular ring around the shoot apical meristem and cephalium. Hence a cross-section of the same shoot (which unfortunately I did not make) might show what looks like a circle of vascular traces, i.e., another vascular cylinder. If this hypothesis of slow development of late- forming juvenile tissues is true, growth of new photosynthetic tissue should only occur for a fixed number of years after cephalium formation. While cephalium formation may start when the shoot is roughly spherical and end with the photosynthetic portion of the shoot that is substantially wider than tall, this growth can only occur for a determinate number of years. Furthermore, there will be fewer juvenile axillary buds each year in a cephalium-bearing plant, meaning fewer vascular traces to the epidermal juncture between juvenile and cephalium portions of the shoot. Thus, the apparent singular vascular trace to this juncture (red lines in Fig. 5b, d) should become less visible over time, as the photosynthetic portion of the shoot becomes wider and more disc-shaped. This second hypothesis can be summarized as a heterochronic shift in development of the portion of Discocactus shoots that were epidermal mitotic products of the shoot apical meristem prior to cephalium formation, at which time these juvenile (vegetative) tissues develop slowly. Once a shoot transitions from juvenile to reproductive (cepha- lium) phases, epidermal mitotic products of the shoot apical meristem return to developing at a normal pace, as was found in younger juvenile stems in Discocactus. In other words, early juvenile (vegetative) tissues develop quickly, late juvenile (vegetative) tissues develop slowly, and reproductive (cephalium) tissues develop quickly. Discussion Although Discocactus cephalia have sunken margins, this sunken nature does not appear to be homologous to the sunken lateral cephalia of Espostoa Britton & Rose, Coleocephalocereus Backeberg, or other cacti with true lateral cephalia. Lateral pseudocephalia, as in Cephalo- cereus senilis Pfeiff., C. columna-trajani (Karw. ex Pfeiff.) P. V. Heath, and Micranthocereus streckeri Van Heek & Van Criek., are not sunken (Vazquez-Sanchez et al. 2005, 2007; Gorelick 2013). True lateral cephalia are sunken in a cleft within vegetative (photosynthetic) tissue, but that cleft is relatively flat (see Fig. 7 in Gorelick 2013). By contrast, Discocactus cephalia have margins that are substantially depressed into the vegeta- tive tissue. In radial sections of Discocactus, the boundary between vegetative and reproductive tissue is “W”-shaped, whereas in cross sections of lateral cephalia, the boundary between vegetative and reproductive tissues is “U”-shaped (Gorelick 2013). Developmentally the clefts of lateral cephalia are different from clefts of terminal (apical) cephalia of Discocactus insofar as they 2014] GORELICK; HYPOTHESES FOR SUNKEN MARGINS OF DISCOCACTUS CEPHALIA 199 grow in horizontal versus vertical orientations, respectively. I cannot yet discern which of the above two hypotheses explains the peculiar morphology of Discocactus cephalia, with their sunken margins, nor can we discern whether all Discocactus species display this morphology, which is very different from that of Melocactus. Drastic hetero- chronic shifts (hypothesis 2) are more likely than concentric vascular cambia produced by a branching cambium (hypothesis 1), but neither hypothesis has been tested. Some tests of the two hypotheses should include the following. First, Discocactus development should be monitored over many years to determine whether new vegetative tissue is created indefinitely after cephalium formation (hypothesis 1; concentric vascular cambium) or stops after several years (hypothesis 2; Mount St Helens). Second, de^ tailed anatomical studies should be carried out to follow the vascular traces going from the primary vascular cylinder to the epidermal juncture between the cephalium and photo synthetic parts, using phloroglucinol (1,3,5-benzenetriol) and hydrochloric acid to follow lignified xylem. It might thus be possible to discern whether there is a bifurcated vascular cambium (hypothesis 1) or merely a dense mass of vascular traces to not yet fully developed axillary buds (hypothesis 2). Third, a cross section through shoot that goes through the sunken margins of the cephalium might reveal whether the vascular traces going from the vascular cylinder to the sunken margins form a complete circle or set of discrete bundles. While this cross-section will not necessarily support or reject the first hypothesis because the hypothesized nascent concentric vascular cambi- um could have inter-fascicular parts, it would provide additional morphological details. Conclusion In some (possibly all) Discocactus species, vegetative shoots continue growing after cepha- lium formation. To do this, Discocactus cephalia have a peculiar morphology that has not previously been documented in cacti. Their sunken cephalium appears in radial section as a “W”-shaped border between photosynthetic and reproductive portions of the shoot. Two hypoth- eses are consistent with these morphological traits: (1) bifurcation of the vascular cylinder resulting in concentric vascular cambia with a circular meristem at the base of the cephalium and (2) heterochronic shifts that cause later development of the photosynthetic portion of the shoot to proceed much more slowly than either early development of the photosynthetic portion of the shoot or development of the reproductive portions of the shoot. Given that heterochronic shifts are far more common than circular meristems in all plants and concentric vascular cambia have never been documented in cacti, the second hypothesis is ceteris paribus more likely, but no data yet exists to test which (if either) hypothesis is correct. Acknowledgments Thanks to the Natural Sciences and Engineer- ing Research Council of Canada (NSERC) for funding, Geoff Stein for the photo of his Discocactus placentiformis in cultivation, Jim Mauseth for numerous helpful comments (espe- cially for the second hypothesis, the “Mount St Helens hypothesis”), and an anonymous reviewer. Literature Cited Buining, a. F. H. 1980. The genus Discocactus Pfeiffer: a revision of known, and description of new, species. Succulenta, Nederlands. Eckardt, N. a. and D. Baum. 2010. The podostemad puzzle: the evolution of unusual morphology in the Podostemaceae. Plant Cell 22:2104. Gorelick, R. 2013. Coleocephalocereus purpureus has a cephalium; Micranthocereus streckeri has a pseudocephalium (Cereeae, Cactoideae, Cacta- ceae). Bradleya 31:142-149. . 2014. Axillary branching of lateral cephalia in Cactaceae is not constrained by tilting of shoot apices. Haseltonia 19:13-16. . In Press. Strange stem architecture of pachy- caulous morning glories. Cactus and Succulent Journal. AND M. Machado. 2012. Axillary branching of lateral cephalia of Coleocephalocereus (Cactaceae). Haseltonia 17:35-41. HernAndez-HernAndez, T., H. M. HernAndez, J. A. De-Nova, R. Puente, L. E. Eguiarte, and S. Magallon. 2011. Phylogenetic relation- ships and evolution of growth form in Cactaceae (Caryophyllales, Eudicotyledoneae). American Journal of Botany 98:44-61. Mauseth, J. D. 2006. Structure-function relationships in highly modified shoots of Cactaceae. Annals of Botany 98:901-926. , T. Terrazas, M. VAzquez-SAnchez, and S. Arias. 2005. Field observations on Backebergia and other cacti from Balsas Basin, Mexico. Cactus and Succulent Journal 77:132-143. Robert, E. M. R., N. Schmitz, I. Boeren, T. Driessens, K. Herremans, J. De Mey, E Van de Casteele, H. Beeckman, and N. Koedam. 2011. Successive cambia: a developmental oddity or an adaptive structure? PLoS ONE 6:el6558, doi: 10. 1 37 1/journal. pone. 001 6558. Rutishauser, R. and K. A. Huber. 1991. The developmental morphology of Indotristicha ramo- sissima (Podostemaceae, Tristichoideae). Plant Sys- tematics and Evolution 178:195-223. Terrazas, T., S. Aguilar-Rodriguez, and C. T. Ojanguren. 2011. Development of successive cambia, cambial activity, and their relationship to physiological traits in Ipomoea arborescens (Con- volvulaceae) seedlings. American Journal of Bota- ny 98:765-774. 200 MADRONO [VoL 61 VAzquez-SAnchez, M., T. Terrazas, and S. Arias. 2005. Morfologia y anatomia del cefalio de Cephalocereus senilis (Cactaceae). Anales del Jardin Botanico de Madrid 62:153-161. . 2007. Morphology and anatomy of the Cephalocereus columna-trajani cephalium: why tilting? Plant Systematics and Evolution 265:87- 99. Madrono, VoL 61, No. 2, pp. 201-217, 2014 FIRE EFFECTS IN A MONTANE WETLAND, CENTRAL CASCADE RANGE, OREGON John A. Christy Oregon Biodiversity Information Center, Institute for Natural Resources, Portland State University-INR, P.O. Box 751, Portland, OR 97207-0751 john.christy@pdx.edu Cynthia N. McCain Willamette and Siuslaw National Forests (retired), P.O. Box 1 148, Corvallis, OR 97339 Sarah E. Greene Corvallis Forestry Sciences Laboratory (retired), 3200 SW Jefferson Way, Corvallis, OR 97331 Jennifer D. Lippert Willamette National Forest, 211 E. 7th Ave., Eugene, OR 97401 Abstract The stand-replacing Charlton, OR wildfire of 1996 burned several vegetation transects established before the fire, providing an opportunity to document recovery in one upland and five wetland vegetation zones over a 12-year period. Total mean percent cover of vegetation recovered to about 48% of pre-burn levels by year 12. Total cover increased slightly in Carex fen, but decreased in the other five zones. Fire damage was inversely proportional to moisture gradient, and recovery more or less directly proportional to the gradient. Upland Tsuga mertensiana forest sustained the greatest damage and showed the least recovery by year 12, followed by seasonally-flooded Poly trichum commune moss bed, Xerophyllum tenax ecotone, Vaccinium uliginosum shrub swamp, grass-dominated fen, and Carex fen. A transient soil crust of early-seral bryophytes occurred in upland Tsuga mertensiana forest. Cover of shrub, forb, grass, and moss layers declined in vegetation zones at the drier end of the moisture gradient, but increased in grass and sedge-dominated zones at the wetter end of the gradient, at the expense of forb, grass, and sedge layers. Key Words: Cascade Range, fen, fire, peatland, vegetation, wetland. Evidence of fire in wetlands of the Pacific Northwest is widespread, but few regional studies have examined the role of fire in these habitats. Hansen (1941, 1944) and Torgerson et al. (1949) reported charcoal and burned peat at depths of 1-2 m in organic soils along the coasts of Oregon and Washington. Martin and Fren- kel (1978) and Kunze (1994) reported evidence of fire in fens along the coast and in the southern Puget Trough, and charred snags are present in peatlands at all elevations (Christy, unpublished data). Heinselman (1963) and others clearly demonstrated how fire can regulate vegetation succession in wetlands, but information about its effect in wetlands of the Pacific Northwest is sparse. Repeat photogra- phy and transect studies indicate that in the absence of fire, both cover and diversity of herbaceous peatland species decline with in- creasing competition from woody vegetation, as these habitats convert to shrub swamp and forested wetland (Schultz 1989; Guerrant et al. 1998; Hebda et al. 2000; Christy 2005a; Cramer 2005; Ratchford et al. 2005; Sanders et al. 2007; Tolman 2006, 2007). Frenkel at al. (1986) described fen vegetation sampled in 1976 at the Torrey-Charlton Research Natural Area (RNA) in the central Cascade Range of Oregon. As part of the USD A Forest Service Pacific Northwest Region’s RNA base- line monitoring program, ecologists resampled Frenkel’s transect and established three others in 1993 to document ecotone conditions across wetlands. In 1996 the stand-replacing Charlton wildfire, part of the Moolack Fire Complex, burned 37.7 km^ surrounding the study site (Fig. 1). The fire killed more than 95 percent of the forest over about three-quarters of the burn area (Gardner and Whitlock 2001; Acker et al. 2013), and many areas burned to mineral soil. Wetlands in the RNA sustained surface damage, but peat did not burn because it was wet enough to prevent ignition. After the fire, the four transects were reestablished and resampled peri- odically over the next 12 yr. Although fen vegetation in a number of other montane wetlands in Oregon has been documented in detail (Seyer 1979, 1981, 1983; Wilson 1986; Titus and Christy 1997; Christy and Titus 1998; Murray 2000; Christy 2001, 2003, 2005b, Greene 202 MADRONO [Vol. 61 Plot locations, Torrey -Charlton RNA, and 1996 Charlton Burn Fig. 1. Torrey Lake plots (white dots), Torrey-Charlton RNA boundary (black line), and extent of 1996 Charlton Burn (light-colored area). and Schuller, unpublished data), none have pre- and post-burn data available. Site-specific data on pre-disturbance vegetation, intensity of disturbance, and condition of survivors can facilitate interpretation and prediction of community response to disturbance (Attiwill 1994; Turner et al. 1997; Zobel and Antos 1997). In contrast to charcoal, pollen, and carbon storage analyses that document long-term changes wrought by fire and climate change, the study at Torrey-Charlton RNA provides scarce short-term pre- and post-burn data for a fen in the Pacific Northwest. Although a single-disturbance study has obvious limitations because of pseudo-replica- tion (Turner et al. 1997), fire effects observed at the RNA will facilitate further studies in the region. Study Site Frenkel et al. (1986) described details of the study site, located at 43°47'46"N, 122°00'36"W. The Torrey-Charlton RNA was formally estab- lished by the USDA Forest Service in July 1998 (Salix Associates 1998), with the study area located in the Torrey Lake Unit of the RNA (Fig. 1). The site is located entirely within the Waldo Lake Wilderness Area, and is surrounded by other wilderness and roadless areas. Hydrology Snowpack, summer precipitation, and temper- ature varied over the 12-year sampling period, typical for montane wetlands in the Cascade Range. Carex fen, normally the wettest vegeta- tion zone sampled at the site, exhibited varying amounts of bare ground, muck, or water in August and September, the driest part of the sampling years. Bare ground, muck, and water covered 58% of the plots in 1993, 96% in 1997, and 78% in 2001, consistent with precipitation recorded at Oakridge, Oregon, about 38 km WSW of the study area (Western Regional Climate Center 2010). During the sampling period, 1331 mm of precipitation were recorded at Oakridge in the 1992-1993 water year, 2014] CHRISTY ET AL.; FIRE EFFECTS IN A MONTANE WETLAND 203 Table 1. Sample Sizes, per Sampling Year. Vegetation zone Sample year Postfire year Number transects Number plots Tsuga mertensiana forest 1993 preburn 4 67 Tsuga mertensiana forest 1997 1 4 67 Tsuga mertensiana forest 2001 5 3 45 Tsuga mertensiana forest 2008 12 2 34 Polytrichum commune moss bed 1993 preburn 1 6 Polytrichum commune moss bed 1997 1 1 6 Polytrichum commune moss bed 2001 5 1 6 Polytrichum commune moss bed 2008 12 1 6 Xerophyllum tenax ecotone 1993 preburn 3 16 Xerophyllum tenax ecotone 1997 1 3 16 Xerophyllum tenax ecotone 2001 5 3 16 Xerophyllum tenax ecotone 2008 12 1 13 Vaccinium uliginosum shrub swamp 1993 preburn 3 31 Vaccinium uliginosum shrub swamp 1997 1 3 31 Vaccinium uliginosum shrub swamp 2001 5 3 31 Vaccinium uliginosum shrub swamp 2008 12 1 11 Grass-dominated fen 1993 preburn 3 6 Grass-dominated fen 1997 1 3 6 Grass-dominated fen 2001 5 2 4 Grass-dominated fen 2008 12 1 2 Carex fen 1993 preburn 4 45 Carex fen 1997 1 4 45 Carex fen 2001 5 4 45 Carex fen 2008 12 2 26 1623 mm in 1996-1997, 111 mm in 2000-2001, and 1196 mm in 2007-2008. Methods Sampling. The original 1976 transect of Frenkel et al. (1986) at Torrey Lake Fen was reestablished by Frenkel and USD A Forest Service ecologists in 1993. This transect and three others established nearby were sampled that same year following USD A Forest Service regional RNA ecotone transect protocols (USFS, unpublished), which differed slightly from those used by Frenkel et al. (1986). All transects were oriented perpendicular to wet- land gradients, spanned the widths of the subject wetlands, and each endpoint was located in upland Tsuga mertensiana (Bong.) Carriere forest. Species cover was observed in 20 X 50 cm microplots, and recorded in absolute percent cover. Microplots along Frenkel’s original tran- sect matched locations of the originals at 0.5 and 1 m intervals, and the transects established in 1993 were sampled at 1 m intervals. Mature trees were tagged in a belt 5-10 m wide along each transect, recording species and dbh, but percent canopy cover was not recorded. For this variable we used the plot-based estimate from Frenkel et al. (1986), which approximated conditions in all four transects in 1993. Because the reestablished transect may not have replicated the exact location of the original, we used the data from 1993 as our pre-burn benchmark. Community composition and zonation observed in 1993 remained essentially unchanged from Frenkel’s original descriptions. All transects were resam- pled periodically over the next 12 yr, but not all transects were resampled in a given year because of budget and time constraints (Table 1). Data analysis. We sorted plots from all transects into six vegetation zones that had been demarcated at the time of pre-burn sampling in 1993 (Table 2), and then summarized plot data as means per zone, per sample year. We used the same Tsuga mertensiana and Xerophyllum tenax (Pursh) Nutt, zone concepts of Frenkel et al. (1986), but to simplify interpretation of fire response we combined their two sedge types and two wet shrub types to one of each. Per vegetation zone and sample year, we summarized (1) percent cover and frequency of dominant species, (2) total vascular plant cover, and (3) total cover per vegetation layer. We omitted species with mean cover of less than 10 percent for all sample years from analysis of dominant species response, but included them in analysis of total vascular plant cover, and cover per vegetation layer. We omitted percent cover of abiotic features (litter, bare ground, fire evidence, etc.) in data analysis, but included observations of same in qualitative descriptions of fire effects. Values indicating recovery from fire represent both positive and negative percent change between 1993 and the most recent sample year, added to 100. 204 MADRONO [Vol. 61 Table 2. Crosswalk of Vegetation Zones Used Here and by Frenkel et al. (1986). Current study Frenkel et al. (1986) Tsuga mertensiana forest Polytrichum commune moss bed Xerophyllum tenax ecotone Vaccinium uliginosum shrub-swamp Grass-dominated fen Carex fen Tsuga mertensiana / Vaccinium scoparium Forest Not present in transect of Frenkel et al. Xerophyllum tenax Fringe Kalmia microphyila / Sphagnum Bog, Vaccinium occidentale / TrifoUum longipes Thicket Not present in transect of Frenkel et al. Carex rostrata Reedswamp, Carex sitchensis Fen The sampling period spanned 12 yr. We selected data from 1993, 1997, 2001, and 2008 because they were the most complete, with two exceptions: (1) data for grass-dominated fen from 2008 were excluded because only two plots were sampled, and (2) the Polytrichum commune Hedw. moss bed was not sampled in 2001, so we used data from 2005 instead. Botanical nomenclature. Botanical nomencla- ture follows the Oregon Flora Project’s Oregon Vascular Plant Checklist (Cook et al. 2013). Names of plant associations follow McCain and Diaz (2002) and Christy (2004). Results For the six vegetation zones over the 12-year sampling period, mean percent cover and fre- quency of dominant species are shown in Table 3. IVIean and total percent cover for each vegetation layer and vegetation zone, the percent change between pre-burn and most recent sample year, and rate of recovery are shown in Figure 2 and Table 4. For convenience, vegetation zones are ordered in this paper along a hydrological gradient from driest to wettest. Tsuga mertensiana forest (Figs. 3, 9, 10; Tables 3 and 4). Before the fire, wetlands at the study site were surrounded by upland forest of old“growth Tsuga mertensiana with a canopy cover of about 80% (Frenkel et al. 1986). Cores taken in 1993 from trees around the edge of the fen indicated that the forest originated from a stand-replacing fire that occurred between 1700 and 1750 {Tsuga mertensiana 8-63 cm dbh, 91- 290 yr old, Pinus monticola Douglas ex D. Don 46 cm dbh, 250 yr old, Pinus contorta var. latifolia Engelm. 8-30 cm dbh, 70-234 yr old). Slow growth and small diameters are typical of this forest type. Of all vegetation zones sampled, fire effects were most severe in Tsuga mertensiana forest, and recovery of total cover was the slowest of any zone, reaching about 19% of pre-burn levels by year 12. All trees were killed by flame heights of up to 15 m, and 55-98% of duff and organic material were consumed, except near the edge of the wetland where loss was 0-50%. Burned snags remained abundant in the RNA ten years after the fire (Acker et al. 2013). Density of tree seedlings in severely burned plots, dominated by Pinus contorta var. latifolia that was extremely sparse even before the fire, was only 4% of that in unburned plots by year six. Most shrubs were killed and had not re-established by year two, and by year 12 total shrub recovery was only 21% of pre-burn levels. Vaccinium scoparium Leiberg ex Coville had trace amounts resprouting by year one, but only 2-3% cover by year five. Within 2 m of the wetland margin, V. scoparium and V. deliciosum Piper resprouted by year two but did not approach pre-burn abundance until year seven. After an initial drop, cover of forbs, grasses, and sedges increased slightly by year 12. At year one the native post-disturbance forb Chamerion angustifolium var. canescens (Alph. Wood) N, H. Holmgren & P. K. Holmgren established on some transects with very low cover in up to 25% of microplots, and increased to 2- 3% by year two. After a slight drop of 1% from pre-burn cover, the moss layer increased to almost 9% by year 12. The liverwort Marchantia polymorpha L. {sensu lato) was widespread and conspicuous on burned soil in intermittent streambeds and small depressions in years one and two, but by year three it was replaced by the moss Ceratodon purpureus (Hedw.) Brid. with localized cover of 5-30%, and by year seven the moss Polytrichum juniperinum Hedw. was con- spicuous. Fire effects continued to impact this zone after year 12 as accelerating decay of dead trees contributed increasing amounts of coarse wood. Poly trichum commune moss bed (Figs. 4, 11; Tables 3 and 4). Before the fire, extensive monotypic beds of Poly trichum commune occurred in shallow, seasonally-flooded glacial potholes (Christy 2004). This vegetation zone was located just below the fringing upland Tsuga mertensiana forest, and had a total mean percent cover of 83%. The fire scorched or killed 60-99% of Poly trichum mats in the single transect sampled, and similar effects were observed in nearby potholes. Some stands exhibited only 1% cover at year one, but increased to 10% at year two, regenerating from uninjured tissue within thicker, moist mats. Polytrichum recovered to some extent by year nine, but by year 12 it had declined to only 21% of pre-burn levels. Graminoids (Carex exsiccata. 2014] CHRISTY ET AL.: FIRE EFFECTS IN A MONTANE WETLAND 205 Table 3. Mean Percent Cover and Frequency of Dominant Species, by Vegetation Zone and Sample Year. Estimate of canopy cover for Tsuga mertensiana forest from Frenkel et al. (1986). Sample year Vegetation 1993 1997 2001 2008 zone Dominant Species Mean Freq Mean Freq Mean Freq Mean Freq Tsuga mertensiana forest Tsuga mertensiana 80.00 100.00 0.00 0.00 0.00 0.00 0.00 0.00 Vaccinium scoparium 17.00 12.30 0.04 2.00 2.70 3.00 3.67 4.00 Xerophyllum tenax 5.70 3.00 1.00 1.30 3.40 1.30 8.67 3.00 Moss 1.40 13.00 1.00 3.00 11.00 11.70 6.58 5.00 Polytrichum commune moss bed (“2001” = 2005) Calamagrostis canadensis 0.00 0.00 0.00 0.00 0.00 14.29 1.14 42.86 Carex exsiccata 0.14 14.29 0.00 0.00 9.43 42.86 12.73 71.43 Deschampsia cespitosa 0.00 0.00 0.01 14.29 21.43 42.86 5.00 71.43 Polytrichum commune 72.86 100.00 3.74 100.00 21.43 57.14 15.00 57.14 Xerophyllum tenax ecotone Xerophyllum tenax 60.30 5.30 27.00 5.00 42.70 4.30 48.38 13.00 Vaccinium scoparium 17.50 1.00 0.03 0.30 5.00 0.70 0.00 0.00 Gaultheria humifusa 0.00 0.00 0.00 0.00 5.80 0.30 0.00 0.00 Sphagnum sp. 19.70 2.00 0.70 0.30 0.80 0.70 0.31 2.00 Moss 4.00 3.30 0.20 LOO 4.30 1.30 2.00 9.00 Vaccinium uliginosum shrub swamp Vaccinium uliginosum 21.40 10.30 11.40 7.70 11.10 5.70 18.00 12.00 Kalmia microphylla 5.90 9.00 1.90 6.30 4.10 6.70 2.36 7.00 Vaccinium deliciosum 2.50 1.70 0.10 0.60 4.40 1.70 1.29 2.00 Xerophyllum tenax 1.30 1.00 6.70 0.70 9.60 1.00 0.00 0.00 Podagrostis thurberiana 3.70 2.60 0.20 1.00 0.00 0.00 0.00 0.00 Sphagnum sp. 42.30 8.30 26.50 6.00 29.30 4.70 15.00 8.00 Moss 13.40 8.70 4.20 2.70 10.20 6.00 10.72 12.00 Grass-dominated fen Calamagrostis canadensis 22.80 1.30 5.00 0.60 2.50 1.50 n/a n/a Deschampsia cespitosa 15.80 0.70 4.20 0.70 0.00 0.00 n/a n/a Sphagnum sp. 30.00 0.70 25.00 0.70 60.00 1.00 n/a n/a Vaccinium uliginosum 3.30 1.00 1.30 0.70 7.30 1.00 n/a n/a Carex exsiccata 1.00 1.30 12.70 1.30 4.50 1.00 n/a n/a Moss 9.80 1.00 0.00 0.00 0.00 0.00 n/a n/a Carex fen Carex aquatilis var. dives 6.80 7.00 2.50 5.00 4.25 3.00 2.50 4.50 Carex buxbaumii 4.30 4.50 1.50 3.30 6.30 1.80 3.60 5.50 Carex exsiccata 25.60 10.30 5.00 7.00 4.70 3.50 5.90 7.00 Deschampsia cespitosa 3.60 3.50 1.00 1.50 0.10 0.30 0.90 1.50 Sphagnum sp. 2.50 3.80 2.40 1.80 5.20 3.30 10.70 8.00 Moss 3.40 6.00 0.40 1.00 2.70 3.30 0.10 1.00 Eleocharis acicularis [L.] Roem. & Schult., Gly~ ceria R. Br., Calamagrostis canadensis [Michx.] P, Beauv.) appeared where none had existed before the burn, with mean covers of 7% and 8%, respectively. Xerophyllum tenax ecotone (Figs. 5, 9, 10; Tables 3 and 4). Before the fire, another up- land-wetland ecotone was characterized by a 3- 5 m wide band of the forb Xerophyllum tenax and shrubs that occurred around margins of wetlands (Frenkel et al. 1986). Xerophyllum tenax was the dominant species with a mean cover of 60%. This zone was severely burned by the fire, and much of the vegetation at the drier upland edge was killed. Total cover in year 12 reached 64% of pre-burn levels, the third slowest of the zones sampled. All growth forms except graminoids exhibited re- duced cover. Pre-burn covers of up to 80% Xerophyllum in some transects were reduced to 15% in year one and 40% in year two. Browned or charred Xerophyllum at the wetter end of the zone resprouted 14 d after the fire, and some stands recovered fully by year one. By year 12, shrub recovery was only 42% of pre-bum levels. Rubus lasiococcus A. Gray and Vaccinium mem- branaceum Douglas ex Torr. at the wetter end of the zone recovered to near pre-burn levels by year two. Gaultheria humifusa (Graham) Rydb., Kal- mia microphylla [Hook.] A. Heller, Spiraea 206 MADRONO [Vol. 61 B ^ r c ~ 3 □ Tree @ fVIoss ■ Shrub ^ Sedge □ Grass □ Forb Fig. 2. Overall fire response, 1993-2008. A. Mean total percent cover by vegetation zone and sample year. B. Change in mean total percent cover by vegetation layer, vegetation zone, and sample year. Carex — Carex fen; Grass — Grass-dominated fen; Polcom — Polytrichum commune moss bed; Tsumer — Tsuga mertensiana forest; Vaculi — Vaccinium uliginosum shrub swamp; Xerten — Xerophyllum tenax ecotone. splendens Baumann ex K. Koch, Vaccinium deliciosum, V. scoparium and V. uliginosum L. died or had trace resprouting by year one, and only 5% cover by year five. Grasses showed a slight increase in cover, and sedges appeared where none had existed in 1993, but cover of both was less than 2%. The moss layer was killed by the fire, and recovery to only 13% of pre-burn levels by year 12 was the poorest for any layer in any of the vegetation zones. Vaccinium uliginosum shrub swamp ( Figs. 6, 9, 10; Tables 3 and 4). Before the fire, Vaccinium uliginosum shrub swamp (= Vaccinium occidentale A. Gray and Kalmia microphylla [Hook.] A. Heller communities of Frenkel et al. 1986) had the highest mean total cover of shrubs (35%) of any zones sampled, with Sphagnum L. covering more than 40% of the moss layer. Total percent cover of vascular plants recovered to 92% of pre- burn levels by year 12, but all growth forms except tree and sedge exhibited diminished cover. Overall, shrubs recovered 84% by 2008. Many were top-killed immediately after the fire but resprouted by year one, except for Vaccinium scoparium that was not detected until year five. Vaccinium deliciosum was inconspicuous until year four. Vaccinium uliginosum recovered to 84% of pre-burn levels by year 12, but Kalmia microphylla and Vaccinium deliciosum had recov- ered to only 41% and 52%, respectively. Tree islands within Vaccinium uliginosum shrub swamp were far enough away from the upland that they did not burn, and these contained the only surviving trees along the transects. Total forb and grass recovery by 2008 were 76% and 86%, respectively. Chamerion angustifolium established in shrub swamp by year one but had nearly disappeared by year two. Sedge showed a 106% increase, but total cover was less than 9%. In contrast to vascular plants, the moss layer had the lowest (47%) recovery of any layer in this zone, and the third poorest overall recovery after the Xerophyllum tenax ecotone and Polytrichum commune moss bed. Aulacomnium palustre and Polytrichum commune had returned to 80% of pre-burn levels by year 12, but the once-dominant Sphagnum capillifolium had recovered only 36% of its former abundance. Grass-dominated fen (Fig. 7; Tables 3 and 4). Before the fire, grass-dominated fen occurred as patches of Calamagrostis canadensis and Deschampsia cespitosa within a matrix of Vaccin- ium uliginosum shrub swamp and Carex fen. The moss layer had a mean cover of 40%, dominated by lawns of Sphagnum capillifolium with a mean cover of 30% and absolute cover in individual stands up to 90%. Photos taken immediately after the fire showed that Calamagrostis canadensis, Deschampsia cespitosa, Kalmia microphylla. Spi- raea splendens, and Vaccinium uliginosum were badly burned. Total mean percent cover of vascular plants recovered to 50% of pre-burn levels by year five, with forb and grass layers recovering 63% and 28%, respectively. In con- trast, the shrub, sedge, and moss layers increased total mean cover by 24% (to 7% cover), 238% (to 5% cover), and 51% (to 60% cover), respectively, over pre-burn levels. Carex fen (Figs. 8, 9, 10; Tables 3 and 4). Before the fire, Carex fen {= Carex rostrata Reedswamp and Carex sitchensis Fen of Frenkel et al. 1986) was a mix of Carex aquatilis var. dives, Carex buxbaumii, and Carex exsiccata occurring in seasonally to perennially flooded 2014] CHRISTY ET AL.: FIRE EFFECTS IN A MONTANE WETLAND 207 Table 4. Mean and Total Percent Cover by Vegetation Layer, Vegetation Zone, and Sample Year. Estimate of canopy cover for Tsuga mertensiana forest from Frenkel et al. (1986). Percent change calculated using preburn (1993) and most recent sample year available per vegetation zone. Vegetation Vegetation zone layer 1993 Sample year 1997 2001 2008 Percent change Recovery Tsuga mertensiana forest Tree 80.00 0.00 0.00 0.00 -100.00 0.00 Shrub 19.37 0.09 4.16 4.09 -78.89 21.11 Forb 4.13 0.96 6.18 5.01 21.31 121.31 Grass 0.07 0.18 1.11 1.91 2628.57 2728.57 Sedge 0.07 0.00 0.44 0.12 71.43 171.43 Moss 1.13 0.82 10.22 8.56 657.52 757.52 TOTAL 104.77 2.05 22.11 19.69 -81.21 18.79 Polytrichum commune moss bed (‘ Grass ‘2001” = 0.00 2005) 0.02 25.00 6.50 n/a n/a Sedge 0.00 0.00 1.00 8.37 n/a n/a Moss 82.50 3.03 51.33 17.50 -78.79 21.21 TOTAL 82.50 3.05 77.33 32.37 -60.76 39.24 Xerophyllum tenax ecotone Shrub 16.88 1.84 9.19 7.10 -57.94 42.06 Forb 67.38 45.6 57.33 54.17 -19.61 80.39 Grass 0.75 0.13 1.06 1.62 116.00 216.00 Sedge 0.00 0.00 1.25 1.15 n/a n/a Moss 18.19 0.38 3.76 2.33 -87.19 12.81 TOTAL 103.2 47.95 72.59 66.37 -35.69 64.31 Vaccinium uliginosum shrub-swamp Tree 2.03 0.00 2.75 6.67 228.58 328.58 Shrub 34.94 18.88 24.91 29.51 -15.54 84.46 Forb 16.35 12.21 14.64 12.35 -24.46 75.54 Grass 11.13 4.75 11.23 9.59 -13.84 86.16 Sedge 8.19 3.41 9.78 8.67 5.86 105.86 Moss 63.47 41.88 44.53 29.59 -53.38 46.62 TOTAL 136.11 81.13 107.84 96.38 -29.19 70.81 Grass-dominated fen Shrub 5.83 1.33 7.25 n/a 24.36 124.36 Forb 2.00 0.33 1.25 n/a -37.50 62.50 Grass 39.50 9.17 11.25 n/a -71.52 28.48 Sedge 1.33 12.67 4.50 n/a 238.35 338.35 Moss 39.83 25.00 60.00 n/a 50.64 150.64 TOTAL 88.49 48.5 84.25 n/a -4.79 95.21 Carex fen Shrub 0.05 0.02 0.18 0.93 1760 1860.00 Forb 8.24 8.92 3.55 9.93 20.51 120.51 Grass 3.63 1.55 0.33 4.41 21.49 121.49 Sedge 40.92 11.59 23.72 29.42 -28.10 71.90 Moss 5.96 3.57 13.53 15.22 155.37 255.37 TOTAL 58.80 25.65 41.31 59.91 1.89 101.89 TOTAL All Vegetation Zones 573.87 208.33 405.43 358.97 -52.13 47.87 depressions. It is the wettest of all zones sampled. Photos taken immediately after the burn showed that Carex fen did not carry a fire and showed no obvious fire effects, and the moss layer charred or burned only where brands landed on it, or where preexisting logs burned into the fen. By year 12, total mean percent cover recovered to 102% of pre-burn levels. Sedges recovered to only 72% of pre-burn levels, and total mean cover increased very slightly for shrubs and grass (less than 1%), forbs (1%), and moss (9%). After a decline immediately following the fire, mean cover of Sphagnum increased 166% by year 12. Discussion By year 12, overall recovery of all six vegetation zones had reached only about 48% 208 MADRONO [VoL 61 Fig. 3. Fire response in Tsuga mertensiana forest, in mean percent cover. A. Dominant species. B. Vegetation layers. Tsumer — Tsuga mertensiana; Vacsco — Vaccinium scoparium; Xerten — Xerophyllum tenax. Figures do not include canopy cover of mature trees (see text). Obscured in 3 A: Tsumer 1997 = 0. of the pre-burn baseline, and persistent reduced cover in some layers indicated that fire effects were still evident 12 yr after the event (Figs. 2, 9, 10; Table 4). Field observation of the six vegetation zones immediately after the fire indicated that the upland Tsuga mertensiana forest had sustained the most severe fire damage, followed in diminishing order by Polytrichum commune moss bed, grass-dominated fen, Xer- ophyllum tenax ecotone, Vaccinium uiiginosum shrub swamp, and Carex fen. However, recovery did not parallel the immediate post-fire impacts observed in 1996. Upland Response. The fire was a catastrophic event for upland Tsuga mertensiana forest, killing all trees and most of the shrubs (Figs. 2, 10, 11). The 250-300 yr fire return interval observed in the RNA was intermediate for reports of 100- 460 yr (Dickman 1984; Dickman and Cook 1989; Gardner and Whitlock 2001). After an initial influx of Pinus contorta var. latifolia, recruitment of additional species at Torrey Lake Fen will be very slow (Acker et al. 2013). Burned forests of Tsuga mertensiana on the Olympic Peninsula of Washington took 55-88 yr to reestablish densities of fewer than 600 trees per hectare (Agee and Smith 1984). Impacts from stand-replacing fire on forest trees and shrubs have been relatively well documented in the Pacific Northwest, and effects in the RNA (Table 5) generally followed patterns reported elsewhere. High mortality and slow recovery of Gaultheria humifusa, Vaccinium deli- Fig. 4. Fire response in Poly trichum commune moss bed, in mean percent cover. A. Dominant species. B. Vegetation layers. Calcan- - Ualamagrostis canadensis; Carexs — Carex exsiccata; Desces — Deschampsia cespitosa; Polcom — Polytrichum commune. Obscured in 4A: Desces 1997 = 0, Polcom 1997 = 3.74. 2014] CHRISTY ET AL.: FIRE EFFECTS IN A MONTANE WETLAND 209 Fig. 5. Fire response in Xerophyllum tenax ecotone, in mean percent cover. A. Dominant species. B. Vegetation layers. Gauhum — Gaultheria humifusa’, Sphagn — Sphagnum spp.; Vacsco — Vaccinium scoparium; Xerten — Xerophyllum tenax. Obscured in 5 A: Vacsco 1997 = 0.03. ciosum, and Vaccinium scoparium were similar to those reported by Matlack et al. (1993); Turner et al. (1997); Zobel and Antos (1997); and Hill and Vander Kloet (2005). Shallow-rooted species are more vulnerable to hot fires than deeper-rooted species, and soil moisture at the time of the burn is critical in determining shrub survival. The seed bank is often consumed in hot fires and recruitment is mediated by distance from surviv- ing seed sources. In stand-replacing situations, resprouting therefore remains the only viable means for short-term recovery of understory species. Despite the fact that fire effects were most severe in the Tsuga mertensiana forest, the amount of change in the understory was relative- ly small (Fig. 2b) because the shady pre-burn understory was never abundant. Increases in forbs, grasses, sedges, and moss over pre-burn levels can be attributed to release from canopy shading. In addition, the surge in bryophyte cover can be attributed to the flush of post-burn “fire mosses” mediated by pH and nutrients from ash. The appearance of a post-fire soil crust dominated by bryophytes has not been reported from the Cascade Range, and differs from the composition of pioneer species on newly-exposed soils (e.g., Zobel and Antos 1997). Transition from an initial bloom of the liverwort Marchantia polymorpha {sensu lato) in years one and two to the mosses Ceratodon purpureus by year three and Poly trichum juniperinum by year seven (Fig. 12) is A B > o u c 0 o Q. □ Tree ■ Moss □ Shrub il Grass □ Sedge 0 Forb 1993 1997 2001 2008 Fig. 6. Fire response in Vaccinium uliginosum shrub swamp, in mean percent cover. A. Dominant species. B. Vegetation layers, Kalmic — Kalmia microphylla; Podthu — Podagrostis thurburkma; Sphagn — Sphagnum spp.; Vacdel — Vaccinium deliciosum; Vaculi — Vaccinium uliginosum; Xerten — Xerophyllum tenax. Obscured in 6 A: Kalmic 1997 = 1.9, Vacdel 1993 = 2.5, Vacdel 1997 = 0.1. 210 MADRONO [VoL 61 Fig. 7. Fire response in grass-dominated fen, in mean percent cover. A. Dominant species. B. Vegetation layers. Calcan — Calamagrostis canadensis; Carexs — Carex exsiccata; Desces — Deschampsia cespitosa; Sphagn — Sphagnum spp.; Vaculi — Vaccinium uliginosum. similar to sequences observed elsewhere (Cremer and Mount 1965; DeBenedetti and Parsons 1984; Maltby et al. 1990; Tesky 1992; Matthews 1993; Thomas et al 1994; Fryer 2008). Crusts formed by these species presumably play a role in ecosystem recovery similar to biological soil crusts reported from arid environments (e.g., Belknap et al. 2001; Ponzetti et al. 2007) by stabilizing soil, fixing atmospheric nitrogen, and creating seedbeds for invading vascular plants. In burned forests and peatlands, recovering vascular vegetation eventually shades out soil crusts, and their component species become much less common. Maltby et al. (1990) observed a decline in crusts between years six and 10, similar to what happened at the RNA. Wetland Response. Frenkel et al. (1986) docu- mented the extent of upland tree cover and shading around the edge of wetlands in the RNA. Removal of forest cover influences adjacent wetlands by increasing snowpack, soil moisture, runoff, streamflow, turbidity, light, and temper- ature (Moore and Bellamy 1974; Tiedemann et al. 1979; Pyne et al. 1996). Changes in nutrient status from, runoff can boost productivity in wetlands, causing relatively short-lived changes in species composition. Pyne et al. (1996) estimated that hydrology and water quality can recover to pre- burn conditions within 5-10 yr. Recovery of wetland vegetation has been estimated to occur within 10-30 yr (Rowe and Scotter 1973; Tallis 1983; DeBenedetti and Parsons 1984; Kuhry B ■ Moss M Shrub □ Sedge O Grass 0 fcrb Fig. 8. Fire response Carex fen, in mean percent cover. A. Dominant species, B. Vegetation layers. Calcan — Calamagrostis canadensis; Caraqd — Carex aquatiiis var. dives; Carbux — Carex buxbaumii; Carexs — Carex exsiccata; -Deschampsia cespitosa; Sphagn — Sphagnum spp; Obscured in 8 A: Carexs 2001 = 4.7, Sphagn 1993 = 2.5, Sphagn 1997 = 2.4. 2014] CHRISTY ET AL.: FIRE EFFECTS IN A MONTANE WETLAND 211 Fig. 9. Original 1976 transect, view to north. A. 1981, before fire. B. 2011, 15 years after fire. Scale intervals in A = 10 cm. Transect length = 55 m. 1994; Ratchford et ah 2005; Kuhry and Turunen 2006). As expected, fire effects at Torrey Lake Fen diminished with increasing soil moisture and distance from burned upland, but effects persisted even in the wettest zones. Polytrichum commune moss bed, Xerophyllum tenax ecotone, Vaccinium uliginosum shrub swamp, and grass-dominated fen burned more severely than Carex fen. Vaccinium uliginosum shrub swamp and Carex fen exhibited few immediate posLflre impacts at an average of 26 m (range = 16 to 31 m) from the perimeter of the burn, but 12 yr after the fire percent cover had not recovered to pre-burn levels, indicating that fire effects were more pervasive than initially apparent. Almost all vegetation layers showed reduced percent cover in Polytrichum commune moss bed, Xerophyllum tenax ecotone, Vaccinium uliginosum shrub swamp by year 12 (Table 4). In contrast, grass- dominated fen showed increases in shrub, sedge and moss layers, but decreases in forb (38%) and grass (72%) layers, while Carex fen showed increases in all layers except sedge, which declined 28%. These responses may be attributable in part to normal patch dynamics of rhizomatous popu- lations as detected with repeat photography in other fens (e.g., Christy and Titus 1998). The Polytrichum commune moss bed exhibited the second-greatest change in vegetation and had the second-poorest overall recovery rate by year 12. Survival of Poly trichum was mediated by moisture content in the moss bed and underlying soil at the time of the fire. Maltby et ah (1990) regarded beds of Poly trichum commune as a long- lived, fire-initiated system that eventually con- verts to vascular plants. However, in the Cascade Range, beds of Polytrichum commune persist in depressional wetlands as an integral component of old-growth Tsuga mertensiana forest (Christy, unpublished data). In the study area, their destruction was the only known wetland com- munity where fire reset vegetation to a completely different serai stage dominated by graminoids (Fig. 11). Recruitment and Recolonization. Survival of upland species around the edge of wetlands may be critical for recovery of upland communities. Compared to the upland Tsuga mertensiana forest, wetland zones exhibited high resistance to fire, and their present serai stage at the RNA for the most part is independent of serai stages in upland forest. No non-native species were present in the transects before or after the 1996 fire. Early establishment of new plants was limited to Chamer- ion angustifolium var. canescens and Pinus con- torta var. latifolia. Chamerion was the only post- disturbance invader, present within one year in all communities except Carex fen and Vaccinium 212 MADRONO [Vol. 61 Fig. 10. Original 1976 transect, view to south. A. 1981, before fire. B. 2006, 10 years after fire. Transect length = 55 m. 2014] CHRISTY ET AL.: FIRE EFFECTS IN A MONTANE WETLAND 213 Fig. 11. Polytrichum commune moss bed community. A, 1981, before fire. B. 2005, nine years after fire. 214 MADRONO [Vol. 61 Table 5. Fire Effects on Plant Species at Torrey-Charlton RNA. Recovery times are minimal estimates from individual plots. Species Fire effect Vascular Plants Abies amabilis Abies lasiocarpa Calamagrostis canadensis Chamerion angustifolium Deschampsia cespitosa Gaultheria humifusa Kalrnia microphylla Lonicera caerulea Picea engelmannii Pinus contorta var. latifoUa Pinus monticola Rubus lasiococcus Spiraea splendens Tsuga mertensiana Vaccinium deliciosum Vaccinium membranaceum Vaccinium scoparium Vaccinium uliginosum Xerophyllum tenax Bryophytes Ceratodon purpureus Marchantia polymorpha (s.l) Polytrichum commune Polytrichum juniperinum Sphagnum capillifolium Complete kill Complete kill Top kill or complete kill, slow to recover, increaser in wettest zones Ephemeral increaser Top kill or complete kill, slow to recover Top kill or complete kill, very slow to recover Top kill or complete kill, recovery 7 yr Top kill or complete kill, very slow to recover Complete kill Complete kill, except on tree islands 16-24 m from burned perimeter Complete kill Top kill, recovery 2 yr Top kill or complete kill, very slow to recover Complete kill, except on tree islands 16-24 m from perimeter Top kill or complete kill, recovery 7 yr Top kill or complete kill, recovery 2 yr Top kill or complete kill, recovery 7 yr Top kill or complete kill, recovery 7 yr Top kill or complete kill, recovery 9 yr Increaser yr 3, ephemeral Increaser yr 1-2, ephemeral Top kill or complete kill, recovery 9 yr Increaser yr 7 Top kill or complete kill, recovery 6 yr uliginosum shrub swamp. Three years after the fire, it was most abundant in the Xerophyllum tenax ecotone, but with an average cover of only 3%. In contrast, post-fire densities in uplands at Yellowstone National Park reached 4.8 sprouts/ m^ two years after the fire and 23.4 sprouts/m^ after five years (Turner et al. 1997). At Torrey- Charlton RNA, seed sources for Chamerion may have been limited, or post-fire surface conditions may not have been favorable for establishment. Chamerion was most consistently found in the Xerophyllum tenax ecotone, but its failure to expand may be due to environmental factors or Fig. 12. Soil crust {Ceratodon purpureus. Polytrichum Juniperinum) in 2005, nine years after fire. Scale intervals = 10 cm. competition, because most species there quickly recovered to near pre-burn abundance. Around the margins of the wetlands at the RNA, survival of shrubs and herbs was limited to resprouting, extension of rhizomes, or other vege- tative mechanisms from surviving plants. Increased light and nutrients available after the fire provided only limited advantage to surviving species, because of their restriction to moisture gradients. However, survival of Vaccinium cespitosum, V. deliciosum, V. membranaceum, V. scoparium, and Rubus lasio- coccus indicates that these habitats may serve as important sources for dispersal and recoloniza- tion of uplands and other wetlands (Hill and Vander Kloet 2005). Recruitment of burned snags in the wetlands will provide substrate for estab- lishment of new shrub swamp, sphagnum hum- mocks, and tree islands. Conclusions Because wetland soils at Torrey Lake Fen were hydrated at the time of the fire, wetland plant communities sustained relatively superficial dam- age compared to the catastrophic effect on upland Tsuga mertensiana forest. Fire damage was inversely proportional to moisture gradient, and recovery of both vascular plants and bryophytes more or less directly proportional to the gradient. Fire effects persisted 12 yr after the fire in the form of reduced total percent cover, and changes in the percent cover of component vegetative layers. Cover of shrub, forb, grass, and 2014] CHRISTY ET AL.; FIRE EFFECTS IN A MONTANE WETLAND 215 moss layers declined in vegetation zones at the drier end of the moisture gradient, but increased in grass and sedge-dominated zones at the wetter end of the gradient, at the expense of forb, grass, and sedge layers. Acknowledgments Dr. Robert Frenkel generously contributed his data and knowledge about Torrey-Charlton RNA, and his 20 years of work at this site made this project possible. The USDA Forest Service supported collection of baseline data in the RNA. Botanists and ecologists from the Willamette and Siuslaw National Forests, the Oregon Biodiversity Information Center, and Salix Associates provided expertise in the field. Todd Wilson provided background information on the RNA. Denny Albert, Joe Rocchio, and an anonymous reviewer provided helpful advice that greatly im- proved the manuscript. Literature Cited Acker, S. A., J. Kertis, H. Bruner, K. O’Connell, AND J. Sexton. 2013. Dynamics of coarse woody debris following wildfire in a mountain hemlock {Tsuga mertensiana) forest. Forest Ecology and Management 302:231-239. Agee, J. K. and L. Smith. 1984. Subalpine tree reestablishment after fire in the Olympic Moun- tains, Washington. Ecology 65:810-819. Attiwill, P. M. 1994. The disturbance of forest ecosystems: the ecological basis for conservative management. Forest Ecology and Management 63:247-300. Belknap, J., J. H. Kaltenecker, R. Rosentreter, J. Williams, S. Leonard, and D. Eldridge. 2001. Biological soil crusts: ecology and manage- ment. United States Department of the Interior (USDI) Bureau of Land Management Technical Reference 1730-2, BLM/ID/ST-0 1/00 1+1 730. Unit- ed States Department of the Interior Bureau of Land Management Printed Materials Distribution Center, Denver, CO. Christy, J. A. 2001. Wetland vegetation of Fanno Meadows Preserve, Polk County, Oregon. Unpub- lished report to The Nature Conservancy. Oregon Natural Heritage Information Center, Portland, OR. . 2003. Summary of vegetation studies, 1980- 2002, Gold Lake Bog Research Natural Area, Willamette National Forest, Lane County, Oregon. Unpublished report to United States Department of Agriculture Forest Service. Oregon Natural Heritage Information Center, Oregon State Uni- versity Institute for Natural Resources, Corvallis, OR. . 2004. Native freshwater wetland plant associa- tions of northwestern Oregon. Oregon Natural Heritage Information Center, Oregon State Univer- sity Institute for Natural Resources, Corvallis, OR. . 2005a. Sphagnum fens on the Oregon coast: diminishing habitat and need for management. Unpublished report to United States Fish and Wildlife Service. Oregon Natural Heritage Infor- mation Center, Oregon State University Institute for Natural Resources, Corvallis, OR. . 2005b. Vegetation and stand history of Henderson Fen, Sycan Marsh Preserve, Lake County, Oregon. Unpublished report to The Nature Conservancy. Oregon Natural Heritage Information Center, Oregon State University Institute for Natural Resources, Corvallis, OR. and j. H. Titus. 1998. Vegetation of Hidden Lake Fen, Blue River Ranger District, Willamette National Forest, Oregon. Unpublished report to USDA Forest Service, Oregon Natural Heritage Program, Portland, OR. Cook, T., S. C. Meyers, and S. Sundberg (eds.). 2013. Oregon Vascular Plant Checklist. Version 1.3. Website http://oregonflora.org/checklist.php (accessed 29 August 2013). Cramer, J. R. 2005. The Biscuit Fire and the flora of serpentine fens: differences in species composition between burned and unburned fens. M.S. thesis. University of Vermont, Burlington, VT. Cremer, W. and a. B. Mount. 1965. Early stages of plant succession following the complete felling and burning of Eucalyptus regnans forest in the Florentine Valley, Tasmania. Australian Journal of Botany 13:303-322. Debenedetti, S. H. and D. J. Parsons. 1984. Postfire succession in a Sierran subalpine meadow. Amer- ican Midland Naturalist 111:118-125. Dickman, a. W. 1984. Fire and PhelUnus weirii in a mountain hemlock {Tsuga mertensiana) forest: postfire succession and the persistence, distribution, and spread of a root-rotting fungus. PhD disserta- tion. University of Oregon, Eugene, OR. AND S. Cook. 1989. Fire and fungus in a mountain hemlock stand. Canadian Journal of Botany 67:2005-2016. Frenkel, R. E., W. H. Moir, and J. A. Christy. 1986. Vegetation of Torrey Lake Mire, central Cascade Range, Oregon. Madrono 33:24—39. Fryer, J. L. 2008. Polytrichum juniperinum. In Fire Effects Information System (FEIS), [Online] Unit- ed States Department of Agriculture Forest Ser- vice, Rocky Mountain Research Station, Fire Sciences Laboratory, Missoula, MT. Website http://www.fs.fed.us/database/feis/ (accessed 29 Au- gust 2013). Gardner, J. J. and C. Whitlock. 2001. Charcoal accumulation following a recent fire in the Cascade Range, northwestern USA, and its relevance for fire-history studies. The Holocene 11:541-549. Guerrant, E. O., S. T. Schultz, and D. K. Imper. 1998. Recovery plan for the endangered western lily {Lilium occidentale). U.S. Fish and Wildlife Service, Portland, OR. Hansen, H. P. 1941. Paleoecology of two peat deposits on the Oregon coast. Oregon State Monographs, Studies in Botany 3:1-31. . 1944. Further pollen studies of peat bogs on the Pacific coast of Oregon and Washington. Bulletin of the Torrey Botanical Club 71:627-636. Hebda, R. j., K. Gustavson, K. Golinski, and A. M. Calder. 2000. Burns Bog ecosystem review synthesis report for Burns Bog, Fraser River Delta, Southwestern British Columbia, Canada. Environ- mental Assessment Office, Victoria, BC. Heinselman, M. L. 1963. Forest sites, bog processes, and peatland types in the Glacial Lake Agassiz 216 MADRONO [Vol. 61 Region, Minnesota. Ecological Monographs 33:327-374. Hill, N. M. and S. P. Vander Kloet. 2005. Longevity of experimentally buried seed in Vaccin- ium: relationship to climate, reproductive factors and natural seed banks. Journal of Ecology 93: 1167-1176. Kuhry, P. 1994. The role of fire in the development of Sphagnum-dominsitQd peatlands in western boreal Canada. Journal of Ecology 82:899-910. AND J. Turunen. 2006. The postglacial development of boreal and subarctic peatlands. Pp. 25-46 in R. K. Wieder and D. H. Vitt (eds.), Boreal Peatland Ecosystems. Ecological Studies 188:1^35. Kunze, L. M. 1994. Preliminary classification of native, low elevation, freshwater wetland vegeta- tion in western Washington. Washington Natural Heritage Program, Department of Natural Re- sources, Olympia, WA. Maltby, E., C. J. Legg, and M. C. F. Proctor. 1990. The ecology of severe moorland fire on the North York Moors: effects of the 1976 fires, and subsequent surface and vegetation development. Journal of Ecology 78:490-518. Martin, R. C. and R. E. Frenkel. 1978. Preserve analysis: Blacklock Point. Oregon Natural Area Preserves Advisory Committee. State Land Board, Salem, OR. Matlack, G. R., D. j. Gibson, and R. E. Good. 1993. Clonal propagation, local disturbance, and the structure of vegetation: ericaceous shrubs in the Pine Barrens of New Jersey. Biological Conserva- tion 63:1-8. Matthews, R. F. 1993. Marchantia polymorpha. In Fire Effects Information System (FEIS), [Online] United States Department of Agriculture Forest Service, Rocky Mountain Research Station, Fire Sciences Laboratory, Missoula, MT. Website http://www.fs.fed.us/database/feis/ (accessed 29 Au- gust 2013). McCain, C. and N. Diaz. 2002. Field guide to the forested plant associations of the westside central Cascades of northwest Oregon; Willamette N.F., USFS; Mt. Hood N.F., USFS; Salem District, BLM; Eugene District, BLM. USDA Forest Service Technical Paper R6-NR-ECOL-TP-02- 02. Moore, P. D. and D. J. Bellamy. 1974. Peatlands. Springer- Verlag, New York, NY. Murray, M. 2000. Wetland plant associations of the western hemlock zone in the central coastal and Cascade Mountains. Oregon Natural Heritage Program, Portland, OR. PoNZETTi, J. M., B. McCune, and D. a. Pyke. 2007. Biotic soil crusts in relation to topography, cheatgrass and fire in the Columbia Basin, Washington. The Bryologist 110:706-722. Pyne, S. j., P. L. Andrews, and R. D. Laven. 1996. Introduction to wildland fire. 2nd ed. John Wiley and Sons, Inc, New York, NY. Ratchford, j. S., S. E. Wittman, E. S. Jules, A. M. Ellison, J. J. Gotelli, and N. J. Saunders. 2005. The effects of fire, local environment and time on ant assemblages in fens and forests. Diversity and Distributions 11:487-497. Rowe, J. S. and G. W. Scotter. 1973. Fire in the boreal forest. Quaternary Research 3:444-464. Salix Associates. 1998. Establishment record for Torrey-Charlton Research Natural Area within Willamette and Deschutes National Forests, Lane and Deschutes Counties, Oregon. Unpublished report to USDA Forest Service, Eugene, OR. Sanders, N. J., N. J. Gotelli, S. E. Wittman, J. S. Ratchford, A. M. Ellison, and E. S. Jules. 2007. Assembly rules of ground-foraging ant assemblages are contingent on disturbance, habitat and spatial scale. Journal of Biogeography 34: 1632-1641. Schultz, S. T. 1989. Status report on Lilium occiden- tale Purdy. Unpublished report to Endangered Species Program, Oregon State Department of Agriculture. University of British Columbia, Van- couver. Seyer, S. C. 1979. Vegetative ecology of a montane mire, Crater Lake National Park, Oregon. M.S. thesis. Oregon State University, Corvallis, OR. . 1981. Survey of vegetation of 18 lakes in Wallowa-Whitman National Forest, Oregon. Un- published report to USDA Forest Service Forest Science Laboratory, Corvallis, OR. . 1983. Ecological analysis, Multorpor Fen Preserve, Oregon. Unpublished report to The Nature Conservancy, Oregon Field Office, Port- land, OR. Tallis, J. H. 1983. Changes in wetland communities. Pp. 311-347 in A. J. P. Gore (ed.). Mires: swamp, bog, fen and moor. Ecosystems of the World, Vol. 4. Tesky, j. L. 1992. Ceratodon purpureus. In Fire Effects Information System (FEIS), [Online] United States Department of Agriculture Forest Service, Rocky Mountain Research Station, Fire Sciences Labo- ratory, Missoula, MT. Website http://www.fs.fed. us/database/feis/ (accessed 29 August 2013). Thomas, P. A., M. C. F. Proctor, and E. Maltby. 1994. The ecology of severe moorland fire on the North York Moors: chemical and pH constraints on moss establishment from spores. Journal of Ecology 82:457-474. Tiedemann, a. R., C. E. Conrad, J. H. Dieterich, J. W. Hornbeck, W. F. Megahan, L. A. ViERECK, and D. D. Wade. 1979. Effects of fire on water, a state-of-knowledge review. United States Department of Agriculture Forest Service General Technical Report WO- 10. Titus, J. H. and J. A. Christy. 1997. Vegetation of Big Marsh, Deschutes National Forest, Oregon. Unpublished report to Deschutes National For- est. Oregon Natural Heritage Program, Portland, OR. Tolman, D. a. 2006. Characterization of the ecotone between Jeffrey pine savannas and Darlingtonia fens in southwestern Oregon. Madrono 53:199- 210. . 2007. Soil patterns in three Darlingtonia fens of southwestern Oregon. Natural Areas Journal 27:374-384. ' Torgerson, E. F., j. McWilliams, and C. J. McMurphy. 1949. Soil survey, Astoria area, ^ Oregon. United States Department of Agriculture ; Bureau of Plant Industry, Soils, and Agricul- tural Engineering. Series 1938, Number 20:1- 80. Turner, M. G., W. H. Romme, R. H. Gardner, AND W. W. Hargrove. 1997. Effects of fire size 2014] CHRISTY ET AL.: FIRE EFFECTS IN A MONTANE WETLAND 217 and pattern on early succession in Yellowstone National Park. Ecological Monographs 67: 411^33. Western Regional Climate Center. 2010. Month- ly total precipitation (inches) for Oakridge Fish Hatchery, Oregon, 1914-2010. Website http:// www.wrcc.dri.edu/cgi-bin/cliMAIN.pl7or62 1 3 (ac- cessed 15 December 2010). Wilson, C. E. 1986. Floristic and edaphic aspects of vegetational patterns in subalpine mires of the Cascade Mountains of Oregon. M.S. thesis. Uni- versity of Oregon, Eugene, OR. ZOBEL, D. B. AND J. A. Antos. 1997. A decade of recovery of understory vegetation buried by volcanic tephra from Mount St. Helens. Ecological Monographs 67:317-344. Madrono, VoL 61, No. 2, pp. 218-226, 2014 HOW DOES SIMULATED GOPHER DISTURBANCE AFFECT THE ESTABLISHMENT OF HOLCUS LANATUS L. (POACEAE) IN CALFORNIA COASTAL PRAIRIE? Meredith A. Thomsen University of Wisconsin-La Crosse, Department of Biology, 1725 State St., La Crosse, WI 54601 mthomsen@uwlax.edu Carla M. D’ Antonio University of California, Santa Barbara, Department of Ecology, Evolution and Marine Biology, Santa Barbara, CA 93106 Abstract Disturbance often has the net effect of promoting invasive plant establishment, but the precise nature of the relationship between disturbance and invasion can depend on community context. We used simulated gopher mounds in bare and monoculture plots of three California native grasses {Bromus carinatus var. maritimus [Piper] C. L. Hitch., Festuca rubra L., and Calamagrostis nutkaensis [J. Presl] Steudel) to test the effects of mounds on seedling establishment and survival of the European perennial grass Holcus lanatus L. Soil disturbance treatments were crossed with manipulations of the plant canopy (shade cages in bare and Bromus plots vs. pinning back grass leaves in Festuca and Calamagrostis plots) to separate some of the positive and negative effects of natural mounds. Mean PAR measured in February at the soil surface varied from 1 1 to 969 pm/m^s. As predicted, shade structures that decreased light availability but also increased soil moisture generally increased Holcus seedling establishment and survival in bare and Bromus plots. In contrast, Holcus seedling establishment increased in response to increased light availability and soil disturbance in Festuca and Calamagrostis plots, even where water availability was lower. Thus, the relative importance of light and water availability across plot types appeared to determine the effect of disturbance on invasive plant establishment. Ultimately, Holcus survival was low on mounds in bare plots and on unshaded mounds in Bromus plots, and similarly low numbers of Holcus seedlings survived across all treatment combinations in Festuca and Calamagrostis plots. Therefore, our results do not support the hypothesis that gopher mounds explain the invasion of Holcus in native-dominated coastal prairie sites. Key Words: Bromus, coastal prairie, Calamagrostis, Festuca, Holcus, invasion. The importance of canopy gaps to seedling establishment and species turnover in grassland vegetation has been well documented (e.g.. Watt 1947; Platt 1975; Goldberg 1987; Foster et ah 2002). For example, mounds created by Thom- omys bottae (Eydoux and Gervais 1836) (pocket gophers) are common in California grasslands and can significantly affect plant community composition (e.g., Hobbs and Mooney 1991; Stromberg and Griffin 1996). The population growth of exotic species is one type of vegetation change that frequently occurs as the result of small-scale disturbances (e.g., Foster et al. 2002; Dethier and Hacker 2005; Milbau et al. 2005). In California grasslands, gopher activity can pro- mote the dominance of exotic annual grasses (Hobbs and Mooney 1991; Stromberg and Griffin 1996; Kotanen 1997; Seabloom et al. 2005), exotic forbs (Gerlach and Rice 2003), an invasive perennial succulent (D’ Antonio 1993), and exotic perennial grasses (Peart 1989b). Yet soil disturbances do not always promote non- native species in California grasslands: the context in which disturbances occur is important in determining whether they will be colonized by invading species (DiVittorio et al. 2007). The net effect of a disturbance is the result of positive and negative factors affecting plant performance (Sousa 1984; Goldberg 1987; Cahill and Casper 2002). In California coastal prairie, light (Peart 1989b) and nitrogen (Kotanen 1997; Canals et al. 2003) availability are both higher on mounds than in undisturbed vegetation, and root competition is likely to be lower (Cahill and Casper 2002). Conversely, mound soils have lower bulk density and are warmer and drier than undisturbed soils (Kotanen 1997; Canals et al. 2003), which likely inhibits seedling establishment, particularly in low rainfall years. In addition, local plant composition is likely to influence the net effect of gopher mounds. For example, if a disturbance removes or reduces the cover of a competitive resident, the positive effects of disturbance likely outweigh the negative ones, thereby facilitating invasion. Conversely, tall neighbors could shade gopher mounds, reducing the harsh moisture and temperature conditions of an open mound and reinforcing the 2014] THOMSEN AND D’ANTONIO: HOLCUS AND DISTURBANCE IN COASTAL PRAIRIE 219 positive effects of disturbance. Thus, to accurate- ly predict the overall relationship between distur- bance and invasibility, we must understand how the positive and negative effects of disturbance vary with the community context in which they occur. Holcus lanatus L. (velvet grass or Yorkshire fog) is a problematic invader of California coastal prairie (Pitcher and Russo 1988) and is listed by the California Invasive Plant Council as a moderate impact invader (CaLIPC 2013). Grime (1979) characterized Holcus as a ruderal-peren- nial herb, with a capacity for rapid spread. Holcus was found to be competitively dominant to three native and two other exotic perennial grasses in a greenhouse experiment (Thomsen et al. 2006a) and can produce seed in its first year of growth (Peart 1989b; Thomsen et al. 2006b). An earlier field experiment indicated that patches of the native grasses Festuca rubra L. and Calamagrostis nutkaensis (J. Presl) Steudel are highly resistant to Holcus invasion (Thomsen and D’Antonio 2007). In contrast, Holcus rapidly established in bare plots and plots dominated by the native perennial grass Bromus carinatus var. maritimus (Piper) C. Hitch. The spread of Holcus in native-dominated coastal prairie sites may result from its ability to enter patches dominated by less resistant species. Alternatively, we hy- pothesized that gopher mounds could promote Holcus seedling establishment by creating gaps in the cover of otherwise competitive species. Here, we examine the interactive effects of shading and soil disturbance on the establish- ment and survival of Holcus seedlings in bare and monoculture plots in a California coastal prairie. Crossed shading and soil disturbance treatments separate some of the positive and negative effects of gopher mounds. Measure- ments of light, nitrogen (N), and water availabil- ity were made across treatments to gain insight as to the mechanism of the disturbance effects. Our earlier experiment indicated that high light availability is associated with invasibility in bare and Bromus plots, while low light availability contributed to the resistance of Festuca and Calamagrostis patches (Thomsen and D’Antonio 2007), Thus, we predicted that altered light availability would have opposite effects in these plots. Furthermore, since Holcus readily invades bare and Bromus plots in the absence of dis- turbance, and since Holcus is water-limited in some coastal prairie sites (Thomsen et al. 2006b), we predicted that the drier conditions on mounds would decrease Holcus performance relative to undisturbed areas of those plots, particularly when unshaded. In contrast we predicted that gopher disturbance in plots of resistant species like Festuca would have a net positive effect on Holcus establishment. Methods The experiment was conducted in grassland at the University of California Bodega Marine Reserve (BMR), 80 km north of San Francisco in Sonoma County, California (38°18'N, 123°03'W). Climate at the site is typical for California coastal prairie; mean annual temper- ature is 12°C and mean annual rainfall is 85 cm, 83% of which falls between November and March. European annual grass species comprise much of the cover in the BMR grassland, although native perennial grass species are also common (Kolb et al. 2002). Holcus is of conservation concern at BMR, dominating ap- proximately 20% of the grassland area (J. Soanes, Bodega Marine Reserve, personal communica- tion) and forming nearly monospecific stands. Holcus spreads readily via decumbent tillers, forming loosely tufted patches (Thompson and Turkington 1988); plants at BMR average 30 cm in height. Forty -four 1 m^ plots were established in an area of the BMR dominated by Holcus, as part of the experiment described in Thomsen and D’An- tonio (2007). Plots were separated by 1.5 m aisles that were mowed annually to prevent Holcus seed inputs. Each plot was randomly assigned to one of four “cover type” treatments: bare plots and monocultures of the three native grass species Bromus carinatus var. maritimus, Festuca rubra, and Calamagrostis nutkaensis. Species were cho- sen to represent the diversity of form among California native grasses, particularly variation in foliage density, phenology, and canopy height. Bromus individuals averaged four cm tall in experimental plots and have a sparse, prostrate growth form, while Festuca and Calamagrostis are upright bunchgrasses, averaging 30 vs. 45 cm tall (Thomsen and D’Antonio 2007). Seeds were collected at BMR {Bromus, Calamagrostis, Hol- cus) and at a nearby coastal prairie remnant {Festuca) and grown in the greenhouse using sterilized soil in 49-mL Fir Cell Cone-Tainers^'^ (Steuwe and Sons, Corvallis, OR). Holcus was removed from all plots using hand tools in December 1999; resprouting plants and new germinants were treated with glyphosate herbicide (Roundup^^, Monsanto Corporation, St. Louis, MO) in January 2000. Native grass seedlings were planted between January 17 and February 5, 2000 at densities reflective of their adult sizes: 100, 64, and 49 plants per plot for Bromus, Festuca and Calamagrostis, respectively. Cover types were maintained by hand weeding (planted plots) or annual glyphosate application (bare plots). Plots were sown with Holcus seeds in November 2001. In addition, fences were con- structed around the plots used in the experiment described here in summer 2000. Fences were meant to exclude burrowing animals, but the 220 MADRONO [Vol. 61 treatment was unsuccessful and soil disturbance was similar across fenced and unfenced plots. Fences were therefore removed in October 2002. At the same time, plots were hand-weeded to remove all vegetation in bare plots and all vegetation other than planted individuals in the other cover types. In November 2002, four 30 cm-diameter circular subplots were established per plot. Two canopy treatments (shaded vs. unshaded) and two soil treatments (undisturbed vs. mound) were crossed and randomly assigned to each subplot. The method used to create shaded or unshaded conditions varied with cover type. In Bromus and bare plots, square shade structures, 35 cm on a side and 50 cm tall, were constructed using a double layer of 1 cm-mesh polypropylene fencing (McMaster-Carr, Los Angeles, CA). A gap at the bottom allowed access to insects and small mammals. In Festuca and Calamagrostis plots, the plant canopy itself provided substantial shading over the soil surface. To create shaded and unshaded subplots the grass canopy was either left in place with leaves sometimes tied together to get more consistent shading over the subplot, or was pinned back using long wires (unshaded). In all cover types, mound subplots were constructed using soil collected with a bucket auger in each plot, outside of the subplots (two 15 cm by 25 cm cores per plot). Soil was coarse- sieved to 1 cm to homogenize soil texture and large clumps of root material were removed. For each mound, half the collected soil was poured into a cylindrical frame; removing the frame resulted in a mound 30 cm wide, the approximate size of natural gopher mounds in other coastal prairie sites (Peart 1989b; Di Vittorio et al. 2007). Holcus seeds were collected from the surround- ing population in August 2002. Subsamples were germinated in the lab to assess the number of viable seeds per mg and calculate the mass needed to obtain 2000 Holcus seeds/m^ (141 /subplot), which is one third of the Holcus seed rain measured in native-dominated patches in a heavily invaded coastal prairie site (Peart 1989a). Two days after the first rains in November 2002, seeds were added to all subplots by sprinkling seeds evenly inside a 30 cm-diameter frame; wet conditions prevented movement after seed additions (M. Thomsen personal observation). To estimate light availability across shading and disturbance treatments, photosynthetically active radiation (PAR) was measured at the soil surface using a hand-held quantum sensor (LI- COR Environmental, Lincoln, NE). Light mea- surements were taken in the center of each subplot within two hours of solar noon on a single clear day in February 2003. In March and August 2003, soil cores were collected from four randomly selected subplots per cover, mound, and shading combination. A pair of 2 cm by 7.5 cm cores were collected and combined from each sampled mound, I and a single 2 cm by 1 5 cm core was collected from ! each undisturbed subplot selected for sampling. All i cores were transported on ice and refrigerated prior ' to extraction. Soils were sieved through a 2 mm mesh and a weighed subsample was dried at 100°C and re-weighed to determine gravimetric percent water content. Approximately 13 g wet soil (10-12 g dry mass) was added to 50 mL 2M KCl and shaken for 45 min. Extracts were filtered through pre- washed Whatman 40 Quantitative Grade Filter Paper (Whatman Group, Middlesex, UK) and ij frozen until being analyzed for ammonium and nitrate concentration on a Lachat Autoanalyzer (Lachat Instruments, Loveland, CO). Ion concen- ' trations were corrected for molecular weight, percent water content and grams of non-gravel soil per cm^ to calculate pg available N per cm^ of soil. In mid May 2003 (end of wet season), all Holcus seedlings in the center 100 cm^ area of each subplot were counted and marked with a ; toothpick. Since some seedling mortality could have occurred before this time point, seedling numbers at this time are regarded as an estimate of seedling establishment rather than emergence. Subplots were re-censused in August 2003 (peak dry season) for living Holcus seedlings to evaluate patterns of survival across experimental treat- ments; no new unmarked seedlings were detected at this time point, and living plants at this time point generally were robust. All statistical analyses were conducted using IMP 10 (SAS Institute, Cary, NC). May and August seedling numbers and N availability data were square root transformed, light availability data were log transformed, and percent water content data were arcsine-square root trans- formed to meet the assumptions of normality and equality of variances. The mean number of Holcus seedlings per plot (averaging across subplots) was determined for May and August, and the effect of cover type on Holcus seed- ling numbers was determined using one-way ANOVA. Analyses of shading and soil treatment effects on seedling numbers were conducted separately for each species because of the different methods used to achieve shaded conditions in bare and Bromus as opposed to Festuca and Calamagrostis plots; a sequential Bonferroni technique (Rice 1989) was used to assure an overall error rate of <0.05 for each outcome variable. Establishment, survival, and light data were analyzed using a randomized complete block ANOVA model (plot as the blocking factor), with shading and disturbance treatments crossed with- in each plot. Because N and water were sampled in four randomly-selected subplots per treatment combination, the effects of experimental manip- ulations on those data were analyzed as a randomized incomplete block design. Pairwise differences were evaluated using Tukey tests. 2014] THOMSEN AND D’ANTONIO: HOLCUS AND DISTURBANCE IN COASTAL PRAIRIE 221 Soil condition Fig. 1. Establishment of Holcus seeds added to subplots in May (a) and the survival of seedlings until August of their first year (b). White bars represent means for unshaded/open canopy subplots and gray bars represent shaded/closed canopy subplots. Results Holcus Establishment and Survival Cover type significantly affected Holcus seed- ling establishment in May (F3 40 ^ 3.38, P < 0.05) and survival until August (^3^40 — 9.30, P < 0.0001). Tukey tests indicated no significant pairwise differences in May seedling numbers across cover types. In August however, there were on average seven surviving Holcus seedlings per 100 cm^ in Bromus plots, significantly more than the three found in Festuca and Calama- grostis plots. August seedling numbers in bare plots were intermediate. Within bare plots, shade structures increased Holcus seedling establishment in May from an average of 4.5 to 7.6 per 100 cm^, and undis- turbed areas of bare plots had more than twice as many Holcus seedlings as mounds (Fig. 1, Table 1). Shading and disturbance interacted to determine the number of Holcus seedlings that established in Bromus plots. There were an average of 1 1 Holcus seedlings in Bromus undisturbed and shaded mound subplots, more than three times as many as were found on unshaded mounds. Shading and soil disturbance interacted differently in their effects on Holcus establishment in Festuca and Calamagrostis plots. Undisturbed, shaded subplots had an average of two Holcus seedlings per 100 cm^, while all other subplots averaged eight Holcus seedlings in Festuca plots and seven in Calamagrostis plots. The number of Holcus seedlings sampled in bare plots in August was significantly affected by canopy condition and soil disturbance (Fig. 1, Table 1). Shade structures nearly doubled the number of Holcus survivors in bare plots, to an average of seven seedlings per 100 cm^; mounds decreased the number of seedlings in August by more than one half Shading and disturbance treatments interacted in their effect on Holcus seedling numbers in August in Bromus plots. Unshaded Bromus mounds had fewer than one- third as many Holcus seedlings as the other treatments, which averaged 9.5 seedlings. The number of Holcus seedlings sampled in August in Festuca and Calamagrostis plot types averaged three seedlings per 100 cm^ and was not significantly affected by shade manipulation or soil disturbance treatments. Light, N, and Water Availability Light at the soil surface was significantly lower inside shade structures constructed in bare and Bromus plots, by approximately 60% in bare plots and 90% in Bromus plots (Table 1, Fig. 2). In Bromus plots, mound creation increased PAR 222 MADRONO [VoL 61 Table 1 . Results of Significance Testing for the Effect of Experimental Manipulations on Holcus Seedling Numbers and on PAR at the Soil Surface. Data were analyzed separately for each species, and bolded P values indicate significant effects after a sequential Bonferroni technique was applied to assure P < 0.05 for each outcome variable. Canopy refers to shade structures created in bare and Bromus plots, and to whether the !, canopy was left in place or pinned back in Festuca and Calamagrostis plots. Soil refers to simulated gopher mounds in all plot types. I # May #August PAR Fi,30 P Fi,30 P Fi,30 P Bare Canopy 25.98 <0.0001 20.98 <0.0001 7.6 <0.01 Soil 73.34 <0.0001 44.6 <0.0001 0.07 0.79 Canopy*Soil 0.16 0.68 0.48 0.49 2.32 0.14 Bromus Canopy 1.76 0.19 2.9 0.1 222.15 <0.0001 Soil 10.03 <0,01 10.59 <0.01 38.29 <0.0001 Canopy*Soil 12.48 <0.01 11.23 <0.01 2.78 0.11 Festuca Canopy 23.58 <0.0001 5.72 0.23 142.7 <0.0001 Soil 10.69 <0.01 4.97 <0.05 7.2 <0.05 Canopy*Soil 18.59 <0.001 4.03 0.05 4.09 0.01 Calamagrostis Canopy 14.03 <0.001 6.45 <0.05 150.7 <0.0001 Soil 0.38 0.5 2.16 0.15 25.89 <0.0001 Canopy*Soil 12.04 <0.01 2.12 0.16 0.31 0.58 by 60%. Similarly, PAR in Festuca and Calama- grostis subplots in which the canopy had been pulled back were approximately 50 and 25 times higher, respectively, than in closed-canopy sub- plots. Soil disturbance also increased PAR in Calamagrostis plots; light availability was ap- proximately four times greater on mounds as compared to undisturbed subplots. Experimental manipulations had no effect on N availability in March (Table 2, Fig. 3). In August, soil N availability was unaffected by shading and disturbance treatments in bare and Bromus plots. August N levels were approximate- ly doubled on Festuca mounds relative to undisturbed areas, and nearly three times as great on mounds in Calamagrostis plots as compared to undisturbed subplots. Shading treatment interacted with disturbance in determining March water availability in bare plots, so that undisturbed areas and shaded mounds had similarly higher soil moisture levels, but soil moisture on unshaded mounds was approximately one-third as great (Table 3, Fig. 4). Shading increased March water avail- ability by 38% in Bromus plots, 46% in Festuca plots, and 27% Calamagrostis plots. At the same time, mounds decreased water availability by approximately 50% in Bromus, Festuca, and Calamagrostis plots. In August, shading in- creased water availability by 30% in Bromus plots, but had no effect on water availability in other cover types. In contrast, mounds signifi- cantly decreased water levels by approximately 60% in all plot types. Discussion As we predicted, the effects of shading and disturbance treatments depended on resident identity. Holcus establishment was generally increased by shade structures in bare and Bromus plots. The open conditions in these plot types likely increase summertime water deficits, so the positive effect of reduced water stress inside Soi! condition Fig. 2. Availability of photosynthetically active radiation at the soil surface in experimental subplots. White bars represent means for unshaded/open canopy subplots and gray bars represent shaded/closed canopy subplots. 2014] THOMSEN AND D’ ANTONIO: HOLCUS AND DISTURBANCE IN COASTAL PRAIRIE 223 Table 2. Results of Significance Testing for the Effect of Experimental Manipulations on March AND August Nitrogen Availability. Data were analyzed separately for each species; bolded P values indicate significant effects after a sequential Bonferroni adjustment was applied to assure an overall P < 0.05 for each month. Canopy refers to shade structures created in bare and Bromus plots, and to whether the canopy was left in place or pinned back in Festuca and Calamagrostis plots. Soil refers to simulated gopher mounds in all plot types. March N August N P P Bare Canopy Fi,,,.9 = 1.38 0.26 Fi,ii.2 = 8.31 0.01 Soil F,,i2 = 0.70 0.42 Fi,ii.3 = 11.29 0.01 Canopy*Soil Fi,ii.9 = 0.03 0.87 Fi,ii.2 = 1.82 0.2 Bromus Canopy Fi94 - 9.38 0.02 Fi,n.6 = 1.61 0.23 Soil Fi;9.7 = 7.95 0.01 Fi,7.o = 7.89 0.03 Canopy*Soil Fi,9.5 = 4.19 0.07 Fyio.S = .33 0.57 Festuca Canopy Fi,i2 = 8.31 0.58 Fi,io.i = 6.97 0.02 Soil Fi 12 = 0.50 0.01 Fi,„.9 = 55.38 <0,0001 Canopy*Soil Fi;i2 = 0.31 0.49 Fi,io.i - 3.64 0.09 Calamagrostis Canopy Fi,ii.i = 2.5 0.14 Fi,12.0 = 0.11 0.7 Soil Fi,ii.i = 1.05 0.33 Fi,i2.o = 25.26 <0.001 Canopy*Soil Fi, 10.04 = 0.31 0.59 Fi,i2.o = 0.06 0.8 the shade structures may have outweighed the negative effect of decreased light availability. Shade structures may have also functioned as wind shelters, which have been shown to affect plant performance at BMR (Lortie and Cushman 2007). In contrast, increased light availability in Festuca and Calamagrostis subplots in which the canopy had been pulled back increased Holcus establishment, despite lower water availability. Management actions such as mowing, grazing, and burning, which will increase light penetration to the soil surface in native bunchgrass stands, should be carefully evaluated in terms of their potential to create opportunities for Holcus establishment. The effects of soil disturbance also differed between bare and Bromus vs. Festuca and Calamagrostis plots. Mounds decreased Holcus establishment and survival in bare subplots and unshaded Bromus subplots, as might be expected given the open, harsh nature of the treatments. Soil condition Fig. 3. Nitrogen availability in experimental subplots in March (a) and August (b). White bars represent means for unshaded/open canopy subplots and gray bars represent shaded/closed canopy subplots. 224 MADRONO [Vol. 61 Table 3. Results of Significance Testing for the Effect of Experimental Manipulations on March AND August Water Availability. Data were analyzed separately for each species; bolded p values indicate significant effects after a sequential Bonferroni adjustment was applied to assure an overall P < 0.05 for each month. Canopy refers to shade structures created in bare and Bromus plots, and to whether the canopy was left in place or pinned back in Festuca and Calamagrostis plots. Soil refers to simulated gopher mounds in all plot types. March water August water P P Bare Canopy Fi,h.o = 27.48 <0.001 Fi,io.7 0.12 0.74 Soil Fi,9.8 = 62.53 <0.0001 Fi, 10.97 — 88.98 <0.0001 Canopy*Soil Fi, 11.04 — 18.90 <0.0001 Fi,10.7 - 1.14 0.31 Bromus Canopy Fi,io.o ~ 15.50 <0.01 Fi,i2.o - 37.97 <0.0001 Soil F,,8.2 = 88.97 <0.0001 Fi9o = 180.09 <0.0001 Canopy*Soil Fi,9.o = 5.42 0.05 Fi,ii.3 = 3.26 0.09 Festuca Canopy Fi,9.8 = 10.12 <0.01 Fi,6.5 = 8.18 0.26 Soil Fi,ii.8 = 26.6 <0.001 Fi 119 = 188.50 <0.0001 Canopy*SoiI Fi,9.8 = 2.8 0.12 f’i,6.5 = 3.07 0.13 Calamagrostis Canopy Fi,n.5 = 13.36 <0.01 Fi,io.9 = 4.05 0.07 Soil F,,,,.5 = 78.95 <0.0001 Fi,io.9 = 99.28 <0.0001 Canopy*Soil Fi,9.9 = 2.47 0.15 Fi,9.i = 0.14 0.02 The lower water availability on mounds in bare and Bromus plots may explain this pattern. We found no effect of a watering treatment on Hoicus invasion in our earlier work at the BMR (Thomsen and D’Antonio 2007), but this work did not involve gopher mounds. Also, Hoicus invasion was facilitated by water addition in a more interior and drier prairie site (Thomsen et al. 2006b). Mounds increased Hoicus seedling establishment in shaded Festuca and Calama- grostis subplots. The result is difficult to explain, since resource availability was either similar between mound and undisturbed shaded areas (light and March N levels) or lower on mounds (March water availability). It is possible that the soil disturbance directly affected Hoicus seed germination, resulting in higher rates of seedling establishment. Soil condition Fig. 4. Water availability in experimental subplots in March (a) and August (b). White bars represent means for unshaded/open canopy subplots and gray bars represent shaded/closed canopy subplots. 2014] THOMSEN AND D’ANTONIO: HOLCUS AND DISTURBANCE IN COASTAL PRAIRIE 225 Other researchers have found that gopher mounds increase N availability relative to undis- turbed areas in California grasslands (Canals et al. 2003; Eviner and Chapin 2005); similarly, we detected higher N availability on mounds in Festuca and Calamagrostis plots in August. Given the low water availability on mounds in August, however, Holcus seedlings are unlikely to benefit from increased nutrient availability at that time. In a greenhouse study, we found that Holcus was a dominant competitor against native and exotic perennial grass seedlings, regardless of N level (Thomsen et al. 2006a). Thus, even if mound creation increases N availability at times not sampled here, it may not strongly affect Holcus performance relative to native species. Overall, the results of this experiment do not support the hypothesis that small-scale distur- bances created by fossorial mammals are a dominant factor in promoting the invasion of California coastal prairie by Holcus lanatus. Open canopy microsites and soil disturbance decreased Holcus seedling numbers in bare and Bromus plots, and had no effect on Holcus survival over summer in Festuca and Calamagrostis plots. Instead, the highest overall number of Holcus seedlings in August was documented in undis- turbed Bromus plots. Thus, low competitive resistance from certain species provides Holcus with better conditions for establishment than those found on mounds. Bromus may even facilitate Holcus establishment through mild shading with insignificant competition, or through another unidentified mechanism. The relatively high establishment and survival we found in bare plots further suggests that larger- scale disturbances, such as those created by badgers or feral pigs, could promote Holcus invasion in coastal prairie (Kotanen 2004). Cush- man et al. (2004) found that pig disturbance decreased the biomass of mature exotic perennial grasses in a coastal prairie site, but seedling establishment data were not reported. Two other studies have examined Holcus colo- nization of gopher mounds in California coastal prairie, with somewhat contrasting results. DiVittorio et al. (2007) found that Holcus seedlings were second only to Air a caryophylla L. in their abundance on experimental gopher mounds in a coastal prairie site near the BMR. Their measurements were taken in May, which is also when we detected high numbers of Holcus seedlings mounds in Festuca and Calamagrostis subplots; longer-term seedling survival was not monitored. Peart (1989b) added Holcus seeds to natural gopher mounds in coastal prairie patches dominated by exotic annual grasses and found that, relative to undisturbed areas, mounds increased Holcus survival and seed production during the first year of growth. The contrast between our results and those of Peart (1989b) could be the result of weaker competition from annual grasses (as opposed to native perennials in our study), or annual- or site-level climatic variation; a cooler summer or an overall wetter site could have ameliorated the water stress experienced by seedlings on mounds. The higher seed density (80,000 per m^) used by Peart (1989b) may also have contributed to the higher levels of Holcus seedling establishment seen on mounds in that study. Here, the seed density (2000 per m^) was chosen to mimic Holcus seed dispersal from invaded areas into native grass- dominated patches, and is similar in magnitude to the 3856 Holcus seeds/m^ measured by DiVittorio et al. (2007) across grassland plots averaging 30% relative cover of Holcus. DiVit- torio et al. (2007) found that exotic plant dominance on mounds was positively correlated with propagule supply of exotic species, in keeping with the prediction that high propagule supply can overcome both biotic (e.g., competi- tion) and abiotic (e.g., water deficit) forms of resistance to invasion (D’ Antonio et al. 2001). Conclusions The combined results of this study and of Thomsen and D’Antonio (2007) highlight how patterns of invasive plant establishment vary across the landscape. Here, we found that invader establishment and the net effect of disturbance on invasion depends on resident identity. In Thom- sen and D’Antonio (2007) we documented the same pattern of Holcus seedling establishment across native plot types, and further illustrated how native species differences also affected the influence of underlying variation in environmen- tal conditions. Holcus seedlings in Bromus plots benefited due to the better abiotic conditions found lower on a hillslope, but in Calamagrostis plots, the low areas with beneficial conditions were protected by competitive residents (Thom- sen and D’Antonio 2007). Accounting for natural variation in the potential for invasive plant establishment across the landscape will allow us to fine-tune management strategies and more successfully decrease the abundance and spread of invasive species. Acknowledgments This work was supported by a National Science Foundation Graduate Research Fellowship, a Uni- versity of California Natural Reserve System Mildred E. Matthias Student Research Grant, and a Marshal and Nellie Alworth Memorial Fund Scholarship. Thanks to the D’Antonio and Sousa labs at the University of California, Berkeley, and to Peter Connors, Rico Tinsman, and Jackie Sones at the Bodega Marine Reserve for technical advice and logistical assistance. 226 MADRONO [Vol. 61 Literature Cited Cahill, J. F. and B. B. Casper. 2002. Canopy gaps are sites of reduced belowground plant competition in a productive old field. Plant Ecology 164:29-36. Cal-IPC. 2013. Invasive Plant Inventory. California Invasive Plant Council, Berkeley, CA. Website http://www.cal-ipc.org (accessed 06 May 2013). Canals, R. M., D. J. Herman, and M. K. Fire- stone. 2003. How disturbance by fossorial mam- mals alters N cycling in a California annual grassland. Ecology 84:875-881. Cushman, J. H., T. A. Tierney, and J. M. Hinds. 2004. Variable effects of feral pig disturbances on native and exotic plants in a California grassland. Ecological Applications 14:1746-1756. D’Antonio, C. M. 1993. Mechanisms controlling invasion of coastal plant communities by the alien succulent Carprobrotus edulis. Ecology 74:83-95. , J. Levine, and M. Thomsen. 2001. Ecosystem resistance to invasion and the role of propagule supply: a California perspective. Journal of Med- iterranean Ecology 2:233-245. Dethier, M. N. and S. D. Hacker. 2005. Physical factors vs. biotic resistance in controlling the invasion of an estuarine marsh grass. Ecological Applications 15:1273-1283. DiVittorio, C. T., j. D. Corbin, and C. M. D’Antonio. 2007. Spatial and temporal patterns of seed dispersal: an important determinant of grassland invasion and restoration potential. Eco- logical Applications 17(2):31 1-316. Eviner, V. T. AND F. S. Chapin III. 2005. Selective gopher disturbance influences plant species effects on nitrogen cycling. Oikos 109:154-166. Foster, B. L., V. H. Smith, T. L. Dickson, and T. Hildebrand. 2002. Invasibility and compositional stability in a grassland community: relationships to diversity and extrinsic factors. Oikos 99:300-307. Gerlach, j. G. and K. j. Rice. 2003. Testing life history correlates of invasiveness using congeneric plant species. Ecological Applications 13:167-179. Goldberg, D. E. 1987. Seedling colonization of experimental gaps in two old-field communities. Bulletin of the Torrey Botanical Club 1 14:139-148. Grime, J. P. 1979. Plant strategies and vegetative processes. John Wiley & Sons, Ltd., Chichester, England. Hobbs, R. J. and H. A. Mooney. 1991. Effects of rainfall variability and gopher disturbance on serpentine annual grassland dynamics. Ecology 72:59-68. Kolb, A., P. Alpert, D. Enters, and C. Holzapfel. 2002. Patterns of invasion within a grassland community. Journal of Ecology 90:871-881. Kotanen, P. M. 1997. Effects of experimental soil disturbance on revegetation by natives and exotics in coastal Californian meadows. Journal of Ap- plied Ecology 35:631-644. . 2004. Revegetation following soil disturban- ce and invasion in a Californian meadow: a 10- year history of recovery. Biological Invasions 6:245-254. Lortie, C. and j. H. Cushman. 2007. Effects of a directional abiotic gradient on plant community dynamics and invasion in a coastal dune system. Journal of Ecology 95:468^81. Milbau, a., L Nijs, F. De Raedemaecker, D. Reheul, and B. de Cauwer. 2005. Invasion in grassland gaps: the role of neighborhood richness, light availability and species complementarity during two successive years. Functional Ecology 19:27-37. Peart, D. R. 1989a. Species interactions in a succes- sional grassland. 1. Seed rain and seedling recruit- ment. Journal of Ecology 77:236-251. . 1989b. Species interactions in a successional grassland. III. Effects of canopy gaps, gopher mounds and grazing on colonization. Journal of Ecology 77:267-289. Pitcher, D. and M. J. Russo. 1988. Element stewardship abstract for Holcus lanatus. The Nature Conservancy, Arlington, Virginia, USA. Platt, W. J. 1975. The colonization and formation of equilibrium plant species associations on badger disturbances in a tall grass prairie. Ecological Monographs 45:285-305. Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43:223-225. Seabloom, E. W., O. N. Bjornstad, B. M. Bolker, AND O. J. Reichman. 2005. Spatial signature of environmental heterogeneity, dispersal and compe- tition in successional grasslands. Ecological Mono- graphs 75:199-214. Sousa, W. P. 1984. The role of disturbance in natural communities. Annual Reviews of Ecology and Systematics 1 5 : 353-39 1 . Stromberg, M. R. and J. R. Griffin. 1996. Long- term patterns in coastal California grasslands in relation to cultivation, gophers, and grazing. Ecological Applications 6:1189-1211. Thompson, J. D. and R. Turkington. 1988. The biology of Canadian weeds. 82. Holcus lanatus L. Canadian Journal of Plant Science 68:131-147. Thomsen, M. A., J. D. Corbin, and C. M. D’Anto- nio. 2006a. The effect of soil nitrogen on compe- tition between native and exotic perennial grasses from northern coastal California. Plant Ecology 186:23-35. , C. M. D’Antonio, K. B. Suttle, and W. P. Sousa. 2006b. Ecological resistance, seed density, and their interactions determine patterns of inva- sion in a California coastal grassland. Ecology Letters 9:160-170. AND . 2007. Mechanisms of resistance to invasion in a California grassland: the roles of competitor identity, resource availability, and environmental gradients. Oikos 116:17-30. Watt, A. S. 1947. Pattern and process in the plant community. Journal of Ecology 35:1-22. Madrono, VoL 61, No. 2, pp. 227-230, 2014 A NEW SPECIES OF TRITELEIA (THEMIDACEAE) FROM THE SOUTHERN SIERRA NEVADA Ed Kentner 1917 S Myers St. #D, Oceanside, CA 92054 ed@kentnerbo tanical . com Kim Steiner 1791 Inverness Dr., Petaluma, CA 94995, and School of Biological and Conservation Sciences, University of KwaZulu-Natal, P. Bag XOl, Scottsville, Pietermaritzburg 3209, South Africa Abstract Triteleia piutensis E. Kentner & K. Steiner is described as a new species from the Piute Mountains of the southern Sierra Nevada in Kem County, California where it is known from only two small populations. Both populations are located in openings in piny on-juniper woodland on volcanically derived soils. It is distinguished from other Triteleia species that have stamens attached at two different levels by its erect, yellow, campanulate flowers with recurved lobes, its short ovary stalk, and a geographic range restricted to the extreme southern Sierra Nevada Mountains. A revised key to Triteleia species with stamens attached alternately at two levels is provided. Key Words: Piute Mountains, Sierra Nevada, Themidaceae, Triteleia. Triteleia Douglas ex Lindl. (Themidaceae) is a genus of 15 species that is widely distributed in western North America with a center of diversity in the Klamath region of northern California and southern Oregon (Fires 2003). Until recently, Triteleia has been assumed to be closely related to Brodiaea Smith and Dichelostemma Kunth based on the presence of an extended perianth tube (Keator 1989). However, despite the extend- ed tube and similarities with those genera in other morphological characters (e.g., corm, leaf, an- thers, stamen appendages, scape pubescence, seeds, and chromosome number [Hoover 1955]), analyses of plastid DNA sequences by Pires and Sytsma (2002) suggest that Triteleia is more closely related to Bloomeria Kellogg than to either Brodiaea or Dichelostemma. These analyses have also indicated that these genera are best placed in the Themidaceae, rather than in the Amaryllida- ceae or Liliaceae (Pires and Sytsma 2002). The most important diagnostic characters within Triteleia are features of the androecium and gynoecium. The insertion of the stamens relative to the perianth, the lengths of the filaments, and the presence or absence of apical filament appendages are major features used in distinguishing among the species, as are the relative lengths of the gynophore (ovary stalk) and ovary. Other useful taxonomic characters include the overall size and shape of the perianth, the relative length of the perianth tube and lobes, and the length of the pedicels (Hoover 1941; Pires 2003; Pires and Keator 2012). However, in several Triteleia species, these characters can be highly variable, and in at least one species, T. grandiflora Lindley, diploid and tetraploid indi- viduals have been shown to differ in floral morphology (Barkworth 1977). Chromosomal changes, including polyploidy and aneuploidy, are common in Triteleia and seem to be the primary mode of evolution within the genus (Burbanck 1941; Lenz 1975, 1976). While most Triteleia species occur in northern California and the Pacific Northwest, three species - T. hyacinthina (Lindl.) Greene, T. laxa Benth., and T. dudleyi Hoover - extend into southern California in the San Gabriel Mountains and foothills of the San Bernardino Mountains (Consortium of California Herbaria 2013). Two species, T. grandiflora and T. hyacinthina, range east into Montana and Wyoming, and T. lemmoniae (S. Watson) Greene is found only in central Arizona (Pires and Sytsma 2002). Two island endemics are known, T. guadalupensis L. W. Lenz from Guadalupe Island off the coast of Baja California, and T. Clementina Hoover, on San Clemente Island off the southern California coast. Hoover (1941) noted that Triteleia is represented in the northern part of its range primarily by species of wide distribution, while in the south the species have localized distributions and are widely separated geographically. On the mainland, Triteleia is uncommon south of the Tehachapi Mountains (Hoover 1955) and its southern limit is demarcated by a handful of sites in the San Gabriel and San Bernardino Moun- tains (Consortium of California Herbaria 2013). During a botanical survey in the Piute Moun- tains of the southern Sierra Nevada in 2010 a Triteleia population was discovered that could 228 MADRONO [Vol. 61 Fig. 1. Photograph of Triteleia piutensis (short- statured form pictured from the population north of Emerald Mountain, Kem County, California). Note the erect, yellow, campanulate flowers with recurved perianth lobes. Photograph courtesy of Jason Brooks. not be referred to any known species using Keator (1993), Pires (2003), or Munz (1974). Photographs and descriptions of the plants were sent to several experts on the genus (Chris Pires, Glenn Keator, and Lee Lenz) with the consensus that it could be an undescribed species. In February 2013, we learned that another population of this species had been collected by Eve Laeger in 2001 about nine kilometers (km) south of the population that was discovered in 2010. This new species is described herein, and a revised key to the Triteleia species with unequally inserted stamens is provided. Taxonomic Treatment Triteleia piutensis E. Kentner and K. Steiner, sp. nov. (Figs. 1, 2).— TYPE: USA, California, Kem Co., Southern Sierra Nevada Mountains, approximately 3.2 km by air north of Emerald Mountain on ridgeline between Back Canyon and Indian Creek, UTM NAD 83 Zone 11 3841 13E, 3905222N/ 35°17'0.4"N, 1 18°16'27.6'W (WGS 84), 1655 m, 3 June 2010, O. Singh 1149 (holotype: UCR). Corm with coarse fibrous coat, 0.7-1, 5 dm below ground level; Leaves 1-2, 2-8 mm wide, 15^14 cm long; scape smooth to slightly scabrous at base, 1 .2- 14 cm; bracts purplish; pedicels 2-25 mm; perianth 11-21 mm, yellow with a central abaxial ± green to brown/purple stripe that broadens distally, tube 5.5- 10 mm, tapered at base, lobes 6-13 mm, ascending and recurved, not spreading; stamens inserted alternately at two levels with the upper level 1- 2 mm above the lower level, unequal in length, short filaments 1. 7-4.5 mm and long filaments 3-5 mm, linear to slightly thickened at base, anthers 1.8- 2.5 mm, lacking appendages, pollen white; ovary 2-6 X stalk, green, longer than the stalk in fruit. Fig. 2. Line drawings (by Fred Roberts) of Triteleia piutensis. A. Tall-statured form; B. Short-statured form; j C. Whole plant; D. Corm; E. Leaf shape and cross section; F. Inflorescence detail; G. Inflorescence bracts; H. Flower top view; 1. Flower side view; J. Flower internal j detail; K. Stamen attachment detail; L. Developing fruit | enclosed in tepals; M. Mature fruits shown with and without enclosing tepals; N. Mature seed. All scale bars | equal 1 cm except for A and B (1 dm), and N (2 mm). ' Paratype: USA, California, Kern Co., South- ern Sierra Nevada Mountains, approximately ' 15.3 km by air northeast of Tehachapi on the ridge west of Horse Canyon about 100 m east of a prominent area of white volcanic soils, UTM NAD 83 Zone 11 381169E, 3896484N / 35°12T5.6'TS1, 118°18T9.6"W (WGS 84). 1585 m, 16 May 2001, E. Laeger 3965 (UCR). Figs. 1, 2. ! Etymology The specific epithet is named for the Piute Mountain range in the extreme southern Sierra i Nevada, where the plants were discovered. These Mountains were unglaciated during the Ice Ages (Hill 2000), and are known to support many rare and uncommon vascular plants. Taxa endemic to the range include Hesperocyparis nevadensis (Abrams) Bartel, Eriogonum kennedyi Porter ex S. Watson var. pinicola Reveal, Streptanthus cordatus Nutt. var. piutensis J. T. Howell, and possibly the new Triteleia species described here, j 2014] KENTNER AND STEINER: A NEW TRITELEIA FROM THE SIERRA NEVADA 229 although the extent of its distribution remains unknown. Distinguishing Characteristics In live plants, the most obvious characteristic distinguishing T. piutensis from its congeners are its erect, bright yellow, campanulate flowers, with ascending recurved perianth lobes that do not spread at anthesis and enclose the capsule in fruit. The pedicels are ascending and the flowers and fruit are held erect. Among the yellow-flowered Triteleia species known to occur in the vicinity of the southern Sierra Nevada, it is easily distin- guished from T. ixioides (W.T. Aiton) Greene and T. dudleyi, which differ in having stamens inserted on the perianth at a single level and filaments with distinctive shapes and/or tip appendages (Fires and Keator 2012). Taxonomic Relationships Triteleia piutensis is morphologically most similar to T. crocea (Alph. Wood) Greene, a species assumed to be restricted to the Klamath and High Cascade Ranges of northern California and southwest Oregon (Fires 2003), but with the interesting exception of a single 1954 collection from Kern County {Hoover 8337 [RSA, SD, UC]) discussed below. Both species have stamens attached alternately at two levels and linear filaments of alternate lengths. The filament lengths are similar in the two species, but the anthers are slightly larger in T. piutensis (1 .8-2.5 mm vs. 1.5-2 mm in T. crocea [Hoover 1941; Fires and Keator 20 1 2]) . The species differ in the relative length of the ovary stalk, which is about equal to, or slightly longer than the ovary in T. crocea (Hoover 1941) and distinctively short in T. piutensis, with the ovary 2-6 X longer than the stalk. The campanulate perianth of T. piutensis is distinctive and contrasts with the widely spreading perianth lobes of T. crocea. The two species also differ in scape height (0.12-1.4 dm in T. piutensis vs. 1-3 dm in T. crocea), although this character may be influenced by growing conditions in T. piutensis, as discussed below. Hoover’s collection of T. crocea from Kern County remains somewhat of a mystery as it represents a range extension of more than 650 km. The site of his collection is located about 45 km north of the Triteleia populations described here, and it was initially suspected that Hoover’s collection could represent another population of T. piutensis. However, duplicates of Hoover’s collection (SD71153, RSA201599) have ovary stalks equal in length to the ovaries, and do not closely resemble T. piutensis in perianth shape or habit. Hoover (1955) noted that his Kern County collection of T. crocea had smaller anthers than specimens from northern California and Oregon, but could find no other characters distinguishing the southern plants. While the possible existence of T. crocea in Kern County warrants further study, the available material appears to be distinct from T. piutensis. Fhenology Triteleia piutensis flowers from May to June and fruits in June to July. Habitat, Ecology, and Conservation Implications Triteleia piutensis is currently known from two populations in the extreme Southern Sierra Nevada Mountains separated by about nine km in distance and ca. 70 m in elevation. The southern population, discovered by Eve Laeger in 2001, occurs on the ridge west of Horse Canyon, 15.3 km northeast of Tehachapi in fine volcanic soils among scattered boulders. The vegetation consists of open woodland dominated by Juniperus californica Carriere with occasional Pinus monophylla Torr. & Frem. Other associated species include Ericameria linear if oHa (DC) Ur- batsch & Wussow, Ericameria teretifolia (Durand & Hilg.) Jeps., Hesperoyucca whipplei (Torr.) Trek, Lewisia rediviva Fursh, Allium cratericola Eastw., Poa secunda J. Fresl, and Bromus tectorum L. This population occurs on private land about 100 m west of a parcel under the jurisdiction of the Bureau of Land Management (BLM). Several hundred plants were observed in two patches on 16 May 2001, but only about a 30 plants could be found when the site was revisited on 27 April 2013, a year in which conditions were quite dry. The northern population, discovered by Jason Brooks in 2010, is located 3.2 km north of Emerald Mountain on an approximately 0. 1 5 acre flat ridge top opening in a woodland dominated by P. monophylla and Quercus john-tuckeri Nixon & C. H. Mull. The exposed heavy clay soils of the opening are underlain with a hardpan derived from volcanic rocks. Associated species include Poa secunda, Perideridia pringlei (J. M. Coult. & Rose) A. Nelson & J. F. Macbr., Calochortus kennedyi Forter, and sparse annuals like Micro- steris gracilis (Hook.) Greene, Rigiopappus iepto- cladus A. Gray, and B. tectorum. In contrast to the plants of the southern population which have scapes that average about 12 cm in length (Fig. 2A), all of the individuals in the northern population have inflorescences born nearly at ground level on short scapes averaging about 2 cm in length (Fig. 2B), perhaps due to the clay soils and hardpan at the site. This short-statured population occurs on land managed by the BLM within a few meters of the boundary fence and the adjacent private property. In two successive years (3 June 2010 and 23 June 2011), about 75 and 120 individuals were observed, respectively. 