VOLUME 58, NUMBER 2 APRIL-JUNE 201 1 1 > H 1^3 Bc>T / J "v ' I jm s ' •v > I t Invasive European Annual Plants Impact a Rare Endemic Sunflower Jolene R. Moroney, Paula M. Schiffman, and Christy A. Brigham 69 Pollen Siring Success in the California Wildflower Clarkia UNGUICULATA (ONAGRACEAE) Nancy L. Smith-Huerta and Frank C. Vasek 78 Smaller Olea europaea Fruits have More Potential Dispersers: Implications for Olive Invasiveness in California Clare E. Aslan and Marcel Rejmanek 86 Systematics, Phylogeny, and Evolution of Papaver californicum and Stylomecon heterophylla (Papaveraceae) Joachim W. Kadereit and Bruce G. Baldwin 92 Morphological Comparisons of White Fir and Red Fir Dwarf Mistletoes in the Sierra Nevada and Southern Cascade Mountains Robert L. Mathiasen 101 Is Cylindropuntia xfosbergii (Cactaceae) a Hybrid? Michael S. Mayer, Anastasia Gromova, Kristen Hasenstab-Lehman, Molly Lippitt, Mia Barnett, and Jon P. Rebman 106 Gabbro Soils and Plant Distributions on Them Earl B. Alexander 1 1 3 Further Floristics on Late Tertiary Lacustrine Deposits in the Southern Arizona Deserts John L. Anderson 123 Erratum 129 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. Corresponding Editor — Timothy Lowrey Museum of Southwestern Biology MSC03 2020 University of New Mexico Albuquerque, NM 87131-0001 madrono@unm.edu Copy Editor — Richard Whitkus Department of Biology Sonoma State University 1801 E. Cotati Avenue Rohnert Park, CA 94928-3609 whitkus @ sonoma.edu Book Editor — Jon E. Keeley Noteworthy Collections Editors — Dieter Wilken, Margriet Wetherwax Board of Editors Class of: 2011 — Jamie Kneitel, California State University, Sacramento, CA Kevin Rice, University of California, Davis, CA 2012 — Gretchen LeBuhn, San Francisco State University, CA Robert Patterson, San Francisco State University, CA 2013 — Eric Roalson, Washington State University, WA Kristina Schierenbeck, California State University, Chico, CA 2014 — Brandon Pratt, California State University, Bakersfield, CA Tom Wendt, University of Texas, Austin, TX CALIFORNIA BOTANICAL SOCIETY, INC. Officers for 20 1 1 -20 1 2 President : V. Thomas Parker, Department of Biology, San Francisco State University, San Francisco, CA 94132, parker@sfsu.edu First Vice President : Andrew Doran, University and Jepson Herbaria, University of California, Berkeley, CA 94720, andre wdoran @ berkeley.edu Second Vice President : Marc Los Huertos, Division of Science & Environmental Policy, California State University, Monterey Bay, Seaside, CA 93955, mloshuertos@csumb.edu Recording Secretary. Mike Vasey, Department of Biology, San Francisco State University, San Francisco, CA 94132, mvasey@sfsu.edu Corresponding Secretary. Heather Driscoll, University Herbarium, University of California, Berkeley, CA 94720, hdriscoll @ berkeley.edu 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, C A 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: Chelsea Specht, Department of Plant and Microbial Biology, University of California, Berkeley, CA 94720-2465, cdspecht@berkeley.edu; Ellen Simms, Department of Intergrative Biology, 3060 Valley Life Sciences Bldg., #3140, University of California, Berkeley, CA 94720, esimms@berkeley.edu. Staci Markos, University and Jepson Herbaria, University of California, Berkeley, CA 94720, smarkos@berkeley.edu. Graduate Student Representatives: Ben Carter, Department of Integrative Biology and University Herbarium, University of California, Berkeley, CA 94720, bcarter@berkeley.edu. Webmaster: Susan Bainbridge, Jepson Herbarium, University of California, Berkeley, CA 94720-2465, sjbainbridge@berkeley.edu. ® This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Madrono, Vol. 58, No. 2, pp. 69-77, 2011 INVASIVE EUROPEAN ANNUAL PLANTS IMPACT A RARE ENDEMIC SUNFLOWER JOLENE R. MORONEY1 AND PAULA M. SCHIFFMAN Department of Biology, California State University, 18111 Nordhoff Street, Northridge, CA 91330-8303 jmoroney@ucla.edu Christy A. Brigham Santa Monica Mountains National Recreation Area, 401 West Hillcrest Drive, Thousand Oaks, CA 91360 Abstract Pentachaeta lyonii A. Gray is a state- and federally-listed endangered species, endemic to heavily invaded southern California grasslands. Recent population extirpations resulting in a decrease in range size have prompted investigation into the effects of invasive annual plants on this species. The goals of this study were (1) to examine the impacts of competition from non-native species from three different functional groups (annual grasses, early-season forbs, and late-season forbs) on P. lyonii success in the field and in pots, (2) to determine which non-native species/functional groups have the greatest competitive effect on P. lyonii , and (3) to evaluate the environmental conditions that contribute to the displacement of P. lyonii by non-native