230 MADROto [VoL 61 Despite intensive botanical surveys in 2010, 2011, and 2012 of several thousand acres adjacent to the northern population, no new occurrences of T. piutensis were found. Although there are public lands to the north of the population in the Sequoia National Forest and a checkerboard of BLM lands nearby, much of the area surrounding the site is remote and privately owned and has been poorly explored botanically. The flowers of T. piutensis are quite showy and distinctive, yet the species does not appear to have been previously collected despite intensive exploration of the southern Sierra Nevada by notable plant collectors and botanists such as Ernest Twisselmann and James Shevock. We conclude that T. piutensis is a rare narrow endemic restricted to the Piute Mountains in the southern Sierra Nevada. Nevertheless, additional populations may remain to be discovered. Both of the known populations of this species occur in the vicinity of the Tehachapi wind resource area where several large wind energy developments have recently been constructed. If additional populations exist, T. piutensis is likely to face increasing threats from development unless developers, consultants, and land managers are made aware of its existence, and surveys to determine its presence are conducted during the environmental review for proposed projects in the area. Key To Triteleia Species With Stamens Attached Alternately At Two Levels A revised key, after Hoover (1941) and Pires and Keator (2012), for the Triteleia species with stamens inserted at two levels is provided below. 1 . Ovary > stalk; perianth tube rounded or short-tapered at base 2. Perianth 17-35 mm, white to blue-purple; filaments triangular; CaR, KR, ......... Triteleia grandiflora 27 Perianth 11-21 mm, yellow; filaments linear; s SNF. ...................... Triteleia piutensis r Ovary <= stalk; perianth tube acute to long- tapered at base 3. Stamens unequal in size 4. Pedicel 7-20 mm; perianth 12-19 mm, bright yellow or pale blue; ovary green ..... Triteleia crocea 4' Pedicel 20-180 mm; perianth 15-28 mm, white, often flushed violet abaxially; ovary bright yellow. .................................................... Triteleia peduncularis 3' Stamens ± equal in size 5. Perianth 16-27 mm, lavender; anthers ±1.5 mm, purple; s Chi (San Clemente Island) ........................................................... Triteleia Clementina 5' Perianth 18M7 mm, blue, blue-purple, or white; anthers 2-5 mm, white to ± blue; NW, CaR, SN, CW, TR .................................................. Triteleia laxa Acknowledgments We acknowledge and thank Jason Brooks, Eve Laeger, Chloe Scott, Onkar Singh, and Catherine Schnurrenberger whose careful botanical field work contributed to the discovery of this new taxon. We thank Nick Jensen and Fred Roberts for providing useful comments and suggestions on the manuscript, and Fred Roberts for preparing the illustration. We also thank the herbarium staff at SD, RSA, and UCR for allowing access to their facilities and materials. Literature Cited Barkworth, M. E. 1977. Intraspecific variation in Brodiaea douglasii Watson (Liliaceae). Northwest Science 51:79-90. Burbanck, M. P. 1944. Cytological and Taxonomic Studies in the genus Brodiaea. II. Botanical Gazette 105:339-345. Consortium of California Herbaria (CCH). 2013. Data provided by the participants of the Consortium of California Herbaria. Website http.7/ucjeps.berke- ley.edu/consortium/ [accessed 5 February 2013], Hill, M. 2000. Geologic story. Pp. 37-69 in G. S. Smith (ed.), Sierra East: edge of the great basin. University of California Press, Berkeley, CA. Hoover, R. F. 1941. A systematic study of Triteleia. American Midland Naturalist 25:73-100. . 1955. Further observations on Brodiaea and some related genera. Plant Life 11:13-22. KeatO'R, G. 1989. The brodiaeas. Four Seasons 8:4-1 1. . 1993. Triteleia. Pp. 1206-1208 in J. C. Hickman (ed.), The Jepson manual: higher plants of Califor- nia. University of California Press, Berkeley, CA. Lenz, L. W. 1975. A biosystematic study of Triteleia (Liliaceae): 1. Revision of the species of section Caliiprora. Aliso 8221-258. . 1976. A biosystematic study of Triteleia (Lilia- ceae): 2, Chromosome numbers and karyotypes of the species of section Caliiprora. Aliso8353-377. Munz, P. 1974. A flora of southern California. University of California Press, Berkeley, CA. Pires, J. C. 2003. Triteleia. Pp. 321-339 in Flora of North America Editorial Committee (eds.). Flora of North America North of Mexico, VoL 26: Magnoliophyta: Liliidae: Liliales and Orchidales. Oxford University Press, New York, NY. AND K. J. Sytsma. 2002. A phylogenetic evaluation of a biosystematic framework: Brodiaea and related petaloid monocots (Themidaceae). American Journal of Botany 89:1342-1359. and G. Keator. 2012. Triteleia. Pp. 1512-1514 in B. G. Baldwin, D. H. Goldman, D. J. Keil, R. Patterson, T. J. Rosatti, and D. H. Wilken (eds.), The Jepson manual: vascular plants of California, 2nd ed. University of California Press, Berkeley, CA. Madrono, VoL 61, No. 2, pp. 231-243, 2014 CYLINDROPUNTIA CHUCKWALLENSIS (CACTACEAE), A NEW SPECIES FROM RIVERSIDE AND IMPERIAL COUNTIES, CALIFORNIA Marc A. Baker College of Liberal Arts and Sciences, School of Life Sciences, Arizona State University, P.O. Box 874501, Tempe, AZ 85287-4501 mbaker6@asu.edu Michelle A. Cloud-Hughes Desert Solitaire Botany and Ecological Restoration, San Diego, CA 92103 Abstract A gynodioecious hexaploid {n = 33), Cylindropuntia chuckwallemis M. A. Baker & M. A. Cloud- Hughes, is newly described. Populations of C. chuckwallensis extend from the Eagle Mountains of Joshua Tree National Park, through the Chuckwalla Mountains of Riverside County, to the north side of the Chocolate Mountains in Imperial County, California, USA, and occur on a variety of substrates primarily between 400-1600 m (1312-5250 ft) elevation. Of the flowering individuals studied, 38% produced only pollen-sterile flowers. Flower color in C. chuckwallensis ranges from dark red-purple (33%) through orange (54%) to yellow (13%). For most individuals (93%) the style and filaments are dark red to light pink. Morphological measurements were made for 15 populations of Cylindropuntia, including four of C. chuckwallensis, four of C. echinocarpa, three of C. multigeniculata, and four of C. acanthocarpa. Multivariate analyses indicated that C. chuckwallensis possesses a unique combination of characters. Fewer than 3% of the 121 C chuckwallensis individuals sampled were misclassified by discriminate function analysis, one as C. echinocarpa, and three as C. multigeniculata. Key Words: Cylindropuntia chuckwallensis, Cactaceae, endemism, evolution, gynodioecy, new species, polyploidy, speciation. The study of Cylindropuntia (Engelm.) F. M. Knuth continues to engage biologists attempt- ing to understand the origins and relationships of its taxa and their underlying evolutionary mechanisms. Common processes associated with the genus include eupolyploidy, hybrid- ization, and apogamy (Baker and Pinkava 1987). Aneuploidy appears to be rare in Cactaceae and has not been documented in Cylindropuntia (Baker et al. 2009). Vegetative reproduction is common and is the only form of apogamy recorded for the genus. However, adventive embryony has been reported for Opuntia Desv. (Davis 1966) and may be respon- sible for the primary mode of reproduction in some species such as the pentaploid allopolyploid O. X charlestonensis Clokey (pro. sp.) (Beard 1937; Baker et al. 2009). There are several hexaploid species known for Cylindropuntia. These include C arbuscula (En- gelm.) F. M. Knuth, C. calmalliana (J. M. Coult.) F. M. Knuth, C sanfelipensis (Rebman) Rebman, and C. wolfii (L. D. Benson) M. A. Baker (Pinkava and McLeod 1971; Pinkava et al. 1992, 1998). All of these, except C arbuscula which is largely, if not entirely, clonal, are gynodioecious (Rebman 1998; Rebman and Pinkava 2001) with some individuals having perfect flowers and others having pollen-sterile flowers. One primarily octoploid species, C. molesta (Brandegee) F. M. Knuth, is also gynodioecious (Pinkava et al. 1998; Rebman and Pinkava 2001). Correlation between polyploidy and subdioecy has also been observed in other cactus genera. Most Consolea Lem. species, for example, are both hexaploid and subdioecious (Negron-Ortiz 2007). In Echinocereus Engelm. and Mammillaria Torr. & A. Gray, subdioecy is associated only with tetraploid and hexaploid taxa (Remski 1954; Johnson 1980; Parfitt 1985; Pinkava et al. 1985, 1998; Baker 2006). However, at least two Opuntia taxa, O. stenopetala Engelm. and O. grandis Pfeiff., are diploid and subdioecious. Thus, at least in some cactus genera, gender dimorphism has been documented only for polyploids, with the majority of subdioecious taxa being hexa- ploid. Selection pressure for floral dimorphism may be greater in polyploids because of a decrease in self-incompatibility (Marks 1966). Male-sterile mutants may be favored as a response to inbreeding depression resulting from this reduction in self-incompatibility (Miller and Venable 2000). Pressure for outbreeding is also a logical consequence for an almost certainly low population size for incipient allopolyploids. 232 MADRONO [VoL 61 We herein describe a new hexaploid gynodioe- cious Cylindropuntia species, Cylindropuntia chuckwallensis M. A. Baker & M. A. Cloud- Hughes, from the Chuckwalla Mountains region of Riverside and Imperial counties, California, and test the significance of differences among mean values for certain morphological characters among populations of this new species and the two most morphologically similar species, C echinocarpa (Engelm. & J. M. Bigelow) F. M. Knuth and C multigeniculata (Clokey) Backeb. Populations of C. acanthocarpa have been included as an outgroup. Populations of Cylindropuntia multigeniculata have been referred to as a nothospecies, C X multigeniculata Backeb. (Pinkava 1999, 2003), with the putative parents being C echinocarpa and C whipplei (Engelm. & J. M. Bigelow) F. M. Knuth. However, the determination of hybrid status was made without supporting data and disregarded data from Trushell (1985). Some of the confusion lies in the identification by Pinkava (2003) as compact forms of C echinocarpa of what we now consider to be Arizona populations of C multigeniculata (Baker 2002). Except for the somewhat spinier fruits, the Arizona individuals are morphologically similar to those at the type locality of C multigeniculata. Populations of C multigeniculata possess several characters that are not intermediate between those of the putative parents, and the means of these characters are significantly different from those of either of the putative parents. Morphologically, the effects of hybridization within the Opuntioideae appear to be additive. Milhalte and Sestras (2012) found a range of heritability between 0.909 and 0.948 among a suite of morphological characters for Fi indi- viduals of artificial hybridization within Cacta- ceae, indicating that the characters analyzed had a strong genetic determinism. Statistical analy- ses of natural cactus hybrids have had similar results, where the morphology of hybrids exhibits the expected degree of intermediacy between that of putative parents (Baker and Pinkava 1984; Powell et al. 1991; Vite et al. 1996; Powell and Weedin 2004). However, with respect to hybridization in Cactaceae, flower pigmentation appears to be less predictable (Powell 2002). Specimens of Cylindropuntia chuckwallensis have been identified historically as C. echino- carpa. Philip Munz, David Keck, and M. French Gilman collected samples of C. chuckwallensis at Corn Springs in 1922 (POM11971, POM 13960), while Willis L. Jepson (JEPS66875) and Frank Peirson (RSA65624) collected samples at Cottonwood Springs on consecutive days in 1928, with Peirson noting the “peculiar dull reddish purple” of the flowers on the herbarium label. Materials and Methods Fieldwork was conducted between May 2011 and May 2013. Figure 1 presents an overview of the study sites for Cylindropuntia chuckwallensis, and descriptions of all study sites are presented in Table 1. Samples of 30-35 individuals were measured for each population. In studies of the Cactaceae, 30 individuals account for approxi- mately 90% of the variation within most charac- ters in a single population and provide a good balance between field effort and statistical robustness (Baker and Butterworth 2013). In order to account for geographical effects, we suggest that at least three populations be sampled for any putative taxon and that these populations be spread over the geographical range as widely as possible. This precept stems from the fact that geographically isolated populations, even within the same taxon, often have a rather uniform morphology. With discriminant analyses, indi- viduals within such populations are sometimes classified correctly for that population nearly 100% of the time. If all such populations were recognized as separate taxa, the number of taxa within the Cactaceae would be unmanageable. Thus, this type of analysis may not be suitable for potential taxa with only one or two known populations or with only a few known individu- als. We define a population here as a collection of individuals that, as far as can be determined, exchange genes frequently among themselves but rarely, if ever, exchange genes with other populations. In addition to the four populations sampled for our new taxon, we also sampled five populations of Cylindropuntia acanthocarpa (Engelm. & J. M. Bigelow) F. M. Knuth, four populations of C. echinocarpa (Engelm. & J. M. Bigelow) F. M. Knuth, and three populations of C. multigenicu- lata (Clokey) Backeb. Cylindropuntia acantho- carpa was sampled as a morphological outgroup. Populations of C. echinocarpa, which are sym- patric with the new species, and those of C multigeniculata, which are allopatric, were includ- ed because of their morphological similarity to the new species. Thirteen stem characters were measured for 452 individuals among the 15 populations (Table 2). It was assumed that all individuals within a single population belong to a single taxon except for the co-existing populations of C. acanthocarpa and C. echinocarpa west of Needles, California, and the Joshua Tree Na- tional Monument and Graham Pass Road populations of C chuckwallensis, where there were a few easily-distinguished individuals of C. echinocarpa. All characters were measured three times per individual, and the values for each character were averaged before any statistical analyses were performed. Spine and sheath thick- 2014] BAKER AND CLOUD-HUGHES: A NEW CYLINDROPUNTIA FROM CALIFORNIA 233 HS^SOW 115M0'W ns^aow 115'’20'W 115°10’W 115°0’W Fig. I. Geographic distribution of Cylindropuntia chuckwallensis and locations of study sites. nesses were measured to the nearest 0.01 mm using digital calipers. Repeated measurements were performed to increase the number of possible values for each character. Data were then imported into Systat® 10 (Systat Software, Inc., Chicago, IL), SPSS® 12 (IBM, Inc., Armonk, NY), and NTSYSPC® (Exeter Soft- ware, Setauket, NY). Discriminate function analysis (DFA) was performed to ascertain the correct classification of individuals for each taxon, and MANOVA was used to determine which character means differed significantly among the various taxa. Results Field surveys indicate that populations of Cylindropuntia chuckwallensis are restricted to areas in and around the Chuckwalla Mountains of Riverside County, California. To date, the known northern extent of the species is in the Eagle Mountains of Joshua Tree National Park, and the southern extent is on the north side of the Chocolate Mountains in the northernmost por- tion of Imperial County (Fig. 1). The total range is approximately 50 km north to south and 60 km east to west. There are approximately 3200 hectares of known occupied habitat for C. chuckwallensis. Although densities of individuals are as high as five or more individuals per hectare, the average density is probably closer to 2.5 individuals per hectare. Cylindropuntia chuckwallensis is apparently gynodioecious. Of 151 individuals observed in flower, 38.4% produced only pollen-sterile flowers, and 61.6% produced only pollen-fertile flowers, with pollen-fertile individuals producing seeds. Flower color ranged from dark red-purple (33.1%) through orange (54.3%) to yellow (12.6%). For most individuals (92.7%), the style and filaments were dark red to light pink, at least toward their apices (Fig. 2). Chromosome determinations of n = 33 were made for 13 individuals of Cylindropuntia chuck- wallensis; and n = 11 for one individual of C. acanthocarpa, one individual of C. echinocarpa, and five individuals of C. multigeniculata (Table 1). One chromosome determination was made for an individual of C. chuckwallensis that was not associated with the sites listed in Table 1: California, Riverside County, along Salt Creek, 17.5 km WNW of Black Butte, 18 km SE of Chiriaco Summit, 33.5866°, —115.5521°, M. A. Baker 17533 (ASU, UCR). A chromosome deter- mination of n = 11 was made for two individuals (M. A. Baker 15259 [ASU, RSA]; M. A. Baker 15260 [ASU, US]) at sample site 11 (Table 1), 234 MADRONO [VoL 61 < Os a M 2^ o ^ £ w Cd !- PQ 3: S H O O cu O II fi K X > < U • g-=S O Q > g^ll s s ffl 2 ^ < - 5 0) ^ S| -I ^ g s z x> Cl 0,00 « a uf 3 2 Z >. , O (N C d ON < o as u > C O a . hJ Cl St 00 0) ^ cd ^ ^ ^ -J “d ^ ca < < z .S H < d. o o o m o o m o in o in o o m o o G o o r- os r- m m r- o r- os os o B C in in 00 OS so in (N o rM Os rxi so (N OS 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 .. —1 m m T=( m ,—1 m 1 m 1— ( m T—i. m ,—1 m m ’—1 m t=H m .-H m i—i m m >— 1 o T T T T T T T T 7 7 7 7 7 7 7 7 sy o o O o m o o in o o o o Zh m m m m m m m m en m m m m m : K K =q 3 ^ K eK ^ K gS so" -R kR RRoa § .a O M •-, St m o •Sw a g c u in O 00 U r O ^ U ^ 9J _3 2 ^ a M U 00 o 3 N K o 00 CN in fv-) ^ ^ ‘n •n K K 00 in fn so ^ CN s K K K (n tv ■n in 00 ^ in "n S t\ Cxj 05 CQ CQ op 05 OQ I'S ^ aq J CQ 3 3 3 3 3 3 3 3 s 0 .K 1 su u u u 2 .s; SL u s K I 2 o su u s s so I •SP s s .u "S 2 0 1 « u u u g 2 2 u ^ Q II Z « g 2 c> "S s 2 Q g 2 Si "S R Sj S3 u u 2014] BAKER AND CLOUD=HUGHES: A NEW CYLINDROPUNTIA FROM CALIFORNIA 235 Table 2, Description of Characters Measured. Node is used here in reference to where branches originate from trunks or larger branches, not a stem node or areole. Description mean length of three ultimate mature stem segments mean diameter at widest portion of three ultimate mature stem segments mean length of three inter-branch spaces along the main trunk or, if unavailable, those along primary branches mean number of branches at each of three trunk nodes, if unavailable or not easily visible, those along primary branches mean length of top-most penultimate tubercle from three stem segments mean width of top-most penultimate tubercle from three stem segments mean height of top-most penultimate tubercle from three stem segments mean number of central spines per areole from tubercles as described above mean number of radial spines per areole from tubercles as described above mean length of longest central spine per areole from tubercles as described above mean length of longest radial spine per areole from tubercles as described above mean thickness of longest central spine per areole from tubercles as described above mean thickness of the sheath of the longest central spine per areole from tubercles as described above Abbreviation Character STEML stem length (mm) STEMDIA stem diameter (mm) DBTRBRCH distance between trunk branches (cm) BRANPTRND branches per trunk node TUBE tubercle length (mm) TUBW tubercle width (mm) TUBH tubercle height (mm) NOCSP central spine number NORSP radial spine number CSPL central spine length (mm) RSPL radial spine length (mm) CSPTH central spine thickness (0.01 mm) SHTH central spine sheath thickness (0.01 mm) which were morphologically intermediate be- tween C acanthocarpa and C multigeniculata. Multivariate Analysis Discriminant function analysis resulted in a 99,1% correct classification of individuals within their pre-assigned taxa (Table 3), A single indi- vidual pre-classified as Cylindropuntia chuckwah lensis was assigned by the DFA as C echinocarpa, and three were assigned as C. muitigenicuiata. All four groups were significantly different from each other at the P = 0,001 level (Table 4). Results from MANOVA indicated that mean values for seven of the 13 continuous morphological char- acters for C chuckwallensis were significantly different from those of C. echinocarpa and C muitigenicuiata (Table 5). Descriptive statistics for the multivariate dataset are presented in Table 6, Discussion Results support the inclusion of western and eastern populations of Cylindropuntia muitigeni- cuiata within one taxon. Predicted group mem- bership correctly classified all individuals within C muitigenicuiata populations, including those of the western population (population 11). Since not a single C muitigenicuiata individual was incorrectly classified as C echinocarpa, the evidence clearly supports the inclusion of the western population within C muitigenicuiata. Since Pinkava's (2000) hybrid hypothesis was based on the inclusion of these western popula- tions under C echinocarpa, there is little, if any, remaining evidence to support his hypothesis. Cylindropuntia chuckwallensis possesses a unique combination of characters when com- pared to its closest morphological relatives. It does not, however, possess any characters that are unique within the genus. This observation, combined with its polyploid nature, especially in comparison to its non-polyploid relatives, sug- gests that C chuckwallensis is an allopolyploid. Its morphology indicates that the majority of its chromosomes came from a species similar to C. muitigenicuiata. The range of tepal colors be- tween plants and the red coloration of the styles and stigmas indicate that the remaining chromo- somes may have been contributed by C. acantho- carpa or a similar extinct ancestor (Fig. 3). Because C chuckwallensis is morphologically more closely allied to C muitigenicuiata, it is likely that C muitigenicuiata contributed a larger number of chromosome sets than did C. acanthocarpa. Thus, C chuckwallensis may have originated as triploid hybrid resulting from an unreduced gamete from C. muitigenicuiata and a normal gamete from C acanthocarpa, or possibly a tetraploid individual of C muitigenicuiata hybridizing with a diploid C acanthocarpa to form hybrid triploids. The triploid may then have produced two unreduced gametes that united to form the hexaploid C chuckwallensis. If Cylindropuntia chuckwallensis has not un- dergone extensive evolutionary changes since these hypothetical events, then its morphology would be expected to align loosely with the formula (2M + lA)/3 = C, where M = C. muitigenicuiata, A = C. acanthocarpa, and C = C chuckwallensis. This assumes that the genetic effects are additive and that there are no threshold effects for the morphological characters sampled. 236 MADRONO [Vol. 61 Fig. 2. Flower longitudinal sections (left images) and face views (right images) for: A., Cylindropuntki accmthocarpa {M. A. Baker 17541. J); B., C echinocarpa (M. A. Baker 17531); C., pollen-sterile C. chiickwallensis (M. A. Baker 17534); pollen fertile D., C. chiickwallensis (M. A. Baker 17525.6); E., C. cbuckwallensis (M A. Baker 17525.7); F., C chuckwallensis {M. A. Baker 17535), bar = 1 cm. The mean values for each character from the morphological analysis were inserted into the formula, and the resulting values for only four characters of the hypothetical hybrid were well- aligned with those of C chuckwallensis: branch- es per internode, central spine number, central spine thickness, and sheath thickness. The remaining nine characters were not a close match, which, combined with the occurrence of a pink style and gynodioecy, indicates that the origin of C chuckwallensis is more complex than that of an allopolyploid of recent origin. Alternatively, the origin may be more recent and the genetic effects from hybridization not additive. However, non-additive genetic effects for morphological characters do not appear to be the normal situation in Cactaceae. An allopolyploid origin is further supported by the occurrence of tetraploidy in its close relative C whipplei (Engelm. & J. M. Bigelow) F. M. Knuth var. enodis (Peebles) Backeberg and by the occurrence of putative hybrids, i.e., morpholog- ical intermediates, between C multigeniculata and C. acanthocarpa that have been observed in the field at more than one site by the authors and by Clokey (1943). The occurrence of tetraploidy suggests a history of unreduced gametes, and the occurrence of putative hybrids suggests that hybridization between C. multigeniculata and C. acanthocarpa is not only possible but also probably frequent. An alternative to the allopolyploid hypothesis is that Cylindropuntia chuckwallensis is an auto- polyploid of an extinct or as yet undiscovered ancestor. The geographical disjunction between populations of C. chuckwallensis and those of C. multigeniculata, its differing morphology, and its gynodioecy suggest that, whether C chuckwak lensis is an allopolyploid or autopolyploid, it originated in the distant past. Similarities between the morphology of Cylim dropuntia chuckwallensis and C. echinocarpa indicate the two species may have similar origins. Putative hybrids between C acanthocarpa and C. multigeniculata are similar morphologically to C. echinocarpa, suggesting that C echinocarpa may also be of hybrid origin. Regardless of its beginning, the large geographic and morpholog- ical range of C echinocarpa suggests a distant origin and that it should not be treated as a nothospecies. The coincidence of flower colors, hexaploidy, gynodioecy, and geographic proximity among Cylindropuntia chuckwallensis, C. calmalliana, C. sanfelipensis, and C. wolfii is remarkable and warrants further investigation. It is possible that the four species share a common ancestor, most likely equal or similar to C. acanthocarpa. Morphologically, both C. calmalliana and C sanfelipensis, and to a lesser extent, C. wolfii, are much more similar to C. acanthocarpa than is C. 2014] BAKER AND CLOUD-HUGHES: A NEW CYLINDROPUNTIA FROM CALIFORNIA 237 Table 3. Classification Results of Discriminant Functions Analysis: predicted Group Membership; 99.1% OF Original Grouped Cases were Correctly Classified. Predicted group membership c. c c. c. Taxon chuckwaliensis echinocarpa acanthocarpa multigeniculata Total By number of C. chuckwaliensis 1 1 7 individuals C echinocarpa 0 C acanthocarpa 0 C. multigeniculata 0 1 119 0 0 0 3 121 0 0 119 120 0 120 0 92 92 By percentage of C chuckwaliensis 96.7 individuals C. echinocarpa 0 C acanthocarpa 0 C. multigeniculata 0 0.8 100 0 0 0 2.5 100 0 0 100 100 0 100 0 100 100 chuckwaliensis. This may indicate that they may have received a greater genetic input from C acanthocarpa in comparison to that of C. chuckwaliensis, that the hybridization/polyploidi- zation events were greatly staggered in time, or that the genetic effects simply were not additive. Since, as far as we know, gynodioecy is lacking in C. acanthocarpa, the development of gynodioecy among the four hexaploids may have been selected for by inbreeding depression rather than as a result of mutually acquired genes. Taxonomic Treatment Cylindropuntia chuckwaliensis M. A. Baker & M. A. Cloud-Hughes, sp. nov. (Figs. 2 and 3). — Type: USA, California, Riverside Co., Chuckwalla Mountains, along Corn Springs Wash, 2.5 km WNW of Corn Spring, 550 m (1830 ft) elev., 33.6327°N, 115.3527°W, 9 Apr 2012, Marc A. Baker 17534 with Michelle A. Cloud- Hughes (holotype: ASU; isotype: SD). Generally low, densely branched shrubs mostly broader (to 2+ m) than tall (to 1+ m), trunks one to several, decumbent to ascending, occa- sionally erect; the bases of older individuals with a gray matting of dead stems and spines; spacing between branch whorls generally less than 3 cm long; branches mostly in whorls of four or more; mature ultimate stems mostly less than three times long (x = 56 mm) as wide (x = 23 mm); tubercles averaging 11 mm long and approximately as wide as high (x = 8 mm); central spines 4—10 (x = 7), averaging 27 mm long; radial spines 6-12 (x = 8), averaging 13 mm long or less than half the length of the centrals; often difficult to differentiate between radial and central spines; thickest central spine averaging 0.42 mm with a sheath averaging 0.65 mm wide; flowers 3-5 cm long, 4-6 cm broad, tepals 1.5- 3 cm long, pale yellow-green to orange to purple- red; style 1.5-3. 5 cm long, generally pale to dark pink-purple, at least distally, although very pale green in approximately 10% of individuals; filament color generally matching style color; fruits obovate to suborbicular, thin-walled, with- out pulp, 15-18 mm wide, 15-30 mm long, areoles 30-55, up to 15 thin spines per areole, drying the first year; seeds about 50, white, irregularly discoid, mostly smooth or faces slightly convex or faceted, 1.5 mm thick, 3 mm in diameter. Etymology The species is named for its occurrence primarily within the Chuckwalla Mountains. Taxonomic Relationships The habit of Cylindropuntia chuckwaliensis most closely resembles that of C. multigeniculata, with individuals mostly broader than tall, older individuals with several trunks, short spaces between trunk nodes, and with several stems per trunk or branch node. Mean values for spine thickness in C. chuckwaliensis are intermediate between those of C. multigeniculata and C Table 4. Wilks’ Lambda for the Results of Discriminant Functions Analysis. Test of function(s) Wilks’ lambda Chi-square df Sig. 1 through 3 0.007 2217.382 39 0.000 2 through 3 0.118 944.368 24 0.000 3 0.420 384.309 11 0.000 238 MADRONO [Vol. 61 Table 5. Character Means by Species. Means in bold are significantly different (P < 0.01) from means of all other spedes. Species Character C. multigeniculata C. chuckwallensis C. echinocarpa C. acanthocarpa Distance between trunk branches 1.4 2.9 8.9 17.2 Branches per trunk node 5.5 4.5 2.4 1.8 Stem length 41.3 56.0 57.6 108.4 Stem diameter 16.4 22.6 24.0 25.9 Tubercle length 6.7 10.7 12.8 27.4 Tubercle width 3.4 7.8 6.3 7.0 Tubercle height 3.8 7.6 6.2 7.1 Central spine number 6.6 7.0 7.3 7.5 Radial spine number 6.5 8.0 8.0 9.9 Central spine length 20.8 26.8 25.7 28.3 Radial spine length 8.7 13.3 13.5 16.5 Central spine thickness 28.9 41.9 55.2 68.1 Central spine sheath thickness 49.1 65.2 89.4 108.4 echinocarpa but are significantly different from both species. Tubercles in C chiickwallensis are significantly wider and taller than those of the other two species (Tables 6 and 7). The styles and filaments in individuals of C chuckwallensis are generally pale to dark pink-purple, at least distally, although very pale green in approximately 10% of individuals. CyUndropuntia acanthocarpa shares this character, at least in its filaments. Both C. multigeniculata and C. echinocarpa have white to pale green styles and filaments. Distribution, Conservation, and Ecology Populations of CyUndropuntia chuckwallensis occur primarily In the area of the Chuckwalla Mountains extending to the north side of the Chocolate Mountains, and the eastern base of the Orocopia Mountains (Salt Creek), with satellite populations in the Cottonwood and Eagle Moun- tains in Joshua Tree National Park. Although the total range is approximately 50 km north to south and 60 km east to west, recorded individuals are confined within isolated areas totaling approximate- ly 3200 hectares or 32 km^. Since our results estimate an average density of 2.5 individuals per hectare, a rough estimate for the total number of individuals within known populations is 3200 hectares X 2.5 individuals per hectare or 8000 individuals. Individuals of CyUndropuntia chuckwallensis occur primarily between an elevational range of 400-1600 m in rocky, gravelly, sandy, and silty substrate on desert pavement and along washes and slopes of varying aspects and inclinations. Vegetation associated with populations of CyUndropuntia chuckwallensis includes sparse to moderately dense desert scrub. Associated vascu- lar plant species are listed by site in Table 8. The most commonly occurring vascular plant associ- ates are Ambrosia dumosa (A. Gray) W. W. Payne, Fouquicria splendens Engelm., Hilaria rigida (Thurb.) Benth. ex Scribn., and Simmondsia chincnsis (Link) C. K. Schneid., which occurred at all five data collection sites. Ambrosia salsola (Torr. & A. Gray) Strother & B. G. Baldwin, CyUndropuntia echinocarpa, C. ramosissima, Echi~ nocereus engelmannii, Eriogonum inflatum Torr. & Frem., Larrea tridentata (Sesse & Moc. ex DC.) Fig. 3. Habit and habitat for CyUndropuntia chuck- wallensis. Top image: M A. Baker 17534; bottom image: M. A. Baker 17535, bar = 1 dm. Table 6. Descriptive Statistics for all Species. Plant height (HT), plant width (W), primary branch angle (BANG), and internode length (IL) in cm; branches per internode (BPI), stem length (SL), stem diameter (SD), tubercle length (TL), tubercle width (TW), tubercle height (TH), number of central spines (CCN), number of radial spines (RSN), length of central spines (CSL), length of radial spines (RSL), thickness of central spines (CSTH), and thickness of sheaths (SHTH) in mm except for discrete characters. 2014] BAKER AND CLOUD-HUGHES: A NEW CYLINDROPUNTIA FROM CALIFORNIA 239 in in 00 q q q q q q m q oo q q q q m 7—7 ON NO q T in q in d u NO 7-4 in d A r4 U 7-4 d od NO NO NO d CN in rn rn rn d q o q in 7-7 q q L- q q q q q q i> q t" m q A On u od NO in u in d 7—4 d 7-4 U od (N rn 04 (N rn o NO O' NO 7-4 m n n q r- q r-- t-H q O' q NO 7—7 7—7 00 q o q in q 7-7 m q 00 d so G NO in NO od CN rn CN CN in (N d ■d rn rn 7—7 NO O' in 04 r-l m n (N 7—7 CN! a ^ w u u in Os a\ m a.§ u u (U o a k a 6 D J U U § in in as Os ^ § P a.g « w > a ^ w ^33 U U in m w as as m 1 1 g « (U a 6 P h-i U U § in in On ON GO (N On a •S 1 2 o 1 a a G u s a 5 U U 240 MADRONO [Vol. 61 Table 7. Comparison of Diagnostic Characters among Cylindropuntia chuckwallensis, C. echinocarpa, AND C MULTIGENICULATA. Values for quantitative characters given as the range within 95% confidence levels. Species Character C. chuckwallensis C. echinocarpa C. multigeniculata Internode length 2.6-3. 1cm 8. 4-9.4 cm 1.2-1. 5 cm Branches per internode 4.3^.7 2.3-2.6 5.3-5.8 Tubercle length 10.4-11.0 mm 12.4-13.3 mm 6. 3-7.0 mm Tubercle width 7.5-7.8 mm 6. 1-6.4 mm 3. 3-3. 5 mm Tubercle height 7.4-7.9 mm 6.0-6. 5 mm 3. 6-3. 9 mm Central spine thickness 0.40-0.43 mm 0.52-0.58 mm 0.28-0.30 mm Sheath thickness 0.63-0.67 mm 0.86-0.93 mm 0.47-0.52 mm Color of style and filaments pale to dark pink-purple, at least distally, rarely very pale green nearly white to pale green nearly white to pale green Gynodioecious yes no no Chromosome number n = 33 « = 11 « = 11 Coville, Parkinsonia florida (Benth. ex A. Gray) S. Watson, and Senegalia greggii (A. Gray) Britton & Rose occurred at four of the five sites. Another cholla, C. bigelovii, occurred at three of the sites. There are currently few^ if any conservation concerns associated with populations of Cylin- dropuntia chuckwallensis. No damage from off- road vehicles has been observed, and there seem to be no obvious herbivores other than those that eat the flowers and seeds. Phenology Flower buds of individuals of Cylindropuntia chuckwallensis emerge mid-March and anthesis occurs about four weeks later. Fruits generally mature about six to eight weeks after anthesis. Paratypes: USA, California. Riverside Co.: Corn Spring, N side of Chuckwalla Mts., 33.6233°N, 115.3255°W, 17 Dec 2011, M. A. Baker 17507 with M. A. Cloud-Hughes (ASU); Corn Spring, N side of Chuckwalla Mts., 33.6257°N, 115.3284°W, 17 Mar 2013, M. A. Cloud- Hughes 087 (RSA, SD); Corn Springs Wash, 830 m ENE of Aztec Spring, 33.6364°N, 115.3667°W, 19 Dec 2011, M A. Baker 17510 with M. A. Cloud-Hughes (ASU); NE end of Chuckwalla Mts., 1.1 km W of Corn Spring Campground, 33.6246°N, 11 5.338 UW, 25 Mar 2012, M A. Baker 17525.1 (ASU, RSA); Same location, 25 Mar 2012, M. A. Baker 17525.2 (ASU, SD); Same location, 25 Mar 2012, M. A. Baker 17525.4 (ASU, US); Same location, 25 Mar 2012, M. A. Baker 17525.5 (ASU, SD); Same location, 25 Mar 2012, M. A. Baker 17525.6 (ASU, ASC); Same location, 25 Mar 2012, M. A. Baker 17527.7 with M. A. Cloud-Hughes (ASU, SD); Chuckwalla Mts., along Corn Springs Wash, 2.5 km WNW of Corn Spring, 33.6328°N, 1 15.3521°W, 9 Apr 2012, M A. Baker 17535 with M. A. Cloud-Hughes (ASU, SD); Chuckwalla Mts., 500 m NW of Aztec Well, 33.636UN, 115.3839°W, 15 Apr 2013, M. A. Baker 17726 (ASU, RSA); Same location, 15 Apr 2013, M. A. Baker 17727 with M. A. Cloud- Hughes (US, SD); Chuckwalla Mts., vicinity of Corn Spring, , 33.62°N, 115.3rw, “dry slopes,” 9 Apr 1922, M. French Gilman s. n. (RSA 12621!); Chuckwalla Mts., vicinity of Corn Springs, 33.76°N, 115.37°W, 9 Apr 1922, P. A. Munz 5017 with D. D. Keck, box labeled “common on gravelly slopes, brown-red flowers” (UC409468!), sheet labeled “abundant about Aztec Well, on lower slopes, flowers reddish-brown” (POM 11 97 11); along Salt Creek, 17.5 km WNW of Black Butte, 33.5866°N, 115.552rW, 9 Apr 2012, M A. Baker 17533 with M. A. Cloud-Hughes (ASU, UCR); N side of Chuckwalla Mts., 1 km west of Chuckwalla Spring, 33.477rN, 115.2242°W, 1 May 2011, M. A. Baker 17305 with M. A. Cloud- Hughes, G. Rink, D. Robertson and N. Kramer (ASU, RSA); N side of Chuckwalla Mts., 1.7 km NNE of Chuckwalla Spring, 33.4917°N, 115.2098°W, 9 Apr 2012, M. A. Baker 17536 (ASU, RSA); Same location, 9 Apr 2012, M. A. Baker 17536.1 with M. A. Cloud-Hughes (ASU, US); Joshua Tree National Park, 5.5 km N of Cottonwood Spring, 33.7856°N, 1 15.804 UW, 20 Mar 2013, M. A. Baker 17720.2 (RSA, JOTR); Same location, 20 Mar 2013, M. A. Baker 17720.3 with M. A. Cloud-Hughes (ASU); Cot- tonwood Mts., Joshua Tree National Park, 450 m SW of Cottonwood Spring, 33.7336°N, 115.813UW, 20 Mar 2013, M. A. Baker 17721.1 with M. A. Cloud-Hughes (RSA, JOTR); Cottonwood Mts., Joshua Tree National Park, 450 m SW of Cottonwood Spring, 33.7338°N, 115.8134°W, 15 Apr 2013, M Harding 134 with M. A. Baker with M. A. Cloud-Hughes (RSA, JOTR); Eagle Mts., Joshua Tree National Park, 7.3 km ENE of Cottonwood Spring, 33.7442°N, 115.7499°, 11 Apr 2013, M. Harding 130 (JOTR!); Eagle Mts., Joshua Tree National Park, 5 km ENE of Cottonwood Spring, 33.751 1°N, 115.7553°W, 11 Apr 2013, M Harding 132 (JOTR!); Cottonwood Mts., Joshua Tree Na- tional Park, Cottonwood Spring, 33.74°N, 115.81°W, 25 Apr 1928, W. L. Jepson 12622b 2014] BAKER AND CLOUD-HUGHES: A NEW CYLINDROPUNTIA FROM CALIFORNIA 241 Table 8. List of Associated Vascular Plant Species for Cylindropuntia chuckwallensis, by Site. Site Chuckwalla Indian Com spring Salt Species spring well wash creek Acamptopappm sphaerocephalm (Harv. & A. Gray) A. Gray Acmispon rigidus (Benth.) Brouillet Adenophyllum porophylloides (A. Gray) Strother Ambrosia dumosa (A. Gray) W. W. Payne x Ambrosia salsola (Torr. & A. Gray) Strother & B. G. Baldwin Bahiopsis parishii (Greene) E. E. Schill. & Panero Bebbia juncea (Benth.) Greene x Coryphantha alversonii (J, M. Coult.) Orcutt Coleogyne ramosissima Torr. Cylindropuntia bigelovii (Engelm.) F. M. Knuth x Cylindropuntia echinocarpa (Engelm. & J. M. Bigelow) F. M. Knuth X Cylindropuntia munzii (C. B. Wolf) Backeb. (pro hybr.) Cylindropuntia ramosissima (Engelm.) F. M. Knuth x Ditaxis lanceolata (Benth.) Pax & K. Hoffm. Ditaxis neomexicana (Miill. Arg.) A. Heller Echinocactus polycephalus Engelm. & J. M. Bigelow x Echinocereus engelmannii (Parry ex Engelm.) Lem. x Encelia farinosa A. Gray ex Torr. x Ephedra calif ornica S. Watson Ephedra nevadensis S. Watson x Eriogonum inflatum Torr. & Frem. x Eriogonum fasciculatum Benth. Euphorbia poly car pa Benth. x Euphorbia setiloba Engelm. ex Torr. Eagonia laevis Standi. x Eerocactus cylindraceus (Engelm.) Orcutt x Eouquieria splendens Engelm. x Eunastrum hirtelium (A. Gray) Schltr. Hibiscus denudatus Benth. x Hilaria rigida (Thurb.) Benth. ex Scribn. x Hyp t is emoryi Torr. x Kramer ia bicolor S. Watson Krameria erecta Willd. ex Schult. x Larrea tridentata (Sesse & Moc. ex DC.) Coville x Lycium andersonii A. Gray x Mammillaria tetrancistra Engelm. x Marina parryi (Torr. & A. Gray) Barneby Mirabilis laevis (Benth.) Curran Olneya tesota A. Gray x Opuntia basilaris Engelm. & J. M. Bigelow Parkinsonia florida (Benth. ex A. Gray) S. Watson x Peucephyllum schottii A. Gray x Phoradendron calif ornicum Nutt. Porophyllum gracile Benth. Psorothamnus schottii (Torr.) Barneby x Psorothamnus spinosus (A. Gray) Barneby Scutellaria mexicana (Torr.) A. J. Paton Senegalia greggii (A. Gray) Britton & Rose x Senna armata (S. Watson) H. S. Irwin & Barneby Simmbndsia chinensis (Link) C. K. Schneid. x Sphaeralcea ambigua A. Gray Stephanomeria pauciflora (Torr.) A. Nelson Tetracoccus hallii Brandegee Trixis californica Kellogg Yucca schidigera Ortgies Ziziphus obtusifolia (Hook, ex Torr. & A. Gray) A. Gray Cottonwood canyon X x X X X X X X X X X X X X X X X X X X X X X X X X X X X X 242 MADRONO [VoL 61 (JEPS66875 digital image!); Cottonwood Mts., Joshua Tree National Park, near Cottonwood, 33.74°N, 115.8rW, 8 May 1971, A K Shevock 1035 (RSA303531!); Cottonwood Mts., Joshua Tree National Park, hills N and NE of Cotton- wood Spring, “Flowers a peculiar dull reddish purple, sometimes with a yellowish shade,” 33N4°N, 115„8rW, 24 Apr 1928, K W. Peirson 7901 (RSA65624!); Cottonwood Mts., Joshua Tree National Park, Cottonwood Road, “quartz outcrop,” 915 m (3000 ft) elev., 21 May 1941, /. E. Cole 763 (JOTR1119!) (The specimen includes a single flower with dark purple-red stamen filaments and tepals. According to Mitzi Harding (pers. comm. 2013), this locality corresponds to hillside overlain with quartz gravel and rocks 1 km west of Smoke Tree Wash, 5.5 km N of Cottonwood Spring, 945 m [3100ft]); Imperial Co.: S side of Chuc Walla Mts., 3.8 km SW of Graham Pass, 33.4271°N, 115.1492°W, 10 Apr 2012, M A. Baker 17537 with M A. Cloud- Hughes (ASU, UNLV); S side of Chuckwalla Mts., along Graham Pass Road, 1.7 km E of its intersection with Bradshaw Trail, 33.4263°N, 115.1532°W, 20 Mar 2013, M A Baker 17719 J (ASU, RSA); Same location, 20 Mar 2013, M A. Baker 17719.2 (ASU, SRSC); Same location, 20 Mar 2013, M A. Baker 177193 (ASU, SD); Same location, 20 Mar 2013, M A. Baker 17719.4 with M A. Cloud-Hughes (ASU, UCR). Key to the Spiny-Fruited Cylindropuntia of the United States Varieties of Cylindropuntia acanthocarpa and C. californica are not treated here. Morphologi- cal intermediates between C acanthocarpa and C echinocarpa, C. ganderi and C californica., and C echinocarpa and C californica are common in areas of sympatry. 1. Ultimate mature stem segment less than 1 cm in diameter, tubercles less than 1-2 mm high. .................................................... Cylindropuntia ramosissima r. Ultimate mature stem segment greater than 1.5 cm in diameter, tubercles greater than 1-2 mm high 2. Ultimate mature stem segments generally less than 10 cm long, tubercles less than three times long as wide 3. Generally erect shrubs (except when infested by rodents), generally with a single ascending to erect main trunk, distance between trunk nodes generally at least 10 cm long, generally with one or two branches per trunk node .................................... Cylindropuntia echinocarpa 37 Generally low shrubs, generally with several decumbent to ascending main trunks, distance between trunk nodes generally less than 5 cm long, generally with three or more branches per trunk node 4. Style and stamen filaments generally with at least some purple-pink coloration; inner tepals pale orange-yellow to purple-red, fruits always very spiny and drying within a few weeks, Riverside and Imperial counties, California. .......................... Cylindropuntia chuckwallensis 47 Style and stamen filaments with no purple-pink coloration, inner tepals always pale green- yellow; fruits often not heavily spined, at least for westernmost populations, not drying for several months or more, perhaps drying more quickly in easternmost populations, Clark County, Nevada and Mohave County, Arizona. ........... Cylindropuntia multigeniculata 27 Ultimate mature stem segments generally more than 10 cm long, tubercles more than three times long as wide 5. Erect open shrubs, generally with a single ascending to erect main trunk that branches well above the base ............................................... Cylindropuntia acanthocarpa 57 Mostly matted shrubs, older individuals generally with several main trunks that branch near or below the base of the plant 6. Stamen filaments purple-red, inner tepals orange-yellow to purple-red, stems very stout, generally greater than 4 cm diameter, areoles of older stems with proliferating spines — narrowly endemic to extreme southeastern San Diego and extreme southwestern Imperial counties, Califor- nia ......................................................... Cylindropuntia wolfii 67 Stamen filaments pale yellow-green, inner tepals yellow to green-yellow, stems thin, generally less than 2,5 cm diameter, areoles of older stems occasionally with proliferating spines — widespread from coastal mountains and nearby deserts of southwestern California. 7. Spines not overlapping or only slightly overlapping those of adjacent areoles, fruits weakly spined, coast and coastal mountains, southern California .............. ...................................... Cylindropuntia californica 77 Spines overlapping those of adjacent areoles for half their length or more, fruits strongly spined, deserts of eastern San Diego, western Imperial, and south edge of Riverside counties, California .............................................. Cylindropuntia ganderi Acknowledgments Mitzi Harding, Biological Science Technician, Joshua Tree National Park, assisted with Park populations, conducted surveys for Cylindropuntia chuckwallensis within the Eagle Mountains area, and provided digital photos from the JOTR herbarium. Tasha La Doux, Administrative Director, Sweeney Granite Mountains Desert Research Center, also helped with JOTR specimens and discussed Southern California Cylindro- puntia. Michael Vamstad, Wildlife Ecologist, Joshua Tree National Park, made it possible to obtain a permit 2014] BAKER AND CLOUD-HUGHES: A NEW CYLINDROPUNTIA FROM CALIFORNIA 243 to collect specimens (permit number: JOTR-2012-SCI- 0024). Jason Brookes, Ashland, Oregon, was the first to bring to our attention to the Cottonwood Springs population within the Park. Andrew Doran, Adminis- trative Curator, University and Jepson Herbaria, University of California, Berkeley, provided digital images of herbarium specimens from JEPS. Literature Cited Baker, M. A. 2002. Phenetic analysis of Cylindro- puntia multigeniculata (Clokey) Backb. (Cactaceae) and its relatives. Report prepared for the United States Fish and Wildlife Service, 17 March 2002, 35 pp. . 2006. A new florally dimorphic hexaploid, Echinocereus yavapaiensis sp. nov. (section Triglo- chidiatus, Cactaceae) from central Arizona. Plant Systematics and Evolution 258:63-83. AND C. Butter WORTH. 2013. Geographic distribution and taxonomic circumscription of populations within Coryphantha section Robustis- pina (Cactaceae). American Journal of Botany 100:984-997. AND D. J. PiNKAVA. 1987. A cytological and morphometric analysis of a triploid apomict, Opuntia X kelvinensis, (subgenus Cylindropuntia, Cactaceae). Brittonia 39:387^01. , , J. P. Rebman, B. D. Parfitt, and A. D. Zimmerman. 2009. Chromosome numbers in some cacti of western North America. VIII. Haseltonia 15:117-134. Beard, E. C. 1937. Some chromosome complements in the Cactaceae and a study of meiosis in Echinocereus papillosus. Botanical Gazette 99:1- 20. Clokey, I. W. 1943. Notes on the flora of the Charleston Mountains, Clark County, Nevada. V. Cactaceae. Madrono 7:69. Davis, G. L. 1966. Systematic embryology of the Angiosperms. Wiley & Sons, Inc., New York, NY. Johnson, M. A. T. 1980. Further cytological investi- gations in Mammillaria proUfera and other Mam- millaria species. Cactus and Succulent Journal of Great Britain 42:43-A7. Marks, G. E. 1966. The origin and significance of intraspecific polyploidy: experimental evidence from Solanum chacoense. Evolution 20:552- 557. Mihalte, L. and R. E. Sestras. 2012. The plant size and the spine characteristics of the first generation progeny obtained through the cross-pollination of different genotypes of Cactaceae. Euphytica 184:369-376. Miller, J. S. and D. L. Venable. 2000. Polyploidy and gender dimorphism in plants. Science 289:2335-2338. Negron-Ortiz, V. 2007. Chromosome numbers, nuclear DNA content, and polyploidy in Consolea (Cactaceae), an endemic cactus of the Caribbean Islands. American Journal of Botany 94:1360- 1370. Parfitt, B. D. 1985. Dioecy in North American Cactaceae: a review. Sida 11:200-206. Pinkava, D. j. 1999. Cactaceae Part Three. Cylindro- puntia. Journal of the Arizona-Nevada Academy of Science 32:32^7. . 2003. Cylindr opuntia. Pp. 102-118 in Flora of North America Editorial Committee (eds.), Flora of North America North of Mexico. Vol. 4: Magnoliophyta: Caryophyllidae, Part 1. Oxford University Press, New York, NY. , M. A. Baker, B. D. Parfitt, M. W. Mohlen- BROCK, and R. D. Worthington. 1985. Chro- mosome numbers in some cacti of western North America: V. Systematic Botany 10:471^83. and M. G. McLeod. 1971. Chromosome numbers in some cacti of western North America. Brittonia 23:171-176. , B. D. Parfitt, M. A. Baker, and R. D. Worthington. 1992. Chromosome numbers in some cacti of western North America: VI, with nomenclatural changes. Madrono 37:98-113. , J. P. Rebman, and M. A. Baker. 1998. Chromosome numbers in some cacti of western North America: VIL Haseltonia 6:32-41. Powell, A. M. 2002. Experimental hybridization between Echinomastus intertextus and E. warnockii (Cactaceae). Haseltonia 9:80-85. AND J. F. Weedin. 2004. Cacti of the Trans- Pecos & adjacent Areas. Texas Tech University Press, Lubbock, TX. , A. D. Zimmerman, and R. A. Hilsenbeck. 1991. Experimental documentation of natural hybridization in Cactaceae: origin of Lloyd’s Hedgehog Cactus, Echinocereus X Iloydii. Plant Systematics and Evolution 178:107-122. Rebman, J. P. 1998. A new cholla (Cactaceae) from Baja California, Mexico. Haseltonia 6:17-21. and D. j. Pinkava. 2001. Opuntia cacti of North America: an oveiwiew. Florida Entomologist 84:474-483. Remski, M. F. 1954. Cytological investigations in Mammillaria and some associated genera. Botani- cal Gazette 116:163-171. Trushell, N. 1985. A systematic revision of the Opuntia whipplei complex (Cactaceae). M.S. Thesis, Arizona State University, Tempe, AZ. ViTE, F., E. PoRTiLLA, J. A. Zavala-Hurtado, P. L. Valverde, and a. DIaz-Solis. 1996. A natural hybrid population between Neobuxbaumia tetetzo and Cephalocereus columna-trajani (Cactaceae). Journal of Arid Environments 32:395^05, Madrono, Vol. 61, No. 2, p. 244, 2014 REVIEW Annotated Checklist of the Vascular Plants of Santa Cruz County, California. Second edition. By DY- LAN Neubauer. 2013. California Native Plant Society, Santa Cruz Coun- ty Chapter, Santa Cruz, CA. 166 pp. Price $15.00 (spiral bound). Santa Cruz County is the second smallest county in California, yet nearly twenty percent of the California flora can be found natively there. This makes botanical references of the area relevant to most botanists in California. Dylan Neubauer has just produced an exceptional resource with her new edition of the Annotated Checklist of the Vascular Plants of Santa Cruz County, California. Second Edition. For a check- list to be useful, it needs to simply list all of the plants found in an area, but Neubauer’s checklist goes far beyond that. The checklist begins with colored maps of Santa Cruz County with public lands and special botanical areas marked. Santa Cruz County has been divided into 17 floristic regions and each is marked on a map, and referenced for each taxa. The checklist includes 1594 taxa (native and non-native) found within the county, and Neubauer has made every attempt to reference a voucher specimen for each one. All taxa are recognized under the revised taxonomy published recently in the new Jepson Manual (TJM2) and the corresponding eFlora (Baldwin et al. 2012). However, you will not be lost with this checklist if you only know the older names, as the index lists all synonyms and serves as a way to relearn taxa that have undergone recent name revisions. The meat of this checklist, and what sets it apart from a simple list of plants, lies in the book’s eight appendices. For land managers and those working with rare plants Appendix 1 provides a list of all rare taxa and their rarity ranking according to Federal, State, and CNPS listings. Appendix 2 lists the 16 endemic taxa to Santa Cruz County. Appendix 3 lists those taxa that appear to be unfortunately extirpated from the county. Despite the small size of Santa Cruz County, botanical discovery still abounds, and Appendix 4 lists those taxa that are not currently recognized but deserve taxonomic distinction within the county. Some of these taxa will be published as new species, or serve as the basis of studies of local adaptation within Santa Cruz. The extensive research conducted by Neubauer to produce this checklist is evident throughout, but especially in Appendix 7. Here, Neubauer, lists taxa that do not occur in the county despite past misconceptions. Most of these taxa are the result of misidentiflcations, which Neubauer confirmed herself by traveling to many herbaria in California, or are waifs, or otherwise unverified in the county. This appendix alone will prevent reoccurring identification errors, and could provide a starting point for future discovery as waifs become more persistent. The checklist ends with Appendix 8, the ‘Notes’ section, which is the most extensive and valuable portion of this checklist. The Notes section is what makes this checklist more like an actual flora. These notes, provided by Neubauer and a host of local experts, offer a wealth of information about most taxa in the county. These notes provide a discussion of taxonomic issues, or key characters to look at during identification. If two species are commonly confused, the Notes section will call out their differences. This section also covers natural history details about taxa, how they vary throughout the county, where specific forms are found, and speculation on how some have arrived to the county. For anyone interested in a particular taxon. Appendix 8 is a great place to start your investigations. The checklist is spiral bound and perfect for use in and out of the field. The checklist is an excellent resource for botanical investigations, a weekend hike, or nursery inventories. I highly recommend it. Contact Cindy Hudson (cindy@ centralcoastwilds.com) from the Santa Cruz CNPS Chapter to obtain a copy. — Jenn Yost, Biological Sciences Department, Califor- nia Polytechnic State University, San Luis Obispo, CA 93407. jyost@calpoly.edu. Literature Cited Baldwin B. G., D. H. Goldman, D. J. Keil, R. Patterson, T. J. Rosatti, and D. H. Wilken (eds.). 2012. The Jepson Manual: vascular plants of California, 2nd ed. University of California Press, Berkeley, CA. Madrono, Vol. 61, No. 2, p. 245, 2014 REVIEW Wiidflowers of Orange Coun- ty and the Santa Ana Mountains. By Robert L. Allen and Fred M. Roberts Jr. 2013. La- guna Wilderness Press, Laguna, CA. 497 pp. ISBN 978-0-9840007- L5. Price $35.00, paperback This is a remarkable publication representing a love of nature and a dedication to sharing it. The authors have assembled an information-rich guide that is profusely illustrated with diagrams, maps, and attractive and informative photo- graphs. It is far more than a wildflower guide. The authors “know” Orange County and the Santa Ana Mountains region and they lovingly describe its special places, its plant life, and some of its wildlife. The introductory chapter includes engaging discussions of the geology and geography of the region, plant features, plant classification, plant names, and plant communities. A section on Watching Wiidflowers includes practical tips for observing, documenting, photographing, drawing, and gardening with wiidflowers plus safety tips for the outdoors. One thing I would like to have seen in the introduction is a discussion of the authors’ criteria for deciding what to include or exclude; cattails, for instance are included as wiidflowers whereas rushes, with more readily recognizable floral structures, are not. I wondered about some other inclusions or exclusions as well. The book is up-to-date, following modern classifications of major plant groups and is aligned for the most part with the second edition of the Jepson Manual (Baldwin et al. 2012). Although the arrangement of taxa within major groups is mostly alphabetical the authors have, in some cases, grouped morphologically similar genera to make comparisons easier. Plant de- scriptions are provided at various taxonomic levels along with habitat and locality information and natural history notes. The book does not include taxonomic keys, so identifications may be a challenge to users without an acquaintance with plant families. Scientific names used in the book are, for the most part, in agreement with those in the Jepson Manual. No standardization exists, however, for common names, and thus the choice of which common names to present to the public is a challenge to authors of a book of this nature. I noted various instances where the authors and I differ in our preferred common names. In some instances the authors include more than one common name for a species. Because of the negative connotation that may accompany the word weed, the authors have avoided many common names incorporating this word — thus tarweeds become tarplants, pygmyweeds become pygmy stonecrops, etc. Common chickweed — which is indeed a weed — is unchanged. The authors provide etymologies for specific epithets and for some generic names. An unexpected, unadvertised, and most infor- mative feature of this book is the description of guilds of animals, especially insects, that are associated with particular species or groups of plants. These are accompanied by excellent photos — often stunning close-ups of insects and informative discussions. Scientific and common names are provided for the animals along with descriptions of their interactions with the plants and other natural history notes. Following the taxonomic treatments of plants and the occasional animal guild is a discussion of Where to go Wildflower Watching. This includes maps and directions plus highlights of what may be observed. In such an urbanized area it is very useful to discover that there are so many special places to see wiidflowers and other natural features. End matters include a detailed list of referenc- es used in preparation of the guide, a glossary, an index to common and scientific names of plants and animals, and an index to plant common names organized by flower color. I look forward to using this book, and I highly recommend it. The authors are to be congratu- lated on an outstanding achievement. — David J. Keil, Biological Sciences Department, California Polytechnic State University, San Luis Obispo, CA 93407, dkeil@calpoly.edu. Literature Cited Baldwin B. G., D. H. Goldman, D. J. Keil, R. Patterson, T. J. Rosatti, and D. H. Wilken (eds.). 2012. The Jepson Manual: vascular plants of California, 2nd ed. University of California Press, Berkeley, CA. Madrono, VoL 61, No. 2, pp. 246-247, 2014 REVIEW Baja California Plant Field Guide, 3rd Edition. By JON P. Rebman, and Nor- man C. Roberts. 2012. San Diego Natural History Museum and Sunbelt Pub- lications, San Diego, CA. 452 pp. ISBN 9780916251 185. Price $34.95 (paperback) Even the casual traveler to Baja California, Mexico cannot ignore its flora. From the window of a car passing down the main highway the eyes are greeted with boojums and magnificent cacti, led by towering cardons, bearded old man’s cacti, and candelabra cacti. There are also agaves, elephant trees, and stately tree yuccas all dressing a dramatic landscape comprised of granite boulders, volcanic ridges, and sediment covered plains, often with the ocean glistening in the background. For someone trained in botany it is a magical land. Half of the wonder lies in the varied and interesting flowers that can only be found on foot and away from the vehicle. Few people know the flora of the Baja California Peninsula like the authors of the new edition of a Baja California Plant Field Guide. Norman Roberts, who sadly did not live to see the edition reach print, wandered the peninsula for nearly six decades. Jon Rebman, the curator of the San Diego Natural History Museum Herbarium, is an intense student of the peninsu- lar flora and seems to spend as much time on the south side of the border as the north. His energy and enthusiasm has given us a wealth of knowledge about Baja California plants since he came to the San Diego Natural History Museum in the mid 1990’s. The Baja California Plant Field Guide in one edition or another has been available since 1975 (Coyle and Roberts 1975; Roberts 1989). These books have been constant companions of mine on my travels along the peninsula. Less complex and technical than Ira Wiggins’ Flora of Baja California (1980), this field guide has rendered the flora of the peninsula accessible to a wide array of travelers from the lay to the professional. This latest edition of the Baja California Plant Field Guide is much expanded, at a hefty 452 pages. It is the best of the three editions. In addition to commonly encountered trees and shrubs, this edition also includes many annual and perennial herbs, something that was lacking in previous editions. The organization follows previous editions and is a relatively standard format for plant identification guides with an introduction followed by entries describing groups and individual species of plants. The introduction has been entirely rewritten and greatly expanded. I highly recommend reading it from start to finish as it provides a wonderful overview of the natural history of Baja California as it pertains to its flora and individual plant species. Various experts contributed to writing the new introduction, including Exequiel Ezcurra, Thomas A. Demere, Pedro P. Garcillan, and Charlotte Gonzales- Abraham. The introduc- tion is accompanied by photographs, satellite images, and maps which add considerably to the discussion. The first ten pages are devoted to the climate of Baja California, followed by 1 1 pages explaining the geology of the peninsula, and 13 pages discussing the phytogeography (vegetation) of the peninsula. Thirteen ecoregions are dis- cussed including the California Mountains region (Sierra Juarez and Sierra San Pedro Martir), Pacific Islands, Central Desert, Central Gulf Coast, La Giganta Ranges, Viscaino Desert, Magdalena Plains, and two Cape ecoregions. Each of the ecoregion accounts includes a representative photograph, and a description that includes the general location, climate, and list of representative species. The Spanish name coun- terparts are also provided for each ecoregion. Jon Rebman penned a section discussing plant endemism on the peninsula, which could be as high as 30 percent. Rebman highlights the cacti as an example, which is not only one of the most significant elements of the peninsula but includes 93 endemic taxa (a 72 percent rate of endemism!). The introduction is rounded out with a discussion on non-native plants and conservation. The rest of the book is devoted to plant field guide and species accounts. The entries are organized starting with primitive forms (a brief two page entry for nonvascular plants, lichens, and bryophytes) and flows toward more ad- vanced forms. The family arrangement is the same as in the second edition of the Jepson Manual (Baldwin et al. 2012) with a few exceptions. Amaranthaceae and Chenopodiaceae are united while Lotus L. has not been split. These differences are more likely due to timing than to author intent. Over 700 different plant taxa in 111 plant families are treated in this section. The photo- graphs alone are well worth the price of the book. 2014] BOOK REVIEW 247 Thumbing though the high quality photographs shows us many unfamiliar and interesting plants that are unlike anything we see in California. Most are beautiful shots and often highlight characters that are useful for identification. This makes for a considerably larger book then previous editions, which might be a little less convenient in the field but well worth the extra weight. Most groups are represented by the more common species or those taxa travelers are most likely to encounter. Each entry is titled by Latin name, synonym if relevant, English name, and Spanish name. A brief description of the plant is provided along with its distribution on the peninsula. In many cases, especially if only one or two taxa in a genus have accounts, a summary of the total number of species in the genus occurring in Baja California is provided, I always appreciate these summaries. If a plant in the field does not appear to match the photos and descriptions, it is nice to know there are more taxa to consider. Other interesting facts are often included, such as plant use. The ferns and their relatives make up a relatively short section, primarily represented by members of the Brake family (Pteridaceae). However, ferns as a group are less likely to be encountered by travelers. Most of the gymno- sperm taxa (18 of 23), which are primarily pines and cypresses of the northern mountains, are provided with entries. The first flowering plant group is the Magno- liidae-Piperales clade. This group only includes only a few plants, such as pipevine (Aristolochia- ceae) and lizard-tail (Saururaceae). Next follow the Monocots and then the Eudicots. The Agave family (Agavaceae) garners the most attention of the Monocots, with 13 accounts. All four members of Nolina Michx. are covered (though I would like to have seen a photograph of Nolina palmer i S. Wats, with flowers or fruit). The grasses are weakly covered, with only six accounts, but there are other resources for Baja California grasses such as Gould and Moran’s The Grasses of Baja California (1981), although out of print. The Eudicot accounts are the heart of the book with over 300 pages devoted to over 600 taxa. My immediate reaction was to check out the oaks (Fagaceae), which seem well represented. I was interested to see species I am not familiar with such as Quercus albocincta Trel. and Q. brandegeei Goldman from Baja California Sur. There are plenty of sunflowers (28 pages, Asteraceae) and legumes (34 pages, Fabaceae). The authors have paid special attention to the cacti (Cactaceae), devoting over 40 pages to this significant and diverse group. Some of the real gems are groups or genera that Californians are less familiar with. Baja California has a much greater diversity of spurges (Euphorbiaceae) then one would find in California and the book reflects this. Only a single species of elephant tree (Burseraceae) occurs in California. This guide provides accounts for six species. Overall, the editing was well done, and the photographs are of high quality. While I have not read every entry, I did not see many significant errors or omissions. There are a few. Koeberlinia spinosa Zucc., for example, is stated as being found in the United States in Arizona and Texas without mention of California where it is known from the lower desert. I would like to have seen some groups covered a little more thoroughly (e.g., grasses, dudleyas), and as an artist, I do miss some of the line drawings of the older editions. One plate in the last edition illustrated the leaves of most of the elephant tree species {Bursera Jacq. ex L. and Pachycormus Coville ex Standi.), which I found useful. The illustration is absent in this edition. I certainly appreciate the challenge this book presented the authors who clearly had to make some sacrifices to keep the book manageable in terms of size. I highly recommend this book for anyone with even a casual interest in Baja California. If you are planning a trip and interested in plants, it is a must have tool. Even if you can’t travel to Baja California, with this book in hand you can almost believe that you have been magically transported there. — Fred Roberts, 722 Point Arguello, Oceanside, CA 92058. antshrike@cox.net. Literature Cited Baldwin B. G., D. H. Goldman, D. J. Keil, R. Patterson, T. J. Rosatti, and D. H. Wilken (eds.). 2012. The Jepson Manual: vascular plants of California, 2nd ed. University of California Press, Berkeley, CA. Coyle J. and N. C. Roberts. 1975. A field guide to the common and interesting plants of Baja California. Natural History Publications Compa- ny, La Jolla, CA. Gould F. W. and R. Moran. 1981. The grasses of Baja California, Mexico. San Diego Society of Natural History Memoir 12, San Diego, CA. Roberts N. C. 1989. Baja California plant field guide. Natural History Publications Company, La Jolla, CA. Wiggins I. L. 1980. Flora of Baja California. Stanford University Press, Stanford, CA. Madrono, Vol. 61, No. 2, pp. 248-249, 2014 REVIEW Flora of the Four Corners Region. Vascular Plants of the San Juan River Dra- inage: Arizona, Colorado, New Mexico, and Utah. By Kenneth D. Heil, Steve L. O'Kane, Jr., Linda Mary Reeves, AND Arnold Clifford. 2013. Monographs in sys- tematic botany from the Missouri Botanical Gar- den, Vol. 124, Missouri Botanical Garden Press, St. Louis, MO. xvi + 1098 pp. ISBN 9784-930723-84-9 (clothbound). Price $72.00 The Four Corners Region is the relatively littie-known area of the southwestern United States where four states meet at a common point. Characterized by stark topographic features and scenic vistas, this portion of the Colorado Plateau is drained by the San Juan River, a major tributary of the Colorado River. Elevations range from 1 130 m at the mouth of the San Juan River to 4292 m at the summit of Mt. Eoius in the San Juan Mountains of Colorado, The Four Corners Region is separated by considerable distances from ma.jor botanical centers and is divided among four states whose published floras were produced by botanists working at different institutions at different times. There has long been a need for a unified modern flora of the region that spans the state boundaries. And this book fills that void. Flora of the Four Corners Region is a multi-authored work that treats 120 families, 697 genera, 2117 species, and 186 (additional) infraspecific taxa. Forty-one of the terminal taxa are endemic or nearly so. Much care went into the development and production of this flora, and the authors and editors are to be congratulated for their efforts and the outstanding product that has resulted. Multiple features make this a very attractive volume. A double-page map of the Four Comers area inside the front covers presents topography, state and county boundaries, and major localities. A second map inside the back covers presents the San Juan River and its tributaries. Several beautiful botanical watercolors grace the introduc- tory pages. Introductory material includes brief sections on the scope and history of the project that produced the flora, discussions of the conventions used in the book, a family-by-family summary of the taxonomic representation of the flora, and overviews of the regional geology and climate. Expanded descriptions of plant communities and vegetation are followed by discussions of plant migration routes and non-native plants. Four watercolor plates illustrating fifteen vegetation associations complete the introduction. The taxonomic section of the book begins with a dichotomous key to families derived from the family key in Colorado Flora: Western Slope (Weber and Wittmann 2001). Taxonomic treat- ments are grouped as Ferns and Fern Allies (paraphyietic, including both lycophyte and fern families), Gymnosperm.s, and Angiosperms. Within these major groups, taxa are arranged alphabetically from family down. The full alpha- betic merger of angiosperm families may be a bit jarring for those expecting the traditional sepa- ration of dicots and monocots or the more recent separation of angiosperms into monophyletic lineages as in the second edition of California's Jepson Manual (Baldwin et al. 2012). Keys are bracket-formatted . Each generic and lower level taxonomic treatment includes scientific name, derivation of the generic name or epithet, and one or more common names. Full descriptions are provided for each taxon, and use of typeface conventions makes the descriptions easy to follow; character headings such as leaves, flowers, fruits, etc. are bold-faced, all caps, and each is spelled out, with the exception of inflorescence, which for some reason, is abbreviated INFL, For each terminal taxon a brief statement of habitat is followed by the geographic distribution in the Four Corners region. Elevations are given both in meters and feet, and both flowering and fruiting times are indicated. A statement of extralimital distribution is often followed by anecdotal comments. Excel- lent line drawings appear here and there for some taxa, but are very unevenly scattered among families. Inclusion of synonyms from older floras and other literature is eclectic. Taxonomic treatments are uneven in their adherence to the recent changes resulting from molecular phylogenetic studies, such as the Angio- sperm Phylogeny Group III classification (APG III 2009). Some family treatments are in accord with the updated classifications. Adoxaceae, for instance, is accepted in its modem circumscription, including Sambucus. Araceae includes Lemna- ceae. The traditional Scrophulariaceae is broken apart with most taxa dispersed to Orobanchaceae, Phrymaceae, and Plantaginaceae. Treatments of a number of angiosperm families, however, follow traditional circumscrip- tions. Acer L. is retained in Aceraceae, which in APG III (2009) is merged into the Sapindaceae. 2014] BOOK REVIEW 249 Liliaceae, as treated in the Four Comers flora, is a highly polyphyletic assemblage with members assignable in APG III (2009) to multiple families in two orders; alternate family assignments are briefly discussed and were added by the editors to the key to genera of Liliaceae (si.). Hydrophyl- laceae and Boraginaceae are maintained as separate families with no mention that they are sometimes merged. Asclepiadaceae is kept sepa- rate from the otherwise paraphyletic Apocyna- ceae; a comment under Apocynaceae acknowl- edges that there is strong evidence for combining the two. This should have been an easy prepub- lication merger. Portulacaceae is retained in the traditional circumscription with acknowledge- ment that most genera probably should be treated as a separate family (Montiaceae). In a similar fashion traditional Primulaceae is retained though alternate family placements for some genera are acknowledged. Celtis L. is in its traditional spot in Ulmaceae rather than in Cannabaceae; Sarcobatus Nees is retained in Chenopodiaceae rather in its own family; Hyper- icum L. is in Clusiaceae rather than Hypericaceae; Peganum L. remains in Zygophyllaceae instead of Nitrariaceae; Proboscidea Schmidel is placed in Pedaliaceae rather than Martyniaceae. A similar tension exists in some families between updated and traditional generic circumscriptions. The taxonomic treatment is followed by 20 pages of color photos of plant species (six per page, arranged alphabetically by binomial), a taxonomically arranged list of the line drawings that are dispersed through the text along with credits to the artists who created them, a 32 page un-illustrated glossary, 24 pages of literature cited, one and one-half pages of general referenc- es, and an index to the accepted scientific names and common names. Synonyms are not indexed. A few miscellaneous comments and quibbles. This is not a field manual for the backpack; it measures 8.75 X 11.25 X 2.25 inches and weighs 6.35 pounds. A full-page watercolor on p. xiv is labeled as Centaurea maculosa Lam.”, but the name used in the taxonomic treatment is '"Cen- taurea stoebe subsp. micranthos’^ (S.G. Gmel. ex Gugler) Hayek. Etymologies of epithets and generic names for the most part appear to be accurate. The derivation of the specific epithet for Pinus contorta Douglas ex Loudon is given as ‘‘twisted together, referring to the needles” (p. 90). However the common name given by John Loudon, the botanist who described P. contorta, is “the twisted-branched pine” (Loudon 1838, p. 2292), which is descriptive of the wind-pruned form of the tree found by David Douglas along the coast of Washington and Oregon. The glossary definitions of spine, thorn, and prickle are accurate, but application of these terms in taxonomic treatments sometimes is not; a bit more editorial oversight would have helped here. The “seed” in the drawing of Coleogyne ramo- sissima Torr. is actually the fruit. But the errors are few and pale before the remarkable achieve- ment this volume represents. — David J. Keil, Biological Sciences Department, California Polytechnic State University, San Luis Obispo, CA 93407. dkeil@calpoiy.edu. Literature Cited Angiosperm Phytogeny Group (APG III). 2009. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG III. Botanical Journal of the Linnean Society 161:105-121. Baldwin B. G., D. H. Goldman, D. J. Keil, R. Patterson, T. J. Rosatti, and D. H. Wilken (eds.). 2012. The Jepson manual: vascular plants of California, 2nd ed. University of California Press, Berkeley, CA. Loudon J. C. 1838. Arboretum et fmticetum Britannicum, Vol. 4: from Garryaceae, p. 2031 to the end. Longman, Orme, Brown, Green, and Longmans, London. Weber W. A. and R. C. Wittmann. 2001. Colorado Flora: Western Slope, 3rd ed. University Press of Colorado, Boulder, CO. Madrono, VoL 61, No. 2, p. 250, 2014 NOTEWORTHY COLLECTION CALIFORNIA Astragalus rattanii var. jepsonianus A. Gray (Barneby) (FABACEAE). — San Benito Co., Bureau of Land Management Clear Creek Management Area, approximately 3.5 airline km (2.2 mi) ESE of Tucker Mountain peak and 3.7 airline km (2.3 mi) WSW of Sampson Peak, large, isolated grassland area 0.5 airline km (0.4 mi) west of Larious Creek, 1035 m (3395 ft, peak flowering, early fruiting, 17 April 2013, R. O’Dell s.n. (DAV200437, DAV200438); Same locality, late flowering, peak fruiting, 2 May 2013, R. O’Dell s.n. (JEPSl 11497, JEPSl 11498); Same locality, 17 April 2013 and 2 May 2013, R. O’Dell s.n. Species identifi- cation was confirmed by Aaron Liston at Oregon State University based on OSC240800 and OSC240801 (both collected 2 May, 2013). Eight suboccurrences all located within 0.7 km (0.4 mi) of each other were observed 17 April through May 2 2013 including; Lat. 36.405606, Lon. — 120.747638 (est. 100 plants); Lat. 36.405218, Lon. —120.747573 (est. 100 plants); Lat. 36.404789, Lon. -120.747725 (est. 200 plants); Lat. 36.403071, Lon. -120.747033 (est. 50 plants); Lat. 36.401249, Lon. -120.745413 (est. 50 plants); Lat. 36.400904, Lon. -120.745873 (est. 200 plants); Lat. 36.400036, Lon. —120.744984 (est. 200 plants); and Lat. 36.401269, Lon. -120.744218 (est. 300 plants). The occurrence is located at the northern edge of the New Idria serpentine mass. The suboccurrences occur on vertic clay soil (Climara clay soil series) derived from serpentine and greywacke. Astragalus rattanii var. jepsonianus grows on small areas of steep, sparsely vegetated, naturally eroded slopes within a highly invaded (Avena L. and Bromus L.) annual grassland. Associated species on the naturally eroded slopes include Acanthomintha obovata Jeps. subsp. obovata. Allium crispum Greene, Allium howellii Eastw. var. sanbenitense (Traub) Traub & Ownbey, Benitoa occi- dentalis (H. M. Hall) D. D. Keck, Calochortus clavatus S. Watson var. pallidus (Hoover) P.L. Fiedl. & Zebell, Caulanthus flavescens (Hook.) Payson, Deinandra halli- ana (D.D. Keck) B.G. Baldwin, Eriogonum argillosum J. T. Howell, Lupinus microcarpus Sims, Madia radiata Kellogg, and Poa secunda J. Presl. Previous knowledge. Astragcdus rattanii var. jepsonia- nus (Jepson’s milkvetch) is an herbaceous annual endemic to the California Coast Ranges and is CRPR list IB. 2 (CNDDB 2013). Previously, the species was only known to occur in the North Coast Ranges including Colusa, Glenn, Lake, Napa, Mendocino, and Tehama counties (Liston 1992; Calflora 2013; CNDDB 2013). The species is typically found growing in grasslands or open areas in chaparral on clay soils derived from serpentine (Liston 1992; Baldwin et al. 2012; Calflora 2013; CNDDB 2013). The species has a serpentine affinity rating of 4.3 (broad endemic/strong indicator; Safford et al. 2005). Significance. The San Benito county occurrence of Astragalus rattanii var. jepsonianus extends the known range of the species approximately 285 km (177 mi) south to include the South Coast Ranges. Additional specimens examined. Characteristics of individuals from the San Benito County occurrence were compared to individuals collected from three occurrences at Walker Ridge in Colusa County. The three Colusa County occurrences are either approxi- mately or exactly the same as those represented by the herbarium specimen collections CHSC46306 (approxi- mately 0.5 km [0.4 mi] distant), UCD83528 (exactly the same), and CHSC27668 (exactly the same). Several newly pressed and dried (<1 mo old) specimens from the San Benito County and Colusa County occurrences were collected and compared by Ryan O’Dell. Flower petal coloration of individuals from the San Benito County occurrence differs slightly but consistently from individuals from Colusa County occurrences in having lighter purple coloration towards the tip of the banner with more purple coloration toward the side edges of the banner. All other characteristics examined including leaf shape, fruit shape and length, and seed shape, color, and surface texture appear to be identical between San Benito County and Colusa County individuals. Habitat features including clay soil at least partly derived from serpentine and grassland as the dominant vegetation type are similar between San Benito County and Colusa County occurrences. — Ryan O’Dell, Bureau of Land Management, 20 Hamilton Ct., Hollister, CA 95023. rodell@blm.gov. Acknowledgements I thank Aaron Liston at Oregon State University for confirming identification of the species. Literature Cited Baldwin, B. G., D. H. Goldman, D. J, Keil, R. Patterson, T. J. Rosatti, and D. H. Wilken (eds.). 2012. The Jepson Manual: vascular plants of California, 2nd ed. University of California Press, Berkeley, CA. Calflora. 2013. Astragalus rattanii var. jepsonianus. Calflora, Berkeley, California, USA. Website http://www.calflora.org (accessed 31 May 2013). California Natural Diversity Database (CNDDB). 2013. Astragalus rattanii jepsonianus. California Department of Fish and Wildlife, Sacramento, Cali- fornia, USA. Website http.7/www.dfg.gov/biogeodata/ cnddb/ (accessed 31 May 2013). Liston, A. 1992. Isozyme systematics of Astragalus sect. Leptocarpi subsect. Californici (Fabaceae). Systematic Botany 17:367-379. Safford, H. D., J. H. Viers, and S. P. Harrison. 2005. Serpentine endemism in the California flora: a database of serpentine affinity Madrono 52:222- 257. Madrono, VoL 61, No. 2, p. 251, 2014 NOTEWORTHY COLLECTION COLORADO Muscari neglectum Guss. ex Ten. (Hyacintha- ceae). — Boulder Co., City of Boulder, between Boulder Creek and Boulder Creek Bike Path near 4th St., 40.013978°, —105.291850°, riparian, numerous, 15 June 2013, R. Utz s.n. (COLO); City of Boulder, 21st St. between Walnut and Pearl St., 40.0197°, —105.2679°, uncommon along ditch, 23 May 2013, M. W. Denslow 2767 (COLO). Jefferson Co., embankment above Coors Beer railroad, accessed from Tucker’s Gulch Trail, Golden, 10 April 1998, S. Smookler s.n. (KHD). Previous knowledge. Muscari neglectum is native to Asia, North Africa, and Europe and is commonly cultivated in many parts of the world as an ornamental (Bailey and Bailey 1976; Wraga and Placek 2009). It has been reported as introduced from 27 states in the eastern and southern United States and Alaska (Straley and Utech 2002; USDA, NRCS 2013). The plant has not previously been reported outside cultivation from Colorado (Harrington 1964; Hartman and Nelson 2001; Snow 2009; Weber and Whitman 2012) or Boulder County specifically (Hogan 1993; Weber 1995). Significance. Collections of M. neglectum were made in June 2013 in a wooded riparian zone of Boulder Creek and along an irrigation ditch in the Front Range foothills of Boulder County, Colorado. The specimens collected did not appear to be persisting from cultivation and both were growing in disturbed riparian areas. We also observed this species in untended residential lawns throughout the city of Boulder. A specimen originally identified as M. bo try aides (L.) Mill, (common grape-hyacinth) was collected from Jefferson County, Colorado in 1998 {Smookler s.n. [KHD]). This species has a broader introduced distri- bution in the western United States, but also is not officially noted in Colorado flora lists (e.g., USDA, NRCS 2013). Muscari botry aides and M. neglectum are most easily separated by flower number and shape; M. botryoides typically has fewer than 20 globose flowers, whereas M. neglectum has 20-40 ovoid flowers. We have determined that Smookler s.n. is actually M. neglectum. Thus it appears that M. neglectum has been established in Colorado for at least 1 5 years with a distribution that extends beyond Boulder County. — 'Ryan M. Utz and Michael W. Denslow, National Ecological Observatory Network, Inc., 1685 38th St., Suite 100, Boulder, CO 80301. ‘rutz(^neoninc.org. Literature Cited Bailey, L. H. and E. Z. Bailey. 1976. Hortus third: a concise dictionary of plants cultivated in the United States and Canada. Macmillan Publishing Company, New York, NY. Harrington, H. D. 1964. Manual of the plants of Colorado, for the identification of the ferns and flowering plants of the state, 2nd ed. Sage Books, Chicago, IL. Hartman, R. L. and B. E. Nelson. 2001. A checklist of the vascular plants of Colorado. Rocky Moun- tain Herbarium, Laramie, WY. Hogan, T. 1993. A floristic survey of the Boulder Mountain Park, Boulder, Colorado. Natural His- tory Inventory No. 13, University of Colorado Museum, Boulder, CO. Snow, N. 2009. Checklist of vascular plants of the Southern Rocky Mountain Region (Version 3). Bernice P. Bishop Museum, Honolulu, HI. Straley, G. B. and F. H. Utech. 2002. Muscari. Pp. 316-318 in Flora of North America Editorial Committee (eds.), Flora of North America North of Mexico, Vol. 26: Magnoliophyta: Liliidae: Liliales and Orchidales. Oxford University Press, New York, NY. USDA, NRCS. 2013. The PLANTS Database, Nation- al Plant Data Center, Greensboro, NC. Website http://plants.usda.gov (accessed 23 August 2013). Weber, W. A. 1995. Checklist of vascular plants, Boulder County, Colorado. Natural History In- ventory No. 16, University of Colorado Museum, Boulder, CO. AND R. C. Wittmann. 2012. Colorado flora, eastern slope. 4th ed. University Press of Colorado, Boulder, CO. Wraga, K. and M. Placek. 2009. Review of taxons from genus Muscari cultivated in Department of Ornamental Plants in Szczecin. Herbal Polonica 55:348-353. Volume 61, Number 2, pages 151-251, published 9 May 2014 > ■■'.*<' f.'ltaf:. ’•' •^. » . ,»•' ^KinfiZlh ■ *’ ■ ^ ^»1:C :A’*^y.';r'y^''j-.' 'TT'jvy^y ^ 0| . aV-: >s. 4*. w ' I '/ ^ M... *j^.. .♦iii*h}^‘^T7~|M^^ vt; :HJ'pW- ■i'j';*ii’.i ! ii/i^iijtr '')f- •?y'''i(>.’y'j ‘ v»M.-v-‘. -y. tuur*‘*)»^:‘ <«-■••*; !'■(»• > if ^ 'lO,. ■ : Oj-'wJjJffeuii'-fl }V vj^'- V *Vt 4JM^«*>fi(l: , /M.v ' , . •i't)*;4llj;i> .t*!lf,Ui..o,,'.'. .J ,-ftl »>.- ■ '■€: i.'.'*w%5V9**q -rfiv p - ■ ,M. (■•^- f ', . w/, ‘. pM i.u.Mtf ’ 'lilsifa , /, ,:‘ i?v-;^ :- . '^>...41 'U-t! »'./K;;'. luH ^ ♦o-:l..*i- 1 V ,f.r ,-svipct.’Spi»-^'^ itt(i.‘ /'.t'stf ■:■} 4' ’ M*i I V -V ■ ' . /f ■y’i Si0k/iffl iVb'/'ii'-- ' f; >'»' .'*t '.’•''‘■/i/. i . '-cl. mO". •. *<'i u;» i^h ■■;■■- fr>c» j^r ■ . . v- ■» j, ••» -'•. •uil .«I5^ •! . ^ife.;^f^>/v\>. -....t^. .<» -t- , ' ■•'‘i'^ '^R 'j . Ija. ' '’•»'■ 1i»1, Ui't ' A.'' I I -i- i J ■' ■ ■ •' •’'‘‘ ■■' ■- ‘^. ■' o -■ i V i«. U''. . s ,f fc .1 WwVKSS. ..}>#'**■ jyi »i u . •. ^ni *..'■ ■ -.:. -oi ■ e^i •> 4}>miL' V* V-JD", ' ,'."f • " .Vl:vw; ■ ■ •Hv; H ♦ ‘ ui' t -rSki”'" '^Cii ipo^f ji.. juissri I] r<*&' » '. l-j • > t.-.* * ■ i;' f'..,.. . ;(<•• K^,,,.it0‘-4'riw t ^ 'l,r • : ,.,/'^,f .»fcV, '.114 Tt, ■*...- ♦•H..' > 'H- • - » ' »0.'w»' ' ■ ',■ , ■Oi.-'r>, V ov-. ' • 1 ■.’c-. 4— . «*#... ■ ^♦.:- Id •. ■ . .1'^ W ^ -lit, 5llfc ' •■ '«(#- vr- • , .M,>»-vI ) . . . i iitiilliiiii 4r, ||i;. ■ i^j -v- u ^ ^*r0 Mrtl^rtHiniitfiT 1 r.. fe:'.. .^V iTf? viid; !j«iii]iii '-'•t . r V 'i^. yi t.'i -’vr..fc .r». • ' ' '■ ‘ J f it '>“3. -vifM'i/t j* , ' . ' ;. ■ *, ’ • m' t ^ >' ■■ <.•'*• ■ . - •■ • . ■ . *:•, , ’ ' ■•:'; ■ ^.fiU‘7ut'r - \u*M. r;^v. ..- ' !••' ■ / :i -fiV',; * :. , -< ; ’ ■, ■■• > .'.?■» V ,; Mr. iJ!*t u-">'.. \ , , ' . : -‘(r' ‘ ►' ./ Wv'"-. \ 1 ■ • H- V. » '• . ■«'* «'*■'»■, f .| '•.. ,! ■ .>■ , 1, Aj,i» -«fa-L% r'. ■ -•■-.• '■ ' -*^-0 V, ..M . *^17 T.rd-f ■ ■ ■“ *■ % ■' ■ • . ■ ‘"'U^ _ ■ • i • V »t,»r .. 'x ' ’ ■'■- V .t, ';■ 11 ‘ - .V » fvu>r : • ■ ' '■ , I #V-* ?. a- ■ ■'. ' ‘'i' .. ;,. '' V't' '■^ *:■• t . ^f, \ Subscriptions— Membership The California Botanical Society has several membership types (individuals ($40 per year; family $45 per year; emeritus $32 per year; students $27 per year for a maximum of 7 years). Late fees may be assessed. Beginning in 2011, rates will increase by $5 for all membership types except life memberships, for which rates will increase by $100, and student memberships, which will not show a rate increase. Members of the Society receive Madrono free. Institutional subscriptions to Madrono are available ($75). Membership is based on a calendar year only. Life memberships are $850. Applications for membership (including dues), orders for subscriptions, and renewal payments should be sent to the Membership Chair. Requests and rates for back issues, changes of address, and undelivered copies of Madrono should be sent to the Corresponding Secretary. Information for Contributors Submit manuscripts to http://www.edmgr.com/madrono/. If you are submitting a manuscript for the first time to the Madrono PeerTrack website, you must first register by clicking "register now" and following the instructions. Manuscripts by authors having outstanding page charges will not be sent for review. Manuscripts may be submitted in English or Spanish. English-language manuscripts dealing with taxa or topics of Latin America and Spanish-language manuscripts must have a Spanish Resumen and an English Abstract. For all articles and short items (NOTES, NOTEWORTHY COLLECTIONS, POINTS OF VIEW, etc.), follow the format used in recent issues for the type of item submitted. Detailed information for contributors, including instructions for manuscript preparation, is available at the California Botanical Society website, www.calbotsoc. org; follow the link for Madrono. Authors are encouraged to include the names, addresses, and e-mail addresses of two to four potential reviewers with their submitted manuscript. All California Botanical Society members current in the volume year that their contributions are published are allowed five free pages per volume year. Additional pages will be charged at the rate of $40 per page. Joint authors may apply all or a portion of their respective five-page allotments to a jointly-published article. Partial pages will be charged as full. 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 color plates and other complex matter (including illustrations, charts, maps, photographs) will be charged at cost. Author’s changes after typesetting @ $4.50 per line will be charged to authors. Page charges are important in maintaining Madrono as a viable publication, and timely payment of charges is appreciated. 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. VOLUME 61, NUMBER 3 JULY-SEPTEMBER 2014 MADRONO A WEST AMERICAN JOURNAL OF BOTANY Contributions Toward a Bryoflora of Nevada: Bryophytes New for THE Silver State. Part III. John C. Brinda, Lloyd R. Stark, James R. Shevock, and John R. Spence ............................. ................... ........... 253 Recognition of TWo Species in Eremocarya (Bor agin ace AE)j^f Evidence From Fornix Bodies, NuTLEFs^^CDROLLiiS-p^AND Biogeography Michael G. Simpson, Regina At Sinip^'^ibn R Rebman, and Ronald B ................. 259 The Role of Soil CoMPETiTioN^^'^iCHE ^ ^ WipE^&ADjto^ Lo^I/HiGH Ae^SITY ANNUAL Ai^EkACEAE WNv -fe- CaREX XERQPHILA (CYP®ACEAE),*/i NW^S^DGE FRoWA^-QfilPARRAL OF Northern CaTifornia Peter F Zika, Lawrence P. Ja^kwM^'dpm Barbara ^^iUon ..................299 SiAf 'Phylogenetic Relationships and Crossing Data Reveal a New Species of Nemophila (Bor agin ace ae) Camille M. Barr ...................................................................................... 308 Nevada.. 316 PUBLISHED QUARTERLY BY THE CALIEORNIA BOTANICAL SOCIETY Madrono (ISSN 0024-9637) is published quarterly by the California Botanical Society, Inc., and is issued from the office of the Society, Herbaria, Life Sciences Building, University of California, Berkeley, CA 94720. Subscription information on inside back cover. Established 1916. Periodicals postage paid at Berkeley, CA, and additional mailing offices. Return requested. Postmaster: Send address changes to Madrono, Kim Kersh, Membership Chair, Uni- versity and Jepson Herbarium, University of California, Berkeley, CA 94720-2465. kersh@berkeley.edu. Editor — Matt Ritter Biological Sciences Department Cal Poly, San Luis Obispo 1 Grand Avenue San Luis Obispo, CA 93407 madronoeditor@gmail.com Editorial Assistant — Genevieve Walden Book Editor — Matt Ritter Noteworthy Collections Editor — David Keil Board of Editors Class of: 2014 — Brandon Pratt, California State University, Bakersfield, CA Tom Wendt, University of Texas, Austin, TX 2015 — Anna Jacobson, California State University, Bakersfield, CA CALIFORNIA BOTANICAL SOCIETY, INC. Officers for 2014-2015 President. Dr. Mark S. Brunell, Department of Biological Sciences, University of the Pacific, Stockton, CA 95211, mbrunell@pacific.edu Eirst Vice President: Andrew Doran, University and Jepson Herbaria, University of California, Berkeley, CA 94720, andre wdoran @ berkeley.edu Second Vice President: J. Travis Columbus, Rancho Santa Ana Botanic Garden, Claremont Graduate University, 1500 North College Avenue, Claremont, California 91711, j.travis.columbus@cgu.edu Recording Secretary: Nancy Morin, P. O. Box 716, Point Arena, CA 95468, Nancy.Morin@nau.edu Corresponding Secretary: Sheryl Creer, Department of Biology, San Francisco State University, San Francisco, CA 94132, secretary@calbotsoc.org Treasurer: David Margolies, California Botanical Society, Jepson Herbarium, University of California, Berkeley, CA 94720, dm@franz.com The Council of the California Botanical Society comprises the officers listed above plus the immediate Past President, V. Thomas Parker, Department of Biology, San Francisco State University, San Francisco, CA 94132, parker@ sfsu.edu; the FJ/Yor of Madrono, Matt Ritter; the Membership Chair, Kim Kersh, University and Jepson Herbaria, University of California, Berkeley, CA 94720, kersh@berkeley.edu; Council Members: Staci Markos, University and Jepson Herbaria, University of California, Berkeley, CA 94720, smarkos@berkeley.edu; and Dylan Burge, California Academy of Sciences, dylan.o.burge@gmail.com; Graduate Student Representatives: Jessica Orozco, Rancho Santa Ana Botanic Garden, 1500 N. College Ave., Claremont, jessica.orozco@cgu.edu; and Adam Schneider, University and Jepson Herbaria, University of California, Berkeley, CA 94720, acschneider@berkeley.edu; Administrator: Lynn Yamashita, University of California, Berkeley, CA 94720, admin@calbotsoc.org; Webmaster: Will Freyman, University of California, Berkeley, CA 94720, freyman@berkeley.edu. @ This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Madrono, VoL 61, No. 3, pp. 253-258, 2014 CONTRIBUTIONS TOWARD A BRYOFLORA OF NEVADA: BRYOPHYTES NEW FOR THE SILVER STATE. PART III John C. Brinda Missouri Botanical Garden, P.O. Box 299, St. Louis, MO 63166-0299 Lloyd R. Stark School of Life Sciences, University of Nevada, Las Vegas, NV 89154-4004 James R. Shevock Department of Botany, California Academy of Sciences, 55 Concourse Drive, Golden Gate Park, San Francisco, CA 94118=4503 j she vock@calacademy . o rg John R. Spence National Park Service, Glen Canyon National Recreation Area, Page, AZ 86040=1507 Abstract Forty-two mosses and three liverworts are reported as new for Nevada. The genera Amblyodon, Campy liadelphus, Isopterygiopsis, Mesoptychia, Myurella, Plagiobry aides, Plagiobryum, Platydictya, Pseudo campy Hum, and Preissia are new for the state of Nevada. Splachnobryum obtusum and Pseudoleskeella rupestris are now confirmed for Nevada based on recently collected material. Grimmia texicana is excluded from the Nevada checklist based on a herbarium label locality error. The bryoflora of Nevada with 355 taxa is comprised of two hornworts, 49 liverworts, and 304 mosses. Key Words: bryophyte inventory. The most recent published checklist of Nevada bryophytes listed 310 taxa (Brinda et al. 2007). Other species new to science occurring in Nevada have subsequently been published (Greven 2010; Medina et al. 2011; Spence and Shevock 2012). In the third installment of this series we add 45 taxa as new for Nevada based on voucher specimens. Each taxon has a short entry for specimen location, herbaria citation, and notes on habitat and substrate. Field collecting and inventory efforts continue to document new taxa for the Nevada bryoflora (Shevock et al. 2005; Spence et al. 2006). Many of the species reported herein have been reviewed, examined, or determined by persons involved in the bryophyte volumes of the Flora of North America Project (FNA 2007, 2014). Liverworts Lophozia heterocolpos (Thedenius) M. Howe [Scapaniaceae] Elko Co.: Humboldt National Forest. East Humboldt Wilderness, base of cirque, head- waters of Fourth Boulder Creek, 16 Aug 2007, 9475 ft, Shevock 30298 (CAS) [determined by Bill Doyle], On seepy metamorphic rock walls and ledges in a subalpine conifer forest. Mesoptychia badensis (Gottsche) L. S5derstr5m & Vaha [Jungermanniaceae] Clark Co.: Toiyable National Forest, Mt. Charleston Wilderness, South Loop Trail, 28 Sep 2008, 10,040 ft, Brinda 2361b (MO, SIU) [confirmed by Vadim Bakalin]. Preissia quadrata (Scopoli) Nees [Marchantiaceae] Clark Co.: Toiyabe National Forest, Mt. Charleston Wilderness, South Loop Trail, 28 Sep 2008, 10,040 ft, Brinda 2359 (UNLV) and Brinda 2361a (CAS, MO, UNLV). Mosses Aloina ambigua (Bruch & W.P. Schimper) Lim- pricht [Pottiaceae] Clark Co.: Valley of Fire State Park, Mouse’s Tank Trail, 4 Mar 2006, 2030 ft, Brinda 724 (MEXU, UNLV) and Brinda 726 (CAS, MEXU, UNLV) [confirmed by Claudio Delgadillo]. Amblyodon dealbatus (Swartz ex Hedwig) Bruch & W.P. Schimper [Meesiaceae] Elko Co.: Humboldt National Forest, East Humboldt Wilderness, canyon wall above beaver ponds near East Humboldt Highland Trail, Third Boulder Creek Basin, 15 Aug 2007, 9275 ft, Shevock 30394 (CAS, NY, UNLV) [determined by Bill Buck]. On moist rock soil and litter over rock terraces and cliff face with Primula Linnaeus, Saxifraga Lin- naeus, and Veratrum Linnaeus within a subal- pine forest of Pinus flexilis E. James. 254 MADROto [VoL 61 Anoectangiiim handelii Schiffner [Pottiaceae] Clark CO'.: Spring Mountains: cliff above Pine Creek, 4 Apr 2005, 4500 ft, Brinda and Shevock 112 (CAS, UNLV); hill east of Blue Diamond, 26 Feb 2006, 3350 ft, Brinda 695a (UNLV); White Rock Spring, 26 Feb 2006, 4825 ft, Brinda 703 (UNLV); Gateway Canyon, 26 Feb 2006, 3865 ft, Brinda 721 (UNLV); south fork of Pine Creek, 27 Apr 2006, 4265 ft, Brinda 1012 (UNLV); Virgin Mountains, Sandstone Bluffs above the Corral North of Quail Point, 28 Jan 2006, 4265 ft, Brinda 587 (MO, UNLV) [confirmed by Richard Zander]; Valley of Fire, Mouse’s Tank, 4 Mar 2006, 2040 ft, Brinda 727b (UNLV); Valley of Fire, Mouse’s Tank Trail, 10 Mar 2007, 1965 ft, Brinda 1284 (UNLV); Frenchman Mountain, Great Un- conformity, 5 Mar 2006, 2170 ft, Brinda 782 (UNLV); Muddy Mountains, Bowl of Fire, 17 Feb 2007, 2095 ft, Brinda 1265 (CAS); Muddy Mountains, Bowl of Fire, 17 Feb 2007, 2095 ft, Brinda 1269 (UNLV). Brachythecium turgidum (C.J. Hartman) Kind- berg [Brachytheciaceae] Elko Co.: East Humboldt Wilderness, Humboldt National Forest, slopes south of Third Boulder Creek toward Birdeye Lake, 1 1 Aug 2007, 9250 ft, Shevock 30324 (CAS, UNLV). This rather glossy appearing species displays a distribution pattern of widely disjunct occur- rences across the Intermountain West. The Nevada occurrence is from vertical rock walls with dripping water in a subalpine forest. Bryum calobryoides J.R. Spence [Bryaceae] Clark Co,: Spring Mountains, Rainbow Moun- tain Wilderness, south fork of Pine Creek, 29 Apr 2008, 4690 ft, Brinda 2263 (CAS, UNLV); Spring Mountains, Rainbow Mountain Wil- derness, south fork of Pine Creek, 29 Apr 2008, 4690 ft, Brinda 2279 (MO, NY). Elko Co.: along Jarbidge River at Lower Bluster Camp- ground, Humboldt National Forest, 25 Aug 2002, 6600 ft, Shevock and Glazer 22740a (CAS). Small plants can occur among other mosses so it can be easily overlooked. The orbicular-shaped leaves are, nonetheless, high- ly diagnostic. Campyliadelphus chrysophyllus (Bridel) Chopra [Amblystegiaceae] Lander Co.: Toiyabe Range, Toiyabe National Forest, headwaters of Big Creek, 30 May 2002, 8100 ft, Shevock and Glazer 22352 (CAS, UNLV) [determined by David Toren], Didiodontium olympicum Renauld & Cardot [Dicranaceae] Elko Co.: East Humboldt Wilderness, Humboldt National Forest, slopes south of Third Boulder Creek toward Birdeye Lake, 14 Aug 2007, 9200 ft, Shevock 30342 (CAS, UNLV); East Humboldt Wilderness, Humboldt National Forest, slopes south of Third Boulder Creek toward Birdeye Lake, 14 Aug 2007, 9200 ft, Shevock 30303 (C AS); Humboldt National Forest, Ruby Moun- tains, Lamoille Canyon along Ruby Crest Trail |j to Dollar Lakes, 20 Jul 2002, 8975 ft, Shevock t and Glazer 22483a (CAS) [determined by David Toren]. Open subalpine metamorphic or gra- nitic soils over rock terraces. Didymodon eckeliae Zander [Pottiaceae] Carson City.: Clear Creek Range just south of highway 50 and 2.3 miles west of highway 395, 5 Apr 2002, 5200 ft, Shevock 21965 (CAS). Douglas Co.: Red Canyon, northern edge of Burbank Canyon Scenic Area, eastern slopes of I Pine Mountain, 7 Apr 2002, 6000 ft, Shevock et I al 22020 (CAS). Lander Co.: Toiyabe National | Forest, Kingston Canyon Road near the old • Kingston Guard Station, 5 Sep 2011, 7530 ft, ; Brinda and Banner 3154 (CAS, MO). Granitic | and marble rocks within pinyon pine wood- ; lands and Jeffrey pine forests. j Drepanodadus polygaeius (Bruch & W.P. Schim- j per) Hedenas [Amblystegiaceae] Carson City.: Sierra Nevada, Lake Tahoe Basin, highway 28 at Bliss Creek between Skunk Harbor and Secret Harbor, Shevock 22429 (CAS, UNLV). Lander Co.: Kingston Creek, j Toiyabe Range, Toiyabe National Forest, 30 May 2002, 7400 ft, Shevock and Glazer 22338 (CAS, UNLV) [determined by David Toren]. In wet meadow and fen habitats. Encalypta alpina Smith [Encalyptaceae] Elko Co,: East Humboldt Wilderness, East Hum- boldt Mts, Humboldt National Forest, wall of cirque just above headwaters of Fourth Boulder Creek, 16 Aug 2007, 9575 ft, Shevock 30434 (CAS, UNLV) [determined by David Toren; confirmed by Diana Horton]. On moist rock ledges and terraces mixed among Blepharos- toma trichophyllum (Linnaeus) Dumortier, Dis- tichium capiiiaceum (Hedwig) Bruch & Schim- per, and Piatydictya jungermannioides (Bridel) H.A. Crum, in a subalpine conifer forest. A highly unexpected addition to the bryoflora of Nevada considering the distribution pattern of widely disjunct occurrences coupled with spe- cific microhabitat requirements. Entosthodon bolanderi Lesquereux [Funariaceae] Clark Co.: Spring Mountains, Red Rock Canyon National Conservation Area. South Fork Pine Creek just above mouth of canyon, 4 Apr 2005, 4200 ft, Shevock and Brinda. 26535 (CAS) [determined by David Toren]. Entosthodon tucsoni (E.B. Bartram) Grout [Fu- { nariaceae] I Clark Co,: Spring Mountains, Red Rock Canyon j National Conservation Area, South Fork Pine | 2014] BRINDA ET AL.: BRYOPHYTES NEW FOR THE SILVER STATE. PART III 255 Creek, 4 Apr 2005, 4200 ft, Brinda and Shevock 100 (CAS, UNLV); Spring Mountains, Red Rock Canyon National Conservation Area, South Fork Pine Creek, 27 Apr 2006, 4200 ft, Brinda and Shevock 999 (UNLV); Spring Mountains, Red Rock Canyon National Con- servation Area, South Fork Pine Creek, 27 Apr 2006, 4200 ft, Shevock and Brinda 26535a (CAS) [determined by David Toren], Fabronia ciliaris (Bridel) Bridel [Fabroniaceae] Clark Co.: South Virgin Mountains, Gold Butte, 21 Apr 2007, 3970 ft, Brinda 1477 (CAS, UNLV). FissMeos bryoides Hedwig [Fissidentaceae] White Pine Co,: Egan Range, Sawmill Creek Canyon 5.4 mi west of Cave Valley, 31 Jul 1979, 8000 ft, Lavin, Mozingo and Ryser M-220 (NY) [confirmed by Ronald Pursell]. Gemmabryum demaretianum (Arts) J.R. Spence [Bryaceae] Clark Co.: McCullough Mountains, pass on highway 164, 16 Oct 2005, 4815 ft, Brinda 394, 397 (UNLV); El Dorado Mountains, near highway 165, 23 Feb 2006, 3300 ft, Brinda 647 (UNLV). Esmeralda Co.: White Mountains, road to Middle Canyon, 3 Jul 2005, 7035 ft, Brinda 225 (UNLV). Gemmabryum klinggraeffii (W.P. Schimper) J.R. Spence & H.P. Ramsay [Bryaceae] Clark Co,: University of Nevada, Las Vegas, Main Campus, lawn Between Physics and Health Sciences Buildings, 27 Jan 2011, 2030 ft, Brinda and Greenwood 2615 (UNLV). Gemmabryum violaceum (Crundwell & Nyholm) J.R. Spence [Bryaceae] Churchill Co.: Stillwater Range between Sham- rock Canyon and Fondaway Canyon, 3 May 2006, 1500 m, Tonenna 38 (CAS); Stillwater Range between Shamrock Canyon and Fond- away Canyon, 3 May 2006, 1500 m, Tonenna 43 (CAS). Clark Co.: Spring Mountains, Bonanza Peak Trail, 4 Jul 2005, 8140 ft, Brinda 296 (UNLV); Virgin Mountains, Cabin Creek, 21 Apr 2006, 4560 ft, Brinda 916a (UNLV). Lincoln Co.: Hiko Range, highway 93 about 3 miles E of highway 375, 3 Apr 2005, 4050 ft, Brinda and Shevock 12a (UNLV); Hiko Range, highway 93 about 3 miles E of highway 375, 3 Apr 2005, 4050 ft, Brinda and Shevock 15b (UNLV). Nye Co.: Death Valley National Park, Grapevine Mountains, Amargosa Range above Strozzi Ranch site south of Wahguyhe Peak, 1 1 Jan 2002, 6800 ft, Shevock and York 21687 (CAS, UNLV). Grimmia vaginulata K. Kellman [Grimmiaceae] Clark Co.: Frenchman Mountain, below north summit, 30 Jan 2007, 3745 ft, Brinda 1251 (UNLV); Lake Mead NR A, Northshore Drive south of the Muddy Mountains, 1 1 Mar 2007, 1615 ft, Brinda and Rosentreter 1301 (UNLV) [confirmed by Ken Kellman]; South Virgin Mountains, along Gold Butte Road east of Tramp Ridge, 21 Apr 2007, 2935 ft, Brinda and Shevock 1490 (UNLV) [confirmed by Ken Kellman]; Virgin Mountains, Whitney Pockets, 21 Apr 2007, 3060 ft, Brinda and Shevock 1510b (UNLV); Lake Mead NRA, sandstone outcrops along the south side of Northshore Road, about 1 km east of West End Wash, 31 Mar 2013, 1730 ft, Brinda and Stark 4448 (CAS, MO, NY, UNLV). This recently de- scribed species (Kellman 2011) was known only from the type locality in coastal central California. The occurrences for Nevada are therefore a remarkable extension in both distribution and habitat for this species. Grimmia vaginulata is not uncommon on the red sandstone outcrops of southern Nevada. Gymnostomum calcareum Nees & Hornschuch [Pottiaceae] Clark Co.: BLM Red Rock Canyon National Recreation Area. North Fork Pine Creek, 4 Apr 2005, 4300 ft, Shevock, Brinda, and Stark 26507 (CAS, UNLV) [determined by David Toren]; BLM Red Rock Canyon National Recreation Area. North Fork Pine Creek, 4 Apr 2005, 4300 ft, Brinda, Shevock, and Stark 78 (UNLV). Elko Co.: Ruby Mountains, Humboldt Na- tional Forest, between Liberty Lake and Liberty Pass, 31 Jul 2004, 10,325 ft, Shevock 26039a (CAS) [determined by David Toren]. Gymnostomum viridulum Bridel [Pottiaceae] Clark Co.: Spring Mountains, Gateway Canyon, 26 Feb 2006, 3865 ft, Brinda 720 (CAS, MO, UNLV) [determined by Richard Zander]. This species resembles and occupies the same microhabitats as G. calcareum but the axillary gemmae in G. viridulum readily distinguish this species. HomomalMum mexicanum Cardot [Hypnaceae] Clark Co.: Red Rock Canyon National Recrea- tion Area, Willow Spring Picnic Area, 2 Apr 2003, 4825 ft, Shevock and Stark 23970 (CAS, UNLV); Red Rock Canyon National Recrea- tion Area, Willow Spring Picnic Area, 2 Apr 2003, 4825 ft, Shevock and Stark 23972 (CAS) [confirmed by David Toren], On sandstone boulders. White Pine Co.: Snake Range, Great Basin National Park, Grey Cliffs off of Baker Creek Road toward Grey Cliffs Group camp- ground, 27 Jul 2003, 7200 ft, Shevock, Glazer and Clifton 24343a (CAS). On lime- stone walls mixed with Hypnum vaucheri Lesquereux. Hygroamblystegium noterophilum (Sullivant & Lesquereux ex Sullivant) Warnstorf [Amblys- tegiaceae] 256 MADRONO [VoL 61 Elko Co,: Ruby Valley, 20 May 1935, 6000 ft, fV. W. Bennett s,n. (NY 714904) [confirmed by Ryszard Ochyra]. Hygrohypnum bestii (Renauld & Cardot) Broth- erus [Amblystegiaceae] Elko Co,: East Humboldt Wilderness, Humboldt National Forest, slopes between Boulder Lake and fens with beaver ponds above East Humboldt Highline Trail, 15 Aug 2007, 9490 ft, Shevock 30389 (CAS). Esmeralda Co,: White Mountains, Inyo National Forest, Morris Creek, 28 May 2002, 6800 ft, Shevock and Glazer 22292 (CAS). Washoe Co.: Lake Tahoe Basin, shaded streamlet above Incline Village, 29 Jul 2002, 7500 ft, Shevock 22607 (CAS) [determined by Dave Jamieson]. On rocks in splash zone of stream. Hygrohypnum styriacum (Limpricht) Brotherus [Amblystegiaceae] Elko Co.: East Humboldt Wilderness, Humboldt National Forest, base of cirque, headwaters of Fourth Boulder Creek off of East Humboldt Highline Trail, 16 Aug 2007, 9475 ft, Shevock 30418 (CAS, MO, NY, UNLV) [determined by David Toren]. On seepy metamorphic rock walls and ledges with Primula and Saxifraga within a subalpine Pinus flexilis forest. Imbribryum mildeanum (Juratzka) J.R. Spence [Bryaceae] Humboldt Co,: Santa Rosa Range, Humboldt National Forest, Lye Creek about 2 mi west of Martin Creek Ranger Station, 23 June 2004, 7915 ft, Nachlinger and Tiehm 2640 (CAS, MO, NY, UNLV, WTU). Lander Co.: Sheep Creek Range, Plateau north of Battle Creek, 12 June 2002, 6595 ft, Nachlinger and Tiehm 2446 (CAS, MO, NY, UBC, UNLV, WTU). White Pine Co.: Antelope Range, Lookout Spring, 10 June 2002, 6680 ft, Nachlinger and Tiehm 2441 (CAS, MO, NY, UBC, UNLV, WTU). Isopterygiopsis pulchella (Hedwig) Z. Iwatsuki [Hypnaceae] Elko Co.: Jarbidge Mountains, Humboldt Na- tional Forest, road paralleling Jarbidge River above Pine Creek Campground just above junction with Fox Creek, 25 Aug 2002, 6700 ft, Shevock and Glazer 22762 (CAS) [determined by David Toren], Jaffueliobryum raui (Austin) Theriot [Grimmia- ceae] Clark Co.: Virgin Mountains, near Red Bluff Spring, 19 Feb 2006, 2100 ft, Brinda 645b (UNLV); Gold Butte Backcountry Byway Road at Whiskey Pocket west of Whiskey Pass, 21 Apr 2007, 3025 ft, Shevock and Brinda 29783 (CAS, UNLV) [determined by David Toren]. Myurella tenerrima (Bridel) Lindberg [Pterigy- li nandraceae] Elko Co.: East Humboldt Mountains, East j|: Humboldt Wilderness, Humboldt National iij Forest, along East Humboldt Highline Trail i just below saddle, Fourth Boulder Creek | Basin, 16 Aug 2007, 9635 ft, Shevock 30406 | (CAS, UNLV) [determined by David Toren]. | Crevice of metamorphic rock outcrops in a jj' subalpine forest of Pinus flexilis. Orthotrichum pilosissimum Medina, Lara & | Garilleti [Orthotrichaceae] Esmeralda Co.: Grapevine Mountains, Amargosa Range, Death Valley National Park, just south of Strozzi Ranch site and Brier Spring, 14 Dec j 2001, 6400 ft, Shevock, Davis and Davis 21581a \ (CAS); White Mountains, Inyo National For- I est. Morris Creek, north base of Boundary 'f Peak, 31 May 2002, 6800 ft, Shevock 221 - (CAS). Eureka Co.: Toiyabe National Forest along road 004, 9.4 mi east of Monitor Valley, i| May 2002, 7500 ft, Shevock and Glazer 22368 || (CAS). Humboldt Co.: Santa Rosa Mountains, j Humboldt National Forest, Buffalo Creek, * j Oct 2008, 5200 ft, Garilleti et al s.n. (CAS, NY, VAL). Mineral Co.: Anchorite Hills, Toiyabe National Forest, Box Canyon, 28 Oct 2008, 7000 ft, Lara et al s.n. (CAS, UC). j Nye Co.: Alta Toquima Wilderness, Toiyabe ; National Forest, Pine Creek above Pine Creek | Campground, May 2002, 7700 ft, Shevock and Glazer 22397 (CAS). This species was recently ^ described for Nevada (Medina et al. 201 1). We , have added it here in an effort to track all taxa j new for Nevada since the published checklist j (Brinda et al. 2007). ■ '| Philonotis caespitosa Juratzka [Bartramiaceae] | Elko Co,: East Humboldt Mountains, East I Humboldt Wilderness, Humboldt National ! Forest, cirque below Greys Peak west of Angel ' Lake, 21 Jul 2002, 8800 ft, Shevock and Glazer \ 22582 (CAS, UC, UNLV) [determined by Dan | Norris; confirmed by David Toren]. On moist to wet soil, high elevations. Plagiobryoides viuosula (Cardot) J.R. Spence ■ [Bryaceae] j Churchill Co.: Stillwater Range between Sham- i rock Canyon and Fondaway Canyon, 3 May j 2006, 1500-1700 m, Tonenna 28 (CAS); Still- water Range between Shamrock Canyon and Fondaway Canyon, 3 May 2006, 1500-1700 m, i Tonenna 37 (CAS). Lander Co.: Toiyabe Range, Toiyabe National Forest, along forest ; road 002, 3.5 miles above Kingston Guard i Station, 30 May 2002, 30 May 2002, Shevock and Glazer 22345 (CAS). This species has also i recently been documented for Death Valley i National Park at Grapevine Springs in adja- i cent California and therefore likely to be j 2014] BRINDA ET AL.: BRYOPHYTES NEW FOR THE SILVER STATE. PART III 257 discovered in adjacent Esmeralda and Clark counties. Plants occur in wet riparian areas or seeps with high salt concentrations generally associated with either Didymodon tophaceus (Bridel) Lisa or Eudadium verticillatum (Hed- wig) Bruch & Schimper. Plagiobryum zierii (Dickson ex Hedwig) Lindberg [Bryaceae] Elko Co.: East Humboldt Mountains, Humboldt National Forest, cirque above falls west of Angel Lake, 3 Aug 2004, 8500 ft, Robertson 8669 (CAS, UC); East Humboldt Wilderness, East Humboldt Mts., base of cirque, headwa- ters of Fourth Boulder Creek, 16 Aug 2007, 9475 ft, Shevock 30426 (CAS, NY, UC, UNLV) [confirmed by Spence]. Wet soil over metamorphic walls and ledges with Salix, Primula, and Saxifraga in a subalpine Pinus fiexiiis forest. Platydictya jungermannioides (Bridel) H. Crum [Hypnaceae] Clark Co.: Toiyabe National Forest, Spring Mountains National Recreation Area, Mary Jane Falls, 3 Oct 2004, 9045 ft, Brinda 410 (UNLV); Toiyabe National Forest, Mt. Charleston Wilderness, wash below Big Falls, 3 Oct 2004, 9055 ft, Brinda 414 (CAS). Elko Co.: East Humboldt Mountains, East Hum- boldt Wilderness, Humboldt National Forest, avalanche chute with steep rocky canyon, Third Boulder Creek Basin, 15 Aug 2007, 9275 ft, Shevock 30396 (CAS, MO, NY, UNLV); wall of cirque just above headwaters of Fourth Boulder Creek, 16 Aug 2007, 9575 ft, Shevock 30434b (CAS). Mixed with Encalypta alpina, Blepharostoma trichophyllum, and Dis- tichium capillaceum. White Pine Co.: Snake Range, Great Basin National Park, Snake Creek near Shoshone Campground, 26 Jul 2003, 8200 ft, Shevock, Glazer and Clifton 24294a (CAS); on soil bank of creek mixed with Mnium marginatum (Dickson ex Wither- ing) Palisot de Beauvois; upper slopes of Blue Canyon, 27 Jul 2003, 9300 ft, Shevock, Glazer and Clifton 24318a (CAS); on soil bank of creek mixed with Mnium marginatum; upper slopes of Blue Canyon, 27 Jul 2003, 9300 ft, Shevock, Glazer and Clifton 24325 (CAS, MO, NY, UNLV) [confirmed by David Toren]. This is a very small moss that can easily be overlooked. It frequently occurs in mixed populations among other larger mosses and liverworts. Microhabitat is damp soil and rock recesses primarily in shaded places. Pseudocampylium radicale (P. Beauvois) Vander- poorten & Hedenas [Amblystegiaceae] Elko Co.: Jarbidge Mountains, Humboldt Na- tional Forest, along forest road 748 at Seven- tysix Creek, Major Gulch, 24 Aug 2002, 6250 ft, Shevock and Glazer 22678 (CAS, UNLV); Freighters Defeat along road 752 north of town of Jarbidge, 26 Aug 2002, 5650 ft, Shevock and Glazer 22803 (CAS, NY, UNLV). Humboldt Co.: Pine Forest Range, BLM Winnemucca Dist., upper slopes of Wood Canyon, 5 Aug 2005, 7150 ft, Shevock and Nachlinger 27501 (CAS, UNLV) [determi- nations by David Toren]. Along stream cours- es. This plant can readily be separated from Amblystegium serpens by the presence of paraphyllia on the stems. An earlier name for this species is Amblystegium radicale. Pseudoleskeella rupestris (Berggren) Hedenas & Sdderstrom [Leskeaceae] Clark Co.: Red Rock Canyon National Recrea- tion Area, North Fork Pine Creek, 4 Apr 2005, 4500 ft, Shevock, Brinda and Stark 26526 (CAS). In the Nevada checklist (Brinda et al. 2007) this species was reported from Nevada but we had not previously seen a specimen to validate its occurrence in the state. Ptychostomum pacMcum J.R. Spence & Shevock [Bryaceae] Washoe Co.: Sierra Nevada, Tahoe Meadows, Toiyabe National Forest, 13 Sep 2009, 8740 ft, Wishner 9512 (CAS, UC). This species was recently described as new to science (Spence and Shevock 2012). We have included it here in an effort to track all taxa new for Nevada since the published checklist (Brinda et al. 2007). This species is likely to remain a very rare species in Nevada although it is widespread in fens in the Californian portion of the Sierra Nevada. Splachnobryum obtusum (Bridel) C. Muller Hal. [Splachnobryaceae] Clark Co.: Lake Mead National Recreation Area, Gold Strike Canyon Hot Springs, 4 Apr 2010, 985 ft, Brinda 2613 (CAS, MO, NY, UNLV) and Webb s.n. [Mar 2010] (UNLV); Virgin Basin, Lake Mead at water’s edge, 5 miles north-northeast of Middle Point, 22 Apr 1941, Clover 6269 (NY!). In Brinda et al. (2007) we had doubts about this taxon actually occurring in Nevada and it was assumed the record was likely to be a herbarium processing error. However, with this collection we confirm this taxon for the Nevada bryoflora. Tortella alpicola Dixon [Pottiaceae] Clark Co.: Spring Mountains, north-facing Ra- vine above South Fork of Pine Creek, 27 Apr 2006, 4250 ft, Brinda 1018 (CAS, MO, UNLV); Sheep Range, near the end of (dosed) Sawmill Road, 6 Aug 2006, 7578 ft, Brinda 1213 (MO, UNLV). Lincoln Co.: Mormon Mountains, Upper South Fork Toquop Wash, 14 May 2005, 6860 ft, Brinda 132 (CAS, MO, UNLV) [confirmed by Patricia Eckel]. 258 MADRONO [Vol. 61 Warnstorfia fluitans (Hedwig) Loeske [Calliergo- naceae] Washoe Co.: Sierra Nevada, Carson Range, Big Meadows South of Verdi, north end of Meadows, 25 Sep 1998, 8715 ft, Tiehm 12764 (NY) [determined by David Toren]. Warnstorfia pseudostraminea (C. Muller Hal.) Tuomikoski & T. Koponen [Calliergonaceae] White Pine Co.: Snake Range, Great Basin National Park, Alpine Lakes Trail from Wheeler Peak trailhead to Stella Lake, 27 Aug 2005, 10,070 ft, Shevock and Giazer 27619 (CAS, UNLV) [determined by David Toren]. Weissia ligulifolia (E.B. Bartram) Grout [Pottia- ceae] Clark Co.: Bureau of Land Management, Las Vegas Dist., Virgin Mountains, White Rock site at mouth of Cabin Canyon, 15 Nov 2002, 3950 ft, Shevock et al. 23632 (CAS, MO, UNLV) [determined by Richard Zander]; ‘Nay’ Canyon, 28 Feb 2006, 4720 ft, Brinda 583 (UNLV); near Red Bluff Spring, 19 Feb 2006, 2035 ft, Brinda 618 (UNLV); El Dorado Mountains, near highway 165, 23 Feb 2006, 3300 ft, Brinda 649 (UNLV); Spring Moun- tains, Gateway Canyon, 26 Feb 2006, 3865 ft, Brinda 715 (UNLV); Valley of Fire State Park, Mouse’s Tank Area, 4 Mar 2006, 2095 ft, Brinda 743B (UNLV). Lincoln Co.: Hiko Range, highway 9 about 3 mi east of junction with highways 375 and 318, 3 Apr 2005, 4050 ft, Brinda and Shevock 3 (UNLV); Hiko Range, highway 9 about 3 mi east of junction with highways 375 and 318, 3 Apr 2005, 4050 ft, Brinda and Shevock 7 (UNLV); Hiko Range, highway 9 about 3 mi east of junction with highways 375 and 318, 3 Apr 2005, 4050 ft, Shevock and Brinda 26468 (CAS, UNLV); Hiko Range, highway 9 about 3 mi east of junction with highways 375 and 318, 3 Apr 2005, 4050 ft, Shevock and Brinda 26476 (CAS); South Pahroc Range, Pahroc Summit, 3 Apr 2005, 5000 ft, Shevock and Brinda 26498 (CAS, NY, UNLV) [determined by David Toren]. Excluded Species Grimmia texicana Greven [Grimmiaceae] Clark Co.: 2 mi W of Las Vegas. This record is based on a single herbarium specimen at FH, examined and subsequently cited as this species in Greven (2010). The collector’s name and date of collection were omitted in the 2010 publication. According to Greven (2010), this species is related to Grimmia arizonae. We have determined that the collection cited by Greven (2010, and personal communication) actually was collected near Las Vegas, New Mexico rather than Las Vegas, Nevada. Therefore, the literature reference of Grimmia texicana occur- ring in Nevada is erroneous and this species is hereby excluded from the Nevada bryoflora. Acknowledgments We thank the various land management agencies throughout Nevada for providing collecting permits for this long-term inventory project. Identifications or confirmations provided by Vadim Bakalin, Bill Buck, Claudio Delgadillo, Bill Doyle, Patricia Eckel, Diana Horton, Dave Jamieson, Ken Kellman, Dan Norris, Ryszard Ochyra, Ronald Pursell, David Toren, and Richard Zander are greatly appreciated. We are also appreciative of comments provided by two anonymous reviewers that enhanced the final version. Literature Cited Brinda, J. C., L. R. Stark, J. R. Shevock, and J. R. Spence. 2007. An annotated checklist of the bryophytes of Nevada, with notes on collecting history in the state. The Bryologist 110:673-705. Flora of North America Editorial Committee (eds.). 2007. Flora of North America. North of Mexico, Vol. 27. Mosses Part 1. Oxford University Press, New York and Oxford. Flora of North America Editorial Committee (eds.). 2014. Flora of North America. North of Mexico, Vol. 28. Mosses Part 2. Oxford University Press, New York and Oxford. Greven, H. 2010. Grimmia texicana sp. nov. (Grim- miaceae) from Texas and its separation from Grimmia arizonae. The Bryologist 113:360-364. Kellman, K. 2011. Grimmia vaginulata (Bryopsida, Grimmiaceae) a new species for the central coast of California. Madrono 58:190-198. Medina, R., F. Lara, V. Mazimpaka, J. R. She- vock, and R. Garilletl 2011. Orthotrichum pilosissimum (Orthotrichaceae), a new moss from arid areas of Nevada with unique axillary hairs. The Bryologist 114:316-324. Shevock, J. R., J. R. Spence, and L. R. Stark. 2005. Contributions toward a bryoflora of Nevada: bryophytes new for the Silver State. Part I. Madrono 52:66-71. Spence, J. R., L. R. Stark, and J. R. Shevock. 2006. Contributions toward a bryoflora of Nevada: bryophytes new for the Silver State. Part II. Madrono 53:400M03. AND J. R. Shevock. 2012. Ptychostomum pacificum (Bryaceae) a new fen species from California, Oregon, and western Nevada, USA. Madrono 59:156-162. Madrono, VoL 61, No. 3, pp. 259-275, 2014 RECOGNITION OF TWO SPECIES IN EREMOCARYA (BORAGINACEAE): EVIDENCE FROM FORNIX BODIES, NUTLETS, COROLLA SIZE, AND BIOGEOGRAPHY Michael G. Simpson, Regina A. Dowdy, and Lee M. Simpson Department of Biology, San Diego State University, San Diego, CA 92182 msimpsoni@mail.sdsu.edu Jon P. Rebman San Diego Natural History Museum, P.O. Box 121390, San Diego, CA 92112-1390 Ronald B. Kelley Department of Chemistry and Biochemistry, Eastern Oregon University, One University Boulevard, La Grande, OR 97850-2890 Abstract Eremocarya (Boraginaceae), a recently resurrected segregate of the genus Cryptantha, has generally been recognized as containing a single species, E. micrantha, with two varieties. Here we present evidence that these two varieties are distinct in a number of features and that they should be treated as separate species: Eremocarya lepida and E. micrantha. Eremocarya lepida differs from E. micrantha in having a significantly greater corolla limb width, nutlet length, maximum nutlet width, and maximum nutlet width: apical nutlet width. Eremocarya lepida also has prominent yellow fornices near the apex of the corolla throat, whereas fornices are absent and the fornix region lacks pigmentation in E. micrantha. In addition, we report the discovery of clusters of minute (ca. 0.1 mm long), transparent, stalked, ellipsoid structures born near the apex of the inner corolla tube that are associated with the five corolla fornices, these being unique to E. lepida. These structures, which we term “fornix bodies,” are of unknown chemistry and function, but they may possibly have a role in the pollination of the showier, larger-flowered E. lepida. In addition to these morphological characters, the two species differ in distribution, elevation, and plant community /vegetation. Eremocarya lepida occurs at higher elevation in chaparral, coniferous woodland, and high desert scrub of southern California and northern Baja California, Mexico. Eremocarya micrantha occurs at lower elevations in desert habitats of Arizona, California, New Mexico, Texas, Oregon, and Utah in the United States, and Baja California and Sonora in Mexico. All of these data strongly support recognition of two species in Eremocarya. Key Words: Biogeography, Boraginaceae, Cryptantha micrantha, Cryptantha micrantha var. lepida, Eremocarya lepida, Eremocarya micrantha, fornix/fornices, fornix bodies. In 1859 Torrey named the flowering plant Eritrichium micranthum Torrey (family Boragina- ceae), as part of the Report on the United States and Mexican Boundary Survey (Holotype G. Thurber 181, Apr 1851; NY 00335240). Torrey diagnosed this new species as a small, canescent- hispid annual with slender, much-branched stems (Fig. lA), linear, obtuse leaves, bracteate and crowded flowers (Fig. IB), and minute corollas, less than “a line” long [i.e., less than ca. 2.1 mm] (Fig. IB, D), the corollas lacking appendages (Fig. IE). Nutlets are described as “about one- third of a line long [i.e., ca. 0.7 mm], narrowly oblong, shining,” glabrous, with a prominent inner sulcus [ventral groove] (Fig. IG; however, note both glabrous and papillate nutlets here), the nutlets adhering to the whole length of the column [gynobase] (Fig. IF). (Note: a “line” is assumed to be 1/12 in., or approximately 2.1 mm) Subsequent to Torrey ’s publication. Gray (1878) named Eritrichium micranthum var. lepi- dum A. Gray (Holotype: D. Cleveland s.n., GH 00097023). Compared to E. micranthum [var. micranthum]. Gray described var. lepidum as “less slender and more hirsute,” with a corolla that is “larger, its expanded limb 2 or 3 lines [ca. 4- 6 mm] in diam. (Fig. 2A, B), the appendages or folds in the throat very manifest (Fig. 3B-E); nutlets nearly a line [i.e., a little less than 2 mm] long, puncticulate-scabrous” (Fig. 2E), Gray contrasts Eritrichium micranthum Torrey var. micrantha as having corolla lobes “one to two- thirds of a line long,” [i.e., ca. 1. 4-2.1 mm broad] (Fig. ID), stating that the corolla is “obscurely appendaged at the throat” (Fig. IE). Gray described the nutlets of var. micrantha as “half to two-thirds of a line long” [ca. 1-1.4 mm], and “smooth and shining or dull and puncticulate- scabrous” (conforming to the variation seen in Fig. IG). In 1885 Gray classified the two varieties of Eritrichium in the genus Krynitzkia, as K. micrantha (Torrey) A. Gray [var. micrantha] and K. micrantha var. lepida A. Gray, commenting 260 MADRONO [Vol. 61 Fig. 1. Eremocarya micrantha. A-C. Field photographs {Simpson 3126, SDSU 19604). A. Whole plant, a small annual. B. Inflorescence. Note small flowers. C. Root, with red pigmentation. D-E. Flowers, rehydrated from dried herbarium material of same collection {Simpson 3126, SDSU 19604). D. Flower, side view, showing calyx and funnelform corolla. E. Open corolla, showing pistil, anthers (note two apical appendages and different levels of attachment), and absence of fornices. F. Fruiting calyx, with persistent nutlet attached to gynobase. G. Nutlets of heteromorphic individual {Purer 4943, SD 39169): “rough” (densely papillate) large/solitary nutlet at left (far left, dorsal view; middle, ventral view) and one of three smooth small/consimilar nutlets at right (dorsal view only). 2014] SIMPSON ET AL.: RECOGNITION OF TWO SPECIES IN EREMOCARYA 261 Fig. 2. Eremocarya lepida. A-B. Field photographs. A. Inflorescence {Simpson 2816, SDSU 17572). B. Flower close-ups {Rebman, 29 Mar 2007). C-D. Mature fruits {Simpson 2369, SDSU 17281). C. Fruiting calyx, with persistent nutlet attached to gynobase. D. Another fruiting calyx, showing all four, homomorphic nutlets. E. One of four homomorphic nutlets {Simpson 2816, SDSU 17572), in dorsal (left), ventral (middle), and side (right) views. 262 MADRONO [Vol. 61 Fig. 3. Eremocarya lepkki. A. Whole flower, rehydrated from dried herbarium material {Simpson 8VI94J, SDSU 5431), showing relatively large, overlapping corolla lobes. (Note: corolla has become detached from and is elevated slightly above, dower base.) B-F. Flowers, photographed from fresh material {Simpson 3724, SDSU 20490). B. Corolla, side view, with yellow, invaginated fornices. C. Corolla, face view, showing lobes and yellow center and fornices at upper throat. D. Close-up of C, showing fornix bodies protruding into corolla throat. E. Corolla from flower at anthesis, opened to show 13 transparent fornix bodies arising from each fornix. Note anthers (each with two apical appendages) arising at multiple levels, below. F. Corolla from slightly post-anthesis flower, opened to show fornix bodies, these apparently covered with pollen grains. 2014] SIMPSON ET AL.: RECOGNITION OF TWO SPECIES IN EREMOCARYA 263 that “the two forms are confluent.” Two years later, Greene (1887) named the new genus Eremocarya (Greek eremos, desert or solitary, and carpos, nut), transferring these two taxa in new combinations: E. micrantha (Torrey) Greene and E. lepida (A. Gray) Greene. Greene charac- terized the new genus as small, hirsute-canescent annual herbs with roots having a “deep purple stain” (Fig. 1C) [probably naphthoquinone based alkannins or shikonins; see Papageorgiou et al. 1999], “leaves all in a radical, rosulate tuft, the numerous racemose branches [actually scopioid cymes, a type of monochasium] repeatedly dichotomous and conspicuously leaf-bracted,” Greene described the fruits as having short, filiform pedicels, with persistent, campanulate calyces that are 5-parted to the base (Fig. IF), the “segments [i.e., sepals] nerveless and not hispid- bristly.” The style is described as “enlarged in fruit and persistent,” and the nutlets as “neither margined nor carinate” [i.e., the margin is rounded], “erect, attached for their whole length, the groove open, little dilated and not furcate at the base” (Figs. IG, 2E). Greene gave no rationale for deviating from Gray in treating the two former varieties as separate species. In 1909 Rydberg segregated the rough-nutlet form of Eremocarya as a separate species, E. muricata Rydberg; this was apparently named because the type material of Eremocarya mi- crantha (basionym Eritrichium micranthum Tor- rey) has smooth (“shining”) nutlets, warranting description of this new, rough-nutlet form. However, Macbride (1916) rejected Rydberg’s new species, citing a co-type specimen of his Eremocarya muricata with smooth (not rough) nutlets. Macbride agreed with the ranking of Gray’s earlier classifications for the two original taxa, lowering Greene’s Eremocarya lepida to the rank of variety, as E. micrantha var. lepida (A. Gray) J. F. Macbride. Macbride implied that this demotion in rank was related to the common variation in nutlet morphology between the two forms, pointing out that in both varieties “herbarium material seems to indicate that the smooth- and rough-fruited forms grow intermin- gled, even in the same population,” providing “no specific value in this genus.” Johnston (1923) transferred varieties micrantha and lepida to the genus Cryptantha, as C. micrantha (Torrey) L M. Johnston [var. mi- crantha] and C. micrantha var. lepida (A. Gray) I. M. Johnston. Johnston (1925) corroborated Maebride’s observation about nutlet morphology in stating that “nutlets of C micrantha are exceptionally variable.” Despite this, Brand (1931) described Eremocarya abramsiana Brand, based on an obvious specimen of lepida having smooth nutlets. Subsequent to Johnston’s 1923 treatment, the classification of these two taxa as varieties of Cryptantha micrantha has been accepted by most botanists in almost all floras for regions where the two occur (Munz and Keck 1968; Munz 1974; Cronquist 1984; Kelley and Wilken 1993; Kelley et al. 2012). However, Mathew and Raven (1962) elevated Johnston’s varieties of Cryptantha mi- crantha to the rank of subspecies, as C micrantha subsp. micrantha and C m. subsp. lepida (A. Gray) K. Mathew & P. H. Raven, a change that has not been widely accepted. Based on a recent molecular phylogenetic study, Hasenstab-Lehman and Simpson (2012) resurrected the genus Eremocarya (and three other segregate genera of Cryptantha) because it comprised a clade within their subtribe Cryp~ tanthinae separate from Cryptantha s.s., a classi- fication we accept here. In fact, at least two relatively recent floristic treatments also recog- nized Eremocarya as separate from Cryptantha: Abrams 1951 (in his Illustrated Flora of the Pacific States, citing only E. micrantha) and Wiggins 1980 (in his Flora of Baja California, citing both varieties of E. micrantha). The two varieties of Eremocarya micrantha {Cryptantha m.] have been separated based primarily on corolla limb size and color and size of the five corolla throat “fornices” (singular “fornix,” also known as “appendages”), the latter constituting invaginations of the corolla tissue, infolded toward the central floral axis along a common radius with the corolla lobes and slightly protruding into the upper throat. The key to these taxa, treated as varieties of Cryptantha micrantha, from Kelley et al. (2012) reads: Corolla limb 1.5-M- mm diam., appendages larger than minute, yellow ........ var. lepida Corolla limb 0,5-1 .2 mm diam., appendages minute, ± white ................ var. micrantha Mathew and Raven (1962) found both taxa to have a common chromosome number of n == 12. Thus, they did not argue for their classification as separate species. These authors stated that the two taxa “have not been found growing togeth- er” and “appear to be largely geographic entities best recognized as subspecies.” Kelley et al. (2012) described the geographic and ecologic separation of the two entities. Variety lepida occurs in “mountain slopes, flats, valleys, granite-based gravelly soils, generally conifer forest, also chaparral, foothill woodland, Joshua-tree woodland, 300-2800 m,” flowering “March- August.” In the United States this taxon is almost entirely restricted to California, occur- ring in the southern Sierra Nevada, Tehachapi Mountains, Transverse Range, Peninsular Range, and the northern area and the region of the Great Basin east of the Sierra Nevada Mountains. Variety lepida also occurs in multiple 264 MADRONO [Vol. 61 Table 1. Principal Components Analysis Loadings for Characters Used in Two Analyses; Six Characters/Five Characters. Percent of total variance explained: axis 1 = 72.467/71.487, axis 2 = 18.159/ 19.261, and axis 3 = 5.298/5.04. Component loadings Character 1 2 3 Corolla limb width 0.91/0.897 0.132/0.084 0.336/-0.425 Nutlet length 0.879/0.874 0.265/0.338 -0.304/0.235 Nutlet maximum width 0.928/0.928 0.255/0.247 -0.154/0.07 Nutlet apical width 0.703476483 0.934/0.863 0.055/-0.051 Nutlet apical width : Nutlet maximum width 0.944/0.951 -0.251/-0.187 -0.131/0.09 Fornix body length 0.936/N.A. -0.034/N.A. 0.263/N.A. populations in Baja California, Mexico (Wiggins 1980; Baja Flora 2013). Variety micrantha occurs in “desert flats, washes, sandy to fine-gravelly soils, <1900 m” (Kelley et al. 2012). The range of var. micrantha overlaps with but is much more widespread than var. lepida, the former occurring in southeastern California, the Great Basin, and desert regions of Arizona, New Mexico, Nevada, southeastern Oregon, western Texas, Utah, and also in Baja California and Sonora, Mexico (Baja Flora 2013; Plant Resources Center 2013; SEINet 2013). The purpose of this article is to present evidence that these two taxonomic entities (referred to below as simply “micrantha” and “lepida”) should be treated at the rank of species. We present several morphological features (in- cluding one thought to be new to science) and cite more detailed biogeographic evidence for their classification as separate species. Materials and Methods Flerbarium specimens were obtained and stud- ied from the following herbaria: California Academy of Sciences (CAS), San Diego Natural Flistory Museum (SD), San Diego State Univer- sity (SDSU), and the University of California, Berkeley (JEPS, UC). A total of 352 herbarium specimens were sampled, annotated, and record- ed for latitude/longitude and elevation (or these estimated from label data). From a randomly chosen subset of 45 of these specimens (approx- imately half for each form), dried flowers of both taxa were boiled for 2-3 min and placed on a piece of clear, double-stick tape on a microscope slide. Corolla limb width of the boiled, re- expanded flower was measured and the corolla throat was slit and the two edges peeled back, followed by staining with a drop of 0.5% toluidine blue. The corolla throat fornix region was observed, and the presence, size (length and width), and number per fornix of peculiar “fornix bodies” (see Results section) were measured with a video-interfaced dissecting microscope, using ImageJ software (Rasband 1997-2007, see Abramoff et al. 2004). From the same 45 specimens, 3-4 mature fruits ' were detached and the nutlets removed and placed in dorsal (abaxial) view. The length, i maximum width (below the middle), and width at 1/4 relative distance from the apex were measured using ImageJ. Nutlet data were segre- gated based on fruit heteromorphism. If selected fruits contained heteromorphic nutlets, the single (“odd”) large nutlet was tabulated separately from the generally three smaller (“consimilar”) !| nutlets, the latter values averaged. If fruit nutlets ; were homomorphic, measurements of all four were averaged. All measured nutlet parameters were averaged per herbarium specimen. Bivariate plots were prepared for nutlet length I (mm) versus corolla limb width (mm) and for nutlet length (mm) versus the ratio of maximum i nutlet width to width at 1/4 relative distance from the apex (this width to width ratio an estimate of the degree of attenuation of the apical portion of the nutlet). In addition, bivariate plots were ! prepared for elevation (m) versus corolla limb width (mm) and for elevation (m) versus nutlet length (mm). A principal components analysis (PCA) was conducted on samples having complete data for six characters: 1) corolla limb width; 2) length of “fornix bodies” (see below; if bodies absent, a zero was assigned); 3) nutlet length; 4) nutlet maximum width; 5) nutlet width 1/4 relative distance from the apex; and 6) ratio of maximum nutlet width to width at 1/4 relative distance from the apex. A second PCA was conducted using these characters except for fornix body length; this was done to compare the effect of this novel feature on the distinctiveness of the two taxa. Variables were standardized by subtracting the total mean for a feature from each individual measurement, then dividing by the total standard deviation. This transformation results in all variables having a mean of zero and a standard deviation of 1 . The resulting factor scores of this PCA were plotted for the 1st versus 2nd components and 2nd versus 3rd components (only the former illustrated and discussed), and component loadings were tabulated (Table 1). All statistical analyses were performed in SYSTAT, 2014] SIMPSON ET AL.: RECOGNITION OF TWO SPECIES IN EREMOCARYA 265 Version 11 (Systat Software, Inc., San Jose, CA USA, http://www. systat.com). To visualize character distributions by taxon, box plots showing the median and the four quartiles of distribution were prepared for 1) corolla limb width (mm); 2) nutlet length (mm); 3) nutlet maximum width (mm); 4) nutlet apical width (1/4 from apex, mm); 5) the ratio of nutlet maximum width: width 1/4 from apex; and 6) elevation (m). Each of these features was evaluated for statistically significant differences by taxon with a t-test. Statistical differences between the two taxa for a particular character were tabulated and the variation in these features illustrated in box plot diagrams, using Systat. A map was prepared showing the distributions of all specimens examined and annotated to variety (indicated with an exclamation mark in Appendix 1) plus specimens not examined by us but identified to variety in databases (data from the CCH 2013, SEINet 2013, and BajaFlora 2013), a total of 554 specimen collections. Specimen records identified only to species (as Cryptantha micrantha, with no variety indicated) were not mapped, with the exception of 14 records: 11 records from northern Nevada and southeastern Oregon from the CPNH (2013) and three Texas records from the Plant Resources Center (2013). However, based on ranges cited in the literature, we feel confident that these represent what has generally been recognized as Cryptantha micrantha var. micrantha, what we are indicating as “micrantha.” Our total mapping records include specimens from ARIZ, ASC, BCMEX, CAS, CIC, DES, DH, HCIB, IRVC, JEPS, JOTR, LL, MWI, OSC, POM, RSA, SD, SDSU, TEX, UC, UCD, UCR, UNM, VVC, and WILLU (acronyms of herbaria after Holmgren and Holmgren 1998 onwards). In addition, maps were prepared of San Diego County from the San Diego County Plant Atlas (2013) and Baja Flora (2013) databases, but using only specimens verified by us; these maps show more detailed representations of plant community and vegeta- tion types. Results We observed that the corolla of “lepida” is generally rotate, i.e., with horizontal, orbicular lobes (Fig. 3A), whereas the corolla of “mi- crantha” tends to be more infundibular, with generally ascending, oblong lobes (Fig. ID). We confirmed, as originally described by Gray (1878), that the fornices of “lepida” are promi- nent and have a yellowish pigmentation (Figs. 2B, 3C). We also note that the corolla tube is also yellow, a feature probably missed in earlier descriptions given the tube is not normally visible because it is covered by the calyx. However, in “micrantha” we detected no evident fornices and no yellow pigmentation in the fornix region (Fig. IB, E), We also report the observation of anthers at different levels and the presence of anther apical appendages for both “micrantha” (Fig. IE) and “lepida” (Fig. 3E), which to our knowledge has not been previously described. We do not yet know if any of these androecial features are unique to these taxa within the Cryptanthinae; a detailed study of the corolla and androecium morphology of the complex will be the topic of another study. An interesting discovery is the presence of unusual and distinctive structures attached to the fornices of “lepida”, but absent in all “micrantha” specimens observed. These struc- tures, which we term “fornix bodies,” are ellipsoid, transparent (in fresh material), and stalked (Fig. 3D-F); they occur in groups of about three (ranging from 1-4, rarely 5), arising from the middle-lower portion of each of the five fornices of a corolla and positioned well above the anthers (Fig. 3E, F). The fornix bodies have a mean length of 0. 1 1 mm (not including the stalk) and an average width of 0.08 mm. Viewed from a face-view of the corolla throat opening and from corolla longitudinal sections, these bodies appear pendant, with a horizontal to reclined orientation (Fig. 3D-F). We point out that the fornix bodies of “lepida” are evident in live material under high magnification (even with a strong hand lens), but are more difficult to see in dried material. A bivariate plot of nutlet length versus corolla limb width shows morphological separation between the two taxa (Fig. 4A). A bivariate plot of nutlet length versus the ratio of nutlet maximum to width 1/4 from the apex also shows separation between the two taxa, but with more of a continuum (Fig. 4B). In either plot, no appreciable difference is noted between samples of “micrantha” having homomorphic versus heteromorphic nutlets. A bivariate plot of eleva- tion versus corolla limb width shows moderate separation between the two taxa (Fig. 4C), but one of elevation versus nutlet length shows more of a continuous grade (Fig. 4D). The PCA shows a discrete separation between “lepida” and “micrantha” utilizing six characters (Fig. 4E). The first principal component, explain- ing 72% of the overall variance, corresponds to size, with five characters (corolla limb width, fornix body length, nutlet length, nutlet maxi- mum width, and ratio of nutlet maximum width; nutlet apical width) loading heavily, at 0.879- 0.944 (Table 1). This separation between taxa persists in the PCA analysis that excludes fornix body length (Fig. 4F), with very similar variance and component loading values (Table 1). Based on our sampling of dried herbarium material, the corolla limb width of “lepida”, with a mean of 2.4 mm, is significantly larger (P < 0.01) than that of “micrantha”, mean = 1.0 mm 266 MADRONO [Vol. 61 nJ 1.1 S 1 « M A A-^ ^ A A A B, 13 A- A 01)' 1.2 s J 1.1 A Homo lepida ® 1 OHomo micrantha 3 ©Hetero L micrantha ^ 0-9 • Hetero S micrantha 0.8 Corolla Limb Width (mm) B :Fc o # ^ Aa^ A Homo lepida ©Homo micrantha •Hetero L micrantha • Hetero S micrantha Nutlet Width-maximumi Width-apical c« b 1500 O) B ^ V A ^ A A A A A A ^A . AA A A Corolla Limb Width (mm) OJ ■ B' g 1000 ‘■S a 3 D CD A A A aA^ A A^^ %^Vizona :y ■ r| (^g Beach California ■Pocolello,. 'f'' ‘ Fig. 6. Distribution map of Eremocarya lepida (white triangles) and E. micrantha (black circles), derived from georeference data of BajaFlora (2013), CCH (201 3), CPNH (2013) Plant Resources Center (2013), and SEINet (2013). A. Full range of taxa, showing much wider distribution of E. micrantha. B. Close-up of southern California. Note differences in elevation between the two forms, with E. lepida found in higher elevation, mountainous regions. Map data from ©Google 2013, INEGI Data. 270 MADRONO [VoL 61 I Chaparral Coastal Sage Scrub Mountains Lower Colorado Desert 100 km B Succulent Scrub 1 1 7° Vtf 29° N Centr.i! Desei I I i Fig. 7. Distribution map of Eremocarya lepida (white triangles) and E. micrantha (black circles). A. San Diego i County, derived from georeference data of the San Diego County Plant Atlas (2013). Note distribution of E. lepida | in chaparral, high-desert scrub, and coniferous woodland; distribution of E. micrantha is primarily in creosote bush j scrub. B. Baja California, Mexico, derived from verified SD specimens only. Note distribution of E. lepida in j chaparral and mountainous areas. Distribution of E. micrantha is limited to the Lower Colorado Desert, Central j Desert region, and a single collection in coastal sage scrub. j 2014] SIMPSON ET AL.: RECOGNITION OF TWO SPECIES IN EREMOCARYA 271 morphology (see Simpson and Hasenstab 2009; Hasenstab-Lehman and Simpson 2012; Kelley et al. 2012). Observation of Eremocarya micrantha specimens from Baja California, Mexico reveals some lower elevation populations with relatively large corollas, but lacking fornix bodies and having a nutlet morphology typical of this taxon, these identified as E. micrantha but not included in our quantitative analyses. A study of these unusual populations will be the subject of a future study. The two taxa also show significant differences in plant community and habitat occurrence, with Eremocarya lepida at higher elevation in high desert/desert transition, montane chaparral, or woodland habitats and E. micrantha occurring at lower elevations on the desert floor in desert scrub and arroyo vegetation. In fact, the two are significantly different in elevation, although with considerable overlap; however, we realize that an elevation parameter is a numerical-based corre- late for the nonparametric concept of plant communities. The distribution map of the two taxa generally corroborates these differences in habitat, with only a few exceptions. We know of no other flowering plant that has structures similar to the fornix bodies described here in Eremocarya lepida. However, they may have been missed previously because they are so small and because features of the corolla are often not described in detail among members of the Boraginaceae; fruit characters have tradition- ally been viewed as more important taxonomi- cally. From our personal observations, they are absent from the related and superficially similar Greeneocharis circumscissa, as well as from any observed members of Cryptantha s.s. or of any observed members of Amsinckia, Cryptantha s.s., Harpagoneiia, Johnstonella, Oreocarya, Pecto- carya, or Plagiobothrys of the subtribe Cryp- tanthinae (Simpson, work in progress). Interest- ingly, Cohen (2013) cited that “glands inside corolla” are present in most species of the Boraginaceae, including the Cynoglosseae to which Eremocarya belongs. However, these corolla glands are unlike those in Eremocarya lepida (Cohen, personal communication). The chemical makeup and function of the observed fornix bodies in E. lepida are unknown. They might somehow function as part of a pollination mechanism. For example, these fornix bodies might exclude certain visiting insects from entering the corolla tube subsequently conserving pollen or nectar for true pollinators; in fact, they would appear to partially block the proboscis of a visiting pollinator (Fig. 3D). Or, they may provide some essential nutrient or resource to a pollinator, making the flowers more “attractive” for visitation and increasing pollination success. However, the fornix bodies actually appear to be rather persistent, as they appear to remain attached in older flowers (Fig. 3F) or in dried herbarium material (even though appearing deflated). We plan to study both the chemical makeup and function of these unique structures. In addition, understanding the phylogeographic relationships of these two taxa will be the goal of a future study. Taxonomic Treatment Eremocarya Greene, Pittonia l(4):58-59. 1887. — TYPE Eremocarya micrantha (Torrey) Greene, Pittonia 1:59. 1887. Eremocarya micrantha (Torrey) Greene, Pittonia 1(4):59. 1887. Eritrichium micranthum Torrey, Report on the United States and Mexican Boundary [Emory] 2(1): 141. 1859. Krynitzkia micrantha (Torrey) A. Gray, Proceedings of the American Academy of Arts and Sciences 20:275. 1885. Cryptantha micrantha (Torrey) I. M. Johnston, Contributions from the Gray Herbarium of Harvard University 68:56. 1923. —TYPE: USA, Texas, El Paso, April 1851, G. Thurber 181 (holotype, designated by Cron- quist 1984: NY 335240). Eremocarya muricata Rydberg, Bulletin of the Torrey Botanical Club 36:677. 1909.— TYPE: USA, Utah, Southern Utah, 1874, Parry 164 (holotype EGO). Eremocarya lepida (A, Gray) Greene, Pittonia 1(4):59. 1887. Eritrichium micranthum Torrey var. lepidum A. Gray, Synoptical Flora of North America 2(1): 193. 1878. Krynitzkia micrantha (Torrey) A. Gray var. lepida (A. Gray) A. Gray, Proceedings of the American Academy of Arts and Sciences 20:275. 1885. Eremocarya micrantha (Torrey) Greene var. lepida (A. Gray) J. F. Macbride, Proceedings of the American Academy of Arts and Sciences 51(10):545. 1916. Cryptantha micrantha (Tor- rey) I. M. Johnston var. lepida (A. Gray) I. M. Johnston, Contributions from the Gray Her- barium of Harvard University 68:57. 1923. Cryptantha micrantha (Torrey) I. M. Johnston subsp. lepida (A. Gray) K. Mathew & P. H. Raven, Madrono 16(5): 170. 1962. Eremocarya abramsiana Brand, Pp. 77 in A. Engler (ed.), Das Pflanzenreich IV, 252 (Heft 97): Borraginaceae-Borraginoideae-Cryp- tantheae. Verlag von Wilhelm Engelmann, Leipzig. 1931. — TYPE: USA, California, San Bernardino County, near Pine Lake, Bear Valley, 5 August 1902, L. Abrams 2904 (holotype GH97011; isotypes DS8945, POM 158081, UC 153888, UC407303). Key The following revised key to the two species of Eremocarya (modified from Kelley and Simpson, in prep.) may be used to separate these taxa. 272 MADROto Corolla limb 0.5-1. 2 mm diam., center white, fornices absent, fornix region white, lacking ellipsoid bodies; nutlets ca. 1-1.1 mm long, apex narrowly acute; plants at matu- rity generally wider than tall ...... E. micrantha Corolla limb 1.5^ mm diam., center yellow, fornices conspicuous, yellow, each with a basal cluster of tiny (ca. 0.1 mm long), pendant, ellipsoid “fornix bodies;” nutlets L2_1.4 mm long, apex acuminate; plants at maturity generally taller than wide ... E. lepida Acknowledgments We thank the CAS, SD, SDSU, UC, and JEPS herbaria for loans of herbarium sheets and for allowing us to remove material for this study. We thank Phil Unitt for allowing us to use the map of San Diego County vegetation in Figure 7A (this based on those compiled by the U.S. Forest Service, Anza-Borrego Desert State Park, and the San Diego Association of Governments). We thank BajaFlora (2013) and Char- lotte Gonzalez- Abraham for use of the phytogeograph- ic map of Baja California in Figure 7B. We also thank the database resources provided by BajaFlora (includ- ing the Baja California Botanical Consortium, BCBC), California Consortium of Herbaria (CCH), Consortium of the Pacific Northwest Herbaria (CPNH), Southwest Environmental Information Network (SEINet), and University of Texas Plant Resources Center. Finally, we are very thankful to reviewers Jim Cohen and Matt Guiiliams, whose helpful comments have greatly improved this paper. Literature Cited AbrAmoff, M. D., P. J. Magelhaes, and S. J. Ram. 2004. Image processing with ImageJ. Biophotonics International 1:36-42. Abrams, L. 1951. Illustrated flora of the Pacific States: Washington, Oregon, and California. Vol. Ill Geraniaceae to Scrophulariaceae, Geraniums to Figworts. Stanford University Press, Stanford, CA. BajaFlora. 2013. The Flora of Baja California, San Diego Natural History Museum. Website http:// www.bajaflora.org (accessed on 8 Sep 2013). Brand, A. 1931. Eremocarya abramsiana. Pp. 77- in A. Engler (ed.). Das Pflanzenreich IV, 252 (Heft 97): Borraginaceae-Borraginoideae-Cryptantheae. Ver- lag von Wilhelm Engelmann, Leipzig. Consortium of California Herbaria (CCH). 2013. Website ucjeps.berkeley.edu/consortium (accessed on 8 Sep 2013). Cohen, J. I. 2013. A phylogenetic analysis of morpho- logical and molecular characters of Boraginaceae: evolutionary relationships, taxonomy, and patterns of character evolution. Cladistics 10.1111/cla. 12036:1-31. Consortium of Pacific Northwest Herbaria (CPNH). 2013. Website http://www.pnwherbaria. org (accessed on 8 Sep 2013). Cronquist, A. 1978. Once again, what is a species? Pp. 3-20 in J. A. Ramberger (ed.), Biosystematics in Agriculture, Allanheld & Osmun, Montclair, NJ. . 1984. Cryptantha micrantha. Pp. 266 in A, Cronquist, A. H. Holmgren, N. H. Holmgren, J. L. Reveal, and P. K. Holmgren, (eds.), Intermountain flora, vascular plants of the intermountain west, [Vol 61 j USA. Vol. 4. Subclass Asteridae (except Astera- [ ceae). New York Botanic Garden Press, Bronx, : NY. [ , 1988. The Evolution and classification of flowering plants, 2nd ed. New York Botanic Garden Press, New York, NY. [ Gray, A. 1878. Eritrichium micranthum. Pp. 193 in i Synoptical Flora of North America Vol. 2, part 1, 2nd ed. Smithsonian Institute, Washington, DC. . 1885. Contributions to the botany of North America: a revision of some Borragineous genera. Pp. 257-310 Proceedings of the American Acade- my of Arts and Sciences. University Press, Boston. Greene, E. L. 1887. Some west American Asperifoliae: IF Pittonia 1:55-60. Hasenstab-Lehman, K. E. and M. G. Simpson. 2012. Cat’s eyes and popcorn flowers: phylogenetic systematics of the genus Cryptantha s.l. (Boragina- ceae). Systematic Botany 37:738-757. Holmgren, P. K. and N. H. Holmgren. 1998 ! onwards. Index Herbariorum. New York Botanical ; Garden Virtual Herbarium, http://sciweb.nybg.org/ | science2/IndexHerbariorum.asp. Accessed on 8 Sep j 2013. ; Johnston, I. M. 1923. Studies in the Boraginaceae IT , Novelties and new combinations in the genus ! Cryptantha. Contributions from the Gray Herbar- j ium of Harvard University New Series 68:52-57 j {Cryptantha micrantha, pp. 56-57). . 1925. Studies in the Boraginaceae IV. The ! North American species of Cryptantha. Contribu- tions from the Gray Herbarium of Harvard j University 74:1-1 14. i Kelley, R. B. and M. G. Simpson. In prep. | Eremocarya. Pp. Flora of North America North of Mexico. Flora of North America Editorial i Committee. 16+ vols., New York and Oxford. , , and K. E. Hasenstab. 2012. Cryp- tantha. Pp. 455^68 in B, G. Baldwin, D. H. Goldman, D. J. Keil, R. Patterson, T. J. Rosatti, | and D. H. Wilken. (eds.). The Jepson Manual: j Vascular Plants of California, 2nd ed. University of j California Press, Berkeley, Los Angeles, London. I Kelley, W. A. and D. Wilken. 1993. Cryptantha. j Pp. 369-378 in J. C. Hickman (ed.), The Jepson j Manual: Higher Plants of California. University of , California Press, Berkeley. Macbride j. F. 1916. III. Certain Borraginaceae [sic], i new or transferred. Proceedings of the American * Academy of Arts and Sciences 51:541-548, Mathew, K. and P. H. Raven. 1962. A new species of Cryptantha (section Circumscissae) from California and two recombinations (section Circumscissae and j Angus t if oliae). Madrono 16:168-171. | Munz, P. a. 1974. A Flora of Southern California. University of California Press, Berkeley. and D. D. Keck. 1968. A California flora and supplement. University of California Press, Berke- ley, CA. Papageorgiou, V. P., A. N. Assimopoulou, E. A. Couladouros, D. Hepworth, and K. C. Nico- : LAOU. 1999. The chemistry and biology of alkan- nin, shikonin, and related naphthazarin natural products. Angewandte Chemie-International Edi- tion 38:270-300. ■ i Plant Resources Center. 2013. University of Texas 1 Herbarium Plant Resources Center. Austin, Texas. 2014] SIMPSON ET AL.: RECOGNITION OF TWO SPECIES IN EREMOCARYA 273 http://www.biosci.utexas.edu/prc. Accessed on Sep- tember 2, 2013. Rasband, W. S. 1997-2007. Image!. Bethesda, MA. http://rsb.info.mh.gov/ij. Rydberg, P. A. 1909. Studies on the Rocky Mountain flora. Bulletin of the Torrey Botanical Club 36:675-698. San Diego County Plant Atlas. 2013. http://www. sdplantatlas.org. Accessed on 5 Sep 2013. SEINet. 2013. Southwest Environmental Information Network. http//:swbiodiversity.org/seinet/index. php. Accessed on August 4, 2013. Simpson, M. G. and K. E. Hasenstab. 2009. Cryptantha of Southern California. Crossosoma 35:1-59. Torrey, J. 1859. Eritrichium micranthum. Report on the United States and Mexican Boundary Survey, under the order of Lieut. Col. W. H. Emory: Botany of the boundary 2:29-270 + 261 plates (Species treatment, p. 141). Wiggins, I. L. 1980. Flora of Baja California. Stanford University Press, Stanford, CA. Appendix 1 Voucher Specimens Used in This Study, Listed Alphanumerically by Collector AND Collection Number; Date Listed for Collections without a Collection Number (S.N.). Bold Font = Collectors/vouchers for which data were measured in quantitative study. ! = Voucher verified by the authors. Eremocarya lepidai Abrams 2089 (POM 49193); Abrams 2904 (UC 153888!, UC 407303!); Abrams 3594 (POM 156574); Alder son s.n. 1893 (UC 180103!); Almeda 6544 (CAS 831263!, SD 130518); Anderson 26 (SDSU 10277!); Andre 8145 (UCR 196107); Angel 597 (SD 207714); 9 Jul 1895, Anthony s.n. (UC 1601969!); Bacigalupi 3059 (UC 917090!); Balls 13489 (RSA 50324); Beauchamp 2288 (SD 85288); Beauchamp 2336 (SD 85342); Beauchamp 2742 (SD 85816); Beauchamp 2861 (SD 128344); Bell 3801 (RSA 798987); Bourell 2980 (CAS 741818!); Boyd 11382 (UCR 148355); Boyd 2302 (RSA 508006, UCR 120623); 10 May 1893, Brandegee s.n. (UC 78818!); May 1889, K. Brandegee s.n. (UC 79423!); 16 Jun 1894, T. S. Brandegee s.n. (UC 78822!); 20 Jun 1904, T. S. Brandegee s.n. (UC 174930!); 30 Apr 1894, T. S. Brandegee s.n. (UC 79425!); 9 Jun 1951, Brattstrom s.n. (SD 44489!); Breisch 270 (SD 194947!); Breisch 42 (SD 172723); Breisch 449 (SD 201471!); Carter 5705 (UC 1443877!); 29 Jul 1897, Chandler s.n. (UC 24835!); Charlton 1951 (RSA 483571); Charlton 2118 (RSA 489621); Chisaki 637 (UC 1035182!); 7 Apr 1929, Clark s.n. (RSA 499653); 7 Apr 2004, Clarke s.n. (UCR 239826); 9 Jun 1991, Clarke s.n. (UC 1931683!, UC 1931683); Clemons 1041 (SD 118370); Clemons 1289 (SD 118945!); Clemons 2230 (SD 126687); Clemons 584 (SD \\563\y, 1882, Cleveland s.n. (UC 193915!); 4 Jul 1884, Cleveland s.n. (UC 78823!); Clokey 6847 (UC 857183!); 21 Jun 1910, Condit s.n. (UC 455977!); Cooper 1249 (RSA 499654); Copp 14 (CAS 840487!, SD 134020!); Copp 14 (SD 134020); May 1933, Cota s.n. (SD 15615!); Cowan 1255 (RSA 790339); Cowan 1258 (RSA 790340); Cur to 285 (SD 121048); Davidson 2819 (RSA 499651); Davidson s.n. (RSA 499652); 5 May 1998, Delgadillo s.n. (SD 154820!); Dempster 4260 (JEPS 44382!); Dunkle 4056 (UCR 208013); Eastwood 2591 (UC 1601675!); Eastwood 3173 (UC 1601674!); Eastwood 9422 (CAS 26923!); Eastwood 9539 (CAS 26924!); Eggleston 19739 (POM 50736); Elvin 2451 (IRVC 28850, UCR 127092); Everett 23093 (UC 1080253!); Eosherg 10680 (UC 551773!); Eosberg 10701 (UC 551800!, UCR 47681); Fosberg 1101 (UCR 47710); Gander 173-44 (SD 11111); Gander 204- 28 (SD 11400); Gander 5123 (SD 20635, SDSU 5412!, UC 930876!); Gander 5603 (SD 21350); Gander 5627 (SD 21374); Gander 5672 (SD 21429); Gander 5680 (SD 21437); Gander 5881 (SD 21651); Gander 7218 (SD 24716); Gander 8687 (SD 27588); Gander 9109 (SD 28534); Gander 9356 (SD 28771); Gander s.n. 10 Jun 1933 (SD 4195!); 11 Jun 1933, Gander s.n. (SD 4194!); Griesel 718 (RSA 164774); Grimmell 282 (CAS 26926!); Gross 1121 (RSA 679981); Gross 3299 (RSA 735603); Gross 4040 (RSA 750186); Gross 4067 (RSA 749784); Gross 4477 (RSA 763620); Gross 4713 (RSA 769838); Gross 5511 (RSA 776148); Gross 944 (RSA 679944); Hall 1120 (UC 64426!); Hall 2051 (UC 24832!); Hall 2487 (UC 24831!); Hcdl 3002 (UC 56860!); Hall 5135 (UC 63333!); Hall 6540 (UC 100875!); Hall 7588 (UC 111522!); 24 Jun 1901, Hall s.n. (UC 24833!); May 1901, Hall s.n. (UC 24830!); 2 Mar 1987, Hawks s.n. (UCR 47194); Helmkamp 4096 (UCR 110966); Helmkamp 4113 (UCR 110994); Hendrickson 15 (SD 137396); Hendrickson 1953 (SD 196338!); Hendrickson 3043 (SD 205616); Hendrickson 3771 A (SD 214888); Henrickson 19925 (RSA 656795); Hirshberg 210 (SD 135173); Honer 3419 (RSA 761037); Honer 3594 (RSA 764252); Howe 547 (SD 113114); Howell s.n. 17 Jun 1971 (CAS 862225!); Jackson 79-2 (SD 114923); 18 Jun 1922, Jaeger s.n. (DH 140398!); Jepson 11796 (JEPS 67692!); Jepson 1295 (JEPS 67968!); Jepson 1308b (JEPS 67969!); Jepson 17119 (JEPS 67973!); Jepson 4888 (JEPS 67970!); Jepson 4896 (JEPS 67971!); Jepson 8751 (JEPS 67972!); Jones 3182 (UC 380868!, UC 881598!); Jones 7065 (POM 72314); 27 Apr 1882, Jones s.n. (POM 72315); Kamb 1155 (UC 1178517!); Kamb 1908 (UC 1051659!); KasapUgil 3228 (UC 1516007!); Keck 21 (POM 97229); Keck 84 (POM 97232); La Doux 2619 (JOTR 32503); La Doux 2813 (JOTR 33922); 18 Feb 1988, LaPres.n. (UCR 50418); 23 Jun 1975, Latting s.n. (UCR 157301); Levin 1674 (SD 119269!); Lint 2037 (RSA 463700); Macias 463 (SD 165072!); Marsden 106 (SD 163799); Marsden 400 B (SD 207715); Marsden 491 (SD 205615!); Marsden 578 (SD 205614); Marsden 80 (SD 163798); Meyer 160 (UC 489041!); Meyer 428 (JEPS 67966!); 8 Jun 2005, Misenhelter s.n. (UCR 202235); Moldenke 25474 (SD 82945); Moran 10923 (SD 53840!); Moran 11126 (SD 54693!); Moran 12717 (DC 1345885!); Moran 13869 (SD 64660!); Moran 13898 (SD 64661!); Moran 14381 (SD 79677!); Moran 14475 (SD 79678!); Moran 14906 (SD 72336!, UC 1361697!); Moran 15001 (SD 69225!); Moran 21288 (SD 86898!); Moran 22039 (SD 91906!); Moran 22063 (SD 91864!); Moran 22143 (SD 91487!); Moran 22958 (SD 95519!); Moran 23334 (SD 96974!); Moran 24086 (SD 97110!); Moran 27342 (SD 103646!); Moran 27427 (SD 103440!); Moran 27448 (SD 103457!); Moran 28894 (SD 105525!); Moran 30665 (SD 111068!); Moran 3430 (UC 1112358!. UC 1112775!, SD 48056!); Moran 8112 (SD 60715!, UC 1199949!, UC 1298136!); Munz 10646 (POM 96716); Munz 5421 (POM 13001, UC 218237!); Munz 5475 (POM 13197, UC 1601673!); Munz 5722 (UC 1601676!, UC 218205!); 274 MADRONO [Vol. 61 Mimz 5839 (POM 12804, UC 218109!); Munz 8843 (POM 48717); 23 Apr 2003, Myers s.n. (UCR 177832); 7 Apr 2003, Myers s.n. (UCR 162407); Nenow 1103 (SD 2210381); Nenow 1301 (SD 220259); Olmsted 3503 (RSA 171377); Otis 7 (SD 201473!); Parish 3245 (UC 78819!); Parish 6943 (UC 1666991); 1 May 1888, Parish s.n. (UC 248281); May 1887, Parish s.n. (JEPS 679601); Peirson 370 (JEPS 679671); Peirson 4483 (RSA 79959); Pigniolo 1053 (SD 158770); Pitzer 662 (RSA 511347); Powell 63 (UC 11818471); Provance 1693 (UCR 111545); Purer 6443 (SD 39201); Purer 6651 (SD 12594, SD 39200); Purpus s.n. May-Oct 1898 (UC 248291); Raven 11122 (UC 12799661); Raven 12580 (UC 11124431); Raven 14241 (UC 12411721); 18 May 1959, Raven s.n. (RSA 145650); Ray K-33 (UC 11784011); Rebman 11407 (SD 171650, UCR 157576); Rebman 12152 (SD 174759); Rebman 17601 (SD 197223!); Rebman 17646 (SD 197222); Rebman 19601 (SD 210102); Rebman 21146 (SD 2130301); Rebman 23390 (SD 221601); Rebman 23610 (SD 223361); Rebman 23658 (SD 225483); Rebman 3902 (SD 144398); Rebman 7210 (SD 1558241); Rebman 7228 (SD 155825, UC 17877491, UCR 155754); Rebman 9019 (SD 160678, UCR 155757); Jun 1897, Reinhardt s.n. (UC 248361); Roos 3648 (UCR 25966); Roos 3705 (UCR 23585); Roos 6194 (RSA 659687); 18 Jul 1965, Roos s.n. (RSA 662062); 26 Jun 1965, Roos s.n. (CAS 9056661, RSA 450247); Ross 2598 (RSA 524630, UC 15842241, UCR 162554); Sanders 16811 (UCR 86673); Sanders 16841 (UCR 86811); Sanders 204 (UCR 18398); Sanders 26195 (IRVC 29584, RSA 712187, UCR 126911); Sanders 31384 (UCR 163844); Sanders 35608 (UCR 193489); Sanders 35640 (UCR 193421); Shevock 1074 (CAS 7133891); Simpson 2369 (SD 180702!, SDSU 17281!); Simpson 2369 (SDSU 17281); Simpson 2816 (SDSU 17572!); Simpson 3109 (SD 208180, SDSU 18628!); Simpson 3184 (SDSU 19533!); Simpson 3320 (SDSU 19612!); Simpson 6VI91AB (SDSU 5418!); Simpson 6VI91AC (SDSU 5388!); Simpson 8VI94J (SDSU 5431); Soza 1695 (UC 19296251); Spencer 1301 (POM 46993); Spencer 1347 (POM 46994); Spencer 1864 (POM 12212); 30 Apr 1918, Spencer s.n. (UC 4729281); Sullivan 453 (SD 201472); Sweet 266 (SD 178709!); Thomas 2953 (VVC 2007); Thomas 4533 (VVC 2728); Thorne 38080 (RSA 633317); Thorne 49838 (RSA 281171); Thorne 55789 (SD 1241271); Thorne 55937 (SD 1241281); Thorne 60451 (SD 1249831); Thorne 61669 (SD 1249811); Thorne 63025 (RSA 792627); Twisselmann 8587 (UCR 79440); van der Werff 3743 (SD 106838); van der Werff 3824 (SD 106780); van der Werff 3825 (SD 106773); Vestal s.n. 21 May 1963 (MW 100013450, SEINET 3449082); Vinton s.n. 15 Mar 2003 (SD 1826651); Vinton s.n. 16 May 2003 (SD 1826661); Wallace 4028 (RSA 788859); Weatherby 665 (RSA 125536); Wedherg 909 (SDSU 5046!); White 11980 (RSA 741225); White 12204 (RSA 741175); White 12501 (RSA 737129); White 13209 (RSA 752009); White 13309 (RSA 751051); White 1833 (UCR 83423); White 2046 (UCR 100037); White 2129 (UCR 157678); White 5 1 80 {VCK 139929); White 5375 (UCR 144099); White 7379 (RSA 653544); White 7489 (RSA 653505); Wiggins 10043 (UC 7187531); Wiggins 11809 (SD 473011, UC 10605361); Wiggins 13053 (UC 10092981); Wiggins 15799 (UC 13032711); Wiggins 1843 (UC 10195861, UC 10196091); Wiggins 1993 (UC 10196111); Wiggins 2250 (UC 10195881, UC 10196121); Wiggins 2350 (UC 10196131); Wiggins 5289 (UC 8571811); Wiggins 5311 (UC 6608091); Wiggins 7813 (UC 651 1 141); Wiggins 9808 (UC 7188731); Wilder 4319 i (POM 318740); Williams- Anderson 26 (SDSU 10277); Wogium 3055 (RSA 608561); Wolfinger 133 (SD 2176121); Wood 1739 (RSA 764200); Woodcock 35 \ (UC 4863671); Yates 6623 (UC 5786631, UC 10638211). ' Eremocarya micranthai Anonymous 2625 (SDSU 5400!); Atwood 17557 (UC 17194411); Axelrod 307 (UC 10638241); Bacigalupi 6236 (JEPS 226141); Barth 1298 (SD 226395); Barth 1477 (SD 226396); Barth 407 ^ (SD 169356!); Beal 739 (JEPS 183951); Beatley 3715 \ (DH 5953771); Beauchamp 1799 (SD 85465); Beauchamp ! 1990 (SD 141844); Beauchamp 2192 (SD 854161); Bell i 657 (RSA 779621); 27 April 1950, Benioff s.n. (UC i 10845211); Boyd 1429 (UC 15630961); Brainerd 1090 ' (CIC36366); 18 Apr 1889, T. S. Brandegee s.n. (UC 793061); 14 Apr 1891, C 5'. Brandegee s.n. (UC 794261); !' 18 Apr 1895, T. S. Brandegee s.n. (UC 794241); i Palmetto Spring [Palm Springs], T. S. Brandegee s.n. (UC 794271); Bright 12847 (UC 5985401); Burch 1IV95C (SDSU 14056); Charlton 1547 (UC 15521691); Clemons \ 1634 (SD 120961 1); Clemons 1951 (SD 122433); Clemons ] 2020 (SD 1226121); 20 May 1880, Cleveland s.n. (UC 788211); Clokey 4709 (UC 8571821); Clokey 5821 (SD ; 34115, UC 8571801); Clokey 5926 (CAS 3809041, UC 9004631); Clokey 8729 (UC 9004621); Clokey 8730 (UC 9004601); Clokey 8731 (CAS 3809031, UC 9004611); Cole | 905 (UC 13299361); 21 Apr 2008, Cooper s.n. (UCR 193452); Correll 38534 (LL 00034693, UC 13682201); Cowan 2266 (SD 1270801); Crampton 2599 (UC ; 12781051); Cronquist 10103 (UC 13296201); Cronquist ' 10192 (DH 5928321); Crum 1863 (UC 6388051); Cusick . 2020 (ORE76325); Davy 2287 (UC 611351); De Groot 6587 (SD 2190001); 14 Mar 1995, Delgadillo s.n. (SD ; 1650731); Dole 13 (UC 15787951); Duran 2743 (UC i 12975851); 17 Apr 1931, Eastwood s.n. (CAS 1902801); Elmer 3682 (POM 49420); Elvin 4661 (UCR 175753); Elvin 514 (IRVC 29314, UCR 175753); Ertter 6891 (UC 15625251); Felger 17377 (SD 96180!); Felger 96-164 (ARIZ 371716); Eerguson 4579 (UC 15543381); Eerris \ 9523 (UC 6048971); Fosberg 10625 (UC 5518121); Fosberg 10651 (UC 5518591); Fosberg 3055 (UC ‘ 7050611); Fraga 347 (UC 19274051, UCR 200056); Gallup 193 (SDSU 5425!); Gander 134 (SD 10507!); Gander 21 (SD 1 1172); Gander 7087 (SD 24511); Gander 7182 (SD 24606); Gander 9 (SD 10507); Gentry 8897 (SD 864101); Goodding 128-52 (UC 10253141); Goodding 2144 (UC 1336541); Goodding 2203 (UC 1336531); Gould 1594 (UC 8571761); Gould 1657 (UC 8571771); Gowen 551 (JEPS 116719!); Grant 445a-6796 (UC 166838!); Grant 445a-6796 (UC 1749311); 15 Mar 2006, Green s.n. (RSA 725855, UCR 170399); Gregory 1333 (SD 172724); Gregory 667 (SD 158771!); Gross 5 14 A (RSA 757924); Guilliams 566 (SDSU 18702); Guilliams 602 (SDSU 18956!); Hall 5938 (UC 1008661); Halse 5950 (UC 17791371); Harbison s.n. 4 May 1939 (SD 25137!); 5 May 1939, Harbison s.n. (SD 251381); Harrison 7719 (CAS 1934071); Heckard 4553 (JEPS 764971); Hendrick- son 2640 (SD 203297!); Hendrickson 2715 (SD 203298); Hendrickson 2784 (SD 205617!); Hendrickson 3530 (SD 214185); Hendrickson 408 (SD 172725); Hendrickson 4588 (SD 210829!); Hendrickson 4695 (SD 210828); Hendrickson 494 (SD 172721); Hendrickson 514B (SD 172722); Hill 84 (ASC 77993); Hitchcock 24241 (UC 12871071); Holmgren 7008 (UNM 54121); Holmgren 8229 (UC 10186201); Hoover 3136 (UC 7624621); Howe 2625 (SD 50962); Howe 2903 (SDSU 5419!); Howe 453 (SD 113115); Hughes 208 (UCD 84936); Jepson 15465 (JEPS 679491); Jepson 17103 (JEPS 679481); Jepson 2014] SIMPSON ET AL.: RECOGNITION OF TWO SPECIES IN EREMOCARYA 275 17120 (JEPS 67947!); Jepson 17180f (JEPS 67946!); Jepson 18163 (JEPS 67945!); Jepson 19504 (JEPS 67944!); Jepson 19550 (JEPS 67943!); Jepson 20295 (JEPS 67964!); Jepson 20406 (JEPS 67942!); Jepson 20539 (JEPS 67965!); Jepson 5428 (JEPS 67956!); Jepson 5954 (JEPS 67955!); Jepson 7147 (JEPS 67954!); Jepson 8603 (JEPS 67953!); Jepson 8733 (JEPS 67952!); Jepson 8831 (JEPS 67951!); Jepson 8952 (JEPS 67950!); Johnston 27 R (UC 306233!); Jones 5023 (UC 133627!); 15 May 1920, Jones s.n. (UC 407942!); 2 Apr 1921, Jones s.n. (UC 407941!); 21 May 1884, Jones s.n. (UC 380863!); Jonsson 1634 (SD 120961!); Jomson 2020 (SD 122612!); Junak 1653 (SD 125584, UCR 58705); Keck 4141 (UC 604178!); Kennedy 1832 (UC 174706!); 15 Apr 1935, Krames s.n. (JEPS 67957!); 24 Feb 1935, Krames s.n. (JEPS 67958!); La Doux 1034 (JOTR 28753); La Doux 3052 (JOTR 33840); Apr 1881, Lemmon s.n. (UC 907598!); Apr 1886, Lemmon s.n. (UC 907601!); Maguire 10423 (RM 147998, UC 553403!); Maguire 16263 (UC 604223!); Maguire 4959 (UC 533114!); Maguire 4961 (UC 5331131); Maguire 4963 (UC 533181!); 18 Jun 2008, Mancuso s.n. (NYl 109936); Mansfield 533 (CIC33195); Mansfield 7084 (CIC34501); Marisa Sripracha 514 (UCR 134590); Marsden 538 (SD 205618!); Marsden 7 (SD 159786); Mason 2555 (UC 1393323!); Mason 6867 (UC 573134!); McVaugh 8193 (TEX 00034694); Mistretta 4629 (RSA 771752); Moran 10317.5 (SD 54531!); Moran 10859 (SD 53783!); Moran 12311 (SD 65317!); Moran 12496 (SD 65316!); Moran 12717 (SD 65315!); Moran 19624 (SD 92326!); Moran 20886 (SD 88930!); Moran 22958 (SD 95519!); Moran 30352 (SD 1 10836!); Moran 30772 (SD 1 1 1259!); Moran 6529 (SD 47530); Moran 8112 (SD 60715!); Morefield 3262 (UC 1534942!); Morefield 3264 (UC 1534944!); Morefield 3304 (UC 1535144!); Morefield 3596 (UC 1545605!); Morgan K83 (SDSU 5421!); Munz 15644 (UC 1022816!); Nelson 1284 (UC 595642!); Nenow 149 (SD 174381); Nielsen 2004011 (CIC32531); Otting 1589 (CIC36340); Palmer 371 (UC 79304!, UC 79305!); Parish 2815 (POM 3709); Parish 6943 (UCR 208136); Parish 8464 (JEPS 67961!); Parish s.n. May 1887 (UC 193914!); Parish s.n. May 1895 (JEPS 67959!); Peck 25598 (OSC79575, UC 801 189!, WILLU27441); Peebles 6975 (CAS 252113!); Peirson s.n. 16 Eeb 1920 (RSA 79960); Pitzer 2041 (UCR 106267); Purer 4943 (SD 39196!); 9 Apr 2008, Rado s.n. (UCR 196042); Raven 11621 (UC 1114560!); Rebman 11304 (SD 168432); Rebman 11358 (SD 168431); Rebman 1484 (SD 137251!); Rebman 21541 A (SD 213031!); Rickard 1853 (DH 562256!); Robbins 3306h (UC 981781!); Rose 37152 (UC 857178!); Rose 40319 (UC 857179!); 10 Mar 1940, Rose s.n. (CAS 275467!); Ross 4964 (UC 1871251!); Rue 91-14 (SD 133159); Salvato 3221 (UCR 203176); Salywon 1050 (SD 188428!); Sanders 23840 (UCR 116613); Sanders 23845 (UCR 116628); Sanders 23943 (UCR 116813); Sanders 24021 (UCR 116727); Sanders 34502 (UCR 192610); Sanders 34881 (UCR 194156); Sanders 36698 (UCR 210808); Sanders 37504 (UCR 214414); Sanders 38262 (UCR 215039); Sanders 7702 (UCD 119901); 13 Mar 1982, Scheidlinger s.m (SDSU 18155!); Schreiber 990 (UC 608476!); Schweich 762 (UC 1980484!); Shreve 10089 (UC 664989!); Simpson 3126 (SDSU 19604!); Simpson 3670 (SDSU 20043!); Simpson 5IV97C (SDSU 12434!); Smith 149 (JEPS 67963!); Smith 64 (JEPS 67962!); Spencer 1918 (UC 857184!); Spencer 514 (POM 47458); Sripracha 6 (UCR 134590); 12 Mar 1927, Stason s.n. (UC 573124!); Stoughton 970 (RSA 777445); Strother 1261 (UC 1434007!); Sweet 509 (SD 200747!); Sweet 514 (SD 200747); Swinney 3449 (UCR 180969); Swinney 3652 (RSA 719189, UCR 180763); Tavares 478d (UC 1250315!); Taylor 16546 (UC 1731360!); Thorne 60111 (SD 124982!); Thorne 62954 (RSA 760357); Thoruber 443 (UC 128172!); Tiehm 11059 (OSC168584); Tiehm 16118 (ID162176- bcl41 194); Tourney 93 (UC 78825!); Tourney s.n. 15 Apr 1894 (UC 24837!, UC 78824!); 3 Apr 1894, Tourney s.n. (UC 220606!); True Jr. 304 (UC 1537677!); Turner 21- 237 (TEX 00300342); Turner 68-94 (SD 78609!); Vasek 600320-01 (UCR 1373B); Vinton s.n. 18 Mar 2003 (SD 182664!); Webster 18216 (SD 95971!); Wiggins 11544 (UC 754314!); Wiggins 2013 (SD 48710, UC 1019590!); Wiggins 262 (SD 94536!); Wiggins 9566 (UC 665774!); Wilder 10-116 (SD 218101!); Wilder 10-248 (SD 218100!); Wilson 98 (UC 61134!); Wilson s.n. May 1893 (UC 61136!); Witham 706 (SD 80848); Wojtan 4IV92C (SDSU 5394!); Wolf3202 (RSA 4003); Wooten s.n. 19 Apr 1905 (DH 137011!, SD 67704!, UC 112658!, UC 480465!); Wotjan 4IV92C (SDSU 5394); Wright 7772 (UC 24834!); Yates 5 394 (UC 573635!); Yates 6400 (UC 573262!, UC 1063825!). Madrono, Vol. 61, No. 3, pp. 276-289, 2014 THE ROLE OF SOIL CHEMISTRY IN THE GEOGRAPHIC DISTRIBUTION OF CEANOTHUS OTA YENSIS (RHAMNACEAE) Dylan O. Burge California Academy of Sciences, 55 Music Concourse Drive, San Francisco, California, 95118 dy Ian . o . burge@gmail . com Abstract Ceanothus otayensis McMinn (Rhamnaceae) was previously known only from metavolcanic- derived soils of the northern Peninsular Ranges — predominantly the San Ysidro Mountains — in San Diego County, California, and adjacent Baja California, Mexico. Recently, a new population of C otayensis was discovered on sedimentary soils at Marine Corps Air Station Miramar, 25 km northwest of the next nearest known population. Sedimentary deposits at the new locality are thought to produce unusual soils. It is possible that the disjunct distribution of C. otayensis is a response to soil conditions, a phenomenon frequently seen in other members of Ceanothus, for instance on serpentine. The present study uses soil chemistry data for seven populations and subpopulations of C. otayensis (metavolcanic: n = 5; sedimentary: n = 2), as well as 22 populations of closely related Ceanothus, to determine whether soils of C. otayensis are chemically distinct from those of closely related Ceanothus, and answer the following question: are sedimentary-derived soils at the new locality chemically similar to metavolcanic-derived soils that support all other known populations of the species? Soils of C otayensis proved to be chemically distinct from soils of closely related Ceanothus, with significantly lower levels of nitrate, sulfur, and conductivity. Sedimentary and metavolcanic soils of C otayensis proved to be chemically indistinguishable from one another (P < 0.05), with low levels of all assayed nutrients other than Ca, suggesting a chemical similarity among the soils of C. otayensis that may help explain its disjunct distribution. Population size estimates indicate that the new disjunct locality at Marine Corps Air Station Miramar supports about 75 adult individuals. Key Words: California, ecology, edaphic, metavolcanic, Mexico, Miramar, Otay Mountain. Plant-soil interactions are a primary driver of plant distribution, as well as a potentially potent force in plant evolution (Stebbins 1942; Krucke- berg 2002; Kay et al. 2011). Unusual soils — such as serpentine — have long been known to support unusual plant communities, including many species that are endemic to such soils, and likely specialized via local adaptation (Gankin and Major 1964; Kruckeberg 1986; O’Dell and Rajakaruna 2011). Although several such “edaphic endemic” taxa have been examined using modern soil chemistry and experimental tools (Baldwin 2005; Sambatti and Rice 2006; Burge et al. 2013), little is presently known about how soil conditions influence adaptation and geographic distribution in such plants, particu- larly those that are specialized to soils other than serpentine. By examining specific cases of edaphic endemism using soil chemistry data, it may be possible to discern general trends in the evolution and maintenance of soil associations in edaphi- cally specialized species or plant communities. The present contribution focuses on Ceanothus otayensis McMinn, a member of Ceanothus subgenus Cerastes S. Wats. Ceanothus otayensis is a low-stature, ascending shrub that is morpho- logically similar to Ceanothus perplexans Trel. and Ceanothus crassifoUus Torr. (Boyd and Keeley 2002), and was once thought to have arisen via hybridization between these two species (McMinn 1942). However, neither of the putative parent species is present within the geographic range of C otayensis, and genetic evidence does not support a hybrid origin (Burge et al. 2011). Ceanothus otayensis was previously thought to occur only on metavolcanic-derived soils of the Peninsular Ranges in southern San Diego Coun- ty, California, and adjacent Baja California, Mexico (McMinn 1942; Wilken 2006). Ceanothus otayensis, along with a small group of species endemic to the San Ysidro Mountains, has been treated as a specialist on metavolcanic-derived soils. Such soils are thin, extremely rocky, and very fast-draining (Bowman 1973). However, very little research has focused on the chemical and physical properties of these soils, or their potential influence on plant life. During a visit to the San Diego Natural History Museum, the author came upon a specimen of C. otayensis collected at Marine Corps Air Station Miramar (hereafter MCAS Miramar; Roberts & Dossey 6209; Appendix 1), 25 km to the northwest of the next nearest known population (Figs. 1 and 2). In 2009, the author visited the locality and was able to locate a small population of these plants (estimated at the time 2014] BURGE: CEANOTHUS OTAYENSIS SOILS 111 Groups ^ S San Ysidro Mtns 48° N Legend O C. otayensis localities — 100 km (background) if San Miguel Mountain + Otay Mountain 120° W / MCAS Miramar Group 2 Ik N San Ysidro Mtns. ^ ^ ^ ^ San Diego United — Mexico "'7*^ Fig. 1 . Global distribution of C. otayensis. Locality data for C. otayensis taken primarily from the Consortium of California Herbaria (http://ucjeps.berkeley.edu/consortium); additional data from Jon Rebman (personal communication); only unique localities mapped (Appendix 2). Inset map of topography is from Jarvis et al. (2008). 278 MADRONO [VoL 61 Fig. 2. Sampling and soil map. Soil sampling locations indicated by open circles (Table 2). Inset map shows C otayensis sampling in San Diego County; polygons for soils adapted from GIS layers in Soil Survey Geographic (SSURGO) database for San Diego County (USDA 2007). 2014] BURGE: CEANOTHUS OTAYENSIS SOILS 279 to contain less than 100 individuals) on sedimen- tary slopes of upper San Clemente Canyon on Kearney Mesa {Burge 1179; Appendix 1). The presence of C. otayensis on low-elevation (130 m) sedimentary-derived soils at MCAS Miramar suggests that the plant may not be a soil specialist, as previously assumed. The present study uses chemical analysis of soils to answer two outstanding questions on the soil chemistry associations and distribution of C. otayensis: 1) Are the soils occupied by C otayensis chemically distinct from those occupied by other members of Ceanothus subgenus Ceras- tes in California? and 2) Are the sedimentary- derived soils at the new MCAS Miramar locality chemically similar to the metavolcanic-derived soils that support other known populations of C otayensisl Population size estimates for C otayensis at MCAS Miramar are also presented, and conservation implications for this rare species are discussed. Materials and Methods Determining Geological Formations and Soil Types To determine the general soil and geology associations of C otayensis, a GIS approach was applied to available georeferenced specimens of the species. Occurrence data for sites other than those reported by DOB (Appendix 1) were obtained from the Consortium of California Herbaria (CCH 2013). Records were selected only if they explicitly provided latitude and longitude data on the label. Duplicates were removed using location; records with the same location — based on latitude and longitude round- ed to three decimal places — were removed, resulting in a list of high-quality localities (Appendix 2). Geological formations and soil types at individual sites were determined using a GIS approach, combined with reference to the geological and soil literature. Latitude and longitude data were used in the program DIVA-GIS V 7.5 (Hijmans et al. 2001) to infer general geology from the digital version of the 1:750,000 Geologic Map of California (Jennings et al. 1977; Saucedo et al. 2000). Local geology was inferred by visual examination of higher resolution physical or digital maps, including the Geological Maps of the Otay Mesa, La Mesa, and Jamul 7.5' Quadrangles (Kennedy and Tan 1977). Local soils were inferred using data from the Soil Survey Geographic (SSURGO) database for San Diego County, California, obtained from the United States Department of Agriculture Natural Resource Conservation Ser- vice (USDA 2007). Soil Sampling To quantify the soil chemistry associations of C. otayensis and closely related members of Ceanothus subg. Cerastes, soils were collected from 29 plant populations in California (Table 2, Fig. 1). The aim of this sampling was to obtain soils from as many C otayensis populations as possible, and from a representative diversity of soils occupied by other members of Ceanothus subg. Cerastes. For the species other than C otayensis, an effort was made to maximize the diversity of substrates represented in the dataset, and to sample from as many purported edaphic- endemic taxa as possible. (Wilken 2006). Seven of the resulting samples were from populations or subpopulations of C otayensis (5 metavolcanic from the San Ysidro Mountains; 2 sedimentary, from MCAS Miramar), and 22 were from species other than C otayensis, representing 13 of the 23 currently recognized members of Ceanothus subg. Cerastes (Wilken 2006) and all but one of the approximately eight edaphic-endemic members of this group (Table 2, Figs. 1 and 2). Sampling of soil was carried out in April and May 2009, and again in April 2013. The soil sampling procedure follows that of Burge and Manos (2011). Soil Chemistry Assays Soil chemistry analyses were done by the Texas A&M University Soil, Water, and Forage Testing Laboratory. Methods were as described by Burge and Manos (2011). In all, 13 soil properties were assayed (Table 3), including pH, nitrate (NOs"), electrical conductivity, major nutrients (P, K, Ca, Mg, and S), micronutrients (Cu, Fe, Mn, and Zn), and sodium (Na). Analysis of Soil Chemistry Data Soil chemistry data were treated in a multivar- iate statistical framework. Differences between sedimentary (MCAS Miramar; n = 2) and metavolcanic (San Ysidro Mountains; n = 5) soils of C. otayensis, as well as differences between the soils of C. otayensis and other species of Ceanothus subg. Cerastes from California (n = 22), were summarized and tested using principal component analysis (PCA) and generalized ca- nonical discriminant analysis (Gittins 1985). Principal components analysis was done in R version 2.10.1 (R Development Core Team 2013), using the “ecodist” package of Goslee and Urban (2007). Soil chemistry variables were transformed into Z~scores before analysis. Following analysis, the first two principal components were visualized in bivariate space and the contribution of different soil chemistry variables to the components was assessed using vector loading. Generalized 280 MADRONO [VoL 61 Fig. 3. Distribution of C otayensis at MCAS Miramar. Map shows detail of upper (eastern) San Clemente j Canyon, with core population of C. otayensis and outlying individuals. Background image from USGS (2012). * canonical discriminant analysis was done in R (R Development Core Team 2013), using the “can- disc” package of Friendly and Fox (2009). Wilks’ Lambda was used to test for significant differences between the two groups based on the 13 soil chemistry variables (Friendly 2007). Estimation of Population Size at MCAS Miramar A transect approach was used to estimate the size of the C otayensis population at its disjunct locality on MCAS Miramar. With the exception of a few outlying shrubs, C otayensis at MCAS Miramar is restricted to a patch of extremely dense chaparral at the head of a tributary of San Clemente Canyon (Fig. 3). Because this patch is surrounded by open scrub, it was possible to walk | the perimeter of the patch and count individual i adult plants. Two observers (DOB and K. j Zhukovsky) walked the margin of the chaparral i patch (Fig. 3), pausing approximately every 5 m to | tally the number of new individuals observable in a line across the canyon. In addition, the surround- ing portions of Kearny Mesa and upper San Clemente Canyon were explored on foot (Fig. 3), and additional outlying individuals of C otayensis were noted. The position of isolated individuals ^ was recorded using a hand-held GPS and the ■ WGS84 datum. All work for the population size ' estimate was carried out on 18 April 2013. ! 2014] BURGE: CEANOTHUS OTAYENSIS SOILS 281 Results Geological Formations and Soil Types Using the CCH database and collections ob- tained specifically for this project, 19 unique C. otayensis localities were identified (CCH 2013, Appendix 2). Fourteen of these are in the Peninsular Ranges of southern San Diego Coun- ty, California and northern Baja California, Mexico (San Ysidro Mountains and San Miguel Mountain; Fig. 1), and two are at the recently discovered locality at MCAS Miramar in northern San Diego County. Inferred geological and soil formations for these sites (Appendix 2) indicate that three major groupings of sites are present (Table 1): 1) localities at MCAS Miramar, 2) localities on San Miguel Mountain and the California portion of the San Ysidro Mountains, and 3) one locality in northern Baja California, Mexico, in the southern- most San Ysidro Mountains (Fig. 1). At the first set of localities, rocks are classified as sedimentary; soils are classified as Terrace escarpments and Redding gravelly loam (Appendix 2; Table 1). At the second set of localities, by contrast, the rocks are classified as volcanic or metavolcanic; soils are classified as San Miguel-Exchequer rocky silt loams and meta- morphic rock land (Table 1). The final locality, from Baja California, Mexico, is apparently on soils derived from andesite, a volcanic rock type (Appendix 1; Table 1). However, this rock is likely metavolcanic, based on the geological setting and field observations by the author (2009). Soil Chemistry In comparison to the soils of other Ceanothus species assayed, the soils of C. otayensis have, on average, lower pH, lower electrical conductivity, lower concentrations of nitrate, and lower levels of every assayed nutrient other than potassium, sodium, zinc, and manganese (Table 3). These differences are significant in the case of conduc- tivity, nitrate, sulfur, and sodium (Student’s paired t-tests, P < 0.03). In comparison to C otayensis from metavolcanic soils of the San Ysidro Mountains (n = 5), the soils of C. otayensis from MCAS Miramar (n = 2) have nearly identical levels for all of the assayed chemical properties (Table 3); there were no significant differences for any of these properties (Student’s paired t-tests, P > 0.05). Principal components analysis and canonical discriminant analysis provide a summary of the results for the 13 soil chemistry variables (Fig. 4). In the case of PCA, the first two principal components account for 50% of variance, with 30% on the first principal component and 20% on the second. The first principal component is strongly positively correlated with conductivity (vector loading = 0.45) and iron (vector loading = 0.44), and strongly negatively correlated with phosphorous (vector loading == —0.13). In canonical discriminant analysis, three groups (1, C. otayensis MCAS Miramar; 2, C. otayensis San Ysidro Mountains; and 3, other species; Table 2) were used to transform soil chemistry variables into canonical space (Fig. 5). A single coordinate axis accounted for 100% of the variance, and the Wilks’ Lambda test (approximate F = 4.06) allowed for rejection of the null hypothesis of no difference between the means for the three groups (P = 0.005). Exami- nation of the canonical discriminant plot (Fig. 5) indicates that this pattern is driven by the difference in chemistry between soils of C. otayensis and the other species. To test this, a two-tailed t-test was carried out using vector loadings from the first and second principal components of the PCA (Fig. 4), using the two C. otayensis localities (MCAS Miramar and San Ysidro Mountains) as groups. This test indicated that the two groups are not distinguishable on the basis of the first or second principal component axes (P = 0.95 and P == 0.39, respectively). Population Size at MCAS Miramar A total of 74 individual shrubs of C otayensis were observed during the transect work (Fig. 3). All individuals were mature, many of them in heavy fruit at the time of observation. No seedlings were observed. Many of the C. otayensis individuals were senescent. A few isolated indi- viduals of C. otayensis were observed in the vicinity of the core population at the head of San Clemente Canyon (Fig. 3; Burge 1378; Appendix 1; Table 2), but no other populations were located at MCAS Miramar. Discussion Soils of C otayensis from the two disjunct population centers (Figs. 1 and 2; San Ysidro Mountains and MCAS Miramar) are chemically indistinguishable (Fig. 5), which is surprising given the very different geological origin of these materials (Table 2); those from the San Ysidro Mountains region have arisen from metavolcanic rocks, while those from MCAS Miramar are from sedimentary rocks, including Quaternary alluvium and more ancient conglomerate (Ap- pendix 2). The chemical similarity of these soils may help to explain the disjunct occurrence of C. otayensis at MCAS Miramar, 25 km northwest of other known populations (Figs. 1 and 2). Substrate Associations Results presented here indicate that edaphic conditions experienced by C. otayensis represent a cohesive and distinctive subset of conditions Table 1. Rock and Soil Types for 19 C otayensis Localities. Group refers to the three general groupings of localities mentioned in the body of the paper 282 MADRONO § .-S 00 ^ -3 iS ^ ^ ^ cd 3 ^ .S o 0 ? ^ - P u ^ oo & Cd 3 C/3 &iU S < « s ^ •a |£o 1 ^ ^ I" S 5 >. ;3 ^ ^ ^ « I s ^0 o g 0 « p f 32” « 'S S ^ 00 £ o " d S 5 >. M • A U n3 C 5-1 nil .3^ w ^ o oT 3 - g’O S S 3 ^ o P ^ ’Bb' . o (U § 0-S2 C •2 o p\ 3 3^ ^ Oi bfl C S § 3 d) o o ■iUli 00000000 (ZltZJC«C/3t«C«C«C« ssssssse 00000000 00 00 fi fl fl e GO CO CO 00 O Z B ^ fi E ^ a> ^ o g O a - - o (U > cd u 5-1 w &J3 >.>>>> >^ A. A. ^ A4 ^ Ad ^ 000000 000000 5-< i-i 5-1 M 5..< !-i 5_( $=( 1.^ tH 000 O' cr CT" 000 ^ ^ ^ 000 XXX WWW 1 ! t 3 3 3 ^ ^ p bO ao ao >. >, o o o o 0 0 !-i U c/5 c/5 s a cd cd O O A. ^ Ad 0 OD 00 a a cd cd O O >. Av Ad Ad o o O o 000 3 3 3 000 ..3 ^ ^ 000 XXX WWW 1 t i 3 3 3 333 60 ao GO 0 0 3 3 3^ a" ^ o o o o o X X ^ W W p W W 3 ^ ^ & 60 W) o a 3 3 0 0 ^33 ^ o" cr •^00 y ^ 43 Poo o w W 3 w w ^ o W) w) ^ u u 3 0 0 23 3 3 o o y ^ A3 Poo ^ X X p w w ■43 A, 3, o 60 60 S§§§§sss I i§§ iss 3 3 3 3 3 3 3 3 3 3 3 3 m m m m m 3 3 3 3 CO m o 3 3 ^33 ^ m m 3 3 3 3 c/5 c/5 >>>>>>>>>>> sssssssssse 00 3 3 0 O > > > > > 1 w s a a a a 3 3H-,h-3l— ^H-,^-5l--,l— ))— 5>-3l— J^-JP— sl-n>^l--5>--S g s ^^^ggg^ggggggggg ££ .'a ^ ^ ~ •^^333 3 3 3 3 3 ■<-‘+-*0 0 0 Ol IZ) Ad rS ^ ■3-3000 3 3 > > > ■-o 'O 5 -2 -M 3 3 0 0 0 w w S ^ ^ _p _p _p _p _p ^p ■3 ‘3 ‘3 ’3 '3 ’3 p _p _p _p ,0 ,p ,p a '3 ’3 ’3 '3 ’3 '3 0000000000000 >>>>>>>>>>>>> 3333333333333 0000000000000 ^nnnnnnnnnnnnnnnnS |■|•SSSS2SSS2SSSSSS2H 3 3 o o a a o o 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 _o 'o '0 'o ■o "o ’0 "o '0 '0 '0 ’0 ■q '0 ■q Z '0 > > > > > > > > > > > > > > > > > <3 Os ^ ^ t\ Os t\ n Os Os oo oo o ^ <>6 Os Os '"s ^ ^ ^ ^ ^ So Sb 60 feb So t\ ^ ^ ^ tx rr ^ K S -C) &..&», J*. cu -Q C/5C/5C/5cnc/5C/5C/5C/5C/5C/5C/5C/5C/5C^ Cd Cd a a 3 3 o o 5-1 U -O o o u 5-1 -O ^3 o o Ui U "O ^ C/5 C^ C/5 C/5 C/5 C/5 V5 m << U U SS;zi^^^^;z;;^;z;:zi2;;^:2 3 3 3 3 m c/5 3 3 3 3 C/5 m 3 3 3 3 C/5 m 0000 5h 5h i-< s.< 3 iS ."H ra ’(/) WD OP^ CO >H >N >H >^ 3 3 3 3 3 3 3 3 C/5 C/5 C/5 (75 O O Ui U >. >H 3 3 3 3 C/5 m SS 3 dl a 2sl o ^ ^ 'y 'y ^ 1*^ 60 60 >H i g § § § P-y l-y 3 3 ^ 7^ m m m [Vol. 61 ! 2014] BURGE: CEANOTHUS OTAYENSIS SOILS 283 Table 2. Soil Sampling, Species from Fross and Wilken (2006); Code D. O. Burge collecting code; Locality country, state, and county for sampling (also see Appendix 1); all collections by DOB, all deposited at DUKE; Soil Type geological material from which soil is derived, with codes from Geological Map of California in parentheses (Jennings et al. 1977). Species Code Locality Elevation (m) Geology C cuneatus 847 USA; San Luis Obispo Co., CA 310 Ultramafic (um) 891 USA; Monterey Co., CA 50 Sedimentary (Q, M) 918 USA; Santa Cmz Co., CA 240 Sedimentary (M) 982 USA; Riverside Co., CA 730 Granite (grMz) 984 USA; San Diego Co., CA 1000 Granite (grMz) 1030 Mexico; Sierra San Pedro Martir 855 Granite (grMz) 1070 USA; San Bernardino Co., CA 460 Sedimentary (Q) 1071 USA; Los Angeles Co., CA 615 Granite (grMz) 1132 USA; Kem Co., CA 1060 Granite (grMz) 1263 USA; Santa Barbara Co., CA 155 Sedimentary (Q) C. divergens 833 USA; Sonoma Co., CA 200 Ultramafic (um) C. divergens 943 USA; Lake Co., CA 1000 Volcanic (Qv) C. ferrisae 835 USA; Santa Clara Co., CA 180 Ultramafic (um) C. fresnensis 1138 USA; Fresno Co., CA 1710 Granite (grMz) C. gioriosus 907 USA; Marin Co., CA 220 Granite (grMz) C. jepsonii 900 USA; Tehama Co., CA 1200 Ultramafic (um) C. maritimus 887 USA; San Luis Obispo Co., CA 23 Sedimentary (Q) C. masonii 913 USA; Marin Co., CA 450 Sedimentary (Kjfm, Mzv) C. ophiochilus 798 USA; Riverside Co,, CA 660 Granite (grMz) C. otayensis 983 USA; San Ysidro Mts,, San Diego Co., CA 900 Volcanic (Mzv) 985 USA; San Ysidro Mts., San Diego Co., CA 800 Volcanic (Mzv) 1053 USA; San Ysidro Mts., San Diego Co., CA 430 Volcanic (Mzv) 1179 USA; MCAS Miramar, San Diego Co., CA 130 Sedimentary (Ec) 1391 USA; San Ysidro Mts., San Diego Co., CA 690 Volcanic (Mzv) 1395 USA; San Ysidro Mts., San Diego Co., CA 1016 Volcanic (Mzv) 1398 USA; MCAS Miramar, San Diego Co., CA 130 Sedimentary (Ec) C. perpiexans 795 USA; San Diego Co., CA 950 Granite (grMz) C. purpureus 905 USA; Napa Co., CA 410 Volcanic (Tv) C. roderickii 1288 USA; El Dorado Co., CA 340 Gabbro (gb) from those supporting other Ceanothus species in California and adjacent Mexico (Fig, 4). Soils of C. otayensis contain notably small amounts of nitrate and P, low availability of which is known to result in disorders affecting the growth and reproduction of crop plants (Brady and Weil 2002), The soils also have low pH and conduc- tivity, which is probably due to generally low levels of nutrient and other ions (Table 3). Geospatial data on geological formations and soil types show that the geology and soils of C otayensis from the San Ysidro Mountains and San Miguel Mountain are highly consistent (Table 2; Appendix 2), with metavolcanic geology giving rise to soils from two series, the San Miguel- Exchequer rocky silt loams and Metamorphic rock land. Metavolcanic rocks of the San Ysidro Mountains and San Miguel Mountain are part of Table 3. Summary Statistics for Soil Chemistry Variables. Ceanothus otayensis San Ysidro Mts., n = 5; Ceanothus otayensis MCAS Miramar, n = 2; Other Ceanothus species, n = 22; all variables reported as average plus or minus standard deviation; conductivity reported as gmol/cm; nitrate and all other levels reported as ppm. Ceanothus otayensis Ceanothus otayensis Other Ceanothus Variable San Ysidro Mts. MCAS Miramar species pH 5.91 0.20 5.46 ± 0.21 6.03 H- 0.72 Conductivity 68 ± 25 67 ± 7.07 127.46 -h 115.55 Nitrate 2.01 ± 1.46 1.32 ± 1.12 12.42 ± 17.83 P 6.49 1.08 8.78 4.10 28.40 -h 35.15 K 203.34 -h 59.55 204.87 50.92 172.92 -h 96.56 Ca 1197.84 -h 369.56 967.48 ± 81.31 1222.10 -1- 721.59 Mg 201.33 -H 51.58 237.70 H- 42.52 697.86 -4- 1091.01 S 6.02 -h 1.83 6.73 -H 0.99 10.91 -h 3.79 Na 35.32 ± 3.89 33.58 1.75 81.26 -h 63.56 Fe 10.25 -h 3.08 17.86 ± 5.96 23.03 H- 14.90 Zn 0.60 -1- 0.41 1.50 4- 0.34 0.88 -t- 1.27 Mn 9.37 2.66 13.38 2.25 10.68 -h 7.14 Cu 0.14 -H 0.04 0.34 0.05 0.38 ± 0.30 284 MADRONO [VoL 61 -6-4-2 0 2 4 First PC Fig. 4. Plot from principal components (PCA) analysis of soil chemistry. Biplot for first two principal components of PCA for 29 soil samples. Arrows represent direction and magnitude of loading on principal component axes. Symbols: Con = electrical conductivity; N03 = nitrate. a narrow band of such geological materials extending northward from northern Baja Califor- nia, Mexico into San Diego County, including some minor peaks and uplands to the west and northwest of San Diego (e.g., Black Mountain). Metavolcanic-derived soils of the San Ysidro Mountains and San Miguel Mountain are known to host a diverse flora and a number of near- endemics, such as Arctostaphylos otayensis Wies. & Schreib. (Ericaceae), Lepechinia ganderi Epling (Lamiaceae), Hosackia crassifolius Benth. var. otayensis (Moran ex Isely) Brouillet (Fabaceae), and notable populations of other rare species, such as Brodiaea orcuttii (Greene) Baker (Themi- daceae), Calochortus dunnii Purdy (Liliaceae), Stipa diegoensis Swallen (Poaceae), Fremontoden- dron mexicanum Davidson (Sterculiaceae), Quer- cus cedrosensis C.H. Muller (Fagaceae), and Hesperocyparis forbesii (Jeps.) Bartel (Cupressa- ceae). Many of these species are also known from gabbro-derived soils of San Diego County, which often support a unique suite of rare species (Zedler 1995, Alexander 2011). The presence of these rare taxa on both metavolcanic and gabbro rock suggests that there may be some chemical similarity between soils derived from this mate- rial, as the present paper shows for the metavol- canic-derived and sedimentary soils of C. otayen- sis. Future work should compare the substrate associations of these species. Such work would likely aid in the management of these rare and poorly-known taxa. 2014] BURGE: CEANOTHUS OTAYENSIS SOILS 285 Canonical scores Structure other spp. q otayensis C. otayensis San Ysidro MCAS Mountains Miramar Fig. 5. Results of PCA Plot from generalized canonical discriminant analysis (CD A) of soil chemistry. Plot of single canonical axis from CDA comparing means of 1) C. otayensis from MCAS Miramar, 2) C. otayensis from the San Ysidro Mountains, and 3) other Ceanothus species. Symbols: Con = electrical conductivity; NOS = nitrate. Sedimentary soils like those at the MCAS Miramar locality are also known to support a diverse and unusual flora, possibly due to the low pH that prevails on some soils (Crocker 1956). However, most of these species are vernal-pool endemics or associates, Ceanothus otayensis, a shrub normally associated with chaparral habi- tats, thus provides an unusual counterpoint to the typical pattern of endemic plant diversity at MCAS Miramar. Although additional research is clearly needed, particularly targeted surveys of plant diversity in and near MCAS Miramar, there are some intriguing examples of other plants disjunct between MCAS Miramar and the San Ysidro Mountains; the endangered herb Monardella viminea Greene is found on MCAS Miramar and a few other scattered localities in San Diego County and adjacent Baja California, Mexico, while the closely related Monardella stoneana Elvin & A.C. Sanders is found only on metavolcanic soils of the San Ysidro Mountains. It is possible that other examples of this disjunction exist, but have not been detected by botanists. The present study shows that there is a chemical similarity between the soils of C otayensis from the San Ysidro Mountains and those at MCAS Miramar, which in turn suggests that other plants of the San Ysidro Mountains, or other rare plants of southern San Diego County, might be expected to occur there. Such examples of disjunction due to soil cross-toler- ance are well known in the botanical literature (reviewed in Brady et al. 2005), but few cases have been examined in a soil chemistry or experimental context. In considering the ecology of C. otayensis and other rare plants from the region of the border between Mexico and the Unites States, it is 286 MADRONO [Vol. 61 i important to also consider populations from Baja California, Mexico. In the case of C otayensis, a population of the species is known to occur near the summit of Cerro Jesus Maria in Baja California, one of the southern peaks of the San Ysidro Mountains (Fig. 1). Here, it appears to occur on andesite-derived (volcanic) soils (Ap- pendix 1). It would be helpful to obtain soil samples from this locality for future work involving the evolution and ecology of C. otayensis. It would also be helpful to analyze soil samples from the populations of plants that have been reported from San Miguel Mountain in San Diego County {Keeley 27100-27117, Rebman 24367; Appendix II). Evolution of Soil Specificity Although my work does not directly address the evolutionary history of the new, disjunct C. otayensis population on Kearney Mesa at MCAS Miramar, it is possible to speculate on the historical events that may have led to the present distribution of the species. One possibility is that the population at MCAS Miramar is the product of a recent dispersal event from the south, and has been able to persist in the new, unusual habitat due to the amenable chemistry of the soils. Alternatively, plants at MCAS Miramar may be relictual, isolated by the loss of plant populations in the region separating MCAS Miramar from the San Ysidro Mountains and San Miguel Mountain, It is intriguing to note that metavolcanic rocks, as well as sedimentary formations of the same type found at the MCAS Miramar locality, are broadly distributed in a continuous band from the San Ysidro Mountains region to north of MCAS Miramar (Fig. 2). It is possible that some of these areas support other small, undiscovered populations of C. otayensis. The soils of C. otayensis are chemically similar to soils that support another narrowly-endemic Ceanothus species, Ceanothus roderickii Knight, which is known from gabbro-derived soils of the central Sierra Nevada (Fig. 4; Burge and Manos 2011). Recent research on C. roderickii indicates that the chemistry of its soils may provide a powerful agent of natural selection, possibly leading to reduced gene flow with closely related species, local adaptation, and speciation (Burge et al. 2013). If this hypothetical mode of divergence and speciation is common in Ceanothus, it may help explain the origin of C. otayensis, as well as several other narrowly-endemic Ceanothus spe- cies from southern California and nearby Baja California, including Ceanothus ophiochilus S. Boyd, T. S. Ross, & Arnseth on pyroxenite-rich gabbros of Riverside County (Boyd and Arnseth 1991), and Ceanothus holensis S. Boyd & J, Keeley on metavolcanic rocks of Baja Califor- nia’s Cerro Bola (Boyd et al. 2002). In general, it would be worthwhile to make a broader test of soil chemistry associations across the ten or so supposedly edaphic-endemic mem- bers of Ceanothus subg. Cerastes, to determine whether there is a consistent syndrome of edaphic adaptation across the group. Such work should ideally include common garden experiments, reciprocal transplant experiments, and more detailed analyses of other soil parameters that are potentially related to the chemical parame- ters, such as moisture availability. This would help to overcome the limitations of this and many similar studies, which often rely on purely observational data from a small suite of chemical soil properties. Status of the New Population and of the Species Ceanothus otayensis is on list IB. 2 of the California Native Plant Society Inventory of Rare and Endangered Plants (CNPS 2014), meaning that it is rare in California as well as in its other areas of distribution (in this case Baja California, Mexico). Ceanothus otayensis is facing a number of significant threats throughout its geographic range. The major threat to the persistence of the species is likely fire regime; C. otayensis is an obligate seeding species that requires fire for recruitment (Wilken 2006). However, too fre- quent fire has been shown to eliminate obligate seeding species from chaparral habitats in south- ern California (Zedler 1995). Major wildfires have burned large portions of the San Ysidro Moun- tains during the past decade, including the Cedar i Fire (ignited 25 October 2003; burned 582 km^; ! CAL FIRE 2014), and the Witch Creek Fire (ignited 21 October 2007; burned 801 km^; CAL ' FIRE 2014). These were two of the largest fires in the recorded history of California (CAL FIRE ! 2014), and both burned substantial portions of C. otayensis habitat in the San Ysidro Mountains ! and nearby Peninsular Ranges, including portions in Baja California, Mexico. Short fire intervals of this kind will probably result in the extirpation of ' C. otayensis populations that are unable to reach sexual maturity between fires. Though populations of C otayensis in the San Ysidro Mountains are large and seem reasonably secure, the smaller populations at San Miguel Mountain and MCAS Miramar could suffer drastic losses in the event of future fires. For example, virtually the whole of San Miguel Mountain was burned in the 2007 Witch Creek Fire (CAL FIRE 2014), including the portion of the mountain where C. otayensis was most recently recorded {Keeley 27100-27117, Rebman 24367', Appendix 1). As discussed above, the new record of the species from MCAS Miramar represents a significant expansion of the ecological and geographic range known for the species. The 2014] BURGE: CEANOTHUS OTAYENSIS SOILS 287 new locality at MCAS Miramar brings the total number of population centers for the species to three, one in the San Ysidro Mountains, a second on San Miguel Mountain, and a third, now, at MCAS Miramar (Fig. 1). Nevertheless, the pop- ulation at MCAS Miramar is extremely small, seemingly consisting of no more than 74 mature plants. Although the size of the nearest C otayensis population on San Miguel Mountain is also very small (probably no more than 100 individuals as of 2010, Jon Rebman personal communication), population sizes in the San Ysidro Mountains are generally significantly larger, consisting of thousands of plants, becom- ing a co-dominant member of the chaparral plant community at higher elevations. The very small size of the C otayensis population at MCAS Miramar makes it especially vulnerable to loss. In such populations, chance events such as wildfire, flooding, or disease introduction can lead to drastic losses of individuals, and possibly local extinction. In addition, the population genetic status of this small population is unknown. If the population is derived via a recent dispersal event from the southern populations, it may suffer from a lack of genetic diversity due to a population bottleneck (Ellstrand and Elam 1993). On the other hand, if the plant at MCAS Miramar is a relictual population, isolated by the loss of plant populations in the region separating Kearny Mesa from the San Ysidro Mountains, it may harbor unique genetic variation not present in the southern populations. Future work should aim to quantify population genetic variation across the range of C. otayensis, especially in isolated population centers, such as those at MCAS Miramar and San Miguel Mountain. Acknowledgments The author thanks Katherine Zhukovsky, Jon Reb- man, and Dieter Wilken for providing constructive criticism of drafts. Assistance with field work and field logistics was provided by Katherine Zhukovsky, JoEllen Kassebaum, and Joyce Schlachter. Funding was provided by the American Society of Plant Taxonomists, The Hunt Institute for Botanical Docu- mentation, Duke University, and a National Science Foundation grant to DOB and PSM (NSF 000457253). Literature Cited Alexander, E. B. 2011. Gabbro soils and plant distributions on them. Madrono 58:113-122. Baldwin, B. G. 2005. Origin of the serpentine-endemic herb Layia disco idea from the widespread L. glandulosa (Compositae). Evolution 59:2473-2479. Bowman, R. H. 1973. Soil survey of San Diego Area, California. United States Department of Agricul- ture. Soil Conservation Service, Washington, D.C. Boyd, S. and J. E. Keeley. 2002. A new Ceanothus (Rhamnaceae) species from northern Baja Califor- nia, Mexico. Madrono 49:289-294. , T. Ross, AND L. Arnseth. 1991. Ceanothus ophiochilus (Rhamnaceae): a distinctive, narrowly endemic species from Riverside county, California. Phytologia 70:28-41. Brady, N. C. and R. R. Weil. 2002. The nature and properties of soils, 13th ed. Pearson Education, Upper Saddle River, NJ. Burge, D. O. and P. S. Manos. 2011. Edaphic ecology and genetics of the gabbro-endemic shrub Ceanothus roderickii (Rhamnaceae). Madrono 58:1-21. , D. M. Erwin, M. B. Islam, J. Kellermann, S. W. Kembel, P. S. Manos, and D. H. Wilken. 2011. Diversification of the genus Ceanothus (Rhamnaceae) in the California Floristic Province. International Journal of Plant Sciences 172:1137- 1164. , R. Hopkins, Y.-H. E. Tsai, and P. S. Manos. 2013. Limited hybridization across an edaphic disjunction between the gabbro-endemic shrub Ceanothus roderickii (Rhamnaceae) and the soil- generalist C cuneatus. American Journal of Bota- ny. 100:1883-1895. California Department of Forestry and Fire PROTECTION (CAL FIRE). 2014. CAL FIRE Incident Information. Website http://cdfdata.fire. ca.gov/incidents/incidents_statsevents (accessed 2 June 2013). California Native Plant Society, Rare Plan Program (CNPS). 2014. Inventory of Rare and Endangered Plants (online edition, v8-02). Califor- nia Native Plant Society. Sacramento, CA. Website http://www.cnps.org/inventory (accessed 02 May 2014). Consortium of California Herbaria (CCH). 2013. Data provided by the participants of the Consortium of California Herbaria. Website ucjeps.berkeley. edu/consortium (accessed 16 July 2013). Crocker, R. L. 1956. The acid soils of the San Diego Mesa, California. Journal of Soil Science 7:242- 247. Ellstrand, N. C. and D. R. Elam. 1993. Population genetic consequences of small population size: implications for plant conservation. Annual Re- view of Ecology and Systematics 24:217-242. Friendly, M. 2007. HE plots for multivariate general linear models. Journal of Computational and Graphical Statistics 16:421M44. and j. Fox. 2009. Candisc: generalized canonical discriminant analysis. R package version 0.5-16. Website http://cran.r-project.org/ (accessed July 10, 2013). Wilken, D. 2006. Ceanothus in the wild. Pp. 131-246 in D. Fross and D. Wilken (eds.), Ceanothus. Timber Press, Portland, OR. Gankin, R. and J. Major. 1964. Arctostaphylos myrtifolia, its biology and relationship to the problem of endemism. Ecology 45:792-808. Gastil, G. G., R. P. Phillip, and E. C. Allison. 1971. Reconnaissance geology of the State of Baja California. Geological Society of America Memoir 140. Gittins, R. 1985. Canonical Analysis: a review with applications in Ecology. Springer-Verlag, Berlin, Germany. Goslee, S. C. and D. L. Urban. 2007. The ecodist package for dissimilarity-based analysis of ecolog- ical data. Journal of Statistical Software 22:1-19. Jarvis, A., H. 1. Reuter, A. Nelson, and E. Guevara. 2008. Hole-filled SRTM for the globe 288 MADRONO Version 4, available from the CGIAR-CSI SRTM 90m Database (http://srtm.csi.cgiar.org). Jennings, C. W., R. G. Strand, and T. H. Rogers. 1977. Geologic map of California, 1:750,000: California Division of Mines and Geology, Sacra- mento, CA. Kay, K. M., K. L. Ward, L. R. Watt, and D. W. SCHEMSKE. 2011. Plant Speciation. Pp. 71-96 in S. P. Harrison and N. Rajakaruna (eds.), Serpen- tine: The Evolution and Ecology of a Model System. University of California Press, Berkeley, CA. Kennedy, M. P. and S. S. Tan. 1977. Geology of La Mesa, Jamul, and Otay Mesa 7,5' Quadrangles. California Division of Mines and Geology, Sacra- mento, CA. Kruckeberg, a. R. 1986. An essay: the stimulus of unusual geologies for plant speciation. New Phy- toiogist 11:455M63. . 2002. Geology and plant life. University of Washington Press, Seattle, WA. Legendre, P. and L. Legendre. 1998. Numerical Ecology. Elsevier, Amsterdam, The Netherlands. Mason, H. L. 1946. The edaphic factor in narrow endemism. 1. The nature of environmental influ- ences. Madrono 8:209-240. McMinn, H. E. 1942. Part 11. A systematic study of the genus Ceanothus. Pp. 131-279 in M. van Rensselaer and H. E. McMinn (eds.), Ceanothus. Santa Barbara Botanical Gardens, Santa Barbara, CA. O'Dell, R. E. and N. Rajakaruna. 2011. Intraspe- cific variation, adaptation, and evolution. Pp. 97- 137 in S. P. Harrison and N, Rajakaruna (eds.). Serpentine: The Evolution and Ecology of a Model System. University of California Press, Berkeley, CA. R Development Core Team. 2009. R: a language and environment for statistical computing. R Founda- tion for Statistical Computing, Vienna, Austria. Website http://cran, r-project.org/ (accessed July 10, 2013). Sambatti, J. B. M. and K. J. Rice. 2006. Local adaptation, patterns of selection, and gene flow in the Californian serpentine sunflower {Helianthus exilis). Evolution 60:696-710. Stebbins, G. L. 1942. The genetic approach to problems of rare and endemic taxa. Madrono 6:241-258. United States Department of Agriculture Natural Resource Conservation Service (USDA). 2007. Survey Geographic (SSURGO) database for San Diego County, California. U.S. Department of Agriculture, Natural Resources Conservation Service, Fort Worth, TX. the Pacific Ocean 3(1). Washington: Government Printing Office. United States Geological Survey (USGS). 2012. uses 0.3m Orthoimage 11SMS875345, San Diego, California. U.S. Geological Survey, Sioux Falls, SD. WiLKEN, D. 2006. Ceanothus in the wild. Pp. 131-246 in D. Fross and D. Wilken, Ceanothus. Timber Press, Portland, OR. Zelder, P. H. 1995. Plant life history and dynamic specialization in the chaparral/coastal sage shrub flora in southern California. Ecological Studies 108:89-115. [VoL 61 1 Appendix 1 f Sampled Ceanothus Populations. jl For each sampled population (see Table 2), the format ' is as follows: collector name and number (herbarium of ’ voucher specimen deposition), description of locality, county, US or Mexican state. j i Ceanothus cumeatus (Hook.) Nutt. — D.O. Burge 847 \ (DUKE), Irish Hills Natural Reserve, San Luis Obispo Co., CA. D.O. Burge 891 (DUKE), Fort Ord Military Reservation, Monterey Co., CA. D.O. Burge 918 | (DUKE), Henry Cowell Redwoods State Park, Santa i Cruz Co., CA. D.O. Burge 982 (DUKE), Tucaiota | Creek watershed, Riverside Co., CA. D.O. Burge 984 '\ (DUKE), Morena Valley, San Diego Co., CA, D.O. , Burge 1030 (DUKE), Sierra San Pedro Martir, Baja i California, Mexico. D. O. Burge 1070 (DUKE), Rialto Municipal Airport, San Bernardino Co., CA. D.O. Burge 1071 (DUKE), Sierra Pelona Mountains, Ruby ■ Canyon, Los Angeles Co., CA. D.O. Burge 1132 \ (DUKE), Clear Creek watershed, south of Ball | Mountain, Kem Co., CA. D.O. Burge 1263 (DUKE), i Solomon Hills, north of Gradosa Ridge, Santa Barbara I Co., CA. j Ceanothus divergens Parry — D. O. Burge 833 (DUKE), Southeast flank of Mount Hood, Sonoma Co., CA. D.O. Burge 943 (DUKE), Boggs Mountain Demonstration State Forest, Lake Co., CA. Ceanothus ferrisiae McMinn — D. O. Burge 835 (DUKE), Anderson Lake County Park, Santa Clara Co., CA. Ceanothus fresnemis Dudley ex Abrams — D. O. Burge 1138 (DUKE), Big Creek Watershed, Fresno Co., CA. Ceanothus gloriosus J.T. Howell — D.O. Burge 907 (DUKE), Point Reyes National Seashore, Inverness Ridge, Marin Co., CA. Ceanothus jepsonii Greene — D. O. Burge . 