plants. In the field, at two sites over two years, control plots were paired with plots in which non-native competitors were clipped at the soil surface. In pots, individual P. lyonii plants were grown in competition with all three groups of non- native competitors at both high and low density. In both the field and in the pots, all three non-native plant groups had negative effects on P. lyonii reproductive potential, with Centaurea melitensis L. having the greatest effect. The effects on P. lyonii height were variable among non-native competitors and between years. Comparisons made of environmental features of sites where P. lyonii has been extirpated to those where it persists suggested that the presence of annual grass is associated with P. lyonii extirpation. Management of P. lyonii presents a challenge considering the tendency of this species to coexist with non-native annual plants due to their common disturbance-dependence, and the ubiquity of European annuals in P. lyonii habitat. Key Words: California grassland, competition, conservation, endangered species, non-native invasive plants, Pentachaeta lyonii , rare plants, restoration. Pentachaeta lyonii A. Gray (Asteraceae) is a state and federally listed endangered species endemic to southern California grasslands (Fotheringham and Keeley 1998). Following its extirpation at sites in the southern part of its range, P. lyonii became restricted to 21 popula- tions in the Santa Monica Mountains and Simi Hills, persisting entirely within the increasingly suburban northern Los Angeles and southern Ventura counties. Historically, P. lyonii had a wider distribution in the Los Angeles basin, Santa Catalina Island, and San Diego (Hickman 1993), but as many as 15 populations have been extirpated within recent decades, and many of the remaining populations appear to be in decline (Brigham 2007). The U.S. Fish and Wildlife Service (1999) recovery plan for P. lyonii identifies competition from invasive non-native plants as a possible cause of the species’ decline. 1 Present address: Department of Ecology and Evo- lutionary Biology, University of California, Los An- geles, 621 Charles E. Young Drive South, Los Angeles, CA 90095-1606. In the case of a rare endemic such as P. lyonii, which has already lost at least 45% of its populations in recent decades, and with the remaining populations geographically isolated by fragmentation, competitive pressures could contribute to local declines and possibly to its ultimate extinction. Annual surveys done by the National Park Service of both P. lyonii numbers and the presence of invasive species have indicated a possible relationship between invasion and de- clines, but no competition studies have been done previous to the present work. Invasive species do play an important role in native species diversity declines (Hobbs and Mooney 1998; Simberloff 2005). Impacts include the alteration of ecosys- tem functioning (Evans et al. 2001), altered disturbance regimes (Brooks et al. 2004) and competition for resources (Dyer and Rice 1997; Eliason and Allen 1997; Brooks 2000). Rare native plants can be particularly vulnerable to competition from invasive plants (Huenneke and Thomson 1995; Walck et al. 1999; Miller and Duncan 2003; Kingston et al. 2004; Thomson 2005; Corbin et al. 2007), especially those species 72 MADRONO [Vol. 58 LO CD -Q E Z3 C CO o E 0) CO _ CE I Grass Erodium Centaurea COMPETITOR Fig. 1 . Differences in number of inflorescences on P. lyonii plants between paired field plots. Pairs consisted of non-native competitors removed, and competitors present. Results are shown for two different seasons. Box plots are centered at the median, and whiskers indicate the octiles. Paired t-tests showed significant differences in all cases (P < 0.01). Centaurea and grasses, (P < 0.05), but not significant between Centaurea and Erodium (P = 0.5688). Grasses and Erodium did not differ significantly in their competitive effects on P. lyonii inflorescence number (P = 0.1426). Effects from competition on height of P. lyonii differed among the non-native functional groups and between years (Fig. 2). In four of the six cases, differences in height were not significant for competition vs. no-competition plants. Sig- nificant differences were found between treat- ments in Erodium plots in 2004, where those released from competition grew taller than those in control plots (P = 0.01, Fig. 2). In contrast, in 2005, plants competing with annual grasses grew taller than those in plots with competitors removed (P < 0.01, Fig. 2). In 2004, P. lyonii plants growing without competition from Cen- taurea were marginally taller than those in control plots (P = 0.064, Fig. 2). Annual grasses had a greater effect on P. lyonii height in the field than Centaurea or Erodium (P < 0.05). Pot Competition Experiment Under the more controlled research conditions of the pot experiment, competition from all non- native groups had a negative effect on P. lyonii reproductive potential. Pentachaeta lyonii plants produced significantly fewer inflorescences when Grass Erodium Centaurea COMPETITOR Fig. 2. Differences in height (mm) of P. lyonii plants between paired field plots. Pairs consisted of non-native competitors removed, and competitors present. Results are shown for two different seasons. Box plots are centered at the median, and whiskers indicate the octiles. Asterisks indicate cases in which paired t-tests showed significant differences between treatments. 