900 (DUKE), Mendocino National Forest, at roadside 12.5 road miles (20 km) west of Paskenta, Tehama Co., CA. Ceanothus maritimus Hoover — D. O. Burge 887 (DUKE), Roadside on Hwy 1, 0.5 road miles (0.8 km) north of bridge over Arroyo de los Chinos, San Luis Obispo Co., CA. Ceanothus masonii McMinn — D. O. Burge 913 (DUKE), Golden Gate National Recreation Area, Bolinas Ridge, Marin Co., CA. Ceanothus opMocMlus S. Boyd, T.S, Ross & Arn- seth — D. O. Burge 798 (DUKE), Agua Tibia Wilderness Area, Cleveland National Forest, Riverside Co., CA. Ceanothus otayensis McMinn — D. O. Burge 983 (DUKE), Otay Mountain, 6.6 road miles from Otay Lakes Road via Minewawa and Otay Mountain Truck Trails, San Diego Co., CA. D.O. Burge 985 (DUKE), Otay Mountain, 7.4 road miles from Otay Lakes Road via Minewawa and Otay Mountain Truck Trails, San Diego Co., CA. D.O. Burge 1053 (DUKE), San Ysidro Mountains, roadside on Otay Mountain Truck Trail, 1 .7 road miles (2.7 km) from Alta Road, San Diego Co., CA. D.O. Burge 1179 (DUKE), Marine Corps Air Station Miramar, Kearny Mesa, San Diego Co., CA. D.O. Burge 1391 (DUKE), San Ysidro Mountains, northern slope of Otay Mountain, upper slopes of Little Cedar Canyon, San Diego Co., CA. D.O. Burge 1395 ' (DUKE), San Ysidro Mountains, Otay Mountain, 2014] BURGE: CEANOTHUS OTAYENSIS SOILS 289 Doghouse Junction, San Diego Co., CA. D.O. Burge 1398 (DUKE), Marine Corps Air Station Miramar, Kearny Mesa, San Diego Co., CA. Ceanothus perplexans Trel. — D.O. Burge 795 (DUKE), Chariot Canyon, San Diego Co., CA. Ceanothus purpureus Jeps. — D. O. Burge 905 (DUKE), Atlas Road, Napa Co., CA. Ceanothus roderickii W. Knight — D.O. Burge 1288 (DUKE), South Fork American River watershed, El Dorado Co., CA. Appendix 2 Geology and Soil Associations of C. OTAYENSIS. For each population, the format is as follows: collector name and number (herbarium of voucher specimen deposition), description of locality, county, US or Mexican state (state geology; local geology; soil). General geology for California is from the Geologic Map of California (Jennings et al. 1977); Mexican locality is from Reconnaissance Geology of the State of Baja California (Gastil et al. 1971). Local geology is from Geology of La Mesa, Jamul, and Otay Mesa 1.5' Quadrangles (Kennedy et al. 1977); data for some San Ysidro Mountains localities are not available for these maps, but are likely to be the same as the other San Ysidro Mountains localities, as indicated. Soil, is from the Soil Survey Geographic (SSURGO) database for San Diego County. Ceanothus otayensis — Moran 17863 (CAS), Summit of Otay Mountain, San Diego Co., CA (state geology: volcanic [Mzv]; local geology: Metavolcanic [KJmv]; soil: Metamorphic rock land [MrG]. Moran 23785 (RSA), 1 1/2 miles east of Doghouse Junction, Otay Mountain, San Diego Co., CA (state geology: volcanic [Mzv]; local geology: Metavolcanic [KJmv]; soil: San Miguel-Exchequer rocky silt loams [SnG]. Pratt s.n. (UCR), Along the main road up Otay Peak, San Diego Co., CA (state geology: volcanic [Mzv]; local geology: Metavolcanic [KJmv]; soil: San Miguel-Exchequer rocky silt loams [SnG]. Pierce s.n. (UCR), SW side Otay Mtn, just east of head of Johnson Cyn, San Diego Co., CA (state geology: volcanic [Mzv]; local geology: Metavolcanic [KJmv]; soil: San Miguel-Exchequer rocky silt loams [SnG]. Elvin 1306 (CAS), West Otay Mountain, north of Otay Truck Trail, San Diego Co., CA (state geology: volcanic [Mzv]; local geology: Metavolcanic [KJmv]; soil: San Miguel-Exchequer rocky silt loams [SnG]. Rebman 6742 (RSA), West side of Otay Mountain, San Diego Co., CA (state geology: volcanic [Mzv]; local geology: Metavolcanic [KJmv]; soil: San Miguel-Exchequer rocky silt loams [SnG]. Sanders 26436 (RSA), Otay Mountain, lower ‘Copper Canyon’, San Diego Co., CA (state geology: volcanic [Mzv]; local geology: Metavolcanic [KJmv]; soil: San Miguel-Exchequer rocky silt loams [SnG]. Betzler 515 (RSA), Otay Mountain, Minnewawa Truck Trail about 3.5 miles from top, San Diego Co., CA (state geology: volcanic [Mzv]; local geology: Metavolcanic [KJmv]; soil: San Miguel-Exchequer rocky silt loams [SnG], Rebman 11043 [RSA], Otay Mountain Ecological Reserve, San Diego Co., CA (state geology: volcanic [Mzv]; local geology: Metavolcanic [KJmv]; soil: San Miguel-Exchequer rocky silt loams [SnG]. Burge 983 [DUKE], Otay Mountain, 6.6 road miles from Otay Lakes Road via Minewawa and Otay Mountain Truck Trails, San Diego Co., CA (state geology: volcanic [Mzv]; local geology: Metavolcanic [KJmv]; soil: Metamorphic rock land [MrG]). Burge 985 (DUKE), Otay Mountain, 7.4 road miles from Otay Lakes Road via Minewawa and Otay Mountain Truck Trails, San Diego Co., CA (state geology: volcanic [Mzv]; local geology: Metavolcanic [KJmv]; soil: Metamorphic rock land [MrG]). Burge 1053 (DUKE), San Ysidro Moun- tains, roadside on Otay Mountain Truck Trail, 1.7 road miles (2.7 km) from Alta Road, San Diego Co., CA (state geology: volcanic [Mzv]; local geology: Metavol- canic [KJmv]; soil: San Miguel-Exchequer rocky silt loams [SnG]). Burge 1391 (DUKE), San Ysidro Mountains, northern slope of Otay Mountain, upper slopes of Little Cedar Canyon, San Diego Co., CA (state geology: volcanic [Mzv]; local geology: Metavol- canic [KJmv]; soil: San Miguel-Exchequer rocky silt loams [SnG]). Burge 1395 (DUKE), San Ysidro Mountains, Otay Mountain, Doghouse Junction, San Diego Co., CA (state geology: volcanic [Mzv]; local geology: Metavolcanic [KJmv]; soil: Metamorphic rock land [MrG]). Keeley 27112 (RSA), N face of San Miguel Mtn, 8 km SW of Hwy 94 at bridge near Jacumba, San Diego Co., CA (state geology: volcanic [Mzv]; local geology: Metavolcanic [KJmv]; soil: San Miguel-Ex- chequer rocky silt loams [SnG]). Rebman 24367 (SD), San Miguel Mountain, east of the Sweetwater Reser- voir, along Miller Ranch Road, San Diego Co., CA (state geology: volcanic [Mzv]; local geology: Metavol- canic [KJmv]; soil: San Miguel-Exchequer rocky silt loams [SnG]). Burge 1179 (DUKE), Marine Corps Air Station Miramar, Kearny Mesa, San Diego Co., CA (state geology: Eocene conglomerate [Ec]; local geology: Lindavista Formation [Ql]; soil: Terrace escarpments [TeF]). Burge 1398 (DUKE), Marine Corps Air Station Miramar, Kearny Mesa, San Diego Co., CA (state geology: Eocene conglomerate [Ec]; local geology: Lindavista Formation [Ql]; soil: Redding gravelly loam [RdC]). Burge 1063 (DUKE), San Ysidro Mountains, Cerro Jesus Maria, Baja California, Mexico (Geology: Andesite [Tpa]). Madrono, VoL 61, No. 3, pp. 290-298, 2014 COMPETITION AND NICHE REQUIREMENTS OF COREOPSIS TINCTORIA: A WIDESPREAD BUT LOCAL HIGH DENSITY ANNUAL ASTERACEAE S. A. Elliott and O. W. Van Auken Department of Biology, The University of Texas at San Antonio, San Antonio, TX 78249 oscar.vanauken@utsa.edu Abstract Coreopsis tinctoria (coreopsis, calliopsis, plains coreopsis, or golden wave) is an annual herbaceous Asteraceae with a broad geographic distribution mostly in the central and western United States. It co-occurs with Bouteloua curtipendula (sideoats grama) or other native C4 grasses. When grown together, C. tinctoria response variables decreased significantly in the presence of B. curtipendula aboveground, belowground, and total dry mass. The response variables included mean plant height, number of flower buds per plant, flowers per plant, as well as aboveground, belowground, and total dry mass. The presence of B. curtipendula belowground dry mass caused the greatest suppression of C. tinctoria belowground dry mass. When B. curtipendula tops were clipped to reduce grass aboveground dry mass (simulated herbivory), the percent survival of C. tinctoria plants increased from one percent in the no-clipping treatment to 18% in the neighbor removal treatment (100% clipped). Coreopsis tinctoria does not appear to be a good competitor in the presence of B. curtipendula and seems to be restricted to gaps or patches in disturbed grasslands where competition from perennial grass neighbors is reduced. Key Words: Belowground competition, aboveground competition, simulated herbivory, C4 grass, Bouteloua curtipendula, sideoats grama, vegetation gaps, dry mass. The North American Prairies once covered approximately 3.63 X 10^ km^ of the North American Continent. Today only a small percent (—1%) of the original prairies remain with most being converted to pasture or cropland (Barbour et al. 1999). On a large scale these still intact grassland communities seem stable and relatively uniform, but on a small scale, they more closely resemble a mosaic of successions or miniature successions started by disturbances that create gaps or patches in the communities (see Begon et al. 2006; Smith and Smith 2012). Fire, burrowing animals, large or small grazers (resulting in trampling of vegetation or intense grazing), and drought are often sources of disturbance in grassland habitats (Begon et al. 2006). These disturbance gaps or patches appear to serve as a reset mechanism (Pickett and White 1985) to reopen these mature communities to early sue- cessional species. The patch dynamics theory of community structure (see Begon et al. 2006) suggests that patches created by disturbance are initially recolonized by early successional species and then proceed in time toward a mature community. Vegetation gaps vary in shape, size, degree of vegetation removal (intensity), timing (recurring or uncommon), and spatial arrange- ment (McEvoy et al. 1993). Individual patches or gaps, created at different times, differ in their composition of early and late successional species. Thus, these disturbance patches or gaps should be viewed both temporally and spatially. Establishment of early successional species increases with reduced competition (Cohn et al. 1989; Cahill and Casper 2002; Mazia et al. 2013), increased soil surface light levels, and soil resources (Bush and Van Auken 1986; Van Auken and Lohstroh 1990; Bush and Van Auken 2010). These characteristics are found in disturbances, clearings, openings, or gaps. Many annuals are opportunistic, early successional species that colonize grassland gaps created by disturbances (Begon et al. 2006). As time passes, other species, usually mid-successional or late-successional spe- cies, establish in these gaps or patches. These later species have different characteristics and require- ments compared to the early successional species. With the change in species composition, from early to mid or late, the community characteristics would also change. Late successional communi- ties do not have the highly repetitive disturbance characteristics of early communities (Bond 2008; Chaneton et al. 2012; Mazia et al. 2013). Important factors in determining the success of seed or plant invasion into grassland communi- ties are: (1) seed or propagule arrival or pressure, (2) community type and characteristics of the community being invaded or encroached, (3) the disturbance regime including type, size and frequency, and (4) biotic interactions (Bond 2008; Chaneton et al. 2012; Mazia et al. 2013). To be successful, early species or gap species usually have to produce large numbers of easily dispersed seeds. Once these seeds disperse into the grassland community they either germinate and establish immediately in newly created gaps or become part of the seed bank, surviving in the soil until a disturbance occurs (Baskin and 2014] ELLIOTT AND VAN AUKEN: COMPETITION IN COREOPSIS 291 Baskin 200 1). Once a seed has reached a site and germinated, intra- and interspecific competition are major factors determining establishment and future success (Mazia et al. 2013). Competition (a negative biotic effect) has been shown to cause major effects on plant growth, reproduction, and survival within natural communities and is often cited as a major factor in determining community composition and structure (Harper 1977; Connell 1983; Schoener 1983; Grace and Tilman 1990; Grellier et al. 2012). Other factors have also been shown to influence community composition and structure, but which factor or factors are more important is undetermined (Callaway 1995; Bert- ness and Leonard 1997; Callaway and Walker 1997; Van Auken and Bush 1997; Wilson and Nisbet 1997; Pearson et al. 2011; Busch et al. 2012). Competition’s role in determining a species’ distribution and community structure has been frequently examined. For example, the distribu- tion of Heterostipa neomexicana (Thurb. ex J. M. Coult.) Barkworth, New Mexico feathergrass {=^Stipa neomexicana [Thurb ex J. M. Coult.] Scribn.) was limited by competition from neigh- boring grasses that restricted it to less favorable grassland microhabitats (Gurevitch 1986). Nas~ sella leucotricha (Trin. & Rupr.) R. W. Pohl {=^Stipa leucotricha Trin. & Rupr., Texas winter- grass) a C3 grass, was a better competitor than the co-occurring C4 perennial grass, Schizachyr- ium scoparium (Michx.) Nash (little bluestem), at high light and nitrogen levels (White and Van Auken 1996). At low light and nitrogen levels competition was equal. However, Schizachyrium scoparium is probably a better competitor than Nassella leucotricha only at high temperatures. With other C4 species, Schizachyrium scoparium appears to be an equal or better competitor in low nutrient soils (Bush and Van Auken 2010). Consequently, a combination of abiotic and biotic factors appears to determine species distributions. Density of Coreopsis lanceolata L., a herbaceous perennial, was negatively correlated with density of other species (Folgate and Scheiner 1992). Plants grew, survived, and reproduced better in areas where competitors were removed, especially if nutrients were added. Species that appear to be poor competitors may be able to avoid competition by completing their life cycle before or after a competitor or by growing in areas where competition is reduced or absent (Eddy 2013). Many herbaceous C3 species grow and complete their life cycle early in the growing season before C4 plants start their growth. Poor competitors, including many annu- als, grow in gaps or patches where their competitors, C3 or C4 perennial grasses, are not present or cannot complete their life cycle (Van Auken 2000; Cahill and Casper 2002). Coreopsis tinctoria Nutt, is an annual herba- ceous Asteraceae found throughout North Amer- ican Prairies in the U.S., southern Canada, and northern Mexico (Strother 2006). In some local areas density is high, but reasons for this are not clear and in many areas its density is very low (Correll and Johnston 1979). Grassland distur- bance gaps appear to be areas where C tinctoria is able to establish, grow, and complete its life cycle. The purposes of the experiments reported here were to examine factors responsible for the growth and success of C. tinctoria in a native, southern, C4 grassland. We hypothesized a reduction in the growth and dry mass production of C. tinctoria in the presence of a C4 grass due to poor competitive abilities of C tinctoria. We wanted to know if grass aboveground (shoot) or belowground (root) mass interference was equal or more important in suppressing C. tinctoria growth. We hypothesized that C. tinctoria growth characteristics would be reduced if the above- ground (shoot), belowground (root), or both aboveground and belowground dry mass of the grass was high. We expected the greatest reduc- tion of C tinctoria dry mass to be caused by belowground grass dry mass. We also hypothe- sized that survival of C tinctoria seedlings in the field increase as grass dry mass decreases due to simulated herbivory. Methods Field interspecific competition experiment. A field interspecific competition experiment be- tween C tinctoria (target) and Bouteloua curti- pendula (Michx.) Torr. (matrix) was conducted to examine competition between the two species including potential differences among above- ground (shoot) and belowground (root) compe- tition or a combination of both shoot and root competition. The experiment was started in April 1997 and was harvested 51-52 d later. The study was conducted at a field site on the University of Texas at San Antonio campus (98°37'47.93"W, 29°34'56.88"N). The soil in the experimental plot was a low nutrient, Patrick Series Mollisol, classified as a clayey over sandy, carbonatic- thermic, typic calciustoll (Taylor et al. 1966). The field site was dominated by B. curtipendula with Cirsium texanum Buckley and Helianthus annuus L. present as low-density species. The site was enclosed by a 1.8 m high chain-link fence to prevent large animal herbivory. Tap water was used to maintain the soil at field capacity. Four competition treatments of the matrix species with 10 replicates each were established in 40, 1 m X 1 m plots at the study site, including no roots/no shoots (NR/NS), roots/no shoots (R/NS), no roots/shoots (NR/S), and roots/shoots (R/S) (Gerry and Wilson 1995; McPhee and Aarssen 2001). 292 MADRONO [Vol. 61 The no roots/no shoots (NR/NS) treatment was established by removing root (belowground) and shoot (aboveground) competition of B. curtipendula (Van Auken and Bush 1997). Root competition was removed by means of a 20 cm root excluder (20 cm deep X 10 cm diameter X 3 mm thick plastic PVC pipe). The root excluders were hammered into the ground to the level of the soil surface at the center of each NR/NS or NR/S plot and all live surface vegetation within the root excluder was manually removed. Shoot competi- tion was removed with a 1 m X 1 m wire mesh (4 cm X 4 cm grid size). The mesh was placed over existing vegetation in plots (predominately B. curtipendula) and secured with 15 cm iron spikes at each corner. A 10 cm diameter opening was cut in the center of the mesh over the root excluder. All vegetation outside the root excluder was pulled away from the root excluder and pushed under the mesh, leaving no upright shoots to shade the target C tinctoria plant. The roots/no shoots (R/NS) treatment was established by removing only shoot competition. To mark the treatment location, a root excluder (2 cm deep X 10 cm diameter X 3 mm thick plastic PVC pipe) was hammered into the ground to the level of the surface soil at the center of each R/NS plot. Shoot competition was then removed by means of shoot excluders as described above. The no roots/shoots (NR/S) treatment was established by removing only root competition. A 20 cm root excluder was used as described above and no shoot excluder was used. The roots/shoots (R/S) treatment was estab- lished by allowing both root and shoot compe- tition. A 2 cm root excluder was used as described above with no shoot excluder. Coreopsis tinctoria seedlings were started from seed in 4 X 4 X 4 cm Jiffypot® (70% sphagnum peat moss and 30% wood fiber, Kristiansand, Norway) in a fiberglass greenhouse on the University Campus about 250 m from the field site. Five seeds were placed in each pot and covered with approx- imately two mm of soil. Any seeds germinating in excess of one were removed. After two weeks the seedlings in their associated Jiffypot® were trans- planted to the field plot and one was randomly placed into the center of each root excluder. Plots were watered daily or as needed (depending on weather conditions) over the course of the experi- ment to maintain the soil at field capacity. Since previous work at the study site was affected by the presence of rabbits, and rabbit droppings were visible, a wire mesh cage (10 cm diameter X 30 cm height; 1 cm X 1 cm grid size hardware cloth) was placed over all transplants to exclude rabbits and rodents (Van Auken and Bush 1997). Upon harvesting after 51-52 d, height (cm), number of flower buds, and number of flowers were measured or counted and recorded for each C. tinctoria plant. Shoots were then clipped at the soil surface and dried to a constant mass in a forced air oven at 75°C. Root excluders were removed from the soil and placed into a 167 L plastic trash barrel containing a mixture of j 1 5 liters tap water and 400 ml of sodium i hexametaphosphate ([NaPOsJe) to loosen the soil from the roots. After soaking in the mixture for 24 h, the root excluders were removed and soil ; was washed from the roots by rinsing with a gentle stream of tap water. Roots free of soil were ! blotted dry, wrapped in aluminum foil with holes for ventilation and dried to a constant mass in a forced air oven at 75°C. Roots were then ashed at ' 600°C for three hours (Bohm 1979) to determine i the mass of any adhering soil particles and ash- free root dry mass was calculated. Shoot, ash-free root, and total dry mass were 'i measured per plant for C. tinctoria. All remaining vegetation {B. curtipendula and low-density spe- cies, including Cirsium texanum and Helianthus annuus) in each 1 m X 1 m plot was clipped at the soil surface, stored in paper bags, and allowed to air dry to a constant mass in a fiberglass i greenhouse. Shoot dry mass was weighed to determine per plot dry mass and then mean dry mass was calculated per treatment. Bouteloua curtipendula shoot dry mass per plot was analyzed with a one-way ANOVA with compe- j tition treatments (NR/NS, R/NS, NR/S, and R/S) as the independent variable (Sail et al. 2012). Significant differences were not detected (P > 0.05). A multivariate one-way analysis of variance (MANOVA) was completed for all C. tinctoria measured parameters. When the MANOVA was significant (Pillai’s Trace, P < 0.05), separate one-way ANOVAs were performed with treat- ment (NR/NS, R/NS, NR/S, and R/S) as the independent variable. If significant differences were detected among any of the response variables and the treatment variable, a Scheffe Multiple Comparison test was employed for pairwise comparisons among individual treat- ments. In addition, a Shapiro-Wilk test was used to test for normality and Levene’s test was used to test for equal variance (Sail et al. 2012). Field simulated neighbor herbivory experiment. Coreopsis tinctoria survival in an established grassland was also examined in a field experiment as a function of B. curtipendula shoot height. The experiment was conducted at the same field site described above and was started in September 1997 and harvested after 61 d. Three levels of grass height were established in three, 0.6 m X 2.1 m (1.26 m^), study plots dominated by B. curtipendula with associated low-density species including C. texanum and H. annuus. Clipping or simulated herbivory of B. curtipendula shoot height was used as an alternative to density or biomass. Clipping has a negative overall effect on 2014] ELLIOTT AND VAN AUKEN: COMPETITION IN COREOPSIS 293 NR/NS R/NS NR/S R/S COM PEWION TREATMENT Fig. 1. Mean Coreopsis tinctoria height in cm (A), number of flower buds per plant (B), and the number of flowers per plant (C) as a function of competition treatment. Treatments were no roots/no shoots (NR/ NS), roots/no shoots (R/NS), no roots/shoots (NR/S), and roots/shoots (R/S) with 10 replications each. There were significant differences in all response variables (MANOVA, Pillia’s trace, P = 0.0001) and significant increases in all response variables in the absence of competition (NR/NS, one way ANOVAs, P < 0.01 for all). Different lower case letters indicate significant differences among treatments (Scheffe Multiple Com- parison Test) and lines are — 1 SD. plant mass. Treatments were no herbivory (0% clipping), partial herbivory (clipping and removal of 50% of the normal grass height), and maximum herbivory or total grass height and biomass removal (100% clipping). Bouteloua curtipendula shoot biomass in the study plots was estimated by clipping all vegetation at the soil surface in five, 0.5 m X 0.5 m, quadrats adjacent to the study plots. Average shoot dry mass per quadrat was determined after drying to a constant mass in a forced air oven at 75°C. Mean (±SD) B. curtipendula aboveground bio- mass was 476 ± 69 g/m^ in the no herbivory plots. Coreopsis tinctoria seedlings were started with excess seed in 4 X 4 X 4 cm Jiffypot® (70% sphagnum peat moss and 30% wood fiber, Kristiansand, Norway) in a fiberglass greenhouse as above. After two weeks seedlings were thinned to one plant per pot and transplanted into the field in their associated Jiffypot®, One hundred seedlings, five rows of 20 seedlings each, were planted 10 cm apart and 10 cm from the edge in each plot and marked to aid in locating the plants. Plots were watered daily or as needed (depending on weather conditions) over the course of the experiment to maintain soil field capacity. Every four days each plot was examined and seedling mortality was recorded. Coreopsis tinctoria survival in the three herbiv- ory treatments (0%, 50%, and 100% clipping) was compared using a two-tailed test. Significant differences in C. tinctoria survival were found among the three herbivory treat- ments; therefore, three one-tailed tests were completed to examine significant differences among individual treatments (Mendenhall and Beaver 1994). Results Field interspecific competition experiment. To- tal competition (aboveground shoot and below- ground root) and differences in aboveground and belowground competition were examined with C. tinctoria as the target species and B. curtipendula as the matrix or competitor species. Initial mean B. curtipendula aboveground (shoot) dry mass (± one SD) in the forty, 1 m X 1 m, study plots was 287.5 ± 42.5 g/m^ with no significant differences among treatments (One-way ANOVA, F = 1.61, P > 0.05). A MANOVA was used to do an overall comparison of all measured response variables in the field interspecific competition experiment. Pillai’s trace was significant (P < 0.0001, Figs. 1 and 2) indicating that competition from B. curtipendula significantly affected C tinctoria measured response variables. Coreopsis tinctoria mean height was significantly greater when plants were grown in the no competition treatment (NR/ 294 MADRONO [Vol. 61 COMPETITION TREATMENT COMPETITION TREATMENT NR/NS R/NS NR/S R/S COMPETITION TREATMENT Fig. 2. Mean Coreopsis tinctoria aboveground dry mass (A), ash free root dry mass (B), and total dry mass per plant (C) as a function of competition treatment. Treatments were no roots/no shoots (NR/NS), roots/no shoots (R/NS), no roots/shoots (NR/S), and roots/ shoots (R/S) with 10 replications each. There were significant differences in all response variables (MAN- OVA, Pillia’s trace, P = 0.0001) and significant increases in all response variables in the absence competition (NR/ NS, one way ANOVAs, P < 0.001 for all). For ash free root dry mass, when roots of the grass were removed (NR/NS or NR/S) there was an increase in C. tinctoria root dry mass. Different lower case letters indicate significant differences among treatments (Scheffe Mul- tiple Comparison Test) and lines are +1 SD. NS, Fig. lA, one-way ANOVA; F = 22.39, P < 0.0001, Scheffe Multiple Comparison Test). Mean height of C tinctoria decreased 40-58% from 53 cm in the no competition treatment to 22-32 cm with root or shoot competition, or when both root and shoot competition was present. Mean height was not significantly different among treatments with root or shoot competition or both roots and shoots present (R/ NS, NR/S, or R/S, P > 0.05, Scheffe Multiple Comparison Test) (Fig. lA). ; The mean number of flower buds per plant was significantly greater when C. tinctoria plants were grown with no competition (Fig. IB, one-way i ANOVA; F = 6.03, P < 0.01, Scheffe Multiple Comparison Test). The mean number of flower buds per plant decreased 67-99%, from approx- i imately three flower buds per plant in the no competition treatment to less than one flower bud per plant with root competition, shoot competition, or when both root and shoot competition was present. The mean number of flower buds per plant was not significantly different among treatments with root or shoot or both root and shoot competition present (P > 0.05, Scheffe Multiple Comparison Test). The | mean number of flowers per plant was signifi- cantly greater when plants were grown in the no competition treatment (Fig. 1C, one-way AN- OVA; F = 7.12, P < 0.001, Scheffe Multiple Comparison Test). The mean number of flowers per plant decreased 90-99% from approximately nine flowers per plant in the no competition treatment to less than one flower per plant with root competition, shoot competition, or when both root and shoot competition was present. The mean number of flowers per plant was not significantly different among treatments with root or shoot or both root and shoot competition present (P > 0.05, Scheffe Multiple Comparison Test). Coreopsis tinctoria mean aboveground or shoot dry mass per plant was significantly greater when plants were grown in the no competition treat- ment (Fig. 2A, one-way ANOVA, F = 73.49, P < 0.0001, Scheffe Multiple Comparison Test). Mean shoot dry mass per plant decreased 74- 90%, from 1.49 g in the no competition treatment, to 0.15-0.39 g when root competition, shoot competition, or when both root and shoot competition was present with no significant differences among them (P > 0,05, Scheffe Multiple Comparison Test). Significant differenc- es in C. tinctoria mean belowground or ash-free root dry mass per plant were also detected. Mean ash-free root dry mass per plant was greatest in the no root, no shoot competition treatment (0.36 g) and least in the root only competition treatment (R/NS, 0.07 g) (Fig. 2B, one-way ANOVA; F= 58.90, P < 0.0001, Scheffe Multiple Comparison Test). Significant reductions in mean 2014] ELLIOTT AND VAN AUKEN: COMPETITION IN COREOPSIS 295 PERCENT CUPPING Fig. 3. Percent survival of Coreopsis tinctoria seed- lings as a function of aboveground grass height reduction (clipping, 0%, 50%, 100%). Survival of C. tinctoria was significantly affected by clipping (two- tailed test, = 19.75, df = 2, P < 0.01). Survival was significantly greater in the 1 00% clipping treatment than the 50% treatment (one-tailed test, — 7.53, df=l, P < 0.01) and 0% treatment (one-tailed X^ test, X^ = 15.20, df = 1, P < 0.01). Different lower case letters indicate significant differences between survival in the clipping treatments (one-tailed X^ test). ash-free root dry mass occurred when grass roots were present (R/NS and R/S) compared to the no root treatments (NR/NS and NR/S), but there were no differences between either of the two treatments with roots or those two treatments without roots (P > 0.05, Scheffe Multiple Comparison Test). Results for mean total C. tinctoria dry mass per plant and mean shoot dry mass per plant were similar (Fig. 2C). Highest mean total dry mass was 1.85 g per plant in the NR/NS treatment. Total C tinctoria dry mass per plant decreased significantly with the presence of root or shoot or both root and shoot total grass dry mass per plant (Fig. 2C, one-way ANOVA, F = 58.89, P < 0.0001). However, C. tinctoria total dry mass per plant was not significantly different among treatments with root or shoot or both root and shoot grass dry mass present (P > 0.05, Scheffe Multiple Comparison Test). Mean C tinctoria total dry mass per plant was between 0.20 and 0.53 g when either grass roots or shoots or both were present, a reduction of 71-89%. Field simulated neighbor herbivory experiment. Coreopsis tinctoria survival as a function of aboveground grass dry mass or competition in an established grassland was also examined. Mean (± one SD) B. curtipendula aboveground or shoot dry mass per quadrat in the study plots (estimated in five, 0.5 m X 0.5 m, quadrats adjacent to the study plots) was 475.98 ± 58.67 g/m^. Simulated herbivory or clipping significant- ly affected C. tinctoria seedling survival in the field (Fig. 3, two-tailed test, = 19.75, df == 2, P < 0.01). Coreopsis tinctoria survival was 1% at 0% grass clipping or no aboveground grass removal and was not significantly different than survival at 50% grass clipping which resulted in 4% C tinctoria survival (one-tailed test, X^ = 2.66, df = 1, P > 0.05). However, C tinctoria survival at 100% grass clipping (zero grass aboveground dry mass) was significantly greater than at 50% clipping (one-tailed X^ test, X^ = 7.53, df = 1, P < 0,01) and 0% clipping (one- tailed test, = 15.20, df = 1, P < 0.01). Greatest survival after 61 d was 18% in the 100% simulated herbivory or grass clipping treatment (100% neighbor removal). Discussion Plants compete for aboveground and below- ground resources simultaneously and determin- ing which resources are more important has been difficult to resolve (Harper 1977; Grime 1979; Wilson and Keddy 1986; Tilman 1988; Grace and Tilman 1990; Keddy 2001). Many experimental grassland studies have shown that aboveground and belowground competition reduces dry mass in most if not all associated species, but aboveground dry mass may be more important in reducing community diversity (Lamb et al. 2009). Light levels are reduced more than 50% under the grassland matrix shading neighbors and limiting germination or growth of many herbaceous species in grassland communities (Bush and Van Auken 1987; Collins et al. 1998; Haag et al. 2004). Belowground biomass in grassland habitats can be two to four times higher than aboveground biomass and as much as 80% is found in the top 25 cm of soil (Risser et al. 1981; Bush and Van Auken 1991; Cahill 2003). As a result of the intense competition experi- enced in intact grasslands, species with poor competitive abilities are sometimes restricted to unproductive sites where competition is minimal (Cahill and Casper 2002). Grime (1977) suggested that a trade-off exists between a species compet- itive ability and its stress tolerance. Habitat specificity of Coreopsis lanceolata appears to be related to both its tolerance of low soil nutrients and poor competitive ability, supporting the theory of a trade-off (Grime 1977). With neighbors removed C. lanceolata was capable of growth and survival on both productive sites and stressful sites with reduced nutrient soils. Mor- phological traits, growth rates, survivorship, and reproduction of C. lanceolata were all reduced with competition (Folgate and Scheiner 1992) and it was restricted to vegetation gaps or non- productive habitats. It appears that distribution of Coreopsis tinctoria is restricted to disturbed habitats or to early season growth as a result of the spatial and temporal competition it encounters in established 296 MADRONO [Vol. 61 grasslands. In the field, C tinctoria growth requirements or niche requirements probably limit it to large or small disturbances in grasslands communities, as was shown for C. lanceolata (Folgate and Scheiner 1992). After a disturbance such as grazing, fire, or a large scale, multi-year drought C4 grasses are reduced (Barbour et al. 1999). Annuals with seeds in the seed bank, like C tinctoria, exploit the lack of competition occurring until the C4 grasses recover their former dominant community posi- tion. During the early growth period, soil nutrients from decomposition are higher and available for uptake by early season C3 herba- ceous annuals like C. tinctoria, while the C4 grasses are still mostly dormant (McKinley and Blair 2008). Coreopsis tinctoria seems to be tolerant of lower temperatures than associated C4 grasses and grows well both in cooler latitudes and early in the growing season (Cahill 2003), but not during the hot, dry, southwestern summers (Enquist 1987). Coreopsis tinctoria seems to be able to grow in a southern grassland matrix in the absence of gaps by establishing early in the growing season (Eddy 2013), before the C4 grasses start their growth (Cahill 2003). Results from field experiments presented here lead to the same conclusions, C. tinctoria is a poor competitor and only found in disturbed sites or in gaps, probably in the early southern growing season or after droughts (Figs. 1 and 2). When C tinctoria was examined in plots that were clipped to simulate grass herbivory (Fig. 3), survival was greatest in plots with all above- ground grass removed. When C. tinctoria was grown with B. curtipendula in an experiment to examine the importance of aboveground, below- ground, and total grass biomass, almost all of C tinctoria response variables increased with almost all grass reduction treatments. Aboveground and total grass dry mass appeared to be equally detrimental to C tinctoria’ s response variables (Figs. 1 and 2). However, belowground dry mass of B. curtipendula in any combination reduced the belowground dry mass of C. tinctoria. Species restricted to stressful, unproductive sites are not necessarily excluded under all conditions. With grassland disturbances such as grazing, mowing, fire, and extended drought, gaps are opened in the matrix and offer a mechanism of entry into productive established communities to species with reduced competitive abilities. Both aboveground and belowground competition from dominant matrix species is reduced in gaps or patches, making establishment and growth possible for poor competitors (Cahill and Casper 2002). Grazing and mowing have been shown to create aboveground and below- ground gaps (Bush and Van Auken 1987; Collins et al. 1998; Haag et al. 2004), increasing surface light levels by 100% and decreasing belowground biomass of dominant matrix species (Collins 1987; Van Auken and Bush 1989, 1990; Bush and Van Auken 1991; Collins et al. 1998; Cahill and Casper 2002). High species diversity in C4 grasslands is characterized by large numbers of C3 forbs. Seeds of many of these C3 forbs remain dormant in the seed bank until either one or a combination of these disturbances open a gap in the grassland matrix that allows their germination and estab- lishment (Collins et al. 1998). This mechanism of entry into established grasslands by utilizing gaps as sites of establishment is apparently also an important mechanism used by early successional species (Collins 1987; Van Auken and Bush 1989; Bush and Van Auken 1991, 1995; Bond 2008; Chaneton et al. 2012; Mazia et al. 2013). Previous unpublished greenhouse studies with Coreopsis tinctoria (Elliott 1999) have shown that both aboveground and belowground grass dry mass reduce the dry mass of C tinctoria, suggesting that grass competition or a lack of competition was important for the success of C tinctoria. Mortality of C. tinctoria was approxi- mately 70% in a northern C3 grassland, with neighboring grass roots apparently inconsequen- tial, but biomass of C tinctoria was reduced by more than 80% in the presence of the grass roots (Cahill 2003), and effects of insect herbivory was mixed (Haag et al. 2004). In addition to gaps, the timing of growth of two potentially competing species from the same habitat seems to be important in determining their local distribution and density (Eddy 2013). A C4 grass like B. curtipendula has high water and nitrogen use efficiencies, high light and temperature require- ments, and grows best late in the growing season, especially in hot southern grasslands (Fay et al. 2003; Weatherford and Myster 201 1). A target C3 woody plant {Prosopis glandulosa Torr.) grown in competition with B. curtipendula had 1 0-20 times more total dry mass when planted two months before the C4 grass but was suppressed 99.9% if planted with the grass but two months after the grass was started (Bush and Van Auken 1991). Coreopsis tinctoria density is low in the matrix of southwestern grasslands and possibly missed in cursory surveys, especially after the main C4 grasses start their growth, and C. tinctoria has completed its life cycle (Correll and Johnston 1979). High densities of C. tinctoria have only been noted in relatively large displays, seemingly in large gaps or disturbances where the C4 grasses are reduced or early in the growing season (Elliott and Van Auken, personal observation). Differ- ential timing of growth, intense competition from associated species, inherent lack of competitive ability, and tolerance of low nutrient soils all result in local variability in density of C. tinctoria despite its widespread distribution. 2014] ELLIOTT AND VAN AUKEN: COMPETITION IN COREOPSIS 297 Conclusions In the field, Coreopsis tinctoria grows poorly or not at all with B. curtipendula, a perennial C4 southern grass. Coreopsis tinctoria appears to be a species that requires large or small disturbances or gaps or it grows early in the growing season before the C4 grasses become active at higher temperatures. When found in native communi- ties, it seems to be a good indicator of disturbances. These disturbances are caused by man, herbivory, burrowing of small mammals, fire, or drought, or combinations of these factors. Acknowledgments We thank J. Jackson, S. Engelken, B. Snyder, V. V. Jackson, C. Rellstab, J. K. Bush, D. Jaffe, R. Le Baron and J. McAdams for their help at various times and in various ways during this study and the preparation of this manuscript. Literature Cited Barbour, M. G., J. H. Burk, W. D. Pitts, F. S. Gilliam, and M. W. Schwartz. 1999. Terrestrial plant ecology, 3rd ed. Benjamin Cummings, Menlo Park, CA. Baskin, C. C. and J. M. Baskin. 2001. Seeds; ecology, biogeography, and evolution of dormancy and germination. Academic Press, San Diego, CA. Begon, M., C. R. Townsend, and J. L. Harper. 2006. Ecology: from individuals to ecosystems. Blackwell Publishing, Malden, MA. Bertness, M. D. and G. H. Leonard. 1997. The role of positive interactions in communities: lessons from intertidal habitats. Ecology 78:1976-1989. Bohm, W. 1979. Methods of studying root systems. Springer-Verlag, New York, NY. Bond, W. J. 2008. What limits trees in C4 grasslands and savannas? Annual review of Ecology, Evolu- tion, and Systematics 39:641-659. Busch, M., C. Knight, N. Mazia, K. Hodara, E. Muschetto, and E. j. Chaneton. 2012. Rodent seed predation on tree invader species in grassland habitats of the Inland Pampa. Ecological Research 27:369-376. Bush, J. K. and O. W. Van Auken. 1986. Light requirements of Acacia smallii and Celtis laevigata in relation to secondary succession on floodplains of south Texas. American Midland Naturalist 115:118-122. . 1987. Light requirements for growth of Prosopis glandulosa seedlings. Southwestern Natu- ralist 32:469^73. . 1991. Importance of time of germination and soil depth on growth of Prosopis glandulosa (Leguminosae) seedlings in the presence of a C4 grass. American Journal of Botany 78:1732-1739. . 1995. Woody plant growth related to planting time and clipping of a C4 grass. Ecology 76:1603-1609. . 2010. Competition between Schizachyrium scoparium and Buchloe dactyloides: the role of soil nutrients. Journal of Arid Environments 74:49-53. Cahill, J. F. Jr. 2003. Lack of relationship between below-ground competition and allocation to roots in 10 grassland species. Journal of Ecology 91:532-540. and B. B. Casper. 2002. Canopy gaps are sites of reduced belowground plant competition in a productive old field. Plant Ecology 164:29-36. Callaway, R. M. 1995. Positive interactions among plants. Botanical review 61:306-349. and L, R. Walker. 1997. Competition and facilitation: a synthetic approach to interactions in plant communities. Ecology 78:1958-1965. Chaneton, E. J., N. Mazia, W. B. Batista, A. G. Rolhauser, and C. M. Ghersa. 2012. Woody plant invasions in Pampa grasslands; a biogeo- graphical and community assembly perspective. Pp. 115-144 in Randall. W. Myster (ed.), Ecotones between forest and grassland. Springer, New York, NY. Cohn, E. J., O. W. Van Auken, and J. K. Bush. 1989. Competitive interactions between Cynodon dactylon and Acacia smallii seedlings at different nutrient levels. American Midland Naturalist 121:265-272. Collins, S. L. 1987. Interactions of disturbances in tallgrass prairie: a field experiment. Ecology 68:1243-1250. , A. K. Knapp, J. M. Briggs, J. M. Blair, and E. M. Steinauer. 1998. Modulation of diversity by grazing and mowing in native tallgrass prairie. Science 280:745-747. Connell, J. A. 1983. On the prevalence and relative importance of interspecific competition; evidence from field experiments. American Naturalist 122:661-696. CORRELL, D. S. AND M. C. JOHNSTON. 1979. Manual of the vascular plants of Texas. Texas Research Foundation, Renner, TX. Eddy, K. C. 2013. Are temporal gaps necessary for the establishment of the native wildflower, Coreopsis tinctorial M.S. thesis. The University of Texas at San Antonio, San Antonio, TX. Elliott, S. A. M. 1999. Coreopsis tinctoria: germina- tion requirements and competitive ability with a perennial grass. M.S. thesis. The University of Texas at San Antonio, San Antonio, TX. Enquist, M. 1987. Wildflowers of the Texas Hill Country. Lone Star Botanical, Austin, TX. Fay, P. a., A. K. Knapp, J. M. Blair, J. Carlisle, D. B. T. Danner, and J. K. McCarron. 2003. Rainfall timing, soil moisture dynamics, and plant responses in a mesic tallgrass prairie ecosystem. Pp. 147-163 in J. F. Weltzin and G. R. McPherson, (eds.), Precipitation regimes and terresterial ecosystems: a North American Perspective. Uni- versity of Arizona Press, Tucson, AZ. Folgate, L. a. and S. M. Scheiner. 1992. Distribu- tion of a restricted locally abundant species; effects of competition and nutrients on Coreopsis lanceo- lata. American Midland Naturalist 128:254-269. Gerry, A. K. and S. D. Wilson. 1995. The influence of initial size on the competitive response of six plant species. Ecology 76:272-279. Grace, J. B. and D. Tilman. 1990. Perspectives on plant competition. Academic Press, San Diego, CA. Grellier, S., S. Barot, J. L. Janeau, and D. Ward. 2012. Grass competition is more important than seed ingestion by livestock for Acacia recruitment in South Africa. Plant Ecology 213:899-908. 298 MADRONO Grime, J. P. 1977. Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. American Naturalist 111:1169-1194. . 1979. Plant strategies and vegetation processes. John Wiley & Sons, Inc., Chichester. Gurevitch, J. 1986. Competition and the local distribution of the grass Stipa neomexicana. Ecol- ogy 67:46-57. Haag, J. J., M. D. Coupe, and J. F. Cahill Jr. 2004, Antagonistic interactions between competition and insect herbivory on plant growth. Journal of Ecology 92:156-167. Harper, J. L. 1977. Population biology of plants. Academic Press, New York, NY. Keddy, P. A. 2001. Competition, 2nd ed. Kluwer Academic Publishers, Dordrecht. Lamb, E. G., S. W. Kembel, and J. F. Cahill Jr. 2009. Shoot, but not root, competition reduces community diversity in experimental mesocosms. Journal of Ecology 97:155-163. Mazia, N., P. M. Tognetti, and E. D. Cirino. 2013. Patch identity and the spatial heterogenity of woody encroachment in exotic-dominated old-field grasslands. Plant Ecology 214:255-266. McEvoy, P. B., N. T. Rudd, C. S. Cox, and M. Huso. 1993. Disturbance, competition, and her- bivory effects on ragwort Senecio jacobaea popu- lations. Ecological Monographs 63:55-75. McKinley, D. C. and J. M. Blair. 2008. Woody plant encroachment by Juniperus virginiana in a mesic native grassland promotes rapid carbon and nitrogen accrual. Ecosystems 11:454-468. McPhee, C. S. and L. W. Aarssen. 2001. The separation of above- and below-ground competi- tion in plants. A review and critique of methodol- ogy. Plant Ecology 152:119-136. Mendenhall, W. and R. J. Beaver. 1994. Introduc- tion to probability and statistics, 9th ed. Duxbury Press, Belmont, CA. Pearson, D. E., R. M. Callaway, and J. L. Maron. 2011. Biotic resistance via granivory: establishment by invasive, naturalized, and native asters reflects generalist preference. Ecology 92:1748-1757. Pickett, S. T. A. and P. S. White. 1985. The ecology of natural disturbances and patch dynamics. Academic Press, Orlando, FL. Risser, P. G., E. C. Birney, H. K. Blocker, S. W. May, W. j. Parton, and J. A. Wiens. 1981. The True Prairie ecosystem. Hutchinson Ross Publica- tion Company, Stroudsburg, PA. Sale, J., A. Lehman, M. Stephens, and L. Creighton. 2012. JMP start statistics: a guide to [VoL 61 j statistics and data analysis using JMP. SAS [ Institute Inc., Cary, NC. j| SCHOENER, T. W. 1983. Field experiments on interspe- jj cific competition. American Naturalist 122:242- 285. l! Smith, T. M. and R. L. Smith. 2012. Elements of ! Ecology, 8th ed. Benjamin Cummings, New York, ! NY. ' Strother, J. L. 2006. Coreopsis tinctoria. Pp. 185, 194, | 197-198 in Flora of North America Editorial i Committee (eds.), Flora of North America North of Mexico, Vol 21: Asteraceae. Oxford University Press, New York, NY. Taylor, F. B., R. B. Hailey, and D. L. Richmond, | 1966, Soil survey of Bexar County, Texas. United j States Department of Agriculture, Soil Conserva- tion Service, Washington, D.C. Tilman, D. 1988. Plant strategies and the dynamics and structure of plant communities. Princeton University Press, Princeton, NJ. Van Auken, O. W. 2000. Characteristics of inter- canopy bare patches in Juniperus woodlands of the ■ southern Edwards Plateau, Texas. Southwestern j Naturalist 45:95-1 10. AND J. K. Bush. 1989. Prosopis glandulosa growth: influence of nutrients and simulated grazing of Bouteloua curtipendula. Ecology 70: 512-516. . 1990. Interaction of two C3 and C4 grasses with seedlings of Acacia smallii and Celtis laevigata. Southwestern Naturalist 35:31C-32L . 1997. Growth of Prosopis glandulosa in response to changes in aboveground and below- ground interference. Ecology 78:1222-1229. and R. j. Lohstroh. 1990. Importance of canopy position for the growth of Celtis laevigata seedlings. Texas Journal of Science 42:83-89. Weatherford, J. L. and R. W. Myster. 2011. Interactive effects of species, simulated grazing, and below-ground resources on competitive out- come among three prairie grasses. Journal of the Torrey Botanical Society 138:107-119. White, J. J. and O. W. Van Auken. 1996. Germina- tion, light requirements, and competitive interac- tions of Stipa ieucotricha (Gramineae). Southwest- ern Naturalist 41:27-34. Wilson, S. D. and P. A. Keddy. 1986. Species competitive ability and position along a natural stress/disturbance gradient. Ecology 67:1236-1242. Wilson, W. G. and R. M. Nisbet. 1997. Cooperation and competition along smooth environmental gradients. Ecology 78:2004-2017. Madrono, Voi. 61, No= 3, pp. 299-307, 2014 CAREX XEROPHILA (CYPERACEAE), A NEW SEDGE FROM THE CHAPARRAL OF NORTHERN CALIFORNIA Peter F. Zika WTU Herbarium, Box 355325, University of Washington, Seattle, WA 98195-5325 Zikap@comcast.net Lawrence P. Janeway The Chico State Herbarium, California State University, Chico, CA 95929-0515, and Feather River Ranger District, Plumas National Forest, 875 Mitchell Avenue, Oroville, CA 95965 Barbara L. Wilson Carex Working Group, 1377 NW Alta Vista Drive, Corvallis, OR 97330 Abstract Carex xerophila Janeway & Zika is described from gabbro and serpentine soils on the west slope of the northern Sierra Nevada in California. It is documented from four populations in or on the margins of chaparral and open forest. The new species is assigned to Carex section Acrocystis Dumort,, and a key is provided for all California representatives of the section. Carex xerophila differs from C globosa Boott in its uniformly short, erect basal peduncles and Sierra foothills habitats. Carex xerophila differs from C brainerdii Mack, by its green leaves that are not densely papillose below, its relatively shorter perigynium beaks, and its lower elevation montane habitats. Carex xerophila has more strongly nerved scales and perigynium faces than C rossii Boott. Key Words: California, Carex section Acrocystis, chaparral sedge, gabbro, Pine Hill, ultramafic. There are 140 indigenous species of Carex known from California, which includes 13 endemics restricted to the state (Zika 2012; Zika et ai. 2012, 2013). In California, there are seven representatives of Carex sect. Acrocystis Du- mort., united by their short spikes, plump pubescent perigynia, and an arillate perigynium stipe that promotes dispersal by ants. These are C brainerdii Mack., C. brevicaulis Mack., C deflexa Homem. var. boottii L. H, Bailey, C globosa Boott, C. inops L. H. Bailey subsp. inops, C. rossii Boott, and C serpenticola Zika. The subtle distinctions between species of Carex section Acrocystis are slow to be untangled, and continue to yield surprises and novelties in North America (Zika et al. 1998; Werier 2006; Sorrie et al. 2011). Among California specimens of Carex sect. Acrocystis are some puzzling populations from the northern Sierra Nevada (Janeway 1992; Oswald 2002). After studying these sedges in the field, we describe them here as a new species. Taxonomic Section Carex xerophila Janeway & Zika, sp. nov. (Figs. 1-3).— TYPE: USA, California, El Dor- ado Co., Bureau of Land Management Cam- eron Park Unit, Pine Hill Preserve, N of Route 50, 3.3 air km W of reservoir on Sawmill Creek, Shingle Springs, 450 m, 17 May 2012, P, F. Zika 25874 and L. P. Janeway (holotype: WTU; isotypes: CAS, CHSC, GH, JEPS, MICH, OSC, RSA, US). Species propria, differt a Carex rossii glumis perigyniisque plus plurinervosis, a C brainerdii foliis juvenilibus gramineo-viridibus epapillosis statim dignoscenda, ceteram pedunculis basalibus brevioribus rigentibus erectis a C globosa recedens. Plants densely to loosely cespitose, forming tufts up to one m in diam. Proximal sheaths scaly, bladeless, red to dark purple, smooth to scabrous on the prominent nerves, veins sometimes per- sisting and ladder-fibrillose. Distal sheaths with translucent faces, often red-streaked, apices concave; ligules obtuse, mostly longer than wide, margins scabrous; distal sheaths bearing blades. Leaves in numerous sterile rosettes, 4-10 on proximal 1/4 of stem, some blades taller than fertile stems, blades 6-42 mm long, 1.7-3. 8 mm wide, V-shaped in cross-section, often folded, green not blue-green, 5-10 prominent veins, scabrous adaxially, blade margins scabrous, tips filiform- triangular, densely scabrous, apex blunt or truncate, bristly (20 X), Stems (culms) 9.5- 35 cm tall, triangular, scabrous. Inflorescences with 1-3 basal spikes 7-11 mm long, on erect filiform scabrous peduncles 1-7.5 cm long, 0,3- 0.4 mm wide (Fig. 3A), the basal spikes with 2-5 perigynia, mostly pistillate, occasionally androgynous with 2-3 terminal staminate scales; non-basal portion of the inflorescences on elongate stems comprised of 1-4 spikes, the spikes 13-38 mm long, overtopped by foliage; 300 MADRONO [Vol. 61 Fig. 1 . Carex xerophila (A-F, H-M all from the type, Zika 25874 and Janeway. G from Zika 25871 and Janeway). A. Habit, with foliage overtopping fertile shoots. B. Detail of shoot base, showing basal scale, and ladder-fibrillose sheath, with detail of scabrous veins. C. Sharply triangular and scabrous leaf tip. D. Pistillate scale with scabrous mid-veins. E. Staminate scale, with scabrous mid-veins. F. Trigonous achene, with detail of papillose surface. G. (from Zika 25871 and Janeway) Fresh perigynium, side view, showing curved beak; base pale, plump, and arillate. H, Dried perigynium, front view, showing shallow teeth at tip of beak and dried base forming a stipe. I. Dried 2014] ZIKA ET AL.: CAREX XEROPHILA 301 Fig. 2. Carex xerophila {Zika 25871 and Janeway) showing arillate base on fresh perigynia, presumably enhancing dispersal by ants. The perigynium nerves are difficult to see until the perigynia are dried, at this scale. Scale bar = 3 mm. lowest non-basal inflorescence bract often emar- ginate and purple at base, blade green, leafy, 8- 45 mm long, often shorter than inflorescence; lowest non-basal lateral spikes pistillate, each with 1-5 perigynia (Fig. 3B), sessile or on peduncles to 5 mm long; distal spike staminate, 12-21 mrn long, 2. 0-3.0 mm wide, 8-17 flowered; stamens 3 per scale, filaments white, anthers yellow, linear, 2. 7^.2 mm long, apical appendage 0.1 mm long, bristly (20x). Staminate scales 4.8- 7.1 mm long, dark purple, sometimes scabrous or sparsely pubescent, with prominent or keeled pale midveins, and 2-15 usually inconspicuous fine parallel lateral veins, apex acute, occasionally rnucronate, mucro to 0.3 mm, often bristly (20x). Pistillate scales ovate to broadly ovate, clasping at base, 3.2-4-.9 mm long, 2. 8^.0 mm wide, glabrous, scabrous, or sparsely pubescent, green central band with (l-)3 strong midveins, the midveins often keeled and scabrous, the broad dark purple lateral bands with 6-12 inconspicu- ous fine parallel veins, distal margins entire to sparsely ciliate, apex acute, or emarginate and rnucronate, the mucro to 0.7 mm and scabrous or ciliate. Perigyeia obovate, pubescent, 3. 4-4.9 mm long, 1.4— 2.1 mm wide; stipe fleshy, oily, turgid and pale when fresh (Fig. 2) then shriveling, 0.9- 1.9 mm long; body plump, rounded-trigonous, usually with 10-16 prominent veins when mature; beak distinct, slightly flattened, erect or slightly incurved, scabrous and pubescent, measured from inflection point 0. 5-0.9 mm long, includ- ing teeth 0. 1-0.4 mm long. Acheees obovoid, Fig. 3. Carex xerophila {Zika 25872 and Janeway). A. Erect basal pistillate spikes. B. Elongate stem and inflorescence, with lateral pistillate spike and terminal staminate spike. rounded-trigonous, papillose (20 X), brown, 2.0- 2.5 mm long, 1.4— 2.0 mm wide. Paratypes: USA, CALIFORNIA, Butte Co.: Magalia Serpentine, E side of Coutolenc Road, 707 m, 14 May 2006, L. P. Janeway 8595 (CHSC, JEPS, MICH, NY, OSC, TRT, WTU); Same site, over mature, 22 Nov 2009, L. P. Janeway 9968 (CHSC); Same site, 29 Apr 1989, B. Castro 307 and G. Kuenster (CAS, CHSC, WS); Same site, 8 Jul 1996, a F. Hrusa 13106 and F D. Wilfred (HSC); Same site, 16 Apr 1984, V. Oswald 1159 (CHSC); Same site, 6 May 1978, M A Taylor 1578 (CHSC, MO); Same site, 16 May 2012, P. F. Zika 25868 and L. P. Janeway (BRY, CDA, CHSC, DAY, MO, OSC, RSA, WTU); Isolated block of serpentine between the E end of Hollywood Drive, Magalia, and the upper end of Magalia Reservoir, 29 May 2005, L. P. Janeway 8397 (CAS, CHSC, GH, JEPS, MO, NY, OSC, TRT, US, WTU); Same site, 4 Jun 1989, G. Kuenster SM. (CAS, CHSC, WS); Same site, 7 May 1997, G. F. Hrusa 13726 and T D. Wilfred (CDA, DAY); Same site, 8 May 1997, G. F. Hrusa 13761 and T D. Wilfred (CDA). El Dorado Co,: Cameron Park Unit, Pine Hill Preserve, 1.3 km W of junction of Hwy 50 and Mother Lode Drive, 457 m, 10 May 2006, L. P. Janeway 8589 (CHSC, HSC, OSC, SBBG, TRT, perigynium, top view, rounded-trigonous. J. Basal spike and peduncle. K. Non-basal inflorescence, with two lateral pistillate spikes and terminal staminate spike. L. Notched sheath mouth on inflorescence bract (non-basal inflorescence). M. Sheath mouth and ligule, stem removed from view. 302 MADRONO [Vol. 61 Fig. 4. Carex xerophila distribution map based on herbarium specimens. County abbreviations: A = Amador; B = Butte; Ca = Calaveras; Co = Colusa; E = El Dorado; G = Glenn; N = Nevada; Pla = Plaeer; Plu = Plumas; Sac = Sacramento; San = San Joaquin; Si = Sierra; Su = Sutter; T = Tehama; Yo = Yolo; Yu = Yuba. UCR, WTU); Same site, 12 Jun 2010, L. P. Janeway 9995 and B. Castro (CHSC, OSC, WTU); Same site, 29 Mar 1972, D. W. Taylor 1378 (DAV). Nevada Co.: South Ponderosa Road, W of Grass Valley, 670 m, 19 Apr 2008, K. /. Callahan 8 (CHSC, JEPS, OSC, TRT, WTU); E slope of American Ranch Hill, ca. 5.6 km SW of Grass Valley, 685 m, 31 May 2010, L. P. Janeway 9985 and C. Brinkhurst, B. Castro, K. Callahan, and Bill Wilson (CHSC, MICH, OSC, WTU); Same site, 1 May 1969, G. H. True 4900 (CAS); Same site, 11 May 1973, G. H. True 7465 and J. T Howell (CAS); Highway 20 roadside, W of Grass Valley, 762 m, 20 Apr 2007, D. G. Kelch 7.167 (CDA, CHSC); Osceola Ridge ca. 0.6 km N of Highway 20, along | Pipeline Road, ca. 4.8 km W of Grass Valley, 720 m, 31 May 2010, L. P. Janeway 9981 and i K. Callahan, C. Brinkhurst, B. Castro, and Bill Wilson (CHSC, MICH, OSC, WTU); Same site, ' 735 m, 16 May 2012, P. F. Zika 25872 and L. P. ' Janeway (CAS, CHSC, HSC, JEPS, OSC, SD, WTU). Yuba Co.: Forsythe Road, 2.7 road km S | of New York House Road, above Prince Albert Creek, 604 m, 24 May 2006, L. P. Janeway 8626 , (CHSC, DAV, JEPS, MICH, OSC, TRT, US, WTU); Ponderosa Way, 1.3 km north of La | Porte Road and Brownsville, 686 m, 16 May ! 2012, P. F. Zika 25870 and L. P. Janeway (CAS, i CHSC, GH, OSC, WTU); Ponderosa Way 1 .3 km ' SE of Robinson Mill Road and Forbestown Road, 771 m, 24 May 2006, L. P. Janeway 8616 \ (CHSC, HSC, MICH, OSC, RSA, TRT, US, WTU); Chaparral adjacent to Brownsville dump, just E of junction of Jiggs Road and Ponderosa Way, 3 air km W of Ruff Hill, Brownsville, 715 m, i 16 May 2012, P. F. Zika 25871 and L. P. Janeway \ (CHSC, JEPS, MO, OSC, RM, SBBG, UCR, ' WTU). I ! Distribution, Habitat, and Ecology j Carex xerophila is recorded from Butte, El j Dorado, Nevada, and Yuba counties in Califor- i nia (Fig. 4). We know of four population centers, | at elevations of 450-770 m, in the northern Sierra j Nevada Mountains. Plants in Butte Co. are on peridotite (serpentine) bedrock and in mixed I conifer forest of the High Sierra Nevada district (Baldwin et al. 2012). The other populations are essentially in the transition from Sierra Nevada Foothills to High Sierra Nevada district (Baldwin et al. 2012), and occur on gabbro-derived soils (Alexander 2011; Burge and Manos 2011; Alex- ander 2012). The plants grow in full sun to partial shade, on dry soils, in open forest, scrub, at the edge of thickets, and in chaparral (Figs. 5, 6), often with or near Hesperocyparis macnabiana (A. Murray) Bartel. In some dense stands of chaparral C. xerophila can dominate the under- story (Fig. 6C). At the Pine Hill site, a wildfire in July 2007 removed the woody vegetation on a gentle northerly slope. Carex xerophila is spar- ingly present on the adjacent shrubby ridgeline, but absent in the recently burned ground. Notable associates of Carex xerophila include: Adenostoma fasciculatum Hook. & Arn., Arbutus nienziesii Pursh, Arctostaphylos viscida Parry, Calochortus monophyllus (Lindl.) Lem., Calyste- gia stebbinsii Brummitt, Ceanothus cuneatus (Hook.) Nutt., C. lemmonii Parry, C roderickii W. Knight, Cercis occidentalis Torr. ex A. Gray, Danthonia unispicata (Thurb.) Munro ex Ma- coun, Eriodictyon californicum (Hook. & Arn.) Torr., Frangula californica (Eschsch.) A. Gray subsp. tomentella (Benth.) Kartesz & Gandhi, 2014] ZIKA ET AL.: CAREX XEROPHILA 303 Fig. 5. Carex xerophila habit, both from El Dorado Co. {Zika 25874 and Janeway). A. A large tussock. B. Another large tussock in chaparral, with Lawrence Janeway. Hesperocyparis macnabiana, Iris hartwegii Baker, Lomatium marginatum (Benth.) J. M. Coult. & Rose, Melica torreyana Scribn., Packera layneae (Greene) W. A. Weber & A. Love, Perideridia bacigaiupii T. I. Chuang & Constance, Picker- ingia montana Nutt., Pinus ponderosa Lawson & C. Lawson, P. sabiniana Douglas ex D. Don, Quercus chrysolepis Liebm., Q. garryana Douglas ex Hook. var. semota Jeps., Q. keiioggii Newb., Q. wislizeni A. DC., Rhamnus ilicifoUa Kellogg, Salvia sonomensis Greene, Sanicula bipinnatifida Douglas, Sisyrinchium helium S. Watson, Trite- leia hyacinthina (Lindl.) Greene, and Wyethia reticulata Greene. Fig. 6. Carex xerophila habitat. A. Nevada Co. {Zika 25872 and Janeway). B. Yuba Co. {Zika 25871 and Janeway). C. Chaparral sedge dominates under a canopy of Arctostaphylos viscida in Yuba Co. {Zika 25871 and Janeway). D. El Dorado Co. {Zika 25874 and Janeway). 304 MADRONO [Vol. 61 Phenology Car ex xerophila fruits mature from late March to early June. Etymology Carex xerophila is named for the xeric sites it inhabits. We propose the common name chapar- ral sedge. Discussion Among the California representatives of Carex sect. Aero cyst is, multi-nerved perigynia are found only in C xerophila, C. globosa and C hrainerdii. In the field, basal spikes of C xerophila are held on stiffly erect and short peduncles, always less than eight cm long. Carex globosa differs in its weak, and ultimately arching, elongate basal peduncles. Although some C. globosa basal peduncles may be short, some spikes are borne on peduncles that are more than 10 cm long and lax at maturity, presenting the arillate perigynia to dispersers on the ground. Although none of these features is unique, in combination one can often readily distinguish C. xerophila by its tendency to have narrower leaves and pistillate scales that are more commonly scabrous, a perigynium body that is never globose, and shorter perigynium beaks (Table 1), as well as a much denser growth form and a different habitat specialization. Where C. globosa forms loose tall clumps in mesic or damp forest and openings along the coast, C xerophila creates denser shorter tufts in xeric situations in the foothills of the northern Sierra Nevada. Carex brainerdii differs from all other Aero- cystis in California with its young foliage that is blue-green and papillose on the lower surface. The blades retain their color with age, though the papillae may wear off the older leaves. In contrast, Carex xerophila foliage is green and scabrous, not blue-green or papillose. Carex rossii and C deflexa var. boottii are similar in habit to C xerophila. All three form rather dense tufts of numerous vegetative shoots. Carex rossii and C. deflexa var. boottii both have perigynia that lack veins on the faces, and their pistillate scales typically have a single vein, not three as usually seen in C xerophila. In addition, both grow at higher elevations than C xerophila, as discussed below. Three species of Carex sect. Acrocystis are found in dry habitats along the western slope of the northern Sierra Nevada: Carex rossii, C, hrainerdii, and C xerophila. Carex rossii is found at higher elevations, in mixed conifer forest and above, while C hrainerdii grows at middle elevations, overlapping some with C. rossii. Carex xerophila is usually found at lower elevations and further west than C brainerdii species, only Carex xerophila is found in the xeric chaparral habitat and the adjacent open forests, and is tolerant of the severe constraints created by gabbro and serpentine soils. Some Carex sect. Acrocystis from southern California appear to be related to this small group of species with multi-nerved perigynia. A few populations from the San Bernardino and San Gabriel Mountains may represent an unde- scribed taxon, and are tentatively separated in the key below as Carex sp. “A.” They resemble C. xerophila, but appear to differ in scale and perigynium characters. More study and better specimens are needed to confirm the consistency of these morphological differences. Carex geophila Mack, and C pityophila Mack, were reported from California (Crins and Rettig 2002; Taylor 2010), but these reports appear to be based on depauperate specimens of C. rossii and C. deflexa var. boottii (Zika et al. 2012). Both C geophila and C pityophila have essentially vein- less perigynium faces (Crins and Rettig 2002), as well as single-nerved pistillate scales, and thus would not be confused with C. xerophila. Carex geophila ranges from Arizona to Texas, north to Colorado, and south to Guatemala; C pityophila is restricted to Utah, Colorado, and New Mexico (Crins and Rettig 2002). Conservation Implications Much of the chaparral habitat on the lower slopes of the Sierra Nevada has been altered by fire suppression, development, agriculture, grazing, and off-road vehicles. This is true near all known localities for Carex xerophila. More field surveys are needed to search for additional populations, and to determine population sizes and trends. However, based on our preliminary investiga- tions, it appears C. xerophila is an uncommon plant in a declining habitat, and in need of conservation attention. At the Pine Hill Preserve site, a number of rare plants are found in the remnants of the chaparral plant community (Wilson 1986; Hunter and Horenstein 1991; Hinshaw 2008; Wilson et al. 2009), including five federally threat- ened or endangered taxa: Calystegia stebbinsii, Ceanothus roderickii, Fremontodendron decumbens R. M. Lloyd, Galium californicum Hook. & Arn. subsp. sierrae Dempster & Stebbins, and Packera layneae (USFWS 1996). Packera layneae also occurs with Carex xerophila at some of the Brownsville sites, and Calystegia stebbinsii is in the chaparral near the Osceola Ridge sites. Federal listing of these chaparral taxa in the northern Sierra Nevada emphasizes that Carex xerophila habitat is sharply reduced and threatened. 2014] ZIKA ET AL.: CAREX XEROPHILA 3 ^ M cti f4' « .9 S 2> :5 s -2 <„ « o .p^ ^ w .9 d P 1 = « « a Pi _g u ^ < ^ w a Ci- o Oc§ m Z 'd W « fei g 55 s o m iH] S 3 ^ Pm 5 m ^ ^ S §1 s ^ W -p < « m i' A H 'd o O 4) ^ -M 8 o § a O 'd Pi ^ U-f u g in o^ i> o^ S T' 7 -T 3 g cn in A O A A o o ■VO O ^ ?2 'a •g rn ni vq > m* A A A , O m o o 23 'd ^ 'o ^ M in d, I ^ in '« A CN I I m ov .9 i aO.^ d w-d- ai ^ .9^ ^ m A A « « « ca JD r=M !— 4 d d o o M B C« w .3 .3 » O d d A A >^3 M bO W) bOrd rd C G G "m 0) 0) « .d =S 0H A A A A -a 0) A Q o d in s i s 4=4 d I 0) m 13 o ^ O ti d s 0) ov 5-1 (NS “i d Is 13 g a ^ A o d g ■4-i 4) A '•p 2 ^ 2 A O 0) 2 1 9a S ^ a A 13 o m A 305 306 MADRONO [Vol. 61 Key to Carex sect. Acrocystis of California, based on Zika et al. (2012) In the key below, mature specimens with ripe perigynia are needed for reliable identification. 1. Stems unisexual . .C. serpenticola (2) 1 ' Stems with both pistillate and staminate flowers 2. Basal spikes absent 3. Pistillate scales and proximal staminate scales green, red, or purple with hyaline margin 0.4- 0.8 mm wide; terminal spike staminate; stems always monoecious, never on serpentine. .... C. inops subsp. inops y Pistillate scales and proximal staminate scales dark purple with hyaline margin 0. 1-0.2 mm wide; terminal spike variable, staminate, pistillate, or gynecandrous; stems monoecious or dioecious; always on serpentine C. serpenticola (2) 2' Basal spikes present 4. Perigynia veinless or nearly so except for two strong marginal ribs; pistillate scales with l(-3) prominent veins, lacking fine veins 5. Inflorescence bracts inconspicuous and shorter than the inflorescence on elongate, non- basal stems (occasionally shoots produced in a second flush of growth will have elongate inflorescence bracts); proximal sheaths disintegrating into stiff fibers; perigynia 1. 5-2.1 mm wide; habitat coastal dunes and headlands C. brevicaulis 5' Inflorescence bracts conspicuous and usually longer than the inflorescence on elongate, non-basal stems; proximal sheaths not disintegrating into stiff fibers; perigynia 1-1.7 mm wide; habitat widespread, coastal and montane 6. Perigynia 3. 1^.5 mm long, beaks 0.7-1. 7 mm long, beak teeth 0.2-0. 5 mm long; stems usually ascending, scabrous; habit loosely to densely cespitose; rhizomes often stout, 1.1-3 mm diam .C. rossii 6' Perigynia 2.3-3. 1 mm long, beaks 0.4-0. 8 mm long, beak teeth 0. 1-0.2 mm long; stems usually spreading or arching, smooth to scabrous; habit loosely cespitose; rhizomes often slender, 0. 8-2.0 mm diam. . C. deflexa var. boottii 4' Perigynia with 10-20 strong veins across the faces, usually extending to mid-body or beyond, in addition to the two marginal ribs; pistillate scales with 3-5 prominent veins, plus up to 14 less conspicuous veins 7. Foliage blue-green when fresh, strongly papillose abaxially when dry, at least on new growth (40 X); perigynium body elliptic or barrel-shaped ... .C. brainerdii 1' Foliage green when fresh, smooth to scabrous abaxially when dry, never papillose; perigynium body obovoid to subglobose 8. Some basal pistillate spikes on long weak arching peduncles more than 10 cm long; perigynia up to 2.3 mm wide; perigynium stipes up to 2.2 mm long; habitat coastal mesic forest and openings C. globosa 8' All basal pistillate spikes on stiffly erect peduncles less than 8 cm long; perigynia up to 1.9 mm wide; perigynium stipes up to 1.6 mm long; habitat inland, dry forest, savanna and chaparral 9. Perigynium stipe longer than beak; pistillate scales usually with 6-12 fine inconspicuous nerves; perigynium beak 0. 5-0.9 mm long; fruiting from late March to early June; 450-735 m elev., northern Sierra Nevada ............ C. xerophila 9' Perigynium stipe equaling or shorter than beak; pistillate scales usually with 3-6 fine inconspicuous nerves; perigynium beak 1.0-1. 3 mm long; fruiting in June, 1900-2400 m elev., San Gabriel and San Bernardino Mountains C sp. “A” Acknowledgments We thank the curators and staff from the following herbaria for assistance and loans of specimens: CAS, CDA, CHSC, DAV, DS, GH, HSC, JEPS, NEBC, ORE, OSC, POM, RSA, UCR, WILLU, and WTU. Cindy Brinkhurst, Karen Callahan, Samantha Hillaire, and Bill Wilson helped guide Janeway to populations of chaparral sedge. Access and collecting permits were granted by Graciela Hinshaw of the Mother Lode Field Office, Bureau of Land Management. Jan Kirschner kindly translated the diagnosis into Latin. Tom Ruehli created the range map. Krista Anandakuttan drew the plate and assisted with the other graphics. Nick Otting and Dick Brainerd of the Carex Working Group provided invaluable feedback with the research and manuscript. Julie Nelson, as always, was generous with encouragement and logistical support. Literature Cited Alexander, E. B. 2011. Gabbro soils and plant distributions on them. Madrono 58:113-122. . 2012. Comment on the gabbro soils of Pine Hill. Madrono 59:1. Baldwin, B. G., D. H. Goldman, D. J. Keil, R. Patterson, T. J. Rosatti, and D. H. Wilkin (eds.). 2012. The Jepson manual: vascular plants of California, 2nd ed. University of California Press, Berkeley, CA. Burge, D. O. and P. S. Manos. 2011. Edaphic ecology and genetics of the gabbro-endemic shrub Ceano- tlms roderickii (Rhamnaceae). Madrono 58:1-21. Crins, W. j. and j. F. Rettig. 2002. Carex Lin- naeus sect. Acrocystis. Pp. 532-545 in Flora of North America Editorial Committee (eds.). Flora of North America north of Mexico. Vol. 23, 2014] ZIKA ET AL.: CAREX XEROPHILA 307 Magnoliophyta: Commelinidae (in part): Cypera- ceae. Oxford University Press, New York, NY. Hinshaw, G. 2008. Pine Hill Preserve management plan. Bureau of Land Management, Folsom Field Office, CA. Website www.blm.gov/pgdata/etc/ medialib/blm/ca/pdf/folsom/plans.Par.441 14. File. dat/FINAL_PHPmgmtPlan0708.pdf (accessed 5 September 2013). Hunter, J. C. and J. E. Horenstein. 1991. The vegetation of the Pine Hill area (California) and its relation to substratum. Pp. 197-206 in A. J. Baker, J. Proctor, and R. D. Reeves (eds.), The vegetation of ultramafic (serpentine) soils. Intercept, Andover, U.K. Janeway, L. P. 1992. Cyperaceae of Butte County, California. Part 1 : Carex. Studies from the herbar- ium, California State University, Chico 9:1-79. Oswald, V. H. 2002. Selected plants of northern California and adjacent Nevada. Studies from the Herbarium, California State University, Chico 11:1^51. SoRRiE, B. A., P. D. McMillan, B. van Eerden, R. J. Leblond, P. E. Hyatt, and L. C. Anderson. 2011. Carex austrodeflexa (Cyperaceae), a new species of Carex sect. Acrocystis from the Atlantic Coastal Plain of the southeastern United States. Journal of the Botanical Research Institute of Texas 5:45-51. Taylor, D. W. 2010. Flora of the Yosemite sierra, being a transect flora of the central Sierra Nevada, including all of Tuolumne, Mariposa and Madera counties, the Mono Basin, and adjacent areas of Mono County. Published by the author, Aptos, CA. United States Fish and Wildlife Service (USFWS). 1996. Endangered and threatened wildlife and plants; determination of endangered status for four plants and threatened status for one plant from the central Sierran foothills of Califor- nia, Federal Register 61:54346-54358. Werier, D. a. 2006. Carex reznicekii, a new wide- spread species of Carex section Acrocystis (Cyper- aceae) from eastern North America. Sida 22: 1049-1070. Wilson, J. L. 1986. A study of plant species diversity and vegetation patterns associated with the Pine Hill gabbro formation and adjacent substrata, El Dorado County, California. M.S. thesis. California State University, Sacramento, CA. , D. R. Ayres, S. Steinmaus, and M. Baad. 2009. Vegetation and flora of a biodiversity hotspot: Pine Hill, El Dorado County, California, USA. Madrono 56:246-278. ZiKA, P. F. 2012. Carex orestera (Cyperaceae), a new sedge from the mountains of California. Novon 22:118-124. , A. L. Hipp, AND J. Mastrogiuseppe. 2012. Carex. Pp. 1308-1339 in B. G. Baldwin, D. H. Goldman, D. J. Keil, R. Patterson, T. J. Rosatti, and D. H. Wilkin (eds.). The Jepson manual: vascular plants of California, 2nd ed. University of California Press, Berkeley, CA. , L. P. Janeway, B. L. Wilson, and L. Ahart. 2013. Carex cyrtostachya (Cyperaceae), a new species of sedge endemic to the Sierra Nevada of California. Journal of the Botanical Research Institute of Texas 7:25-35. , K. Kuykendall, and B. Wilson. 1998. Carex serpenticola (Cyperaceae), a new species from the Klamath Mountains of Oregon and California. Madrono 45:261-270. Madrono, VoL 61, No. 3, pp. 308-315, 2014 PHYLOGENETIC RELATIONSHIPS AND CROSSING DATA REVEAL A NEW SPECIES OF NEMOPHILA (BORAGINACEAE) Camille M. Barr 703 Turtle Crest Drive, Irvine, CA 92603 For correspondence contact Stephen Weller, sgweller@uci.edu Abstract I report a new species of Nemophila growing sympatrically in some localities with N. menziesii Hook, & Arn. var. atomaria (Fisch. & Mey.) H. P, Chandler {sensu Constance 1941) in Northern California. The new species, N. hoplandensis, differs from N. menziesii var. atomaria in both floral and vegetative color. The corollas of N. menziesii var. atomaria range from deep blue to a bluish white. Nemophila hoplandensis has large and unusually bright white corollas and vegetative structures that are a brighter green than those of N. menziesii var. atomaria. I provide support from controlled crosses for treating these populations as a new species; despite sharing pollinators N. hoplandensis is reproductively isolated from N. menziesii var. atomaria via failure of hybrid seed production following artificial crosses. Molecular phylogenetic analyses also clearly differentiate N. hoplandensis within the genus. The newly described species, currently known only from Mendocino and Napa counties, has a range that is more restricted than and lies within the range of N. menziesii var. atomaria. Key Words: Boraginaceae, Hydrophyllaceae, Nemophila menziesii var. atomaria, phylogeny, sympatry. Nemophila menziesii Hook. & Am., previously placed in the family Hydrophyllaceae but recently moved to an expanded Boraginaceae (APG 1998; Olmstead et a!. 2000), has a complicated taxo- nomic history. Commonly known as baby blue- eyes, N. menziesii has been split into multiple species and varieties in some treatments, but these subdivisions have been rejoined by other authors (Nuttall 1822; Bentham 1834; Fischer and Meyer 1835, 1846; Gray 1875; Eastwood 1901; East- wood 1902; Chandler 1902; Chandler 1907; Brand 1913; MacBride 1919; Chittenden and Turrill 1926; Constance 1941). Nemophila Nutt, is currently split into eight species, and N. menziesii into three varieties (for a summary of the genus see Helfgott 2000). In this paper I describe a new white-flowered species of Nemophila that con- tains individuals previously thought to lie at the far end of the range of the floral color polymorphism seen in N. menziesii var. atomaria (Fisch. & C.A. Mey.) H. P. Chandler. Nemophila menziesii is distributed predominantly in Califor- nia but extends into Oregon and Mexico. Corolla color varies extensively in N. menziesii var. atomaria {sensu Constance 1941), which has whiter corollas in northern areas, and bluer corollas in the southern regions of its range (Barr 2004, see also Plate 1 in Chittenden and Turrill 1926). In central and northern California corolla color varies even within populations, ranging from blue to white. I discovered populations of what I initially assumed to be N. menziesii var. atomaria in Mendocino and Napa counties that contained individuals with unusually bright and relatively invariant white corollas, and slightly brighter green vegetative parts than observed in other populations of N. menziesii var. atomaria (Fig, 1). In some areas these white individuals occurred sympatrically with N. menziesii var. atomaria. In areas of sympatry I observed pollinators moving between white-flowered indi- viduals and N. menziesii var. atomaria. Nemophila menziesii var. atomaria is widely distributed in the coastal ranges of central and northern California. The new species is restricted to Mendocino and Napa counties based on my extensive field observations throughout northern California and southern Oregon and examination of the collections in the Jepson and Humboldt State University herbaria. Here I present molec- ular phylogenetic and crossing data in support of the hypothesis that these populations of plants with bright white corollas represent a new species of Nemophila, and describe this newly-discovered species. Methods Study Populations I collected tissue samples of Nemophila men- ziesii var. atomaria from throughout its range in California, from Riverside County to Mendocino County (Fig. 2, Table 1). I collected Nemophila heterophyiia Fisch. & C.A. Mey and Phaceiia californica Cham, from Mendocino and Marin Counties, respectively. I obtained samples of other species and genera as follows: N. maculata Benth. ex Lindl. (commercial source); N. aphylla 2014] BARR; A NEW SPECIES OF NEMOPHILA IN CALIFORNIA 309 Fig. 1 . Nemophila hoplandensis (white corolla) and N. menziesii var. atomaria (blue/white corolla). (L.) Brummitt (dried) from Diane Ferguson, at the Herbarium of Louisiana State University, LA; N. phacelioides Nutt, (dried), from Beryl Simpson, School of Botanical Sciences, Austin, TX.; N, parviflora Douglas ex Benth. (dried) from James West in Texas. DNA Analysis I sequenced two nuclear regions (the rDNA internal transcribed spacer regions I [primers ITS2,5] and II [primers ITS3,4]; Gueidan et al. 2007), and three intergenic chloroplast spacer regions (EF, located between the trnL 3' exon and trnF; L1L2, located between trnT and the trnL 5' exon a [Taberlet et al. 1991]; and TrnS-trnG [GS], located between trnS and trnG [Hamilton 1999; Table 2]). PCR and Sequencing I extracted total DNA primarily from fresh samples frozen after collection in the field, although in a few cases I used samples that had been dried in silica-gel, or herbarium specimens. I performed extractions using either Qiagen DNEasy Plant kits (Qiagen, Valencia, CA, USA) or the CTAB method of Doyle and Doyle (1987). I used Touch-down PCR (Korbie and Mattick 2008) to sequence both strands and then proofread the resulting chromatographs using Sequencher software (Gene Codes Corporation, Inc., Ann Arbor, MI, USA). I combined the complementary strands into one sequence using MEGA5.1 (Tamura et al. 2011) and aligned the sequences using ClustalX (Thompson et al. 1997) with manual correction as necessary. I inserted gaps, which were generally short (1-13 nucleo- tides), in my alignments because they appeared to be a phylogenetically informative trait (Egan and Crandall 2007 [2008]). Fig. 2. Map of California showing collection sites. Filled circles indicate that only Nemophila menziesii var. atomaria was found in this location. Open circles indicate that N. hoplandensis was present as well. Population abbreviations: Bodega Bay, CA (BB); Covelo, CA, Rock Mill (RM); Covelo, CA, Train Trestle (TT); Figueroa Mountain, Los Padres National Forest, CA (FM); Hastings Reserve, Carmel Valley, CA (Hast); Hopland Research and Extension Center, Hopland, CA (HOP); La Panza, Los Padres National Forest, CA (La Panza); Lake Berryessa, Napa, CA (LBer); Manchester State Beach, Manchester, CA (Manch); Motte Rimrock Reserve, Perris, CA (MR); Mount Tamalpais, Golden Gate National Recreation Area, CA (MT); Point Reyes National Seashore, Marin County, CA (PR); Skyline Wilderness Park, Napa, CA (SWP); Toro Canyon, CA (Toro). Phylogenetic Analysis To ensure that my phylogenetic analyses were robust to choice of gene, I analyzed four different datasets: a) concatenated nuclear sequences, b) concatenated chloroplast sequences, c) concate- nated nuclear and chloroplast sequences, and d) a single nuclear locus, ITS2,5. All of the concate- nated datasets contained individuals from N. menziesii var. atomaria and from the new species, which was previously considered N. menziesii var. atomaria. I used MEGA to determine the best fit nucleotide substitution model. The resulting models were T92 for EF, GS, L1L2, and TN93 for ITS2,5, and T92 with a gamma distribution for ITS 3,4. I used MEGA, with 1000 bootstrap replicates, to construct maximum likelihood trees of the concatenated sequences. To confirm the robustness of my phylogenetic analyses to choice of algorithm I also constructed an ITS2,5 tree 310 MADRONO [Vol. 61 Table 1. Description of Specimens. Sample name, population (Figure 2 location abbreviation) and collector (given if other than the author, C Barr), county, and taxon. Sample name Population / Collector County Taxon Naph Baton Rouge Parish, Louisiana / Diane Ferguson Baton Rouge Parish, Louisiana Nemophila aphylla NhetHopFMS Hopland (HOP) Mendocino, CA Nemophila heterophylla NhetCovRMdr Hopland, Old Toll Road (HOP) Mendocino, CA N. heterophylla HLwhl Hopland, Hog Lake (HOP) Mendocino, CA Nemophila hoplandensis Hop 1124 Hopland (HOP) Mendocino, CA N. hoplandensis HoplllO Hopland (HOP) Mendocino, CA N. hoplandensis HLwhl Hopland, Hog Lake (HOP) Mendocino, CA N. hoplandensis WINEl Hopland, in vineyard (HOP) Mendocino, CA N. hoplandensis NwhHopFMS Hopland, Foster Meadow Sign (HOP) Mendocino, CA N. hoplandensis BB1018 Bodega Bay (BB) Sonoma, CA Nemophila menziesii var. atomaria FMl Figueroa Mtn. (FM) Santa Barbara, CA N. menziesii var. atomaria FM2 Figueroa Mtn. (FM) Santa Barbara, CA N menziesii var. atomaria HastNH-1 Hastings, Newt Hill (Hast) Monterey, CA N menziesii var. atomaria Hast3-1 Hastings (Hast) Monterey, CA N menziesii var. atomaria Hast8-2a Hastings (Hast) Monterey, CA N menziesii var. atomaria HLbll Hopland, Hog Lake (HOP) Mendocino, CA N menziesii var. atomaria HLbl2 Hopland, Hog Lake (HOP) Mendocino, CA N menziesii var. atomaria Hopll75 Hopland, Main (HOP) Mendocino, CA N menziesii var. atomaria LaPanzal La Panza San Luis Obispo, CA N. menziesii var. atomaria LBerMCl Lake Berryessa (LBer) Napa, CA N. menziesii var. atomaria LH787 Point Reyes, Lighthouse (PR) Marin, CA N menziesii var. atomaria Manch6 Manchester State Beach (Manch) Mendocino, CA N menziesii var. atomaria Manch9 Manchester State Beach (Manch) Mendocino, CA N. menziesii var. atomaria MR2005b Motte Rimrock (MR) Riverside, CA N menziesii var. atomaria MR330 Motte Rimrock (MR) Riverside, CA N menziesii var. atomaria MtTam2 Mount Tamalpais (MT) Marin, CA N. menziesii var. atomaria MtTam6 Mount Tamalpais (MT) Marin, CA N. menziesii var. atomaria NB934 Point Reyes, North Beach (PR) Marin, CA N. menziesii var. atomaria RM4b Covelo, Rock Mill (RM) Mendocino, CA N. menziesii var. atomaria SWP2bll Skyline Wilderness Park (SWP) Napa, CA N. menziesii var. atomaria Torol Toro Monterey, CA N. menziesii var. atomaria TT5 Covelo, Train Trestle (TT) Mendocino, CA N menziesii var. atomaria NmHopFMS Hopland (HOP) Mendocino, CA N menziesii var. atomaria Nmac Commercial Seed Packet Nemophila maculata Nphac Travis County, Texas / Beryl Simpson Travis County, Texas Nemophila phacelioides PhacKB Point Reyes, Kehoe Beach (PR) Marin, CA Phacelia californica using MrBayes version 3.1.2. (Ronquist and Huelsenbeck 2003). Greenhouse Crosses I examined the ability of populations of the putative new species to hybridize with popula- tions of N. menziesii var. atomaria using con- trolled crosses. I first germinated seeds collected throughout the range of N. menziesii var. atomaria in a growth chamber. I planted the resulting seedlings in a pollinator-free greenhouse in 2004 at the University of California, Irvine, where I cross-pollinated individuals from differ- ent populations to determine reproductive com- patibility. I used several plants of the new species in self and cross pollinations within the new species as well as in crosses to individuals of N. menziesii var. atomaria and N. maculata. In most cases I was able to carry out reciprocal crosses. In all crosses I used hermaphroditic plants that were emasculated before anthesis. I collected pollen in microcentrifuge tubes and performed crosses, using a toothpick, on either the same day as collection, or within the next two days using refrigerated pollen. Results Phylogeny. All phylogenetic analyses grouped individuals of the newly described species as a monophyletic clade separated from N. menziesii var. atomaria with strong statistical support. This clade is shown within a grey box on the maximum likelihood consensus tree resulting from analysis of concatenated nuclear data (Fig. 3); the trees produced by other analyses were similar. All sequences of the new species were identical for the genes surveyed. Intra- and interpopulation crosses. Two individ- uals of the new species were used as female 2014] BARR; A NEW SPECIES OF NEMO PHI LA IN CALIFORNIA 311 Table 2. Loci and Primers Used for Generation of Sequences Used in Phylogenetic Analysis. Locus Primer Location Reference ITS5 ITS2 GGA AGT AAA AGT CGT AAC AAG G GCT GCG TTC TTC ATC GAT GC ITS II Gueidan et al. 2007 ITS3 ITS4 GCA TCG ATG AAG AAC GCA GC TCC TCC GCT TAT TGA TAT GC ITS I Gueidan et al. 2007 trnH psbA ACT GCC TTG ATC CAC TTG GC CGA AGC TCC ATC TAC AAA TGG Hamilton 1999 trnS trnG GCC GCT TTA GCT CAC TCA GC GAA CGA ATC ACA CTT TTA CCA C Intergenic spacer between trnS and trnG Hamilton 1999 trnLl trnL2 CGA AAT CGG TAG ACG CTA CG GGG GAT AGA GGG ACT TGA AC Intergenic spacer between trnT and the trnL 5' exon Taberlet et al. 1991 trnL trnF GGT TCA AGT CCC TCT ATC CC ATT TGA ACT GGT GAC ACG AG Intergenic spacer between the trnL 3' exon and trnF Taberlet et al. 1991 parents in crosses to other taxa {N. maculata and N. menziesii var. atomarid). These crosses pro- duced very few capsules (proportion of flowers setting fruit 0-0.11) compared to results from within-species crosses of three individuals of the new species (proportion of flowers setting fruit O. 69-1.0) or self pollinations (proportion of flowers setting fruit 0.29-0.9; Table 3). When the new species was used as the male parent in crosses to N. maculata very few capsules were produced (Table 3). The new species was not used as the male parent in crosses to N. menziesii. Nemophila maculata produced few capsules when crossed as a female parent to N. menziesii (Table 3). Discussion I found strong evidence that populations of plants with unusually bright white corollas in the central Coastal Ranges, and previously consid- ered N. menziesii var. atomaria, represent a distinct species described here as Nemophila hoplandensis. Reproductive isolation in plants is not necessary to delimit different species (Riese- berg and Carney 1998; Rieseberg and Willis 2007), although these populations are almost completely reproductively isolated from N. men- ziesii var. atomaria. Phylogenetic analysis sup- ports recognition of a new species because the N. hoplandensis samples form a monophyletic clade distinct from samples of N. menziesii var. atomaria and all other species of Nemophila. Variation in corolla color within N. menziesii var. atomaria, which ranges from blue to a very light bluish white, had previously obscured the distinct nature of the white-flowered species. Phylogenetic analysis (Fig. 3) suggests that N. hoplandensis is more closely related to N. maculata and N. heterophylla than to N. menzie- sii. Similar levels of reduced capsule produc- tion occurred following crosses between N. hoplandensis and N. maculata and between N. hoplandensis and N. menziesii (Table 3). Study of the evolutionary divergence of these species might be further explored through additional crosses of these species and N. heterophylla. Speciation The differentiation between N. hoplandensis the N. menziesii var. atomaria is interesting given their physical proximity. I observed individual bees visiting both species, eliminating pollinator- mediated selection as a method of speciation (Levin 1971). Post-zygotic isolation is generally viewed as inefficient because it leads to pollen waste and perhaps stigma clogging (Grant 1981; Pascarella 2007). Thus, over time, closely located species are expected to evolve pre-mating isola- tion to prevent this inefficiency. The apparent preference of N. hoplandensis for drier microhab- itats in some locations (C. Barr, personal observation) may have evolved to prevent polli- nator sharing and increase pre-zygotic isolation. In other areas, however, individuals from both taxa are distributed evenly throughout patches, demonstrating that the extent of microhabitat differentiation varies. Future study of this group might also focus on the genetic differentiation of sample MR2005b from other individuals of N. menziesii var. atomaria (Fig. 3). This individual was grown from seed collected at the Motte Rimrock reserve near Riverside, CA, the southernmost N. menzie- sii sample in my study. Plants from Motte Rimrock have been classified as N. menziesii var. integrifolia Parish (Constance 1941), howev- er, they differ phenotypically from other N. menziesii plants in having small flowers with unusually dark blue/purple corollas. They also have a more diffuse architecture, with almost climbing stems, than seen in other populations of N. menziesii. Reduced seed production following crosses between plants from Motte Rimrock and Figueroa Mountain, near Santa Barbara CA, 312 MADRONO [Vol. 61 18 25 £7 68 98 43 70 80 -™RM4b - NmHopFMS ManchO -Hop1175 HLbll IVIanchS LaPanzal 6^FM1 IfM2 HastNH-1 Tbro1 Hast3-1 Hast8-2a TT5 LBerMCI BB1018 LH787 MtTamB NB934 SWP2bl1 MR2005b 97 90 Nemophila menziesii Nmac Nemophila maculata NhetHopFMS NhetCovRMdr HLwh1 “ Hop1124 HoplllO Sd NwhHopFMS WINE1 1 74 Nemophila heterophylla Nemophila hoplandensis Naph Nemophila aphylla Nphac Nemophila phacelioides PhacKB Phacelia californka I — I 0.02 Fig. 3. Unrooted bootstrap consensus tree of concatenated nuclear data constructed using maximum likelihood. The scale bar represents the number of substitutions per site. The gray box shows specimens of N. hoplandensis. All specimens of TV. menziesii used in phylogenetic analysis were TV. menziesii var. atomaria. 2014] BARR: A NEW SPECIES OF NEMOPHILA IN CALIFORNIA Table 3. Capsules/Flower Resulting from Crossing Studies. 313 Dam Sire Flowers pollinated Caps collected Caps/ flower N. maculata Nmac N. menziesii var. atomaria Manch9 7 0 0.00 N. menziesii var. atomaria NB934 5 0 0.00 N. maculata self 17 10 0.59 N. hoplandensis Hopl 1 10 9 1 0.11 N. menziesii var. atomaria TT5 4 0 0.00 N. menziesii var. atomaria LBMCl N. menziesii var. atomaria Manch6 9 9 1.00 N. menziesii var. atomaria Manch9 11 7 0.64 N. menziesii var. atomaria LH787 17 10 0.59 N. menziesii var. atomaria NB934 10 7 0.70 N. menziesii var. atomaria RM4b 15 8 0.53 N. menziesii var. atomaria MtTam2 9 6 0.67 N. menziesii var. atomaria TT5 5 0 0.00 N. menziesii var. atomaria HLbll 15 8 0.53 N. menziesii var. atomaria Hop 11 75 3 0 0.00 N. hoplandensis Hopl 1 10 N. menziesii var. atomaria Manch9 9 1 0.11 N. maculata Nmac 9 1 0.11 N. hoplandensis self 13 9 0.69 N. hoplandensis Hopl 124 13 9 0.69 N. hoplandensis HLwhl 11 8 0.73 N. menziesii var. atomaria HLbll 1 0 0.00 N. hoplandensis Hopl 124 N. hoplandensis Hopl 1 10 13 12 0.92 N. hoplandensis self 17 5 0.29 N. hoplandensis HLwhl 11 10 0.91 N. menziesii var. atomaria HLbll 4 0 0.00 N. menziesii var. atomaria Hopl 175 N. menziesii var. atomaria self 6 3 0.50 N. menziesii var. atomaria HLbll 10 8 0.80 N. menziesii var. atomaria NB934 15 6 0.40 N. menziesii var. atomaria MtTam6 7 0 0.00 N. menziesii var. atomaria RM4b 5 4 0.80 N. hoplandensis HLwhl N. hoplandensis self 10 9 0.90 N. hoplandensis Hop 1110 7 7 1.00 N. hoplandensis Hop 1 1 24 2 2 1.00 geographically the closest population in my sample to Motte Rimrock (C. Barr, personal observation) suggests that this population has diverged from other populations of N. menziesti. Taxonomic Treatment Nemophila hoplandensis C.M. Barr, sp. nov. (Fig. 1).— TYPE: USA, California, Mendo- cino County, University of California ANR Hopland Research and Extension Center, off of Highway 175, in clearings of California Oak and Bay Laurel, in loamy soil near main office of field station, 39°8.50'N, 123°04.9rW, 25 Mar 2007, C. M. Barr HopMainl-1 (holotype: JEPS; isotypes: RSA, HSC, PH, DAY, UCSB). Annuals, forming a rosette during winter rains beginning around December, bolting in early spring with warmer temperatures to become branched (diffuse or dense, depending on density of neighboring plants and intensity of sun). Leaves bright green, both surfaces hirsute. Basal leaves 9-67 mm long, including a petiole which can equal the length of the lobed section of the leaf, 4-19 mm wide, pinnately lobed, lobes 4-8, lightly pointed, sometimes again pinnate; cauline leaves similar, smaller distally. Flowers: 1-28 per plant, 17-24 mm diam., axillary or in terminal helicoid cymes; calyx rotate, sepals acuminate, shorter than petals, sinuses often with spreading or reflexed appendages smaller than calyx lobes; corolla shallowly cupped to almost flat, bright white with purple dots on the center of each lobe and/or variable thin blue or purple veins from the center to edge, but no large segments of blue or purple; petals 14-25 mm, with an irregular tooth at each end; filaments growing during anthesis, anthers oblong and light tan to dark brown; ovary globose, style 2-6 mm long, cleft about 1/ 4 mm its length, segments growing and reflexing after anthesis. Fruit: capsule brown, 7-9 mm; seeds brown, 1-13, nearly globose, 3^ mm diam., shallowly pitted; elaiosome white-yellow. Paratypes: USA, CALIFORNIA. Mendocino Co.: off Rt. 75, Hopland, 25 Mar 2004, C. M. Barr HopMain\vhl-4 (RSA); C M. Barr Hop- Mainwhl-5 (PH); C M Barr HopMamwhl-6 (UCSB); C M Barr HopMainbB-l (JEPS); C. M Barr HopMambB~2 (RSA); C. M. Barr HopMainbB-3 (PH); C. M. Barr HopMambB-4 314 MADRONO [Vol. 61 (HSC); C. M. Barr HopMainbI3-5’, C. M. Barr HopFM\vh7-l (RSA); C. M. Barr HopFMwh7-2 (HSC); Hopland, California, off Rt. 75, 11 Mar 2004, C.M. Barr HopMain\vhl-7 (UCR); C M. Barr HopMaimvhl-8 (DAV); open meadow off of Mountain House Road, 2 mi from Hwy 101, 10 Mar 2004, C. M. Barr CirAR2-l (JEPS); C M. Barr CirAR2-2 (RSA); C. M. Barr CirAR2~3 (PA); C M. Barr CirAR2-4 (HAS); Old Toll Road, about 2 mi from intersection with Rt. 75, 38 56.972N, 123 02. 814W, 9 Mar 2004, C.M Barr s.n. (JEPS); C. M Barr OTRlwhFl (RSA); C M. Barr OTRlwh4-2 (PH); C. M Barr OTRhvh4-3 (HSC); Hopland, California, off Rt. 75, 9 Mar 2004, C.M Barr HopLMwhSA (JEPS); C M Barr HopBOGwh6-l (JEPS); C M Barr HopBOG\vh6-2 (RSA); C. M. Barr Hop- BOGwli6-3 (PH); C M. HopBOGwh6-4 (HSC); Old Toll Road, about 2.55 mi from intersection with Rt. 75, 38 57.021N, 123 02. 303W, 9 Mar 2004, C. M Barr OTR2bl9G (JEPS); C. M. OTR2bl9~2 (PH). Etymology. The species is named for the University of California Hopland Research and Extension Center, at the center of the distribution of this species. Corolla color is the most distinctive morpho- logical character of Nemophila hopkmdensis. As noted by Simpson et al. (2001), corolla color is often lost during drying of herbarium specimens, and this may in part explain why the distinct morphology of N. hopkmdensis had previously gone unnoticed. In the field, corollas of N. hoplandensis are brighter white outside the central area that is typically white, or white with purple spots and/or veins that radiate from the center to the edge of each petal. These bright white flowers are less variable than white individuals of the other taxa (see Barr 2004 for examples of color variability in N. menziesii var. atomaria\ some plants of N. menziesii var. atomaria rival those of N. hoplandensis in whiteness of corollas; also see Plate 1 in Chittenden & Turrill (1926) for examples of corolla color variation in Nemophi- la). Vegetative portions of N. hoplandensis are also somewhat brighter green than that of other Nemophila species; this difference is also lost upon drying. Acknowledgments I gratefully acknowledge Kerry Heise from the University of California ANR Hopland Research and Extension Center, who originally led me to populations of N. menziesii at the station. I also thank Robert Patterson for stumbling upon my first Nemophila population. I thank Jim McLister and Dan Harwig for much help in the field and greenhouse. I thank Bruce Baldwin, Mike Hardig, and Jim West for information about populations in the San Francisco Bay Area, and Mark Stromberg and Joseph Messin for help at the Hastings and Motte Rimrock field stations. respectively. Steve Weller, Steve Frank, and Ann Sakai provided advice, kind general encouragement, and greenhouse space. I also thank Doug Taylor and Lila Fishman for providing tools for the sequencing. Whit Farnum was very helpful with collecting the molecular data. Gwendolynne Barr provided useful housing and help with figure imaging, and Diane Campbell gener- ously allowed me use of her growth chamber. I am very grateful to Robin Bush for preparing the Bayesian ITS2,5 tree, and William R. Anderson for his advice on the name for this species. I thank the wonderful UC Hopland, Bodega Bay, Hastings, and Motte Rimrock field stations and Point Reyes National Seashore for access to populations, plants, and housing and for general assistance. Finally, I thank John Strother of the Jepson Herbarium and Robin Bush and Steve Weller of the University of California, Irvine, for comments on the manuscript. I performed DNA extraction, PCR and sequencing in the laboratories of Dr. Douglas Taylor, University of Virginia, and Dr. Lila Fishman, Univer- sity of Montana. Funding for this work came partly from the Explorer’s Club and a NSF Dissertation Improvement Grant (DEB-0073550). Literature Cited Angiosperm Phylogeny Group (APG). 1998. An ordinal classification of the families of flowering plants. Annals of the Missouri Botanical Garden 85:531-553. Barr, C. M. 2004. Hybridization and regional sex ratios in Nemophila menziesii. Journal of Evolu- tionary Biology 17:786-794. Bentham, G. 1834. Review of the order of Hydro- phyllaceae. Transactions of the Linnean Society 17:267-282. Brand, A. 1913. Hydrophyllaceae. Pp. 1-210 in A. Engler (ed.), Das Pflanzenreich IV, Vol. 251 (Heft 59). Verlag von Wilhelm Engelmann, Leipzig, Austria. Chandler, H. P. 1902. A revision of the genus Nemophila. Botanical Gazette 34:194—215. . 1907. Notes on two California Nemophilas. Botanical Gazette 44:318-382. Chittenden, R. J. and W. B. Turrill. 1926. Taxonomic and genetical notes on some species of Nemophila. Bulletin of Miscellaneous Informa- tion (Royal Gardens, Kew) 1:1-12. Constance, L. 1941. The genus Nemophila Nutt. University of California Publications in Botany 19:341-345. Doyle, J. J. and J. L. Doyle. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19:11-15. Eastwood, A. 1901. Some small-flowered Species of Nemophila from the Pacific Coast. Bulletin of the Torrey Botanical Club 28:137-160. . 1902. New species of Nemophila from the Pacific Coast. Bulletin of the Torrey Botanical Club 29:471^74. Egan, A. N. and K. A. Crandall. 2007 [2008]. Incorporating gaps as phylogenetic characters across eight DNA regions: Ramifications for North American Psoraleeae (Leguminosae). Molecular Phylogenetics and Evolution 46:532-546. Fischer, F. E. and C. A. Meyer. 1835. Index Seminum [St. Petersburg] 2:42M3. . 1846. Sertulum petropolitanum. 2014] BARR: A NEW SPECIES OF NEMOPHILA IN CALIFORNIA 315 Grant, V. 1981. Plant Speciation. 2nd ed. Columbia University Press, New York, NY. Gray, A. 1875. A conspectus of the North American Hydrophyllaceae. Proceedings of the American Academy 10:312-332. Gueidan, C., C. Roux, and L. Lutzoni. 2000. Using a multigene analysis to assess generic delineation and character evolution in the Verrucariaceae (Eurotiomycetes, Ascomycota). Mycological Re- search 112:1307-1318. Hamilton, M. B. 1999. Four primer pairs for the amplification of chloroplast intergenic regions with intraspecific variation. Molecular Ecology 8:521-523. Helfgott, D. M. 2000. Evolution and molecular systematics of Nemophila Nutt. (Hydrophyllaceae) and the woody Bencomia alliance (Rosaceae). Ph.D. Dissertation. The University of Texas at Austin, Austin, TX. Korbie, D. J. and J. S. Mattick. 2008. Touchdown PCR for increased specificity and sensitivity in PCR amplification. Nature Protocols 3:1452-1456. Levin, D. A. 1971. The origin of reproductive isolating mechanisms in flowering plants. Taxon 20:91-113. MacBride, j. F. 1919. Reclassified or new spermato- phytes, chiefly North American. Contributions from the Gray Herbarium, new series 59:28-39. Nuttall, T. 1822. Description of rare plants recently introduced into the gardens of Philadelphia. Journal of the Academy of Natural Sciences of Philadelphia 2:179-181. Olmstead, R. G., K. j. Kim, R. K. Jansen, and S. J. Wagstaff. 2000. The phylogeny of the Asteridae sensu lato based on chloroplast ndh¥ gene sequenc- es. Molecular Phylogeny and Evolution 16:96-1 12. Pascarella, j. B. 2007. Mechanisms of prezygotic reproductive isolation between two sympatric species, Gelsemium rankinii and G. sempervirens (Gelsemiaceae), in the southeastern United States. American Journal of Botany 94:468-76. Rieseberg, L. H. and S. E. Carney. 1998. Plant hybridization. New Phytolologist 140:599-624. and j. H. Willis. 2007. Plant speciation. Science 317:910-914. Ronquist, F. and j. P. Huelsenbeck. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19:1572-1574, Simpson, B. B., D. M. Helfgott, and J. L. Neff. 2001. A new cryptic species of Nemophila (Hydro- phyllaceae) from Texas and the lectotypification of N. phacelioides Nuttall. Lundellia 3:30-36. Taberlet, P., L. Gielly, G. Pautou, and J. Bouvet. 1991. Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Molecular Biology 17:1105-1109. Tamura, K., D. Peterson, N. Peterson, G. Stecher, M. Nei, and S. Kumar. 2011. MEGA5: Molecular evolutionary genetic analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Molecular Biology and Evolution 28:2731-2739. Thompson, J. D., T. J. Gibson, F. Plewniak, F. Jeanmougin, and D. G. Higgins. 1997. The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Research 25:4876-82. Madrono, VoL 61, No. 3, p. 316, 2014 NOTEWORTHY COLLECTIONS NEVADA Agave deserti Engelm. (AGAVACEAE). — Clark Co., Newberry Mountains, in a south-facing bowl northwest of unnamed peak with radio facility, ca. 1.25 km southwest of Christmas Tree Pass, UTM 11+ 7045 16E, 3903573N (NAD83) (35.2544N, -1 14.75 19W), with Coleogyne ramosissima Torr., Eriogonum wrightii Ton*, ex. Benth., Ferocactus cylindraceiis (Engelm.) Orcutt, Piniis monophylla Torr. & Frem., and Qiierciis turbinella Greene, 1422 m, in a granitic bedrock area with occasional patches of soil, roughly 50 rosettes present, 8 April 2013, T. Embrey 159 with S. E. Henke (UNLV, ARIZ) (USGS 1982). ' Previous knowledge. Agave deserti (desert agave) occurs in California, Arizona, Baja California, and mainland Mexico. A search of SEINet, UNLV, RENO, and Consortium of California Herbaria did not uncover any collections from Nevada (CCH 2013, SEINet 2013, K. Birgy, Univ. of Nevada, Las Vegas, personal communication, 5 May 2013, A. Tiehm, Univ. of Nevada, Reno, personal communication, 1 June 2013). Significance. The collection is the first report of A. deserti in Nevada. To the east, A. deserti is known from the Kingman vicinity in Arizona, while the closest and northernmost site in California is at the southern end of the Ivanpah Mountains. The collection from Nevada occurs along roughly the same latitude as the north- ernmost California and Arizona sites, with the excep- tion of a disjunct locality in northern Arizona from Gunsite Point 205 km to the north of the Kingman vicinity {Phillips 79-627 [MNA]) (SEINet 2013). Fur- ther searches in the more inaccessible reaches of the Newberry Mountains to the south of the current locality may reveal more A. deserti sites; to the north and east the Newberry Mountains are within the Lake Mead National Recreation Area boundary, and these portions have been well botanized (Holland 1982). Subsequent treatments of the flora of Nevada should include A. deserti (Kartesz 1987). Lorandersonia SALICINA (S. F. Blake) Urbatsch, R. P. Roberts & Neubig (ASTER ACEAE). —Clark Co., “Goldstrike Hot Springs Canyon”, Black Canyon, Lake Mead National Recreation Area, UTM 1 1+, ~30 m, 702751E, 3986350N (NAD83) (~36.0005N, -114.7505W) with Eucnide urens Parry, Maiircmdella antirrhiniflora (Humb. & Bonpl. ex Willd.) Rothm., and Pleiirocoronis pluriseta (A. Gray) R.M. King & H. Rob., —266 ni, on north-facing exposed schist, plants extending up the canyon wall, 9 April 2013, T. Embrey 161 with S. E. Henke (ARIZ, Lake Mead National Recreation Area herbarium) (USGS 1983). Previous knowledge. Lorandersonia salicina (willow glowweed) occurs in Arizona and Nevada. This plant has been reported to occur only within the Grand Canyon (Huisinga et al. 2006), although several records exist outside of the Grand Canyon (SEINet 2013). Only three previous records document this plant within Nevada (K. Birgy, Univ. of Nevada, Las Vegas, personal communication, 5 May 2013). Significance. The collection is the southernmost known site of L. salicina. This plant occurs 19 km to the northeast in James Bay, Arizona. A previous collection by the author, Embrey 127 (Lake Mead National Recreation Area herbarium), from the River Mountains in Nevada 8 km to the northwest of the current collection also documents this species further to the west than previously known (USGS 1987). The two other records from Nevada {Landau 1575 and Landau 1578 [UNLV]) are from sites directly to the north of the James Bay, Arizona location. The current locality in the Black Canyon is roughly 75 km from the terminus of the Grand Canyon and this collection indicates that this species has a range that extends into the lower Colorado River area. — Teague Embrey, Botanist, WestLand Resources, Inc., 4001 E. Paradise Falls Drive, Tucson, AZ 85712. tembrey(^westlandresources.com. Acknowledgements I extend a thank you to Shelley McMahon and Ellen Dorn at the University of Arizona herbarium for their help with processing the collections, and Alice Newton and Toshi Yoshida at Lake Mead National Recreation Area for their assistance with the research permit. I would also like to thank Chris Roberts for the initial tip on the A. deserti locality information, and Shannon Henke for her companionship and support on the many hikes we took together. Literature Cited Consortium OF California Herbaria (CCH). 2013. Data provided by the participants of the California Consortium of Herbaria. Website http://ucjeps. berkeley.edu/consortium (accessed 15 November 2013). Holland, J. S. 1982. A floristic and vegetative analysis of the Newberry Mountains, Clark County, Nevada. M.S. thesis. University of Nevada, Las Vegas, NV. Huisinga, K., L. Makarick, and K. Watters. 2006. River and desert plants of the Grand Canyon. Mountain Press Publishing Company, Missoula, MT. Kartesz, J. T. 1987. A flora of Nevada. Ph.D. dissertation. University of Nevada, Reno, NV. Southwest Environmental Information Network (SEINet). 2013. Website http://swbiodiversity.org/ seinet/index.php (accessed 15 November 2013). United States Geological Survey (USGS). 1982. 1: 100,000-scale metric topographic map of Davis Dam, Nevada-Arizona-California. U.S. Depart- ment of the Interior, USGS, Denver, CO. . 1983. 1: 100,000-scale metric topographic map of Boulder City, Nevada-Arizona. U.S. Depart- ment of the Interior, USGS, Denver, CO. . 1987. 1: 100,000-scale metric topographic map of Lake Mead, Nevada-Arizona. U.S. Department of the Interior, USGS, Denver, CO. Volume 61, Number 3, pages 253-316, published 27 July 2014 Subscriptions— Membership The California Botanical Society has several membership types (individuals ($40 per year; family $45 per year; emeritus $32 per year; students $27 per year for a maximum of 7 years). Late fees may be assessed. Beginning in 2011, rates will increase by $5 for all membership types except life memberships, for which rates will increase by $100, and student memberships, which will not show a rate increase. Members of the Society receive Madrono free. Institutional subscriptions to Madrono are available ($75). Membership is based on a calendar year only. Life memberships are $850. Applications for membership (including dues), orders for subscriptions, and renewal payments should be sent to the Membership Chair. Requests and rates for back issues, changes of address, and undelivered copies of Madrono should be sent to the Corresponding Secretary. Information for Contributors Submit manuscripts to http://www.edmgr.com/madrono/. If you are submitting a manuscript for the first time to the Madrono PeerTrack website, you must first register by clicking "register now" and following the instructions. Manuscripts by authors having outstanding page charges will not be sent for review. Manuscripts may be submitted in English or Spanish. English-language manuscripts dealing with taxa or topics of Latin America and Spanish-language manuscripts must have a Spanish Resumen and an English Abstract. For all articles and short items (NOTES, NOTEWORTHY COLLECTIONS, POINTS OF VIEW, etc.), follow the format used in recent issues for the type of item submitted. Detailed information for contributors, including instructions for manuscript preparation, is available at the California Botanical Society website, www.calbotsoc. org; follow the link for Madrono. Authors are encouraged to include the names, addresses, and e-mail addresses of two to four potential reviewers with their submitted manuscript. All California Botanical Society members current in the volume year that their contributions are published are allowed five free pages per volume year. Additional pages will be charged at the rate of $40 per page. Joint authors may apply all or a portion of their respective five-page allotments to a jointly-published article. Partial pages will be charged as full. 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 color plates and other complex matter (including illustrations, charts, maps, photographs) will be charged at cost. Author’s changes after typesetting @ $4.50 per line will be charged to authors. Page charges are important in maintaining Madrono as a viable publication, and timely payment of charges is appreciated. 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. SMITHSONIAN LIBRARIES \ iiiiiiiiiii 1 II II 1 3 9088 0 7i 3 40 10 J r' I VOLUME 6 1 , NUMBER 4 OCTOBER-DECEMBER 20 1 4 MADRONO A WEST AMERICAN JOURNAL OF BOTANY CONTENTS NEW SPECIES Geographic And Seasonal Variation In Chaparral Vulnerability To Cavitation Anna L. Jacobsen, R. Brandon Pratt, Stephen D. Davis, and Michael F. Tobin 317 Functional Trait Differences Between Weedy^And Non A^edy Plants In Southern California Evan D. MacKinnon, R. Brandon Prdtt^ridAnna LJmp^sen ............ 328 Variation In Seed Charactpisti'c|And Growth JfeR?^THrsTilES (C ardue AE : Aster ace a^)i^^®^rMa^^01egon D. F Spencer, 339 Vernal Pool BTgrf-DrcKS AsPARi^^^E: RoherP^^rest^i:.'':::^di!j'.l'i^f:^^^ .................... 350 VASCUL^^^^^^^^(^^^OSTBl^^^mNAL\lWT-^®|riM*ADERA Melanie Arneit, J^nM. Huber, KathrenmurreM\Sje}^^rf and Sylvia HaultaiM^^. • * Secondary Dispersal Of Willow ^^^^-^^ling On^^^r Into Safe Sites John M. Boland 388 A Newly Described Serpentine-Endemic Ceanothus (Rhamnaceae) From Coastal Marin County, California V. Thomas Parker 399 PUBLISHED QUARTERLY BY THE CALIFORNIA BOTANICAL SOCIETY Madrono (ISSN 0024-9637) is published quarterly by the California Botanical Society, Inc., and is issued from the office of the Society, Herbaria, Life Sciences Building, University of California, Berkeley, CA 94720. Subscription information on inside back cover. Established 1916. Periodicals postage paid at Berkeley, CA, and additional mailing offices. Return requested. Postmaster: Send address changes to Madrono, Kim Kersh, Membership Chair, Uni- versity and Jepson Herbarium, University of California, Berkeley, CA 94720-2465. kersh@berkeley.edu. Editor — Matt Ritter Biological Sciences Department Cal Poly, San Luis Obispo 1 Grand Avenue San Luis Obispo, CA 93407 madronoeditor@gmail.com Editorial Assistant — Genevieve Walden Book Editor — Matt Ritter Noteworthy Collections Editor — David Keil Board of Editors Class of: 2014 — Brandon Pratt, California State University, Bakersfield, CA Tom Wendt, University of Texas, Austin, TX 2015 — Anna Jacobson, California State University, Bakersfield, CA CALIFORNIA BOTANICAL SOCIETY, INC. Officers for 2014-2015 President: Dr. Mark S. Brunell, Department of Biological Sciences, University of the Pacific, Stockton, CA 95211, mbrunell@pacific.edu First Vice President: Andrew Doran, University and Jepson Herbaria, University of California, Berkeley, CA 94720, andrewdoran@berkeley.edu Second Vice President: J. Travis Columbus, Rancho Santa Ana Botanic Garden, Claremont Graduate University, 1500 North College Avenue, Claremont, California 91711, j.travis.columbus@cgu.edu Recording Secretary: Nancy Morin, P O. Box 716, Point Arena, CA 95468, Nancy.Morin@nau.edu Corresponding Secretary: Sheryl Creer, Department of Biology, San Francisco State University, San Francisco, CA 94132, secretary@calbotsoc.org Treasurer: David Margolies, California Botanical Society, Jepson Herbarium, University of California, Berkeley, CA 94720, dm@franz.com The Council of the California Botanical Society comprises the officers listed above plus the immediate Past President, V. Thomas Parker, Department of Biology, San Francisco State University, San Francisco, CA 94132, parker@sfsu.edu; the Editor of Madrono, Matt Ritter; the Membership Chair, Kim Kersh, University and Jepson Herbaria, University of California, Berkeley, CA 94720, kersh@berkeley.edu; Council Members: Staci Markos, University and Jepson Herbaria, University of California, Berkeley, CA 94720, smarkos@berkeley.edu; and Dylan Burge, California Academy of Sciences, dylan.o.burge@gmail.com; Graduate Student Representatives: Jessica Orozco, Rancho Santa Ana Botanic Garden, 1500 N. College Ave., Claremont, Jessica.orozco@cgu.edu; and Adam Schneider, University and Jepson Herbaria, University of California, Berkeley, CA 94720, acschneider@berkeley.edu; Administrator: Lynn Yamashita, University of California, Berkeley, CA 94720, admin@calbotsoc.org; Webmaster: Will Freyman, University of California, Berkeley, CA 94720, freyman@berkeley.edu. @ This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Madrono, VoL 61, No. 4, pp. 317-327, 2014 'g^THSOW^ DEC 1 7 2014 GEOGRAPHIC AND SEASONAL VARIATION IN CH?rPARgAL VULNERABILITY TO CAVITATION Anna L. Jacobsen* and R. Brandon Pratt Department of Biology, California State University, Bakersfield, 9001 Stockdale Hwy, Bakersfield, CA 93311 ^ ajacobsen@csub.edu Stephen D. Davis Natural Science Division, Pepperdine University, 24255 Pacific Coast Hwy, Malibu, CA 90263 Michael F. Tobin Department of Natural Sciences, University of Houston Downtown, 1 Main St, Houston, TX 77002 Abstract Resistance of stem xylem to water stress-induced cavitation and embolism among chaparral shrub species in California has been extensively studied, providing the opportunity to examine broad patterns in cavitation resistance. We used previously published as well as unpublished vulnerability to cavitation curve data from 16 chaparral shrub species of southern California to examine the variability of cavitation resistance across sites, regions, and seasons. Additionally, these data provided a unique opportunity to address a recent methodological debate within the field of plant hydraulics. We found that different methods, specifically a centrifuge method and a dehydration method, produced similar results (P = 0.184). Vulnerability to cavitation varied seasonally, with species exhibiting greater susceptibility to water-stress induced cavitation during the wet season (P = 0.003). Cavitation resistance did not differ among sites that were less than 10 km apart even though these sites differed in their coastal exposure, precipitation, and temperatures (P = 0.476). However, across larger geographic distances and with increased climatic divergence, cavitation resistance significantly varied (P = 0.005), with populations from a higher rainfall mountain range exhibiting greater susceptibility to cavitation. These data suggest that species may be particularly susceptible to the onset of early summer drought before xylem has hardened. Variation in cavitation resistance may be limited locally, but broadly dispersed species may diverge in cavitation resistance across their range. Maintaining populations that vary in cavitation resistance may be an important component of species conservation planning in an era of increased climatic variability. Key Words: Cavitation resistance, chaparral, drought, embolism, water potential, water relations, water stress, xylem. Cavitation and subsequent embolism of xy- lem conduits can occur when plants experience water stress (Davis et al. 2002; Tyree and Zimmermann 2002) and results in reduced hydraulic transport efficiency. This may be particularly harmful during periods of extreme or protracted drought when catastrophic hy- draulic failure may lead to plant dieback and death (McDowell et al. 2008). Cavitation- induced hydraulic failure has been linked to plant dieback and mortality during both long- and short-term droughts (Rice et al. 2004; Anderegg et al. 2013; Paddock et al. 2013; Pratt et al. 2014) and when plants were experimentally exposed to water stress (Pratt et al. 2008). Xylem cavitation resistance is an important plant functional trait that varies among ecosys- tems (Maherali et al. 2004; Choat et al. 2012) and, at smaller scales, among plant communities (Jacobsen et al. 2007b; Hacke et al. 2009). For species with broad distributions, selection at drier sites and during dry years may lead to intraspe- cific variation among populations in some species (Mencuccini and Comstock 1997; Kavanagh et al. 1999; Kolb and Sperry 1999b; Pratt et al. 2012), although previous studies have also found that some climatically divergent populations do not vary greatly in cavitation resistance (Men- cuccini and Comstock 1997; Matzner et al. 2001; Stout and Sala 2003; Lamy et al. 2011). Species may also be phenotypically plastic in their xylem vulnerability to cavitation, although this has been relatively little studied (Holste et al. 2006; Beikircher and Mayr 2009; Mayr et al. 2010; Fichot et al. 2010; Awad et al. 2010; Plavcova and Hacke 2012). Co-occurring species of chaparral shrubs in southern California exhibit highly divergent levels of cavitation resistance (Davis et al. 1999; Jacobsen et al. 2007a; Pratt et al. 2007b) and include some of the most cavitation resistant angiosperms ever measured (Maherali et al. 318 MADRONO [Vol. 61 Fig. 1. Vulnerability to cavitation curves were analyzed from species occurring in southern California that had been measured at multiple locations. Data were divided into two regions, either the Santa Monica Mountains or the San Gabriel Mountains. Within the Santa Monica Mountains, vulnerability curves were further divided into subpopulations sampled at sites located near the coast and those sampled from inland sites. Selected highways and cities from the study area are indicated on the map and the mean annual precipitation for the sites and regions (as reported in the previously published studies included in Table 1) are shown. 2004). However, large scale mortality of chapar- ral shrubs has been reported during extreme drought years, particularly at a desert-chaparral ecotone (Paddock et al. 2013). Chaparral shrubs exhibit multiple life history types, largely defined by their post-fire response. This includes species that resprout from under- ground storage structures following fire, species that recruit post-fire through the germination of fire-cued seeds, and species that employ both of these strategies. These differential life history types are related to large functional differences between some chaparral shrub species at both the adult (Jacobsen et al. 2007a; Pratt et al. 2007a, b) and seedling (Pratt et al. 2008, 2010, 2012) stages. Furthermore, drought-induced mortality of seed- lings may be common, particularly among post- fire seeding species whose seedlings must survive the protracted summer dry period that occurs in the Mediterranean-type climate region of south- ern California (Frazer and Davis 1988; Thomas and Davis 1989). Selection for higher cavitation resistance may therefore be expected at drier and more interior sites, where mortality has been observed, compared to species occurring in more mesic or coastal sites that experience less water stress. These patterns may become more pro- nounced across broader, regional scales. The cavitation resistance of chaparral shrubs has been extensively studied, particularly in southern California (Jarbeau et al. 1995; Redfeldt and Davis 1996; Davis et al. 1999a, 2002; Jacobsen et al. 2005, 2007a, b; Pratt et al. 2007b), which enabled us to evaluate the impact of differences between sites and regions on vulnerability to cavitation within repeatedly measured species of this plant community. Using previously published and some previously un- published data we examined if cavitation resis- tance of chaparral shrubs 1) varied locally among shrub subpopulations at sites that were located in close proximity but differed in their ocean exposure (coast vs. inland) and 2) varied region- ally between shrub populations from two moun- tain ranges occurring in southern California (Santa Monica Mts. vs. San Gabriel Mts.) (Fig. 1). We predicted that populations from coastally exposed sites would be less cavitation resistant than inland sites, due to the stronger maritime climatic influence along the coast (Vasey et al. 2012). We predicted that popula- tions from the Santa Monica Mountains would be more cavitation resistant than the San Gabriel Mountains due to higher mean annual precipita- tion occurring in the San Gabriel Mountains, consistent with previous studies examining the impact of precipitation or watering treatments on cavitation resistance (Mencuccini and Comstock 1997; Kolb and Sperry 1999b; Helms 2009; Awad et al. 2010). This analysis was complicated by the recent suggestion that some methods used to construct 2014] JACOBSEN ET AL.: CHAPARRAL CAVITATION RESISTANCE 319 vulnerability to cavitation curves may not pro- duce reliable data (Choat et al. 2010; Cochard et al. 2010). In particular, it has been suggested that centrifuge-based data may produce a measure- ment artifact, especially in species with long vessels. The dataset compiled for the present study provided an ideal test of this suggestion, because both dehydration and standard centri- fuge curves have been used to construct vulner- ability curves on the same chaparral species at the same sites and during the same season. Addi- tionally, many chaparral shrub species have long maximum vessels, including several species with maximum vessel lengths greater than one meter in length (Jacobsen et al. 2012). We also evaluated the impact of the season (wet vs. dry) during which curves were constructed, because this has previously been reported to impact vulnerability to cavitation (Kolb and Sperry 1999a; Jacobsen et al. 2007b). Methods Previously published vulnerability curves con- ducted on chaparral shrub species from the Santa Monica Mountains and the San Gabriel Moun- tains of southern California were compiled (Jarbeau et al. 1995; Redfeldt and Davis 1996; Davis et al. 1999a; Davis et al. 2002; Jacobsen et al. 2005, 2007a, b, Pratt et al. 2007b). Pre- viously unpublished vulnerability curves mea- sured by the authors from 2004-2012 were also included in analyses (Table 1). This included data for 16 species for which curves were reported that had been measured on the same species across multiple sites (coastal vs. inland within the Santa Monica Mts.), regions (Santa Monica Mts. vs. San Gabriel Mts.), and/or seasons (wet vs. dry) and using two different methods (dehydration vs. centrifuge) (Fig. 1). Wet season vulnerability curves were defined as those that were measured between December and May for any sample year across all of the studies listed above, when plants were hydrated in the field (Jacobsen et al. 2008). Rainfall typically occurs during November to April at these sites and regions, with plant water status declining in July following the use of soil moisture reserves (Jacobsen et al. 2008). Dry season vulnerability curves were defined as those that were measured during July to October when plants are dehydrated in the field and were no longer growing (Jacobsen et al. 2007b). Dehydration vulnerability curves were con- structed using the methods described in Jacobsen et al. (2007a). Briefly, large branches, longer than the longest vessels as determined via air injection, were collected in the field. Branches were allowed to dehydrate for varying periods of time in the laboratory before being tightly bagged overnight to allow them to equilibrate. Water potentials of branches were measured the next morning and then branches were cut underwater, alternately at each end, to excise a 0.10-0.14 m central stem segment for measurement. This sampling proto- col reduces the negative pressure in the xylem prior to the extraction of the measured sample, which may be important in order to avoid artifoct when sampling some species (Wheeler et al. 2013). Hydraulic conductivity (Kh) of stem segments was measured both immediately fol- lowing excision of the stem segment and follow- ing a one hour flush at 100 kPa using a degassed and ultra-filtered solution (Jacobsen et al. 2007a). The percentage loss in hydraulic conductivity (PLC) was determined for each stem segment (Sperry et al. 1988). Centrifuge vulnerability curves were construct- ed using the methods described in Pratt et al. (2007a). Stem segments 0.14 m or 0.27 m in length were excised, underwater, from larger branches that had been harvested in the field. Segments were then flushed as described above before being subjected to increasing negative water potentials by being spun in a custom centrifuge rotor (Alder et al. 1997; Tobin et al. 2013). Hydraulic conductivity was measured following each spin and these were used to calculate the PLC at each imposed water potential relative to the initial flushed value. Vulnerability curves, whether generated using dehydration or centrifuge based data, were plotted as the PLC with declining water potential. All data from previous studies were replotted and reanalyzed for the current study. Water potential and PLC data from each curve and study were fit using a Weibull curve (Microsoft Excel 2010, Microsoft, Redmond, WA). This curve was then used to calculate the water potential at 50% loss in hydraulic conductivity (P50) for each curve. The P50 value was used as an estimate of species cavitation resistance, as this is the most com- monly compared and reported value in species comparisons (Choat et al. 2012). Additionally, the alpha (shape parameter), which determines if a curve is concave or convex, and the beta (scale parameter), which determines the “stretch” of the curve along the x-axis, for the Weibull fit were recorded. These fit parameters were compared later, if we found that there was a significant shift in the P50 between our comparison groups, to determine if the change in cavitation resistance was due to a shift in the shape or scale of the vulnerability to cavitation curve. Differences in the P50 of species as impacted by site, region, season, or method were analyzed using paired t-tests with data paired by species (Release 16.1.0, Minitab, State College, PA). For these comparisons, data were matched by the parameters that were not being evaluated in the current comparison and a single pair of mean P50 was generated for each species (e.g., the site comparisons were conducted on species data 320 MADRONO [Vol. 61 c/l •<-' I u o 1 W a ^ U ^ ’3 +1 iH ^ ^ Vk O ^ O j_ m D g'W IS s; (D ^ 5^ a I 00 u p n O . w C/i*' S . c 'o ^ o « ^•ag- £Q « I a ^ ’s J c 2 - H < H > < u o H I S! a o c a « 3= CO "d u 'd a 03 a a G ’O 0 03 < w z J Z) > U. ^ 0|1 g - - 1 ^ G 2 G CD z a ^ z I -Z G ^ a 2 33 -^2 2 G G S I ^ -d 0^ s u ^ - -d -J <13 " CQ G d3 < 3 ^ E- > G in o ^ . d -G Q :! d! « d G d 03 CO S O ■d G, 03 d [G- O . . O \o ai o^ - Ov ^ o 3 .2 ^ d > . t3 d -d ^ ^ 03 w G G d 03 £1 « _g g.2o |2 S d a Si s'l^ O d r, (N 'd d^ __ a 'd ON d ^ -H ^ -a d C -S HG 03 G 03 ^ a ^ a o 3^ a +-< d d i-j ■d .. 3 5 a a -Z G d s a 8 a - § a 1^ +1 +1 bO < I >2 “i 3.1 s: d ~§ <3 03 ,0 y CO P cd TO iT) Q S 'd (N ^ - G* ^ G ^ G 4D ^ d 3 •G o .a 0 2 o Q t— I d »-H [G. cd a m O •a ^ 2 ^ p ^ 9" 3 a ON G On a X) d d2 o o ; d-_ 1 rn (N r- a r-- 00 cn d- d- d a d d- d fSi (N m ro (N m ro fNl m ro ro m -H O 0.05 for all; see slope and intercepts with 95% confidence intervals in Fig. 4). Similarly, the intercepts of regressions were not different from zero in any case (Fig. 4; P > 0.05 for all). Discussion Methodological Considerations We found that both centrifuge and dehydra- tion based vulnerability curves produced similar estimates of cavitation resistance. This is consis- tent with several recent studies that have shown that these methods produce similar results (Jacobsen and Pratt 2012; Sperry et al. 2012; 324 MADRONO [Vol. 61 Tobin et al. 2013). Some prior reports of disagreement between methods may be due to the use of a different centrifuge technique (Cochard et al. 2010) or the confounding of technique comparisons with the pooling of data generated from both flushed and non-flushed curves (Cochard and Delzon 2013). Dehydration curves are most often based on percentage loss in conductivity (PLC) obtained by taking initial, or native, measurement of conductivity (K^) fol- lowed by the flushing of samples to measure the maximum potential conductivity (Ksmax) (howev- er, see Jacobsen and Pratt 2012). These curves would not be expected to be comparable to centrifuge curves generated using non-flushed stems, because non-flushed curves begin at the native rather than the .^smax (see Sperry et al. 2012). This results in non-flushed centrifuge curves exhibiting a flat initial portion of their curve, until the point that the centrifuged pressure exceeds the native pressure of the samples, which often results in a sigmoidal in shape. Non-flushed centrifuge curves may dras- tically differ in shape compared to flushed curves (Sperry et al. 2012). Thus, the agreement of methods in the present study may be at least partially due to the careful matching of samples and the inclusion of only flushed-sample centri- fuge curves. Another issue of repeated concern with centri- fuge-based curves has been the influence of open vessels (i.e., vessels that are longer than the measured sample length and therefore do not contain a terminal vessel element within the sample) (Choat et al. 2010; Cochard et al. 2010). This issue has been experimentally exam- ined using both alteration of the number of open vessels (Jacobsen and Pratt 2012) and short and long vesselled comparisons (Sperry et al. 2012; Tobin et al. 2013) and open vessels have not been found to impact vulnerability curves when the standard Alder et al. (1997) centrifuge technique is used. Additionally, this issue has been tied to the shape of vulnerability curves, with particular concern raised about “r” shaped (or exponential shaped) vulnerability curves (Cochard and Del- zon 2013), although careful study of r-shaped curve species, including single vessel air injection measures, have confirmed the validity of r-shaped curves (Sperry et al. 2012; Christman et al. 2012; Tobin et al. 2013). Chaparral species exhibit a wide range in the shape and scale of their vulnerability curves and are also very divergent in the length of xylem vessels found among species, providing the opportunity for us to evaluate both the impact of long vessels on curves as well as the validity of vulnerability curves of differing shapes. Centri- fuge and dehydration based vulnerability curves were not different across species. This included species with maximum vessel lengths as long as 1.4 m (Ro), 1.1m (Fc), and 1.2 m (Ri) (Jacobsen et al. 2012; See Table 1 for species codes), which would clearly have vessels open through the measured samples. There was also methodolog- ical agreement between species with shorter maximum vessel lengths, including 0.3 m (Af), 0.5 m (Cm), and 0.3 m (Cs) (Jacobsen et al. 2012; See Table 1 for species codes). Additionally, when matched for region and season, we found good agreement between dehydration and centri- fuge curves, including for species with r-shaped curves. Consistent with the findings of previous studies (Sperry et al. 2012; Christman et al. 2012; Tobin et al. 2013), this suggests that these curves describe a genuine species strategy. This finding highlights the diversity of hydraulic strategies employed by plants, including many chaparral shrub species. Seasonal Effects Cavitation resistance differed depending on the season during which measurements were con- ducted. Across species there was a general shift toward increased resistance from the wet season to the dry season. This is consistent with a general “hardening” of the xylem as plants reduced growth, produced latewood, and transitioned into the dry season during which they experience lower water potentials. This same pattern has been described previously in arid and semi-arid plant communities (Kolb and Sperry 1999a; Jacobsen et al. 2007b; Helms 2009). Changes in xylem function seasonally have also been de- scribed for other hydraulic traits and in other plant communities, including large seasonal changes in hydraulic conductivity (Sperry et al. 1987; Tibbetts and Ewers 2000; Jacobsen et al. 2007b; Choat et al. 2010). This suggests that plant hydraulics may be quite variable intra-annually and that measurements of plant hydraulics conducted at a single sampling time may not present a complete picture of the hydraulic strategy of a species. Additionally, measurement of PLC or conductivity across seasons may not be comparable to measurements, such as vulner- ability curves, that are measured at a single time point. Finally, increased resistance later in the season suggests that plants may be particularly sensitive to early season drought and this may be an important area of research in predicting plant responses to more variable climate patterns. Geographic Variability Across shorter distances and between subpop- ulations within a single mountain range (i.e., sites), we did not find that cavitation resistance varied even though both the mean annual precipitation and temperature differed among coastal and inland sites. This is consistent with 2014] JACOBSEN ET AL.; CHAPARRAL CAVITATION RESISTANCE 325 recent research on a single chaparral lineage, Arctostaphylos Adans., that found that cavitation resistance did not differ between foggy coastal sites and non-foggy inland sites along the central coast of California (Jacobsen and Pratt 2013), even though the water potentials of plants at these sites differed (Vasey et al. 2012). It may be that across relatively short distances (<10 km), enough genetic material is exchanged to prevent subpopulations from diverging in response to local climatic conditions. This may also partially explain the high mortality observed among some species during drought where they occur at the drier end of their range (Paddock et al. 2013). Across larger distances and among disjunct populations occurring within two mountain ranges, cavitation resistance significantly differed. Populations occurring in the San Gabriel Moun- tains, which receive considerably more mean annual precipitation (855 mm) than the Santa Monica Mountains (460 mm), were more vulner- able to cavitation. This was consistent with our prediction and with a previous study that compared two chaparral shrub species between these two regions (Helms 2009). This is also consistent with previous studies that have de- scribed populations from more mesic areas as being more vulnerable to cavitation in some species (Mencuccini and Comstock 1997; Kava- nagh et al. 1999; Kolb and Sperry 1999b) as well as findings from shrub species occurring in other Mediterranean-type climate regions (Pratt et al. 2012). Additionally, the San Gabriel Mountains sites experience winter freezing and freeze-thaw events that may impact the hydraulics of plants from the San Gabriel Mountains (Davis et al. 1999b; Cordero and Nilsen 2002), although it is not clear the impact this may have on the present comparison because freezing also occurs in some sites within the Santa Monica Mountains (Pratt et al. 2005). Finally, soil properties have also been shown to impact vulnerability to cavitation (Sperry and Hacke 2002) and it should be noted that soil composition was different between these regions, with the Santa Monica Mountains composed predominantly of sedimentary derived soils and the San Gabriel Mountains composed predominantly of plutonic derived soils (Schoen- herr 1992). Hydraulic Trait Diversity It is not clear whether these differences represent phenotypic plasticity in xylem function or if they represent genetic differences, and it would require establishment of a common garden with plants from both regions to determine. However, these data suggest that there is intraspecific variability in cavitation resistance among these chaparral shrub species. California climate is expected to warm significantly under most climate change scenarios and patterns of precipitation are expected to change, although California topography makes fine-scale predic- tions difficult (Cayan et al. 2008). Thus, preser- vation of hydraulic trait diversity among chap- arral species and populations may be key to long- term resilience of chaparral shrub species and communities. Conclusions Chaparral species vary in their cavitation resistance seasonally and among populations that occurred in regions with varying mean annual precipitation, suggesting that the timing of water stress and the traits of different populations are both important factors likely to influence the long-term resilience of chaparral communities. Additionally, chaparral species are highly vari- able in their vulnerability to cavitation, including the presence of some highly vulnerable species that exhibit r-shaped '(or exponential shaped) vulnerability curves as well as more resistant species. More research is needed to more fully understand these very different, yet co-occurring, hydraulic strategies. Acknowledgments RBP thanks NSF (IOS-0845125) and the Andrew W. Mellow Foundation for support of this project. ALJ thanks NSF (lOS- 1252232). Literature Cited Alder, N. N., W. T. Pockman, J. S. Sperry, and S. Nuismer. 1997. Use of centrifugal force in the study of xylem cavitation. Journal of Experimental Botany 48:665-674. Anderegg, W. R. L., L. PlavcovA, L. D. L. Ander- EGG, U. G. Hacke, J. A. Berry, and C. B. Field. 2013. Drought’s legacy: multiyear hydraulic dete- rioration underlies widespread aspen forest die-off and portends increased risk. Global Change Biology 19:1188-1196. Awad, H., T. Barigah, E. Badel, H. Cochard, and S. Herbette. 2010. Poplar vulnerability to xylem cavitation acclimates to drier soil conditions. Physiologia Plantarum 139:280-288. Baldwin, B. G., D. H. Goldman, D. J. Keil, R. Patterson, T. J. Rosatti, and D. H. Wilken, (eds.). 2012. The Jepson Manual: vascular plants of California, 2nd ed. University of California Press, Berkeley, CA. Beikircher, B. and S. Mayr. 2009. Intraspecific differences in drought tolerance and acclimation in hydraulics of Ligustrum vulgare and Viburnum lantana. Tree Physiology 29:765-775. Cayan, D. R., E. P. Maurer, M. D. Dettinger, M. Tyree, and K. Hayhoe. 2008. Climate change scenarios for the California Region. Climate Change 87:S21-S42. Choat, B., W. M. Drayton, C. Brodersen, M. A. Matthews, K. A. Shackel, H. Wada, and A. J. McElrone. 2010. Measurement of vulnerability to water stress-induced cavitation in grapevine: a 326 MADRONO [Vol. 61 comparison of four techniques applied to long- vesseled species. Plant, Cell & Environment 33:1502-1512. , S. Jansen, T. J. Brodribb, H. Cochard, S. Delzon, R. Bhaskar, S. J. Bucci, T. S. Field, S. M. Gleason, U. G. Hacke, A. L. Jacobsen, F. Lens, H. Maherali, J. Martinez-Vilalta, S. Mayr, M. Mencuccini, P. J. Mitchell, A. Nardini, j. Pittermann, R. B. Pratt, J. S. Sperry, M. Westoby, I. J. Wright, and A. E. Zanne. 2012. Global convergence in the vulnera- bility of forest to drought. Nature 491:752-756. Christman, M. A., J. S. Sperry, and D. D. Smith. 2012. Rare pits, large vessels and extreme vulner- ability to cavitation in a ring-porous species. New Phytologist 193:713-720. CoCHARD, H. AND S. Delzon. 2013. Hydraulic failure and repair are not routine in trees. Annals of Forest Science 70:659-661. , S. Herbette, T. Barigah, E. Badel, M. Ennajeh, AND A. ViLAGROSA. 2010. Does sample length influence the shape of xylem embolism vulnerability curves? A test with the Cavitron spinning technique. Plant, Cell & Environment 33:1543-1552. Cordero, A. R. and E. T. Nilsen. 2002. Effects of summer drought and winter freezing on stem hydraulic conductivity of Rhododendron species from contrasting climates. Tree Physiology 22: 919-928. Davis, S. D., F. W. Ewers, J. S. Sperry, K. A. PoRTwooD, M. C. Crocker, and G. C. Adams. 2002. Shoot dieback during prolonged drought in Ceanothus chaparral of California: a possible case of hydraulic failure. American Journal of Botany 89:820-838. , , J. Wood, J. J. Reeves, and K. J. Kolb. 1999a. Differential susceptibility to xylem cavitation among three pairs of Ceanothus species in the Transverse Mountain Ranges of southern California. Ecoscience 6:180-186. , J. S. Sperry, and U. G. Hacke. 1999b. The relationship between xylem conduit diameter and cavitation caused by freezing. American Journal of Botany 86:1367-1372. Fichot, R., T. S. Barigah, S. Chamaillard, D. Le Thiec, F. Laurans, H. Cochard, and F. Brignolas. 2010. Common trade-offs between xylem resistance to cavitation and other physiolog- ical traits do not hold among unrelated Populus deltoides X Populus nigra hybrids. Plant, Cell & Environment 33:1553-1568. Frazer, J. M. and S. D. Davis. 1988. Differential survival of chaparral seedlings during the first summer drought after wildfire. Oecologia 76:215- 221. Hacke, U. G., A. L. Jacobsen, and R. B. Pratt. 2009. Xylem function of arid-land shrubs from California, USA: an ecological and evolutionary analysis. Plant, Cell & Environment 32:1324—1333. Helms, A. M. 2009. The effect of severe drought on xylem water transport in two chaparral species in the San Gabriel and Santa Monica Mountains. Honors thesis, Pepperdine University, Malibu, CA. Holste, E. K., M. j. Jerke, and S. L. Matzner. 2006. Long-term acclimatization of hydraulic properties, xylem conduit size, wall strength and cavitation resistance in Phaseolus vulgaris in response to difference environmental effects. Plant, Cell & Environment 29:836-843. Jacobsen, A. L. 2008. Comparative community physiology: nonconvergence in water relations among three semi-arid shrub communities. New Phytologist 180:100-113. . 2013. Vulnerability to cavitation of central California Arctostaphylos (Ericaceae): a new anal- ysis. Oecologia 171:329-334. , F. W. Ewers, R. B. Pratt, W. A. Paddock III, AND S. D. Davis. 2005. Do xylem fibers affect vessel cavitation resistance? Plant Physiology 139:546-556. AND R. B. Pratt. 2012. No evidence for an open vessel effect in centrifuge-based vulnerability curves of a long-vesselled liana {Vitis vinifera). New Phytologist 194:982-990. , , S. D. Davis, and F. W. Ewers. 2007b. Cavitation resistance and seasonal hydrau- lics differ among three arid Californian plant communities. Plant, Cell & Environment 30: 1599-1609. , , F. W. Ewers, and S. D. Davis. 2007a. Cavitation resistance among 26 chaparral species of southern California. Ecological Mono- graphs 77:99-115. , , M. F. Tobin, U. G. Hacke, and F. W. Ewers. 2012. A global analysis of xylem vessel length in woody plants. American Journal of Botany 99:1583-1591. Jarbeau, j. a., F. W. Ewers, and S. D. Davis. 1995. The mechanism of water-stress-induced embolism in two species of chaparral shrubs. Plant, Cell & Environment 18:189-196. Kavanagh, K. L., B. j. Bonds, S. N. Aitken, B. L. Gartner, and S. Knowe. 1999. Shoot and root vulnerability to xylem cavitation in four popula- tions of Douglas-fir seedlings. Tree Physiology 19:31-37. Kolb, K. J. and J. S. Sperry. 1999a. Transport constraints on water use by the Great Basin shrub, Artemisia tridentata. Plant, Cell & Environment 22:925-935. . 1999b. Differences in drought adaptation between subspecies of sagebrush {Artemisia triden- tata). Ecology 80:2373-2384. Lamy, J.-B., L. Bouffier, R. Burlett, C. Plomion, H. CocHARD, AND S. Delzon. 2011. Uniform selection as a primary force reducing population genetic differentiation of cavitation resistance across a species range. PLoS ONE 6:e23476, doi: 1 0. 1 37 1/journal. pone. 0023476 Maherali, H., W. T. Pockman, and R. B. Jackson. 2004. Adaptive variation in the vulnerability of woody plants to xylem cavitation. Ecology 85:2184-2199. Matzner, S. L., K. J. Rice, and J. H. Richards. 2001. Intra-specific variation in xylem cavitation in interior live oak {Quercus wislizenii A. DC.). Journal of Experimental Botany 52:783-789. Mayr, S., B. Beikircher, M. A. Obkircher, and P. Schmid. 2010. Hydraulic plasticity and limitations of alpine Rhododendron species. Oecologia 164: 321-330. McDowell, N., W. T. Pockman, C. D. Allen, D. D. Breshears, N. Cobb, T. Kolb, J. Plaut, J. Sperry, A. West, D. G. Williams, and E. A. Yepez. 2008. Mechanisms of plant survival and 2014] JACOBSEN ET AL.: CHAPARRAL CAVITATION RESISTANCE 327 mortality during drought: why do some plants survive while others succumb to drought? New Phytologist 178:719-739. Mencuccini, M. and J. Comstock. 1997. Vulnera- bility to cavitation in populations of two desert species, Hymenodea salsola and Ambrosia dumosa, from different climatic regions. Journal of Exper- imental Botany 48:1323-1334. Paddock, W. A. S., Ill, S. D. Davis, R. B. Pratt, A, L. Jacobsen, M. F. Tobin, J. Lopez-Portillo, AND F. W. Ewers. 2013, Factors determining mortality of adult chaparral shrubs in an extreme drought year in California. Aliso 31:49-57. PlavcovA, L. AND U. G. Hacke. 2012. Phenotypic and developmental plasticity of xylem in hybrid poplar saplings subjected to experimental drought, nitrogen fertilization, and shading. Journal of Experimental Botany 63:6481-6491. Pratt, R. B., F. W. Ewers, M. C. Lawson, A. L. Jacobsen, M. Brediger, and S. D. Davis. 2005. Mechanisms for tolerating freeze-thaw stress of two evergreen chaparral species: Rhus ovata and Malosma lamina (Anacardiaceae). American Jour- nal of Botany 92:1102-1113. , A. L. Jacobsen, F. W. Ewers, and S. D. Davis. 2()07a. Relationships among xylem trans- port, biomechanics and storage in stems and roots of nine Rhamnaceae species of California chapar- ral. New Phytologist 174:787-798. , , K. A. Golgotiu, J. S. Sperry, F. W. Ewers, and S. D. Davis. 2007b. Life history type and water stress tolerance in nine California chaparral species (Rhamnaceae). Ecological Mono- graphs 77:239-253. , , S. M. Jacobs, and K. J. Esler. 2012. Xylem transport safety and efficiency differ among fynbos shrub life history types and between two sites differing in mean rainfall. International Journal of Plant Sciences 173:474-483. , , R. Mohla, F. W. Ewers, and S. D. Davis. 2008. Linkage between water stress toler- ance and life history type in seedlings of nine chaparral species (Rhamnaceae). Journal of Ecol- ogy 96:1252-1265. , , A. R. Ramirez, A. M. Helms, C. A. Traugh, M. F. Tobin, M. S. Heffner, and S. D. Davis. 2014. Mortality of resprouting chaparral shrubs after a fire and during a record drought: physiological mechanisms and demographic conse- quences. Global Change Biology 20:893-907. , G. B. North, A. L. Jacobsen, F. W. Ewers, and S. D. Davis. 2010. Xylem root and shoot hydraulics is linked to life history type in chaparral seedlings. Functional Ecology 24:70-81. Redtfeldt, R. a. and S. D. Davis. 1996. Physiolog- ical and rporphological evidence of niche segrega- tion between two co-occurring species of Adenos- toma in California chaparral. Ecoscience 3:290-300. Rice, K. J., S. L. Matzner, and W. Byer. 2004. Patterns of tree dieback in Queensland, Australia: the importance of drought stress and the role of resistance to cavitation. Oecologia 139:190-198. Schoenherr, a. A. 1992. A natural history of California. University of California Press, Berke- ley, CA. Sperry, J. S., M. A. Christman, J. M. Torrez-Ruiz, H. Taneda, and D. D. Smith. 2012. Vulnerability curves by centrifugation: is there an open vessel artifact, and are “r” shaped curves necessarily invalid? Plant, Cell & Environment 35:601-610. , J. R. Donnelly, and M. T. Tyree. 1988. A method for measuring hydraulic conductivity and embolism in xylem. Plant, Cell & Environment 11:35^0. and U. G. Hacke. 2002. Desert shrub water relations with respect to soil characteristics and plant functional type. Functional Ecology 16: 367-378. , N. M. Holbrook, M. H. Zimmermann, and M. T. Tyree. 1987. Spring filling of xylem vessels in wild grapevine. Plant Physiology 83:414-417, Stout, D. L. and A. Sala. 2002. Xylem vulnerability to cavitation in Pseudo tsuga menziesii and Pinus ponderosa from contrasting habitats. Tree Physiol- ogy 23:43-50. Thomas, C. M. and S. D. Davis. 1989. Recovery patterns of three chaparral shrub species after wildfire. Oecologia 80:309-320. Tibbetts, T. J. and F. W. Ewers. 2000. Root pressure and specific conductivity in temperate lianas: exotic Celastrus orbiculatus (Celastraceae) vs. native Vitis riparia (Vitaceae). American Journal of Botany 87:1272-1278. Tobin, M. F., R. B. Pratt, A. L. Jacobsen, and M. E. De Guzman. 2013. Xylem vulnerability to cavita- tion can be accurately characterized in a species with long vessels using a centrifuge method. Plant Biology 15:496-504. Tyree, M. T. and M. H. Zimmermann. 2002. Xylem structure and the ascent of sap. 2nd ed. Springer- Verlag, Berlin, Germany. Vasey, M. C., M. E. Loik, and V. T. Parker. 2012. Influence of summer marine fog and low cloud stratus on water , relations of evergreen woody shrubs {Arctostaphylos: Ericaceae) in the chaparral of central California. Oecologia 170:325-337. Warton, D. L, I. J. Wright, D. S. Falster, and M. Westoby. 2006. Bivariate line-fitting methods for allometry. Biological Reviews 81:259-291. Wheeler, J. K., B. A. Huggett, A. N. Tofte, F. E. Rockwell, and N. M. Holbrook. 2013. Cutting xylem under tension or supersaturated with gas can generate PLC and the appearance of rapid recovery from embolism. Plant, Cell & Environment 36:1938-1949. Madrono, VoL 61, No. 4, pp. 328-338, 2014 FUNCTIONAL TRAIT DIFFERENCES BETWEEN WEEDY AND NON-WEEDY PLANTS IN SOUTHERN CALIFORNIA Evan D. MacKinnon*, R. Brandon Pratt, and Anna L. Jacobsen Department of Biology, California State University, Bakersfield, 9001 Stockdale Hwy, Bakersfield, CA 93311 ^evandmackinnon@yahoo.com Abstract Weedy plants may have unique functional traits that distinguish them from other species. We categorized four annual plant species as weedy and five as non-weedy based on their prevalence in disturbed and invaded environments. This designation was tested in a field trial where we scattered equal numbers of viable seeds in 20 different plots and monitored density and cover over three months. The plants a priori designated as weedy had significantly greater cover and densities than species designated as non-weedy. We hypothesized that a suite of functional traits would define the weedy plant habit. We tested this hypothesis by comparing functional traits between weedy and non- weedy plant species. A principal components analysis (PCA) identified three distinct ecological clusters among the analyzed species (weedy forbs, non-weedy forbs, and grasses). The weedy habit was different from the non-weedy habit in several traits (slower growth, heavier diaspores, earlier flowering initiation, and dormant seeds requiring cold-stratification for germination). Weedy annuals in southern California appear to share a suite of traits, suggesting that their success as weeds is linked to adaptive traits. Further understanding of the traits shared among weedy plant species may help screen for native plants that are valuable for ecological restoration of highly invaded landscapes. Key Words: California, functional trait, restoration, weedy plants. Functional traits describe morphological, physiological, or phenological features that influ- ence the resource acquisition and fitness of a species (Violle et al. 2007). Environmental and evolutionary forces drive the set of functional traits exhibited by a given species. Functional traits analyzed across different species can help identify plants with similar suites of traits (Gitay and Noble 1999; Violle et al. 2007). Describing functional traits is of value because, for example, it provides predictive power for how groups of species may respond to environmental change, why some species become invasive, and how species influence ecosystem function. Functional trait studies may also help identify new candidate species for ecological restoration, because the underlying causes of plant success may be trait- based (Funk et al. 2008). We hypothesized that plants with a weedy habit would have similar suites of functional traits and that they would differ from co- occurring less weedy species. In determining which traits to sample, the ecological context of our study location was an important consider- ation. The southern San Joaquin Valley is widely disturbed, primarily by agricultural activities, and is an arid environment with a Mediterranean- type rainfall pattern characterized by cool moist winters and hot dry summers (Germano et al. 2011). The native upland vegetation of the region is primarily saltbush scrub habitat with abundant winter annual forbs. The suite of analyzed traits was based on previous work and selected based on the ecolog- ical context of the study region (see Table 1 for a list of traits and predictions). Plant traits are shaped by the environment and tradeoffs. Hab- itat productivity and disturbance are two major environmental factors that shape plant traits (Grime 1977). In productive habitats, successful plants develop competitive traits associated with resource capture, whereas in low resource habi- tats stress tolerance traits predominate. In highly disturbed habitats ruderal traits dominate, such as the ability to colonize. Tradeoffs among suites of traits preclude plants from being successful in all contexts (Grime 1977; Kimball et al. 2013). For example, the traits that confer competitive ability (rapid growth and resource acquisition) are not compatible with stress tolerance traits (dense and tough tissues). We categorized traits as competitive, stress-tolerant, and ruderal in order to frame our traits within an established framework describing ecological strategies (Grime 1977). Some of our predictions about traits run counter to general patterns among weedy forbs and grasses, but make sense in the context of our study region. For instance, because of the arid environment of the southern San Joaquin Valley, we expected successful plants to be drought-tolerant and have relatively mas- sive seeds, even though these characteristics are not usually associated with invasive weeds (however, see Funk and Vitousek 2007), which 2014] MACKINNON ET AL.: FUNCTIONAL TRAIT DIFFERENCES IN PLANTS 329 Table 1. Traits That are Most Likely Related to Weediness in the San Joaquin Valley. Traits are organized into the life-history strategy they most closely measure, and methods for quantifying each trait are outlined. Included are outcomes that were predicted for the weedy and non-weedy habit. Table and predictions include information from Cleland et al. 2012 (a), Cleland 201 1 (b), Funk and Zachary 2010 (c), Cramer et al. 2008 (d), Funk et al. 2008 (e), Grotkopp and Rejmanek 2007 (f), Venable 2007 (g), Allen 2004 (h), Lake and Leishman 2004 (i), Seabloom et al. 2003 (j), Tyree et al. 2003 (k), Grime 2001 (1), Eriksson 2000 (m), Schutz and Rave 1999 (n), Casper and Jackson 1997 (o), Eliason and Allen 1997 (p), Philippi 1993 (q), Eissenstat 1991 (r), Poorter and Remkes 1990 (s), Venable and Brown 1988 (t), Wilson 1988 (u), Bazzas et al. 1987 (v), Grime 1984 (w), Tiffney 1984 (x), Mack and Pyke 1983 (y), Grime and Hunt 1975 (z). Ecological Method of strategy Trait measurement Trait-specific prediction; Literature review Weedy Non- Weedy Competitive Growth rate Maximum relative growth rate Maximum Maximum total- biomass plant biomass Above-ground Specific leaf area competition Leaf area ratio Leaf weight ratio Below-ground competition Rootishoot ratio Competitive influence Mass of an invasive plant Germination requirements Final percent germination (cold-stratified vs. non- stratified) Ruderal Diaspore mass Time to germination (cold-stratified vs. non-stratified) Diaspore mass Phenology Time to flowering period initiation Stress- Drought- tolerant tolerance Seedling drought- tolerance Rapid growth is a competitive trait higher which allows plants to rapidly occupy space and deplete resources (s, w, z). Large plants are strong competitors higher because they have a greater demand for limiting resources (p, u, w). Greater leaf surface area exposed higher to sunlight allows for more competitive carbon capture and rapid growth (f, i, s, z). Greater investment in photosynthetic higher tissue than in respiring tissues allows for more competitive carbon capture and rapid growth (f, i, s, z). Greater biomass invested in leaves higher than in other tissues allows for more competitive carbon capture which promotes rapid growth (f, i, s, z). Heavy biomass investment in roots higher allows more soil volume to be exploited, so that plants are more competitive for water and nutrients (c, j, o, r, z). Competitive species will reduce the lower biomass of an invasive plant (b, e, j, 1, o). Species that germinate early (without higher cold-induced germination) are more competitive because they use-up space and resources before other plant seeds have germinated (a, e, g, h, n, q, y). Species that germinate rapidly higher under un-stratified conditions exclude other species by growing early (a, e, g, h, n, q, y). Smaller diaspores may be more larger broadly dispersed, and their small size may enable a greater production of total diaspores per plant (d, e, j, m, t, x), but large diaspores have greater energy reserves for rapid growth (s, w, z). Species that flower earlier can shorter complete their lifecycle over a shorter rainfall period (h, i, v). Plants that can persist during periods higher when resources are temporarily scarce are more likely to survive and reproduce during drought (c, f, k, z). lower lower lower lower lower lower higher lower higher smaller longer lower 330 MADRONO [Vol. 61 Table 2. Nine Annual Plant Species were Compared in this Study. These species were used to identify the traits that relate to the “weedy” tendency of some native species to persist in human-disturbed and invaded San Joaquin Valley environments. Scientific names and authorities follow Baldwin et al. 2012. Species Family Abbreviation Habit Bromus madritensis (L.) Husn. subsp. rubens Poaceae Bm Weedy (Invasive) Ambrosia acanthicarpa Hook. Asteraceae Aa Weedy Amsinckia menziesii (Lehm.) A. Nelson & J.F. Macbr. Boraginaceae Am Weedy Heterotheca grandijlora Nutt. Asteraceae Hg Weedy Bromus carinatus Hook. & Arn. var. carinatus Poaceae Be Non-weedy Clarkia imguiculata Lindl. Onagraceae Cu Non-weedy Lasthenia calif ornica Lindl. Asteraceae Lc Non-weedy Layia platyglossa (Fisch.& C.A.Mey) A. Gray Asteraceae Lp Non-weedy Phacelia tanacetifolia Benth. Boraginaceae Pt Non-weedy more typically avoid drought and produce numerous small seeds for wide dispersal (Table 1; Tiffney 1984; Seabloom et al. 2003). To test our hypothesis and predictions, we selected nine species common in southern Cali- fornia and performed a field experiment at a decommissioned cotton field in order to deter- mine if plants deemed weedy based on field observations were indeed more successful when starting with the same number of viable seeds in the soil. We then conducted greenhouse- and laboratory-based measurements to test for func- tional trait differences among five non-weedy species and four weedy species (see Table 1 for traits and predictions). Methods Project Set-up and Species Selection In the spring 2012, we identified native annual plant species within the boundaries of a former cotton field, located on the campus of California State University, Bakersfield. We found three native species and one invasive species that we classified as weedy (Table 2). Species were considered weedy if they met all four of the following criteria: 1. they occurred in disturbed areas around Bakersfield; 2. they were wide- spread in the United States or in California; 3. they were described in Baldwin et al. (2012) as occurring in disturbed areas; and 4. they were described as weeds in botanical literature (Whit- son et al. 2001; DiTomaso and Healy 2007). We categorized three native herbaceous species and a native grass as non-weedy because they did not meet all of the aforementioned criteria (Table 2). Our designation of these species as non-weedy does not indicate that these species are never considered weedy; instead, we used this term to indicate that these species are less weedy than those we have designated as weedy. Seeds from weedy species were collected from specimens growing on campus while seeds for non-weedy species were purchased from a local southern California native seed supplier (S&S Seeds, Carpinteria, CA). We used both grasses and forbs for this study. The invasive grass Bromus madritensis L. subsp. rubens (L.) Husn. was included in our study as a representative of the weedy plant habit, while a native brome species, Bromus carinatus Hook. & Arn. var. carinatus, was included within the non- weedy habit. Including a grass species within each habit helped us focus on weedy plant traits while controlling for those trait differences resulting from different growth forms. We conducted a field experiment in the winter and spring (2013) in order to verify our weedy and non-weedy habit designation. Some species may be successful in disturbed environments largely because of the abundance of their seeds in the soil. The field trial controlled for this because we seeded plots with an equal quantity of viable seeds for each species. Viability was determined from germination experiments. In early October 2012, we prepared a field site by mowing an approximately 200 m^ area of a former cotton field. We randomly identified 20 locations within the area where we cleared the first five cm of topsoil from a one m^ area. In a pilot study, this depth was found to remove much of the existing soil seed bank (E. D. MacKinnon, personal observation). In each plot we added approximately 100 viable seeds for each of the nine species. A thin layer of thatch was applied to plots in the form of store-bought straw, as thatch was found in previous re-seeding efforts at the site to be necessary to prevent seeds from blowing away (E.D. MacKinnon, personal observation). The density of individuals in plots and the canopy cover were used to quantify field success. Data were collected in late March 2013, at the peak of the growing season for most ex- perimental species (Daubenmire 1959). Since we only sampled once, our field data did not capture field-based germination and growth dynamics. Later-growing species did not contribute a large percentage of the total canopy area at this sampling time, but they may have later in the 2014] MACKINNON ET AL.; FUNCTIONAL TRAIT DIFFERENCES IN PLANTS 331 growing season. For these species, density of individuals was more informative than canopy cover. We conducted a greenhouse study where we used seeds from all species to perform several experiments. We designed each experiment to evaluate a specific trait that is likely necessary for survival in disturbed and invaded environments of the southern San Joaquin Valley. For all greenhouse experiments, we used soil collected from the on-campus study plot that was sifted to remove existing seeds, and amended this soil with one-quarter part vermiculite by volume. The soil was fine grain sand and the addition of vermic- ulite helped keep soils from cementing in pots. A climate-controlled greenhouse was used with an automatic overhead irrigation system which kept the soil continually moist. No fertilizers were used. For all experiments, containers were randomized and reshuffled every three days to control for the effects of any greenhouse envi- ronmental heterogeneity (Grime and Hunt 1975). Competitive Characteristics Growth rate and maximum biomass. One hundred individuals of each species were planted into containers at a density of one individual per container (2401 inserts, Growers Solution, Coo- keville, TN, USA). We knew from laboratory germination experiments that species germinated at different rates, thus to achieve similar-aged plants we staggered sowing times (Grime and Hunt 1975; Grime 2001). Seedling emergence was nevertheless sporadic, so each container was labeled with the date of seedling emergence to ensure that each sampling harvest would occur on same-aged seedlings. The initial sample harvest occurred two weeks from the precise emergence date. We randomly selected 10 plants from each species for destruc- tive harvest (Grime and Hunt 1975; Swanbor- ough and Westoby 1996). For harvested plants, we removed the above-ground portion of each plant at the soil surface. To separate the roots from the soil, the entire container was submerged in water, and agitated to gently loosen the soil. Once coarse material was separated, a second rinse removed any remaining soil particles. All root material, including fine roots, was likely accounted for in this process because the fine soil particles did not adhere to the roots. Samples were dried for 48 hr at 60°C in a drying oven (WU-050 14-06 Gravity Convection Oven, Thermo Fischer Scientific, Pittsburgh, PA) before measuring dry weight for roots and shoots (CPA2P Sartorius Analytical Balance, Gettingen, Germany). We repeated this process for 10 individuals of each species at two-week intervals for the entire study. We continued sampling for each species until it had flowered and begun to senesce or until a loss in whole-plant biomass was found between two time periods. This loss in mass between time periods was due to senesced leaves and roots. For each sampling period, we calculated the mean relative growth rate (RGR; Hunt 1978; Grime 2001). This calculation, expressed as g g“‘ day~' or simply day”', is the change in mass over time relative to the initial plant mass, because large plants tend to grow relatively slower than small plants (Hunt 1978). To calculate the maximum relative growth rate (RGRmax)? we used the greatest value for mean RGR obtained over the entire experiment (Daw- son et al. 2011; Grime and Hunt 1975). Allocation to leaves and roots. We harvested samples after six-weeks of growth. We chose this time period because none of the species had yet reached reproductive maturity, so they were at a more comparable stage of development (Grime 2001). For ten replicates, we measured the whole- plant leaf area (LI-3100C Leaf Area Meter, Li"Cor, Inc., Lincoln, NE), using only healthy, fully-developed leaves. After all plant material had spent 48 hr in a 60°C drying oven, we measured the dry mass of leaves, stem, and roots separately. From these measurements, we calculated the specific leaf area (leaf area divided by leaf mass [m^ kg”']), leaf area ratio (leaf area divided by the whole-plant mass [m^ kg”']), and leaf weight ratio (leaf mass divided by whole-plant mass ([g g”']; Hunt 1978; Poorter and Remkes 1990). Using the same samples harvested for leaf traits, we measured the root dry mass after roots had spent 48 hr in a 60°C drying oven. For each individual, we divided the root dry mass by the above-ground dry mass to calculate the root: shoot ratio (Grime 2001). This provided a value for the relative contribution invested in below- ground biomass. Competitive influence on a locally abundant invader. The experimental species Bromus madri- tensis subsp. rubens, which was included in our weedy plant habit, is also a locally abundant invasive species. The ability of other plants to compete with this invasive grass may be key to their success. To generate a competitive interac- tion with this invasive grass, we grew B. madritensis together with another plant in 2.8L containers (product code OGTP, Growers Solu- tion, Cookeville, TN). To ensure equal-aged seedlings for all species, we used laboratory- germinated seeds. Attaining equal-aged plants was a priority, because small initial differences in seedling size can lead to compounding competi- tive effects over time (Wilson 1988). We destruc- tively harvested plants after eight weeks of growth, when all species appeared to be growing vigorously, and many were flowering. This period was chosen because the initial signs of leaf senescence were becoming visible in some species. 332 MADROto [VoL 61 For biomass measurements, we could only analyze above-ground portions because entan- gled roots could not be distinguished between species. Using the dry biomass of B. madritensis shoots, we were able to quantify the relative competitive effect that each plant had on this invasive grass (Gracet 1995; Casper and Jackson 1997). This experiment also included the compet- itive influence that B. madritensis had on itself. A larger biomass of B. madritensis implied the competing species had little effect on the growth of the invasive grass, while a lower mass for B. madritensis suggested the competing species was highly competitive. Early germination ability. We allowed seeds to imbibe in three layers of moist paper (Germina- tion paper, Anchor Paper Company, St. Paul, MN) in sterile plastic Petri plates (Chuanren et al. 2004). We placed 20 seeds for each species, evenly divided into eight Petri plates (160 seeds total for each species) into a refrigerator at 3°C for two- months to stimulate winter conditions and cold- stratify seeds (Skordilis and Thanos 1995). As a control, eight Petri plates, each with 20 seeds, were kept in darkness in a climate controlled laboratory and were not cold-stratified. We monitored germination every day for the first seven days and approximately every week thereafter. We considered seeds germinated when the seed coat was cracked, and part of the embryo had emerged (Shipley and Parent 1991). At the end of the 60-day stratification period, we placed both stratified and non-stratified plates together in a room temperature laboratory. Initially, we monitored germination daily to capture the spike in germination that we expected upon removing seeds from the refrigerator. We continued with approximately weekly sampling to monitor germination for 30 d. Ruderal Characteristics Diaspore mass. We m.easured the mass of the entire unit of dispersal, or diaspore (Grime and Hunt 1975; Bekker et al. 1998). The diaspore included the seed and associated dispersal aids, such as awns, pappus, and spines. We sampled and weighed a total of 160 unique diaspores for each species in order to estimate the mean mass per diaspore for each species. Time to flowering initiation. We grew 12 plants of each species in containers (2401 inserts, Growers Solution, Cookeville, TN). Every two days, we surveyed for flowering. For most species we recorded flower-opening when petals were completely unfolded (Stenstrdm and Molau 1992). For B. madritensis, flowering began when the spikelet emerged and the florets began to overlap (Baldwin et al. 2012). For Ambrosia acanthicarpa Hook., we recorded flowers as open as soon as staminate heads became visible. Stress-tolerance Seedling drought-tolerance. For this experi- ment, we used laboratory-germinated seeds to ensure equal-aged seedlings for all species. For ■ every species, we transplanted a germinated seed into one of 12, 2.8L containers (Product code OGTP, Growers Solution, Cookeville, TN) containing pre-moistened soil. After only two days of automatic overhead irrigation set to three | minutes, twice per day, we moved each container i to a non-irrigated portion of the greenhouse. We ! monitored plants daily, and randomized contain- 1 ers by shuffling during every visit. We recorded a seedling as dead based on curled leaves and stem, and browning of tissue (Tyree et al. 2003). After all seedlings appeared to have died, containers were irrigated for two months to confirm, that all seedlings were indeed dead (Tyree et al, 2003). Statistical Analyses For the field experiment, we analyzed percent cover and the density of individuals. For cover data, we used an ordinal regression (IMP version 9.0.0, SAS Institute, Cary, NC), because cover data were rank ordered into one of five different percent cover categories (2.5%, 15,0%, 37.5%, 62.5%, 85.0%; Daubenmire 1959). Data for the abundance of individuals were analyzed with a generalized linear model (JMP version 9.0.0, SAS Institute, Cary, NC). To determine the traits that were most important to the functional grouping of species, we conducted a principal components analysis (PCA; Funk et al. 2008; Diaz et al. 2004). Variables were standardized mean responses for a species derived from each experiment we performed. Since the assumptions for a PCA were not entirely satisfied, we used this analysis only for a descriptive purpose (McGarigal et al. 2000). For each component, only variables with loading scores greater than 0.3, and species positioned with sampling entity loading scores greater than 1.0 were considered to be impor- tant contributors to the principal components (McGarigal et al. 2000). In addition to a PCA, we also analyzed each experiment individually. This is because each experiment was designed to identify a specific functional difference between weedy and non- weedy species. For the variables LWR, LAR, SLA, RGRmax (roots), root:shoot ratio, and competitive ability, differences among species and between habits were analyzed using ANOVA in JMP version 9.0.0 (SAS Institute, Cary, NC). Variables used in the model were habit (weedy vs. non-weedy) and species nested within habit. We 2014] MACKINNON ET AL.: FUNCTIONAL TRAIT DIFFERENCES IN PLANTS 333 Table 3. Results (Mean ±1 SE) from Individual Trait Analyses Showing Relative Relationships BETWEEN NoN"WEEDY AND WEEDY PLANT HABITS. Sample size represents the number of replicates within a species. Asterisks denote significant differences between weedy and non-weedy habits (* = P < 0.05, = p < 0.01, *** = P < 0.001). Variable Units n Weedy Non-weedy Rmax (total plant)*** day~' 10 2.94 ± 0.75 8.45 ± 1.53 Rmax (shoot)* day“‘ 10 2.24 ± 0.55 5.10 ± 1.02 Rmax (root) day~' 10 4.34 ± 1.51 9.65 ± 3.31 Biomass g 10 0.43 ± 0.09 0.36 ± 0.04 LAR mm^ mg ' 10 7.13 ± 1.68 5.69 ± 0.93 LWR g g”' 10 0.30 ± 0.05 0.24 ± 0.01 SLA m^ kg-> 10 23.63 ± 2.28 24.07 ± 3.93 Root:shoot ratio gg ‘ 10 1.68 ± 0.43 1.66 ± 0.12 Mass of red brome g 7 0.56 ± 0.10 0.49 ± 0.09 Germination proportion*** Stratified % 8 58.13 ± 7.94 76.38 ± 7.21 Un-stratified % 8 33.75 ± 17.03 67.38 ± 13.44 Time to germination Stratified days 8 16.44 ± 2.66 12.05 ± 1.82 Un-stratified days 8 11.91 ± 7.44 2.84 ± 0.90 Diaspore mass** mg 8 3.75 ± 2.25 2.36 ± 1.68 Time to flowering initiation*** days 8-12 39.44 ± 5.03 51.50 ± 4.81 Time to seedling death days 12 29.44 ± 0.99 31.32 ± 1.03 treated species as a random factor to avoid pseudoreplication. Data were transformed as necessary to meet the assumptions of parametric statistics. For all variables that did not satisfy the assumption of normality, we used a Scheirer-Ray Hare test, which served as a non-parametric version of a two-way nested ANOVA (Sokal and Rohlf 1995; Dytham 201 1) using Minitab (Release 16.0, Minitab Inc., State College, PA) and Excel (Release 2010, Microsoft Corp., Bellingham, WA). We used this test for the variables RGR^ax (total), RGRmax (shoot), diaspore mass, days to flowering initiation, and drought tolerance. For the germination experiment, we analyzed the final germination proportion, and the time to germination. These analyses allowed us to deter- mine whether or not the requirement for cold stratification in seed germination differed between the weedy and non-weedy habits. For the final germination proportion, we used a generalized linear mixed model fit with the Leplace approx- imation using the R package lme4 (R foundation for statistical computing; Bates et al. 2009). The model included stratified and non- stratified (re- ferred to as treatment), weedy and non-weedy (referred to as habit), an interaction term between treatment and habit as fixed factors and a repeated measures term as a random factor. Species were treated as random factor to correct the degrees of freedom, and avoid pseudoreplication when test- ing for differences between habits (Sokal and Rohlf 1995). A significant interaction term signi- fied that the effect of the stratification treatment differed between weedy and non-weedy habits. For time to germination, we compared days to germination initiation using a Scheirer-Ray Hare test (Sokal and Rohlf 1995; Dytham 2011). The model included stratified and non-stratified treat- ments, weedy and non-weedy habits, an interaction term between treatment and habit, and species nested within habit. One species, Amsinckia menziesii (Lehm.) A. Nelson & J.F. Macbr., was removed from the data set because it did not germinate under non-stratified conditions. Mann Whitney tests were used to compare stratified and non-stratified treatments for each species. Results Field-based Trial In late March at our field plot, when most plants were fully grown, weedy species contributed greater vegetative cover than non-weedy species (x\ ^ 16.62, P < 0.001). There were also more individuals of weedy species than non-weedy species (Z — 176.23, P < 0.001). This confirmed our original weedy and non-weedy habit categories. Competitive Characteristics Maximum growth rate and maximum total plant biomass of seedlings were assessed as traits indicative of competitive ability in the critical early growth stage. The weedy plant habit had lower maximum relative growth rates (RGR^ax) than the non-weedy habit (Table 3). This was the case for the entire plant (Xj = 36.30, P < 0.001) and shoots (x] = 4.23, P = 0.039), but not for roots (G y = 2.42, P = 0.164). Weedy and non-weedy habits were also not different in maximum biomass attained (G 7 = 0.51, P = 0.499). 334 MADRONO [Vol. 61 Resource allocation to shoots and leaves is an important determinant of growth rates and competitive ability for light capture, while a larger resource allocation to root biomass may allow plants to become stronger competitors for water and soil nutrients. Weedy and non-weedy habits did not differ in their leaf weight ratio (LWR; Fij = 1.07, P = 0.335), leaf area ratio (LAR; Fi 7 = 0.63, P = 0.452), or specific leaf area (SLA; Fi,7 = 0.01, P = 0.932). Weedy and non-weedy habits were also not different in their relative biomass contribution to roots, as shown by their rootishoot ratio {Fij = 0.01, P = 0.964). The ability to compete for space and resources was quantified in terms of the effect each species had on Bromus madritensis subsp. ruhens biomass when grown in the same container. There was no difference between weedy and non-weedy habits in their competitive effect on the invasive grass {Fij = 0.29, P = 0.603). Most species resulted in B. madritensis biomass that was similar to that attained by B. madritensis plants grown under iiitraspecific competition. Intraspecific competi- tion should be strong due to similar functional traits between conspecific individuals, thus most species were strong competitors. The lack of seed dormancy allows early germination and growth, and may confer a competitive advantage. A cold-stratification treatment increased the proportion of germinated seeds to a greater degree for the weedy habit than the non-weedy habit (Z = —3.559, P < 0.001 for interaction of treatment and habit). This indicates that the weedy habit is more specifically cued to germinate under cool and moist conditions, while the non-weedy habit does not exhibit such seed dormancy. These weedy species germinated during, not after the cold stratification period, implying they are cued to germinate at the beginning of winter, when weather conditions are appropriate for growth. Although weedy species benefited from cold-stratification more than non-weedy species, contrasts showed that both habits benefited from cold stratification. Despite greater overall germination percentage under cool temperatures, this treatment tended to slow the germination of most species. Cold- treated seeds took longer to begin germinating than seeds in the non-stratified treatment (X^ ~ 0.99, P < 0.001). This effect was not significantly different between weedy and non- weedy habits (X| = 0.93, P = 0.074). Ruderal Characteristics The ability of species to disperse, a key ruderal trait, was assessed by examining diaspore mass. The weedy habit exhibited heavier diaspores than the non-weedy habit (x^ = 10.21, P = 0.001). Because of their mass, diaspores of these species are unlikely to disperse long distances; however, they may contain greater energy reserves that increase their chances of establishment within an existing com- munity and their ability to compete successfully. Early flowering, a trait that allows ruderal plants to take advantage of short resource pulses, was analyzed by counting the number of days to flowering initiation. Only six species flowered, which allowed a comparison of flowering period initiation among three weedy and three non- weedy species. Among these six, the weedy habit flowered earlier than the non-weedy habit (X“ = 9.45, P = 0.002). Stress-tolerant Characteristics Drought-tolerance, which may help seedlings persist between unpredictable rainfall events, was assessed as the number of days to drought- induced seedling mortality. There was little variation among species in seedling drought- tolerance. There was no difference between weedy and non-weedy habits (x^ = 3.16, P = 0.075). Principal Components Analysis Data were summarized and relationships among traits were assessed using a principal components analysis (Fig. 1 ). These results can be compared to a summary of the results from individual trait analyses (Table 3). With respect to the variables analyzed, the principal components analysis showed that species clustered into distinct groups (Fig. la). All species except Layia platyglossa (Fisch. & C.A. Mey.) A. Gray, Phacelia tanacetifolia Benth., and Amsinckia menziesii (Lehm.) A. Nelson & J.F. Macbr. loaded strongly on the first component; and all species except B. carinatus loaded strongly on the second component. The invasive grass B. madritensis and the native grass B. carinatus formed a group that was separate from forb species. The grasses were characterized by high root allocation, high total plant biomass, low growth rate, and heavy diaspores. Most non-weedy herba- ceous species loaded toward relatively rapid growth, smaller total plant biomass, diaspore size, and rootishoot ratio on the first component. On the second component, non-weedy species were char- acterized by greater drought-tolerance and low leaf investment per plant biomass (Fig. la). Weedy native forbs occupied a range of plant sizes, rootishoot ratios, diaspore mass, and growth rates (first component), but all of them clustered towards higher leaf investment and relative drought intoler- ance (second component). One non-weedy outlier, P. tanacetifolia, did not load strongly on either component. The principal components analysis also revealed relationships among traits (Fig. lb). On the first component, diaspore mass, total plant biomass, and rootishoot ratio correlated negatively to rapid growth rates. On the second component, a tradeoff 2014] MACKINNON ET AL.: FUNCTIONAL TRAIT DIFFERENCES IN PLANTS 335 First Component (38.1%) Fig. 1. Principal components analysis (PCA) completed from the mean species values for each of the variables analyzed. Species (A) and trait (B) vectors are shown as separate panels. The variables flowering phenology and cold-stratification requirement were excluded from this analysis because the data structure was not appropriate for a PCA. Ellipses show functionally similar species’ cluster with respect to the variables analyzed. Each point in panel A represents a species. Species abbreviations are given in Table 2. was observed with leaf investment traits negatively associated with root allocation. Discussion Percent Cover and Density of Individuals under Field-conditions Our weedy and non-weedy habit designation was confirmed by a field experiment, where we found weedy species were more abundant than non-weedy species. Plant abundance was not confounded by seed quantity because we used the same number of viable seeds for each species in an area where the existing seed bank had been eliminated, and observations were limited to one growing season. Our results suggest that the success of these weedy species is not solely determined by seed abundance, and may be attributed to other factors (Seabloom et al. 336 MADRONO [Vol. 61 2003; Lockwood et al. 2005). We hypothesized that functional trait differences were an impor- tant cause of these differences in abundance. Other explanations exist as well, such as differ- ential herbivory, where the seeds of non-weedy plants were preferentially removed by seed predators, but this was not apparent based on our field observations. Weedy and Non-weedy Species We found that plants fell into three groups: 1. grasses, 2. weedy forbs, and 3. non-weedy forbs, with a distinction between weedy and non-weedy habits evident among the forbs (Fig. 1). The weedy forb group differed from non-weedy forbs mainly by their comparatively high leaf area investment. The grass group was characterized by relatively large total plant biomass, large root: shoot ratio, and large diaspores. Individual trait analyses demonstrated that weedy species, when compared to non-weedy species, had slower growth rate, heavier diaspores, earlier flowering period initiation, and a greater cold-stratification requirement for germination (Table 3). The traits that were shared by weedy species are likely linked to the environmental conditions associated with the southern San Joaquin Valley. The slower growth rates found in weedy forbs suggests they are at a competitive disadvantage in teiTns of a lack of a capacity for rapid use of space and resources (Poorter et al. 1990; Grime and Hunt 1975; Eliason and Allen 1997; Dawson et al. 2011). However in San Joaquin Valley habitats competitive ability in terms of rapid growth may not be as important as other factors, and could even be detrimental, causing plants to deplete limited resources faster than they are replenished (Grime and Hunt 1975). The lack of a difference between weedy and non-weedy habits in seedling drought-tolerance suggests that de- spite arid conditions of the southern San Joaquin Valley, seedling drought-tolerance is unlikely to be an important feature differentiating weedy from non-weedy annual plants, which may instead share the tendency to avoid drought by growing only when resources are plentiful. The ability of weedy species to initiate flower- ing earlier than non-weedy species may enable these plants to reproduce quickly, even during a short rainy season (Cramer et al. 2008; Eriksson 2000; Lake and Leishman 2004; Grotkopp and Rejmanek 2007). Based on the heavy diaspores of weedy species, long-distance dispersal ability does not appear to be a key factor in their ability to colonize disturbed areas (Eriksson 2000; Venable and Brown 1988; Tiffney 1984; Drenovsky 2012). However, these heavy diaspores could still be well dispersed if they have specialized structures that allow them to be dispersed by wide ranging animals, and this appears to be the case for some species included in this study. Although seeds of weedy species had greater germination proportion in the cold-stratified treatment, they tended to germinate at cool temperatures, rather than after the cold period had ended. Thus, instead of being cued to germinate with warming spring conditions, which is usually the case in cooler climates, these plants are instead cued to germinate at the start of winter (Keeley and Davis 2007). This is consistent with the occurrence of these plants in a Mediter- ranean-type climate, where cool temperatures usually coincide with adequate moisture for growth. This winter germination ability, com- bined with slow growth rates, could relate to a weedy strategy that enables weedy species to slowly occupy space and resources during cool winter months. This way, weedy species may be well established by the time spring arrives and other species begin to germinate. It would be valuable to compare field-based germination and coverage to test this hypothesis. Conclusions Based on an analysis of nine co-occurring species of the southern San Joaquin Valley of California, there appears to be a weedy suite of traits that describes the success of some annual plant species. Weedy species were characterized by several traits: slower growth, heavier dia- spores, earlier flowering period initiation, and a germination mechanism cued to the cool and moist winters associated with Mediterranean- type climates. The distinctions we found between weedy and non-weedy plants may be used to locate native species from databases that could be useful for restoration (e.g., the TRY database; Kattge et al. 2011). Using native plants that resemble invasive plants functionally may be useful for exposing invasive species to competi- tion, whereas species that differ in functional traits may be selected because they are func- tionally complementary to a particular invader (Funk et al. 2008). Functionally weedy native plants in particular could lead to increased na- tive plant cover and diversity in ecosystems that, due to continuing disturbance, no longer sup- port historic vegetation. It is possible that this form of restoration would eventually improve ecological conditions for more ambitious resto- ration objectives. Acknowledgements Funding for this project was provided by the Student Research Scholars Program and from the Graduate Student-Faculty Collaborative Initiative. Special thanks to L. Maynard Moe for his input and guidance during this project. Evan MacKinnon was supported by an NSF Research Assistantship position during his MS 2014] MACKINNON ET AL.: FUNCTIONAL TRAIT DIFFERENCES IN PLANTS 337 work at California State University, Bakersfield (NSF IOS-0845125). Literature Cited Allen, E. B. 2004. Restoration of Artemisia shrub- lands invaded by exotic annual Bromus: a compar- ison between southern California and the intermountain west. United States Department of agriculture, USD A Forest Service Proceedings, Washington D.C. Baldwin, B. G., D. H. Goldman, D. J. Keil, R. Patterson, T. J. Rosatti, and D. H. Wilken. (eds.). 2012. The Jepson manual: vascular plants of California. 2nd ed. University of California Press, Berkeley, CA. Bates, D., M. Maechler, and B. Dai. 2009. Ime4: linear mixed effects models. R package. Website http://cran.r-project.org/web/packages/lme4i (ac- cessed 01 May 2013). Bazzaz, F. a., N. R. Chiariello, P. D. Coley, and L. F. PiTELKA. 1987. Allocating resources to reproduction and defense: new assessments of the costs and benefits of allocation patterns in plants are relating ecological roles to resource use. Biosciences 37:58-66. Bekker, R. M., j. P. Barker, U. Grandin, R. Kalamees, P. Milberg, P. Poschlod, K. Thompson, and J. H. Willems. 1998. Seed size, shape and vertical distribution in the soil: indicators of seed longevity. Functional Ecology 12:834-842. Casper, B. B. and R. B. Jackson. 1997. Plant competition underground. Annual Review Ecolo- gy, Evolution, and Systematics 28:545-570. Chuanren, D., W. Bochu, L. Wanqian, C. Jing, L. JiE, AND Z. Huan. 2004. Effect of chemical and physical factors to improve the germination rate of Echinacea angustifoUa seeds. Biointerfaces 37:101-105. Cleland, E. E. 2011. Trait divergence and the ecosystem impacts of invading species. New Phytologist 189:649-652. , L. Larios, and K. N. Suding. 2013 [2012]. Strengthening invasion filters to re-assemble native plant communities: soil resources and phenoiogical overlap. Restoration Ecology 21:390-398. Cramer, V. A., R. J. Hobbs, and R. J. Standish. 2008. What’s new about old fields? Land aban- donment and ecosystem assembly. Trends in Ecology and Evolution 23:104—112. Daubenmire, R. 1959. Canopy-coverage method of vegetational analysis. Northwest Science 33:43-64. Dawson, W., M. Fischer, and M. van Kleunen. 2011. The maximum relative growth rate of common UK plant species is positively associated with their global invasiveness. Global Ecology and Biogeography 20:299-306. Diaz, S., J. G. Hodgson, K. Thompson, M. Cabido, J. H. C. CORNELISSEN, A. Jalili, G. Montserrat- MartI, j. P. Grime, F. Zarrinkamar, Y. Asri, S. R. Band, S. Basconcelo, P. Castro-Diez, G. Funes, B. Hamzehee, M. Khoshnevi, N. Perez- Harguindeguy, M. CPerez-Rontome, F. a. Shirvany, F. Vendramini, S. Yazdani, R. Abbas-Azimi, a. Bogaard, S. Boustani, M. Charles, M. Dehghan, L de torres-Espuny, V. Falczuk, j. Guerrero-Campo, A. Hynd, G. Jones, E. Kowsary, F. Kazemi-Saeed, M. Maestro-Martinez, a. Romo-Diez, S. Shaw, B. SiAVASH, P. Villar-Salvador, and M. R. Zak. 2004. The plant traits that drive ecosystems: evidence from three continents. Journal of Vegeta- tion Science 15:295-304. , M. Cabido, and F. Casanoves. 1999. Functional implications of trait-environment link- ages in plant communities. Pp. 338-360 in E. WeiherP. A. Keddy. (eds.), Ecological assembly rules: perspectives, advances, retreats. Cambridge University Press, Cambridge, UK. Ditomaso, j. M. and E. a. Healy. 2007. Weeds of California and other western states. University of California Division of Agriculture and Natural resources, Oakland, CA. Drenovsky, R. E., B. j. Grewell, C. M. D’Anto- Nio, J. L. Funk, J. J. James, N. Molinari, I. M. Parker, and C. L. Richards. 2012. A functional trait perspective on plant invasion. Annals of Botany 110:141-153. Dytham, C. 2011. Choosing and using statistics. 3rd ed. Wiley-Blackwell, West Sussex, UK. Eissenstat, D. 1991. On the relationship between specific root length and the rate of root prolifer- ation: a field study using citrus rootstocks. New Phytologist 118:63-68. Eliason, S. a. and E. B. Allen. 1997. Exotic grass competition in suppressing native shrubland rees- tablishment. Restoration Ecology 5:245-255. Eriksson, O. 2000. Seed dispersal and colonization ability of plants: assessment and implications for conservation. Folia Geobotanica 35:115-123. Funk, J. L., E. E. Cleland, K. N. Suding, and E. S. Zavaleta. 2008. Restoration through reassembly: plant traits and invasion resistance. Trends in Ecology and Evolution 23:695-703. AND P. M. ViTOUSEK. 2007. Resource-use efficiency and plant invasion in low-resource systems. Nature 446:1079-1081. AND V. A. Zachary. 2010. Physiological responses to short-term water and light stress innative and invasive plant species in southern California. Biological Invasions 12:1685-1694. Germano, D. j., G. G. Rathbun, L. R. Saslaw, B. L. Cypher, E. A. Cypher, and L. M. Vreden- BURGH. 2011. The San Joaquin Desert of Califor- nia: ecologically misunderstood and overlooked. Natural Areas Journal 31:138-147. Gitay, H., I. R. Nobel, and J. H. Connel. 1999. Deriving functional types for rain-forest trees. Journal of Vegetation Science 10:641-650. Gracet, j. B. 1995. On the measurement of plant competition intensity. Ecology 76:305-308. Grime, J. P. 1977. Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. The American Naturalist 111:1169-1194. . 1984. The ecology of species, families and communities of the contemporary British flora. New Phytologist 98:15-33. . 2001. Plant strategies, vegetation processes, and ecosystem properties. 2nd ed. John Wiley and Sons Ltd, West Sussex, England. AND R. Hunt. 1975. Relative growth-rate: its range and adaptive significance in a local flora. Journal of Ecology 63:393^22. 338 MADRONO [Vol. 61 Grotkopp, E. and M. RejmAnek. 2007. High seedling relative growth rate and specific leaf area are traits of invasive species; phylogenetically independent contrasts of woody angiosperms. American Journal of Botany 94:526-532. Hunt, R. 1978. Plant Growth and analysis. Studies in Biology no. 96. Edward Arnold Limited, London, UK. Kattge, J., S. Diaz, S. Lavorel, I. C. Prentice, P. Leadley, G. Bonisch, E. Garnier, M. Wes- TOBY, P. B. Reich, I. J. Wright, J. H. C. CORNELISSEN, C. ViOLLE, S. P. HARRISON, P. M. VAN Bodegom, M. Reichstein, B. j. Enquist, N. A. SOUDZILOVSKAIA, D. D. Ackerly, M. Anand, O. Atkin, M. Bahn, T. R. Baker, D. Baldocchi, R. Bekker, C. C. Blanco, B. Blonder, W. Bond, R. Bradstock, D. E. Bunker, F. Casanoves, J. Cavender-Bares, J. Q. Chambers, F. S. Chapin III, J. Chave, D. CooMES, W. K. Cornwell, J. M. Craine, B. H. Dobrin, L. Duarte, W. Durka, J. Elser, G. Esser, M. Estiarte, W. F. Fagan, J. Fang, F. FernAndez-Mendez, A. Fidelis, B. Finegan, O. Flores, H. Ford, D. Frank, G. T. Freschet, N. M. Fyllas, R. Gallagher, W. A. Green, A. G. Gutierrez, T. Hickler, S. Higgins, J. G. Hodgson, A. Jalili, S. Jansen, C. Joly, A. J. Kerkhoff, D. Kirkup, K. Kitajima, M. Kleyer, S. Klotz, j. M. H. Knops, K. Kramer, I. Kuhn, H. Kurokawa, D. Laughlin, T. D. Lee, M. Leishman, F. Lens, T. Lenz, S. L. Lewis, J. Lloyd, J. LlusiA, F. Louault, S. Ma, M. D. Mahecha, P. Manning, T. Massad, B. Medlyn, j. Messier, A. Moles, S. C. Muller, K. Nadrowski, S. Naeem, U. Niinemets, S. Nollert, a. Nuske, R. Ogaya, J. Oleksyn, V. G. Onipchenko, Y. Onoda, J. Ordonez, G. Overbeck, W. Ozinga, S. Patino, S. Paula, J. G. Pausas, j. Penuelas, O. L. Phillips, V. Pillar, H. Poorter, L. Poorter, P. Poschlod, A. Prinzing, R. Proulx, A. Rammig, S. Re- iNSCH, B. Reu, L. Sack, B. Salgado-Negret, J. Sardans, S. Shiodera, B. Shipley, A. Siefert, E. SOSINSKI, J. SOUSSANA, E. SWAINE, N. SWEN- soN, K. Thompson, P. Thornton, M. Wal- dram, E. Weiher, M. White, S. J. Wright, B. Yguel, S. Zaehle, a. E. Zanne, and C. Wirth. 2011. TRY - a global database of plant traits. Global Change Biology 17:2905-2935. Keeley, j. E. and F. W. Davis. 2007. Chaparral. Pp. 339-366 in M. G. BarbourT. Keeler-Wolf, and A. A. Schoenherr (eds.). Terrestrial vegetation of California, 3rd ed. University of California Press, Los Angeles, CA. Kimball, S., J. R. Gremer, T. E. Huxman, D. L. Venable, and A. L. Angert. 2013. Phenotypic selection favors missing trait combinations in coexisting annual plants. The American Naturalist 182:191-207. Lake, J. C. and M. R. Leishman. 2004. Invasion success of exotic plants in natural ecosystems: the role of disturbance, plant attributes and freedom from herbivores. Biological Conservation 117: 215-226. Lockwood, J. L., P. Cassey, and T. Blackburn. 2005. The role of propagule pressure in explaining species invasions. Trends in Ecology and Evolution 20:223-228. MACk, R. N. and D. a. Pyke. 1983. The demography of Bromus tectorunr. variation in time and space. Journal of Ecology 71:69-93. McGarigal, K., S. Cushman, and S. Staeford. 2000. Multivariate statistics for wildlife and ecol- ogy research. Springer, New York, NY. Philippi, T. 1993. Bet-hedging germination of desert annuals: beyond the first year. The American Naturalist 142:474-487. Poorter, H. and C. Remkes. 1990. Leaf area ratio and net assimilation rate of 24 wild species differing in relative growth rate. Oecologia 83:553-559. ScHUTZ, W. AND G. Rave. 1999. The effect of cold stratification and light on the seed germination of temperate sedges {Carex) from various habitats and implications for regenerative strategies. Plant Ecology 144:215-230. Seabloom, E. W., W. S. Harpole, O. J. Reichman, AND D. Tilman. 2003. Invasion, competitive dominance, and resource use by exotic and native California grassland species. Proceedings of the National Academy of Sciences of the United States of America 100:13384-13389. Shipley, B. and M. Parent. 1991. Germination responses of 64 wetland species in relation to seed size, minimum time to reproduction and seedling relative growth rate. Functional Ecology 5:111- 118. Skordilis, a. and C. A. Thanos. 1995. Seed stratification and germination strategy of the Mediterranean pines Pinus brutia and P. halepensis. Seeds Science Research 5:151-160. SOKAL, R. R. AND F. J. Rohlf. 1995. Biometry. 3rd ed. Freeman and Company, Basingstoke, England. Stenstrom, M. and U. Molau. 1992. Reproductive ecology of Saxifragci oppositifolia: phenology, mating system, and reproductive success. Arctic and Alpine Research 24:337-343. Swanborough, P. and M. Westoby. 1996. Seedling relative growth rate and its components in relation to seed size: phylogenetically independent con- trasts. Functional Ecology 10:178-184. Tiffney, B. H. 1984. Seed size, dispersal syndromes, and the rise of the angiosperms: evidence and hypothesis. Annals of the Missouri Botanical Garden 71:551-576. Tyree, M. T., B. M. J. Engelbrecht, G. Vargas, and T. a. Kursar. 2003. Desiccation tolerance of five tropical seedlings in Panama. Relationship to a field assessment of drought performance. Plant Physiology 132:1439-1447. Venable, D. L. 2007. Bet hedging in a guild of desert annuals. Ecology 88:1086-1090. and j. S. Brown. 1988. The selective interac- tions of dispersal, dormancy, and seed size as adaptations for reducing risk in variable environ- ments. The American Naturalist 131:360-384. Whitson, T. D., L. C. Burrill, S. A. Dewey, D. W. Cudney, B. E. Nelson, R. D. Lee, and R. Parker, (eds.). 2001. Weeds of the West. 9th ed. University of Wyoming, Jackson, WY. Wilson, J. B. 1988. The effect of initial advantage on the course of plant competition. Oikos 51:19-24. Madrono, Vol. 61, No. 4, pp. 339-349, 2014 VARIATION IN SEED CHARACTERISTICS AND GROWTH FOR THISTLES (CARDUEAE: ASTERACEAE) IN CALIFORNIA AND OREGON D. F. Spencer' and P-S. Liow USDA-ARS Exotic & Invasive Weeds Research Unit, Davis, CA 95616 * david . spencer@ars . usda. go v M. J. Pitcairn and B. Villegas California Department of Food and Agriculture, Sacramento, CA 95832 Abstract Genera in the tribe Cardueae (Asteraceae) include both non-invasive and invasive species widely dispersed in the western U.S. Seed characteristics are important for seed dispersal, seedling growth, and seedling survival. We determined seed characteristics and their variation from natural populations of 22 taxa of Cardueae. We also measured C and N content of seeds for 1 7 taxa, and conducted a greenhouse growth experiment with five species from this group. We tested the hypothesis that it is possible to distinguish invasive species from non-invasive species based on these characteristics. Seed weight differed significantly (P < 0.0001) among taxa and varied by a factor of 24, from a mean of 1.48 mg for Centaiirea solstitialis L. to 35.63 mg for Cynara carchmculus L. subsp. //r/ve'.s'ct'm’ Wiklund. There were no significant relationships between status as an invasive species and explanatory variables based on logistic regression results (P > 0.05). Seed N, C, and C;N ratio differed significantly (P < 0.0001) among taxa, but these characteristics were not associated with invasiveness. Relative growth rate (RGR) for greenhouse grown plants ranged from 0.010 to 0.030 g g“' day"'. Linear regression results indicate that there was no significant relationship between seed weight and RGR or other measures of plant growth and condition and seed weight. For many of the taxa analyzed in this study, the infomiation on seed weight, nutrient content, and growth measures (RGR, net photosynthesis rate) have not been previously reported. Key Words: Cardims, centaurea, Cirsium, Cynara, Silyhiim. The tribe Cardueae Cass, (as described in Flora of North America) is one of the largest tribes in Asteraceae, consisting of over 2500 species worldwide (Barkley et al. 2006). In North America, Cardueae is represented by only two native genera: Cirsium Mill. (ca. 80 species) and Saussurea DC. (three species) (Keil 2006). Cali- fornia has a particularly high diversity of Cirsium species. The second edition of The Jepson Manual recognizes 39 native taxa (species and infraspecific varieties) (Baldwin et al. 2012). Despite this high amount of native diversity, California has accumulated a number of exotic Cardueae taxa. The second edition of the Jepson Manual (Baldwin et al. 2012) lists the following thistle genera in tribe Cardueae as fully natural- ized in California: Acroptilon Cass., Arctium L., Carduus L., Carthamus L., Centaurea L., Cynara L., Silybum VailL, Volutaria Cass., and several exotic species of Cirsium. A plant's life history and the way it allocates limited resources to growth, defense, and repro- duction can affect population dynamics and invasiveness. Differences in seed characteristics (e.g., size, weight, nutrient content) strongly affect demographic rates (i.e., propagule produc- tion, germination rates, seedling survivorship) and may help identify factors associated with invasiveness in native and exotic taxa. The size and nutrient content of seeds is important for seed dispersal, seedling growth, and seedling survival. When a seed germinates, growth is initially limited by the quantity of nutrients contained within it (Hendry and Grime 1993). Plants from smaller seeds appear to have higher relative growth rates than those from larger seeds (Marah6n and Grubb 1993; Gross 1994). One theory has proposed that small seed size is a characteristic that may identify some invasive species (Rejmanek 1995; Grotkopp 2002); how- ever, this does not always appear to be the case (Daws et al. 2007; Forcella 1985). In some cases, smaller seeds have greater dispersal abilities, but this depends on the dispersal mechanism (wind, water, vertebrate, ant, ballistic) and on plant height (Thomson et al. 2011). In the presence of granivores, small seed size may be advantageous as small seeds may not be as easily detected or handled as larger seeds (Henderson 1990). In contrast, seedlings from larger seeds may have a greater survival probability (Stanton 1984; Dal- ling and Hubbell 2002). Despite the important role, seed size has in the survival and growth of plants there is little information on seed characteristics for several California and Oregon taxa of the Cardueae. Of the 22 taxa examined in this report we were only able to find information on seed size and its 340 MADRONO [VoL 61 variation for four taxa. These include the invasive plant, Centaurea solstitialis L., the potential seed oil bioenergy crop, Cynara cardunculus L., the weedy Silybum marianum (L.) Gaertn., and Cirsium vulgar e (Savi) Ten. Thus, this manuscript presents data on seed weight and its variation from natural populations of 22 taxa of Cardueae. We also present data on the nutrient content (C, N) of seeds for 17 taxa from this group and the results of a greenhouse experiment with five species from this group. We use this information to test the hypothesis that within tribe Cardueae it is possible to distinguish invasive species from non-invasive species based on seed traits (Ordo- nez et al. 2010). As noted by Ordonez et al. (2010) the approach of comparing traits of invasive and native species within a group of related species to identify traits associated with successful plant invasions has been informative. Materials and Methods Seeds were collected by hand from 22 Car- dueae taxa at various sites in California and three sites in Oregon between 1998 and 2002, except that one collection was made in 1988 (Table 1). Nomenclature of these taxa follows Baldwin et al. (2012) unless otherwise noted. Flower heads were placed in paper bags and returned to the California Department of Food and Agriculture laboratory in Sacramento where seeds were removed. Seeds were transferred to small enve- lopes and stored under dry conditions at room temperature. Each seed was weighed individually to 0.1 mg (N == 73-943 seeds per taxon). For 17 taxa either five or 10 seeds from each taxon (216 seeds for Centaurea solstitialis) were analyzed for C and N content (percent of dry weight) using a Perkin-Elmer Model 2400 CHN analyzer with acetanilide used as the standard. We calculated the C:N ratio in order to assess the suitability of seeds as food items by comparing them with a value of 17. For example, C:N values greater than 17 may indicate less desirable food items for insects (McMahon et al. 1974; Karban and Baldwin 1997). Elementary statistics (mean, coefficient of variation, standard error, median, skewness, and kurtosis) for seed weight for each taxon (across sample dates and sample sites) were calculated using the univariate procedure in SAS (2011). The skewness and kurtosis indices were included because they indicate the degree to which a frequency distribution departs from a normal distribution (Sokal and Rohlf 1995). Previous workers use the skewness statistic to summarize size distributions of plants within a population and as such, it could be applied to a population of seeds as well (Rejmanek et al. 1989; Petersen et al. 1990). Herrera (2009) has suggest- ed that skewness and kurtosis may be important in describing plant phenotypes, especially for species producing large numbers of seeds. Statis- tical tests of the null hypothesis that the skewness index and the kurtosis index were not different from zero were performed with a two-tailed t-test as described by Sokal and Rohlf (1995). We calculated an unbalanced analysis of variance using the GEM procedure in SAS to test the hypothesis that seed weight, C, N, and C:N ratio did not differ across taxa. Effects were considered significant at the 0.05 level. We tested the hypothesis that seed character- istics could identify invasive species using logistic regression. We calculated logistic regression with the binary species category, invasive/non-invasive status as listed by the California Invasive Plant Council (Cal-IPC 2006) as the dependent variable and mean seed weight, range of seed weights, skewness of seed weights, kurtosis of seed weights, or the coefficient of variation for seed weight as explanatory variables using the logistic procedure in SAS (2011). Greenhouse experiment. This experiment was performed to determine the effect of seed size on growth rate of five species which varied in mean seed size (Table 2): Centaurea solstitialis, Centau- rea cyanus L., Cirsium douglasii DC., Cirsium under sonii (A. Gray) Petr., and Cynara carduncu- lus L. suhsp. flavescens Wiklund. The experiment was conducted from December 16, 2003 to May 26, 2004 in a greenhouse located at Davis, California. Typical conditions during the exper- iment were mean daily temperature 24°C and mean irradiance from a bank of fluorescent lamps 224 pmol m“^ s“' with a 14:10 light dark cycle. The natural irradiance averaged 167 W m“^ s~’, and increased from <20 W m“^ s“^ to >300 W m“^ s“‘ during the experiment. Twenty-five seeds of each taxon were placed in a small mesh bag and soaked in deionized-water for 24 hours. At planting, one seed of each species was placed in individual 1-L plastic pots (12 cm diameter by 12.5 cm tall) filled to a depth of 1 1 cm with a prepared clay substrate (one week prior to the day of planting, the clay was soaked in deionized water for five days). Pots were ran- domly assigned to positions on the greenhouse bench so that they were about 13 cm apart. There were a total of 125 pots (5 species X 25 pots). Hoagland’s nutrient solution (Hoagland and Arnon 1950), with 75 mg N L“‘, was added once a week. Fifty ml Hoagland’s solution was added in week one, 100 ml per week subsequently. Starting two weeks after planting, digital photos were taken for each pot and leaf areas determined from the pictures using an image analysis program (Sigma Scan Pro, Systat Software, Inc., San Jose, CA). Every four weeks, five pots of each species (except Cirsium andersonii) were randomly select- 2014] SPENCER ET AL.: FUNCTIONAL TRAIT VARIATION IN THISTLES 341 Table L Taxa, Date Collected, and Location for the Seeds Used in this Study. All counties are in California except those followed by “(OR)” which are in Oregon. Species collections on the same date are from different sites within the county. The column labeled “ Inv.” presents the California Invasive Plant Council' rating for impact as an invasive plant: H = high, M = moderate, and L = limited. Dashed lines ( — ) indicate that the species is not currently considered to be invasive. ' California Invasive Plant Council, Invasive Plant Inventory (CAL-IPC). Website http://www.cal-ipc.org/ip/inventory/index.php (accessed 4 June 2013) ^Hickman, J. C. (ed.). 1993. The Jepson Manual: Higher plants of California. University of California Press, Berkeley, CA. Taxon Inv. Date collected County Carduus pycnocephalus Spreng. M 08/12/02 Solano Centaurea cyanus L. — 05/27/99 San Luis Obispo 06/08/99 Shasta Centaurea solstitialis L. H 08/30/00 Siskiyou Centaurea jacea L. nothosubsp. pratensis (W.D.J. Koch) Celak. M 08/16/00 Del Norte Cirsium andersonii Jeps. — 07/20/98 Nevada 08/20/98 Nevada 08/23/00 El Dorado Cirsium brevistylum Cronquist — 08/04/98 Humboldt 08/16/00 Del Norte 08/16/00 Humboldt 07/19/00 Linn (OR) Cirsium occidentale (Nutt.) Jeps. var. californicum — 05/05/99 Kern (A. Gray) D. J. Keil & C.E. Turner 05/05/99 Kern 05/26/99 Santa Barbara 07/13/00 Los Angeles Cirsium occidentale var. candidissimum (Greene) — 07/22/99 Mono J.F. Macbr. 