2011] MORONEY ET AL.: EUROPEAN ANNUALS IMPACT A RARE ENDEMIC 73 a I Control Grass Erodium Centaurea COMPETITOR Fig. 3. Numbers of inflorescences on P. lyonii plants grown in pots with different non-native competitors growing at both high and low density, compared with P. lyonii plants grown in pots without competitors (control). Box plots are centered at the median, and whiskers indicate the octiles. Different letters above the plots indicate significant differences (Tukey HSD). grown in pots with non-native competitors at both low and high densities compared with control plants grown without competition (P < O. 001, Fig. 3). Pentachaeta lyonii plants grown with Erodium or Centaurea at low density had more inflorescences than those growing with Erodium or Centaurea at high density, or those growing with annual grasses (Fig. 3). Similarly, P. lyonii plants growing without competition (control) were significantly taller than plants growing in competition with all three non-native species groups, both at high and low densities (P < 0.001, Fig. 4). With all three invasive com- petitors, P. lyonii plants competing in low-density pots were taller than those in high-density pots; however, these differences were only significant in the Erodium group (Fig. 4). Comparison of Extant Versus Extirpated Sites The logistic regressions suggested that high cover of litter, annual grasses, and non-native plants other than those in our target groups, as well as low cover of Erodium and bare ground were good predictors of P. lyonii extirpation (P < 0.05 in all cases, Fig. 5). Nonmetric multidimen- sional scaling yielded a final stress2 of 0.33747. All of the environmental variables that were correlated with whether the population at a site was extant or extirpated had r2 > 0.2 except for volumetric water content and percent cover of Centaurea. There was a clear separation of extant from extirpated sites (along the horizontal axis). The factors that were most positively correlated with sites supporting extant populations of P. lyonii were percent cover of Erodium , PAR and percent bare ground. Amount of litter and annual grasses were the most negatively correlated with sites with extant populations (Fig. 6). Discussion This study addressed the direct effects of competition from three different functional groups of non-native plants on P. lyonii , and the factors possibly contributing to its extirpa- tion. Competition from all three groups of non- native plants reduced the reproductive potential T ^ o - E E s Control I I High density I I Lowdensily COMPETITOR Fig. 4. Heights of P. lyonii plants grown in pots with different non-native competitors grown at both high and low density, compared with P. lyonii plants growing in pots without competitors (control). Box plots are centered at the median, and whiskers indicate the octiles. Different letters above the plots indicate significant differences (Tukey HSD). 74 MADRONO [Vol. 58 Fig. 5. Comparison of the composition of sites where P. lyonii is extant and sites where it is extirpated. of P. lyonii , and the presence of annual grasses not only resulted in direct competitive interac- tions, but also was implicated as an indirect factor contributing to P. lyonii extirpation. Direct Effects of Competition The competitor removal experiment took place over two field seasons with very different environmental conditions (2003-2004 and 2004— 2005), and examined two possible indicators of competitive effects on P. lyonii : number of inflorescences produced and plant height. Effects of competition on the number of inflorescences remained consistent over both field seasons for all three invasive plant groups, despite the large difference between years in available soil mois- ture. 2004—2005 was an exceptionally wet year (56 cm above average), with four times more rainfall than in 2003-2004. Even with an excess of a potentially limiting resource, competitive inter- ference from non-native plants significantly reduced the reproductive output of P. lyonii. This indicated that these invasive plants have a superior ability to capture other important limiting resources (perhaps nutrients, space, and/or light). These