09/03/98 Plumas 08/18/98 Shasta 08/19/98 Modoc 08/24/00 Mono 08/29/00 Trinity 08/23/00 Alpine Cirsium cymosum (Greene) J.T. Howell — 07/01/99 Nevada var. canovirens /Rydb.) D.J. Keil 08/20/98 Nevada 07/19/00 Lake 08/23/00 Alpine Cirsium crassicaule Jeps. — 06/15/99 Kern Cirsium cymosum (Greene) J. T. Howell 06/08/99 Siskiyou var. cymosum (JFP-1)^ 06/08/99 Siskiyou 06/08/99 Siskiyou 06/29/00 Lassen Cirsium douglasii DC. (JFP-1)^ — 08/20/98 Nevada 08/05/98 Humboldt 08/20/00 Trinity 08/17/00 Humboldt 08/17/00 Humboldt 08/17/00 Humboldt 08/29/00 Trinity 08/30/00 Trinity Cirsium scariosum Nutt. var. loncholepis (Petr.) D.J. Keil — 05/27/99 San Luis Obispo Cirsium occidentale (Nutt.) Jeps. var. occidentale — 06/15/99 San Luis Obispo Cirsium ochrocentrum A. Gray — 08/19/88 Modoc 07/19/00 Lake Cirsium remotifolium DC. — 07/19/00 Linn (OR) Cirsium scariosum Nutt. var. scariosum (JFP-1)- 07/01/99 Plumas 02/09/98 Plumas Cirsium undulatum Spreng, — 07/12/98 (OR) 07/18/00 Wasco (OR) Cirsium occidentale (Nutt.) Jeps. var. venustum (Greene) Jeps. — 06/15/99 Monterey 06/15/99 San Benito 06/15/99 Monterey 06/15/99 Monterey 06/15/99 Kern 08/15/00 Mendocino Cirsium vulgar e (Savi) Ten. M 08/12/02 Yolo Cynara cardunculus L. subsp. Jlavescens Wiklund L 07/10/00 Riverside Silybum marianum (L.) Gaertn. M 07/01/02 Yolo 342 MADRONO [Vol. 61 Table 2. Basic Statistics for 22 Cynareae (Asteraceae) Taxa. Statistics are number of samples (N), mean (Mean), coefficient of variation (CV), minimum (Min.), maximum (Max.), kurtosis (Kurt.), and skewness (Skew). The column labeled “ Inv.” presents the California Invasive Plant Councif rating for impact as an invasive plant: H = high, M = moderate, and L = limited. Dashed lines ( — ) indicate that the species is not currently considered to be invasive. Species are arranged from the smallest mean seed mass to the largest mean seed mass (mg). Differences in seed size among species were significant (P < 0.0001, unbalance analysis of variance calculated by PROC GLM). An asterisk (*) following kurtosis or skewness indicates that the value was significantly different from zero, P < 0.05. The complete scientific names are listed in Table 1. ‘California Invasive Plant Council, Invasive Plant Inventory (CAL-IPC). Website http://www.cal-ipc.org/ip/inventory/index.php (accessed 4 June 2013). Taxon N Mean (mg) CV % Min. Max. Kurt. Skew. Inv. Cen taurea solstitial is 341 1.48 26.6 0.25 2.46 0.05 -0.27* H Cirsium hrevistyliim 190 2.07 24.8 0.74 3.51 0.11 -0.10 — Centaurea jacea nothosubsp. pratensis 100 3.09 17.4 1.05 3.96 1.68* -0.92* M Cirsium vulgare 200 3.13 15.4 1.25 4.14 1.45* -0.82* M Cirsium scariosum var. loncholepis 100 3.98 21.5 1.97 5.65 -0.67 -0.43 — Centaurea cyanus 130 4.13 21.7 1.60 6.40 0.32 0.19 — Cirsium remotifolium 100 4.79 17.6 1.78 6.39 0.66 -0.47 — Cirsium scariosum var. scariosum 136 4.88 30.1 1.43 8.36 -0.34 0.30 — Cirsium crassicaide 100 5.41 19.3 2.62 6.96 -0.21 -0.77* — Cirsium cymosum var. ccmovirens 175 6.32 32.3 1.98 14.73 2.33* 1.17* — Carduus pycnocephalus 100 6.89 12.3 4.43 8.81 0.04 -0.42 M Cirsium occidentale var. occidentale 73 7.10 20.4 3.57 9.46 -0.38 -0.41 — Cirsium douglasii 310 7.22 33.8 1.62 14.78 0.20 0.01 — Cirsium occidentale var. californieum 191 10.82 27.0 4.66 20.42 1.02* 0.88* — Cirsium occidentcde var. venustum 251 12.99 22.8 4.89 20.89 -0.21 -0.14 — C irsiiim cmdersonii 156 13.03 32.0 4.23 22.85 -0.90* 0.28 — C irsiui 1 1 o chro cen trim i 130 13.11 19.7 6.30 19.96 0.38 -0.08 — Cirsium cymosum var. cymosum 220 14.11 28.8 5.62 24.09 -0.56 0.15 — Cirsium occidentale var. candidissimum 411 15.58 26.1 1.59 25.08 -0.73* -0.08 — Cirsiun i imdulatum 130 15.95 19.0 7.80 21.31 -0.16 -0.66* — Silybum marianum 100 23.09 14.5 10.89 29.63 1.12* -0.58 L Cynara eardunculus subsp. flavescens 130 35.63 27.7 11.06 48.29 -0.90* -0.62 M ed and harvested. Because Cirsium cmdersonii grew noticeably slower than the other species, two pots of Cirsium cmdersonii were harvested instead of five pots. The aboveground and belowground parts were separated and their fresh and dry weights detemiined. We calculated the relative growth rate (RGR) for each species using the logarithm of plant total dry weight as the dependent variable in linear regression against days since planting (Hunt 1982). For this, we used only data from the first four sampling dates since plant total dry weight declined after that. Net photosynthesis measurements. We also measured relative leaf chlorophyll using a SPAD meter (Minolta 502, Spectrum Technologies, East Plainfield, IL). A SPAD meter measures the transmittance of red (650 nm) and infrared (940 nm) radiation through a leaf and calculates a relative meter reading, which reflects the chlorophyll content of the leaf (Uddling et al. 2007). For four species {Centaurea cyanus, Cirsium douglasii, Centaurea soIstitiaJis, and Cynara eardunculus subsp. flavescens), on two dates (19 March 2004 and 5 May 2004) we measured net photosynthesis (estimated from carbon exchange rates. An) using a LI-COR 6400. All measurements were made between 10:00 and 14:00 hours. The first fully expanded leaf on a stem was measured to control for effects of leaf age. The measurement chamber was placed in the middle of the leaf on the leaf s upper surface. Conditions in the chamber were allowed to equilibrate until the stability param- eter, total CV%, was near one, after which the appropriate measurements were recorded. Mea- surements were made at 1200 pmol m“- s“' (PAR), mean air temperature in the chamber when the measurements were made was 26.2 1°C on 19 March (standard deviation = 0.3, N = 14) and 28.0°C on 5 May (standard deviation = 0.7, N = 12). All of the SPAD and An measurements for a species were used to calculate the mean SPAD and An for each species. Mean values were used as dependent variables in linear regression versus mean seed weight for each species, to test the hypothesis that mean seed weight would predict values of these functional traits. Results For the entire set of 22 taxa, seed weight varied by a factor of 24, from a mean of 1.48 mg for Centaurea solstitkdis to 35.63 mg for Cynara eardunculus subsp. flavescens (Table 2). Results from analysis of variance indicate that the differences among taxa were significant (P < 2014] SPENCER ET AL.: FUNCTIONAL TRAIT VARIATION IN THISTLES 343 0.0001). Based on the coefficient of variation Cirsium douglasii seed weight exhibited the greatest variation with Cirsium occidentale (Nutt.) Jeps. var. compactum Hoover second. The value for kurtosis index was significantly different from zero, for eight taxa, and the value for the skewness index was significantly different from zero, for seven taxa (Table 2). Mean seed weight for the six taxa considered as invasive {Centaurea solstitialis, Centaurea jacea L. notho- subsp. pratensis (W.D.J. Koch) Celak., Cirsium vulgare, Carduus pycnocephalus L., Silybum marianum, and Cynara carduncuius subsp. flaves- certs') were among the smallest, mid-range, or the largest values for the 22 taxa examined in this study (Table 2). Accordingly, results of logistic regression of invasive status versus the explana- tory variables (mean seed weight, range of seed weights, or the coefficient of variation for seed weight) revealed no significant relationships (P > 0.05) among the explanatory variables and status as an invasive plant species. Differences among taxa for seed N, seed C, and seed C:N ratio were significant (P < 0.0001). The lowest value for seed N content was for Cirsium undulatum (Nutt.) Spreng. (1.77%) and the highest for Cirsium under sonii (3.32%) (Fig. 1). Seed N for two invasive species, Carduus pycnocephalus and Centaurea solstitialis, were greater than 2.5% and two others, Cynara carduncuius subsp. flavescens and Silybum mar- ianum, had values near 2.5%. Tissue N for invasive species ranked intermediate among the range of taxa values. With respect to tissue C, two of the invasive species had the greatest values (Fig. 1). Tissue C for a third invasive species, Silybum marianum was sixth highest, and Cynara carduncuius subsp. flavescens, ranked fifth from the lowest. The C:N ratio ranged from 16.3-29.3. For Centaurea solstitialis and Cynara carduncuius suhsyt. flavescens C:N ratios were near 17, while C:N ratios for Carduus pycnocephalus and Silybum marianum were greater than that value. Based on the 95% confidence limits, the C:N ratio was greater than 17 for nine of the 17 taxa measured. Growth experiment. Plant height and total leaf area increased at different rates following germi- nation (Fig. 2). At the end of nine weeks, Centaurea cyanus plants were tallest and Cirsium douglasii, Cirsium andersonii, and Centaurea solstitialis were the shortest. Between 21 and 42 d after planting Cynara carduncuius subsp. flavescens had the greatest leaf area. After 49 d Centaurea cyanus and Cirsium douglasii had similar leaf areas to Cynara carduncuius subsp. flavescens. The smallest leaf area values were for Cirsium andersonii. Centaurea solstitialis leaf area was intermediate between these groups. The total dry weight of all species increased initially to maximum values around 10-14 wk after planting and then declined reflecting senescence (Fig. 3). Cirsium douglasii plants were the largest at the 14 week harvest. This species allocated more biomass to roots than the others (Fig. 3). Two species, Centaurea cyanus and Centaurea solsti- tialis, produced flowers during this experiment while the others did not. Centaurea cyanus produced flowers by six weeks after planting. Plant dry weight did not appear to be related to mean seed weight (Fig. 3). Relative growth rate (RGR) based on changes in dry weight (Fig. 4) was lowest for Centaurea cyanus (0.010 g g“' day ') and greatest for Cirsium douglasii. (0.030 g g~ ' day ^ ‘ ). Centaurea solstitialis was second highest (0.024 g g”' day^') however, there was no significant linear regression relationship between RGR and mean seed weight for the species examined (Table 3). Net photosynthesis measurements. Net photo- synthesis (An) varied from mean value of 8.4 pmol CO2 m“^ s“* for Centaurea cyanus to 12.5 jumol CO2 m“- s“‘ for Centaurea solstitialis (Table 4). Under these conditions, differences among species were not significant (analysis of variance, F3 25 = 1.48, P = 0.25). There did not appear to be a significant relationship between initial seed size and net photosynthesis (Table 3). SPAD values ranged from 25.8 for Centaurea cyanus to 37.2 for Cynara carduncuius subsp. flavescens (Table 4). Differences among species means were significant (analysis of variance, ^3,21 = 3.94, P = 0.03). There also did not appear to be a significant relationship between seed size and SPAD value (Table 3). Discussion In the present evaluation of 22 Cardueae taxa seed size characteristics were not significantly related to whether or not a taxon was considered invasive, suggesting that for these taxa seed weight is not a good predictor of invasiveness. Small mean seed weight was associated with invasiveness among species in the genus Pinus L. (Rejmanek 1995; Grotkopp 2002); however, this does not always appear to be the case, as others have associated increased seed size with invasive- ness (Daws et al. 2007). For agricultural weeds, some of the most rapidly spreading weeds had relatively large heavy seeds (Forcella 1985). The values for seed weight reported in this paper are similar to the few previous reports for these taxa. Mean seed weight for Centaurea solstiticdis has been previously reported as 1 .42 mg for seeds without a pappus and 1.18 mg for seeds with a pappus (Graebner et al. 2010). Widmer et al. (2007) reported that Centaurea solstitialis mean seed weights varied from 0.85- 1.91 mg with a mean value of 1.22 mg for seeds 344 MADRONO [VoL 61 Fig. 1. Seed N% (A), C% (B), and C:N ratio (C) for 17 Cardueae taxa. Values are the mean ±95% confidence limits. Species are arranged from left to right in the order of increasing seed weight given in Table 2. Bars filled with cross hatch pattern are for species designated as invasive by California Invasive Plant Council (Cal-IPC 2006) (see Table 2). collected at ten worldwide locations. Hierro et al. (2013) included a graph, which showed Centaurea solstitialis seed weights for seeds from Argentina and Turkey. Their values were quite close to the mean reported here (1.48 mg). Ghavami and Ramin (2008) reported that Silybum marianum seeds had an average weight of 23.9 mg for wild type plants or 21.4 mg for “Royston” type plants grown in a pot experiment in Iran. In the present study Silybum marianum seeds were 23.1 mg each on average. Foti et al. (1999) reported that individual Cynara cardunculus L. var. sylvestris Lam. seeds (wild type) were 19 or 21 mg per seed, and seeds of a related crop species {Cynara cardunculus L. var. altilis DC.) were 26 or 29.5 mg per seed in a two year field experiment in Italy. A report by Archontoulis et al. (2010) indicated a range of values for weight of individual Cynara cardunculus subsp. flavescens seeds of 26-56 mg. In our study the range was 1 1^8 mg per seed. 2014] SPENCER ET AL.: FUNCTIONAL TRAIT VARIATION IN THISTLES 345 Fig. 2. Plant height (A) and leaf area (B) during the first nine weeks after planting for five taxa, which differed in seed size {Cirsium andersonii^ Cirsium cyanus, Cirsium douglasii, Centaurea solstitiaiis, and Cynara cardunculus subsp. flavescens). Values are the mean ±1 SE, N for leaf area varied from six to 23 depending on the number of leaves present. N for plant height varied from two to five. with an average value of 35.6 mg per seed. The slightly higher maximum values reported by Archontoulis et al. (2010) may in part be due to the fact that they are from experimental plots where the plants had been raised as crops, implying that some of the growing conditions had been optimized (i.e., weed control, fertilizers added). Considering that the values for Cynara cardunculus subsp. flavescens in our paper are from naturally growing populations growing under ambient conditions, the differences in seed size are small. To the best of our knowledge, the data on seed weight for the remaining taxa reported in this paper represent new information. Variation in seed weight, as measured by the coefficient of variation (CV), in this study was similar to previous reports with the following caveat. Differences in how seed weight was determined make it difficult to compare variation in seed weight with previous studies. Many previous studies report seed weight per 100 seeds or some other number of seeds. Thus, the seed weight is actually a mean and as a result, any estimates of variation based on these means are under-estimates. The CVs for the present data were estimated from individually measured seeds. Nevertheless, the CVs for seeds from the 22 taxa reported here are in line with previous estimates. 346 MADRONO [Vol. 61 2 6 10 14 18 21 2 6 10 14 18 21 2 6 10 14 18 21 2 6 10 14 18 21 2 6 10 14 18 21 Centaurea Centaurea Cirsium Cirsium Cynara cardunculus solstitialis (1) cyanus (4) douglasii (7) andersonii (13) subsp. flavescens (36) Fig. 3. Mean dry weight (grams) allocated to roots, shoots, and flowers over time (two, six, 10, 14, 18, or 21 wk after planting) for five taxa (Cirsium cmdersonii, Cynara carchinculus subsp. jlavescens, Cirsium cyanus, Cirsium douglasH, and Centaurea solstitialis), which differed in seed size. The number in parentheses after the species name is the mean seed mass rounded to the nearest mg. Total dry weight (bar height) differed significantly due to species (P < 0.0001), time (P < 0.0001), and the interaction term, species by time (P < 0.0001) based on analysis of variance. For example, CV for Agropyron intermedium (Flost) Veauv. was 22% (Hunt and Miller 1965), for An t ho X ant hum odoratum L. was 7,2% (Anto- novics and Schmitt 1986), and for Arena sativa L. was 15.3% (Murphy and Frey 1962). Significant values of skewness or kurtosis indicate that the seed weight distributions have more elongated distributions at one end or may have more than one peak. Either kurtosis, skewness, or both for seed weight distributions in this study were significantly different from zero, for 1 1 of the 22 taxa examined. In other studies, significant skewness or kurtosis indices indicate that environmental conditions influence seed weight more than maternal effects (Tungate et al. 2002; Tiscar and Lucas 2010). Seeds from Fig. 4. Relative growth rate (RGR ± standard error) based on changes in dry weight over time for five species {Cirsium andersonii, Cirsium cyanus\ Cirsium douglasii, Centaurea solstitialis, and Cynara cardunculus subsp. jlavescens), which differed in seed size. The number in parenthesis to the right of the taxon name is the mean seed mass rounded to the nearest milligram (mg). RGR was not significantly related to seed mass based on linear regression (Table 3). 2014] SPENCER ET AL.: FUNCTIONAL TRAIT VARIATION IN THISTLES 347 Table 3. Linear Regression Equations Relating Plant Growth and Condition Characteristics to Seed Mass. Units are as follows: RGR (g g""' day“'); leaf area (cnV); SPAD (instrument units); mean plant height (height), (cm); A^ (pmol CO2 m“^ s ’); seed mass (mg). P > F is the probability of obtaining a greater F-value and R- is the proportion of variation explained by seed mass. Equation DF F-value P > F R2 RGR = 0.022 - 0.00011 X seed mass 1, 3 0.09 0.80 0.04 Leaf Area = 41.3 + 0.77 X seed mass 1, 3 10.49 0.08 0.84 SPAD = 30.9 + 0.18 X seed mass 1, 3 0.70 0.49 0.26 Height = 10.8 + 0.05 X seed mass 1, 3 0.19 0.71 0.09 An = 10.61 — 0.048 X seed mass 1, 3 0.39 0.60 0.16 the taxa in this study came from locations, which encompass the entire north to south range examined in this study, and in several different years. No clear-cut pattern relating these factors to skewness or kurtosis of seed weight distribu- tions is apparent from these data. An alternative explanation of significant skewness or kurtosis indices is that they indicate a strategy that is a form of bet hedging. That, is to say that seeds of different dispersal abilities (i.e., different weights) will likely land in a larger variety of microhabitats making it more likely that some will survive, especially in variable environments (Cohen 1966). While such a strategy may seem to favor successful invasions, there is no evidence for it from these data. Tissue C and N values for seeds of Cardueae taxa reported here were similar to values reported for plant species in other published studies, 40% and 2^%, respectively (Mengel and Kirkby 1982). Mattson (1980) reported that N content of different plant tissues ranged from 0. 3-7.0% with highest concentrations (3-7%) occurring in growing tissues or in storage organs such as seeds. There is very little published information on the nutrient content of seeds of the taxa we analyzed. Foti et al. (1999) have published crude protein values for Cyncira cardunciilus subsp. flavescens seeds, which are equal to an N content of 3.21%. This is similar to the value for Cynara carduncuhis subsp. flavescens, 3.0%, reported in the present study and within the 95% confidence limit shown in Figure 1. Seed C content for some of the seeds analyzed in this study was higher than the typical value for plant tissue (40%). Information on seed nutrient content may be useful in understanding interactions between plants and the animals, which consume seeds. For example, plant tissue C:N ratios <17 are believed to be favorable to herbivores (Russell- Hunter 1970; McMahon et al. 1974). Values > 17 are believed to be N limited for invertebrate growth (Russell-Hunter 1970; McMahon et al. 1974). For example insect herbivores forced to compensate for feeding on low quality plant material (e.g., C:N ratio > 17) by consuming greater amounts of it may wear out their mouth parts more quickly (Karban and Baldwin 1997) or may ingest larger than desirable quantities of plant defense chemicals which may adversely impact their performance (Slansky and Wheeler 1992). Thus, it appears that some of the seeds examined may be a less desirable food item in some cases. Under the conditions of the greenhouse experiment, only two species produced flowers. This implies that environmental controls of flowering (e.g., photoperiod) may differ among the five species examined. There is limited data for An (net photosynthesis) values for the four species measured in this study (Archontoulis et al. 2012; Dukes 2002). An values for these species at 1200 pmol CO2 m“^ s~‘ were toward the lower range of values reported for three energy crops whose light saturated An values ranged from 5 to > 35 pmol CO2 s“‘ (Archontoulis et al. 2012). The RGR values for these greenhouse grown plants ranged from 0.01 0-0.030 gg“' day“‘. These values are toward the lower end of the range (0.01-0.13 g g“^ day“‘) previously reported for herbaceous species (Houghton et al. 2013). In the present study, there was no significant relationship between seed weight and RGR. Prior research has shown evidence that in some cases plants with larger seeds have lower RGR (Maranon and Grubb 1993) but not in all cases (Buckley et al. 2003). Other measures of plant growth and condition (e.g., mean plant height. Table 4. Net Photosynthesis (An) and SPAD Values for Four Thistle Species Grown in the Greenhouse Experiment. Values are the mean, standard error, and number of leaves measured. Within each date species are sorted from top to bottom by seed size: 1, 4, 7, and 36 mg. Species SPAD (Units) An (pmols CO2 m^ s ') Centaurea solstitialis 33.1 (1.9, N = 7) 12.51 (1.25, N = 8) Centciurea cyanus 25.8 (5.32, N = 3) 8.40 (1.42, N = 4) Cirsium douglasii 37.0 (1.6, N = 7) 10.61 (1.67, N = 8) Cynara cardimculus sxxhsy. flavescens 37.2 (1.6, N = 5) 9.29 (0.87, N = 6) 348 MADROto [VoL 61 mean leaf area, mean leaf SPAD, mean net photosynthetic rate) were also unrelated to seed weight for the species examined (Table 3). Relative biomass allocated to reproduction by plants including variation in seed size, can be influenced by environmental conditions (Thomp-- son and Stewart 1981; Benech Arnold et al. 1992). Thus, information on seed size variation may contribute to understanding the distribution and abundance of Cardueae taxa. Some have suggested that functional characteristics such as seed size may be useful for predicting whether a species will be an invasive plant (Ordonez et ah 2010). The data on seed size variation, nutrient content, RGR, An, and leaf chlorophyll content for Cardueae taxa exam- ined in this study do not support that hypothesis. This is perhaps not a surprising finding. Grime et al. (1988, see pp. 647) warned that while certain attributes may have predictive value in specific systems, this approach fails to be useful in all cases. They indicate that this is because 1) the same attribute may have different importance to different organisms and 2) natural selection affects more than one of a species’ attributes at any one time resulting in a correlated set of attributes, which reflect ecological specialization. Results from recent studies agree with the warning and imply that it is unlikely that a universal suit of plant attributes exists, which will provide an explanation for the distribution of alien plant species (Tecco et al. 2010; Dawson et al. 2011). Acknowledgments The comments of G. Kyzer and B. Blank who read an earlier version of this manuscript helped improve it. G. Ksander provided technical support. Mention of a manufacturer does not constitute a warranty or guarantee of the product by the U.S. Department of Agriculture nor an endorsement over other products not mentioned. The U.S. Department of Agriculture is an equal opportunity employer. Literature Cited Antonovics, J. and J. Schmitt. 1986. Paternal and maternal effects on propagule size in Anthox- anthum odoratum. Oecologia 69:277-282. Archontoulis, S. V., P. C. Struik, X. Yin, L. Bastiaans, j. Vos, and N. G. Danalatos. 2010. Inflorescence characteristics, seed composition, and allometric relationships predicting seed yields in the biomass crop Cynara cardunculus. Global Change Biology Bioenergy 2:113-129. , X. Yin, j. Vos, N. G. Danalotos, and P. C. Struik. 2012. Leaf photosynthesis and respiration of three bioenergy crops in relation to temperature and leaf nitrogen: how conserved are biochemical model parameters among crop species? Journal of Experimental Botany 63:895-911. Baldwin, B. G., D. H. Goldman, D. J. Keil, R. Patterson, T. J. Rosatti, and D. H. Wilken. (eds.). 2012. The Jepson manual: vascular plants of California. 2nd ed. University of California Press, Berkeley, CA. Barkley, T. M., L. Brouillet, J. L. Strother. 2006. Asteraceae Martinov tribe Cardueae Cassini. Pp. 3- 12, 82-195 in Flora of North America Editorial Committee (eds.) Flora of North America North of Mexico, Vol. 19: Magnoliophyta: Asteridae (in part): Asteraceae, part 1. Oxford University Press, New York, NY. Benech Arnold, R. L., M. Fenner, and P. J. Edwards. 1992. Mineral allocation to reproduc- tion in Sorghum bicolor and Sorghum haiepense in relation to parental nutrient supply. Oecologia 92:138-144. Buckley, Y. M., P. Downey, S. V. Fowler, R. Hill, J. Memmot, H. Norambuena, M. Pitcairn, R. Shaw, A. W. Sheppard, C. Winks, R. Witten- berg, and M, Rees. 2003. Are invasives bigger? A global study of seed size variation in two invasive shrubs. Ecology 84:1434—1440. California Invasive Plant Council (CAL-IPC). 2006. California Invasive Plant Inventory. CaLIPC Publication 2006-02. California Invasive Plant Council: Berkeley, CA. Website www.cal-ipc.org (accessed 4 June 2013). Cohen, D. 1966. Optimizing reproduction in a randomly varying environment. Journal of Theo- retical Biology 12:119-121. Dalling, j. W. and S. P. Hubbell. 2002. Seed size, growth rate and gap microsite conditions as determinants of recruitment success for pioneer species. Journal of Ecology 90:557-568. Daws, M. L, J. Hall, S. Flynn, and H. W. Pritchard. 2007. Do invasive species have bigger seeds? Evidence from intra-and inter-specific comparisons. South African Journal of Botany 73:138-143. Dawson, W., D. F. R. P. Burslem, and P. E. Hulme. 2011. The comparative importance of species traits and introduction characteristics in tropical plant invasions. Diversity and distributions 17:1111- 1121. Dukes, J. S. 2002. Comparison of the effect of elevated CO2 on an invasive species (Centaurea sobtitialis) in monoculture and community settings. Plant Ecology 160:225-234. Forcella, F. 1985. Final distribution is related to rate of spread in alien weeds. Weed Research 25:181-191. Foti, S., G. Mauromicale, S. A. Raccuia, B. Fallico, F. Fanella, and E. MacCarone. 1999. Possible alternative utilization of Cynara spp. 1. Biomass, grain yield and chemical compo- sition of grain. Industrial Crops and Products 10:219-228. Ghavami, N. and A. A. Ramin. 2008. Grain yield and active substances of milk thistle as affected by soil salinity. Communications in Soil Science and Plant Analysis 39:2608-2618. Graebner, R. C., R. M. Callaway, and D. Montesinos. 2012. Invasive species grows faster, competes better, and shows greater evolution toward increased seed size and growth than exotic non-in vasive congeners. Plant Ecology 213:545- 553. Grime, J. P., J. G. Hodgson, and R. Hunt. 1988. Comparative plant ecology: a functional approach to common British species. Unwin Hyman Ltd. London, England. 2014] SPENCER ET AL.: FUNCTIONAL TRAIT VARIATION IN THISTLES 349 Gross, K. L. 1984. Effects of seed size and growth form on seedling establishment of six monocarpic perennial plants. The Journal of Ecology 72: 369-387. Grotkopp, E., M. RejmAnek, and T. L. Rost. 2002. Toward a causal explanation of plant invasiveness: seedling growth and life-history strategies of 29 pine (Pinus) species American Naturalist 159:396- 419. Henderson, C. B. 1990. The influence of seed apparency, nutrient content and chemical defenses on dietary preference in Dipodomys ordii. Oecolo- gia 82:333-341. Hendry, G. A. and J. P. Grime, (eds.). 1993. Methods in comparative plant ecology: a labora- tory manual. Chapman & Hall, London, England. Hickman, J. C. (ed.). 1993. The Jepson Manual: Higher plants of California. University of Califor- nia Press, Berkeley, CA. Hierro, J. L., 6. Eren, D. Villarreal, and M. C. Chiuffo. 2013. Non-native conditions favor non- native populations of invasive plant: demographic consequences of seed size variation? Oikos 122: 583-590. Hoagland, D. R. and D. I. Arnon. 1950. The water- culture method for growing plants without soil. California Agricultural Experimental Station Cir- cular 347, 2nd ed. revised by D. I. Arnon, The College of Agriculture, University of California, Berkeley, CA. Houghton, J., K. Thompson, and M. Rees. 2013. Does seed mass drive the differences in relative growth rate between forms? Proceedings of the Royal Society B: Biological Sciences 280. doi/ 10.1098/rspb.2013.092L Hunt, R. 1982. Plant growth curves: the functional approach to plant growth analysis. Edward Ar- nold, London, England. Hunt, O. J. and D. G. Miller. 1965. Coleoptile length, seed size, and emergence in intermediate wheatgrass (Agropyron intermedium [Host] Beauv.). Agronomy Journal 57:192-195. Karban, R. and I. T. Baldwin. 1997. Induced responses to herbivory. University of Chicago Press, Chicago, IL. Keil, D, j. 2006. Cirsium. Pp. 95-164 in Flora of North America Editorial Committee (eds.). Flora of North America, North of Mexico, Vol. 19: Magnoliophyta: Asteridae (in part), Asteraceae, part 1. Oxford University Press, New York, NY. Maranon, T. and P. j. Grubb. 1993. Physiological basis and ecological significance of the seed size and relative growth rate relationship in Mediterra- nean annuals. Functional Ecology 7:591-599. Mattson, W. J., Jr. 1980. Herbivory in relation to plant nitrogen content. Annual Review of Ecology and Systematics 11:119-161. McMahon, R. F., R. D. Hunter, and W. D. Russell-Hunter. 1974. Variation in aufwuchs at six freshwater habitats in terms of carbon biomass and of carbon: nitrogen ratio. Hydro- biologia 45:391^04. Mengel, K. and E. a. Kirkby. 1982. Principles of Plant Nutrition. 3rd ed. International Potash Institute, Worblaufen-Bern, Switzerland. Milberg, P., L. Anderson, and K. Thompson. 2000. Large-seeded species are less dependent on light for germination than small-seeded ones. Seed Science Research 10:99-104. Murphy, C. F. and K. J. Frey. 1962. Inheritance and heritability of seed weight and its components in oats. Crop Science 2:509-512. Ordonez, A., I. J. Wright, and H. Olff. 2010. Functional differences between native and alien species: a global-scale comparison. Functional Ecology 24:1353-1361. Petersen, T. D., Z. Ning, and M. Newton. 1990. Dynamics of size structures in seedling stands of Fraxinus mandshurica in Northeast China. Annals of Botany 66:255-263. RejmAnek, M. 1995. What makes a species invasive? Pp. 3-13 in P. PysekK. PrachM. Rejmanek and M. Wade. (eds.). Plant invasions: general aspects and sjpecial problems. Workshop held at Kostelec nad Cernymi lesy, Czech Republic, 16-19 September 1993. SPB Academic Publishing, Amsterdam. , G. R. Robinson, and E. RejmAnkovA. 1989. Weed-crop competition: experimental designs and models for data analysis. Weed Science 37:276-284. Russell-Hunter, W. D. 1970. Aquatic productivity. The Macmillan Company, London, England. SAS Institute Inc. 2011, SAS OnlineDoc® 9.3. SAS Institute Inc., Cary, NC. Slansky, F. and G. S. Wheeler. 1992. Caterpillars’ compensatory feeding response to diluted nutrients leads to toxic allelochemical dose. Entomologia Experimentalis et Applicata 65:171-186. SOKAL, R. R. and F. j. Rohlf. 1995. Biometry: the principles and practice of statistics in biological research. 3rd ed. W. H. Freeman and Company, New York, NY. Stanton, M. L. 1984. Seed variation in wild radish: effect of seed size on components of seedling and adult fitness. Ecology 65:1105-1112. Tecco, P. a., S. DIaz, M. Cabido, and C. Urcelay. 2010. Functional traits of alien plants across contrasting climatic and land-use regimes: do aliens join the locals or try harder than them? Journal of Ecology 98:17-27. Thompson, K. and A. J. A. Stewart. 1981. The measurement and meaning of reproductive effort in plants. American Naturalist 117:205-211. Thomson, F. J., A. T. Moles, T. D. Auld, and R. T. Kingsford. 2011. Seed dispersal distance is more strongly correlated with plant height than with seed mass. Journal of Ecology 99:1299-1307. TIscar, P. and M. Lucas. 2010. Seed mass variation, germination time and seedling performance in a population of Pinus nigra subsp. salzamannii. Forest Systems 19:344—353. Tungate, K. D., D. j. Susko, and T. W. Rufty. 2002. Reproduction and offspring competitiveness of Senna obtusifolia are influenced by nutrient availability. New Phytologist 154:661-669. Uddling, j., j. Gelang-Alfredsson, K. Piikki, and H. Pleijel. 2007. Evaluating the relationship between leaf chlorophyll concentration and SPAD-502 chlorophyll meter readings. Photosyn- thesis Research 91:37^6. WiDMER, T. L., F. Guermache, M. Y. Dolgovskaia, AND S. Y. Reznik. 2007. Enhanced growth and seed properties in introduced vs. native populations of yellow starthistle (Centaurea solstitialis). Weed Science 55:465-473. Madrono, Vol. 61, No. 4, pp. 350-366, 2014 VERNAL POOL BLUE DICKS {DICHELOSTEMMA LACUNA-VERNALIS; ASPARAGACEAE: BRODIAEOIDEAE) REVISITED Robert E. Preston ICE International, 630 K Street, Suite 400, Sacramento, CA 95814 robert.preston@icfi.com Abstract Dichelostemma lacuna-vernalis L.W. Lenz was proposed in 1974 for populations of a diminutive Dichelostemma associated with vernal pool terrain. The author of Dichelostemma in the 1993 edition of The Jepson Manual did not accept the new species, reporting that the morphological and ecological characteristics of D. lacuna-vernalis were within the ranges for D. capitatum (Benth.) Alph.Wood. The purpose of this paper was to test the validity of D. lacuna-vernalis using a morphometric analysis of eighteen morphological characters in populations of D. capitatum and D. lacuna-vernalis sampled in the field and by comparing plants of both taxa grown under common garden conditions. The data were subjected to cluster analysis, principal components analysis, and discriminant analysis. The results of the analysis confirm the morphological distinctness of D. lacuna-vernalis and support its recognition as a separate taxon. Based on current taxonomic concepts in the Brodiaeoideae, tMs taxon is best recognized at subspecies rank, as D. capitatum subsp. lacuna-vernalis (L. W. Lenz) D.W. Taylor. Populations of D. capitatum subsp. lacuna-vernalis are distinguished by their short scapes (ca. 15 cm), inflorescences with one or two flowers, short (<4 mm) perianth tubes, and outer perianth lobes that are ovate, decurrent at the base, and wider than the inner perianth lobes. Key Words: California, Dichelostemma, geophyte, taxonomy, Themidaceae. The genus Dichelostemma Kuiith currently consists of five geophyte species endemic to the western USA and northern Mexico (Pires 2002, Piles and Keator 2012). Phylogenetic studies place Dichelostemma in Themidaceae (Fay and Chase 1996; Fay et al. 2000; Pires et ah 2001; Pires and Sytsma 2002) and more recently in subfamily Brodiaeoideae of the Asparagaceae (Chase et al. 2009; Steele et ah 2012). These studies also indicate that Dichelostemma is not monophyletic; one species, Dichelostemma capi- tatum (Benth.) Alph.Wood, is sister to the clade that includes Brodiaea Sm. and the other four species of Dichelostemma. Dichelostemma capitatum has been the subject of nearly perpetual taxonomic confusion since the early 19th century, so much so that Keator (1992) dubbed it a “problematic” species. It is the only hexandrous species in the genus, which prompted Baker (1871) to propose moving it to the genus Muilki S. Watson. On the same basis, Rydberg (1912) proposed placing the species in its own genus, Dipterostemon Rydb., into which he also placed three other taxa now treated as synonyms or subspecies of D. capitatum. Rydberg’s argu- ment that the possession of six stamens was sufficiently diagnostic to warrant segregation of this new genus seemed weakly justified. However, D. capitatum possesses multiple other characters that further differentiate it from other Dichelostemma species. Hoover (1940) observed that D. capitatum produces cormlets at the base of the corms and at the ends of short stalks, whereas all other species of Dichelostemma produce cormlets only at the base of the corm. Keator (1968) noted many additional differences between D. capitatum and the other species of Dichelos- temma in leaf width, pubescence of the scape, arrangement of tracheids in the stem, shape of the seed coat cells, and the seed genuination pattern. Keator (1991) also noted that D. capitatum does not hybridize with other Dichelostemma species, whereas the other species hybridize with each other. In D. capitatum, the six stamens are united at the base of the filaments into a short staminal tube via fusion of the connective tissue, a feature not present in the other species of Dichelostemma (Lenz 1976). Moreover, the staminal tube pos- sesses six lanceolate appendages that extend upward and cover the anthers and style, similar but not homologous to the corona found in other species of Dichelostemma, which is an extension of the perianth (Lenz 1976). Berg (1996) proposed resurrecting Dipterostemon on the basis of em- bryology. Although the embryology of Brodiaea and Dichelostemma is quite similar, the inner integument of the ovule of D. capitatum consists of two cell layers, similar to that of Muilla and Triteieia Douglas ex LindL, but different from the multilayered inner integument that represents a synapomorphy of Brodiaea and the other Diche- lostemma species (Berg 1978, 1996, 2003). The morphological evidence for recognizing Dipteros- temon is fully supported by the molecular data, which show that Dichelostemma is only mono- phyletic if D. capitatum is excluded (Pires et al. 2001; Pires & Sytsma 2002; Nguyen et al. 2008; Steele et al. 2012), 2014] PRESTON: DICHELOSTEMMA LACUN A-VERN ALIS REVISITED 351 Although D. capitatum exhibits a high degree of morphological variation, only a single infra- specific taxon within D. capitatum is currently recognized, D, capitatum subsp. pauciflorum (Torr.) Keator. However, Lenz (1974) proposed that populations of diminutive Dichelostemma capitatum associated with vernal pool terrain be recognized as a new species, D. lacuna-vernalis L.W. Lenz. Lenz characterized the new species as morphologically similar to D. capitatum but differing by having broad, keel-less leaves, 1-3- flowered inflorescences, shorter scapes, and smaller bracts. Lenz also indicated that D. lacuna-vernalis occurred in different habitats than D. capitatum, although he was not specific about the habitat differences, Keator (1991) did not accept the new taxon, reporting that the mor- phological and ecological characteristics of D. lacuna-vernalis were within the ranges for D. capitatum. He suggested that smaller stature and fewer flowers were a result of environmentally- induced phenotypic plasticity, a consequence of stress from growing in seasonally saturated soils, i.e., that plants referable to D. lacuna-vernalis were simply on the low end of the normal range of size variation for D. capitatum. In the treatment of Dichelostemma for The Jepson Manual, Keator (1993) placed D. lacuna-vernalis in synonymy with D. capitatum. Because D. lacuna-vernalis was based on a single population, and because Lenz’ characterization of the differ- ences between D. capitatum and D. lacuna- vernalis was very general, and the protologue lacked a key to differentiate between the species, perhaps Keator was justifiably conservative in not recognizing D. lacuna-vernalis. Fires (2002) and Fires and Keator (2012) concurred with Keator’s treatment of D. capitatum but acknowl- edged that further study of the taxon was warranted. In March 2007, I encountered a population of diminutive Dichelostemma plants in Butte County growing sympatrically with a population of typical D. capitatum. The plants matched Lenz’ description of D. lacuna-vernalis, and on further inspection, I found that in addition to their short stature and few-flowered inflorescences, the perianth tubes were very short, a feature that Hoover (1940) had earlier noted in depauperate plants of D. capitatum. In addition, I observed that the outer perianth lobes were broadly ovate with cordate bases, unlike the oblong, truncate- based perianth lobes of typical D. capitatum. These observations prompted me to initiate a closer comparison of D. capitatum and D. lacuna- vernalis. Keator’s (1991) hypothesis that D. lacuna- vernalis does not warrant taxonomic recognition rested on two assumptions: first, that morpho- logical variation in D. capitatum is continuous from robust plants to depauperate plants, i.e.. there is no morphological discontinuity that that reliably distinguishes D. lacuna-vernalis from D. capitatum; and second, that the primary source of variation among populations stems from a response to environmental factors, i.e., that the morphology of plants assignable to D. lacuna- vernalis is the result of phenotypic plasticity. The purpose of this study is to test Keator’s hypothesis by addressing each of the underlying assumptions: 1) does the range of morphological variation in populations assignable to D. lacuna- vernalis overlap continuously with that of D. capitatum; and 2) do plants from D. capitatum populations and from D. lacuna-vernalis popula- tions differ morphologically when grown under the same environmental conditions? To answer the first question, I sampled populations of D. capitatum throughout northern California, in- cluding populations assignable to D. lacuna- vernalis on the basis of characters proposed by Lenz (1974), and subjected the data to a morphometric analysis. To answer the second question, I collected corms from populations of both putative taxa, grew them in pots in a common garden, and compared their morpholo- gy both to each other and to their source populations. Methods Fopulation Sampling Between 2007 and 2014, I sampled 59 Diche- lostemma populations in northern California, primarily from the eastern Sacramento Valley and adjacent Sierra Nevada foothills, but also from the interior North Coast Ranges and other scattered locations (Appendix 1). For each population, plants were collected with intact corms or were placed in water to prevent the flowers from wilting before measurements were made. I measured scape height, maximum leaf width, number of flowers, maximum bract length, and maximum pedicel length. I dissected one flower from 10-30 plants in each population, using flowers at approximately the same stage of anthesis, to minimize variation due to any change in flower size from the beginning to the end of anthesis. I measured 13 floral characters and noted the shape and position of the floral parts. Common Garden Flants I grew plants in a common garden from corms collected from the populations of D. capitatum and D. lacuna-vernalis sampled for the morpho- logical characters. The corms were planted in 8- inch pots using a commercial potting mix. Flants were grown together outside under conditions of ambient temperature, light, and rainfall, with occasional supplemental watering. I measured the 352 MADRONO [Vol. 61 same set of morphological characters for the garden-grown plants that were sampled for populations in the field. I also measured corm characteristics, including the number and size of cormlets produced. Morphological Analysis Sixteen populations were assigned to D. lacuna-vernalis (L01-L16 in Appendix 1) based on their occurrence in vernal pool terrain and on the following diagnostic characters from Lenz (1974) and from my own personal observations: stems less than 2 dm tall; inflorescence with 1-3 flowers; perianth tube less than or equal to 4 mm long; and, outer perianth lobes broadly ovate with cordate bases. The other 43 populations were assigned to D. capitatum (C01-C43 in Appendix 1). The field-collected data were analyzed using cluster analysis, principal components analysis (PCA), and discriminant analysis (DA). The cluster analysis and PCA were performed using the character means from each population (data matrix provided in Appendix 2), and the DA used the individual measures from each plant sampled. Prior to the analysis, the data were standardized by subtracting the mean of each variable and dividing by the standard deviation. The cluster analysis employed Ward’s method and Euclidean distances. The DA was first performed using two groups, one consisting of plants assigned to populations of D. lacuna-vernalis and the second consisting of the plants assigned to populations of D. capitatum. The DA employed a forward stepwise analysis to identify characters with the highest discriminant power. A classification tree analysis was performed to further test the predictive value of each character and to determine the split between values for D. capitatum and D. lacuna-vernalis. All statistical tests were carried out using the SYSTAT 13 statistics program (SYSTAT Software, Chicago, IL). Results Morphological Analysis Cluster analysis. The cluster analysis found that the populations form several clusters that are largely distinct (Fig. 1). The first cluster (Group 1) includes 21 populations of D. capitatum from scattered locations throughout northern Califor- nia (C1-C21). The second cluster (Group 2) includes the D. lacuna-vernalis populations (Ll- L16) but also includes 22 populations of D. capitatum that share some character states with D. lacuna-vernalis and others with D. capitatum populations in Group 1 . Group 2 is comprised of two subclusters, the first of which (Group 2A) includes populations of D. capitatum from the northern Sierra Nevada foothills and the interior North Coast Ranges and two (C22-C43) popu- lations of D. lacuna vernalis (LI 5, LI 6), and the second of which (Group 2B) contains the remaining populations of D. lacuna-vernalis (L1-L14). Principal components analysis. The plots of the principal component scores graphically illustrate the morphological distinctiveness of D. lacuna- vernalis (Fig. 2a, b). Moreover, the PCA found that the two groups of D. capitatum populations found in the cluster analysis also show little overlap. The first principal component (Factor 1), which accounts for 49.2% of the variation, appears to be a general size factor (Table 1). Populations of the diminutive D. lacuna-vernalis are grouped at the low end of Factor 1 . The more robust populations of D. capitatum, which correspond to Group 1 in the cluster analysis, are grouped at the high end of Factor 1, and populations of D. capitatum in Group 2a in the cluster analysis are in an intermediate position (Fig. 2a). The second principal component, which accounts for 21.8% of the variation, is also a size factor, but it loads primarily on the length and width of the petals. Factor 2 provides little separation of the three groups. The third factor, which accounts for 13.8% of the variation, loads primarily on ovule number but also appears to be a function of the relative lengths of the outer perianth lobes and the style. The D. capitatum populations corresponding to Group 2a in the cluster analysis are differentiated from D. lacuna-vernalis and the more robust D. capitatum populations (Group 1) along the axis of Factor 3 and are not in an intermediate position (Fig. 2b). Discriminants analysis. The DA for all individ- uals from all populations found D. lacuna- vernalis to be morphologically distinct from D. capitatum (Wilks’ X = 0.224, df = 13, 1294, P < 0.001). Thirteen variables contributed significant- ly to the discriminant function, explaining 78% of the variation (Table 2). The DA correctly classi- fied 98% of all individuals, with 99% of the D. lacuna-vernalis individuals correctly classified. Among the plants grouped with D. capitatum, 98% of the individuals were correctly classified; of the 18 misclassified individuals, 16 came from the small-flowered Interior North Coast Range populations. Variables with the highest loadings were length of the perianth tube, style length, width of the outer perianth lobes, scape height, appendage length, and plant height (Table 2). A DA conducted using just the first five variables in Table 2 was nearly as successful as the full model (Wilks’ X = 0.260, df = 5, 1306, P < 0.001), correctly classifying 97% of all individuals. 2014] PRESTON: DICHELOSTEMMA LACUNA-VERNALIS REVISITED 353 Q. o < CM a 3 O (D 0 5 10 15 20 Distances Fig. 1. Cluster tree for 42 populations of D. capitatum and 16 populations of D. lacuna-vernalis, based on cluster analysis of the means of 18 characters. Group 1 includes 21 populations of D. capitatum from scattered locations throughout northern California (C1-C21), and Group 2 includes D. lacuna-vernalis populations (L1-L16) and 22 populations of D. capitatum (C22-C42) from the northern Sierra Nevada foothills and inner North Coast Ranges. Population codes refer to voucher specimens cited in Appendix 1. including 95% of the D. lacuna-vernalis individ- uals and 98% of the D. capitatum individuals. Morphological comparisons. The cluster analy- sis, PCA, and DC all show that D. lacuna-vernalis is morphologically distinct from D. capitatum. Although the range of values for each quantita- tive character measured for D. lacuna-vernalis overlaps with those for D. capitatum, the means for most characters are significantly smaller than 354 MADRONO [Vol. 61 Factor 1 Factor 1 Fig. 2. Plot of factor scores from principal components analysis of 42 populations of D. capitatum and 16 populations of D. lacima-vernalis, based on the means of 18 morphological characters. A. Factor 1 vs. Factor B. Factor 1 vs. Factor 3. Closed squares = D. capitatum. populations in cluster analysis Group 1, open squares = capitatum populations in cluster analysis Group 2, closed circles = D. lacima-vernalis. b rj 2014] PRESTON: DICHELOSTEMMA LACUN A-VERN ALIS REVISITED 355 Table 1. Results of the Principal Components Analysis of 59 Populations of Dichelostemma CAPITATUM and D. LACUN A-VERN ALIS, BASED ON 18 Morphological Characters. Variable Factor 1 Factor 2 Factor 3 Scape height 0.826 -0.278 0.082 Leaf width 0.792 -0.114 0.280 Flowers 0.825 -0.404 0.244 Bract length 0.917 -0.039 0.195 Pedicel length 0.861 -0.244 0.306 Perianth tube length 0.869 -0.225 -0.321 Outer lobe length -0.118 0.735 -0.594 Inner lobe length 0.186 0.814 -0.360 Outer lobe width -0.495 0.790 0.232 Inner lobe width 0.244 0.826 0.259 Appendage length 0.913 0.203 -0.223 Outer filament length 0.379 0.733 0.363 Inner filament length 0.615 0.454 0.462 Outer anther length 0.723 0.348 -0.374 Inner anther length 0.895 0.178 -0.174 Ovary length 0.940 0.063 0.148 Style length 0.773 0.065 -0.569 Ovules -0.127 0.328 0.772 Eigenvalues % of total variance 8.856 3.928 2.490 explained 49.2 21.8 13.8 those for D. capitatwn (Table 3), and the distribution of values is non-unimodal. The most readily observed differences between D. lacuna- vernalis and D. capitatum are the very short perianth tube (relative to the lobes) and the ovate (vs. oblong) outer perianth lobes that are wider than the inner lobes (Fig. 3). Other characters, such as short stature and few flowers per scape, are also found in D. capitatum populations that cluster in Group 2a. Although Lenz (1974) characterized the leaves of D. lacuna-vernalis as quite broad, the leaves of field collected plants were narrower than those of D. capitatum. Based on the DA, scape height, perianth tube length, width of the outer perianth lobe, append- age length, and style length are the most useful characters for discriminating between D. lacima- vernalis and D. capitatum (Table 2). The classifi- cation tree analysis was used to determine the nodal values for each character and to construct a matrix for differentiating between D. capitatum and D. lacumuvernalis (Table 4). In addition, although flower number did not contribute significantly to the DA, it also appears to be a useful discriminator. Two characters not included in the DA, the ratio of the lengths of the perianth tube and perianth lobe and the relative width of the inner and outer perianth lobes, were also found to be highly predictive. Overall, the eight characters presented in Table 4 unambiguously assigned 89% of the sample individuals to the correct taxon, including 98% of the D. lacuna- vernalis individuals and 86% of the D. capitatum individuals. As might be expected from the cluster analysis, almost all of the misclassified individuals (having 5 or more characters in ranges for D. lacuna-vernalis) or ambiguous (having 4 characters in ranges for D. capitatum and 4 in ranges for D. lacima-vernalis) were members of D. capitatum Group 2a. When all of the individuals in a population were considered together, all of the D. lacima-vernalis populations were correctly classified, and only two of the 43 D. capitatum populations were misclassified. Most individuals in one D. capitatum population from the Sierra Nevada foothills (C25) were classified as D. lacuna-vernalis; these plants were small-statured and few-flowered but had large flowers. Another D. capitatum population from serpentine chaparral in the Interior Coast Range (C40) had many individuals classified as D. lacuna-vernalis; these plants were small-statured Table 2. Summary of Stepwise Discriminant Analysis of 18 Characters Measured for 1308 Field Sampled Dichelostemma capitatum and Dichelostemma lacuna-vernalis Individuals, Including the Structure Matrix Correlations Between Characters and the Canonical Discriminant Function Score. Five characters (flower number, bract length, pedicel length, length of the outer filament, and length of the inner anther) did not contribute significantly to the discriminant function. Step Character Wilks’ X Approximate F-ratio Approximate p-value Correlation coefficient 1 Perianth tube length 0.451 1,592.623 <0.001 0.841 2 Outer lobe width 0.355 1,188.106 <0.001 -0.686 3 Ovule number 0.313 954.301 <0.001 -0.586 4 Inner lobe width 0.272 871.634 <0.001 -0.143 5 Outer lobe length 0.26 742.549 <0.001 -0.012 6 Scape height 0.254 637.558 <0.001 0.634 7 Inner lobe length 0.247 567.402 <0.001 0.016 8 Appendage length 0.239 516.051 <0.001 0.634 9 Inner filament length 0.229 485.286 <0.001 0.021 10 Style length 0.226 443.462 <0.001 0.788 11 Ovary length 0.225 405.686 <0.001 0.488 12 Outer anther length 0.224 373.389 <0.001 0.348 13 Leaf width 0.224 345.117 <0.001 0.364 356 MADRONO [Vol. 61 Table 3. Means and Ranges for 19 Characters Measured in Field-Collected Individuals of Dichelostemma capitatum and D. lacuna-vernalis. Within each row, means with different superscripts differ significantly (ANOVA, P < 0.05). Dichelostemma capitatum Dichelostemma lacuna-vernalis n Mean Range n Mean Range Height 881 27.6" 3.1-64.3 378 14.4‘’ 4.4-26. 1 Outer leaf width 870 7.2" 2.5-25.0 378 5.3'^ 2.2-12.5 Flowers 883 5.0" 1-25 399 1.7'’ 1-5 Bract length (maximum) 846 13.5" 6.5-30.0 376 9.9'’ 4.8-16.0 Pedicel length (maximum) 850 5.1" 1-20.0 377 2.6” 0.7-6.0 Perianth length 1016 15.9" 10.8-22.9 387 13.6'’ 9.8-17.8 Tube length 1024 5.8" 3.2-10.3 390 3.6'’ 2.0-6.0 Lobe length (outer) 829 11.1 7.0-16.5 338 11.4” 7.5-14.6 Lobe length (inner) 1158 10.1 6.2-15.5 400 10.0” 6.3-13.3 Lobe width (outer) 857 5.4" 2.9-9. 1 374 7.1” 4.0-9.8 Lobe width (inner) 858 5.6 3.2-9.0 374 5.9” 3.2-9.0 Appendage length 836 6.1" 3.7-9.0 372 4.9” 3.0-6.7 Filament length (outer) 829 2.6 1.4-5. 2 367 2.7” 1. 5-3.8 Filament length (inner) 828 1.7 0.7-3.7 367 1.7” 0.5-2.8 Anther length (outer) 829 2.4 1. 0-^.0 367 2.1” 1.2-3. 0 Anther length (inner) 828 3.9 2.5-6.2 367 3.4” 2.4-4.5 Ovary length 836 4.1" 2.4-6. 1 366 3.5” 2. 0-5.0 Style length 835 6.0" 3. 3-9. 3 366 3.9” 2.5-5.6 Ovules per ovary 862 29.6" 9-54 367 38.2” 21-60 and small-flowered, but the floral proportions were more similar to those of D. capitatum. One unanticipated result of this analysis was that among the sampled populations of D. capitatum there are two morphologically distin- guishable forms. Both forms have long perianth tubes (relative to the lobes) and long styles. However, populations in cluster analysis Group 1 generally have longer scapes and produce many more flowers per scape than populations in Group 2a. As suggested by the cluster analysis, plants in Group 2a populations are more similar in some respects to D. lacuna-vernalis than to Group 1 populations. Plants in Group 2a Fig. 3. Flowers of D. capitatum (left, center), and D. lacuna-vernalis (right). The outer perianth lobes of D. capitatum populations in cluster analysis Group 2 (center) are decurrent at the base, which distinguishes them from populations in Group 1 (left). Scale bar = 5 mm. 2014] PRESTON: DICHELOSTEMMA LACUN A~ VERNA LIS KEWISITEXA 357 Table 4. Character Matrix for Differentiating between Dichelostemma capitate m and D. lacuna- VERNALIS. Character Dichelostemma capitatum Dichelostemma lacuna-vernalis Percent of samples correctly classified Scape height >20 cm <20 cm 78.80% Flower number >2 <2 83.50% Perianth tube length Ratio of tube length to >4.1 mm <4.1 mm 93.20% lobe length Tube length < 2X lobe length Tube length > 2X lobe length 85.90% Outer perianth lobe, width Relative width of inner and <6.2 mm >6.2 mm 82.50% outer perianth lobes Inner > outer Inner < outer 79.00% Appendage length >5.2 mm <5.2 mm 79.70% Style length >4.5 mm <4.5 mm 80.50% populations and in D. lacuna-vernalis populations have ovate ovaries (vs. urn-shaped in Group 1), perianth lobes that spread from the tubes at different levels (vs. from the same level in Group 1), and outer perianth lobes that are decurrent down the perianth tube, below the base of the inner perianth lobes, for one to three mm (vs. not or decurrent less than one mm in Group 1). In other respects, such as scape length, leaf width, and flowers per scape. Group 2a populations of D. capitatum are intermediate between Group 1 populations and D. lacuna-vernalis. In Group 1 populations of D. capitatum, the inner perianth lobes are wider than the outer perianth lobes, but in Group 2a populations of D. capitatum, the inner and outer perianth lobes are the same width, on average. Common Garden Plants Plants grown in a common garden from corms collected from populations of D. capitatum and D. lacuna-vernalis maintained their distinctive morphology (Fig. 4). As in the field-collected samples, pot-grown individuals from populations of D. lacuna-vernalis were shorter, produced fewer flowers per scape, had shorter perianth tubes, and had broader perianth lobes than pot- grown individuals from populations of D. capi- tatum. Moreover, pot-grown individuals from populations of Group 2a populations of D. capitatum maintained their differences and simi- larities to Group 1 populations of D. capitatum and D. lacuna-vernalis. Several other differences between D. capitatum and D. lacuna-vernalis were evident in pot-grown plants that were not readily apparent in the field. Dichelostemma lacuna-vernalis corms produced few cormlets, with only 44.8% of the corms producing cormlets (1.2 cormlets/corm, range 1- 3). In contrast, 60.9% of D. capitatum corms produced cormlets (2.3 cormlets/corm, range 1- 6). Dichelostemma lacuna-vernalis also is capable of producing scapes from smaller corms than D. capitatum. In D. lacuna-vernalis, 46.4% of corms 8-1 1 mm in diameter produced flowering scapes. and 85.6% of corms 11-14 mm in diameter produced flowering scapes. In D. capitatum, only 15.5% of corms 8-11 mm in diameter produced flowering scapes, and only 59.3% of corms 11- 14 mm in diameter produced flowering scapes. Moreover, D. lacuna-vernalis averaged 2.7 scapes per corm, and larger corms produced up to 8 scapes. Dichelostemma capitatum usually pro- duced a single scape and rarely produced more than two scapes per corm (1.7 scapes per corm, range 1-6). Discussion Dichelostemma capitatum is a broadly circum- scribed species that consists of multiple cytotypes, but there has been little support from botanists for the recognition of infraspecific taxa, despite being widely distributed in the southwestern U.S. and northern Mexico, occurring in a broad range of habitats, and exhibiting a high degree of morphological variation (Keator 1968, 1992). Keator (1968) sampled extensively among popu- lations of D. capitatum throughout much of its range to determine whether combinations of morphological characters occurred together con- sistently enough to warrant formal recognition of segregate taxa. Although he found that that chromosome number differed substantially among populations {n = 9, 18, 27, and 36), he also found that the cytotypes generally were not morphologically distinguishable, except that dip- loid cytotypes sometimes had smaller flowers. Keator concluded that there were no character complexes by which to distinguish between cytotypes nor by which geographic races could be differentiated, further stating that “[sjuch characters as flower color, bract shape and color, and pedicel length are variable within popula- tions to such an extent that their use for taxonomic or correlative purposes seems impos- sible (Keator 1968, p. 376).” Keator (1992) characterized the inability to utilize morpholog- ical data to elucidate ecological or evolutionary relationships within D. capitatum as “problemat- ical.” However, Keator’s analysis was based on a 358 MADRONO [Vol. 61 Fig. 4. Box plots comparing morphological characters measured in Dichelostemma capitatum and D. lacuna- vernalis. A) scape height in cm; B) flowers per scape; C) perianth tube length in mm; D) outer perianth lobe width in mm; E) length of decurrent base of outer perianth lobe in mm; F) ratio of outer lobe width to inner lobe width; G) style length in mm; H) ovules per ovary. The boxes represent the second and third quartiles, with the central horizontal lines representing the median; the upper and lower whiskers represent the 95th and 5th percentiles, respectively; outliers are not shown. Cl-f = D. capitatum Group 1 cluster analysis populations, field collected individuals; Cl-g = D. capitatum Group 1 populations, garden-grown individuals; C2-f = D. capitatum Group 2 populations, field collected individuals; C2-g = D. capitatum Group 2 populations, garden-grown individuals; L-f = D. lacumi-vernalis, field collected individuals; L-g = D. lacima-vernalis, garden-grown individuals. limited set of floral characters and did not employ a multivariate statistical approach. Keator (1968, 1991), like Hoover (1940) before him, attributed morphological variation in D. capitatum to environmental plasticity. Both authors rejected Brodiaea insular is Greene {= Di- chelostemma insulare [Greene] Burnham), a taxon based on robust populations from the Channel Islands, citing the common occurrence of robust individuals in many mainland populations, when 2014] PRESTON: DICHELOSTEMMA LACUNA~VERNAL1SKE\\^\TEE> 359 observed growing under favorable conditions. Both authors cautiously recognized D. capitatum var. pauciflorum (Torr.) Hoover { — D. capitatum subsp. pauciflorum [Torr.] Keator) from desert areas of the southwestern U.S. and northern Mexico, noting the presence of populations morphologically intermediate between var. pauci- Jlorum and “typical” D. capitatum. Keator’s caution in recognizing other taxa within D. capitatum is consistent with the tradition among botanists that morphologically similar polyploid cytotypes are rarely named and considered as species separate from their diploid progenitors, primarily because of the practical aspects of differentiating between them in the field or herbarium (Judd et al. 2007; Soltis et al. 2007). The results of the present study demonstrate that, contrary to Keator’s conclusion, the range of morphological variation in populations as- signable to D. lacuna-vernalis does not overlap continuously with that of D. capitatum. Morpho- logical variation is not continuous in D. capita- tum, and variation within populations is much less than among populations. In each of the populations sampled for this study, individuals expressed a discrete range of morphological variation, not the full spectrum of possible phenotypes. Some groups of populations are morphologically more similar to each other than to other groups, i.e., character complexes exist by which groups of populations can be differentiat- ed. Moreover, plants referable to D. lacuna- vernalis are not simply on the low end of the normal range of size variation for D. capitatum. Scape height, flowers per scape, perianth tube length, and style length in D. lacuna-vernalis are clearly outside the normal range for D. capitatum (Fig. 4a, b, c, g). Perianth tube length is rarely more than 4 mm in D. lacuna-vernalis and very rarely less than 4 mm in D. capitatum (Figs. 3, 4c). Also, not all characters are smaller in D. lacuna-vernalis’, the outer perianth lobes are generally wider, not narrower, than those of D. capitatum (Fig. 4d), and the outer perianth lobes are wider than the inner perianth lobes (Fig. 4f). In D. capitatum, the perianth lobes are of equal width or the inner lobes are wider than the outer. In contrast to D. lacuna-vernalis, populations of D. capitatum with short perianths (i.e., popula- tions C37-C43) have longer perianth tubes than D. lacuna-vernalis but relatively shorter perianth lobes. Keators’s second objection to recognizing lacuna-vernalis, his belief that the smaller size of many characters in D. lacuna-vernalis is simply a plastic response to growing in waterlogged, clay soils, is not supported by the results of the common garden study. Dichelostemma lacuna- vernalis individuals maintain their small stature, few flowers, and other diagnostic characteristics, and D. capitatum individuals maintain their larger stature and many flowers, when grown under identical conditions. Moreover, if the distinctive morphology of D. lacuna-vernalis was simply due to environmental conditions, then populations of D. capitatum growing in hetero- geneous environments should contain a mixture of plants with both morphologies, and individu- als with characteristics of D. lacuna-vernalis could be expected to occur in any part of the range for D. capitatum. Instead, D. lacuna-vernalis plants occur in discrete populations within a well- defined geographic distribution along the western base of the Sierra Nevada foothills and adjacent Great Valley, in a narrow elevation band between 30 and 270 m (Fig. 5). That some of the variation observed in Dichelostemma populations may be environmen- tally induced is not in dispute. In other species in Brodiaeoideae, several characters, such as scape height and the number and size of cormlets produced, are influenced by moisture availability, temperature, or plant density (Niehaus 1971; Han et al. 1991; Cocozza et al. 2000). Corm size, which is a function of age and other factors, such as the presence of mycorrhizal fungi (Scagel 2004), also has an effect on the ability to produce flowering scapes, scape size, and the number of flowers (Han et al. 1991; Schlising and Chamberlain 2006). Corm size also appears to have an effect on these characters in Dichelostemma capitatum (unpublished observations). Leaf width appears to be positively correlated with comi size in both D. capitatum and D. lacuna-vernalis, and the observation of smaller leaf widths in D. lacuna- vernalis (Table 1) may be due, at least in part, by the ability of D. lacuna-vernalis to flower from smaller corms. The influence of conn size on reproductive traits in D. capitatum and D. lacuna- vernalis is currently under investigation and will be the subject of a forthcoming paper. The results of the morphological analysis and the common garden study both support Lenz’ (1974) proposal that as D. lacuna-vernalis merits taxonomic recognition. As the DA showed, D. lacuna-vernalis plants can be distinguished from D. capitatum plants with a high degree of reliability, and the populations can be easily recognized in the field by characteristics other than the small stature and few-flowered scapes (Table 4). Less clear, however, is whether it should be recognized at species rank or at an infraspecific rank. Taxonomic circumscriptions within Brodiaeoi- deae traditionally have been grounded on the morphological species concept, with species distinguished on the basis of discrete differences in the shape of the floral parts and with infraspecific taxa delineated on the basis of size differences or the relative position of floral parts treated (Hoover 1939, 1940, 1941; Preston 2010). Dichelostemma lacuna-vernalis is distinguished 360 MADRONO [Vol. 61 120° W Fig. 5. Distribution of Dichelostemma capitatum subsp. lacuna-vernalis in California, USA. from D. capitatum primarily on the basis of size differences (scape height, flower number, and the size of the floral parts), which suggests treating D. lacuna-vernalis as an infraspecific taxon. Evidence for treating D. lacuna-vernalis at species rank, i.e., reproductive barriers between D. lacuna-vernalis and D. capitatum as a consequence of genetic or ecological factors, or both, is currently ambigu- ous. Although D. capitatum is known to consist of multiple cytotypes (Keator 1968), no chromo- some counts for D. lacuna-vernalis have been documented, and no hybridization studies have been done. Under current practice, infraspecific taxa in Brodiaeoideae are recognized as subspe- cies (Niehaus 1971; Keator 1991; Fires 2002a, b; Fires and Keator 2012, Fires and Freston 2012); therefore, D. lacuna-vernalis appears to be best treated at subspecies rank. Taylor (2010) inde- pendently came to the same conclusion and proposed the new combination, Dichelostemma capitatum subsp. lacuna-vernalis (L.W. Lenz) D.W. Taylor, The morphometric analysis unexpectedly found that populations of D. capitatum sampled for this study also form two distinct groups, one of which shared some characteristics with D. lacuna-vernalis and shared other characteristics with “typical” D. capitatum, with some charac- teristics appearing intermediate (Figs. 1-4). The 22 populations of D. capitatum in Group 2a (C22-C43) are recognizable statistically (cluster analysis, FCA) as well as by distinct morpholog- ical characteristics, such as the decurrent bases of the outer perianth lobes (Fig. 3). The populations appear to have a discrete geographic distribution, although this has not been fully investigated. Most of the populations (C22-C36) occur along the western base of the Sierra Nevada foothills parapatrically with D. lacuna-vernalis (sympatric in at least one location) or at higher elevations than D. lacuna-vernalis. Other populations in Group 2a (C37-C43) occur in the interior North Coast Ranges outside of the range of D. lacuna- vernalis. The interior North Coast Range popu- lations are similar to the Sierra Nevada foothill populations in most characteristics but have smaller flowers. The results of this study indicate that other population groups in D. capitatum can be distinguished morphologically and geographical- ly and may merit taxonomic consideration. However, a number of questions remain to be addressed before formal taxonomic recognition of these groups can be proposed. What is the geographic extent of the Group 2a populations? Because the distinctiveness of these populations was not recognized before this study, the full distribution of these populations has not been determined or sampled, as has been the case for D. lacuna-vernalis. Keator (1968) noted that D. 2014] PRESTON: DICHELOSTEMMA LACUNA-VERNALIS REVISITED 361 capitatum populations in southern California mountains have small flowers, and these popu- lations also need to be evaluated. Do these other morphologically recognizable groups correspond to different cytotypes? Keator’s (1968) finding that diploids have smaller flowers than poly- ploids suggests that this may be the case, but additional cytological studies are needed to evaluate that hypothesis. Studies utilizing DNA markers are needed to test the validity of the morphologically recognizable groups and may be needed to determine whether the polyploid cytotypes represent unique lineages or multiple lineages. Dichelostemma capitatum is a common and familiar member of the California flora, but such familiarity appears to have fostered an assumption that the species has been well- characterized. Taxonomic Treatment Dichelostemma capitatum (Benth.) Alph. Wood subsp. lacuna-vernalis (L.W. Lenz) D.W. Tay- lor, FI. Yosemite Sierra 373. 2010. Dichelos- temma lacuna-vernalis L.W. Lenz, Aliso 8: 129. 1974.— TYPE. USA, California, Sacramento Co., Orangevale, 12 Apr 1967, L.W. Lenz 24671a (holotype: RSA235779 [digital image!]; isotypes: RSA235800, RSA457167, RSA457168, RSA457169, RSA457170, RSA457171 [digital images!]). Because the original description was based solely on the type specimen, the description for D. capitatum subsp. lacuna-vernalis is emended here to incorporate data obtained from populations sampled across the range of the subspecies. Perennial herb from a corm; corms up to 25 mm in diam, not deeply seated, sometimes bearing 1-2 offsets. Leaves 2, subulate, thin, flat to concave, keel-less, ca. 2 dm long, 2. 2-8. 8 (-12.5) mm broad at base, margins ciliate. Inflorescence scapose, umbellate, l-3(-5) flow- ered; scape l^(-6) per corm, slender, (4.4-) 6.2- 22.6 (-26.1) cm long; bracts ca. 6 mm wide, (4.8-) 6.2-13.5 (-16.0) mm long, ovate, acuminate, purple; pedicels < 4.4 (<6.0) mm long. Flowers blue-vioiet; perianth (9.8-) 10.7-16.6 (-17.8) mm long; tube campanulate, (2-) 2.4-4. 8 (-6.0) mm long; outer lobes ovate, cordate at base, (7.5-) 9.0-13.8 (-14.6) mm long, (4.0-) 4.9-9.3 (-9.8) mm wide; inner lobes oblong, (6.3-) 7.6-12.3 (-13.3) mm long, (3.2-) 4. 0-7. 9 (-9.0) mm wide; appendages (3-) 3. 7-6.0 (-6.7) mm long; outer stamens (2.8-) 3. 2-5. 4 (-5.6) mm long, filaments (1.5-) 1.6-3. 5 (-3.8) mm long, anthers (1.2-) 1.5- 3.5 (-3.0) mm long; inner stamens (2.9-) 3. 0-4.9 (-5.2) mm long, filaments (0.5-) 0.7-2. 5 (-2.8) mm long, anthers (1.7-) 2.4-4. 3 (^.5) mm long; ovary ovoid, (2.0-) 2. 5-4.5 (-5.0) mm long; style (2.5-) 2. 8-5.0 (-6.5) mm long; ovules (7-) 8-17 (-20) per locule. Fruit a loculicidal capsule, ovoid, ca. 8.5 mm long, 5.5 mm wide, valve apex acute. Seeds black, ovoid to rhomboid, finely striate, 1-1.5 mm long. Phenology Like many other geophytes in the Brodiaeoi- deae (Niehaus 1971; Han et al. 1994; Schlising and Chamberlain 2006; Kannely and Schlising 2014), D. capitatum subsp. lacuna-vernalis forms corms that are dormant in the soil during the summer drought. New leaves emerge soon after the start of the rainy season, generally in October or November. The plants spend the next three to four months producing a new main corm. Blooming in the field occurs from late February to early April, generally two to three weeks sooner than sympatric populations of D. capita- tum subsp. capitatum. However, plants grown in pots in Davis, California, bloomed as early as the first week of January. Seed set follows soon after, and all aboveground parts wither and dry during the summer dormant period. Because the corms produce few offsets, and the subspecies appears to reproduce primarily by seed. Distribution and Ecology Dichelostemma capitatum subsp. lacuna-verna- lis is endemic to the western base of the Sierra Nevada foothills and adjacent Great Valley, ranging from Butte County south to Merced County (Fig. 5). The populations are restricted to a narrow elevation band between 30 and 270 m, which corresponds to the zone of annual precipitation between the 500 and 750 mm isohyets (National Weather Service 2013). Based on soil information obtained for each population from the National Resource Conservation Ser- vice’s Web Soil Survey (http://websoilsurvey.nrcs. usda.gov/), soils in which the populations occur are loamy, usually sandy loams, gravelly loams, or stony loams, most of which are alfisols (Redding, Red Bluff series), enceptisols (Exche- quer series), and ultisols (Mokelumne series). These soils formed in alluvium from mixed sources and are shallow, having a duripan, bedrock, or both present within 0.25 to 1.5 m of the soil surface. The epithet “lacuna-vernalis” refers to the undulating vernal pooLswale terrain in which the species often occurs (Lenz 1974); however, the populations do not grown in vernal pools but in open upland grasslands adjacent to vernal pools, often on mounds, or in grassy swales in oak woodland. Associated species are early spring blooming annuals, including Amsinckia menziesii (Lehm.) Nelson & J.F. Macbr., Minuartia cali- fornica (A. Gray) Mattf., Crassula connata (Ruiz- Lopez & Pavon), Dichelostemma capitatum 362 MADRONO [Vol. 61 subsp. capitatum, Dichelostemma multiflorum (Benth.) A. A. Heller, Erodium botrys (Cav.) BertoL, Eschschohia lobbii Greene, Hypochaeris glabra L., Lasthenia gracilis (DC.) Greene, Layia fremoutii (Torr. & A. Gray) A. Gray, Lepidium nitiduai Torr. & A. Gray, Lomatiurn spp., Lupinus bicolor Lindley, Minuartia californica (A. Gray) Mattf., Plagiobothrys spp., Pkmtago erect a E. Morris, Senecio vulgaris L., Thysanocarpus radi- ans Benth., Trifolium depauperatum Desv., Tri- physaria erkmtha (Benth.) Chuang & Heckard, and Triteleia hyacinthina (Lindley) Greene. Conservation Status Shortly after it was described, D. capitatum subsp. lacuna-vernalis was included in the Cali™ fornia Native Plant Society (CNPS) Inventory of Rare and Endangered Plants of California as “rare and endangered” (Smith et al. 1980). It was reviewed by the U.S. Fish and Wildlife Service as a candidate for listing under the Endangered Species Act, but after the discovery of additional occurrences in the early 1980’s, it was determined not to be threatened or endangered, and subse- quent editions of the CNPS Inventory listed it as “rare, but not endangered.” It was dropped entirely from the CNPS Inventory after being synonymized with D. capitatum in the first edition of The Jepson Manual (Keator 1993). Based on current herbarium records, D. capitatum subsp. lacuna-vernalis is known from only 40 occurrences. Because it has a relatively broad range (285 km) that spans most of the length of the Sacramento Valley and the north end of the San Joaquin Valley, and because the plants appear to be locally common where present, it should not currently be considered rare. However, because it occurs within a narrow elevation band, it has a very limited distribution. Moreover, it occurs within an area that is experiencing substantial population growth, which makes it vulnerable to future habitat loss, and at least three of the known occurrences are located in urbanized areas and appear to have been extirpated (personal observation). There- fore, D. capitatum subsp. lacuna-vernalis should be considered a “watchlist” plant (California Rare Plant Rank 4) by the California Natural Diversity Database. Acknowledgments I am grateful to Craig Martz (California Depart- ment of Fish and Game) for sharing his field notes and unpublished reports from his files. Without his earlier efforts of to locate and document populations of D. lacima-vernalis, this study would have been much more difficult to carry out. I also thank three anonymous reviewers for constructive criticism of the manuscript. Literature Cited Baker, J. G. 1871. A revision of the genera and species of herbaceous capsular gamophyllous Liliaceae. Journal of the Linnean Society, Botany 1 1:349^36. ' Berg, R. Y. 1978. Development of ovule, embryo sac, i and endosperm in Brodiaea (Liliales). Norwegian Journal of Botany 25:1-7. . 1996. Development of ovule, embryo sac, and endosperm in Dipterostemon and Dichelostemma (Alliaceae) relative to taxonomy. American Journal of Botany 83:790-801. . 2003. Development of ovule, embryo sac, and endosperm in Triteleia (Themidaceae) relative to taxonomy. American Journal of Botany 90:937- 948. Chase, M. W., J. L. Reveal, and M. F. Fay. 2009. A subfamilial classification for the expanded aspar- agalean families Amaryllidaceae, Asparagaceae and Xanthorrhoeaceae. Botanical Journal of the Linnean Society 161:132-136. CocozzA, T. M. A., G. Cristiano, B. de Lucia, and E. Maloupa. 2000. Influence of some cultural techniques on corm enlargement in Brodiaea. Acta Horticulturae 541:277-280. Fay, M. F. and M. W. Chase. 1996. Resurrection of Themidaceae for the Brodiaea alliance, and recir- cumscription of the Alliaceae, Amaryllidaceae, and Agapanthoideae. Taxon 45:441^51. , P. J. Rudall, S. Sullivan, K. L. Stobart, A. Y. De Bruun, F. Qamaruz-Zaman, W. P. Hong, J. Joseph, W. J. Hahn, J. G. Conran, AND M. W. Chase. 2000. Phylogenetic studies of Asparagales based on four plastid DNA loci. Pp. 360-371 in K. L. Wilson and D. A. Morrison (eds.). Monocots: systematics and evolution. CSIRO, Collingwood, Victoria, Australia. Han, S. S. 1994. Morphology of flower initiation of brodiaea. Scientia Horticulturae 56:235-243. , A. H. Halevy, R. M. Sachs, and M. S. Reid. 1991. Flowering and corm yield of brodiaea in response to temperature, photoperiod, corm size, and planting depth. Journal of the American Horticultural Society 116:19-22. Hoover, R. F. 1939. A revision of the genus Brodiaea. American Midland Naturalist 22:551-574. . 1940. The genus Dichelostemma. American Midland Naturalist 24:463^76. . 1941. A systematic study of Triteleia. American Midland Naturalist 25:73-100. Judd, W. S., D. E. Soltis, P. S. Soltis, and G. Ionta. 2007. Tolmiea diplomenziesii: a new species from the Pacific Northwest and the diploid sister taxon of the autotetraploid T. menziesii (Saxifra- gaceae). Brittonia 59:217-225. Kannely, a. and R. a. Schlising. 2014. Biology of the geophyte, Triteleia ixioides subsp. anilina (Themidaceae), in coniferous forests of Butte County, California. Madrono 61:87-104. Keator, G. 1968. A taxonomic and ecological study of the genus Dichelostemma (Amaryllidaceae). Ph.D. dissertation. University of California, Berkeley, CA. . 1991. Studies in the genus Dichelostemma. Four Seasons 9(l):24-39. . 1992. Blue dicks brodiaea [Dichelostemma capitatum): a common but problematical species. Four Seasons 9(2):31-42. 2014] PRESTON: DICHELOSTEMMA LACUNA-VERNALIS REVISITED 363 . 1993. Dichelostemma. Pp. 1190-1192 in J, C. Hickman (ed.), The Jepsoii manual: higher plants of California. University of California Press, Berkeley, CA. Lenz, L. W. 1974. A new species of Dichelostemma (Liliaceae) from California. Aliso 8:129-131. . 1976. The nature of the floral appendages in four species of Dichelostemma (Liliaceae). Aliso 8:383-389. National Weather Service. 2013. Average annual Northern California precipitation, 1961-1990. Website http://www.wrh.noaa.gov/sto/norcalpcpn. php (accessed 09 October 2013). Niehaus, T. F. 1971. A biosystematic study of the genus Brodiaea (Amaryllidaceae). University of California Publications in Botany 60:1-67. Nguyen, N. H., H. E. Driscoll, and C. D. Specht. 2008. A molecular phylogeny of the wild onions {Allium; Alliaceae) with a focus on the western North American center of diversity. Molecular and Phylogenetic Evolution 47:1157-1172. PiRES, J. C. 2002a. Dichelostemma. Pp. 328-331 in Flora of North America Editorial Committee (eds.). Flora of North America North of Mexico, Vol. 26: Magnoliophyta: Liliidae: Liliales and Orchidales. Oxford University Press, New York, NY. . 2002b. Brodiaea. Pp. 321-328 in Flora of North America Editorial Committee (eds.). Flora of North America North of Mexico Vol. 26: Magno- liophyta: Liliidae: Liliales and Orchidales. Oxford University Press, New York, NY. AND G. Keator. 2012. Dichelostemma. Pp. 1511-1512 in B. G. Baldwin, D. H. Goldman, D. J. Keil, R. Patterson, T. J. Rosatti, and D. H. Wilken (eds.). The Jepson manual: vascular plants of California, 2nd ed. University of California Press, Berkeley, CA. AND K. J. Sytsma. 2002. A phylogenetic evaluation of a biosystematic framework: Brodiaea and related petaloid monocots (Themidaceae). American Journal of Botany 89:1342-1359. AND R. E. Preston. 2012. Brodiaea. Pp. 1508- 1511 in B. G. Baldwin, D. H. Goldman, D. J. Keil, R. Patterson, T. J. Rosatti, and D. H. Wilken (eds.). The Jepson Manual: vascular plants of California, 2nd ed. University of California Press, Berkeley, CA. , M. F. Fay, W. S. Davis, L. Hufford, J. Rova, M. W. Chase, and K. J. Sytsma. 2001. Molecular and morphological phylogenetic analy- ses of Themidaceae (Asparagales). Kew Bulletin 56:601-626. Preston, R. E. 2010. Brodiaea matsonii (Asparagaceae: Brodiaeoideae) a new species from Shasta County, California. Madrono 57:261-267. Rydberg, P. A. 1912. Studies on the Rocky Mountain flora — XXVI. Bulletin of the Torrey Botanical Club 39:99-111. SCAGEL, C. F. 2004. Soil pasteurization and mycorrhi- zal inoculation alter flower production and corm composition of Brodiaea iaxa ‘Queen Fabiola’. HortScience 39:1432-1437. SCHLISING, R. A. AND S. A. CHAMBERLAIN. 2006. Biology of the geophytic lily, Triteleia laxa (Themidaceae), in grasslands of the northern Sacramento Valley. Madrono 53:321-341. Smith, J. P., Jr., R. J. Cole, and J. O. Sawyer, Jr. (eds.). 1980. Inventory of rare and endangered vascular plants of California, 2nd ed. California Native Plant Society, Berkeley, CA. Soltis, D. E., P. S. Soltis, D. W. Schemske, J. F. Hancock, J. N. Thompson, B. C. Husband, and W. S. Judd. 2007. Autopolyploidy in angiosperms: have we grossly underestimated the number of species? Taxon 56:13-30. Steele, P. R., K. L. Hertweck, D. Mayfield, M. R. McKain, j. Leebens-Mack, and J. C. Pires. 2012. Quality and quantity of data recovered from massively parallel sequencing: examples in Aspar- agales and Poaceae. American Journal of Botany 99:330-348. Taylor, D. W. 2010. Flora of the Yosemite Sierra: being a transect flora of the central Sierra Nevada, including all of Tuolumne, Mariposa and Madera Counties, the Mono Basin, and adjacent areas of Mono County. Published by the author, Aptos, CA. Appendix 1 Voucher Specimens for Populations Sampled FOR Morphological Measurements. Characters Means are Provided in Appendix 2. Dichelostemma capitatum. COl: Butte Co., Upper Bidwell Park, Chico, on the north rim at the level of Horseshoe Lake, 09 Feb 1983, Oswcdd 63 (CHSC). C02: Placer Co., 4 mi E of Lincoln, 16 Mar 2013, Preston 2891 (DAV). C03: Sacramento Co., Fair Oaks, Phoenix Park, 14 Mar 2009, Preston 2653 (DAV). C04: Sacramento Co., along Scott Rd, ca. 0.5 mi N of its jet with Latrobe Rd, 03 Apr 2010, Preston 2738 (DAV). COS: Sacramento Co., along Scott Rd, 1.6 mi S of its jet with White Rock Rd, 03 Apr 2010, Preston 2736 (DAV). C06: Sacramento/Amador Co., ca. 5.5 mi W of lone, along Hwy 104, 24 Mar 2013, Preston 2899 (DAV). C07: Calaveras Co., 1.7 mi NNE of Burson, along Chile Camp Rd, at jet with S Camanche Parkway, 08 Apr 2007, Preston 2424 (DAV). COS: San Joaquin Co., 5 mi NE of Bellota, along Hwy 26, 15 Apr 2007, Preston 2426 (DAV). C09: Calaveras Co., 5 mi NW of Copperopoiis, along Salt Springs Valley Rd, 0.15 mi S of jet with Rock Creek Rd, 15 Apr 2007, Preston 2429 (DAV). CIO: Napa Co., along Hwy 128, 0.4 mi W of Knoxville Rd, 14 Apr 2013, Preston 2904 (DAV). Cll: Colusa Co.; along CA-20, 9.7 miles southwest of Williams, on the north side of Salt Creek, 1 1 Mar 2009, Helmkamp and Helmkamp 14366 (UCR). Cl 2: Solano Co., ca. 4 mi NE of Benicia, along Lopes Rd, 0.7 mi N of its jet with Parish Rd, 22 Mar 2013, Preston 2897 (DAV). C13: Alameda Co., Mission Hills, 0.4 mi SW of Ohlone College, 22 Mar 2013, Preston 2896 (DAV). C14: Santa Clara Co., Anderson Lake County Park, 22 Mar 2013, Preston 2895 (DAV). Cl 5: San Mateo Co., Hillsborough, along Crystal Springs Rd where it crosses under the Junipero Serra Freeway, 28 Mar 2013, Preston 2900 (DAV). Cl 6: Stanislaus Co., mouth of Arroyo Del Puerto, 28 Mar 1935, Sharsmith 1532 (UC). C17: Fresno Co., along Panoche Rd, 2.5 mi W of Interstate 5, 24 Mar 2009 [voucher misplaced, to be recollected]. Cl 8: Fresno Co., 9.5 mi NE of Coalinga, at Skunk Hollow, 24 Mar 2008 [voucher misplaced, to be recollected]. Cl 9: Fresno Co., Clovis, at jet Herndon Ave and Academy Ave, 01 Apr 2013, Preston 2901 (DAV). C20: Fresno Co., Friant, along 364 MADRONO [Vol. 61 Millerton Rd, across from entrance to Friant Dam, 01 Apr 2013, Preston 2902 (DAV). C21: Kern Co., Tehachapi Mtns, ca. 9.3 mi ENE of Lebec, in Bear Trap Canyon, 15 May 2007, Preston 2515 (DAV). C22: Butte Co., top of North Table Mt., W edge of Mountain adjacent to the waterfall at the S branch of Coal Creek, 23 Feb 1979, Joker st 1134 (CHSC). C23: Butte Co., Table Mountain, 0.2 mi SE of Cherokee Rd, 16 Mar 2013, Preston 2894 (DAV). C24: Butte Co., 7.9 mi NNE of Oroville, along E side of Clark Rd, 16 Mar 2013, Preston 2893 (DAV). C25: Yuba Co., along both sides of Hammonton-Smartville Rd, 1.0 mi E of N entrance to Beale AFB, 25 Mar 1982, Martz & Sanner 42 (DAV). C26; Sacramento Co., E of Rancho Cordova, along Scott Rd, 1 .2 mi S of its jet with White Rock Rd, 24 Mar 2013, Preston 2898 (DAV). C27: Sacramento Co., along Scott Rd, 0.95 mi S of Deer Creek, 03 Apr 2010, Preston 2737 (DAV). C28: Sacramento Co., along Hwy 16, at jet with lone Rd, 22 Mar 2014, Preston 2949 (DAV). C29: Sacramento Co., along lone Rd, 2.35 mi S of Flwy 16, 03 Apr 2010, Preston 2739 (DAV). C30: Amador Co., 2.5 mi NW of lone, along Irish Hill Rd, 0.56 mi N of Hwy 104, 08 Apr 2007, Preston 2422 (DAV). C31: Amador Co., 1.9 mi SSE of lone, along Buena Vista Rd, 0.1 mi S of jet with Hwy 88, 08 Apr 2007, Preston 2423 (DAV). C32: Sacramento Co., Rancho Seco County Park, 06 Apr 2014, Preston 2950 (DAV); C33: Calaveras Co., 1.4 mi NNE of Burson, along Chile Camp Rd, 0.7 mi E of Burson Rd, 17 Apr 2010, Preston 2763 (DAV). C34: San Joaquin Co., 3.8 mi E of Clements, on W side of Cord Rd, 0.9 mi S of jet with Hwy 12, 15 Mar 2008, Preston 2587 (DAV). C35: Calaveras Co., 0.75 mi SE of Wallace, on S side of Hwy 12, 22 Mar 2014, Preston 2948 (DAV). C36: Stanislaus Co., north side of Tuolumne River, opposite La Grange, 22 Feb 1941, Hoover 4773 (UC). C37: Lake Co., SE edge of Manning Flat, 03 May 2011, Preston 2856 (DAV). C38: Lake Co., Shaul Valley, in field N of SR 29, 04 May 2011, Preston 2857 (DAV). C39: Napa Co., Lake Hennessy, along Conn Valley Rd at public access site, 16 Apr 2013, Preston 2907 (DAV). C40: Sonoma Co., 0.6 mi E of Occidental, at jet Occidental Rd and Facendini Lane, 29 Mar 2013, Preston 2901 (DAV). C41: Napa Co., Calistoga, E end of old landing strip, on N side, 16 Apr 203, Preston 2905 (DAV). C42: Lake Co., 4.5 mi NW of Middletown, on W side of Hwy 175, 16 Apr 2013, Preston 2908 (DAV). C43: Lake Co., Kelseyville, at SW comer of jet Main St and Douglas Rd, 16 Apr 2013, Preston 2909 (DAV). Dichelostemma lacuna-vevnalis . LOl: Butte Co., E side of State Route 191 at intersection of Pentz Rd, 22 Mar 1982, Martz 30 (DAV). L02: Butte Co., ca. 3.7 mi NW of Oroville, on NW side of PG&E substation, 23 Mar 2007, Preston 2418 (DAV). L03: Butte Co., along State Route 70, 1 mi N of intersection with Palermo Rd, S of Oroville, 16 Mar 1982, Martz 31 (DAV). L04: Yuba Co., along Hammonton-Smartville Rd, 2.4 mi E of Doolittle Drive, 16 Mar 2013, Preston 2892 (DAV). L05: Placer Co., both sides of Dowd Rd, 3 mi S of Sheridan, 24 Mar 1982, Martz 41 (DAV). L06: Yuba Co., along both sides of Wheatland-Smartville Rd, 3.7 mi NE of Wheatland, 26 Mar 1982, Sanner 33 (DAV). L07: Placer Co., west side of Sierra College Blvd, 1.1 miles S of English Colony Rd intersection, 26 Mar 1982, Martz 38 (DAV). L08: Placer Co., on Sierra College Blvd, 2.3 mi N of Douglas Blvd intersection, 07 Mar 1982, Martz 13 (DAV). L09: Sacramento Co., Fair Oaks, Phoenix Park, 31 Mar 2007, Preston 2419 (DAV). LIO; Sacramento Co., W side of Scott Rd, 1 .2 mi S of intersection with White Rock Rd, 26 Mar 1982, Martz 36 (DAV). El 1; Sacramento Co., E side of Scott Rd, 1.6 mi N of intersection with Latrobe Rd, 26 Mar 1982, Martz 35 (DAV). El 2: Sacramento Co., on E side of lone Rd, 2.3 mi S of jet with Hwy 16, 12 Mar 2008, Preston 2586 (DAV). LI 3: Amador Co., 0.7 mi W of Indian Hill, along S side of Hwy 104, 02 Apr 2007, Preston 2420 (DAV). L14: Calaveras Co., 0.75 mi SE of Wallace, on S side of Hwy 12, 15 Mar 2008, Preston 2588 (DAV). El 5: Calaveras Co., W side of County Rd J14 (Milton Rd), 1.3 mi N of intersection with Hunt Rd, 10 Mar 1982, Martz 21 (DAV). L16: Merced Co., ca. 5 mi NE of Planada, at jet Cunningham Rd and S Bear Creek Drive, 09 Mar 2010, Preston 2721 (DAV). 2014] PRESTON: DICHELOSTEMMA LACUNA-VERNALIS REVISITED 365 Appendix 2 Population Means Means for 19 morphological characters measured in field-collected individuals of Dichelostemma capitatum (43 populations) and D. lacuna-vernalis (16 populations). Population codes and vouchers are provided in Appendix 1. Characters: 1 = scape height; 2 = leaf width; 3 = flowers/scape; 4 = bract length; 5 = pedicel length; 6 = perianth length; 7 = perianth tube length; 8 = outer lobe length; 9 = inner lobe length; 10 = outer lobe width; 1 1 = inner lobe width; 12 = appendage length; 13 = outer filament length; 14 = inner filament length; 15 = outer anther length; 16 = inner anther length; 17 = ovary length; 18 = style length; 19 = ovules/ovary. Measures are in mm, except scape height (cm). Characters Population Taxon code 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Dichelostemma capitatum COl 26.7 8.3 5.4 17.5 5.3 17.5 6.6 12.0 10.9 5.8 6.6 7.2 3.1 2.2 2.8 4.7 4.7 7.5 29.9 C02 31.6 9.9 9.4 18.9 8.5 16.0 6.3 10.1 9.7 5.3 5.8 6.1 2.7 2.0 2.5 4.2 4.6 5.4 38.7 C03 40.7 7.8 5.5 16.0 6.1 16.6 6.1 11.2 10.5 5.3 5.9 6.8 2.7 2.0 2.6 4.2 4.6 6.4 31.8 C04 41.1 7.7 6.8 14.4 8.1 15.8 5.5 10.8 10.3 4.8 5.5 5.6 2.7 1.6 2.2 3.8 4.3 5.5 36.0 COS 35.7 7.0 6.8 14.8 6.7 16.7 6.2 11.0 10.5 4.9 5.6 6.7 2.9 2.0 2.5 4.2 4.3 6.3 30.9 C06 24.6 6.9 5.5 14.2 5.8 14.7 5.4 9.9 9.3 4.9 5.5 6.2 2.4 1.6 2.3 4.1 4.0 4.9 40.2 C07 35.6 6.3 4.2 13.2 5.0 14.6 5.6 9.5 9.0 4.3 4.7 5.8 2.5 1.6 2.2 3.7 3.9 5.2 38.8 C08 39.8 8.1 7.4 14.2 7.0 15.6 5.5 10.4 10.2 4.9 5.7 6.2 2.5 1.6 2.5 4.2 4.5 4.9 42.0 C09 39.3 8.0 9.8 16.0 9.6 15.1 5.5 10.2 9.7 5.3 6.1 6.3 2.8 1.9 2.3 3.9 4.4 5.0 36.4 CIO 28.9 7.7 5.9 14.9 5.8 14.4 5.7 9.2 8.7 5.0 5.7 6.3 2.6 2.0 2.4 3.9 4.1 5.4 27.8 Cll 37.5 7.9 6.7 16.9 6.8 16.9 6.9 10.7 10.0 5.5 6.2 7.3 3.2 2.7 2.6 4.3 4.8 7.2 29.5 C12 34.3 8.6 8.1 18.0 5.9 17.5 6.8 10.9 10.7 6.0 6.6 7.5 3.3 2.5 2.6 4.5 4.8 7.1 33.3 C13 24.5 6.4 5.2 18.5 5.5 17.4 7.1 11.0 10.3 5.5 6.0 6.8 2.9 2.3 2.4 4.1 4.4 7.7 33.1 C14 32.1 9.0 6.9 16.2 8.7 17.4 7.5 10.6 9.8 5.2 5.8 6.9 2.5 2.1 2.2 4.0 4.8 6.9 28.6 C15 29.5 6.5 7.2 12.6 5.3 17.7 8.1 11.6 9.6 4.3 4.7 6.2 2.0 1.3 2.5 4.2 4.3 7.8 19.4 C16 33.9 6.6 5.1 14.0 5.7 16.6 6.8 10.6 9.8 5.4 6.2 6.6 2.5 2.1 2.5 4.1 4.8 6.7 30.5 C17 33.2 8.7 8.9 15.0 8.2 15.8 6.4 10.2 9.4 5.1 5.8 6.3 2.4 1.7 2.5 4.1 4.6 6.3 36.6 C18 36.3 11.3 10.1 13.5 8.7 16.3 7.1 10.0 9.2 4.9 5.4 5.9 2.6 2.0 2.4 3.9 4.5 5.7 34.0 C19 43.3 10.0 7.3 17.0 6.9 16.2 6.8 10.0 9.4 4.9 5.4 6.7 2.9 2.2 2.4 4.4 4.6 6.3 35.3 C20 35.7 9.2 8.2 15.2 8.3 15.6 6.2 10.0 9.4 4.9 5.4 6.2 2.3 1.9 2.2 3.9 4.3 5.5 39.0 C21 - - 9.7 13.2 4.8 13.9 5.8 - 8.2 3.6 4.3 4.8 1.9 1.0 2.2 3.5 3.8 4.8 22.7 C22 18.4 7.3 3.7 14.8 5.7 16.8 5.9 12.7 10.8 5.6 5.8 6.1 2.6 1.6 2.6 4.2 4.1 6.8 25.1 C23 13.4 6.8 3.3 14.3 4.6 16.8 6.4 12.3 10.4 5.5 5.5 5.9 2.5 1.4 2.3 3.7 3.7 6.0 24.6 C24 28.4 7.6 2.9 11.7 3.8 16.9 5.5 13.1 11.4 5.5 5.7 6.5 2.6 1.5 2.7 4.1 3.7 7.9 23.4 C25 15.7 5.5 1.4 10.2 2.4 17.5 5.1 14.5 12.4 6.8 6.3 6.4 2.9 1.8 2.9 4.2 4.0 6.1 30.8 C26 21.8 5.1 2.6 10.9 3.1 16.3 5.2 13.1 11.1 5.9 5.6 6.3 2.6 1.6 2.3 3.7 3.8 6.3 28.1 Cll 18.9 3.7 2.2 9.6 2.5 15.2 5.0 12.4 10.2 5.3 5.0 5.7 2.3 1.2 2.3 3.7 3.3 5.7 28.8 C28 26.6 5.5 2.7 10.5 2.8 16.1 5.4 11.6 10.7 5.6 5.4 6.0 2.4 1.2 2.3 3.6 3.7 6.5 27.2 C29 23.9 4.6 2.1 9.2 2.8 16.1 5.6 12.1 10.5 5.9 5.5 5.7 2.2 1.2 2.3 3.6 3.9 6.0 25.8 C30 28.5 4.9 2.7 9.9 2.7 17.0 5.7 12.9 11.3 5.9 5.6 5.9 2.3 1.3 2.2 3.6 3.6 5.9 27.4 C31 25.4 4.7 3.1 9.7 2.8 15.6 5.7 11.2 9.9 5.2 5.1 5.5 2.1 1.1 2.1 3.5 3.5 6.3 23.8 C32 20.0 5.5 2.8 11.5 3.9 16.1 5.6 12.7 10.5 5.8 5.2 6.0 2.4 1.6 2.8 4.5 3.7 6.4 24.4 C33 25.8 5.3 3.5 9.7 2.6 16.2 5.8 12.2 10.4 4.9 4.9 5.7 2.1 1.2 2.3 3.5 3.4 6.3 26.8 C34 18.0 8.3 3.3 13.5 5.0 16.5 6.3 11.9 10.2 6.0 6.2 6.4 2.6 1.5 2.3 3.9 3.9 6.5 26.4 C35 34.0 10.3 4.4 15.1 5.1 18.2 7.1 13.1 11.1 7.0 6.9 6.7 2.7 1.7 2.5 3.8 4.1 7.6 30.9 C36 16.7 4.6 2.8 10.9 2.7 14.8 5.7 11.3 9.1 5.4 5.3 5.6 2.6 1.4 2.2 3.6 3.5 6.1 26.3 C37 20.5 5.4 3.0 10.6 3.2 14.1 5.3 10.2 8.8 5.1 5.4 5.3 2.2 1.4 2.4 3.7 3.9 5.1 20.9 C38 16.0 5.3 2.1 9.3 3.0 12.4 4.5 9.5 7.8 5.5 5.4 4.4 2.1 1.1 2.0 3.2 3.2 4.1 23.1 C39 24.3 12.7 3.7 11.5 4.1 12.9 4.6 9.6 8.3 5.4 5.8 4.6 2.6 1.5 2.2 3.3 3.9 4.3 32.5 C40 13.0 5.4 2.2 9.9 2.5 13.5 4.0 10.7 9.5 5.9 5.7 4.9 2.4 1.5 2.2 3.6 3.7 4.1 27.7 C41 22.2 7.4 2.9 12.5 4.4 14.6 4.9 11.3 9.8 5.4 5.3 5.2 2.9 1.7 2.3 3.8 4.0 4.8 28.4 C42 22.9 6.0 2.9 11.4 3.2 14.3 5.4 11.1 8.9 5.6 5.4 5.4 2.7 1.7 2.3 3.7 3.5 5.2 24.7 C43 17.2 7.7 3.9 12.1 3.6 14.1 4.9 11.2 9.2 5.8 5.5 5.4 2.4 1.3 2.3 3.5 3.8 4.8 27.8 Dichelostemma lacuna-vernalis LOl 7.5 3.9 1.8 11.0 2.8 12.6 4.1 9.8 8.4 5.4 4.7 4.3 2.2 1.5 2.0 3.1 3.0 3.8 30.5 L02 13.7 5.0 1.7 11.3 2.8 14.4 3.6 12.1 10.8 7.7 6.9 5.1 2.9 2.1 2.2 3.4 3.7 4.3 32.7 L03 16.8 4.6 1.1 11.0 2.0 13.7 3.5 11.7 9.9 6.5 5.6 4.9 2.7 1.8 2.2 3.4 3.3 4.3 32.7 L04 10.0 4.4 1.8 9.7 2.4 12.1 3.4 10.1 8.7 6.2 5.1 4.2 2.1 1.3 2.0 3.1 3.1 3.4 35.8 LOS 12.3 4.9 1.8 9.3 2.9 12.8 3.4 10.7 9.4 6.4 5.3 4.6 2.5 1.6 2.0 3.4 3.3 3.6 36.4 L06 17.3 5.2 1.6 9.9 2.8 14.0 3.3 11.8 10.7 7.0 5.9 4.8 2.8 1.8 2.1 3.3 3.3 3.8 37.1 366 MADRONO [Vol. 61 Appendix 2. Continued. Population Taxon code Characters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 L07 15.5 6.9 2.2 11.8 3.7 15.0 4.1 12.3 11.0 8.1 6.9 5.4 3.1 2.1 2.4 3.7 4.1 4.5 42.7 LOS 16.6 6.7 2.1 10.1 2.9 14.0 3.4 12.1 10.6 7.2 6.2 5.0 2.8 1.7 2.2 3.5 3.5 4.0 39.1 L09 12.2 4.4 1.4 8.4 1.8 13.1 3.0 11.2 10.1 7.2 5.9 4.9 2.6 1.7 2.0 3.2 3.3 3.6 34.8 LIO 13.0 5.1 1.6 9.8 2.4 13.8 3.6 11.7 10.1 7.2 6.1 5.1 2.6 1.8 2.1 3.4 3.3 3.9 41.6 Lll 13.2 4.4 1.4 8.3 2.3 13.2 3.6 11.0 9.6 6.9 5.6 4.8 2.5 1.6 2.1 3.4 3.6 3.7 40.4 L12 16.9 5.7 1.4 9.5 2.3 14.4 3.9 12.1 10.5 7.7 6.2 5.0 2.6 1.5 2.3 3.7 3.6 4.0 44.3 L13 16.3 4.5 1.9 8.3 2.3 13.6 3.5 11.1 10.1 6.4 5.0 4.6 2.7 1.4 2.1 3.2 3.2 3.6 43.8 L14 13.9 6.6 1.6 9.0 2.4 13.7 3.9 11.4 9.9 7.7 6.4 4.9 2.6 1.5 2.0 3.3 3.7 3.8 39.0 L15 17.0 5.0 1.7 10.2 2.9 14.6 4.1 12.0 10.5 7.8 6.3 5.1 3.0 1.7 2.3 3.6 3.8 4.1 42.8 L16 16.3 6.2 1.7 9.4 2.3 12.2 3.0 10.7 9.2 6.6 5.5 4.7 2.4 1.5 2.1 3.3 3.2 3.7 37.8 Madrono, VoL 61, No. 4, pp. 367-387, 2014 VASCULAR FLORA OF DEVILS POSTPILE NATIONAL MONUMENT, MADERA COUNTY, CALIFORNIA Melanie Arnett 419 S. 11th Street, Laramie, WY 82070 Ann M. Huber P.O. Box 64, Three Rivers, CA 93271 Kathren Murrell Stevenson P.O. Box 10941, Oakland, CA 94610 Sylvia Haultain Sequoia and Kings Canyon National Parks, Three Rivers, CA 93271 Sylvia_Haultain@nps.gov Abstract Devils Postpile National Monument is 319 ha in size, located in Madera County, California, on the western slope of the Sierra Nevada at an average elevation of 2280 m. Though small in size, the monument supports diverse habitats, including forests, chaparral, riparian corridors, meadows, seeps, and ponds. In 2001, the National Park Service conducted a survey to 1) inventory the vascular flora and document with vouchered specimens; and 2) describe the distribution and abundance of species of special management concern (rare and/or non-native). Methods coupled a species-level inventory with vegetation mapping, covered an estimated 70 percent of the monument’s surface area, and combined broad and targeted search strategies. Survey results yielded a 121 percent increase (from 169 to 373) in the number of documented plant taxa, representing 60 families and 199 distinct genera. Forty-five percent of species were clustered within six families: Asteraceae, Poaceae, Cyperaceae, Brassicaceae, Onagraceae, and Boraginaceae. The survey found three rare and eight non-native taxa previously unknown from the monument and documented nine potential range extensions. Rare species included Cinna bolanderi (new county record), Hulsea brevifolia, and Mimulus laciniatus. Localized infestations of the non-native and invasive Cirsium vulgare were discovered. Control measures for this species were initiated during the field season and continued in subsequent years. Key Words: collections. Devils Postpile National Monument, flora, inventory, monitoring, vegetation survey. Devils Postpile National Monument is man- aged by the National Park Service for its outstanding geologic features and natural beauty. Located in Madera County, California, high on the western slope of the Sierra Nevada near the town of Mammoth Lakes (Fig. 1), the monument attracts hundreds of thousands of visitors each year (average for years 2000-2010 = 118,381). Primary attractions include the hexagonal basalt columns for which the monument is named and the picturesque 30 m drop of Rainbow Falls on the Middle Fork San Joaquin River. Of the 319 ha (798 acres) of Devils Postpile National Monu- ment, 273 ha (674 acres) are designated as part of the Ansel Adams Wilderness. Both the John Muir and Pacific Crest Trails transect the monument. During the summer of 2001, we completed a floristic inventory of the vascular plants in Devils Postpile National Monument and established vegetation plots as part of a vegetation mapping program for Sierra Nevada parks. Both of these efforts are components of the natural resource inventory phase of the National Park Service (NPS) Inventory and Monitoring Program for the Sierra Nevada Network parks (Devils Post- pile National Monument, Sequoia and Kings Canyon National Parks, and Yosemite National Park). The primary purpose of the NPS natural resource inventories is to document the presence of resources in parks, and to assess and document the current condition and knowledge of these resources. Inventories allow comparison of exist- ing conditions to reference conditions or the desired state of parks and establish a solid baseline for making scientifically sound manage- ment decisions (National Park Service 2009). In keeping with the goals of the NPS Inventory and Monitoring Program (Fancy et al. 2009; National Park Service 2009), the two main objectives of this study were: 1) to document the occurrence of at least 90 percent of the species of vascular plants occurring in the monument with vouch- ered specimens; and 2) to describe the distribu- tion and abundance of species of special man- agement concern, specifically, those designated as rare, threatened, endangered or invasive non- natives. Results from this survey were initially reported in an unpublished report submitted to 368 MADRONO [Vol. 61 Fig. 1. Map of Devils Postpile National Monument and surrounding areas. Devils Postpile National Monument is located in Madera County, California, high on the western slope of the Sierra Nevada near the town of Mammoth Lakes. the Sierra Nevada Inventory and Monitoring Network (Arnett and Haultain 2005). Prior to this survey, available data on the flora of Devils Postpile National Monument were limited in the number of years collected, the geographic extent of areas surveyed, the percent- age of vascular plants documented, and the locations of vouchered specimens. The estimated percentage of vascular plants captured by prior surveys was 82%, a figure based on surveys conducted from 1974-1980, an unpublished vascular plant list (Medeiros 1996), and consul- tation with local experts (National Park Service 2001). The flora was known primarily from collections made by park naturalists in four years: 1972, 1977, 1978, and 1980. K. Ann Hoffmann made 77 collections in 1972 represent- ing 77 taxa in 27 families; Joseph L. Medeiros made 42 collections in the period between 1976 and 1980 representing 42 taxa in 21 families; Sandra C. Morey made 131 collections in 1980 representing 120 taxa in 28 families. The species list produced from these investigators included 235 taxa; however, only 169 of these were documented with herbarium vouchers. The labels of these vouchers, housed in the Devils Postpile National Monument Herbarium (DEPO), reveal that 99 percent or more of the collections were made along the Rainbow Falls Trail and in Soda Springs Meadow, both of which are in the eastern half of the monument in the vicinity of Middle Fork San Joaquin River (Fig. 2). These earlier collection efforts documented two rare and two non-native plant species. The two rare taxa, Lupinus duranii Eastw. and Hulsea brevifoUa A. Gray, are both listed in the Inventory of Rare and Endangered Plants of California (CNPS 2014). The current status of these taxa in the monument is discussed under Results/Rare and Endangered Taxa, below. The non-native taxa documented were Taraxacum officinale F.H. Wigg. and Phleum pratense L. Description of the Study Area Geology Devils Postpile National Monument sits at an average elevation of 2280 m (7600 feet [ranging from 2200-2500 m; 7200-8200 feet]) just below the Sierra Crest on the west slope of the Sierra Nevada. The postpile formation for which the monument is named is only one of the many manifestations of the widespread volcanic activity in this part of the Sierra Nevada. The basalt columns of the postpile were formed under a unique set of conditions: ample volume of magma, slow cooling time, and homogeneity of mineral composition (Huber and Eckhardt 1985). Glaciers that flowed down the valley of the Middle Fork San Joaquin River eroded most of the lava flows of the area and exposed the rocks that we see in the monument today. Other major geologic units in the monument include andesite of Mammoth Pass, rhyodacite of Rainbow Falls, and basalt of the Buttresses (Clow and Collum 1985, Fig. 2). The soils are mostly sandy with a thin surface layer of loose pumice, but in the meadows there is some development of organic material. Climate The monument’s climate is a reflection of its proximity to the Sierra Nevada Crest and the Mediterranean-type weather patterns of Califor- nia. Characterized by cold, wet winters and warm, dry summers, the majority of precipitation at the monument falls in the form of snow during the winter months. Temperatures during winter storm events remain at or below freezing. Between-storm diurnal fluctuations are more extreme, with daytime temperatures reaching 15°C, and recorded lows sometimes below — 18°C. Summer temperatures are moderate, with warm days and cool nights. Clear skies and low humidity are typical, but localized afternoon showers and flash floods created by monsoon patterns are possible during the summer months (Balmat and Scott 2010). Vegetation Near the headwaters of the Middle Fork San Joaquin River, this high-Sierra river corridor 2014] ARNETT ET AL.: VASCULAR FLORA OF DEVILS POSTPILE 369 Ranger Station Soda Springs Devils Postpile Kings Creek Trail Rainbow Falls Trail Rainbow Falls Legend o Collection site Trail == Road ^ Middle Fork San Joaquin River M Monument LJ boundary Meadow Basalt of Devils Postpile Andesite of Mammoth Pass Basalt of the Buttresses Ryodacite of Rainbow Falls 0.5 Miles "T 0.5 Kilometers Fig. 2. Map of Devils Postpile National Monument showing major features, major geologic units (adapted from Clow and Collum, 1985) and location of collection sites. The postpile after which the monument is named is composed of hexagonal basalt columns. Other major geologic units in the monument are older in origin and include andesite of Mammoth Pass, rhyodacite of Rainbow Falls, and basalt of the Buttresses. Collections were made wherever a new taxon was encountered. Efforts were made to maximize the number of specimens collected at each site. Each collection site therefore comprises an area approximately 25 m^ in size. Though the gaps between collection sites were searched, no collections were necessary in these areas. supports rich forests and verdant meadows. Species characteristic of both the wetter western and drier eastern slopes of the Sierra Nevada are present in Devils Postpile National Monument because of its proximity to the Sierra Nevada Crest. The monument falls within the central High Sierra Nevada geographic subdivision described by Baldwin et al. (2012). Though small in size, the monument is diverse enough in its topography and geology to support a number of different plant communities, including forests, chaparral, meadows, seeps, ponds, and riparian systems. Forests and Chaparral Pinus contorta Loudon subsp. murrayana (Grev. & Balf.) Critchf. (lodge- pole pine), which is prevalent in lowland com- munities, gives way to upland coniferous forests that include mixes of P. jeffreyi Balf. (Jeffrey pine), P. monticola Douglas ex D. Don (western white pine), Abies magnifica A. Murray bis (red fir), and A. concolor Lindl. (white fir). At the higher, drier elevations, these taxa also occur with occasional Juniperus grandis R.P. Adams (Sierra juniper) and Pinus albicauUs Engelm. (whitebark pine). The understory of these coniferous forests is relatively sparse on mid- to high-slopes, with Carex rossii Boott, Eriogonum nudum Douglas ex Benth., Gayophytum spp., Hieracium albiflorum Hook., Monardella odoratissima Benth. subsp. pallida (A. Heller) Epling, Phacelia hydrophyl- loides Torr. ex A. Gray, Stephanomeria tenuifolia (Raf.) H.M. Hall, and Stipa occidentalis Thurb. ex S. Watson being the most abundant herbs. Dry, sunny forest openings support small patches of chaparral consisting of various mixes of Arctostaphylos nevadensis A. Gray, A. patula Greene, Ceanothus cordulatus Kellogg, Prunus emarginata (Douglas) Eaton, Quercus vaccinifolia Kellogg, Ribes cereum Douglas var. cereum and R. roezlii Regel var. roezlii. Sunny openings also support numerous large stands of Pteridium 370 MADRONO aqiiilinimi (L.) Kuhn var. puhescens Underw. in both moist and dry drainage bottoms. Only one aspen stand of significant size is found in the monument. Meadows. On low-slopes and in drainage bottoms, accumulated moisture allows montane meadow vegetation to flourish. Common species found in the meadows include Car ex spp., Epilobium spp., Equisetum arvense L., Horkelia fiisca Lindl. var. parviflora (Nutt, ex Hook. & Arn.) Wawra, Lilium kelleyanum Lemmon, Lu- pinus polyphyllus Lindl. var. burkei (S. Watson) C.L. Hitchc., Maianthemum racemosiim (L.) Link, M. stellatum (L.) Link, Mimulus spp., Platanthera leucostachys Lindl., Primula Jeffrey i (Van Houtte) A.R. Mast & Reveal, and Trifolium spp. Seeps and Ponds. Seep vegetation is relatively common and contributes significantly to the diversity of the flora of Devils Postpile National Monument. Characterized by high amounts of soil moisture, the seeps support micro-communi- ties including Allophyllum integrifolium (Brand) A.D. Grant & V.E. Grant, Desehampsia dantho- nioides (Trin.) Munro, Hyper ieum anagalloides Cham. & Schltdk, Lithophragma glabrum Nutt., Microsteris gracilis (Douglas ex Hook.) Greene, Mimulus spp.. Polygonum douglasii Greene, and Trifolium monantlium A. Gray. Shallow ponds are common in the monument and further increase habitat diversity within the wetland- upland matrix. Riparian. Alnus incana (L.) Moench subsp. tenuifolia (Nutt.) Breitling, Cornus sericea L. subsp. sericea, and Salix spp. are the dominant woody species along the banks of the Middle Fork San Joaquin River. Populus trichocarpa Torr. & A. Gray (black cottonwood) occasionally dot the banks, while Arnica mollis Hook., Cliamerion angustifoliiim (L.) Holub, Helenium bigelovii A. Gray, and Sphenosciadium capiteUa- tum A. Gray are more common along river banks. Intermittent riparian habitats support species such as Allium validum S. Watson, Carex spp., Mimulus guttatus Fisch. ex DC., M. lewisii Pursh, and Toxicoscordion venenosum (S. Wat- son) Rydb. var. venenosum. Disturbance During 2004, fire-history sampling was con- ducted in the monument. Partial sections from fire-scarred trees were used to date past fire events. Results indicate moderate fire frequency over much of the monument. The mean fire return interval for four sites ranged from 14 to 18 years over a time span from about 1700 to 1860 AD (Caprio et al. 2006), Longer intervals between fires were indicated in the northwest [Vol. 61 i corner of the monument where Tsuga mertensi- I ana (mountain hemlock) and Pinus monticola ' (western white pine) occur. The Rainbow Fire was ignited by lightning on August 20, 1992 south of Devils Postpile National Monument in the Inyo National Forest. Strong winds moved the canopy fire northward up the Middle Fork San Joaquin River from its point of origin near Pond Lily Lake. When the . winds died down, the fire slowed and dropped to the forest floor, burning in surface fuels. Ap- proximately two-thirds of the monument was affected by the fire. In many areas, fire crept along the forest floor, occasionally burning into trees. The southeast portion of the monument shows signs of high-severity, wind-driven fire with high tree mortality. The fire history results indicate fire was not an unusual event in most of the monument’s forest communities; however, the absence of fire for 105 to 120 yr before 1992 was unprecedented and probably contributed significantly to the severity of the Rainbow Fire (Caprio et al. 2006). Most of the human activity and disturbance occurs near the eastern border of the monument, where visitation is concentrated around the postpile formation, Rainbow Falls, camp- grounds, and the ranger station (Fig. 2). Fishing is extremely popular in this part of the Sierra. Soda Springs Meadow and the gravel bars nearest the ranger station are heavily used by fishermen, resulting in the formation of unmain- tained social trails, or “fishing trails”. Horses and pack stock travel throughout the monument on the maintained trails. Visitation in the western half of the monument, especially to the south, is limited to hikers passing through on the King Creek Trail or the John Muir and Pacific Crest trail corridor (Fig. 2). Disturbance is therefore limited in these regions. Methods Search strategies and collection protocols were designed to accomplish the goals of documenting the occurrence of at least 90 percent of the species of vascular plants occurring in Devils Postpile National Monument with vouchered specimens and of describing the distribution and abundance of rare, threatened, endangered native species, and invasive non-native species. The survey was conducted by Melanie Arnett during the 2001 field season, which extended from June 18th through September 7th. To obtain a thorough coverage of the monu- ment and its habitats, Arnett combined broad searches aimed at covering as much area as possible with targeted searches of specific plant communities and habitat types. Because of the monument’s relatively small size, it was possible to cover 70 percent or more of the total surface 2014] ARNETT ET AL.: VASCULAR FLORA OF DEVILS POSTPILE 371 area on foot using this approach. With respect to habitat types, Arnett intermingled breadth with depth by visiting all habitat types found within the monument and thoroughly searching those likely to house species underrepresented in the previously-documented flora. Unique areas were visited repeatedly to capture plants at different phonological stages and to maximize the number of species documented by the survey. Habitat-Specific Targeted Searches Targeted searches enabled Arnett to document the distribution and abundance of both rare and non-native plant taxa occurring within the monument. A preliminary list of 43 “special- status” plants listed as rare by CNPS (Lists 1^) that were either known or considered likely to occur in the monument was used to inform directed searches for rare and or sensitive taxa (CNPS 1994^2001; Jones and Stokes 2001). Only two such taxa had previously been documented in the park. The majority of this list comprised potentially-occurring taxa documented to occur in areas adjacent to the park via searches of several data sources including collections at major California herbaria; the sensitive plant list for Inyo National Forest; and the California Natural Diversity Database (CNDDB 2000) (see Jones and Stokes 2001 for a complete list of data sources and “special-status” criteria). Habitat-specific searches were conducted for non-native invasive species. Population size was estimated for all non-native taxa encountered; invasive species were pulled out by the roots and left to decompose on site. Plants too big to be pulled were chopped down as close to ground level as possible. Seed heads or fruits that could potentially mature were either removed or destroyed, depending on their level of maturity. In addition to these surveys, a team of National Park Service botanists established 57 vegetation plots between July 25th and July 30th, 2001 to ground-truth the draft vegetation map of the monument. The draft map included 35 unique vegetation mapping units representing twelve alliances consistent with Yosemite National Park’s draft vegetation classification report, published after this study (Keeler- Wolf et al. 2012). Plots were established according to stan- dardized sampling protocols developed by the National Park Service vegetation mapping pro- gram (The Nature Conservancy and Environ- mental Systems Research Institute 1994), and were located in a representative area of each delineated polygon. Arnett accompanied the crew, and while the NPS botanists established and intensively sampled the plots, Arnett con- ducted extensive surveys of the corresponding polygons. Their combined efforts ensured that all 35 mapping units were visited over the course of the field season, and that habitats containing under-represented plant communities were searched in a thorough manner. Documentation of Plant Taxa Representative specimens of all vascular plant taxa encountered within the 319 ha of the monument were collected in order to document the vascular flora of Devils Postpile National Monument. At each collection site, UTM coor- dinates were recorded using a Garmin GPS 12 Personal Navigator^^, and notes were taken on the aspect, percent slope, slope position, sub- strate, surface material, elevation, and plant community associated with the site. When possible, all parts of the plant necessary for identification were collected, and enough material for at least two full herbarium sheets was included. The field season included a total of 23 field collection days between June 18th and September 7th. Relative abundance of each taxon was estimated at the end of the field season. Collections were made wherever a new taxon was encountered; however an effort was made to maximize the number of specimens collected at each site in order to increase efficiency. The majority of specimens were collected in June and July (Table 1). On average, 10 specimens per site were collected in June, 3.5 in July, one in August, and two in September. Figure 2 illustrates the location of collection sites, each of which represents an area of approximately 25 m^. Though the large gaps without collection sites were searched, no collections were necessary in these areas. The highest concentration of collec- tion sites was along the river, in riparian habitats and meadows, where species richness was highest. Hickman (1993), Munz (1965 and 1968), Botti (2001), and Cronquist et al. (1977) were the primary references used for identifications. No- menclature originally followed Hickman (1993), and has been updated for this publication to follow the second edition of the Jepson Manual (Baldwin et al. 2012; Jepson Flora Project 2013). A total of 338 specimens were verified at the University and Jepson Herbaria in Berkeley, California (UC & JEPS) during the weeks of 9- 13 July and 10-14 September 2001. The remain- ing 139 specimens were subsequently verified at the Rocky Mountain Herbarium in Laramie, Wyoming (RM). Dr. Allan Smith determined the ferns and Dr. Robert Dorn determined the willows. All other determinations were made by Melanie Arnett. Vouchers from this study are deposited at the DEPO, Jepson, and RM herbaria. Vascular plant collection and distributional data compiled for this study are stored in an Access database, as are vegetation mapping plot data. Spatial data are stored in ArcView project files. Digital data are 372 MADRONO [VoL 61 Table 1. Number of Specimens Collected for Each Day/Month. Total specimens collected Date Number of collection sites Number of specimens collected June 18 Jun 2001 8 34 20 Jun 2001 2 31 22 Jun 2001 4 40 23 Jun 2001 4 60 24 Jun 2001 7 47 25 Jun 2001 3 15 26 Jun 2001 3 30 27 Jun 2001 1 14 June TOTAL 32 271 July 3 Jul 2001 6 41 5 Jul 2001 3 6 15 Jul 2001 5 19 22 Jul 2001 3 17 24 Jul 2001 2 2 25 Jul 2001 7 33 26 Jul 2001 4 11 27 Jul 2001 6 18 28 Jul 2001 6 24 29 Jul 2001 5 11 30 Jul 2001 4 8 July TOTAL 51 190 August 6 Aug 2001 1 1 7 Aug 2001 2 2 18 Aug 2001 3 3 August TOTAL 6 6 September 7 Sep 2001 5 10 (primarily conifers) stored at Sequoia and Kings Canyon National Parks, where they are managed by the park plant ecologist. Results The total number of vascular plant taxa now documented from Devils Postpile National Mon- ument is 373 (List 1), representing a 121 percent increase over previous studies. We collected 507 specimens representing 343 taxa over the course of this study. From these specimens, 935 vouch- ers were produced; one complete set was depos- ited in the DEPO Herbarium, and duplicates, when available, were deposited at the Jepson and the Rocky Mountain herbaria (List 1). This study increased the total number of vouchers in the DEPO Herbarium from 287 to 754. A floristic summary for the monument is shown in Table 2. Two of the 43 potentially-occurring special status plants identified in Jones and Stokes (2001) were found: Hulsea brevifolia and Mimulus laciniatus. Based on a 121 percent increase in the number of recorded vascular plant taxa, and the combi- nation of both targeted surveys and plot-based sampling, we estimate that 90 percent or more of the vascular flora of Devils Postpile National Monument is now documented with vouchered specimens. At least 19 taxa that were previously documented with herbarium vouchers were not encountered during the field season of 2001, but vouchers were examined and confirmed by Arnett. An additional 25 taxa in List 2 were reported in the monument but never vouchered, or, if they were vouchered they were redetermined during the course of this study; these taxa may occur in the monument but they were not encountered during the 2001 field season. Table 2. Floristic Summary of Vascular Plant Taxa of Devils Postpile National Monument. Clade Families Genera Species Native Non-Native Total Species Lycophyta 1 1 1 0 1 Ferns 4 7 8 0 8 Gymnosperms 2 4 8 0 8 Eudicots 39 154 257 5 262 Monocots 13 39 90 3 93 Total 59 205 364 8 372 2014] ARNETT ET AL.: VASCULAR FLORA OF DEVILS POSTPILE 373 Rare and Endangered Taxa No federally or state listed rare or endangered plants are known to occur within Devils Postpile National Monument. Three species cited in the Inventory of Rare and Endangered Plants of California (CNPS 2014) were documented in this study. Two of these species were included in the Jones and Stokes draft list of special status plants prepared for the monument (Jones and Stokes 2001). One taxon that had previously been documented in the monument, Lupinus duranii (List IB), has since been redetermined to Lupinus lepidus var. seiiulus. Cinna bolanderi (Poaceae), known from Fresno, Mariposa, and Tulare counties, is docu- mented from one collection in the monument: Melanie Arnett 8425. This taxon is on the CNPS List IB, and has a threat rank of 0.2. There is only one other record for this taxon in Madera County, M.B. Dunkle 4685 (POM364968 n.v.), collected in 1935 at Reds Meadow. The site at which this specimen was collected typifies the known habitat of Cinna bolandert a moist SSW- facing, spring-fed drainage surrounded by Jeffrey pine/ red fir forest. Vouchers from the specimen collected by Arnett are deposited at both the DEPO and Jepson herbaria. Hulsea brevifolia (Asteraceae), known from El Dorado, Fresno, Madera, Mariposa, Tulare, and Tuolumne counties, is documented from nine collections in Devils Postpile National Monu- ment: Joseph L. Medeiros and W. Eckhardt s.n.; Sandra C. Morey 154; Dieter H. Wilken 8241 (SEINET480163, SD74615, UCD133277, UCSB26631); G. L. Stebbins Jr. 2617 (UC797814); Peter H. Raven 3685 (CAS372031); H. Williams s.n. (UCSB71604); Melanie Arnett 8110 (JEPS99762), Melanie Arnett 8192, and 8401. This taxon is on CNPS List IB; its threat rank is 0.2. It is interesting to note that the morphologic differen- tiation between H. brevifolia and H. mexicana is the number of ray flowers (10-23 and 20-35 respectively) and whether the corolla tube hairs are a mixture of glandular and nonglandular hairs, in the case of the former, or of all glandular hairs, as in the latter (Wilken 2013). This distinction was not apparent in the specimens examined at the UC & JEPS Herbaria. Hulsea mexicana, known in the U.S. from only a single location (San Diego County), grows in volcanic substrates and in burned or disturbed sites. At the time of this survey, H. brevifolia was quite common in volcanic substrates in the post-fire region of the monument. Vouchers from this study are deposited at the DEPO, JEPS, and RM herbaria. Mimulus laciniatus (Phrymaceae), known from Amador, Butte, Fresno, Madera, Mariposa, Plumas, Tulare, and Tuolumne Counties, is documented from two collections in the monu- ment: Melanie Arnett 8026, and 8309. This taxon is on the CNPS watch list (List 4), and its threat rank is 0.3. Both collections of this taxon were made from seeps on granite, which is in accordance with the habitat given in the Jepson eFlora (Jepson Flora Project 2013). Non-native Taxa Of the eight non-native taxa (List 1) that were documented, Cirsium vulgare (bull thistle) was the only taxon that appeared to be rapidly expand- ing. Numerous populations of this taxon, each consisting of 5—250 individuals, were encoun- tered, especially in the area known as the Buttresses (Fig. 2). Removal efforts began for C. vulgare in 2001; follow-up control measures were implemented on all populations of C vulgare in subsequent years (see Methods). The majority of the other non-native taxa were more localized, occurring near the ranger station in and around the meadows that are used for access to fishing (Fig. 2). Potential Range Extensions Based on the range descriptions provided in the Jepson Flora Project (2013), nine range exten- sions were documented with this inventory (List 3). The extension of these ranges included six elevation extensions, which varied from a differ- ence of 200 m to 1000 m in elevation. Four taxa documented through this survey represent new occurrences within the central High Sierra Nevada subdivision of California. Discussion The results of this inventory support Ertter’s (2000) assertion that comprehensive, species-level inventories are not too cumbersome to be of value, as is generally assumed from both a logistical and financial standpoint. The small size of Devils Postpile National Monument provided ideal conditions for coupling a species-level inventory with vegetation mapping in a timely and fiscally responsible manner, as is outlined by the Inventory and Monitoring program and advocated by Charlet (2000). The results of this inventory emphasize the fact that a small-scale inventory can yield large-scale results. The 121 percent increase in the documented flora of the monument brings with it numerous other data that are valuable to managers and researchers in many fields of study. Nine probable range extensions were documented (List 3). Two rare species that were previously unknown in the monument were documented, one of which represents a new Madera county record. An infestation of the invasive thistle Cirsium vulgare in a portion of the monument that seldom, if 374 MADRONO [Vol. 61 ever, sees human visitors was detected in time to enlist support from the National Park Service - California Exotic Plant Management Team and successfully control populations (National Park Service 2011). In addition. List 2 and 4 offer resource managers an idea of which plant taxa to look for in the future in order to continue increasing our knowledge of the flora of Devils Postpile National Monument. The results of this inventory have also assisted the parks within the Sierra Nevada Inventory and Monitoring Network (SIEN) in subsequent resource management planning efforts, including invasive species and resource management plans. Early detection efforts enabled the implementa- tion of rapid response protocols and successful invasive species management (National Park Service 2011). Data from the results of this study will be useful as Devils Postpile National Monument prepares its resource stewardship strategy and will be available for use in interpretive programs that enrich the visitor experience. Opportunities for Future Research As mentioned in the introduction, the flora of Devils Postpile National Monument contains characteristics of both east- and west-side Sierra Nevada floras. There were 373 taxa found in this 319 ha study area, compared to 446 taxa found in a 4400 ha floristic study of the San Joaquin Roadless Area, managed by Inyo National Forest and just east of Devils Postpile National Monument (Constantine-Shull 2000). Other flo- ristic studies in the area include Howald (1983), Bagley (1988), and Taylor (1981). In a compar- ative analysis of 13 high-elevation floras, con- ducted by Constantine-Shull, the flora of the San Joaquin Roadless Area was found to be floristi- cally more similar to other westside floras of the Sierra Nevada than to those of the eastside. Of the westside floras, however, this one had the highest similarity to eastside floras included in the analysis. Devils Postpile National Monument was not included in the analysis, due the absence of grasses, sedges, and willows on the species list at the time. Were the results of this survey to be included in such an analysis, we hypothesize that the flora of Devils Postpile would be in a position close to the San Joaquin Roadless Area with respect to other Sierra Nevada floras. All floras and florulas represent a snapshot in time of the botanical inventory of an area, and at least some changes are expected, and inevitable, over time. Continued exploration of the monu- ment is desirable, especially additional focused searches for taxa documented to occur nearby but not yet found within the monument. Since this study was conducted, the development of the online Consortium of California Herbaria (CCH) has vastly improved the ability to access a large volume of herbaria records. A recent search yielded six taxa collected within the monument that were not encountered during this study and for which vouchers were not verified. These collections were made between 1935 and 1980, and it is possible that these taxa no longer occur within the Monument but should be included among the list of plants to watch for in Devils Postpile. These taxa are included in list 4. Acknowledgements Many people helped make this survey a success. Special thanks go to the Sierra Nevada Network Inventory and Monitoring Program for funding this study. The staff at Devils Postpile National Monument have our gratitude for their contributions: Deanna Dulen, Lisa Bassani, Laura Wilvert, Brook Fisher, Peter Houpt, Vireo Gaines, Julie Raiche, and Monica Buhler. We would like to thank the staff and curators of the University & Jepson Herbaria at UC Berkeley and the Rocky Mountain Herbarium, especially Richard Moe, David Baxter, Linda Vorobik, Allan Smith, Bruce Baldwin, Barbara Ertter, Ronald Hartman, and Ernie Nelson. David Baxter and Richard Moe at the University & Jepson Herbaria were instrumental in help with updating the nomenclature. Thanks also to Robert Dorn, The Ecological Careers Organization, Linda Mutch, Jonathan Nesmith, the Sequoia and Kings Canyon National Parks vegetation mapping crew (Jennifer Akin, Cheryl Bartlett, Sarah Kane, Catie Karplus, Laura Pilewski, Sally Reynolds, and Lorie Werner), Brian Knauss, Peggy Moore, Eric Frenzel, Angela Evenden, Joseph L. Medeiros, and K. Ann Hoffmann. Thanks to Daniel Phipps MeCoy for all his help and support, and to Alice Chung-MacCoubrey for her careful review of and comments on the manuscript. We are also grateful for eomments provided by two anonymous reviewers, which led to significant improve- ments in the manuseript. Literature Cited Arnett, M. and S. Haultain. 2005. Final report to the Sierra Nevada Inventory and Monitoring Network Sequoia and Kings Canyon National Parks, Three Rivers, CA. Unpublished report prepared for the National Park Service. Bagley, M. 1988. Sensitive plant species survey for June Mountain Ski Area and Rodeo Meadows, Inyo National Forest, Mono County, CA. Unpub- lished report prepared for Mammoth/June Ski Resort, Mammoth Lakes, CA. Baldwin, B. G., D. H. Goldman, D. J. Keil, R. Patterson, T. J. Rosatti, and D. H. Wilken, (eds.). 201 2. The Jepson Manual: vascular plants of California, 2nd ed. University of California Press, Berkeley, CA. Balmat, j. and D. K. Scott. 2010. Weather data inventory. Devils Postpile National Monument. Natural Resource Data Series NPS/SIEN/ NRDS— 2010/1 13. National Park Service, Fort Collins, CO. Botti, S. j. 2001. An illustrated flora of Yosemite National Park. Yosemite Association, El Portal, CA. 2014] ARNETT ET AL.: VASCULAR FLORA OF DEVILS POSTPILE 375 California Native Plant Society (CNPS). 1994- 2001. Electronic inventory of rare and endangered vascular plants of California, Version 1.5.1 (Data updated 10 October 2001). Sacramento, CA. , Rare Plant Program. 2014. Inventory of rare and endangered plants (online edition, V8-02). California Native Plant Society. Sacramento, CA. Website http://www.rareplants.cnps.org (accessed 17 April 2014). California Natural Diversity Database (CNDDB). 2000. RareFind 2, Version 2.1.2 (5 September 2000 update). California Department of Fish and Game, Sacramento, CA. Caprio, a. C., M. Keifer, and K. Webster. 2006. Long-term effects of the 1992 Rainbow Fire, Devils Postpile National Monument, California. Extend- ed abstract of oral presentation at the third international fire ecology and management con- gress, Association for Fire Ecology, 13-17 Nov. 2006, San Diego, CA. Charlet, D. a. 2000. Coupling species-level invento- ries with vegetation mapping. Madrono 47:259- 264. Clow, D. W. and K. R. Collum. 1985. Distribution of volcanic rocks: Devils Postpile National Monument and vicinity. Pp. 20-21 in N. K. Huber and W. Eckhardt, Devils Postpile story. Sequoia Natural History Association, Three Riv- ers, CA. Consortium of California Herbaria (CCH). 2014. Data provided by the participants of the Consortium of California Herbaria. Website http://ucjeps.berkeley.edu/consortium/ (accessed 24 May 2014). Constantine-Shull, H. M. 2000. Floristic affinities of the San Joaquin Roadless Area, Inyo National Forest, Mono County, California. M.S. thesis. Humboldt State University, Areata, CA. Cronquist, a. 1977. Intermountain flora: vascular plants of the Intermountain West, USA. Volume Six, Monocotyledons. New York Botanical Gar- den, New York, NY. Denn, M. and D. Shorrock. 2009. Devils Postpile National Monument wetland inventory and con- dition assessment. National Park Service, Water Resources Division, Three Rivers, CA. Ertter, B. 2000. Our undiscovered heritage: past and future prospects for species-level botanical inven- tory. Madrono 47:237-252. Fancy, S. G., J. E. Gross, and S. L. Carter. 2009. Monitoring the condition of natural resources in US national parks. Environmental Monitoring and Assessment 151:161-174. Hickman, J. C. 1993. The Jepson manual: higher plants of California. University of California Press, Berkeley, CA. Howald, a. M. 1983. The Vegetation and Flora of Mammoth Mountain. Unpublished report pre- pared for Mr. Dave McCoy, Mammoth Mountain Ski Area, Mammoth Lakes, CA. Huber, N. K. and W. Eckhardt. 1985. Devils Postpile story. Sequoia Natural History Associa- tion, Three Rivers, CA. Jepson Flora Project (eds.). 2013. Jepson eFlora. Website http://ucjeps.berkeley.edu/IJM.html (ac- cessed 17 April 2014). Jones AND Stokes. 2001. Draft list of the special status vascular plants of Devils Postpile National Mon- ument. Unpublished report. National Park Service files. Sequoia and Kings Canyon National Parks, Three Rivers, CA. Keeler-Wolf, T., P. E. Moore, E. T. Reyes, J. M. Menke, D. N. Johnson, and D. L. Karavidas. 2012. Yosemite National Park vegetation classifi- cation and mapping project report. Natural Resource Technical Report NPS/YOSE/NRTR — 2012/598. National Park Service, Fort Collins, CO. Medeiros, J. 1996. A plant checklist of Devils Postpile National Monument, Madera County, California. Unpublished document. National Park Service files. Sequoia and Kings Canyon National Parks, Three Rivers, CA. Munz, P. a. 1968. A California flora: Supplement. University of California Press, Berkeley, CA. AND D. D. Keck. 1965. A California flora. University of California Press, Berkeley, CA. National Park Service. 2001. Biological inventory plan for the Sierra Nevada Network. Three Rivers, CA. . 2009. Strategic plan for natural resource inventories: FY 2008 - FY 2012. Natural Resource Report NPS/NRPC/NRR— 2009/094. National Park Service, Fort Collins, CO. . 2011. Devils Postpile Monument invasive plant management plan: five year plan 201 1-2016. Devils Postpile National Monument, CA. Taylor, D. W. 1981. Plant checklist for the Mono Basin. Mono Basin Research Group, Contribution No. 3. Lee Vining, CA. The Nature Conservancy and Environmental Systems Research Institute. 1994. Field meth- ods for vegetation mapping. Prepared for the United States Department of Interior National Biological Survey and National Park Service. Avail- able online at http://wwwl.usgs.gov/vip/standards/ fieldmethodsrpt.pdf (accessed 31 July 2014). WiLKEN, D. H. 2013. Hulsea in Jepson Flora Project (eds.), Jepson eFlora, Website http://ucjeps.berkeley. edu/cgi-bin/get_IJM.pl?tid=3566 (accessed 28 May 2014). Appendix 1 List 1 The vascular plant checklist for Devils Postpile National Monument. List is organized by clade following the second edition of The Jepson Manual (Baldwin et al. 2012) and online Jepson eFlora (Jepson Flora Project 2013), then by family and taxon name (alphabetical order within family). Abundance is given for each taxon (local: occurring occasionally in small populations, occasional: occurring occasionally as individuals, locally common: occurs occasionally in larger populations, locally abundant: occurs often in dense stands, uncommon: unlikely to be encountered and sometimes not present in appropriate habitats, and abundant: very likely to be encountered/ nearly always found in appropriate habitats, sometimes forming dense stands). Taxa preceded by an asterisk (*) are non- native, and those by a broken diamond symbol (❖) are listed in the California Native Plant Society’s Inventory of Rare and Endangered Plants of California (CNPS 2014). Habitat or vegetation community type informa- tion is given when available, and reflects information 376 MADRONO [Vol. 61 taken from voucher labels as well as vegetation mapping plot data. Taxa that were not collected in this study are cited from collections made by the following collectors; KAH = K. Ann Hoffmann, SCM = Sandra C. Morey, JLM = Joseph L. Medeiros, IN = Jan Nachlinger; collection numbers were not supplied with the specimens of the latter two collectors. Voucher records from these collectors held in the DEPO herbarium but not referenced here are available upon request from the National Park Service Sierra Nevada Network Inventory and Monitoring Network data manager (http://science.nature.nps.gov/IM/units/sien/) and/or the corresponding author. All collection num- bers higher than 8000 are from collections made by Melanie Arnett (MA) with or without Sylvia Haultain, Linda Mutch, and/or Daniel Phipps McCoy. Collection numbers presented in the format ‘DEPO. [plot num- ber], [collection number]’ represent vegetation mapping plot vouchers. Specimens were deposited in the following herbaria (DEPO = Devils Postpile National Monument Herbarium (DEPO), Jeps. Herbarium (JEPS), and Rocky Mountain Herbarium (RM). Accession numbers assigned by JEPS provided by the participants of the Consortium of California Herbaria (ucjeps.berkeley.edu/consortium/); online posting of accession numbers assigned by RM is in progress and not yet available when this manuscript was prepared for press. Ferns and Lycophytes Dennstaedtiaceae Pteridium aquilinum (L.) Kuhn var. pubescens Un- derw., locally common, mixed red fir - white fir forest, mountain alder thicket, and seasonally flooded riverine habitats, 26 Jul 2001 MA 8400 (DEP0913, JEPS 99704). Equisetaceae Equisetum arvense L., locally common, riverine and palustrine wetlands, black cottonwood forest, mountain alder and willow thickets, 24 Jun 2001, MA 8182 (DEP0917, JEPS99702); E. laevigatum A. Braun, locally common, moist granitic gravels in Jeffrey pine - white fir forest, 15 Jul 2001, MA 8348 (DEP0918, JEPS99703, RM). Pteridaceae Aspidotis densa (Brack.) Lellinger, uncommon, basalt outcrop and associated seep in white fir forest, 26 Jun 2001, MA 8259 (DEPO 1127, JEPS99707, RM); Crypto- gramma acrostichoides R.Br,, occasional, east-facing granite outcrop in Jeffrey pine forest with scattered western juniper, 24 Jun 2001, MA 8214 (DEP01128, JEPS99708); Pellaea hreweri D.C. Eaton, uncommon, west-facing basalt outcrop in lodgepole pine forest, 25 Jul 2001, MA 8391 (DEP01129, JEPS99709, RM). Selaginellaceae Selaginella watsonii Underw., local, granite outcrops, 29 Jul 2001, DEPO.0040.08 (DEPO 1209). Woodsiaceae Athyrium filix-femina (L.) Roth var. cyclosorum Rupr., occasional, black cottonwood forest, mountain alder and willow thickets, 15 Jul 2001, MA 8343 (DEP0914, JEPS99705, RM); Cystopteris fraplis (L.) Bernh., occasional, moist rock crevices, moist slope in mist of falls, 25 Jun 2001, MA 8225 (DEP0916, JEPS99706), 24 Jun 2001, MA 8178 (DEP0915, RM). Gymnosperms Cupressaceae Juniperus grandis R.P. Adams, occasional, steep east- facing slopes with sagebrush, huckleberry oak, and Jeffrey pine, 9 Jul 2001, MA 8478 (DEPO620, JEPS99749). Pinaceae Abies concolor (Gordon & Glend.) Lindl. ex Hildebr., abundant, red fir - white fir forest, lodgepole pine forest and woodland, with Jeffrey pine and huckleberry oak, 9 Jul 2001, MA 8485 (DEP0638, JEPS99931); A. magnifica A. Murray bis var. magnifica, abundant, in pure stands of red fir forest, mixed red fir - white fir forest, lodgepole pine forest, in mixed stands with Jeffrey pine, western white pine, 9 Jul 2001, MA 8481 (DEP0639, JEPS99932); Pinus albicaulis Engelm., uncommon, dry slope in lodgepole pine forest with red fir, 25 Jul 2001, MA 8385 (DEPO640, JEPS99933); P. contorta Douglas ex Loudon subsp. murrayana (Grev. & Balf.) Critchf, locally abundant, forms pure stands of forest and woodland; also occurs as a component of red fir - white fir forests on upland sites, and black cottonwood - mountain alder associations along the river, 9 Jul 2001, MA 8480 (DEP0641, JEPS99934); P. jeffreyi Grev. & Balf., common, forms open woodlands on exposed slopes, emergent from huckleberry oak shrublands, component of red fir - white fir, Jeffrey pine - red fir, and Jeffrey pine - white fir - red fir forests, 9 Jul 2001, MA 8486 (DEP0642, JEPS99935); P. monticola Douglas ex D. Don, local, component of red -fir western white pine forest, Jeffrey pine-huckleberry oak woodland; and Jeffrey pine - white fir - red fir forests, 9 Jul 2001, MA 8479 (DEP0643, JEPS99919); Tsuga mertensiana (Bong.) Carriere, uncommon, north-facing slope with red fir, 24 Jul 2001, MA 8367 (DEP0644). Eudicots Adoxaceae Sambucus nigra L. subsp. caerulea (Raf.) Bolli, occasional, along banks of river, 28 Aug 1977, JLM 129 (DEPO507). Apiaceae Cymopterus terebinthinus (Hook.) Torr. & A. Gray var. californicus (J.M. Coult. & Rose) Jeps., local, top of rhyodacite cliffs in Jeffrey pine - lodgepole pine forest, 8 Jul 2001, MA 8473 (DEP0559, JEPS99800); Heracleum maximum W. Bartram, local, riparian, 28 Jul 2001, MA 8433 (DEPO560, JEPS99801); Ligusticum 2014] ARNETT ET AL.: VASCULAR FLORA OF DEVILS POSTPILE 377 grayi J.M. Coult. & Rose, local, moist areas in riparian understory, 18 Aug 2001, MA 8476 (DEP0561, JEPS99802); Lomatium torreyi (J.M. Coult. & Rose) J.M. Coult. & Rose, uncommon, in vertical cracks of steep basalt column cliffs, 25 Jul 2001, MA 8397 (DEP0562, JEPS99803); Osmorhiza berteroi DC., local, aspen grove, 23 Jun 2001, MA 8165 (DEP0564), 20 Jun 2001, MA 8064 (DEP0563, JEPS99804); O. occidentalis (Nutt.) Torr., uncommon, dry, loose pumice over sandy granitic soil near falls, 24 Jun 2001, MA 8198 (DEP0565, JEPS99805); Oxypolis occidentalis J.M. Coult. & Rose, uncommon, wet meadow; mountain alder and willow thickets, 29 Jul 2001, MA 8461 (DEP0566, JEPS99806); Perideridia parishii (J.M. Coult. & Rose) A. Nelson & J.F. Macbr. subsp. latifolia (A. Gray) T.I. Chuang & Constance, common, seep in red fir - white fir forest, black cottonwood forest, mountain alder and willow thickets, wet meadow, 26 Jul 2001, MA 8405 (DEP0567, JEPS99807); Sium suave Walter, uncommon, moist slope in Jeffrey pine - red fir forest in mucky, loamy sand, mountain alder thickets, 27 Jul 2001, MA 8428 (DEPO570, JEPS99808); Sphenosciadium capitellatum A. Gray, occasional, wet meadows and streamsides, mountain alder thickets, moist areas in lodgepole pine and red fir forests, 24 Jun 2001, MA 8223 (DEP0568, JEPS99809). Apocynaceae Apocynum androsaemifolium L., common, open slopes, dry understory of conifer forests, chaparral, 22 Jun 2001, MA 8083 (DEP0571, JEPS99790). Asteraceae Achillea millefolium L, common, upland and wetland habitats, including conifer forest, riparian forests and shrublands, wet meadows, 24 Jun 2001, MA 8217 (DEP0574, JEPS99791); Ageratina occidentalis (Hook.) R.M. King & H. Rob., local, basalt rocks near the postpiles, sagebrush and bitter cherry shrublands, 29 May 1977, JLM (DEPO108), 6 Aug 1977, JN (DEPO107, DEP0487); Agoseris X elata (Nutt.) Greene (pro. sp.), local, mesic lodgepole pine forest, 29 Jul 2001, MA 8454 (DEP0575); A. monticola Greene, occasional, dry terrace in lodgepole pine forest, 25 Jul 2001, MA 8399 (DEP0576); A. retrorsa (Benth.) Greene, uncommon, Jeffrey pine - white fir forest, 22 Jun 2001, MA 8081 (DEP0577, JEPS99792); Anaphalis margaritacea (L.) Benth. & Hook, f., common, riparian and upland habitats, understory of conifer forests, rock outcrops, 15 Jul 2001, MA 8336 (DEP0578, JEPS 99793); Antennaria corymbosa E.E. Nelson, uncommon, wet meadow and gravel bars, 27 Jun 2001, MA 8272 (DEP0579, JEPS99794, RM); A. rosea Greene subsp. confinis (Greene) R.J. Bayer, common, lodgepole pine forest, red fir forest, wet meadows and seeps, 22 Jun 2001, MA 8083 (DEP0571), 15 Jul 2001, MA 8330 (DEPO802), 25 Jul 2001, MA 8392 (JEPS99795); Arnica chamissonis Less., locally common, meadow edge adjacent to lodgepole pine forest, black cottonwood forest, mountain alder thickets, 28 Jul 2001, MA 8437 (DEPO803, JEPS99796, RM); A. mollis Hook., locally common, moist slope in mist of falls, willow thickets, 25 Jun 2001, MA 8234 (DEPO805), 5 Jul 2001, MA 8324 (DEPO804, JEPS99797); Artemisia douglasiana Besser, locally abundant, mountain alder and willow thickets, 28 Jul 2001, MA 8434 (DEPO806, JEPS99798); A. ludoviciana Nutt, subsp. incompta (Nutt.) D.D. Keck, locally abundant, riparian habitats and meadow edges, 27 Jul 1977, TLM(DEP094); A. tridentata Nutt, subsp. vaseyana (Rydb.) Beetle, uncommon, openings in lodgepole pine forest and Jeffrey pine forest, 9 Jul 2001, MA 8483 (DEPO807, JEPS99799); Chaenactis alpigena Sharsm., uncommon, pumice slope in conifer- ous forest, 19 Aug 1978 JLM (DEP0694); C. douglasii (Hook.) Hook. & Arn. var. douglasii, locally abundant, sandy openings in coniferous forest, often on pumice, 22 Jun 2001, MA 8114 (DEP0815), 22 Jun 2001, MA 8080 (DEP0816, JEPS99785); Cirsium andersonii (A. Gray) Petr., local, post-bum Jeffrey pine - white fir forest, red fir - white fir forest, 22 Jun 2001, MA 8077 (DEP0818, JEPS99787); C. scariosum Nutt., local, wet meadow and gravel bars, 27 Jun 2001, MA 8274 (DEP0819, JEPS99788, RM); *C. vulgare (Savi) Ten., native to Europe, locally common, granite outcrop near sag pond, mountain alder thickets, understory of coniferous forest, 27 Jul 2001, MA 8423 (DEPO820); Ericameria bloomeri (A. Gray) J.F. Macbr., occasional, on dry pumice soil along river, lodgepole pine, red fir, and white fir forests, black cottonwood forest, mountain alder thickets, 16 Sep 1980, SCM 244 (DEPO104), 17 Aug 1977, JLM (DEPO103); E. nauseosa (Pall, ex Pursh) G.L. Nesom & G.I. Baird var. speciosa (Nutt.) G.L. Nesom & G.L Baird, occasional, sandy volcanic gravels in post-burn coniferous forest, 27 Jul 2001, MA 8414 (DEP0817, JEPS99786, RM); E. parryi (A. Gray) G.L. Nesom & G.L Baird var. monocephala (A. Nelson & P.B. Kenn.) G.L. Nesom & G.L Baird, uncommon, basalt outcrop in lodgepole pine forest, 25 Jul 2001, MA 8396 (DEP0572); Erigeron breweri A. Gray var. breweri, local, basalt outcrops, post-burn Jeffrey pine -white fir forest, 22 Jun 2001, MA 8079 (DEPO809, JEPS99770), 27 Jul 2001, MA 8411 (DEPO810, JEPS99789, RM); E. coulteri Porter, occasional, basalt outcrops in Jeffrey pine - red fir forest, 27 Jul 2001, MA 8424 (DEP0821, JEPS99771); E. elmeri (Greene) Greene, occasional, Jeffrey pine forest, cliffs and rock outcrops, 24 Jun 2001, MA 8215 (DEP0823); E. glacialis (Nutt.) A. Nelson var. glacialis, local, dry streambed, moist wash, and riparian habitat, 28 Jul 2001, MA 8448 (DEP0824, JEPS99773), 24 Jun 2001, MA 8218 (DEP0826, JEPS99772, RM), 22 Jun 2001, MA 8286 (DEP0827); E. glacialis (Nutt.) A. Nelson var. hirsutus (Cronquist) G.L. Nesom, local, lodgepole pine forest, dry slopes near river, 3 Jul 2001, MA 8313 (DEP0827, JEPS99774, RM), 24 Jun 2001, MA 8193 (DEP0582); E. lonchophyllus Hook., uncom- mon, meadow edge adjacent to lodgepole pine forest, 28 Jul 2001, MA 8438 (DEP0837, JEPS99733); Eriophyl- lum lanatum (Pursh) J. Forbes var. integrifolium (Hook.) Smiley, uncommon, dry slopes and benches in conifer- ous forest, 22 Jul 2001, MA 8350 (DEP0583, JEPS99775); Eucephalus breweri (A. Gray) G.L. Nesom, occasional, open coniferous woodlands, 24 Jun 2001, MA 8189 (DEPO808, JEPS99780); Eurybia integrifolia (Nutt.) G.L. Nesom, occasional, open coniferous forest, moist wash, 9 Nov 1980, SCM 232 (DEP096); Gnaphalium palustre Nutt., locally common, meadow and pond edges, 27 Jul 2001, MA 8419 (DEP0588, JEPS99779), 25 Jul 2001, MA 8368 (DEP0589); Helenium bigelovii A. Gray, occasional, sandy riparian soils, 28 Jul 2001, MA 8431 (DEP0591, JEPS99760), 24 Jun 2001, MA 8220 (DEPO590); Hieracium albiflorum Hook., abundant, understory of coniferous forest, mountain alder thickets, black cottonwood forest. 378 MADRONO 25 Jun 2001, MA 8236 (DEP0592, JEPS99761); H. horridum Fr., common, basalt rock outcrops, granite cliffs, open pumice flats, 18 Jun 2001, MA 8018 (DEP0593); *>Hidsea brevifoHa A. Gray, common, in post-burn red fir forest with lodgepole pine, 26 Jul 2001, MA 8401 (DEP0594), 24 Jun 2001, MA 8192 (DEP0595, RM), 22 Jun 2001, MA 8110 (DEP0596, JEPS99762); H. vestita A. Gray subsp. vestita, uncom- mon, red fir - white fir forest, exposed pumice, 26 Jul 1972 KAH 184 (DEP0519); *Lactuca serriola L., native to Europe, local, exposed loose pumice, 22 Aug 2001, MA 8477 (DEP0597, JEPS99763); Madia exigua (Sm.) A. Gray, local, basalt outcrop and associated seep in white fir forest, 26 Jun 2001, MA 8249 (DEPO600, JEPS99764); Microseris nutans (Hook.) Sch. Bip., local, loose pumice in red fir - white fir forest, Jeffrey pine woodland, lodgepole pine forest, 5 Jul 2001, MA 8328 (DEPO601, JEPS99766), 24 Jun 2001, MA 8190 (DEPO602, RM), 18 Jun 2001, MA 8033 (DEPO603, JEPS99765); Packera cana (Hook.) W.A. Weber & A. L5ve, local, basalt outcrop at top of postpile formation, 10 Jul 1977, JLM (DEPOllO), 3 Jul 1980, SCM 114 (DEPOll 1); Pseudognaphalium californicum (DC.) An- derb., occasional, loose pumice over sand, 8 Jul 2001, MA 8474 (DEP0584, JEPS99776); P. thermale (E.E. Nelson) G.L. Nesom, common, post-burn red fir - white fir forest, riverbed and gullies, 26 Jul 2001, MA 8409 (DEP0586, JEPS99777), 9 Jul 2001, MA 8482 (DEP0585, JEPS99778, RM); Senecio integerrlmus Nutt. var. exaltatus (Nutt.) Cronquist, common, dry north-facing slopes near river, 24 Jun 2001, MA 8191 (DEPO604, JEPS99767); S. scorzonella Greene, uncom- mon, wet meadow and gravel bars, 27 Jun 2001, MA 8271 (DEP0828, JEPS99768); X triangularis Hook., locally abundant, black cottonwood forest, mountain alder and willow thickets, damp places in understory of coniferous forest, 20 Jun 2001, MA 8068 (DEP0829, JEPS99769); Solidago elongata Nutt., locally abundant, black cottonwood forest, aspen grove; mountain alder and willow thickets, wet meadow, 15 Jul 2001, MA 8337 (DEPO830, JEPS99730); Stephanomeria teniiifolia (Torr.) H.M. Hall, abundant, rocky outcrops, under- story of fir forest with well-developed duff as well as post-burn, 3 Jul 2001, MA 8321 (DEP0831, JEPS99731, RM); Symphyotrichiim bracteo/atum (Nutt.) G.L. Ne- som, uncommon, aspen grove; base of cliffs adjacent to river, 18 Aug 2001, MA 8475 (DEP0812, JEPS99781, RM); S. foliaceum (Lindl. ex DC.) G.L. Nesom var. parryi (D.C. Eaton) G.L. Nesom, uncommon, moist slopes, 27 Jul 2001, MA 8426 (DEP0813, JEPS99783, RM); S. spathulatiim (Lindl.) G.L. Nesom var. spathu- latiim, occasional, wet meadow; willow thickets, 29 Jul 2001, MA 8453 (DEP0814, JEPS99784); ^Taraxacum officinale F.H. Wigg., native to Europe, local, moist riparian areas, wet meadow, 3 Jul 2001, MA 8287 (RM), 27 Jun 200 IM^ 8281 (JEPS99732), 23 Jun 2001, MA 8154 (DEP0835); *Tragopogon dubius Scop., native to Europe, uncommon, post-burn coniferous forest, 22 Jun 2001, MA 8089 (DEP0836); Wyethia mollis A. Gray, occasional, post-burn Jeffrey pine - white fir forest, 22 Jun 2001, MA 8078 (DEP0838, JEPS99734). Betulaceae Alnus incana (L.) Moench subsp. tenuifoUa (Nutt.) Breitung, common, forming dense thickets along riparian corridor, forming understory of black cotton- [Vol. 61 wood forest, 20 Jun 2001, MA 8076 (DEP0839, JEPS99746). Boraginaceae Cryptantha affinis (A. Gray) Greene, locally com- mon, moist areas in red fir forest and red fir - white fir forest, 26 Jul 200 IMA 8410 (DEP0843, JEPS99735, RM), 22 Jun 2001, MA 8095 (DEP0842); C echinella Greene, locally common, dry areas in lodgepole pine forest, red fir - white fir forest, 3 Jul 2001, MA 8318 (DEP0845); C. torreyana (A. Gray) Greene, locally common, in dry area at meadow edge under lodgepole pine, 4 Aug 1980, SCM 186 (DEPO120); Hackelia micrantha (Eastw.) J.L. Gentry, local, moist areas in red fir - white fir forest, streambanks in mountain alder thickets and black cottonwood forest, 22 Jun 2001, MA 8104 (DEP0846, JEPS99736); H. mundula (Jeps.) Ferris, occasional, openings in dry coniferous forest, 27 Jul 2001, MA 8422 (DEP0848), 20 Jun 2001, MA 8053 (DEP0847, JEPS99737, RM); H. velutina (Piper) I.M. Johnst., occasional, rocky slope in riparian habitat, 7 Jun 1980, SCM 128 (DEP0122); Hesper- ochiron pumilus (Griseb.) Porter, uncommon, wet meadow edge, 3 Jul 1980, SCM 109 (DEPO140); Nemopliila spatulata Coville, local, wet meadow; mountain alder thickets, 23 Jun 2001, MA 8129 (DEP0962, JEPS99647); Phacelia eisenii Brandegee, uncommon, moist slope in mist of falls, 25 Jun 2001, MA 8233 (DEP0963, JEPS99648); P. hastata Douglas ex Lehm. var. compacta (Brand) Cronquist, common, sandy benches and outcrops in Jeffrey pine - white fir forest, red fir - western white pine forest, 20 Jun 2001, MA 8067 (DEP0964, JEPS99649); P. hydrophylloides A. Gray, uncommon, lodgepole pine forest, red fir forest, 3 Jul 2001, MA 8306 (DEP0965); P, mutabilis Greene, occasional, Jeffrey pine - white fir forest, lodgepole pine forest, 20 Jun 2001, MA 8051 (DEP0966); Plagiobothrys hispidulus (Greene) I.M. Johnst., locally common, moist to dry meadows, openings in Jeffrey pine - white fir forest, lodgepole pine forest, 30 Jul 2001, MA 8469 (DEP0853), 25 Jul 2001, MA 8381 (RM), 25 Jun 2001, MA 8238 (RM), 22 Jun 2001, MA 8097 (DEP0849, JEPS99817), 22 Jun 2001, MA 8096 (JEPS99738), 18 Jun 2001, MA 8042 (DEP0854); P. hispidus A. Gray, uncommon, post- burn coniferous forest in loose pumice; lodgepole pine forest, 3 Jul 2001, MA 8317 (DEP0841), 3 Jul 2001, MA 8322 (DEPO840, JEPS99751). Brassicaceae Bar bar ea orthoceras Ledeb., local, moist slope in mist of falls, streambanks, 25 Jun 2001, MA 8231 (DEP0867, JEPS99757, RM), 25 Jun 2001, MA 8235 (DEP01141, JEPS99859); Boechera divaricarpa (A. Nelson [pro. sp.]) A. Love & D. Love, common, moist sands and gravels, 3 Jul 2001, MA 8295 (DEP0864, JEPS99755), 18 Jun 2001, MA 8020 (DEP0856, JEPS99756); B. howellii (S. Watson) Windham & Al- Shehbaz, common, lodgepole pine forest, 18 Jun 2001, MA 8015 (DEP0861, JEPS99754); B. pinetorum (Ti- destr.) Windham & Al-Shehbaz, uncommon, rock outcrops, wet meadow; lodgepole pine forest, 25 Jul 2001, MA 8375 (DEP0858, JEPS99752, RM), 23 Jun 2001, MA 8122 (DEP01176, JEPS99842); B. rectissima (Greene) Al-Shehbaz, common, rock outcrops in 2014] ARNETT ET AL.: VASCULAR FLORA OF DEVILS POSTPILE 379 conifer forest, 7 Jun 1980, SCM 130 (DEPOI23), 15 Jun 1972 KAH 136 (DEP021); B. repanda (S. Watson) Al-Shehbaz, common, rock outcrops, conifer forest, 20 Jun 2001, MA 8047 (DEP0863, JEPS99818),18 Jun 2001, MA 8035 (DEP0862); B. retrofracta (Graham) A. L5ve & D. L5ve, common, rock outcrops, red fir forest, red fir - white fir forest, Jeffrey pine woodland; sagebrush scrub, 22 Jun 2001, MA 8111 (DEP0859), 20 Jun 2001, MA 8069 (DEPO860); Cardamine hreweri S. Watson, uncommon, mountain alder and willow thickets, 28 Jul 2001, MA 8432 (DEP0868, JEPS 99758); Descurainia incana (Bernh. ex Fisch. & C.A. Mey.) Dorn, uncommon, shaded Jeffrey pine - red fir forest, mountain alder and willow thickets, 226 Jun 2001, MA 8266 (DEPO870, JEPS99819), 3 Jun 2001, MA 8150 (DEP0869, RM); Dr aha albertina Greene, local, wet meadow; riparian, 23 Jun 2001, MA 8125 (DEP0871); Erysimum perenne (S. Watson ex Coville) Abrams, common, dry understory of coniferous forest, 22 Jim 2001, MA 8113 (DEP0872, JEPS99759); Lepidium densifJorum Schrad., uncommon, on landslide near river, 9 Jul 2001, MA 8484 (DEP0873); Nastur- tium officinale W.T. Aiton, locally common, stream courses, riparian, 22 Jul 2001, MA 8358 (DEP0878, JEPS99825); Phoenicaulis cheiranthoides Nutt., uncom- mon, basalt outcrops, 8 Jun 2001, MA 8472 (DEP0874, JEPS99739); Rorippa curvipes Greene var. curvipes, uncommon, wet meadows and gravel bars, 27 Jul 2001, DEPO.0036.12 (DEPO1208); R. curvisiliqua (Hook.) Bessey ex Britton, local, wet meadow; willow thickets, 25 Jul 2001, MA 8382 (DEP0877), 27 Jun 2001, MA 8279 (DEP0875, JEPS99750); Streptanthus tortuosus Kellogg, common, rocky to sandy soils in understory of coniferous forest, riparian forest, chaparral, 20 Jun 2001, MA 8048 (DEPO605, JEPS99740). Caprifoliaceae Lonicera conjugialis Kellogg, common, lodgepole pine forest, red fir - white fir forest, 3 Jul 2001, MA 8305 (DEPO607, RM), 18 Jun 2001, MA 8013 (DEPO606, JEPS99741); L. involiicrata (Richardson) Banks ex Spreng. var. involucrata, local, riparian areas, moist places in lodgepole pine forest, 3 Jul 2001, MA 8304 (DEPO608, JEPS); Symphoricarpos mollis Nutt., uncommon, Jeffrey pine - white fir forest, 28 Jul 2001, MA 8450 (DEPO609, JEPS); S. rotimdifoiius A. Gray var. rotundifolius, occasional, Jeffrey pine - white fir - red fir forest, sagebrush shrubland; bitter cherry shrubland, 24 Jun 2001, MA 8177 (DEPO610, JEPS99742). Caryophyllaceae Sagina saginoides (L.) H. Karst., locally common, wet meadow; riparian, 23 Jun 2001, MA 8132 (DEP0611, JEPS99744); Silene menziesii Hook., uncommon, black cottonwood forest, mountain alder thickets, 28 Jul 2001, DEPO.0018.06 (DEPO1210); *Spergularia rubra (L.) J.S. Presl & C. PresL, native to Europe, occasional, wet meadow and gravel bars, 27 Jun 2001, MA 8273 (DEP0612, JEPS99815); Stellaria longipes Goldie subsp. longipes, local, wet meadow and streambanks, 28 Jul 2001, MA 8442 (DEP0615), 23 Jun 2001, MA 8137 (DEP0614, JEPS99814); S', umbellata Turcz. ex Kar. & Kir., local, wet meadow; aspen grove, 23 Jun 2001 MA 8167 (DEP0616, JEPS99745), 23 Jun 2001, MA 8151 (DEP0617, RM). Cornaceae Cornus sericea L. subsp. sericea, locally common, moist places near river, 24 Jun 2001, MA 8188 (DEP0618, JEPS99747). Crassulaceae Sedum obtusatum A. Gray subsp. ohtusatum, uncom- mon, rock outcrops, 5 Jul 2001, MA 8329 (DEP0619, JEPS99748). Ericaceae Arctostaphylos nevadensis A. Gray, uncommon, Jeffrey pine - white fir forest, 20 Jun 2001, MA 8055 (DEP0919, JEPS99664); A. pa tula Greene, locally common, red fir - white fir forest, Jeffrey pine woodland; bitter cherry shrubland, 22 Jun 2001, MA 8091 (DEP0921, JEPS99665, RM); Rhododendron columbianum (Piper) Harmaja, uncommon, mesic lod- gepole pine forest, 3 Jul 2001, MA 8312 (DEP0922, JEPS99666); Pterospora andromedea Nutt., occasional, red fir - white fir forest, 26 Jul 2001, MA 8403 (DEP0923); Pyrola picta Sm., occasional, lodgepole pine forest, 25 Jun 2001, MA 8237 (DEP0924); Sar codes sanguinea Torr., occasional, lodgepole pine forest, Jeffrey pine forest, 18 Jun 2001, MA 8012 (DEPO930, JEPS99667). Fabaceae Acmispon americanus (Nutt.) Rydb. var. americanus, locally common, Jeffrey pine - white fir forest, lodgepole pine forest, moist slope in mist of falls, 28 Jul 2001, MA 8449 (DEP0935), 25 Jun 2001, MA 8227 (DEP0934, RM); Hosackia crassifolia Benth. var. crassifolia, locally common. Chaparral, 27 Jul 2001, MA 8417 (DEP0932, JEPS99668); H. oblongifolia Benth. var. oblongifolia, uncommon, moist area in Jeffrey pine - red fir forest, 27 Jul 2001, MA 8427 (DEP0933, JEPS99669); Lupinus albicaulis Douglas ex Hook., locally common, dry understory of Jeffrey pine - white fir - red fir forest, 28 Jul 2001, MA 8452 (DEP0938, RM), 24 Jun 2001, MA 8197 (DEP0937, JEPS99671); L. latifoliiis J. Agardh var. columbianus (A. A. Heller) C.P. Sm., locally common, moist drainages, aspen grove, 22 Jun 2001, MA 8102 (DEP0939, JEPS99821); L. lepidus Douglas ex Lindl. var. sellulus (Kellogg) Barneby, locally common, intermittent stream course and surrounding meadow, 23 Jun 2001, MA <5777 (DEPO940, JEPS99652); L. polyphyllus Lindl. var. burkei (S. Watson) C.L. Hitchc., common, aspen grove; black cottonwood forest, willow thickets, wet meadow; gravel bars, 27 Jun 2001, MA 8270 (DEP0942, JEPS99655), 26 Jun 2001, MA 8290 (DEP0944, JEPS99654), 23 Jun 2001, MA 8162 (DEP0941, RM), 20 Jun 2001, MA 8070 (DEP0943, JEPS99654, RM); Trifolium cyathiferum Lindl., locally common, moist seeps, 26 Jun 2001, MA 8246 (DEP0945, JEPS99811); T. longipes Nutt, subsp. hansenii (Greene) J.M. Gillett, locally common, wet meadow; riparian forest, willow thickets, 23 Jun 2001, MA 8166 (DEP0946, JEPS99656); T. monanthum A. 380 MADRONO [VoL 61 Gray subsp. monanthum, locally common, wet meadow; riparian forest, willow thickets, 23 Jun 2001, MA 8131 (DEP0948), 22 Jun 2001, MA 8093 (DEP0931, JEPS99657, RM); T. wormskioldii Lehm., local, wet meadow edge, 28 Jul 2001, MA 8436 (DEP0949, JEPS99658). Fagaceae Chrysolepis sempervirens (Kellogg) Hjelmq., uncom- mon, west-facing Jeffrey pine - white fir forest, 29 Jul 2001, MA 8463 (DEPO950, JEPS99659, RM); Quercus vaccinifolia Kellogg, locally common, Jeffrey pine - white fir - red fir forest, Jeffrey pine woodland; sagebrush shrubland, 9 Jul 2001, MA 8487 (DEP0951, JEPS99660), Gentianaceae Fraser a speciosa Douglas ex Griseb., uncommon, Jeffrey pine - white fir forest, 28 Jul 2001, MA 8429 (DEP0953, JEPS99810); Gentianopsis holopetala (A. Gray) litis, uncommon, wet meadow, 22 Aug 1977, JLM (DEP0517), 3 Aug 1972 KAH 200 (DEP033); G. simplex (A. Gray) litis, uncommon, wet meadow; mountain alder and willow thickets, 29 Jul 2001, MA 8460 (DEP0952, JEPS99661). Geraniaceae Geranium richardsonii Fisch. & Trautv., local, black cottonwood forest, aspen grove; mountain alder thick- et, 23 Jun 2001, MA 8161 (DEP0954, JEPS99642). Grossulariaceae Ribes cereum Douglas var. cereum, common, Jeffrey pine - white fir forest, lodgepole pine forest, 24 Jun 2001, MA 8199 (DEP0955, JEPS99938), 20 Jun 2001, MA 8054 (DEP0956, JEPS99643); R. inerme Rydb. var. inerme, common, riparian; willow thickets, 15 Jul 2001, MA 8339 pEP0957, JEPS99644); R. nevadense Kel- logg, occasional, riparian; red fir - white fir forest, Jeffrey pine - white fir forest, 15 Jul 2001, MA 8341 (DEPO960, JEPS99645), 24 Jun 2001, MA 8186 (DEP0959, RM); R. roezlii Regel var. roezlii, common, red fir - white fir forest, lodgepole pine forest, Jeffrey pine - white fir forest, mountain alder thickets, 20 Jun 2001, MA 8065 (DEP0961, JEPS99646), 03 Jul 1980, SM 117 (DEP0779), 15 Jun 1972 KAH 137 (DEP0958); R. viscosissimum Pursh, occasional, moist areas in conifer- ous forest, 27 Jul 2001, DEF0.0026J7 (DEPO1207). Hypericaceae Hypericum anagalloides Cham. & Schldl, locally common, riparian; wet meadow, 05 Jul 2001, MA 8327 (DEP0968, RM), 23 Jun 2001, MA 8130 (DEP0967); H. scouleri Hook., uncommon, river bank, 27 Aug 1983 JLM (DEP0144). Lamiaceae Agastache urticifoiia (Benth.) Kuntze, occasional, mountain alder and willow thickets, 3 Jul 2001, MA 8289 (DEP0987, JEPS99641), 26 Jun 2001, MA 8263 (DEP0988, JEPS); Monardella odoratissima Benth. subsp. pallida (A. Heller) Epling, common, bitter cherry shrubland; open slopes, 24 Jun 2001, MA 8195 (DEP0989, JEPS99622); Stachys aibens A. Gray, locally common, riparian forest, mountain alder and willow thickets, 22 Jul 2001, MA 8359 (DEPO990, JEPS99623). Loasaceae Mentzelia dispersa S. Watson, common, dry open slopes, Jeffrey pine - red fir forest, 22 Jun 2001, MA 8090 (DEPO1007, JEPS99631, RM), 18 Jun 2001, MA 8034 (DEPO1006). Montiaceae Calyptridium monospermum Greene, locally common, sandy soils in lodgepole pine forest, Jeffrey pine - white fir - red fir forest, 24 Jun 2001, MA 8221 (DEP01118, JEPS99885, RM); C umbellatum (Torr.) Greene, local, exposed dry pumice, 2 Jul 1980, SCM 108 (DEP0212), 30 Jun 1972 KAH 160 (DEP051); Claytonia rubra (Howell) Tidestr. subsp. rubra, uncommon, moist slope in mist of falls, 25 Jun 2001, MA 8224 (DEP01119, JEPS99886); Lewisia nevadensis (A. Gray) B.L. Rob., uncommon, wet meadow, 19 Jul 1980, SCM 140 (DEP0215), 23 Jun 1980, SCM 102 (DEP0214); K triphylla (S. Watson) B.L. Rob., uncommon, granitic seep, 18 Jun 2001, MA 8031 (DEPO1120); Montia chamissoi (Spreng.) Greene, uncommon, mountain alder and willow thickets, granitic seeps, meadow edge, 03 Jul 2001, MA 8283 (DEP01123, JEPS99887), 23 Jun 2001, MA 8138 (DEP01121, RM), 23 Jun 2001, MA 8136 (DEP01122, JEPS99888), 18 Jun 2001, MA 8032 (DEP01124). Onagraceae Chamerion angustifolium (L.) Holub subsp. circum- vagum (Mosquin) Hoch, locally common, moist open places, gravel bars, mountain alder thickets, 15 Jul 2001, MA 8335 (DEPOlOlO, JEPS99614), 4 Sep 1980, SM 224 (DEPO701), 25 Jul 1972 KH 175 (DEP0551); Circaea alpina L. subsp. pacifica (Asch. & Magnus) Raven, locally common, moist swales, mountain alder thickets, 24 Jun 2001, MA 8202 (DEPO1008, JEPS99612); Epilobium anagallidifolium Lam,, local, moist seeps, mountain alder thickets, 20 Jun 2001, MA 8052 (DEPO1009, JEPS99613); E. brachycarpum C. Presl, locally common, in post-burned coniferous forest, dry sands, gravels in Jeffrey pine forest with scattered western jumper; slope of flaky rhyodacite below falls, 28 Jul 2001, MA 8446 (DEPO1015, JEPS99615), 27 Jul 2001, MA 8415 (DEPO1012, RM), 24 Jun 2001, MA 8211 (DEPOll, JEPS99616, RM); E. canum (Greene) Raven subsp. latifolium (Hook.) Raven, uncommon, exposed cliffs and rocky outcrops, 29 Jul 2001, MA 8458 (DEPO1016, JEPS99617); E. ciliatum Raf. subsp. ciliatum, locally common, wet meadow edge, willow thickets, 25 Jul 2001, MA 8379 (DEPO1017, JEPS99618), 26 Jun 2001, MA 8264 (DEPO1018); E ciliatum Raf. subsp. glandulosum (Lehm.) P. Hoch & Raven, common, mountain alder thickets, moist swales in coniferous forest, 28 Jul 2001, MA 8447 (DEPO1021, JEPS99621, RM), 22 Jul 2001, MA 8360 (DEPO1019, JEPS99619, RM), 20 Jun 2001, MA 8062 (DEPO1020, 2014] ARNETT ET AL.: VASCULAR FLORA OF DEVILS POSTPILE 381 JEPS99620); E. glaberrimum Barbey subsp. fastigiatum (Nutt.) P. Hoch & Raven, local, moist areas, red fir forest, 3 Jul 2001, MA 8311 (DEPO1022, JEPS99602); E. glaberrimum subsp. glaberrimum, locally common, on landslide near river, 26 Jun 2001, MA 8265 (DEPO1023, JEPS99603); E. halUanum Hausskn., locally common, wet meadow, streambanks, 26 Jul 2001, MA 8408 (DEPO1031, RM), 25 Jul 2001, MA 8378 (DEPO1024, JEPS99604), 23 Jun 2001, MA 8141 (DEPO1026), 23 Jun 2001, MA 8118 (DEPO1027, RM), 22 Jun imiMA 8098 (DEPO1028), 22 Jun 2001, MA 8101 (DEPO1030, JEPS99937), 20 Jun 2001, MA 8063 (DEPO1032); E. hornemannii Reichb. subsp. hornemannii, uncommon, moist riverbank, 4 Aug 1980, SCM 191 (DEPO704); E. lactiflorum Hausskn., uncommon, moist riverbank, 28 Jul 2001, DEPO. 0018.19 (DEPO 1202); Gayophytum decipiens F.H. Lewis & JSzweyk., occasional, lodgepole pine forest, Jeffrey pine - white fir - red fir forest, black cottonwood forest, mountain alder thickets, 22 Jun 2001, MA 8115 (DEPO 1034, JEPS99605, RM); G. diffusum Torr. & A. Gray subsp. parviflorum F.H. Lewis & JSzweyk., abundant, red fir - white fir forest, mountain alder thickets, lodgepole pine forest, 23 Jun 2001, MA 8142 (DEPO 1037), 22 Jun 2001, MA 8099 (DEPO 103 5, JEPS99936), 16 Sep 1980, SCM 247 (DEPO 158), 19 Jul 1980, SCM 145 (DEP0157), 25 Jul 1977, JLM (DEP0698); G. heterozygum F.H. Lewis & JSzweyk., local, lodgepole pine forest, red fir forest, Jeffrey pine forest, 22 Jul 2001, MA 8353 (DEPO 1038, JEPS99606), 22 Jul 2001, MA 8354 (DEPO 1039, RM); G. humile A.L. Juss, common, lodgepole pine forest, basalt outcrop; white fir forest, 26 Jun 2001, MA 8247 (DEPO 1040, RM), 22 Jun 2001, MA 8100 (DEPO 1041, JEPS99607),18 Jun 2001, MA 8023 (DEPO 1042); G. racemosum Torr. & A. Gray, common, margin of sag pond, 27 Jul 2001, MA 8418 (DEPO1043, JEPS99608). Orobanchaceae Castilleja applegatei Fernald subsp. pinetorum (Fer- nald) T.I. Chuang & Heckard, common, Jeffrey pine forest, shrubby, rocky hillsides, Jeffrey pine - red fir forest, 24 Jun 2001, MA 8187 (DEP0652), 23 Jun 2001, MA 8152 (DEP0653, JEPS99829); C miniata Hook, subsp. miniata, occasional, mountain alder and willow thickets, 3 Jul 2001, MA 8285 (DEP0654, JEPS99830), 25 Jul 1980, SCM 166 (DEP0783), 25 Jun 1972 KAH 155 (DEP0969); C. peirsonii Eastw., local, sagebrush scrub, 29 Jul 2001, DEPO. 0062. 02 (DEPO 1200); C tenuis (Heller) T.I. Chuang & Heckard, local, aspen grove; basalt outcrop and associated seep in white fir forest, 26 Jun 2001, MA 8250 (DEPOl 186, JEPS99831), 23 Jun 2001, MA 8176 (DEPOl 185, RM); Orobanche fasciculata Nutt., uncommon, loose pumice in burned forest, 3 Jul 2001, MA 8320 (DEP0637, JEPS99930), 29 Aug 1980, JLM (DEPO 165) ; Pedicularis attollens A. Gray, uncommon, wet meadow, 25 Jul 1980, SCM 170 (DEP0253), 28 Jul 1972 KAH 195 (DEP074); P. semibarbata A. Gray, common, red fir - white fir forest, Jeffrey pine - white fir - red fir forest, mountain hemlock forest, 22 Jun 2001, MA 8084 (DEPO670, JEPS99694). Phrymaceae Mimulus breweri (Greene) Coville, locally common, white fir forest, moist drainage near buttresses, 26 Jun 2001, MA 8239 (DEPOl 191, JEPS99834), 22 Jun 2001, MA 8092 (DEPOl 190); M. guttatus Fisch. ex DC., common, wet places, aspen grove, mountain alder thickets, black cottonwood forest, moist slope in mist of falls, basalt outcrop and associated seep in white fir forest, 25 Jun 2001, MA 8230 (DEPOl 193), 26 Jun 2001, MA 8240 (DEPOl 192, JEPS99835), 23 Jun 2001, MA 8158 (DEPOl 194, RM), 3 Jul 1980, SCM 112 (DEP0789); *>M. laciniatus A. Gray, local, seeps on granite outcrops, 3 Jul 2001, MA 8309 (DEP0662, RM), 18 Jun 2001, MA 8026 (DEP0661, JEPS99836); M. leptaleus A. Gray, local, moist pumice slope, 4 Jul 1980, SCM 123 (DEP0788), 15 Jul 1978 JLM (DEP0786); M. lewisii Pursh, local, streambanks and seeps, mountain alder thickets, black cottonwood forest, 20 Jun 2001, MA 8059 (DEP0663, JEPS99837), 1 Jul 1977, JLM (DEP0786); M. moschatus Lindl., locally common, seeps and streambanks, 26 Jun 2001, MA 8241 (DEP0668, RM), 23 Jun 2001, MA 8159 (DEP0664, JEPS99838), 20 Jun 2001, MA 8061 (DEP0667), 20 Jul 1980, SCM 158 (DEPO790); M. pilosus (Benth.) S. Watson, uncommon, dry east-facing slope, 27 Aug 1980, SCM and JLM 212 (DEP0791); M. primuloides Benth, locally common, wet meadows, willow thickets, 23 Jun 2001, MA 8128 (DEP0669, JEPS99692), 3 Jul 1980, SCM 111 (DEP0792), 18 Jun 1972, KAH 151 (DEP0991); M tilingii Regel, locally common, seeps and moist outcrops, 20 Jun 2001, MA 8060 (DEPOl 195, JEPS99693, RM), 20 Jul 1980, SCM 165 (DEP0793), 18 Jun 1972 KAH 152 (DEPO1002). Plantaginaceae Collinsia parviflora Douglas ex Lindl., locally com- mon, moist shady places, 23 Jun 2001, MA 8127 (DEPOl 187, JEPS99832), 1 Jul 1980, SCM 105 (DEP0237), 18 Jun 1972 KAH 146 (DEP067); C. torreyi A. Gray var. wrightii (S. Watson) I.M. Johnst., locally common, conifer forest, moist bench of large granitic outcrop, 20 Jun 2001, MA 8057 (DEPOl 188); Penstemon azureus Benth. var. azureus, local, granite outcrop in moist shady pocket, Jeffrey pine forest, 24 Jun 2001, MA 8213 (DEP0672, JEPS99696, RM), 18 Jun 2001, MA 8043 (DEP0671, JEPS99695), 26 Jul 1972, KAH 193 (DEPO 10 14); P. heterodoxus A. Gray var. cephalophorus (Greene) N.H. Holmgren, locally common, wet meadows and drainages, lodgepole pine forest, 24 Jun 2001, MA 8203 (DEP0673, JEPS99697); P. laetus A. Gray var. laetus, occasional, on pumice and basalt slope; basalt cliffs, 19 Jul 1980, SCM 146 (DEPO800), 27 Jul 1977, JLM (DEP0799); P. new- berryi A. Gray var. newberryi, common, red fir forest, Jeffrey pine - white fir - red fir forest, red fir - western white pine forest, 23 Jun 2001, MA 8147 (DEP0674, JEPS), 3 Jul 1980, SCM 113 (DEPO801), 30 Jun 1972, KAH 161 (DEPO 1036); P. rostriflorus A. Gray, occasional, rock outcrops, Jeffrey pine - red fir forest, bitter cherry shrubland, 22 Jun 2001, MA 8082 (DEP0675, JEPS99698); P. rydbergii A. Nelson var. oreocharis (Greene) N.H. Holmgren, locally common, wet meadow; willow thickets, 27 Jun 2001, MA 8269 (DEP0676, JEPS99699); Veronica americana (Raf.) Schwein. ex Benth., locally common, mountain alder and willow thickets, black cottonwood forest, aspen grove, 23 Jun 2001, MA 8134 (DEP0678, JEPS99700), 28 Aug 1977, JLM (DEP0535); V. serpyllifolia L. subsp. humifusa (Dickson) Syme, locally common, gravel beds, willow thickets, 23 Jun 2001, MA 8135 382 MADRONO [Vol. 61 (DEPO680, RM), 20 Juii 2001, MA 8058 (DEP0679, JEPS99701), 28 Aug 1977, JLM (DEP0536); K wormskjoldii Roem. & Schult., locally common, sand- bars and wet meadows, 26 Jun 1977, JLM and JN (DEP0534), 14 Jim 1972, KAH 132 (DEPO1052). Polemoiiiaceae Allophylhim gilioides (Bentli.) A.D. Grant & V.E. Grant subsp. violacenm (A. Heller) A.G. Day, uncom- mon, on landslide near river, 26 Jun 2001, MA 8267 (DEP01090, JEPS99905, RM); A. integrifolium (Brand) A.D. Grant & V.E. Grant, locally common, basalt outcrop and associated seep in white fir forest, red fir - white fir forest, 26 Jul 2001, MA 8404 (DEPO1091, JEPS99906), 26 Jun 2001, MA 8253 (DEPO1092, JEPS99889, RM); Collomia linearis Nutt., locally abundant, moist understory of red fir - white fir forest, mountain alder and willow thickets, wet meadow, 3 Jul 2001, MA 8296 (DEPO1094), 23 Jun 2001, MA 8140 (DEPO1093, JEPS99890, RM), 4 Aug 1980, SLM 194 (DEP0193); C tinctoria Kellogg, common, moist slope in mist of falls, mountain alder thickets, aspen grove; black cottonwood forest, 25 Jun 2001, MA 8326 (DEPO1095); Ipomopsis aggregcita (Pursh) V.E. Grant subsp. aggregata, locally common, Jeffrey pine - white fir - red fir forest, 28 Jul 2001, MA 8451 (DEPO1097, JEPS99892), 5 Aug 1980, SCM 201 (DEP0196), 9 Jul 1972 KAH 169 (DEP045); Navarretia capiUaris (Kel- logg) Kuntze, uncommon, granitic seep, 18 Jun 2001, MA 8044 (DEPO1096, JEPS99891); N. leptalea (A. Gray) L.A. Johnson, local, moist pumice, 20 Jul 1980, SCM 162 (DEP0195), 15 Jul 1978 JLM (DEP0194); Linanthus pimgens (Torr.) J.M. Porter & L.A. Johnson, common, rock outcrops, north-facing shaded slope in Jeffrey pine - red fir forest, 23 Jun 2001, MA 8148 (DEPOllOO, JEPS99893); Leptosiphon ciliatus (Benth.) Jeps., common, loose pumice, 22 Jun 2001, MA 8109 (DEPOllOl, JEPS99894), 23 Jun 1980, SCM 103 (DEP0199), 18 Jun 1972, KAH 145 (DEP047); Microsteris gracilis (Hook.) Greene, occasional, moist areas, basalt outcrop and associated seep in white fir forest, 26 Jun 2001, MA 8254 (DEPO1102, JEPS99895, RM), 3 Jun 2001, MA 8133 (DEPO1103); Polemonium occidentale Greene, uncommon, mountain alder and willow thickets, 28 Jul 2001, MA 8430 (DEPO1104), 25 Jul 1980, SCM 182 (DEPO200). Polygonaceae Bistorta bistortoides (Pursh) Small, locally common, wet meadows, gravel bars along river, 27 Jun 2001, MA 8280 (DEPOl 111, JEPS99881), 25 Jul 1980, SCM 171 (DEP0753), 4 Jul 1980 SCM 124 (DEP0752), 20 Jul 1972, KAH 174 (DEP0599); Eriogomim nudum Benth. var. dediictum (Greene) Jeps., abundant, mesic lodge- pole pine forest, red fir forest, black cottonwood forest, mountain hemlock forest, mountain alder thickets, 20 Jun 2001, MA 8049 (DEPOl 105, JEPS99896); E. nudum var. nudum, abundant, dry to moist pumice, 4 Sep 1980, SCM 223 (DEP0748), 20 Jul 1980, SCM 153 (DEP0747), 30 Jun 1972, KAH 163 (DEP0822); E. spergulinum A. Gray var. reddingianum (M.E. Jones) J.T. Howell, abundant, rocky basalt outcrops, lodge- pole pine forest, black cottonwood forest, 18 Jun 2001, MA 8014 (DEPOl 106, JEPS99897), 2 Aug 1977, JLM (DEP0749); E. umhellatum Torr. vm . furcosum Reveal, uncommon, dry sands and gravels in red fir - lodgepole pine forest, 22 Jul 2001, MA 8349 (DEPOl 107, JEPS99898, RM); E. umbellatum var. nevadense Gand., uncommon, ledges and cracks in basalt outcrop, 16 Sep 1980, SCM 245 (DEPO750), 9 Jul 1972, KAH 171 (DEP0587); E. wrightii Torr. ex Benth. var. subscapo- sum S. Watson, occasional, volcanic sands and gravels in post-burned forest, granite outcrops, 27 Jul 2001, MA 8413 (DEPOl 108, JEPS99879, RM); Oxyria digyna (L.) Hill, uncommon, Jeffrey pine - white fir - red fir forest with lodgepole and western white pine, 24 Jun 2001, MA 8181 (DEPOl 109, JEPS99880), 16 Sep 1980, SCM 237 (DEPO205); Polygonum douglasii Greene, locally common, basalt outcrop and associated seep in white fir forest, riverbank; mountain alder thickets, 26 Jun 2001, MA 8251 (DEPOl 112); P. polygaloides Meissner subsp. kelloggii (Greene) J. Hickman, locally common, wet meadows and gravel bars along river, 27 Jun 2001, MA 8282 (DEPOl 116, JEPS99883); P. sawatchense Small subsp, sawatchense, locally common, wet meadow, 26 Jul 2001, MA 8407 (DEPOl 113, JEPS99882), 26 Jun 2001, MA 8252 (DEPOl 115, RM), 23 Jun 2001, MA 8143 (DEPO 1114); Rumex salicifolius J. A. Weinm., local, wet meadow; mountain alder and willow thickets, black cottonwood forest, 28 Jul 2001, MA 8439 (DEPOl 117, JEPS99884); R. triangidivalvis (Danser) Rech. f., occasional, wet meadow, 19 Jul 1980, SCM 139 (DEP021 1), 30 Aug 1977, JLM (DEPO210). Primulaceae Primula jeffreyi (Van Houtte) Mast & Reveal, locally abundant, wet meadows, mountain alder and willow thickets, 23 Jun 2001, MA 8126 (DEPOl 126, JEPS 99870, RM), 4 Jul 1980, SCM 121 (DEP0216), 14 Jun 1972, KAH 131 (DEP052). Ranunculaceae Aeon i turn columhianum Nutt., locally common, aspen grove; wet meadow, 23 Jun 2001, MA 8163 (DEPOl 130), 23 Jun 2001, MA 8168 (DEPOl 132, JEPS99871); Aquilegia formosa Fisch., local, aspen grove; mountain alder thicket, black cottonwood forest, 23 Jun 2001, MA 8160 (DEPOl 133, JEPS99872), 25 Jun 1972, KAH 159 (DEP053); Delphinium glaucum S. Watson, locally common, mountain alder thickets, 29 Jul 2001, MA 8456 (DEPOl 134, JEPS99873), 9 Sep 1980, SCM 231 (DEP0218), 26 Jul 1972, KAH 190 (DEP0865); D. gracilentum Greene, occasional, open pumice soils, in disturbed area along trail in pumice- organic soil, 15 Jul 1978, JLM (DEP0219), 18 Jun 197,2 KAH 143 (DEP055); D. nuttallianum Walp., common, basalt outcrop and associated seep in white fir forest, 26 Jun 2001, MA 8242 (DEPOl 135, JEPS99874); D. polycladon Eastw., uncommon, streanibanks and wet drainages, 24 Jun 2001, MA 8204 (DEPOl 136, JEPS99875); Ranunculus alismifolius Benth. var. alis- mellus A. Gray, locally common, wet meadow, 14 Jun 1972, KAH 130 (DEP056); R. cymbalaria Pursh, local, edge of spring near river in wet meadow, 28 Jul 2001, MA 8443 (DEPOl 137, JEPS99876), 4 Jul 1977, JLM (DEPO220); Thalictriim fendleri A. Gray var. fendleri, common, aspen grove; black cottonwood forest, mountain alder thickets, 23 Jun 2001, MA 8169 (DEPOl 138, JEPS99877, RM), 6 Jul 1980, SCM 126 2014] ARNETT ET AL.: VASCULAR FLORA OF DEVILS POSTPILE 383 (DEP0221), 3 Jul 1972, KAH 168 (DEP0623); T. sparsijlorum Fisch. & C.A. Mey., local, mountain alder and willow thickets, 3 Jul 2001, MA 8294 (DEP01139, JEPS99878). Rhamnaceae Ceanothus cordulatus Kellogg, locally common, red fir - white fir forest, bitter cherry shrubland, 22 Jun 2001, MA 8116 (DEPO1140, JEPS99860, RM); Fran- gula rubra (Greene) Grubov, uncommon, east-facing slope in loose pumice on granite capped with basalt, 25 Jun 2001, MA 8235 (DEP01141, JEPS99859). Rosaceae Amelanchier utahensis Koehne, local, red fir - white fir forest, sagebrush scrub, 26 Jul 2001 DEPO.0060.10 (DEP01214); Drymocallis lactea (Greene) Rydb. var. lactea, common, moist, often rocky places, mountain alder and willow thickets, 22 Jun 2001, MA 8103 (DEP01149, JEPS99865), 3 Jul 1972, KAH 167 (DEP061); Geum macrophylhmi Willd., local, wet meadows, streambanks, mountain alder and willow thickets, 3 Jul 2001, MA 8288 (DEPOl 144, JEPS99862), 25 Jul 1980, SCM 181 (DEP0222); Holodiscus discolor (Pursh) Maxim var. microphyllus (Rydb.) Jeps., com- mon, Jeffrey pine - red fir forest, Jeffrey pine woodland; red fir - western white pine forest, 24 Jun 2001, MA 8212 (DEPOl 145), 16 Sep 1980, SCM 243 (DEP0224), 7 Aug 1980, SCM 202 (DEP0223), 3 Aug 1972 KAH 196 (DEP058); Horkelia fusca Lindl. var. parviflora (Hook. & Arn.) Wawra, locally abundant, lodgepole pine woodland; dry meadow edges, 22 Jul 2001, MA 8355 (DEPOl 147, RM), 23 Jun 2001, MA 8120 (DEPOl 146, RM), 19 Jul 1980, SCM 138 (DEP0225), 20 Jul 1972 KAH 173 (DEP059); Ivesia santolinoides A. Gray, local, sandy granite and pumice ledges, 24 Jun 2001, MA 8200 (DEPOl 148, JEPS99864), 15 Aug 1977, JLM (DEP0226); Potentilla glaucophylla Lehm. var. glauco- phylla, uncommon, wet meadow, 26 Jul 2001, DEPO. 0004.01 (DEPO1206); P. gracilis Hook. var. fastigiata (Nutt.) S. Watson, locally common, wet meadow; aspen grove; willow thickets, 23 Jun 2001, MA 8119 (DEPOl 151), 23 Jun 2001, MA 8164 (DEPOl 150, JEPS99866); Prunus emarginata (Hook.) Walp., com- mon, rocky slopes, chaparral, 27 Jul 2001, MA 8416 (DEPOl 155, JEPS99868), 15 Jul 2001, MA 8338 (DEPO 11 54, RM), 22 Jun 2001, MA 8085 (DEPOl 152, JEPS99867); Rosa woodsii Lindl. subsp. gratissima (S. Watson) Roy L. Taylor & MacBryde, uncommon, slope of flaky rhyodacite below falls, 28 Jul 2001, MA 8444 (DEPOl 156, JEPS99849); Sorbus californica Greene, local, moist conifer forest, 24 Jun 2001, MA 8179 (DEPOl 157, JEPS99850); Spiraea splendens K. Koch, occasional, moist rocky places in conifer forest, 26 Jun 2001, MA 8268 (DEPOl 158, JEPS99851). Rubiaceae Galium aparine L, common, moist slope in mist of falls, 25 Jun 2001, MA 8229 (DEPOl 159, JEPS99852); G. bifolium S. Watson, locally common, moist gravels in conifer forest, basalt outcrop and associated seep in white fir forest, 26 Jun 2001, MA 8244 (DEPOl 161), 22 Jun 2001 MA 8094 (DEPOl 160, JEPS99853, RM); G. trifidum L. subsp. subbiflorum (Wiegand) Puff, local. wet meadow; willow thickets, 28 Jul 2001, MA 8441 (DEPOl 163), 23 Jun 2001, MA 8139 (DEPOl 162, JEPS99854); G. triflorum Michx., locally abundant, lodgepole pine forest, wet meadow; mountain alder and willow thickets, 3 Jul 2001, MA 8284 (DEPOl 165, JEPS99855), 20 Jun 2001, MA 8066 (DEPOl 166, JEPS99856, RM), 18 Jun 2001, MA 8030 (DEPOl 167, JEPS99857); Kelloggia galioides Ton*., common, red fir forest, lodgepole pine forest, black cottonwood forest, mountain alder thickets, 19 Jul 1980, SCM 151 (DEP0235), 18 Jun 2001, MA 8029 (DEPOl 168, JEPS99858), 27 Jul 1977, JLM (DEP0234), 26 Jul 1972, KAH 192 (DEPO60). Salicaceae Populus tremiiloides Michx., uncommon, stream terraces, sagebrush scrub, 23 Jun 2001, MA 8171 (DEPOl 170, JEPS99840), 12 Sep 1972 KAH 206 (DEP0936); P. trichocarpa Hook., occasional, riparian; alluvial terrace, 15 Jul 2001, MA 8340 (DEPOl 169, JEPS99839); Salix jepsonii C. Schneider, occasional, Riverbank; moist drainage in Jeffrey pine woodland, 18 Jun 2001, MA 8040 (DEPOl 171); S. lasiandra Benth. var. lasiandra, occasional, wet meadows and seeps, 23 Jun 2001, MA 8122 (DEPOl 176, JEPS99842), 18 Jun 2001, MA 8037 (DEPOl 174); 5’. lemmonii Bebb, common, wet meadow; forming thickets on alluvial terraces, 23 Jun 2001, MA 8157 (DEPOl 172, RM), 18 Jun 2001, MA 8017 (DEPOl 173, JEPS99841); 5. scouleriana Hook., occasional, dry conifer forest, 27 Jul 2001, MA 8421 (DEPOl 177, JEPS99843, RM), 20 Jun 2001, MA 8056 (DEPOl 178, JEPS99844). Saxifragaceae Heuchera rubescens Torr., locally common, dry rocky areas in Jeffrey pine - red fir forest, Jeffrey pine - white fir - red fir forest, bitter cherry shrubland, 23 Jun 2001M^ 8149 (DEPOl 179, JEPS99845), 8 Aug 1980, SCM 204 (DEP0231); Litliophragma glabrum Nutt., local, basalt outcrop and associated seep in white fir forest, lodgepole pine forest, 26 Jun 2001, MA 8248 (DEPOl 180, JEPS99846), 18 Jun 2001, MA 8022 (DEPOl 181); Micranthes nidifica (Greene) Small, locally common, granitic seep; moist slope in mist of falls, 25 Jun 2001, MA 8226 (DEPO650, JEPS99848), 18 Jun 2001, MA 8024 (DEPOl 183), 18 Jun 2001, MA 8027 (DEPOl 184, RM); M. odontoloma (Piper) A. Heller, uncommon, moist bank of basalt columns associated with Arnica mollis and Perideridia parishii, 12 Aug 1977, JLM (DEPO780); Pectiantia breweri (A. Gray) Rydb., uncommon, mountain alder thickets, 30 Jul 2001, MA 8470 (DEPOl 182, JEPS99847). Solanaceae Solanum umbelliferum A. Gray, locally common. South-facing volcanic rock outcrops, bitter cherry shrubland, 22 Jun 2001, MA 8088 (DEP0681, JEPS99812, RM). Urticaceae Urtica dioica L. subsp. holosericea (Nutt.) Thorne, local, loose pumice over sandy soil, 26 Jun 2001, MA 8261 (DEP0684, JEPS99683). 384 MADRONO [Vol. 61 Valerianaceae Valeriana californica A. Heller, local, loose pumice over sandy soil, 24 Jun 2001, MA 8196 (DEP0685, JEPS99684). Violaceae Viola macloskeyi F. Lloyd, occasional, in wet area along trail; willow thickets, 27 Jul 1972, KAH 260 (DEP076). Viscaceae Arceuthobium americanum Engelm., locally common, lodgepole pine, 15 Jul 2001 MA 8347 (DEP0686, JEPS 99685). Monocots Alliaceae Allium validum S. Watson, local, wet meadow; stream courses, mountain alder and willow thickets, 3 Jul 2001, MA 8315 (DEP0992, JEPS99624). Cyperaceae Carex abrupta Mack., locally common, mountain alder and willow thickets, 30 Jul 2001, MA 8466 (DEP0626, JEPS99721), 22 Jul 2001, 8356 (DEPO630, RM), 15 Jul 2001, MA 8342 (DEP0629, JEPS99720); C. athrostachya Olney, locally common, wet meadow; mountain alder and willow thickets, sag pond in bench of rolling granite hill, 25 Jul 2001, MA 8369 (DEP0632, JEPS99725, RM), 25 Jul 2001, MA 8370 (DEP0633, JEPS99724), 25 Jul 2001, MA 8384 (DEP0625, JEPS99723), 25 Jul 2001, MA 8371 (DEP0631, RM), 15 Jul 2001, MA 8345 (DEP0628, JEPS99722); C heteroneura W. Boott, locally common, wet meadow; mesic areas in lodgepole pine forest, mountain alder thickets, 25 Jul 2001, MA 8372 (DEP0635, JEPS99726, RM), 22 Jun 2001, MA 8108 (DEP0879, JEPS99727, RM), 18 Jun 2001, MA 8019 (DEP0636, JEPS99728); C hoodii Boott, locally common, mountain alder and willow thickets, 3 Jul 2001, MA 8299 (DEP0881, RM), 24 Jun 2001, MA 8216 (DEPO880, RM); C integra Mack., uncommon, wet meadow; mountain alder and willow thickets, 3 Jul 2001, MA 8300 (DEP0627, JEPS99711), 23 Jun 2001, MA 8121 (DEP0622, JEPS99710); C. jonesii L.H. Bailey, uncommon, inter- mittent stream channel and surrounding meadow, 24 Jun 2001, MA 8222 (DEP0887), 23 Jun 2001, MA 8144 (DEP0882, JEPS99712, RM); C. lenticularis Michx. var. impressa (L.H. Bailey) L.A. Standi., locally common, wet sands along river; C. lenticularis Michx. var. lipocarpa (Holm.) L.A. Standi., locally common, by water’s edge, mountain alder and willow thickets, sag pond margins, 27 Jul 2001, MA 8420 (DEP0889, JEPS99823, JEPS99827), 3 Jul 2001, MA 8298 (DEPO890, JEPS99688), 3 Jul 2001, MA 8297 (DEP0891, RM), 18 Jun 2001, MA 8036 (DEP0888); C. leporinella Mack., local, mountain alder thickets, 27 Jul 2001, D EP 0.0039 J 4 (DEP01196); C. mariposana L.H. Bailey ex Mack., local, moist drainages, willow thickets, Jun 22 2001, MA 8107 (DEP0892, JEPS99824, JEPS99826, RM); C. microptera Mack., local, black cottonwood forest, willow thickets, 29 Jul 2001, DEPO.0019.21 (DEP01197); C. multicostata Mack., common, lodgepole pine forest, red fir - white fir forest, mountain alder thickets, 30 Jul 2001, MA 8471 (DEP0621, JEPS99691), 25 Jul 2001, MA 8395 (DEP0893, JEPS99689, RM), 25 Jul 2001, MA 8398 (DEP0894, JEPS99690, RM); C. nebrascensis Dewey, locally common, wet meadow, 30 Jul 2001, MA 8467 (DEP0895, JEPS99672, RM), 18 Jun 2001, MA 8016 (DEP0896, JEPS99673); C. nervina L.H. Bailey, locally common, moist to wet places, 24 Jun 2001, MA 8206 (DEP0899, JEPS99674), 18 Jun 2001, MA 8038 (DEPO900); C. pellita Muhl. ex Willd., locally com- mon, aspen grove; mountain alder and willow thickets, 30 Jul 2001, MA 8464 (DEP0883, JEPS99686, RM), 22 Jul 2001, MA 8357 (DEP0884, JEPS99687, RM), 23 Jun 2001, MA 8175 (DEP0885, RM); C rossii Boott, locally abundant, dry conifer forest, wet meadow; mountain alder thickets, 24 Jun 2001, MA 8208 (DEPO902, JEPS99676), 18 Jun 2001, MA 8021 (DEPO901, JEPS99675, RM),; C subfusca W. Boott, common, seasonally moist areas, mountain alder thickets, wet meadow, 25 Jul 2001, MA 8389 (DEPO904, JEPS99678, RM), 24 Jun 2001, MA 8205 (DEPO903, JEPS99677); C. utriculata Boott, locally abundant, wet meadow; willow thickets, 23 Jun 2001, MA 8155 (DEPO905, JEPS99822, JEPS99828); C. vesicaria L., locally abundant, sag pond in bench of rolling granite hills, wet meadow, 26 Jul 2001, DEPO.0024.02 (DEP01198), 25 Jul 2001, MA 8388 (DEPO906, RM), 15 Jul 2001, MA 8346 (DEPO907, JEPS99679); C. whitneyi Olney, uncommon, dry sandy areas in red fir - white fir forest, 28 Jul 2001, DEPO.0038.13 (DEP01199); C. squarrosus L., uncom- mon, basalt outcrop and associated seep in white fir forest, 26 Jun 2001, MA 8257 (DEPO908, JEPS99680); Eleocharis bella (Piper) Svenson, locally common, wet meadow in saturated soil, 25 Jul 2001, MA 8380 (DEPO910, JEPS99681); E. macrostachya Britton, locally common, wet meadow in saturated soil and muck, 25 Jul 2001, MA 8387 (DEP0911, JEPS99662, RM); Scirpus microcarpus J. Presl & C. Presl, locally common, wet meadow; mountain alder and willow thickets, 29 Jul 2001, MA 8462 (DEP0912), 23 Jun 2001, MA 8156 (DEP0624, JEPS99663). Iridaceae Sisyrinchium idahoense E.P. Bicknell var. occidentale (E.P. Bicknell) Douglass M. Hend., uncommon, 8 Sep 1980, SC M 208 (DEP0145). Juncaceae Juncus balticus Willd., locally common, granite seep near river in loamy sand; wet meadow, 15 Jul 2001, MA 8334 (DEPO970, JEPS99633); J. drummondii E. Mey., locally common, moist rocky places in Jeffrey pine - white fir - red fir forest, 20 Jun 2001, MA 8046 (DEP0971, JEPS99634, RM); J. nevadensis S. Watson, common, wet meadow; mountain alder and willow thickets, alder grove; black cottonwood forest, 25 Jul 2001, MA 8383 (DEP0976, JEPS99637), 3 Jul 2001, MA 8293 (DEP0973, JEPS99638), 22 Jul 2001, MA 8362 (DEP0975, RM), 20 Jun 2001, MA 8074 (DEP0972, JEPS99635), 20 Jun 2001, MA 8075 (DEP0974, RM), 18 Jun 2001, MA 8039 (DEP0979), 2014] ARNETT ET AL.: VASCULAR FLORA OF DEVILS POSTPILE 385 18 Jun 2001, MA 8041 (DEP0978, JEPS99636), 20 Jul 1980, SCM 160 (DEP0146), 18 Jun 1972, KAH 150 (DEP034); J. orthophyllus Coville, uncommon, wet meadow; streambanks, 26 Jul 2001, DEPO. 0004.08 (DEPO1203); J. parryi Engelm., locally common, dry slopes in lodgepole pine forest and red fir -• western white pine forest, huckleberry oak shrubland, 25 Jul 2001, DEPO. 0030.02 (DEPO 1204); J. saximontanus A. Nelson, uncommon, mountain alder thickets, 22 Jul 2001, MA 8361 (DEP0982); Luzula comosa E. Mey., uncommon, intermittent streamcourse; wet meadow, 23 Jun 2001, MA 8145 (DEP0983, JEPS99639, RM); L. parviflora (Ehrh.) Desv, locally common, moist bench in slope of large granite outcrop; mountain alder and willow thickets, 3 Jul 2001, MA 8291 (DEP0985, RM), 20 Jun 2001, MA 8073 (DEP0984, JEPS99640). Juncaginaceae Trigiochin maritima L., uncommon, rock ledge along river with granite substrate, 23 Jun 2001, MA 8146 (DEP0986). Liliaceae Calochortus leichtlinii Hook, f., occasional, open areas in conifer forest, edge of landslide, 26 Jun 2001, MA 8262 (DEP0994), 4 Aug 1980, SCM 199 (DEPO 150); Fritillaria atropurpurea Nutt., uncommon, edge of landslide in lodgepole and Jeffrey pine, 18 Jun 2001, MA 8025 (DEP0995, JEPS99625); Lilium kel- leyanum Lemmon, locally common, moist drainages, 24 Jun 2001, MA 8201 (DEP0996, JEPS99626), 26 Jul 1972 KAH 194 (DEP037). Melanthiaceae Toxicoscordion venenosum (S. Watson) Rydb. var. venenosum, local, moist shaded areas, lodgepole pine forest, 3 Jul 2001, MA 8316 (DEPO 1005, JEPS99630), 18 Jun 2001, MA 8028 (DEPO 1004); Veratrum californicum Durand var. californicum, local, willow thickets, moist road verge, 10 Aug 1980, SCM 209 (DEP0155). Orchidaceae Corallorhiza maculata (Raf.) Raf., uncommon, red fir forest, 24 Jul 2001, MA 8366 (DEPO 1044, JEPS99609); Platanthera dilatata (Pursh) Lindl. ex L.C. Beck var. leucostachys (Lindl.) Luer, locally common, wet mead- ow, mountain alder and willow thickets, moist road verge, 23 Jun 2001, MA 8123 (DEPO 1045, JEPS99610); Spiranthes romanzoffiana Cham., local, wet meadow; mountain alder and willow thickets, 29 Jul 2001, MA 8459 (DEPO1046, JEPS99611). Poaceae Agrostis exarata Trin., uncommon, mountain alder and willow thickets, 7/22/2001 MA 8363 (DEPO 1049, JEPS99715); A. humilis Vasey, uncommon, mountain alder thickets, 22 Jul 2001, MA 8365 (DEPO 1054, JEPS99718); A. idahoensis Nash, common, mountain alder thickets, basalt outcrop and associated seep in white fir forest, 26 Jun 2001, MA 8245 (DEPO 1050, JEPS99716); A. scabra Willd., common, mountain alder and willow thickets, granite outcrop in lodgepole pine forest, 22 Jul 2001, MA 8364 (DEPO1053, RM), 3 Jul 2001, MA 8310 (DEPO 1051, JEPS99717); A. variabilis Rydb., local, mountain alder thickets, lodgepole pine woodland; Jeffrey pine woodland, 26 Jul 2001, DEPO.0014.05 (DEPO 121 3); Bromus ciliatus L., occa- sional, wet meadow; stream channels, mountain alder and willow thickets, 30 Aug 1980, SCM 220 (DEPO 170); B. laevipes Shear, locally common, basalt scree in lodgepole pine forest, lodgepole pine woodland, Jeffrey pine woodland, red fir - white fir forest, sagebrush scrub, black cottonwood forest, mountain alder thickets, 22 Jul 2001, MA 8351 (DEPO 1056, JEPS99719, RM), 3 Jul 2001, MA 8319 (DEPO1055, JEPS99920), 22 Jun 2001, MA 8106 (DEPO 1057, RM); B. suksdorfii Vasey, local, aspen grove; lodgepole pine woodland, 23 Jun 2001, MA 8172 (DEPO1058, JEPS99921); Calamagrostis canadensis (Michx.) P. Beauv., locally common, seeps, wet meadow, 15 Jul 2001, MA 8331 (DEPO 1059, JEPS99922, RM); C. stricta (Timm) Koeler subsp. inexpansa (A. Gray) C.W. Greene, locally common, wet meadow; willow thickets, 28 Jul 2001, MA 8435 (DEP0655, JEPS99923, RM); Cinna bolanderi Scribn., uncommon, moist drainage from spring associated with basalt outcrops over granite in Jeffrey pine - red fir forest, 27 Jul 2001, MA 8425 (DEP0656, JEPS99924); Danthonia intermedia Vasey, uncommon, mesic lodgepole pine forest, 26 Jul 2001, DEPO. 0003. 07 (DEPO 1201); Deschampsia cespitosa (L.) Beauv. subsp. cespitosa, local, wet meadow, 30 Aug 1980, SCM 216 (DEPO 173); D. danthonioides (Trin.) Benth, locally common, moist slope in mist of falls, basalt outcrop and associated seep in white fir forest, willow thickets, 25 Jun 2001, MA 8228 (DEP0657, JEPS99925), 26 Jun 2001, MA 8255 (DEP0659, RM); D. elongata (Hook.) Benth., local, wet meadow, 3 Jul 2001, MA 8303 (DEPO660, JEPS99926); Distichlis spicata (L.) Greene, uncommon, granite seep along river, 15 Jul 2001, MA 8332 (DEPO 1060, JEPS99927); Elymus elymoides (Raf.) Swezey var. californicus (J.G. Sm.) J.P. Sm, common, rock outcrops, Jeffrey pine - red fir forest, 25 Jul 2001, MA 8393 A (DEPO 1061); E. elymoides (Raf.) Swezey var. elymoides, common, granite outcrop; sag pond in Jeffrey pine - white fir forest, mountain alder thickets, 20 Jun 2001, MA 8071 (DEPO1062, JEPS99928, RM); E. glaucus Buckley subsp. glaucus, locally abundant, sag pond in bench of rolling granite hills, wet meadow; mountain alder and willow thickets, Jeffrey pine woodland, 25 Jul 2001, MA 8374 (DEPO 1064, JEPS99929), 15 Jul 2001, MA 8344 (DEPO 1063, RM); E. trachycaulus (Link.) Shinn, subsp. trachycau- lus, local, granite seep; slope of flaky rhyodacite just below falls, mesic lodgepole pine forest, wet meadow; willow thickets, 129 Jul 2001, MA 8455 (DEPO 1067), 5 Jul 2001, MA 8333 (DEPO 1066, JEPS99907); E. triticoides Buckley, uncommon, wet meadow, 31 Jul 1977 JEM (DEP0719); Glyceria elata (Nash ex Rydb.) M.E. Jones, locally common, moist bench in slope of large granite outcrop; sag pond in Jeffrey pine - white fir forest, mountain alder and willow thickets, aspen grove; black cottonwood forest, mesic lodgepole pine forest, wet meadow, 29 Jul 2001, MA 8457 (DEPO 1068, JEPS99909), 23 Jun 2001, MA 8173 (DEPO 1070, JEPS99910, RM), 20 Jun 2001, MA 8072 (DEPO 1069, JEPS99908, RM); G. striata (Lam.) Hitchc., uncom- mon, willow thickets, 30 Jul 2001, MA 8465 386 MADRONO [Vol. 61 (DEPO1071); Hordeimi hrachyantherum Nevski subsp. californicufii (Covas & Stebbins) Bothmer, N. Jacobson, & Seberg, abundant, wet meadow; mountain alder and willow thickets, aspen grove, 23 Jun 2001, MA 8170 (DEPO1072, JEPS99911); Melica bulbosa Geyer ex Porter & J.M. Coult., uncommon, rock outcrop in bitter cherry shrubland, 30 Jul 2001, DEPO.0041.14 (DEPO1205); M. harfordii Bob, uncommon, wet mead- ow, 30 Aug 1980, SCM219 (DEPO180); M stricta Bob, occasional, Jeffrey pine forest, 24 Jun 2001, MA 8210 (DEPO1073, JEPS99912); Muhlenbergia andina (Nutt.) Hitchc., uncommon, red fir - white fir forest, 26 Jul 2001, MA 8406 (DEPO1074, JEPS99913); M. filiformis (Thurb. ex S. Watson) Rydb, locally abundant, basalt outcrop and associated seep in white fir forest, wet meadow and gravel bars near river; mountain alder and willow thickets, 3 Jul 2001, MA 8301 (DEPO1077), 27 Jun 2001, MA ^277 (DEPO 1075, RM), 26 Jun 2001, MA 8256 (DEPO1076, JEPS99813); M. richardsonis (Trin.) Rydb, common, Jeffrey pine forest with scattered western juniper; wet meadow; gravel bars near river; mountain alder thickets, 27 Jun 2001, MA 8276 (DEPO1079, RM), 24 Jun 2001, MA 8209 (DEPO1078, JEPS99914); Pbleum alpimim L, common, wet meadow, intermittent streamcourse, willow thickets, 23 Jun 2001, MA 8124 (DEPO1080, JEPS99915); pratense E., native to Eurasia, uncommon, meadow edge, 1 1 Sep 1980, SCM 234 (DEPO730); *Pc>40 m), ping-pong balls were used to simulate the sailing of seeds on the water. Ping-pong balls (Family Maid by Dollar Empire LLC, Vernon, CA; recreational grade; 40 mm diameter; 2.7 g) were good mimics for S. good- dingii seeds because they sailed at the same speed and, most importantly, in the same direction as the seeds. A preliminary test on the Goat Canyon pool showed no significant difference in the direction of dispersal of ping-pong balls and S. gooddingii seeds (Mann-Whitney U = 284, ni = 112 = 24, P > 0.05). To simulate the secondary dispersal of A. gooddingii seeds on the pool, groups of 24 balls were released from 10 locations on the pool (Fig. 1) when westerly winds were blowing, and their dispersal tracks and destinations were mapped. All 240 balls (100%) were retrieved. The tracks of the actual balls were then converted to tracks of virtual balls originating from evenly- spaced points covering the entire pool. A grid of evenly-spaced (five m apart) virtual balls (n = 436) was placed over the map of the actual ball tracks and each virtual ball followed the track of the nearest actual ball to its destination at the pool’s edge. The number of virtual balls landing in each seedling site (with 10 m shoreline length) was counted. Seedling densities — predicted vj. actual. To test the hypothesis that secondary dispersal of seeds influences the patterning of plants in communi- ties, a simple model was developed to predict the density of seedlings found in each of the seedling sites around the Goat Canyon pool. The predict- ed number of seedlings per ni^ in a site (SLpredicted) is the sum of the seedlings that grew from seeds that arrived via hydrochory (SLh), anemochory (SLa), and pleustochory (SLp), i.e., SLpREDiCTED = SLh + SLa + SLp In this study, the number of seedlings growing from seeds that arrived via hydrochory (SLh) was zero because the pool received no inflow during the study period. The number of seedlings growing from seeds that arrived via anemochory (SLa) was estimated to be: SLa = Sa . WcF • Lcf/Wss • Lss where Sa is the average number of seeds caught in the seed traps during the study period (expressed as seeds per m^); Wcf and Wss are the average width of the capillary fringe (i.e., safe site) and the total width of the seedling site during the study period, respectively, as measured perpendicular to the water’s edge (m); and Lcf and Lss are the lengths of the capillary fringe and the seedling site, respectively, as measured parallel to the water’s edge (m). In these calculations Lcf and Lss were set at one m and SLa was expressed as number of seedlings per m^. The ratio Wcf • Lcf / Wss • Lss is required because only a narrow capillary fringe is suitable for seedling recruitment and it moved slowly through the seedling site as the water level declined. The number of seedlings growing from seeds that arrived via pleustochory (SLp) was estimated as: SLp = Sa.F/Wss.Lss where F is the Tetch’ (m^), i.e., the strip of the pool’s surface (one m wide) that was on the path from the source S. gooddingii plants to the seedling site, and Sa, Wss and Lss are as above. In these calculations Lss was set at one m and SLp was expressed as number of seedlings per m^. This equation formalizes the idea that all of the seeds that landed on the fetch during the study period sailed downwind to the seedling site where they germinated and became established. Notice that both SLa and SLp are based on only physical characteristics of the site and the number of seeds caught in the seed traps. The model’s predicted densities of seedlings in the seedling sites were compared to the actual seedling densities using correlation analysis (So- kal and Rohlf 1995). Results Seed Behavior Floating ability. All 60 S. gooddingii seeds floated on the water’s surface and remained 2014] BOLAND: WILLOW DISPERSAL 393 Fig. 3. The speed of S. gooddingii seeds sailing on the water surface. The equation and coefficient of determination are shown for the linear regression line. floating for the full four days. The cottony hairs did not become water-logged, and the seeds were not released from their enveloping coma of hairs. Salix gooddingii seeds therefore float when first landing on water and have the ability to stay afloat for days. Sailing speed. Salix gooddingii seeds were blown downwind while floating on the surface of the water. They sailed swiftly over the surface of the pool even when only light breezes were blowing. The seeds sailed quicker when the breeze was stronger and reached speeds in excess of 5 m per minute during the experiment (Fig. 3; r^ = 0.251, n = 60, P < 0.06). It took only a few minutes for a seed to sail the length of the small pool (20 m). Seed Dispersal and Seedling Densities Seedling densities. Salix gooddingii was the most abundant species within the seedling com= munity that developed in the zone exposed during the 15“day study period at the Goat Canyon pool. Its seedlings formed a band that encircled the entire pool with average densities from 68.8 to 3140 seedlings per m^ (Sites 10 and 5, respectively; Table 1). Seedling densities were lowest along the Off Shore (particularly in Sites 9-12) and highest along the On Shore (particularly in Sites 3-5), and there was a significant difference in densities on the two shores (Mann- Whitney U = 7.4; ni = 7; n2 = 6; P < 0.01). Sediment grain sizes were similar at all sites, with fine sands and silts dominating (Table 1). There was no significant difference in the composition of the sediments on the Off Shore and On Shore (Mann-Whitney U = 23; ni = 7; n2 = 6; P > 0.05). Primary dispersal - aerial seed rain. A total of 28 N. gooddingii seeds were caught in the seed traps at the Goat Canyon pool during the 1 5-day period, for an average of 68.2 (±14.6 std. err.; n = 120) seeds per m^. There was no correlation between the number of seeds in a trap and the distance of the trap from the seed source (r^ = 0.1015; n = 10; P > 0.05) suggesting that, at this distance from the source (>250 m), the seed rain over the pool was sparse and approximately evenly distributed. Secondary dispersal - seeds and seed mimics on the water. Salix gooddingii seeds were blown from their source trees to the Goat Canyon pool when westerly winds blew (Fig. 1). Seeds that landed on the pool continued to be blown by the westerly wind across the surface of the pool to the water’s edge along the On Shore. At the water’s edge seeds often accumulated with other floating plant debris, such as flocculent algae, pollen and leaves. Each day this mix of floating debris was deposited along the On Shore by the largest breaking waves and highest wind surges of the day. When the seeds were washed against the firm shoreline sediment they were separated from their cottony hairs and their dispersal was ended. Because the water level in the Goat Canyon pool was declining at a rate of —one cm per day and exposing —10 cm of shore per day, the stranded debris was left in distinctive lines, known as windrows or wrack lines, —10 cm apart. These windrows extended for tens of meters along the shoreline. Later, the seedlings were distributed in the same distinctive windrows. The ping-pong balls provided a clear view of where seeds sailed on the pool’s surface. All 240 balls (100%) were blown by westerly winds across the pool to the On Shore. Even balls released near the Off Shore ended up on the On Shore. These observations confirmed that sites around the pool were receiving seeds from the pool’s surface and, due to the location of the source trees and 394 MADRONO [Vol. 61 necessary wind direction, On Shore sites were receiving the majority of the seeds that landed on the pool. When the tracks of the actual ping-pong balls were converted into tracks of evenly-spaced virtual balls, nearly all of the virtual balls ended up on the On Shore (Table 1). Because each virtual ball represented a particular area of the pool’s surface, an accumulation of many virtual balls in a site meant that the site had received balls from a large area of the pool’s surface. This area of the pool is the fetch for that seedling site. The configuration of the pool in relation to the source plants meant that some seedling sites had a large fetch whereas others had none at all (Table 1). Salix gooddingii seeds that landed in the dry areas around the Goat Canyon pool were observed to become stuck, either on plants or on rough ground. There was no indication that seeds that landed in these dry areas were later blown to safe sites at the water’s edge. Seedling densities - predicted vs. actual The mathematical model accurately predicted the pattern of seedling densities around the Goat Canyon pool (Table 1). The model predicted that seedling densities would be highest along the On Shore (particularly in Sites 4-6) and lowest along the Off Shore (particularly in Sites 8-12). The predicted and actual number of seedlings at the sites were significantly correlated (r^ = 0.827; n = 13; P < 0.01). An important component of the model was the size of the fetch, and the fetch for each site was positively and significantly corre- lated with the number of seedlings in a seedling site (r“ = 0.570; n = 13; P < 0.01). Where there was a long fetch between the source and the seedling site the site had many seedlings, whereas if there was little or no fetch between the source and the seedling site the site had few seedlings (Table 1). Discussion Sailing on the Water Surface - Pleustochory When S. gooddingii seeds land on a pool, they float on the water’s surface and sail quickly with the breeze across the surface. As there is currently no term for the dispersal of seeds by wind while floating on standing water, I have coined the tenu pleustochory from the Greek words ''pleusf “to sail” or “to float” and '"chore’’" “to move.” This form of dispersal (pleustochory) is differentiated from anemochory, in which seeds are dispersed by wind and carried in air, and from hydrochory, in which seeds are carried downstream by flowing water (van der Pijl 1982). The prefix “pleust” already appears in the scientific literature as pleuston, the name given to the group of plants (e.g., Lemna spp., AzoUa spp., Sargassuni spp.) 2014] BOLAND: WILLOW DISPERSAL 395 and animals (e.g., Velella velella) that are free- floating at the air-water interface of a body of water and are dispersed by winds while floating on water (Wetzel 1983). The term pleustochory should also include the dispersal of seeds and seedlings that float at the water’s surface and are transported by wind-driven water movements (Soomers et ah 2010; Sarneel et ah 2014). Wind-driven seed dispersal along the ground has been studied in many species and situations. It has been documented in temperate grasslands (van Tooren 1988), coastal environments (Wat- kinson 1978), alpine habitats (Chambers et al. 1991), arid and semiarid environments (Aguiar and Sala 1997), on snow (Matlack 1989; Greene and Johnson 1997), as well as in environments disturbed by fire (Bond 1988), human activities (Campbell et al. 2003) and volcanic eruptions (Fuller and del Moral 2003). In general, these studies have found that dispersal by wind over the ground occurs only under certain conditions: when seeds remain mobile for sufficient periods of time; when the ground surface is smooth; when few obstacles impede seed movement; and when winds reach the ground. Such conditions are rarely met for seeds on the ground and, in general, wind-blown seeds usually move only short distances on the ground before becoming trapped on plants or in crevices (Chambers and MacMahon 1994). These conditions are easily met for willow seeds on water, however. The seeds remain mobile for days, the water surface is smooth, there are few obstacles, and winds reach the surface of the water. The result is that willow seeds can quickly travel tens of meters or more on the water’s surface. The seeds of S. gooddingii can float and sail because they are small, light, and surrounded by a coma of long, hollow hairs (Steyn et al. 2004; Seiwa et al. 2008). Many other riparian and wetland species have small, cottony seeds similar to S. gooddingii and are also likely to be able to sail; these species include other willows {Salix spp.), cottonwoods {Populus spp.), cattails (Ty- pha spp.), tamarisk (Tamarix spp.), and many Asteraceae (e.g., Baccharis spp.). Pleustochory therefore may be a common mode of dispersal in many riparian and wetland plant species. Sailing into Safe Sites - Directed Dispersal The secondary dispersal of S. gooddingii seeds by pleustochory results in seeds traveling from the middle of a pool to the edge of a pool. For S. gooddingii and other willows, safe sites are vegetation-free, moist substrates that typically occur at the water’s edge around disturbed pools and along stream banks (Mahoney and Rood 1998; Seiwa et al. 2008). In these safe sites, willow seeds germinate successfully and seedling estab- lishment is high (Boland 2014). The secondary dispersal of 5*. gooddingii seeds by pleustochory therefore results in seed movement from unsuit- able sites (middle of a pool) to safe sites (edge of a pool), and is an example of directed dispersal, i.e., the non-random dispersal to sites especially favorable for successful germination and seedling establishment (Howe and Smallwood 1982; Wenny 2001; Seiwa et al. 2008). Most examples of directed dispersal involve the action of an animal vector, i.e., an ant, bird, or mouse carries the seed away from the parent plant or ground surface to a more suitable germination site (e.g., Herrera and Jordano 1981; Wenny and Levey 1998). Some researchers have even questioned whether it is possible for winds to move seeds in a non-random manner with respect to establishment sites (e.g., Vander Wall and Longiand 2004). But winds can concentrate seeds into safe sites if the safe sites are structurally different to their surroundings; in this study, S. gooddingii seeds were easily blown across the smooth surface of the Goat Canyon pool but became stuck in the rough, muddy edge, which is their safe site. In addition, Seiwa et al. (2008) found under experimental conditions that willow seeds can be blown over dry sand and trapped in wet sand and therefore blown from less suitable to more suitable microsites, although dispersal of this type has not been documented under natural conditions. Also, winds can con- centrate seeds into safe sites when wind speed or direction is changed by an obstacle or vegetation gap. Winds in and around forest gaps were found to carry seeds of shade-intolerant plants from surrounding forested areas into sunny forest gaps, more suitable for their germination and growth (Augspurger and Franson 1988). So, contrary to the prevailing view of Vander Wall and Longiand (2004) and others, winds can disperse seeds non-randomly and can play a role in their directed dispersal into establishment sites. Directed dispersal is difficult to identify in most systems and may be more common and ecologically significant than previously believed (Wenny 2001). Directed dispersal is difficult to identify because: (1) following a dispersing seed from its parent to its final destination is difficult; (2) safe sites are often difficult to specify or are poorly known for most species; and (3) deter- mining which seeds from a cohort germinate successfully can be complicated by dormancy and the presence of other cohorts in the seed bank (Wenny 2001). Willows, however, have charac- teristics that make directed dispersal compara- tively easy to detect: easily-followed seed produc- tion (Stella et al. 2006; Boland 2014); relatively easily-followed primary and secondary dispersal (Gage and Cooper 2005, this study); well-known and easily identified safe sites (Seiwa et al. 2008); and immediate germination with no seed dor- mancy or seed bank (Emery 1988; Stella et al. 396 MADRONO [Vol. 61 2006; Karrenberg et al. 2002). The willows of southern Californian riparian ecosystems are therefore excellent species in which to test dispersal hypotheses. Pleustochory and Seedling Densities The general hypothesis that the secondary dispersal of seeds on pools influences the patterning of plants was confirmed. First, direct observations showed that, when westerly winds were blowing, most of the seeds (and ping-pong balls) that landed on the Goat Canyon pool were blown across the water to the On Shore. Second, seedling densities along the On Shore were very dense - up to 3140 seedlings per m^ or more than 45 times the density of seeds arriving in the aerial seed rain, the only other mode of seed transport into these sites at that time. Third, a model derived from only the number of seeds in the aerial seed rain and physical characteristics of the sites correctly predicted the pattern of seedling densities around the Goat Canyon pool and estimated that 99% of the seedlings on the On Shore came from pleustochory. Finally, the seedlings on the On Shore were growing in distinctive windrows, a sure sign that the seeds had arrived from the water rather than from the air (Glaser 1981). Together, these lines of evidence indicated that secondary dispersal via pleustochory produced the striking seedling density pattern around the Goat Canyon pool. Many authors point to the high mortality rates of willow seedlings during their first summer as an important limiting feature of young riparian communities (e.g., Karrenberg et al. 2002). But the effect of high mortality can be greatly offset by initially high seedling densities; even if 90% of seedlings die in their first summer, enough will survive their first year to form a thriving thicket the following year. In an earlier study in the Tijuana River Valley (Boland 2014), S. gooddingii seed- lings had a 92% mortality rate during their first year but formed dense thickets by the start of their second year (—20 yearlings per m^ and — 1.7 m tall), and went on to form tall, dense stands three years later (Boland, personal observations). Ex- tremely high initial densities of seedlings - in part due to the concentrating effect of pleustochory - greatly lessen the importance of high mortality rates. The effect of pleustochory on seedling densities was obvious at the Goat Canyon pool because the pool was large and unvegetated with a single, isolated seed source. Pleustochory most likely occurs on many water bodies - pools, lakes, even flowing water - when there are source plants nearby, though its effect on seedling densities may be less clear elsewhere. What should be clear at other sites is the effect of pool size. This study showed that the size of the fetch greatly influenced the density of seedlings around the pool. The size of the fetch was important because , all of the seeds that landed on the water within j the fetch quickly sailed downwind to the safe site ; at the water’s edge, and a long fetch “collected” | more seeds from the seed rain than a short fetch, j Consequently there was a positive correlation j between fetch size and seedling densities. As size I of fetch and size of pool are also positively correlated, it follows that pool size will influence the density of seedlings at a site, Pleustochory and Rapid Establishment It is important for willow seeds to disperse and establish quickly. Willow seeds are small and have no food reserves, so will perish within a few days unless they land in a suitable habitat (Karrenberg et al. 2002). In addition, most willow safe sites on the edges of drying pools are available only temporarily. Therefore, willow seeds need to quickly disperse to safe sites, and pleustochory plays a vital role in their speedy delivery. At the Goat Canyon pool, a 5'. gooddingii seed could be blown from the source stand near the pool in the morning and take approximately five minutes to arrive at the center of the pool 420 m away (i.e., primary dispersal via anemochory). The seed would then take approximately 10 min to sail the 47 m on the pool’s surface to the pool’s edge (i.e., secondary dispersal via pleustochory). The seed would likely be stalled as long as five hours at the water’s edge until the highest waves and wind surges deposit it on the shore, thereby separating the seed from its cottony hairs and placing it in its moist recruitment site where it would start to germinate immediately. Evaporation and the resulting declining water levels would then strand the germinated seed in the moist mud above the new waterline. Fifty percent of the seeds would have germinated within approximately 11 hours of being immersed in the moist mud (Boland unpublished data). In this case, the steps from dispersal of the seed from the parent plant to germination in the recruitment site took less than 24 hours to complete, and this may be typical for willows. Chambers and MacMahon (1994) titled their review of secondary dispersal “A day in the life of a seed” and, although their title implies that they were dealing with processes that occur in a day, they were actually dealing with primary dispersal, secondary dispersal, dormancy and germination in many species where these process- es can take months or years. But in *5*. gooddingii all of these processes can really take a single day - probably the most important day in the life of a S. gooddingii seed. Conclusions This paper defines a previously undescribed mode of seed dispersal as pleustochory (dispersal 2014] BOLAND: WILLOW DISPERSAL 397 of seeds by wind while the seeds are floating on water), and distinguishes this mode from ane- mochory (dispersal of seeds by wind while the seeds are in the air) and from hydrochory (dispersal of seeds downstream by flowing water). The purpose of distinguishing pleustochory from the other two modes of dispersal was to highlight its vital role in three aspects of dispersal: (1) its role in the directed dispersal of willow seeds to their safe sites - the vast majority of S. gooddingii seedlings at the Goat Canyon pool were estimat- ed to have arrived via pleustochory; (2) its role in the concentration of seeds (and seedlings) within safe sites - S. gooddingii seedlings at the Goat Canyon pool were significantly more dense along the On Shore (where pleustochory had an effect) than along the Off Shore (where it had little or no effect); and (3) its role in the speedy delivery of short-lived seeds into short-lived germination safe sites, allowing S. gooddingii seeds to become established in recruitment sites in less than 24 hours after release from the parent plant. The findings of this paper should also lead us to view pools in a new way, i.e., as collectors and distributors of floating seeds. S. gooddingii safe sites are typically rare (comprising less than 0.03% of the flood plain surface area in the Tijuana River Valley during May 2014, Boland unpublished data). Being rare, safe sites are small targets for airborne S. gooddingii seeds. But, as this study shows, a pool acts as a collector dish or landing field for seeds that can then sail to safe sites. Because overall pool surface areas are relatively large in relation to those of safe sites (approximately 30 times larger in the flood plain of the Tijuana River Valley in May 2014, Boland unpublished data), pools greatly expand the safe site target area for S. gooddingii seeds, thereby increasing the likelihood that seeds will arrive in a site suitable for establishment. Secondary dispersal via pleustochory therefore helps make willows quintessential pioneer plants, as it enables them to quickly arrive and establish in disturbed habitats where conditions for successful germination are rare in both space and time. Acknowledgments I thank: Michelle Cordrey at the Tijuana River National Estuary Research Reserve for the wind data, GPS measurements, and producing the maps; Chris Peregrin at California State Parks for making the Goat Canyon area available for study; Lisa Ordonez for assistance in the field; Deborah Woodward for assis- tance in the field, helpful discussions, and comments on early drafts of this manuscript; and two anonymous reviewers for thoughtful reviews that greatly improved the manuscript. Literature Cited Aguiar, M. R. and O. E. Sala. 1997. Seed distribution constrains the dynamics of the Pata- gonian steppe. Ecology 78:93-100, Augspurger, C. K. and S. E. Franson. 