results were corroborated in the pot competition study, where growing condi- tions were more controlled, and water was generously provided. The reduction in number of inflorescences across the board indicated that at least three non-native plant groups that Axis 1 Fig. 6. Nonmetric multidimensional scaling of environmental variables. Ordination rotated so Axis 1 correlates maximally with extant (unfilled triangles) versus extirpated (filled triangles) sites. The lengths and directions of the vectors indicate the strengths and directions of the correlations with the axes. 2011] MORONEY ET AL.: EUROPEAN ANNUALS IMPACT A RARE ENDEMIC 75 commonly occur in P. lyonii habitat negatively impact fecundity, potentially reducing time to extinction. The impacts of competition on plant height were less clear. In both seasons P. lyonii plants growing in the field in competition both with Centaurea and Erodium were shorter than those in plots where competitors were removed. The opposite was true for P. lyonii plants growing in competition with grasses. This may reflect differences in root morphology between grasses and forbs. The forbs, Pentachaeta , Centaurea and Erodium , possess taproots, whereas the grasses have fibrous root systems. These rooting differ- ences may cause interspecific variation in the ability of plants to access below-ground resources and thereby affect aboveground growth and competitive dynamics (Gurevitch et al. 2006). They may also account for the difference in effects between wet and dry years. Height differences were greater in the forb groups in 2003-2004, when soil moisture was more limited, than in 2004—2005 when it was more abundant. Functional group phenology may contribute to differences in height response between P. lyonii competing with forbs and P. lyonii competing with grasses. Both Centaurea and non-native grasses typically grow taller than P. lyonii (Hick- man 1993), and can potentially reduce available light. However, the grasses gain height earlier in the season than Centaurea , which spends its first few months as a basal rosette and then matures later in the season. Erodium also grows rapidly early in the season, but it is generally of shorter stature than P. lyonii (Hickman 1993), and likely does not significantly reduce the light available to P. lyonii plants. The large height difference in grass plots between taller P. lyonii plants in control plots and shorter ones in plots from which competitors had been removed in 2004— 2005 may have been due to the abundance of rainfall that year. Greater moisture availability likely resulted in exceptionally fast growth rates of grasses, causing P. lyonii plants to elongate to compensate for early reduction in available light. However, this response does not necessarily indicate superior performance. Shorter P. lyonii plants that were growing without competition from grasses produced more inflorescences than tall plants growing with competition, indicating a trade-off in resource allocation. None of the invasive species had a greater effect on P. lyonii reproduction or height than any of the others in the pot competition experiment, but in the field experiment competi- tion from Centaurea suppressed inflorescence production more than Erodium and significantly more than annual grasses did. Centaurea meli- tensis and P. lyonii are both late-season annuals with basal leaf rosettes and taproots, sharing functional traits in both phenology and morphol- ogy. The bulk of their reproductive efforts occur as, or even after, annual grasses and Erodium are completing theirs. Two co-occurring plant species with similar phenologies and morphologies would be expected to compete for, rather than partition resources (Dukes 2002). Another inva- sive species of Centaurea (C. solstitialis L.) has been shown to increase late-season evapotrans- piration in the communities it has invaded (Dukes 2001). Centaurea melitensis may, similar- ly, deplete water more efficiently than P. lyonii , resulting in reduced late-season resource avail- ability, when water becomes more limiting, and ultimately reduced reproductive capacity. Differences in the phenologies of annual grasses, Erodium , and Centaurea could subject a P. lyonii individual to competition early in development (annual grasses and Erodium) as well as later during flowering ( Centaurea ). Depending on the species composition of the immediate neighborhood of a P. lyonii plant, competitive pressure could affect an individual plant throughout its life cycle. Environmental Factors Associated with Extirpation In the absence of outright destruction of habitat, it is difficult to be certain of the causes of local extinction of this species, but compari- sons of sites with extant populations and sites with extirpated populations can identify environ- mental factors