1988. Input of wind-dispersed seeds into light-gaps and forest sites in a Neotropical forest. Journal of Tropical Ecology 4:239-252. Boland, J. M. 2014. Factors determining the estab- lishment of plant zonation in a southern Califor- nian riparian woodland. Madrono 61:48-63. Bond, W. J. 1988. Pro teas as “tumbleseeds”: wind dispersal through the air and over soil. South African Journal of Botany 54:455M60. Campbell, D. R., L. Rochefort, and C. Lavoie. 2003. Determining the immigration potential of plants colonizing disturbed environments: the case of milled peatlands in Quebec. Journal of Applied Ecology 40:78-91. Chambers, J, C. and J. A. MacMahon. 1994. A day in the life of a seed: movements and fates of seeds and their implications for natural and managed systems. Annual Review of Ecology and System- atics 25:263-292. , , and j. H. Haefner. 1991. Seed entrapment in alpine ecosystems: effects of soil particle size and diaspore morphology. Ecology 72:1668-1677. Emery, D. E. 1988. Seed propagation of native California plants. Santa Barbara Botanic Garden, Santa Barbara, CA. Faber, P. A., E. Keller, A. Sands, and B. M. Massey. 1989. The ecology of riparian habitats of the southern California coastal region: a commu- nity profile. U.S, Department of the Interior, Fish and Wildlife Service, Research and Development National Wetlands Research Center, Biological Report 85 (7.27), Washington, D.C. Fuller, R. N. and R. del Moral. 2003. The role of refugia and dispersal in primary succession on Mount St Helens, Washington. Journal of Vegeta- tion Science 14:637-644. Gage, E. A. and D. J. Cooper. 2005. Patterns of willow seed dispersal, seed entrapment, and seed- ling establishment in a heavily browsed montane riparian ecosystem. Canadian Journal of Botany 83:678-687. Glaser, P. H. 1981. Transport and deposition of leaves and seeds on tundra: a late-glacial analogue. Arctic and Alpine Research 13:173-182. Greene, D. F. and E. A. Johnson. 1997. Secondary dispersal of tree seeds on snow. Journal of Ecology 85:329-340. Harper, J. L. 1977. Population biology of plants. Academic Press, New York, NY. Herrera, C. M. and P. Jordano. 1981. Prunus mahaleb and birds: the high efficiency seed dispersal system of a temperate fruiting tree. Ecological Monographs 51:203-221. Howe, H. F. and J. Smallwood. 1982. Ecology of seed dispersal. Annual Review of Ecology and Systematics 13:201-228. Johnson, E. A. and G. L Fryer. 1992. Physical characterization of seed microsites - movement on the ground. Journal of Ecology 80:823-836. Karrenberg, S., P. j. Edwards, and J. Kollmann. 2002. The life history of Salicaceae living in the active zone of floodplains. Freshwater Biology 47:733-748. Mahoney, J. M. and S. B. Rood. 1998. Streamflow requirements for cottonwood seedling recruitment - an integrative model. Wetlands 18:634-645. 398 MADRONO [Vol. 61 Matlack, G. R. 1989. Secondary dispersal of seed across snow in Betula lenta, a gap-colonizing tree species. Journal of Ecology 77:853-869. Nathan, R. and H. C. Muller-Landau. 2000. Spatial patterns of seed dispersal, their determi- nants and consequences for recruitment. Trends in Ecology and Evolution 15:278-285. Sarneel, J. M., B. Beltman, A. Buijze, R. Groen, AND M. B. SOONS. 2014. The role of wind in the dispersal of floating seeds in slow flowing or stagnant water bodies. Journal of Vegetation Science 25:262-274. Seiwa, K., M. Tozawa, N. Ueno, M. Kimura, M. Yamazaki, and K. Matuyama. 2008. Roles of cottony hairs in directed seed dispersal in riparian willows. Plant Ecology 198:27-35. SOKAL, R. R. AND F. J. Rohlf. 1995. Biometry. W.H. Freeman and Company, New York, NY. SOOMERS, H., D. N. WlNKEL, Y. DU, AND M. J. Wassen. 2010. The dispersal and deposition of hydrochorous plant seeds in drainage ditches. Freshwater Biology 55:2032-2046. Stella, J. C., J. J. Battles, B. K. Orr, and J. R. McBride. 2006. Synchrony of seed dispersal, hydrology and local climate in a semi-arid river reach in California. Ecosystems 9:1200-1214. Steyn, E. M. a., G. F. Smith, and A. E. van Wyk. 2004. Functional and taxonomic significance of seed structure in Salix mucronata (Salicaceae). Bothalia 34:53-59. Takeda, H. 1936. On the coma or hairy tuft on the seed of willows. The Botanical Magazine, Tokyo 50:283-289. van der Pul, L. 1982. Principles of dispersal in higher plants. Springer-Verlag, New York, NY VAN Tooren, B. F. 1988. The fate of seeds after dispersal in chalk grassland: the role of the bryophyte layer. Oikos 53:41^8. Vander Wall, S. B. and W. S. Longland. 2004. Diplochory: are two seed dispersers better than one? Trends in Ecology and Evolution 19:155-161. Watkinson, a. R. 1978. The demography of a sand dune annual: Vulpia fasciculata. III. The dispersal of seeds. Journal of Ecology 66:483^98. Wenny, D. G. 2001. Advantages of seed dispersal: a re- evaluation of directed dispersal. Evolutionary Ecology Research 3:51-74. AND D. J. Levey. 1998. Directed seed dispersal by bellbirds in a tropical cloud forest. Proceedings of the National Academy of Sciences of the United States of America 95:6204—6207. Wetzel, R. G. 1983. Limnology. Saunders, Philadel- phia, PA. Willson, M. F. 1983. Plant reproductive ecology. John Wiley and Sons, New York, NY. Zedler, j. B., C. S. Nordby, and B. E. Kus. 1992. The ecology of Tijuana Estuary, California: a National Estuarine Research Reserve. NOAA Office of Coastal Resource Management, Sanctu- aries and Reserves Division, Washington, D.C. Madrono, Vol. 61, No. 4, pp. 399-406, 2014 A NEWLY DESCRIBED SERPENTINE^ENDEMIC ADr//6/N (RHAMNACEAE) FROM COASTAL MARIN COUNTY, CALIFORNIA V. Thomas Parker Department of Biology, San Francisco State University, 1600 Holloway Avenue, San Francisco, CA 94132 parker@sfsu.edu Abstract Ceanothus decovnutus V. T. Parker is a newly described species found on serpentine outcrops in western Marin County. This taxon has been considered as part of C. jepsonii Greene in recent treatments. Ceanothus decornutus differs in a number of ways from C. jepsonii and suggests relationships to other northern San Francisco Bay species. Beyond the species description and a key to similar taxa north of San Francisco Bay, an analysis of leaf characters is used to indicate possible relationships. Key Words: Ceanothus, chaparral, maritime, Rhamnaceae, serpentine. On a ridge above Platform Bridge Road, SE of Black Mountain and just south of Nicasio Reservoir, in western Marin County, I accompa- nied Roger Raiche and Terri Thomas on 19 April 1991 to see a Ceanothus found on serpentine soils along that ridge. Upon arriving, the Ceanothus L. was in bloom among the rocky serpentinite outcrops; Raiche asked my opinion of the species and if I thought it was C. jepsonii Greene. In contrast to C. jepsonii, the flowers ranged from light blue to white, often with pink in some of them, and those that I checked were all 5-merous in contrast to 6-merous flowers in C jepsonii; that suggested it was not C. jepsonii, but perhaps a different species of which I was unaware. The population currently is known to Marin County botanists and is referred to under C. jepsonii (Howell et al. [2007] but not mentioned in Howell [1970]). In comparison with other poten- tial candidates in the northern San Francisco Bay region, these populations should be separated from C. jepsonii so that their origin and potential relationships with other North Bay species can be investigated. The newly described Ceanothus appears to be a part of the complex of species found mostly on serpentine and volcanic soils in Marin, Sonoma, and Napa counties. This taxon is close to C. jepsonii in appearance, but differs in a few significant characters such as having 5-merous flowers and with horns (out- growths) on the top of the fruit being usually short and rounded in contrast to the lengthy and wrinkled horns found on fruit in C. jepsonii. While the leaves are similar in form to C jepsonii, they average fewer, shorter spines along the edges. Compared to other species in the North Bay Region, the new taxon differs by leaf or fruit characters, for example, the leaves are much shorter than C purpureus Jeps. with far fewer and much shorter spines or teeth, and compared to C. sonomensis J. T. Howell or C. confusus J. T. Howell, C. decornutus is much more ovate, and differs in floral color. In this paper I describe this species and compare a number of characters of this taxon with other northern San Francisco Bay region Ceanothus species. Taxonomic Treatment Ceanothus decornutus V. T. Parker, sp. nov. (Fig. 1). — TYPE: USA, California, Marin Co., serpentinite outcrops on ridge south of Black Mountain and Nicasio Reservoir, 15 April 2013, V. T. Parker 1487 (holotype: CAS; isotype: JEPS). Plant generally erect, ±open, < 1.5 m. Stem: ascending to erect, intricately branched; twigs generally brown, thick. Leaf: opposite, evergreen, generally reflexed at tip; stipules knob-like; petiole < 2 mm; blade 10-20 mm, 9-19 mm wide, ovate to oblong-ovate, often ± folded lengthwise, adaxially green, glabrous, abaxially ± short-strigose between veins, margin thick to ± rolled under, wavy, 7-9-spine-toothed. Inflores- cence: umbel-like, 1—2.5 cm length. Flower: sepals, petals 5, light blue to white, sometimes with pink. Fruit: 4.5-7 mm wide, 3-ridged to lacking, frequently granular-surfaced; horns 0-1.5 mm, thick, generally rounded, bulge-like. Habitat: Rocky, serpentinite outcrops (50) 150-290 m. Along ridge SE of Black Mountain, to the east and paralleling Bolinas Ridge, North Coast Ranges, Marin County. Center of population at 38°03'33.59"N, 122°44'25.82"W. Mar-Apr. Methods Collections of Ceanothus decornutus were made multiple times. Data in this paper reflect collec- tions made 22 March 1991 (K T. Parker 1474), 14 June 1991 (K T. Parker 1475), 24 May 2011 (K T. Parker 1486), and 15 April 2013 (K T. 400 MADRONO [Vol. 61 Fig. 1 . Fruits and leaves of Cecmothus decornutus. 1 A) typical fruits illustrating the thickened mounds for horns at the top of the fruit, leaves folded with significant spines; IB) example of a plant with fruit having small horns, and granular to papillate surfaces. Parker 1487). Other collections were C. sono- mensis J. T. Howell, Sonoma County, 1 July 1991, near Sugar Loaf along Adobe Canyon Road (K T. Parker 1478); C. gloriosus J. T. Howell var. exaltatus J. T. Howell, Marin County, Bolinas Ridge north of Bolinas-Fairfax Road, 18 July 1991 (K T. Parker 1479); C. purpureas Jeps., Napa County, Soda Canyon Road, near Chimney Rock Road, 18 July 1991 { V. T. Parker 1482); C. jepsonii var. jepsonii, Marin County, Pine Mountain Fire Road, near the Azalea Hill parking area, 18 July 1991 (K T. Parker 1481); C. jepsonii var. albiflorus J. T. Howell, Napa County, North of Hwy 121 about 6-8 mi N of Napa, 18 March 1994 {V. T. Parker 1475). I did not include C. confusus nor C. diver gens in the following analyses because of the presence of horns on their fruit and the typical leaf only having spines near the tip of the leaf. Measurements of leaves and fruit were made with digital calipers; measurements of fruit used freshly collected fruit. For C. decornutus and C. jepsonii var. jepsonii, 15 individuals were used, for the other species, ten individuals. For leaf measurements, a mature branch from current years growth was selected arbitrarily and the largest leaf on that branch was measured. Similarly, the largest fruit on an arbitrarily chosen branch was measured for C. decornutus and C. jepsonii var. jepsonii. Analyses were made using R (R Core Team 2012), and consisted of simple descriptive statis^ tics and Principal Components Analysis. The data used in the PCA included leaf length, leaf width, spine number, spine length, and both with and without floral merosity. The results shown are PCA analyses without including floral merosity. Results This species is variable in floral color and fruit morphology. Generally flowers are light blue but 2014] PARKER: NEW MARIN COUNTY CEANOTHUS 401 Fig. 2. Fruits and leaves of Ceanothus jepsonii var. jepsonii. 2A) fruits illustrating the prominent and wrinkled horns. Fruits typically display prominent and wrinkled ridges. 2B) close-up of fruits also illustrating typical leaves, folded with significant spines. frequently there are white flowers; additionally, in some instances pink occurs as part of the floral tube. Similarly, fruit are usually quite similar, with a rounded bulge or thickened horn that is short and lacking prominent ridges, although the fruit are often somewhat granular to papillate on the surface (Fig. lA). If one searches through the population, fruit with more prominent horns and sometimes wrinkled ridges do occur (Fig. IB) similar to but shorter and not as pronounced as in C. jepsonii (Fig. 2). Distribution and Habit This taxon is found among rocky, serpenti- nite outcrops generally along a ridge parallel to and east of Platform Bridge Road, south of Black Mountain and Nicasio Creek. The ridge is principally serpentine grassland (Fig. 3A) with serpentinite outcrops in some places among soils that are clay-dominated Henneke Series serpentine soils (calsoilresource.lawr. ucdavis.edu). Most of the population is along the upper portions of the ridge, above 200 m, but serpentinite outcrops with clusters of Ceanothus plants continue on the northern side of the ridge down to Nicasio Reservoir. The vegetation containing Ceanothus decornutus is essentially a nearly monospecific maritime chap- arral stand, within which are occasional other woody species (e.g., Baccharis pilularis DC. and Umbellularia californica [Hook. & Arn.] Nutt.) intermixed with or surrounded by serpentine grassland (Fig. 3B). 402 MADRONO [Vol. 61 Fig. 3. Habitat of Cecmothiis decormitus. 3 A) Generally the populations are on exposed serpentinite outcrops along a ridge above Nicasio Reservoir. They are visible as the darker grey areas along the top of the ridge. Salt spray/wind-sculpted Umbellularia (Nees) Nutt, are visible toward the back of the photo along the ridge. 3B) Cecmothus individuals are found within the rock outcroppings, usually as the only woody species. Taxonomic Relationships Cecmothus ciecornutus differs from C. jepsonii by the flowers being 5-meroLis while C. jepsonii are 6- merous, and by reduced length of horns on the fruit (hence the name ‘decornutus’ to reflect the fruit being ‘unhorned’ or ‘dehorned’) (Table 1). Addi- tionally, in Marin County, C. jepsonii is dark blue- purple in flower color {\3.y. jepsonii), while flowers of C deeornutus range from white to light blue, sometimes with pink. While there is some variation in flower color and fruit characteristics in C. deeornutus, other morphological characteristics are generally quite consistent within the population, e.g., leaf morphology and branching patterns. Although located within a similar geographic region, C. deeornutus deviates more from C jepsonii and C. gloriosus varieties than expected. When superficially examining the leaf and stem morphology, C. deeornutus appears to be related to C. jepsonii. Both have spiny, ‘holly-like’ leaves, often with the apical spine bent under, with leaves slightly folded along the midrib (Figs. lA, 2B). Similarly, stems found on both are rather stout. Variation in the fruit suggests the potential for a relationship with C. jepsonii. 2014] PARKER; NEW MARIN COUNTY CEANOTHUS 403 Table 1. Fruit Characters (Mean and Standard Deviation) for Ceanotuus Decornvtus and C. JEPSONII = 15). Taxon Fruit width (mm) (SD) Height to rim (mm) (SD) Length of horn (mm) (SD) C. decornutus 5.77 (0.63) 3.82 (0.61) 0.58 (0.39) C. jepsonii var. jepsonii 5.34 (0.49) 3.26 (0.58) 2.05 (0.48) At the same time, other morphological char- acters of Ceanothus decornutus are more consis- tent with the two serpentine and/or volcanic-soil endemics C sonomensis and C purpureus. For example, when considering leaf and fruit charac- ters, C. decornutus often falls allometrically more with C sonomensis or C purpureus, depending on the character, than with either C. jepsonii var. jepsonii or C. jepsonii var. albiflorus (Table 2). Measurements of leaf length (Fig. 4A), leaf width (Fig. 4B), number of teeth on leaves (Fig. 4C), and the length of the teeth (Fig. 4D) suggest that C. gloriosus does not share leaf characters with C. decornutus. The new taxon is close to C purpureus in leaf length and width characters, while closer to C sonomensis in number of teeth and their length (Figs. 4C, 4D). Combining these characters in a PCA analysis also indicates that C. decornutus variation is centered as much within C sonomensis and C purpureus than with either C jepsonii variety for this combination of characters and that C gloriosus is quite distinct from the other taxa as expected (Fig. 5). The first two axes of the PCA account for 70% of the variation, with leaf length and leaf width loaded more on the first axis, and tooth length and number loaded more on the second axis. Fruit morphology also varies among these species. All the species have similar sized fruit, but C. decornutus differs from C gloriosus, C. purpureus and C. sonomensis in often having a roughened, granular surface of the fruit and, short, thickened or highly reduced horns (if they are present at all). Although the fruit horns are thickened similar to C. jepsonii, in C decornutus most often they are reduced to absent and almost always lack the significant wavy ridges. Finally, C. decornutus fruit differs C. jepsonii fruit in another way; C jepsonii fruit generally go through a significant reddish color phase prior to fruit ripening (Howell 1970; Howell et al. 2007) while C decornutus fruit lack that reddish phase. The origin of C decornutus is not clear, for example, and recent molecular genetic research in Ceanothus frequently does not help to define closely related species boundaries (Jeong et al. 1997; Hardig et al. 2000; Burge and Manos 2011). The combined history of research in Ceanothus suggests several hypotheses. Ceanothus decornu- tus could be a taxon that simply has arisen allopatrically from a common ancestor with, for example, C. jepsonii; serpentine endemics have arisen more than once in section Cerastes S. Watson, so that is a possibility. The North San Francisco Bay spiny-leaved section Cerastes species represent a collection of closely related taxa (Hardig et al. 2002), and given that some of the mosaic of characters in C decornutus are similar to C. jepsonii and others similar to C sonomensis or C purpureus, C. decornutus also could have arisen from incomplete lineage sorting (Burge et al. 2011). Not that many characters were examined in this study, so all of these are a possibility. Another hypothesis is that taxa have genetically hybridized and introgressed; this has been the principal model for Ceanothus since early studies showed little reproductive barriers among species (e.g.. Nobs 1963). Nobs (1963) conducted considerable experimental work with Ceanothus taxa and was impressed with the ability of taxa within sections to hybridize, particularly in section Cerastes occurring in the north coast ranges (p. 50, Nobs 1963). Nobs (1963, p. 82) also thought that new forms could arise by natural selection working on recombina- tion swarms of different taxa brought together by environmental change, selecting among pheno- types arising post-fire from persistent soil seed banks. Recent research confirms that these section Cerastes taxa in the North Bay are closely related (Hardig et al. 2002). While hybridization may be possible and hybrids can be seen in the field, other hypotheses are just as or more likely in the Table 2. Leaf Characters (Mean and Standard Deviation) for Select Ceanothus Species (n = 1 5 for THE First Two, n = 10 for the Rest) from the Northern San Francisco Bay Region. Taxon Leaf length (mm) (SD) Leaf width (mm) (SD) Teeth #/side (SD) Length of teeth (mm) (SD) C decornutus 14.65 (2.91) 14.56 (2.80) 3.73 (0.46) 1.43 (0.40) C. jepsonii var. jepsonii 11.96 (1.63) 10.13 (1.82) 4.27 (0.80) 2.02 (0.54) C. jepsonii var. albiflorus 14.28 (1.29) 10.46 (1.43) 4.80 (0.63) 2.09 (0.45) C. sonomensis 13.09 (1.80) 8.38 (1.15) 3.30 (0.48) 1.30 (0.39) C. purpureus 17.51 (2.37) 16.76 (3.07) 5.30 (0.68) 2.50 (0.45) C. gloriosus var. exaltatus 14.31 (2.80) 10.48 (1.62) 7.50 (0.97) 0.90 (0.18) 404 MADRONO [Vol. 61 A C B D species species Fig. 4. Comparison of leaf characteristics among select Ceanothus species from the northern San Francisco Bay region. 4A) leaf length and standard errors; 4B) leaf width and standard errors; 4C) number of teeth per side of leaf with standard errors; 4D) length of teeth with standard errors. Species include son = C. sonomensis\ dec = C. decornutus\ jep = C. jepsonii var. jepsonii; alb = C. jepsonii var. albiflorus\ pur = C. purpiireiis; glo = C. gloriosus var. exaltatiis. origin of these taxa (Hardig et al. 2002; Burge and Manos 2011; Burge et al. 2011, 2013). For example, Hardig et al. (2002) investigated two Ceanothus taxa of presumed hybrid origin and their putative parents but concluded the possibil- ity of allopatric speciation was just as likely. Burge et al. (2011, 2013) also could not geneti- cally fully separate two closely related Ceanothus taxa on adjacent but different soil types and concluded that soil or other abiotic processes were driving species boundaries. Thus, while the variable floral color and fruit morphology in C decornutus may suggest a history of hybridization or introgression, equally likely scenarios are allopatric speciation or incomplete lineage sorting. Currently, determin- ing the exact origin of this entity may be difficult, but this study “opens the way for further studies and observations” of these North Bay taxa as suggested by Howell (1939) when he originally described several of them. Further studies should incorporate more taxa (e.g., C. confusus and C divergens) and morphological characters, as well as including molecular or other data. Ceanothus taxa in the North Bay region represent a potentially ideal set of species to understanding patterns and modes of evolution within Califor- nia’s radiation of species in this genus. 2014] PARKER: NEW MARIN COUNTY CEANOTHUS 405 PC1 Fig. 5. Principal components analysis of leaf characters for select Ceanothus species from the northern San Francisco Bay region. Legend abbreviations as in Figure 4. Special Status Consideration Many of the stands of C. decornutus are found on property owned by the Golden Gate National Recreation Area and managed by Pt. Reyes National Seashore. The total area occupied by Ceanothus patches overall is under 0.25 km^ and the majority of populations are on other adjacent property and some are within a development to the north end of the ridge. Currently, these populations do not seem at risk except within the development should the development expand in the future. Historic cattle management practices in the area have preserved this taxon as well as other rare species co-occurring at the same site both on public lands as well as on private lands nearby. At the same time, this is a rare species with relatively limited distribution, and it should be accorded special status. This will potentially be one of the rarest and most endangered of all the Ceanothus species. Purchasing lands or development rights from willing sellers might be a consideration. Key to C. Decornutus and Spiny-Leaved Erect Ceanothus Shrubs of Northern SF Bay Region 1. Leaves alternate, stipules thin and deciduous {Ceanothus section Ceanothus) V. Leaves generally opposite, stipules thick, corky and persistent {Ceanthothus section Cerastes) 2. Leaf blade generally flat or convex to concave adaxially, leaf margin entire or teeth ± sharp, but not spine-like 3. Leaves broadly ovate to widely elliptic, base round, edge not revolute, usually 13-17 small teeth well distributed along the edge . .C. gloriosus (several varieties) y . Leaves narrowly ovate to oblong, base usually cuneate, edge slightly revolute, 3-5 teeth generally above the middle of the leaf C. confusus 2' . Leaf blade ± wavy to ± folded lengthwise, leaf margin with spiny teeth 4. Leaf blade elliptic to narrowly obovate C. divergens 4'. Leaf blade widely elliptic to ± round 5. Leaves ± spreading, margin 3-5-toothed C. sonomensis 5' . Leaves generally reflexed, margin 7-15-toothed 6. Sepals, petals 6(-8); fruit with prominent wavy ridges, horns wrinkled, 1.5- 3 mm; generally rocky serpentine soils C. jepsonii 6'. Sepals, petals 5; fruit generally smooth, ridges, if present, generally straight, horns 1-1.3 mm; multiple soil types 406 MADRONO [Vol. 61 7. Flowers dark blue to purple, leaves with 11-13 teeth, the teeth 2-3.2 mm long; volcanic substrates C purpureus 7'. Flowers light blue to white, leaves with 7-9 teeth, the teeth 1-2.2 mm long; serpentine soil C. decomutus Acknowledgments Roger Raiche first realized this population might be different from Cecmothus jepsonii and brought it to the attention of Terri Thomas, who now is in charge of natural resources at the Presidio, but worked for the Golden Gate National Recreation Area at that point. Literature Cited Burge, D. O. and P. S. Manos. 2011. Edaphic ecology and genetics of the gabbro-endemic shrub Cecmothus voderickii (Rhamnaceae). Madrono 58:1-21. , D. M. Erwin, M. B. Islam, J. Kellermann, S. W. Kembel, D. H. Wilken, and P. S. Manos. 2011. Diversification of Cecmothus (Rhamnaceae) in the California Floristic Province. International Journal of Plant Science 172:1 137-1 164. , R. Hopkinds, Y. -FI. E. Tsai, and P. S. Manos. 2013. Limited hybridization across an edaphic disjunction between the gabbro-endemic shrub Cecmothus roderickii (Rhamnaceae) and the soil-generalist C cuuecitus. American Journal of Botany 100:1883-1895. Hardig, T. M., P. S. Soltis, D. E. Soltis, and R. B. Hudson. 2002. Morphological and molecular analyses of putative hybrid speciation in Cecmothus (Rhamnaceae). Systematic Botany 27:734—746. , , AND . 2000. Diversification of the North American shrub genus Ceanothus (Rhamnaceae): conflicting phylogenies from nucle- ar ribosomal DNA and chloroplast DNA. Amer- ican Journal of Botany 87:108-123. Howell, J. T. 1939. Studies in Cecmothus - 1. Leaflets of Western Botany 2:159-165. . 1970. Marin Flora, manual of the flowering plants and ferns of Marin County, California. 2nd ed. with supplement. University of California Press, Berkeley, CA. , F. Almeda, W. Follette, and C. Best. 2007. Marin Flora, an illustrated manual of the flowering plants, ferns and conifers of Marin County, California. Revised illustrated edition. California Academy of Sciences and California Native Plant Society (Marin Chapter), San Fran- cisco, CA. Jeong, S-C., a. Liston, and D. D. Myrold. 1997. Molecular phylogeny of the genus Cecmothus (Rhamnaceae) using rheC and ndh¥ sequences. Theoretical and Applied Genetics 94:852-857. Nobs, M. A. 1963. Experimental studies on species relationships in Cecmothus. Carnegie Institution of Washington, Publication 623, Washington, DC. R Core Team. 2012, R: a language and environment for statistical computing. R Eoundation for Statis- tical Computing, Vienna, Austria, ISBN 3-900051- 07-0, Website http://www.R-project.org Madrono, Vol. 61, No. 4, pp. 407-410, 2014 REVIEW Trees in Paradise: A Cal- ifornia History. By Jared Farmer. 2013. W.W. Nor- ton & Company, New York, NY Acknowledge- ments; maps, figures, pho- tographs; appendix of common and scientific names of species; further reading; notes; index, 552 pp. ISBN 978-393-2. $35.00, hardcover This sizable book, well-written and nicely researched, stoutly bound and illustrated with four sizable galleries of photographs printed in black-and-white, is an admirably recounted “natural and unnatural history of California trees” (p. xx). Inevitably, the subject species picked will chafe those most knowledgeable about California botany and forest history who will dispute the trees that author Jared Farmer, a prize-winning historian, chooses to write about in Trees in Paradise: A California History. As Fanner notes in a concise preface, California includes the oldest, tallest, and biggest trees in the United States. Three of the four varieties he singles out — eucalypts, citrus, and palms — are exotics; only the redwood and giant sequoia are native to California, and even those are sometimes described as refugial species, survivors of a more beneficent climate and habitat of long ago. This engaging read of eight chapters has two for each “tree” — one looking at the nineteenth century, the second concerned with the later twentieth century up through today, an organization that adds background and currency to each study. The fact is, this book is rather more a cultural history of California than it is a study of specific trees, and Farmer is interested in the reshaping with exotic species of the state’s landscapes in the interest of a triumvirate of causes: aesthetics (eucalypts), economics (citrus, almost solely the navel orange), and semiotics (palms with their Southern California symbology) (p. 432). Madrono readers may pause, no doubt consumed by the same sorts of doubts that afflicted me in considering Farmer’s choices. What of the bristlecone pine, witness to 10,000 yrs of rigorous climate change? Or of the California bay, or the buckeye? And of the Monterey pine or the same region’s cypress, what? And how, especially, could varied and signature oak species not feature among Califor- nia’s select trees? No doubt Farmer has faced these challenges and occasional sputters of doubt in press tours and interviews. Yet as a reviewer, I’m given to repeat an adage that I heard one of my teachers, James E. Vance Jr., chuff in aggravation when the reviewer of a book Jay had recently published went on in excruciating and self-aggrandizing detail about how he, the review- er, would have taken on the subject: a book review best assesses an author’s success or failure with the topic as written about, rather than teeing off at length about how the reviewer might have chosen to approach the theme. And in bits and pieces throughout Trees in Paradise, Farmer does a good job of laying out the reasoning behind his decision to consider trees that brought a different kind of prosperity, and in particular a distinctive look and feel, to California. To Farmer’s way of thinking, trees connote the California Dream, and the visions of an ideal landscape that colonists coming into California brought with them — though presumably, he is concerned mostly with occupiers of a northern European stripe, perhaps American-British most of all. These settlers were not, he suggests, enamored of a setting relatively bereft of trees. It wouldn’t be hard to argue that Farmer’s predilection for certain trees is bound up in a lot of preconceptions. After all, settlers of Spanish-Mexican origin, who came to Alta California in 1769, were actually very much used to a dehesa of oaks so characteristic of south- western Spain and Portugal, and Mexico has the largest variety of Quercus L. species anyplace on earth (Campos et al. 2013). Farmer claims, “[p]ost-Gold Rush settlers did not feel content with the existing landscape subtly modified by Indians, Spaniards, and Mexicans. It looked deforested. It looked unfinished” (p. 117). Grass- lands abounded; so did coastal shrub lands and chaparral. As the author suggests: From roughly 1 850 to 1 950 — California’s first hundred years as a state — American horti- culturists planted innumerable trees in former- ly shadeless locales. To use an old-fashioned term, they emparadised the land. They import- ed a profusion of ornamental and commercial species and varietals and created moneymak- ing orchards and picturesque tree-lined streets. In short, tree planters staged a landscape revolution. By the mid-twentieth century, eucalypts defined the look of lowland Califor- nia, oranges dominated Southland agriculture, and palms symbolized Los Angeles, (p. xxviii) Certainly, as the Spanish-Mexican era gave way to the “American,” parts of the California landscape underwent a deliberate and profound 408 MADRONO [Vol. 61 renovation; I would simply note that, Native American forest management aside, eight decades of Hispanic presence, use, and coexistence transpired before Anglo-Americans aggressively began trying to undo a Mediterranean landscape. That is a quarrel, though maybe more of a nit- pick. Certainly Farmer’s work deserves a com- mentary for how he treats each of the four “trees” chosen. Redwoods and the giant sequoia — coastal and gallery or riparian forests, and those from the edges of the west-central Sierra — were a formidable challenge for axe-wielders and indus- trial foresters, once an enormous amount of clean, near-perfect, and hugely durable wood was recognized as a resource after 1848. Removing those trees and protecting some remnant of them posed two separate challenges dealt with admi- rably in this book. In fact, new techniques in logging and timber movement had to be invented, just as novel means of preservation and legal protection would develop across a century, from 1870’s to 1970’s. Nonetheless, a hunger for wood and for the trophy rounds that could be sawed from the giant sequoia brought down many a tree, including some name-plaqued and com- memorated by earlier colonizers in the Sierra foothills. As a good history should. Trees in Paradise offers up no shortage of facts. For example, from the Sierra east of Fresno 1890- 1910, “loggers felled roughly one-quarter of all mature sequoias in California — that is, the world,” moving timber by way of a massive flume — 54 mi long and dropping 4200 ft in elevation, that earned Sanger at the terminus of the flume in the San Joaquin Valley the title of “Flumeo- polis of the West” (pp. 44, 43). Lauding the size of trees and scale of harvest was not without creepy undertones: the turn of the twentieth century was high season for eugenicists — foresters, fellers, and scientists included — and the language of race improvement and prodigal- ity was much a part of the redwood revelry. That would take a drastically different turn after World War II, when technology made it easier to remove some of the very largest coastal redwoods, and hijinks of the financial services industry took stock resources, including unhar- vested coastal redwoods held carefully in reserve by select-harvesting companies, and sought to cash out by clear-cutting entire groves of the tallest and oldest trees. The galvanizing effect of the Redwood Summer of 1990 and the Headwaters conflict earns a characteristic Farmer remark: “[i]n the minds of the ‘Freedom Riders for the forest’ — white people easily identified by countercultural accouterments like African talking drums and crocheted Rasta beanies — the rights of wild trees seemed analogous to the rights of black southerners” (p. 96). But “[o]ld- growth protectors were generally tone-deaf to Humboldt County’s culture of producerism” | (p. 97). The MAXXAM era with the looming | figure of Charles Hurwitz, the demise of Pacific | Lumber and the company town at Scotia, the j usurpation of selective logging by clear cutting, j the arrival of EarthFirst!ers and the ascent of j tree-sitter Julia “Butterfly” Hill into the redwood named Luna, are each part of the late twentieth- I century story. And as Farmer notes, times j change, sometimes startlingly: “[o]n the North Coast, America’s THC nerve center, cannabis now rivals redwoods as the region’s leading export” (pp. 104-105). | Various species of eucalyptus were brought into California, though principally the Tasma- nian Blue Gum that accounted for about 90% of California’s eucalyptus plantings. While Farmer doesn’t emphasize it deeply, the introduction | came in no small measure because of active j sharing between Australia and California during ' the Gold Rush years, and the eucalyptus was so I prevalent in Tasmania and Australia that adding [ trees to California, especially in moist areas j where eucalypts, with their volatile oils, was j considered virtuous and health-giving. That turned out to have more to do with the vast amounts of water taken up and transpired by the thirsty trees, drying out a seasonally soaked landscape, but nonetheless, the eucalyptus was for a time considered a hero in nineteenth-century battles against malaria in the Sacramento and San Joaquin valleys. And the trees did in their own way become nearly universal. Several frames from the famed “Migrant Mother” sequence of photographs taken by Dorothea Lange during the Depression actually show a sizable eucalyptus grove behind the shanty where mother and children are sheltered, leftover from a speculative plantation. Much of the intent behind eucalyptus planting was the creation of woodlots — and so many were planted, in such density, that trees are still prominent on the California scene. Specula- tive investor Frank Havens from 1910-1913 purchased sections of the Berkeley-Oakland hills and started a plantation there as part of his Mahogany Eucalyptus and Land Company. It went broke and back into public hands not long after, but eucalyptus species and Monterey pines still carpet the East Bay Hills (pp. 143-144), a major management quandary for the East Bay Regional Park District and the University of California, Berkeley, which now control much of the property and are attempting a return to native species. Leland Stanford, a former gover- nor and railroad magnate, decided to plant part of his Palo Alto Stock Farm to eucalyptus species, but the plan hiccupped with delays until 1916, when a couple of groves of Eucalyptus globulus Labill. were installed, to be amended by other varieties, and “[i]n the mid-twentieth century, university arborists added many other 2014] BOOK REVIEW 409 types of gums to the sprawling [Stanford University] campus; an arboreal census in 1984 counted ninety-four Eucalyptus species” (p. 162). The campus of UC San Diego on Torrey Pines Mesa was established on a former eucalyptus plantation (p. 161), and UC Berkeley’s West Gate is marked by enormous trees, planted 100+ years ago when eucalyptus was considered a boon to the eyes and a benison to the sense of smell. The trees brought good and bad. The good was fundamental: “[t]hat most Californian of mod- ern-day activities, driving alone on a highway, windows down, approaches perfection with the help of a blue gum canopy. The light is breathtaking, the smell invigorating. Who wouldn’t wish for a convertible?” (p. 217). A downside was the 1991 Berkeley-Oakland Hills fire, which I witnessed: “[f]reezes, droughts, and dry downslope winds occur naturally in the East Bay. Not true of blue gum, which covered some 20 percent of the burned area [1520 acres; 2500 single-family dwellings; 25 lives lost] and contrib- uted an estimated 70 percent of the fuel load” (p. 179). As Farmer notes, fear of fire, and a marked distaste for exotics, blossomed in late twentieth-century California into a kind of “botanical xenophobia” (p. 207), and in a place so militantly pluralistic (and rich) as Marin County, which would never consider an anti- immigrant movement, there are regular efforts to purge eucalypts, brooms, and other exotics from the landscape — surely species-ism, of a sort. About citrus there is so much in print that Farmer has his hands full, and having written at length about the economics and symbolism of California orange groves 25 yrs ago (Starrs 1988), I found this the least inspiring quarter of the book. I was partly given pause by the author’s emphasis on a specific time and place: he’s most interested in the Southland from the 1870’s into the 1940’s, and not nearly so curious about the movement north into the San Joaquin Valley of citrus in the early twentieth century, establishing groves of navel and Valencia oranges in Porter- ville, Exeter, Orange Cove, Navelencia, Tulare, and all through fertile soils on bench lands in the eastern San Joaquin Valley. His is a porthole perspective, with those other sites largely left aside (except as home to a startling array of agricultural maladies). The star is Southern California, where oranges were the economic and cultural crop of choice, and palms the symbol of Mediterranean success. Orange culti- vation has Classical roots, in the Garden of the Hesperides, and orange culture was suggestive as an earlier and more blessed world where, in Farmer’s calculation, “[s]ociology mirrored ge- ography. The Citrus Belt — in effect, the wealth — of Southern California occupied the broad, inclined alluvial fans (known locally as ‘benches’ or ‘mesas’) beneath the San Gabriel and San Bernardino mountains” (p. 243). Even today, growers in the San Joaquin Valley speak in near-reverential tones about the California Fruit Growers Exchange, about Sunkist as a marketing and quality-control arm for the citrus industry, and they know all the principals of their local packing houses. Farmer, however, is interested in what was, and what was lost: the formation of irrigation colonies (technically, “mutual water companies,” p. 232), the ways that orange culture drew settlers through three or four generations of Southern California land and life, and — in a lovely phrasing, marked “the transition from tree culture to horticapitalism” (p. 269); in the requirements for climate modification in citrus culture in the Southland; workers were crucial to citrus, less for cultivation than for setting up smudge pots that would moderate the effects of frost or even the rare hard freeze on oranges. Finks drawn between the lighting of 3.3 million Southern California smudge pots in January 1932 and a rising concern about air pollution, allergies, and maladies afflicting field worker are not original, but Farmer artfully connects agricultural excess to urban quality of life concerns. In the 1950’s, high school football teams from Redlands and San Bernardino played for the smudge-pot trophy (p. 317). Equally daunting are horror stories about changes wrought: Riverside was the richest city in per capita income in the United States in 1895; by the early 1970’s, as oranges were all but gone from the scene, it was the “smog capital of the world” (p. 309). Pests abounded in the groves, and required extraordinary counter- measures. Farmer recounts a 1916 attempt to extirpate a scale insect using hydrocyanic acid gas (HCN) that consumed 4 months, eleven work gangs, the tenting of 383,500 trees on 4250 acres, and 1 1 railroad cars of cyanide (p. 286). These are tales not of love, but their tone of excess and peril does leave room for admiration, and Trees in Paradise rightly singles out the magnetic pull of orange groves, drawing visitors, then boarders, and eventually, permanent residents to Southern California who would swell town numbers, and reduce the possibilities of tree-crop agriculture until humans had displaced tree crops as a feature on the land. The 1974 film Chinatown, directed by Roman Polanski from a script by Robert Towne, charted a fictionalized version of the transition, and whether on film or bound page, the results of land use change and urban water seizure were anything but pretty. Eighteenth-century Franciscan friars brought date palms with them from Spain and Mexico to the missions of 1700’s California, in part to guarantee a supply of Palm Sunday fronds for cashiered soldiers, clerics, and Indian peons tilling and ranching along the chain of coastal missions. Palms were then, and remain now, symbolic features. When Henry Huntington 410 MADRONO [Vol. 61 wanted to embellish his estate in San Marino after the San Francisco earthquake of 1906, he had two Canary Island fan palms moved whole from San Francisco. There has long been, it turns out, a land-office business in freighting entire palms from one site to another, in no small measure because palms are not “trees,” as the term would conventionally have it, but monocots in the order Arecales Bromhead that can be exhumed and relocated with relative ease. Most palms in California are single-stemmed, many of them Washingtonia H. Wendl., and part of their story is an aggressive messiness, throwing off fronds, harboring any variety of rats, bats, and insect pests, encouraging vandalism when un- trimmed, and requiring no small amount of management and cleanup. Yet to many Califor- nians, palms are beloved, and symbolic of difference. Vocabulary is one of the wonders of this book. There are “scouts” or “spotters” whose job is locating desirable and available palm trees for transplanting; “palmskinners” adept at moving the trees; the “palmification” of a landscape; and botanists (including Berkeley pioneer, Willis Jepson) who in their disdainful view of palms as “skybusters” were staunch “antipalmists” (pp. 410, 396, 405, 423, 362). As Farmer notes, “[a] California palm is not just a plant. It is a signpost” (p. 408). Though here, again, I would note some provincialism in the author’s take on palms. As with citrus (oranges), his focus is on trees in a very specific area. Southern California. Those who have traveled the state more widely recognize a crucial role that palms, many dating from the late nineteenth century, play in rural life in the state — or even along shaded streets in such successful Ag centers as Modesto, Sacramento, Visalia, and Colusa. And through agricultural reaches of the state are thousands of homes, set back from a county road, whose presence is announced by a two- abreast echelon of palms leading from public road to private manse. Sometimes the house (and barnyard and outbuildings) may not even exist anymore — but the palms, or at least a suggestive number of them — announce what was once a symbol of rural pride and investment in place. Humor can give life, even to a serious book. There is hardly a better moment in Trees in Paradise than Farmer’s discussion of Randy Newman’s “I Love L.A.,” with its allusions to driving, watching, and diversity in the Southern California landscape (pp. 404^05). That album by the multiple Grammy- (nine nominations, five wins) and Oscar-winner (13 nominations, three wins) is titled Trouble in Paradise (1983), and a satirical Newmanesque spirit resounds through the best moments of Trees in Paradise. Two short selections may help round out Farmer’s story: “[o]f the myriad kinds of trees propagated by the million in nineteenth-century California, euca- lypts, citruses, and palms had the most signifi- cant, long-lasting effects” (p. 432). Trees, in his reckoning, embody a view that “[t]he pith of the California Dream is the idea that the Golden State is different, special, unique, unprecedented” (p. xxix). While quibbling about the selection of trees is possible, to do so might not be wise. This is a good telling, and readers could learn a lot by spending some quality time with Fanuer’s narrative of arboreal landscape change as a handmaid or telltale of cultural history. — Paul F. Starrs, Department of Geography, Univer- sity of Nevada, Reno, MS 0154, Reno, NV 89557-0154. starrs@unr.edLi. Literature Cited Campos P., L. Huntsinger, J. L. Oviedo, P. F. Starrs, M. DIaz, R. B. Standiford, and G. Montero (EDS). 2013. Mediterranean working oak woodland landscapes: dehesas of Spain and ranchlands of California, Springer- Verlag, Berlin. Starrs, P. F. 1988. The Navel of California and other oranges: images of California and the orange crate. California Geographer 23:1-41. Madrono, Vol. 61, No. 4, pp. 411-412, 2014 NOTEWORTHY COLLECTION CALIFORNIA Opuntia X CHARLESTONENSIS Clokey (pro. sp.) (CACTACEAE). — San Bernardino Co., Mescal Mountains, Mescal Range 7.5’ USGS quad, 35.43 1°N, 115.529°W (WGS84), 1600 m elevation, 200 m SSE of the Iron Horse Mine, 24 km WSW of Nipton, 21 Oct 2008, M. A. Baker 16749 (RSA 765617!, RSA 765618!). Chromosome number determination by M. Baker of In = 55 during meiosis of microgametogenesis. Approx- imately 50 individuals within a one hectare area along wash on southeast slope in Coleogyne ramosissima Torr. scrub with Aloysia wrightii A. Heller, Aristida purpurea Nutt., Atriplex canescens (Pursh) Nutt., Bouteloiia curtipendula (Michx.) Torr., B. eriopoda (Torr.) Torr., Cylindropuntia acanthocarpa (Engelm. & J.M. Bigelow) F.M. Knuth, Dasyochloa pulchella (Kunth) Willd. ex Rydb., Echinocereus engehnannii (Parry ex Engelm.) Lem., E. mojavensis (Engelm. & J.M. Bigelow) Rum- pier, Encelia frutescens (A. Gray) A. Gray, Ephedra viridis Coville, Eriogonum inflatum Torr. & Frem., Gutierrezia sarothrae (Pursh) Britton & Rusby, Juni- perus californica Carriere, Menodora scoparia Engelm. ex A. Gray, M. spine scens A. Gray, Miihlenbergia porteri Scribn. ex Beal, Opuntia basikiris Engelm. & J.M. Bigelow, O. phaeacantha Engelm., O. polyacantha Haw. var. erinacea (Engelm. & J.M. Bigelow) B.D. Parfitt, Purshia stansburyana (Torr.) Henrickson, Rhus trilobata Nutt., Scutellaria mexicana (Torr.) A.J. Paton, Salvia dorrii (Kellogg) Abrams, Sphaeralcea ambigua A. Gray, Thymophylla pentachaeta (DC.) Small, Tridens muticus (Torr.) Nash, Yucca baccata Torr., and Y. brevifolia Engelm. Previous knowledge. This species was originally collected by Clokey in 1938 (Clokey 1943). Clokey’s collection and all collections of this species prior to M. A. Baker 16749 were from Kyle Canyon on the east side of Mount Charleston in the Spring Mountains of Clark County, Nevada, on limestone slopes at eleva- tions over 2100 m (7000 ft). Opuntia X charlestonensis has a confusing taxonomic history, having been considered a variety of O. phaeacantha (Backeberg 1958) or simply a synonym of O. phaeacantha (USDA NRCS 2014). Benson (1969) placed populations of O. X charlestonensis under the name O. littoralis (Engelm.) Cockerell var. martiniana (L.D. Benson) L.D. Benson, the distribution of which he defined as “southern California in the New York Mountains, eastern San Bernardino County; Nevada in Lincoln and Clark counties; Utah near the Arizona border; northern Arizona” (Benson 1969, pg. 142). He further suggested that “the variety hybridizes with Opuntia erinacea and shades into O. phaeacantha, O. macrorhiza, and especially O. violacea” (Benson 1969, pg. 142). Parfitt (1980) rejected Benson’s circumscrip- tion and provided ample morphological and cytological data that indicated O. martiniana (L.D. Benson) B.D. Parfitt was a species separate from O. littorcdis. However, he made no reference to O. X charlestonensis. More recent taxonomic work has shown that Opuntia X charlestonensis is consistently pentaploid {n = 55/2) (Baker et al. 2009). In all areas where it has been collected, it is sympatric with hexaploid (/? = 33) O. phaeacantha and tetraploid (/? = 22) O. polyacantha var. erinacea. The morphological characteristics of O. charlestonensis are intermediate between these two putative parent species but are more similar to O. phaeacantha. This led Baker et al. (2009) to conclude that O. X charlestonensis is most likely a nothospecies, with its pentaploid genome resulting from the combi- nation of normal gametes of O. phaeacantha and O. polyacantha var. erinacea. Significance. This is the first record of Opuntia X charlestonensis from California. It is also the first record of the species anywhere outside of the Kyle Canyon area of Mount Charleston in the Spring Mountains of Clark County, Nevada. This record represents substan- tial increases in both the geographic and elevational ranges of the species. The California population of Opuntia X charlesto- nensis most likely represents a separate hybridization event between O. phaeaeantha and O. polyacantha var. erinacea. Morphological evidence for a separate hy- bridization event includes the larger, thinner stems of the California population. Within the California population, at least some the local O. phaeacantha individuals appear to have morphology intermediate to that of O. engelmannii Salm-Dyck ex Engelm. This may explain the larger, thinner stems of the local O. X charlestonensis. Due to its taxonomic uncertainty, Opuntia X charlestonensis currently has no special species status in Nevada or elsewhere. The hypothesized hybrid origin of the species also raises doubt regarding the necessity of conservation, although the hybridization events may have occurred in the distant past. Individuals of O. X charlestonensis apparently reproduce via apomictic seeds (Baker et al. 2009), which may account for their large population in Kyle Canyon and the local abundance of individuals in the Mescal Mountains. This putative nothospecies and others like it, such as O. X curvispina Griffiths (pro. sp.), are a challenge to both taxonomists and conservationists because they repre- sent a gray area between re-occurring spontaneous hybrids and well-established species. Further under- standing of whether these oddities are evolutionary dead-ends or incipient pioneers is vital to the under- standing of evolutionary mechanisms. The primary threat to populations of Opuntia X charlestonensis is most likely fire. The Carpenter Fire in summer 2013 in the southern Spring Mountains came within a few kilometers of all of the previously known O. X charlestonensis populations and could easily have devastated the species. It is unclear whether the new population in California falls on BLM land or in the Mojave National Preserve. If it is on BLM land, it may be vulnerable to expansion of the mining operation five km to the north. There are many areas throughout southern Califor- nia where Opuntia phaeacantha and O. polyacantha var. erinacea co-occur, and these areas may well contain undiscovered populations of O. X eharlestonensis. Efforts to find new populations would greatly help elucidate the taxonomic status of this species and determine how best to protect it from the vulnerabilities inherent to species of restricted ranges. 412 MADRONO [Vol. 61 — Marc A. Baker^ Arizona State University, Main Campus, College of Liberal Arts and Sciences, School of Life Sciences, P. O. Box 874501, Tempe, Arizona 85287-4501 USA. 'mbaker6@asu.edu; and Michelle A. Cloud-Hughes^, Desert Solitaire Botany and Ecological Restoration, San Diego, CA 92103. ^mcloudhughes@gmail.com. Literature Cited Backeberg, C 1958. Die Cactaceae, Handbuch der Kakteenkunde, Vol. 1: Einleitung und Beschrei- bung der Peireskioideae und Opuntioideae. G. Fischer, Jena. Baker, M. A., J. P. Rebman, B. D. Parfitt, D. J. PiNKAVA, AND A. D. ZIMMERMAN. 2009. Chro- mosome numbers in some cacti of western North America - VIII. Haseltonia 15:117-134. Benson, L. D 1969. The native cacti of California. Stanford University Press, Stanford, CA. Clokey, I. W 1943. Notes on the flora of the Charleston Mountains, Clark County, Nevada. V. Cactaceae. Madrono 7:67-76. Parfitt, B. D 1980. Origin of Opuntia curvospina (Cactaceae). Systematic Botany 5:408^18. USDA, NRCS. 2014, The PLANTS Database (http:// plants.usda.gov, February 2014). National Plant Data Team, Greensboro, NC 27401-4901 USA. Madrono, VoL 61, No. 4, p. 413, 2014 NOTEWORTHY COLLECTION CALIFORNIA Froelichia gracilis (Hook.) Moq. (AMA- RANTHACEAE). — Shasta Co., along both sides of Collyer Drive and College View Drive (frontage roads on the N and S sides of State Hwy 299) for a distance of 1400 m centered between the Hwy 299/Old Oregon Trail interchange and the Hwy 299/Chum Creek Road inter- change, City of Redding, 40.617896°N, 1 22.3339 16°W, elev. 200 m, 4 Nov 2013, J. Luper 001 (JEPS), identified by Donald Burk and confirmed by Mihai Costea. The population consists of several thousand plants estab- lished along gravelly road shoulders; associates include other ruderal species such as Cynodon dactylon (L.) Pers., Dittrichia graveolens (L.) Greater, Eragrostis cilianensis (All.) Janch., Erodium cicutarium (L.) Aiton, Erodium botrys (Cav.) Bertol., Heiiotropium europaeum L., Kickxia elatine (L.) Dumort., and Senecio vulgaris L. Previous knowledge. Froelichia gracilis (slender snake- cotton, cottonweed) is known as an historical waif in California based on specimens collected along the Santa Fe Railway in San Dimas, Los Angeles Co., June 1955 (CCH 2013). The species is native to the mid western U.S., and extends from Iowa, Nebraska, and Colorado, south to Texas, New Mexico, and Arizona (Kartesz 2013) and into Mexico (McCauley 2003). The species has now spread to most states east of the Mississippi River (Kartesz 2013). Blake (1956) attributed railroads as the principal dispersal mechanism for the eastern spread of the species. Froelichia gracilis is also naturalized in Europe, Japan, the West Indies, and in Queensland, Australia (Csurhes and Zhou 2008). Froelichia gracilis is listed as invasive by the State of Connecticut (Connecticut Invasive Plant Council 2011). Significance. This is the first report in California since 1955, and the first report ever in northern California, representing an 800 km range extension. Based on the number of individuals and distribution along the roadways, this population of F. gracilis appears to have been present for at least several years. Froelichia gracilis is adapted to open disturbed habitats, with sandy or gravelly soils, such as roadsides, railways, and degraded pastures (Blake 1956, Csurhes and Zhou 2008). These conditions are broadly represented throughout Califor- nia. Similar to the spread associated with railways, this occurrence adjacent to State Hwy 299 could contribute to the species’ range expansion in California. McCauley (2003) notes that the species’ adaptation to open sandy or gravelly soils should restrict its spread to open sites with poor soil. A similar finding was made regarding the potential spread of the species in Australia, with overgrazed pastures being identified as likely coloniza- tion sites (Csurhes and Zhou 2008). — John Luper and Donald Burk', ENPLAN, 3179 Bechelli Lane Suite 100, Redding, CA 96002. ^dburk@enplan.com. Acknowledgments We thank Mihai Costea for confirming identification of the species. Literature Cited Blake, S. F. 1956. Froelichia gracilis in Maryland. Rhodora 58:35-38. Connecticut Invasive Plant Council. 2011. Con- necticut Invasive Plant List, October 2011. Consortium OF California Herbaria (CCH). 2013. The Consortium of California Herbaria database. Data provided by the participants of the Consor- tium of California Herbaria. Website http://ucjeps. berkeley.edu/consortium/ (accessed 19 November 2013). Csurhes, S. and Y. Zhou. 2008. Plant Risk Assess- ment. Cotton-tails: Froelichia floridana and F. gracilis. Queensland Government Department of Primary Industries and Fisheries. Kartesz, J. T., The Biota of North America Program (BONAP). 2013. North American Plant Atlas. Chapel Hill, N.C. (maps generated from Kartesz, J. T. 2013. Floristic Synthesis of North America, Version 1.0. Biota of North America Program (BONAP). [in press]). Website: http:// www.bonap.org/napa.html (accessed 19 November 2013). McCauley, R. A. 2003. Froelichia. Pp. 442M45 in Flora of North America Editorial Committee (eds.). Flora of North America north of Mexico. Vol. 4: Magnoliophyta: Caryophyllidae, part 1. Oxford University Press, New York, NY. Madrono, Vol. 61, No. 4, pp. 414-415, 2014 NOTEWORTHY COLLECTION CALIFORNIA Lophocereus schottji (Engelm.) Britton & Rose (CACTACEAE). — Riverside Co., Chuckwalla Moun- tains, Pilot Mountain 7.5' USGS quad, 33.62°N, 1 15.32°W (WGS84), 482 m (1581 ft) elevation, approx- imately 1 km E of Corn Springs Campground, 14 km SE of Desert Center, 26 Oct 2013, M. A. Baker 17802 with M. Cloud-Hughes, A. Karl, and T. Thomas (ASU 289420!); 28 Sep 2013, /. Anderson 2013-05 with D. Bell and K Clark (ASU 289052 [n.v]); 28 Sep 2013, D. Bell 6031 with J. Anderson and K. Clark (RSA 809190 [n.v.]). Two individuals: one approximately 3 m tall X 2 m wide growing in Olneya tesota A. Gray in an east-facing drainage in rocky, gravelly, sandy granitic alluvium; second approximately 1.5 m tall X 1.5 m wide growing at the base of a dead Fouquieria splendens Engelm. at the west-facing base of a granitic outcropping in rocky, gravelly, sandy granitic alluvium with Allionia incarnata L., Ambrosia dumosa (A. Gray) W.W. Payne, A. salsola (Torr. & A. Gray) Strother & B.G. Baldwin, Chamae- syce polycarpa (Benth.) Millsp., Cylindropuntia ramo- sissima (Engelm.) F.M. Knuth, Dasyochloa pulchella (Kunth) Willd. ex Rydb., Encelia farinosa A. Gray ex Torr., Fagonia laevis Standi., Fouquieria splendens, Ferocactus cylindraceus (Engelm.) Orciitt, Krameria bicolor S. Watson, and Olneya tesota. Previous knowledge. This species was originally collected by Schott in Sonora, Mexico in 1855 (MO 2015496!) and published under the basionym Cereus schottii Engelm. (Engelmann 1856). The taxonomy of the species has changed several times and remains a matter of debate (Terrazas and Loza-Cornejo 2002, Arias et al. 2003). It was changed to Lophocereus schottii (Engelm.) Britton & Rose by Britton and Rose in 1909 (Britton and Rose 1909). In 1984 the International Organization for Succulent Plant Study formed the International Cactaceae Systematics Group in an attempt to come to a consensus on taxonomy within the Cactaceae (Anderson 2001). The group lumped Lophocereus (A. Berger) Britton & Rose and several other genera under the genus Pachycereus (A. Berger) Britton & Rose, thereby changing the species name to Pachycereus schottii (Engelm.) D.R. Hunt (Hunt 1987). However, more recent work combining genetics and morphology suggests that Pachycereus is a paraphyletic genus and that several genera, including Lophocereus, should be resurrected (Arias et al. 2003, Arias and Terrazas 2009). Lophocereus schotti is a basally-branched columnar cactus mainly found in Baja California, Sinaloa, and Sonora, Mexico. Until this discovery, the northernmost occurrences were in southern Arizona in Organ Pipe National Monument. These two Corn Springs individuals were found in March 2013 by Buford Crites and Cameron Barrows. The discovery was reported by Kate Barrows, president of the Riverside-San Bernardino chapter of the Califor- nia Native Plant Society (pers. comm., March 2013). Surveys of the area by several groups of botanists have not discovered any other individuals. Significance. This is the first record for Lophocereus schottii in California. The closest record for the species is in Canon de los Torrentes, Baja California, 40 km south of the international border. This record therefore represents a northern range extension of over 140 km. This is also the first record for any Lophocereus species for California. This discovery has occasioned much lively discussion regarding the possible origin of these disjunct plants. The first possibility is that they represent the last of a relict population. This hypothesis is supported by Stebbins and Major (1965), which identified the northern and eastern edges of the Colorado Desert as having a high number of relict species. A second possibility is that the species was brought to the Com Springs area by Native Americans. For thousands of years, Corn Springs served as a temporary home for several tribes of nomadic indigenous peoples (Bureau of Land Management 2014). The native peoples of southern Arizona and northern Baja, where the species is more common, ate the fruit and used the pulp for various medicinal purposes (Rebman and Roberts 2012). It is therefore possible that Lophocereus schottii made it to Com Springs through trading amongst tribes. The third possibility is that miners, who occupied the Com Springs area from the early 19th century through the early 1980’s, planted Lophocereus schottii as an ornamental. There is a homestead area approximately 5 km west of Corn Springs Campground at Aztec Well where there are large, mature, presumably-planted saguaros. A thorough survey of the Aztec Well area has not been possible, due to the hazardous nature of exploring remote desert dwellings. The one homeowner we spoke with at Aztec Well, who was aware of the Lophocereus schottii individuals near the campground, assured the authors that there were no others at Aztec Well and that he had not seen any others in his many years of exploring the area. He also said that he knew of a botanist who was already aware of the individuals but did not reveal a name. His attitude toward unexpected visitors discouraged us from further inquiries. The largest immediate threat to this disjunct popu- lation is collection for horticulture. Both plants are very close to a road, and both plants have had stems removed in the past. Both Lophocereus schottii individuals appear to be in good health. Although it is difficult to estimate the age of cacti, data from long-term growth measurements of Lophocereus schottii at Organ Pipe National Monument (Parker 1988) indicate that the large, mature individual is likely less than 50 years old, while the small, immature individual is probably less than 40 years old. The past poaching of stems on each individual makes accurate age estimates even more difficult. The mature plant flowered in late March 2013 and again in response to heavy summer rains in August/ September 2013. No fruit resulted from either of these flowering periods. Lophocereus schottii is not only an obligate outcrosser, but also it is largely pollinated by a particular moth, Upiga virescens (seiiita moth), which is closely associated with the cactus at every stage of its life cycle (Fleming and Holland 1998). In Fleming and Holland’s 1998 study, pollination by Upiga virescens 2014] NOTEWORTHY COLLECTION 415 accounted for approximately 75% of fruit set in Lophocereus schottii. A single flowering individual of an obligately out- crossing species cannot support a pollinator that relies on fruit development for its own reproduction. Halictid bees, which do occur at Corn Springs (USDA-ARS 2014), are also able to pollinate Lophocereus schottii, though much less successfully than the senita moth (Fleming and Holland 1998). But until the smaller Lophocereus schottii individual reaches sexual maturity, this relict population will remain static. Even if these two plants do eventually cross, their genetic diversity will remain very low. Unfortunately, unless there are several more Lopho- cereus schottii individuals hidden in the hills nearby, in the absence of a substantial propagation effort, this relict occurrence is bound for extinction. — Michelle A. Cloud-hughes^ Desert Solitaire Botany and Ecological Restoration, San Diego, CA 92103. * mcloudhughes@gmail.com; and Marc A. Baker^, Southwest Botanical Research, Chino Valley, AZ 86323. ^mbaker6@asu.edu Acknowledgments The authors would like to thank Andrew Sanders (UC Riverside Herbarium), Kate Barrows and Greg Suba (California Native Plant Society), Duncan Bell, Nick Jensen, and Michael Honer (Rancho Santa Ana Botan- ical Garden), Jim Andre and Tasha La Doux (Sweeney Granite Mountains Research Center), Tom Chester, Tim Thomas, and Neal Kramer for much thought-provoking discussion and debate on these enigmatic plants. Literature Cited Anderson, E. F. 2001. The Cactus Family. Timber Press, Portland, Oregon. Arias, S. and T. Terrazas. 2009. Taxonomic revision of Pachycereus (Cactaceae). Systematic Botany 34:68-83. , , AND K. Cameron. 2003. Phylogenetic analysis of Pachycereus (Cactaceae, Pachycereeae) based on chloroplast and nuclear DNA sequences. Systematic Botany 28:547-557. Britton, N. L. and J. N. Rose. 1909. The genus Cereus and its allies in North America. Contribu- tions from the United States National Herbarium 12:413^38. Englemann, G. 1856. Synopsis of the Cactaceae of the Territory of the United States and adjacent regions. Proceedings of the American Academy of Arts and Sciences 3:259-314. Fleming, T. H. and J. N. Holland. 1998. The evolution of obligate pollination mutualisms: senita cactus and senita moth. Oecologia 1 14:368-375. Hartmann, S., J. D. Nason, and D. Bhatta- CHARYA. 2002. Phylogenetic origins of Lophocereus (Cactaceae) and the senita cactus-senita moth pollination mutualism. American Journal of Bota- ny 89:1085-1092. Hunt, D. R. 1987. Bradleya: Yearbook of the British Cactus and Succulent Society 5:93. Parker, K. C. 1988. Growth rates of Stenocereus thurberi and Lophocereus schottii in southern Arizona. Botanical Gazette 149:335-346. Rebman, j. P. and N. C. Roberts. 2012. Baja California plant field guide. 3rd ed. San Diego Natural History Museum and Sunbelt Publica- tions, San Diego, CA. Stebbins, G. L., Jr. and J. Major. 1965. Endemism and speciation in the California flora. Ecological Monographs 35:1-35. Terrazas, T. and S. Loza-Cornejo. 2002. Phyloge- netic relationships of Pachycereeae: a cladistic analysis based on anatomical-morphological data. Pp. 66-86 in T. H. Fleming and A. Valiente-Banuet (eds.). Columnar cacti and their mutualists: evolu- tion, ecology, and conservation. University of Arizona Press, Tucson, AZ. United States Department of Agriculture, Agricultural Research Service. Bee Biology and Systematics Laboratory: Bee Biology and Systematics Laboratory. Accessed via http://www. gbif.org/occurrence/658 752050 on 2014-07-06. United States Department of the Interior, Bureau of Land Management. 2013. Corn Springs Campground, www.blm.gov/ca/st/en/fo/ palmsprings/corn_springs_campground.html. Ac- cessed April 2014. Madrono, Vol. 61, No. 4, p. 416, 2014 CBS PRESIDENT’S REPORT FOR VOLUME 61 The Council for the California Botanical Society has experienced considerable transitions in this our 101st year as a Society. We have ten new Council members for the next year. These new members were elected at the annual banquet and include, Mark Brunell, University of the Pacific, as the new President; J. Travis Columbus, Rancho Santa Ana Botanic Garden, as the 2nd Vice President; Nancy Morin, Flora of North America Association, as the Recording Secretary; Sheryl Creer, Insignia Environmen- tal, as the Corresponding Secretary; David Margolies, Franz Incorporated, as the Treasurer, Will Freyman, University and Jepson Herbaria, as the webmaster; Jessica Orozco, Rancho Santa Ana Botanic Garden, and Adam Schneider, University and Jepson Herbaria, as the Graduate student representatives; Dylan Burge, California Academy of Sciences, as the new Council Member at Large. Continuing on the council are myself, but now as Past President, Andrew Doran, as 1st Vice President, Kim Kersh, as Membership Chair, Staci Markos, Member at Large, Matt Ritter, as Editor of Madrono, and Lynn Yamashita, as Society Administrator. The perspectives brought by the new Council Members should invigorate the Society yet again. At the same time, we profusely thank the outgoing members for their hard work, especially for organizing the Centennial last year; these include Gene- vieve Walden, Dean Kelch, Anna Larsen, Tom Schweich, Rich Whitkus, Mike Vasey, and Bier Kraichak. Following the Centennial, the Council faced a series of hills to get over rather than the large mountain of the Centennial. As a consequence, we felt a bit at loose ends at first, but ended up with a busy and productive year. Much of the year was involved with transitions in the Council, beginning the organization of the Spring 2015 Graduate Student Meetings at Rancho Santa Ana Botanic Garden, and organizing the Spring 2014 Annual Banquet. A new initiative is to provide support for graduate students to attend the 2015 meeting. For those of you unable to make it, the annual banquet this year was held on Saturday, April 26, at the conference center of the Romberg Tiburon Center for Environmental Studies (San Franciso State University), near Tiburon (Marin County, California). During the day, Marin County locals Doreen Smith and Eva Buxton, along with Mike Vasey and Dean Kelch from the Council, led a field trip to the serpentine areas of Ring Mountain. That was followed by a Social and business meeting, after which we heard a wonderful talk by Dr, Susan Harrison (UC Davis). Her talk was on the “Effects of climate on serpentine plant communities.” Now that the Society is in its second century, we are increasingly digital and online. Please check out the changes in the website of the California Botanical Society (www.calbotsoc.org). The website has links, for example, to a large number of photos from the Centennial on our Facebook page, to YouTube videos of the Symposium talks, to Madrono, as well as information on the upcoming Graduate Student meeting. The central objectives of the Society when it was established were to promote the collection of new information about the flora of the West, to provide a vehicle for disseminating that information, and to educate the public about the value of conserving plant species and habitats found in wild areas throughout our region. The digital evolution of much of these objectives is reflected in our Society’s evolution over the last 10 yr. Because I am transitioning to the Past President position, this will be the last President’s Report you will receive from me. I have thoroughly enjoyed my time as President and the opportunity to work with hard-working and energetic colleagues on the Council. Thank you for this opportunity. At the same time, I want to remind you that our Society depends on the activities and engagement of its membership. We must broaden our reach to meet the challenges of an environment dominated by humans, work towards the conservation of natural habitats, and continue increase of understanding and knowledge of plant biology. V. Thomas Parker December 2014 Madrono, VoL 61, No. 4, p. 417, 2014 EDITOR’S REPORT FOR VOLUME 61 I am pleased to report the publication of volume 61 of Madrono by the California Botanical Society (CBS) in 2014 The publication of Madrono remains on schedule with an average time between initial submission and publication of about 8 months. I hope that Madrono continues to be viewed as the best outlet for western botanists to publish their work in a timely fashion, while reaching an interested and relevant audience. The efforts of numerous individuals are critical to the continued quality of the journal. Among these are our Noteworthy Collections editor, David Keil; the editorial assistant, Genevieve Walden; Steve Timbrook who has long provided the Volume Index and Table of Contents; Annielaurie Seifert at Allen Press; and the CBS executive council. I am also grateful to our contributors for their interesting and insightful manuscripts, and our reviewers who take time from their busy schedules to assess the quality of submitted work. This year we received 56 new manuscripts and 50 were accepted for publication. As the Editor, I have enjoyed my interactions with contributors and reviewers this past year and anticipate continued submissions of novel and exciting work. Matt Ritter September 2014 Madrono, Vol. 61, No. 4, p. 418, 2014 REVIEWERS OF MADRONO MANUSCRIPTS 2014 Denny Albert Krikor Andonian George Argus Peter Ball James Bartolome Robert Boyd Dylan Burge John Callaway Chris Campbell Benjamin Carter Kenton Chambers Gretchen Coffman James Cohen Allison Colwell Robert Cox David Diamond Lance Evans Kern Ewing Ezequiel Ezcurra Jennifer Funk David Gallagher Robert Graham Matt Guilliams Judy Harpel Uwe Hacke Bonnie Heidel Larry Hufford Kimberly Hunter Christopher T. Ivey Patricia Fall Glenn Keator David Keil Kenneth Kellman Jaun Larrain Federico Luebert Lucas Majure Marlon Marchado Jim Mauseth Joseph McAuliffe Dale McNeal Michael Mesler Robert Muller Tom Mulroy Robert Naczi Guy Nesom Ingrid Parker Robert Patterson Bruce Pavlik Chris Pires Jarmila Pittermann Mark Porter Robert Preston Nishanta Rajakaruna Jon Rebman Tony Reznicek Kevin Rice Joe Rocchio Stanley Seller Ernie Schuyler Rick Standiford Mark Stromberg John L. Strother David Torren Claudia Tyler David Wagner Dieter Wilken Dorde Woodruff Jenn Yost Madrono, VoL 61, No. 4, pp. 419-420, 2014 INDEX TO VOLUME 61 Classified entries; Major subjects, key words, and results; botanical names (new names are in boldface); geographical areas; reviews, commentaries. Incidental references to taxa (including most lists and tables) are not indexed separately. Species appearing in Noteworthy Collections are indexed under name, family, and state or country. Authors and titles' are listed alphabetically by author in the Table of Contents to the volume. Agavaceae (see Agave) Agave cieserti, noteworthy collection from NV, 316. Amaranthaceae (see Froelichia) Aquatic plants of Yellowstone Nat. Park, 159. Arizona (see Carnegiea and Erigeron) Asparagaceae (see Dichelostemma) Asteraceae: Baccharis salicifolia, riparian woodland zonation, 48; Coreopsis tinctoria, competition and niche requirements, 290; Erigeron lemmonii chromo some number and reproductive attributes, 9; variation in thistle seed characteristics and growth, 339. Noteworthy collection; Lorandersonia salicina from NV, 316. Astragalus rattanii war. jepsonianus, noteworthy collection from CA, 250. Baccharis salicifoUa, riparian woodland zonation, 48. Boraginaceae; Eremocarya, recognition of two species, 259; Phacelia, phylogenies and chromosome evolution, 16. New taxon; Nemophila hoplandensis, 308. Bouteloua ciirtipendula (see Coreopsis) Bryoflora of NV, contributions toward, 253. Cactaceae; Carnegki gigantea, epidemial browning, gir- dling, damage and bird cavities in northern Sonoran Desert, 115; Cylindropuntia mortality and extreme drought in the Sonoran Desert, 126; Discocactiis, morphology and development of sunken terminal cephalium, 194. Noteworthy collections from CA; Lophocereus schottii, 414; Opuntia X charlestonensis, 411. California; Baccharis salicifoUa, Saiix gooddingii and S. lasiolepis, riparian woodland zonation, 48; Ceanothus otayensis, soil chemistry and geographic distribution, 276; chaparral stem xylem cavitation vulnerability, 317; coastal prairie invasion by Holcus kmatus, 218; conifer foliar analyses on seipentine and gabbro soils in Klamath Mts., 77; Cordylanthus tenuis pallescens., identification and taxonomic status, 64; Devil’s Postpile Nat. Monument, CA, flora, 367; Dichelostemma lacima- vernaiis, morphologically distinct vernal pool terrain sp., 350; Phacelia, phylogenies and chromosome evolution, 16; Quercus agrifolia, Q. douglasii and Q. lobata, regeneration in central coastal CA, 1; Quercus serotiny, 151; riparian woodland plant zonation, 48; Saiix gooddingii seed dispersal, 388; Tritelia ixioides subsp. anilina, biology in coniferous forest, Butte Co., 87; weedy vs. non-weedy species, functional traits, 328. New taxa; Carex xerophila, 299; Ceanothus decornutus, 399; Cylindropuntia chuckwallensis, 231; Nemophila hoplandensis, 308; Tritelia piutensis, 227. Noteworthy collections; Astragalus rattanii var. jepso- nianus, 250; Froelichia gracilis, 413; Lophocereus schottii, 414; Opuntia X charlestonensis, 411; Taxus hrevifolia, 148. Campylopus intrqflexus, ecology and distribution of an introduced moss in western No. Am., 82. Carex xerophila, new sp. from no. CA, 299. Carnegiea gigantea, epidermal browning, girdling, dam- age and bird cavities in northern Sonoran Desert, 115. Cascade Range, OR; Fire effects in a montane wetland, 201. Ceanothus: C. otayensis, soil chemistry and geographic distribution, 276. New taxon; C. decornutus, 399. Chaparral stem xylem cavitation vulnerability, 317. Chromosome counts; Erigeron lemmonii, 12. Coastal prairie invasion by Holcus lanatus in CA, 218. Colorado (see Muscari) Conifer foliar analyses on serpentine and gabbro soils in Klamath Mts, CA, 77. Cordylanthus tenuis subsp. pallescens, identification and taxonomic status, 64. Coreopsis tinctoria, competition and niche requirements, 290. Crataegus douglasii, C. gaylussacia and C siiksdorfii stomata size and ploidy level, 177. Cryptantha (see Eremocarya) Cylindropuntia chuckwallensis, new sp. from CA; 231. Cylindropuntia mortality and extreme drought in the Sonoran Desert, 126. Cyperaceae (see Carex) Desert plant community structure in Baja Calif., MEX- ICO, 105. Devil’s Postpile Nat. Monument, CA, flora, 367. Dichelostemma lacima-vernalis, morphologically distinct vernal pool terrain sp., 350. Dicranaceae (see Campylopus) Discocactus, morphology and development of sunken temiinal cephalium, 194. Editor’s Report for Vol. 61, 417. Eremocarya, recognition of two species, 259. Erigeron lemmonii chromosome number and reproductive attributes, 9. Fabaceae (see Astragalus) Fagaceae (see Quercus) Fire effects in a montane wetland in OR, 201. Floras; Bryoflora of NV, contributions toward, 253; Devil’s Postpile Nat. Monument, 367. Eroelichia gracilis, noteworthy collection from CA, 413. Gabbro soil, foliar analyses of conifers on, 77. Hexalectris colemanii, noteworthy collection from NM, 149. Holcus lanatus, effect of simulated gopher disturbance on establishment in coastal prairie, 218. Hyacinthaceae (see Muscari) 420 Hydrophyllaceae (see Boraginaceae) Invasive plants: Variation in thistle seed characteristics and growth, 339; weedy vs. non-weedy species func- tional traits, 328. Keil, David J., Vol. 61 dedicated to, 419. Keys: Ccirex sect. Acrocystis of CA, 306; Ceanothus, erect with spiny leaves in northern San Francisco Bay region, 405; Cryptantha micrantha vars., 263; Cylindropuntia of USA with spiny fruit, 242; Eremocarya spp., 272; Scouleria in OR, 143; Tritelia spp. with stamens attached alternately at two levels, 230. Klamath Mts., CA, conifer foliar analyses on serpentine and gabbro soils, 77. Lophocereus schottii, noteworthy collection from CA, 414. Lorandersonia salicina, noteworthy collection from NV, 316. MEXICO: Baja California: geomorphic landforms and plant community structure and dominance in central desert region, 105. Moss (see Bryoflora of NV, Campylopus and Scouleria) Muscari neglectum, noteworthy collection from CO, 251. Nemophila hoplandensis, new sp. from CA, 308. Nevada (see Bryoflora) Noteworthy collections: Agave deserti and Loranderso- nia salicina, 316. New Mexico (see Hexalectris) Opuntia X char lest onensis, noteworthy collection from CA, 411. Orchidaceae (see Hexcdectris) Oregon: Fire effects in a Cascade Range montane wetland, 201. New taxon: Scouleria siskiyouensis, 137. Orobanchaceae (see Cordylanthus) Phacelia, phylogenies and chromosome evolution, 16. Plant zonation in riparian woodland in so. CA, 48. Ploidy level and stomata size in Pacific NW Crataegus, 177. President’s Report for Vol. 61, 416. Poaceae (see Bouteloua and Holcus) Quercus: Regeneration of Q. agrifolia, Q. douglasii and Q. lohata in central coastal CA, 1; serotiny in CA, 151. Regeneration (see Quercus) [Vol. 61 Restoration: Weedy vs. non-weedy species functional traits, 328. Reviews: Annotated Checklist of the Vascular Plants of Santa Cruz County, California. Second edition by Dylan Neubauer, 244; Baja California Plant Field Guide. 3''‘' Edition by Jon P. Rebman and Norman C. Roberts, 246; Flora of the Four Corners Region. Vascular Plants of the San Juan River Drainage: Arizona, Colorado, New Mexico, and Utah by Kenneth D. Heil et al., 248; Intermountain Flora: Vascular Plants of the Intermoun- tain West, U.S.A. by N. H. Holmgren, P. K. Holmgren, J. L. Reveal and Collaborators, 146; Trees in Paradise: A California History by Jared Farmer, 407; The Drunken Botanist. The Plants that Create the World’s Great Drinks by Amy Stewart, 144; Wildjlowers of Orange County and the Santa Ana Mountains by Robert L. Allen and Fred M. Roberts, Jr., 245. Rhamnaceae (see Ceanothus) Rosaceae (see Crataegus) Riparian woodland, plant zonation in so. CA, 48. Salicaceae (see Salix) Salix: S. gooddingii and S. lasiolepis riparian woodland zonation, 48; S. gooddingii seed dispersal, 388. Scouleria siskiyouensis, a new rheophytic moss of southern OR, 137. Scoulariaceae (see Scouleria) Seed dispersal in Salix gooddingii, 388. Seed retention (see Serotiny) Serotiny in CA Quercus, 151. Serpentine: Endemic Ceanothus decornutus, 399; foliar analyses of conifers on, 77. Soil chemistry and tree growth, 77. Sonoran Desert (see Carnegiea and Cylindropuntia) Stomata size and ploidy level in Pacific NW Crataegus, 111. Taxaceae (see Taxus) Taxus brevifolia, noteworthy collection from CA, 148. Themidaceae (see Dichelostemma and Tritelia) Thistle, variation in seed characteristics and growth, 339. Tritelia ixioides subsp. anilina, biology in coniferous forest, Butte Co., CA, 87; T. piutensis, new sp. from so. Sierra Nevada, CA, 227. Vernal pools (see Dichelostemma) Water stress: Stem xylem cavitation vulnerability of chaparral spp., 317. Weedy vs. non-weedy species functional traits, 328. Yellowstone Nat. Park, new aquatic plant records, 159. MADRONO Madrono, Vol. 61, No. 4, pp. 421^22, 2014 DEDICATION David J. Keil The California Botanical Society is pleased to dedicate Volume 61 to one of California’s best plant taxonomists, Dr. David J. Keil, in recognition of his tireless efforts toward a greater understanding of the California flora. Dave was born in 1946 with a plant key in his hand, or at least that’s how I imagine it. Over his career he has come to know the California flora so well that he is considered by many to be among the best for sight identification of plants in the state. Dave’s passions for botany were developed at Arizona State University where he earned his B.S. in 1968, and a M.S. in 1970. Dave went on to earn a Ph.D. from Ohio State University in 1973 while studying the relationships of Pec t is in the southwestern U.S. and Mexico. His career is defined by rigorous taxonomic research, new species discovery, the production of floras, and excellent teaching. Dave had numerous teaching appointments after the completion of his Ph.D., so it wasn’t until he joined the faculty at California Polytechnic State University in San Luis Obispo in 1976 that the California botanical community benefited from Dave’s keen eye and botan- ical expertise. His primary research focuses on the systematics of the Asteraceae, which has yielded many papers that have contributed to our understanding of California’s largest plant family. His work has included new species descriptions, new techniques in systematics, new taxonomic treatments, and primary research in ecology. His work spans all of North America, with particular attention to California and Arizona. His generic treatments span all tribes of the Asteraceae including Pectis L., Cirsium Mill., Isocarpha R. Br., and Gnaphalium L. Dave was appointed as the Director of The Robert F. Hoover Herbarium at Cal Poly in 1978, and there he has built an extensive collection. Dave is an avid plant collector, whether it be in his home county, elsewhere in California, or on one of his many family vacations throughout North America. Dave’s family has been actively involved in his collections, from pressing in the field to mounting in the herbarium. Dave’s collections now number at near 33,000, most of which are housed at the Hoover Herbarium at Cal Poly. These collections have served as the basis for his forthcoming update to the flora of San Luis Obispo County. As a result of extensive time spent in the field and herbarium, Dave has authored over 130 species descrip- tions mostly in the Asteraceae, but also in the Poaceae and Ranunculaceae. Four taxa have been named in Dave’s honor: Ancistrocarphus keilii Morefield, Erigeron inornatus (A. Gray) A. Gray var. keilii G.L. Nesom, Wedelia keilii B.L. Turner, and Chrysanthellum keilii B.L. Turner. Few can claim such an impact in California taxonomy. Dave’s extensive fieldwork made him a perfect candidate for involvement with The Jepson Manual Project. Dave authored the Key to California Plant Families and served as the editor and primary author of the Asteraceae in both editions of the manual. Key writing has always been one of Dave’s strengths, and it is a major part of the long lasting legacy he has created throughout his career. For the second edition of The Jepson Manual, Dave authored a new key to families that encompasses the major taxonomic revisions that have taken place since 1993 and served as co-editor for the entire manual. His ability to track nomenclatural changes and translate those changes into meaningful morphological characters in all the major plant families was crucial for the writing of the new family key. Part of what makes Dave’s keys so valuable, is that they are written with field botanists in mind. He is committed to making keys that are useful and attempts to anticipate user misinterpretation on unfamiliar characters. This can only be done if the key writer is familiar with every other possible plant, which Dave usually is. Expect to see many of Dave’s additions to California’s eFlora as he field-tests the current keys in Jepson. Beyond California, Dave has edited and served as a contribut- ing author for three volumes of the Asteraceae for the Flora of North America project. He has published keys to Arizona’s plant families and produced many regional floras. In addition to these highly technical botanical references, Dave has also authored works designed to teach the uninitiated the wonders of California botany. He produced Wildflowers of San Luis Obispo in collaboration with the local California Native Plant Society (CNPS) chapter — a full-color reference for San Luis Obispo’s open space preserves. Many students of California botany may have used one of Dave’s textbooks in their courses. Dave authored California Vegetation with V. L. Holland in 1995 and Vascular Plant Taxonomy with Dirk Walters in 1971, a textbook now in its 5th edition. Both texts have trained generations of California botanists. It is one thing for a botanist to develop expertise on their own, but Dave has been sharing his expertise, knowledge, and excitement about plants with generations 422 MADRONO [Vol. 61 of students at Cal Poly. Over his career Dave taught courses in general botany, plant taxonomy, field botany, and biogeography. Dave won Cal Poly’s Distinguished Teaching Award in 1980. Each spring, for the last 37 years, in a course call Field Botany, Dave traveled around California with a group of students teaching them elements of the California flora. Students have described Field Botany as both the hardest and best course they take during their college careers, and it is often a life changing experience. Many of California’s best land managers, academics, and consultants have gotten their start in botany as a result of Dave’s courses. This year, which was Dave’s 37th and last year teaching Field Botany, we met past course alumni on every field trip. Students have been so inspired by the experience that a near cult following has developed with students returning year after year to tag along, or act as assistants for the course. Songs have been written in honor of Dave and the experiences he creates during the course. For example, the desert camping trip inspired “Little Bushes’’ by Bob Hole, a student in 1989. “Little bushes on the hillside. Little bushes, and they’re all gray-green Little bushes, little bushes And they all look just the same. There’s Eriogonum, and Ephedra And Purshia, and Lepidium, And they’re all just little bushes. And they all look just the same.” Or this song called “The Coastal Salt Marsh” honoring both Matt Ritter and Dave Keil. “Rah, Rah Ritter and salute to Captain Keil, We’re marching off to learn the names of each plant in the field, Juncus, Carex, Scirpus, Typha - all will be revealed. So sing Rah, Rah Ritter and salute to Captain Keil.” Dave served the California Botanical Society as editor of Madrono for three years (1987-1990), was on the editorial board from 1987 to 1994, and is the current editor of Madrono’s Noteworthy Collections section. He is an active CNPS member, and has taught numerous Jepson Workshops on Asteraceae, Rare Plants, the Flora of Kern County, and the Flora of San Luis Obispo County. Personally, Dave has served as an incredible mentor and friend to me. I hope that over my career I can achieve a fraction of what he has done in the classroom, the herbarium, and in the field. I benefit immensely from being able to walk down the hall and be given the answers to difficult questions about Latin, curation, plant morphology, or taxonomy. Watching him work and teach is truly inspiring. Even in his final years of teaching, he is constantly seeking better and novel pedagogical approaches. He still gets excited over seeing new plants, and his passion is infectious. Dave is retiring from Cal Poly this coming year and will focus his efforts toward a new flora for San Luis Obispo County. He will continue working at The Hoover Herbarium and collecting throughout California. It is an honor to work with him. On behalf of all students of California botany, thank you very much, Dave, for your service and inspiration. Jenn Yost Biological Sciences Department California Polytechnic State University San Luis Obispo, CA 93407, jyost(@calpoly.edu Madrono, Vol. 61, No. 4, pp. 423^25, 2014 MADRONO VOLUME 61 TABLE OF CONTENTS Alexander, Earl B., Foliar analyses of conifers on serpentine and gabbro soils in the Klamath Mountains 77 Arnett, Melanie, et al., Vascular flora of Devil’s Postpile National Monument, Madera County, California 367 Bailey, Pamela (see Noyes, Richard D.) Baker, Gail A. (see Rundel, Philip W., et al.) Baker, Marc A. (see Cloud-Hughes, Michelle A., and Marc A. Baker: 2 entries) Baker, Marc A., and Michelle A. Cloud-Hughes, Cylindropuntia chuckwallensis (Cactaceae), a new species from Riverside and Imperial counties, California . . _ 231 Baker, Marc A., and Michelle A. Cloud-Hughes, Noteworthy collection from California 411 Barr, Camille M., Phylogenetic relationships and crossing data reveal a new species of Nemophila (Boraginaceae) . 308 Bobich, Edward G., Nick L. Wallace and Keely L. Sartori, Cholla mortality and extreme drought in the Sonoran Desert 126 Boland, John M., Factors determining the establishment of plant zonation in a southern Californian riparian woodland 48 Boland, John M., Secondary dispersal of willow seeds: Sailing on water into safe sites 388 Brainerd, Richard E. (see Wilson, Barbara L., et al.) Brinda, John C., et al.. Contributions toward a bryoflora of Nevada: Brophytes new for the Silver State. Part III 253 Burge, Dylan O., The role of soil chemistry in the geographic distribution of Ceanothus otayensis (Rhamnaceae) 276 Burk, Donald (see Luper, John, and Donald Burk) Carmen, William J. (see Koenig, Walter D., et al.) Carter, Benjamin E., Ecology and distribution of the introduced moss Campylopus introjlexus (Dicranaceae) in western North America 82 Chau, Kelvin (see McGoey, Brechann V., Kelvin Chau and Timothy A. Dickinson) Christy, John A., et al., Fire effects in a montane wetland, central Cascade Range, Oregon 201 Cipriano, Frank W. (see Walden, Genevieve K., et al.) Cloud-Hughes, Michelle A. (see Baker, Marc A., and Michelle A. Cloud-Hughes: 2 entries) Cloud-Hughes, Michelle A., and Marc A. Baker, Noteworthy collection from California 414 Cloud-Hughes, Michelle A., and Marc A. Baker, Noteworthy collection from New Mexico 149 Creer, Sheryl, and Robert Patterson, Review of The Drunken Botanist: The Plants That Create the World’s Great Drinks by Amy Stewart 144 D’Antonio, Carla M. (see Thomsen, Meredith A., and Carla M. D’Antonio) Danzer, Shelley, and Taly Dawn Drezner, Relationships between epidermal browning, girdling, damage, and bird cavities in a military restricted database of 12,000+ plants of the keystone Carnegiea gigantea in the northern Sonoran Desert 115 Davis, Stephen D. (see Jacobsen, Anna L., et al.) Day, Jonathan (see Willits, Margaret L., and Jonathan Day) Denslow, Michael W. (see Utz, Ryan M., and Michael W. Denslow) Dickinson, Timothy A. (see McGoey, Brechann V., Kelvin Chau and Timothy A. Dickinson) Dowdy, Regina A. (see Simpson, Michael G., et al.) Drezner, Taly Dawn (see Danzer, Shelley, and Taly Dawn Drezner) Elliott, S. A., and O. W. Van Auken, Competition and niche requirements of Coreopsis tinctoria: A widespread but local high density annual Asteraceae 290 Embrey, Teague, Noteworthy collections from Nevada 316 Garrison, Laura M. (see Walden, Genevieve K., et al.) Gorelick, Root, Morphology and development of sunken terminal cephalium in Discocactus (Cactaceae) 194 Greene, Sarah E. (see Christy, John A., et al.) Griswold, Sophie (see Pearse, Ian S., et al.) Haultain, Sylvia (see Arnett, Melanie, et al.) Hellquist, C. Barre (see Hellquist, C. Eric, C. Barre Hellquist and Jennifer J. Whipple) Hellquist, C. Eric, C. Barre Hellquist and Jennifer J. Whipple, New records for rare and under-collected aquatic vascular plants of Yellowstone National Park 159 Huber, Ann M. (see Arnett, Melanie, et al.) Jacobsen, Anna L., et al., Geographic and seasonal variation in chaparral vulnerability to cavitation 317 Jacobsen, Anna L. (see MacKinnon, Evan D., R, Brandon Pratt and Anna L. Jacobsen) Janeway, Lawrence P. (see Zika, Peter F., Lawrence P. Janeway and Barbara L. Wilson) Kannely, Alfred, and Robert A. Schlising, Biology of the geophyte, Triteleia ixioides subsp. anilina (Themidaceae), in coniferous forests of Butte County, California 87 Keil, David J., Review of Flora of the Four Corners Region. Vascular Plants of the San Juan River Drainage: Arizona, Colorado, New Mexico, and Utah by Kenneth D. Heil, et al. 248 424 MADRONO [Vol. 61 Keil, David J., Review of WildjJowers of Orange County and the Santa Ana Mountains by Robert L. Allen and Fred M. Roberts, Jr. 245 Kelly, Ronald B. (see Simpson, Michael G., et al.) Kentner, Ed, and Kim Steiner, A new species of Triteleia (Themidaceae) from the southern Sierra Nevada 227 Knaus, Brian J. (see Wilson, Barbara L., et al.) Knops, Johannes M. H. (see Koenig, Walter D., et al.) Koenig, Walter D. (see Pearse, Ian S., et al.) Koenig, Walter D., et ah, Serotiny in California oaks 151 Liow, P-S. (see Spencer, D. F., et al.,) Lippert, Jennifer D. (see Christy, John A., et al.) Luper, John, and Donald Burk, Noteworthy collection from California 413 MacKinnon, Evan D., R. Brandon Pratt and Anna L. Jacobsen, Functional trait differences between weedy and non-weedy plants in southern California 328 McCain, Cynthia N. (see Christy, John A., et al.) McGoey, Brechann V., Kelvin Chau and Timothy A. Dickinson, Stomata size in relation to ploidy level in North American hawthorns {Crataegus, Rosaceae) 177 Nelson, Julie Kierstead (see Wilson, Barbara L., et al.) Nilsen, Erik T. (see Rundel, Philip W., et al.) Norris, Daniel H. (see Shevock, James R., and Daniel H. Norris) Noyes, Richard D., and Pamela Bailey, Chromosome number and reproductive attributes for Erigeron lemmonii (Asteraceae), a cliff-dwelling endemic of southeastern Arizona 9 O’Dell, Ryan, Noteworthy collection from California 250 Otting, Nick (see Wilson, Barbara L., et al.) Parker, V. Thomas, A newly described serpentine-endemic Ceanothus (Rhamnaceae) from coastal Marin County, California 399 Parker, V. Thomas, President’s Report for Volume 61 416 Patterson, Robert (see Creer, Sheryl, and Robert Patterson) Patterson, Robert (see Walden, Genevieve K., et al.) Pearse, Ian S. (see Koenig, Walter D., et al.) Pearse, Ian S., et al.. Stage and size structure of three species of oaks in central coastal California 1 Pitcairn, M. J. (see Spencer, D. F., et al.,) Pizarro, Desirree (see Pearse, Ian S., et al.) Pratt, R. Brandon (see Jacobsen, Anna L., et al.) Pratt, R. Brandon (see MacKinnon, Evan D., R. Brandon Pratt and Anna L. Jacobsen) Preston, Robert E., Vernal pool blue dicks {Dichelostemma lacuna-vernalis', Asparagaceae: Brodiaeoideae) revisited . 350 Rebman, Jon P. (see Simpson, Michael G., et al.) Ritter, Matt, Editor’s Report for Volume 61 417 Roberts, Fred, Review of Baja California Plant Field Guide, 3''‘‘ Edition by Jon P. Rebman and Norman C. Roberts 246 Rundel, Philip W., et al., Geomorphic landforms and plant community structure and dominance in the central desert region of Baja California, Mexico 105 Sartori, Keely L. (see Bobich, Edward G., Nick L. Wallace and Keely L. Sartori) Schlising, Robert A. (see Kannely, Alfred, and Robert A. Schlising) Sharifi, M. Rasoul (see Rundel, Philip W., et al.) Shevock, James R., and Daniel H. Norris, Scouleria siskiyouensis (Scouleriaceae), a new rheophytic moss endemic to southern Oregon, USA 137 Shevock, James R. (see also Brinda, John C., et al.) Shultz, Leila M. (see Rundel, Philip W., et al.) Simpson, Lee M. (see Simpson, Michael G., et al.) Simpson, Michael G., Review of Intermountain Flora: Vascular Plants of the Intermountain West, U.S.A. Subclass Magnoliidae-Caryophyllaceae, Vol. 2A by N. H. Holmgren et al. 146 Simpson, Michael G., et al., Recognition of two species in Eremocarya (Boraginaceae): Evidence from fornix bodies, nutlets, corolla size, and biogeography 259 Spence, John R. (see Brinda, John C., et al.) Spencer, D. F., et al.. Variation in seed characteristics and growth for thistles (Cardueae: Asteraceae) in California and Oregon 339 Spicer, Greg S. (see Walden, Genevieve K., et al.) Stark, Lloyd R. (see Brinda, John C., et al.) Starrs, Paul F., Review of Trees in Paradise: A California History by Jared Farmer 407 Steiner, Kim (see Kentner, Ed, and Kim Steiner) Stevenson, Kathren Murrell (see Arnett, Melanie, et al.) Thomsen, Meredith A., and Carla M. D’Antonio, How does simulated gopher disturbance affect the establishment of Holcus lanatus (Poaceae) in California coastal prairie? 218 Tobin, Michael F. (see Jacobsen, Anna L., et al.) Utz, Ryan M., and Michael W. Denslow, Noteworthy collection from Colorado .. 251 Van Auken, O. W. (see Elliott, S. A., and O. W. Van Auken) Villegas, B. (see Spencer, D. F., et al.,) 2014] TABLE OF CONTENTS 425 Virginia, Ross A. (see Rundel, Philip W., et al.) Walden, Genevieve K., et al., Phylogenies and chromosome evolution of Phacelia (Boraginaceae: Hydrophylloideae) inferred from nuclear ribosomal and chloroplast sequence data . 16 Wallace, Nick L. (see Bobich, Edward G., Nick L. Wallace and Keely L. Sartori) Walters, Eric L. (see Koenig, Walter D., et al.) Whipple, Jennifer J. (see Hellquist, C. Eric, C. Barre Hellquist and Jennifer J. Whipple) Willits, Margaret L., and Jonathan Day, Noteworthy collection from California 148 Wilson, Barbara L., et al.. Identification and taxonomic status of Cordylanthus tenuis subsp. pallescens (Orobanchaceae) 64 Wilson, Barbara L. (see also Zika, Peter F., Lawrence P. Janeway and Barbara L. Wilson) Yost, Jenn, Dedication of Volume 61 to David J. Keil 421 Yost, Jenn, Review of Annotated Checklist of the Vascular Plants of Santa Cruz County, California. Second Edition by Dylan Neubauer 244 Zika, Peter F., Lawrence P. Janeway and Barbara L. Wilson, Carex xerophila (Cyperaceae), a new sedge from the chaparral of northern California 299 Dates of Publication of Madrono, Volume 61 Number 1, pages 1-150, published 27 March 2014 Number 2, pages 151-251, published 9 May 2014 Number 3, pages 253-316, published 24 July 2014 Number 4, pages 317-425, published 9 December 2014 « i .- :*I- -.^It.*!'* ';' > j.iiV r'A Vj ' ' •' T' ■■XU'Wid-i' •■/ ,«jir<»c« i.-.liHVl* ^ ; 'JsKik:' • u vt;nH • .vir.,< ' -..'.{f,, ] ‘ ' ‘ ‘ n , ». V. 1!-.’ Ul.i » * ' 11^ ■■•'_•' ' ■■ s* ' •! i U iv ; ...> •'. T " V si I'lt *;>„>> ■• . - ■ > « " ■ ... ,. . , i,.. m :H» ' (;" 'iC .'i - • ■■ ^ ..ii 4^^051^ ,_ . ., -. , , . ^ •' ^ •■". ••n:» WunV J .■' •' ' i 1 ..■■■v/':-#^.">5;'V^:M.i-v*/I J .» ■■ v '■■•^' >«■' : S i ■(•■I'V^'Jj '' 1 :iS,i( ')V( tir.. •. •'■ ■■ ' *■' • ■ s - - * .f0aL,%ar.t „, . ,a, ■» * ' , - f^<%w ' ■ ' wot,n. r\. H->4n 4U ^ i v^..y 4- t ¥ .v^' .. « .‘.4i '1 AyA •;: . ’•■ • ' ' oiiJii»cKrv * .'■•■• fkji *.»■ Jt '' M' >•; r S' IT.. ,» nK|n V HT'' ' ;>■. , I ' . ' ,. * •*'•■ » J»r- , i /I. I ■ • W- *■ f • !r^. .. -t*' H' • (KJ-I-J- • 4,'7\ ** rt|i>" ''4‘’1 Vv: i.i'j.^^V I ; f*.- V . »'■■ -'V.* A ‘'V 4,, ■• ?«!*»'* I ♦...t* *■ 0 ^ » ^-S). ,■-. . ■> ifi|||||ii:’ / ■ ,\ '^ ■ U •’li .1 »>.'■■■■* . *» . ■ v*.4‘l *' • . • ^- V- It- ’ «'■ s; ' '• ;:■ ■■■ ' , - . y <--■ .. ] V* Vf vs. .'.t;- .V‘ j-v- ' VW^ IP>“' ff • » ' * Iipir '■ — ‘■‘‘ * * , MaH'.' yW •* sm. 'K ■;- - » ^ n 1^ ■■ i^jr* ’«. .. .'tHiisw -'**'■•. >4H| ■ r' *■'>'• ',' ' »>» N# r 'y ^iif,|jr . * " , '■»» - =’ I ■ .-H',. ..< '^l■('=♦*^ ■* • *• '■ ;'' -" v- .'Yt'ip ,'*1 •■*•-•• ■» .V... pf ; ■■■'•- irt-^!**..' "■'> ■' •* Z,"*. »' t ' - • ' » 'r'. f -(-’f -’ ■ ■•'*■ * sr?' ,v ' 4Eg‘«-.'^.4 ^.iC^ yw |*A Ul ■ -r’^„.r :. ^ -■ '.-w, — . ■* ■"■ T' ■H .,\ -■ •; M ‘'"k IV*' ^ ‘ ■* ' •"' * ^ - . i»>,:. Cw .*#*►» .'^ n..;v v ■ <■ i»« -''fin?; “ Subscriptions — Membership The California Botanical Society has several membership types (individuals ($40 per year; family $45 per year; emeritus $32 per year; students $27 per year for a maximum of 7 years). Late fees may be assessed. Beginning in 2011, rates will increase by $5 for all membership types except life memberships, for which rates will increase by $100, and student memberships, which will not show a rate increase. Members of the Society receive Madrono free. Institutional subscriptions to Madrono are available ($75). Membership is based on a calendar year only. Life memberships are $850. Applications for membership (including dues), orders for subscriptions, and renewal payments should be sent to the Membership Chair. Requests and rates for back issues, changes of address, and undelivered copies of Madrono should be sent to the Corresponding Secretary. Information for Contributors Submit manuscripts to http://www.edmgr.com/madrono/. If you are submitting a manuscript for the first time to the Madrono PeerTrack website, you must first register by clicking "register now" and following the instructions. Manuscripts by authors having outstanding page charges will not be sent for review. Manuscripts may be submitted in English or Spanish. English-language manuscripts dealing with taxa or topics of Latin America and Spanish-language manuscripts must have a Spanish Resumen and an English Abstract. For all articles and short items (NOTES, NOTEWORTHY COLLECTIONS, POINTS OF VIEW, etc.), follow the format used in recent issues for the type of item submitted. Detailed information for contributors, including instructions for manuscript preparation, is available at the California Botanical Society website, www.calbotsoc. org; follow the link for Madrono. Authors are encouraged to include the names, addresses, and e-mail addresses of two to four potential reviewers with their submitted manuscript. All California Botanical Society members current in the volume year that their contributions are published are allowed five free pages per volume year. Additional pages will be charged at the rate of $40 per page. Joint authors may apply all or a portion of their respective five-page allotments to a jointly-published article. Partial pages will be charged as full. 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 color plates and other complex matter (including illustrations, charts, maps, photographs) will be charged at cost. Author’s changes after typesetting @ $4.50 per line will be charged to authors. Page charges are important in maintaining Madrono as a viable publication, and timely payment of charges is appreciated. 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. SMITHSONIAN LIBRARIES 3 9088 01798 8973 BOOK REVIEWS Trees in Paradise: A California Histg Paul F Starrs NOTEWORTHY COLLECTIONS ANNOUNCEMENTS CBS President's Report for Volume 6i .. Edeto^s^eport for Volume"^'! Reviewers of MadronoWIanuscripts 2014. 913 Yt TOl I 10/03/ IGy:^ "615750 S f