correlated with extirpation. The three variables that were the best predictors of extirpation were related. Percent cover of annual grasses, percent cover of litter, and low percent cover of bare ground are all associated with P. lyonii extirpation, and with annual grass invasion and dominance. The persistence of litter in some annual grass species can reduce open patches of bare ground through the winter rainy season, when germination of P. lyonii takes place. The light and moisture conditions at ground level can be drastically altered under a layer of grass litter, possibly precluding germination of P. lyonii. The association of these three variables suggest that grasses are not only important direct competitors with P. lyonii , as indicated in the competitor removal experiment, but also cause indirect competitive pressures strong enough to displace the species locally. Aside from the implication of annual grass presence as a factor in local extirpation of P. lyonii , it is interesting to note the unexpected lack of correlation of C. melitensis, and the negative correlation of Erodium spp. presence with P. lyonii extirpation. The percent cover of C. melitensis was not correlated with sites with either extant or extirpated populations of P. lyonii. Considering the magnitude of this species’ negative effect on P. lyonii reproduction, its co- 76 MADRONO [Vol. 58 occurrence with P. lyonii may not be as common as the co-occurrence of P. lyonii with annual grasses. It may be a much newer phenomenon, as the spread of C. melitensis in southern California wildlands has been increasing in recent years (Cal-IPC 2008). If this is the case, competition from C. melitensis has the potential to pose an even greater threat to P. lyonii populations in the future. The unexpected negative correlation of Erodium with P. lyonii extirpation suggested that it, like Pentachaeta , favors sites without extensive annual grass presence. Management Implications Conservation biology is a crisis discipline (Primack 2010). Because of the unprecedented rate of species extinctions, often managers are forced to take actions to attempt to preserve endangered species without thorough prior in- vestigations into their ecological relationships. In the case of P. lyonii , little is known about its ecology. The disturbed grassland areas where P. lyonii occurs exhibit characteristics that promote invasion by non-native species, and non-native plants have dominated many of the sites with extant P. lyonii populations over a long period of time. In this study, non-native plants were removed individually by hand. These methods would be extremely labor-intensive and expensive for large-scale, long-term management of P. lyonii. However, they are the only methods to date shown to significantly improve reproductive potential for P. lyonii. Until alternative methods have been explored, hand-weeding should be implemented at least in the most threatened sites. Alternative methods should be investigated to facilitate the feasibility of large scale, long-term restoration efforts. The use of monocot-specific herbicides early in the season to eliminate competition from annual grasses may be an option, but effects on the native community as a whole should be studied before implementation. The use of prescribed burning may also be an alternative. However, P. lyonii s ability to tolerate fire is poorly understood, and experiments to evaluate the effects of fire frequency, intensity, and seasonality on both P. lyonii and its associated community should be carried out prior to consideration as a restoration tool. Moreover, large-scale removal of invasive species in habitats where they are established members of the community can have unexpected and undesirable consequences for ecosystems (Zavaleta et al. 2001; Ogden and Rejmanek 2005). Conservation and restoration research in California grassland ecosystems has focused primarily on the native perennial grass, Nassella pulchra (Hitchc.) Barkworth (Stromberg et al. 2007). However, annual forbs are increasingly recognized as a major native component of these systems (Keeley 1990; Schiffman 2000, 2007), contributing greatly to their biodiversity (Kim- ball and Schiffman 2003). In some heavily invaded grasslands, native forbs have been excluded from fertile sites, and persist only in marginal, relatively low-resource refugia, where non-native plants cannot invade (Seabloom et al. 2003). The resulting fragmentation of native forb populations and reduction of population size can potentially contribute to local extinction (Lande 1993). In order to preserve biodiversity in California grassland ecosystems, conservation research efforts must include annual forbs. Although much work remains ahead for the conservation of P. lyonii , this study provides a foundation for the design of a sound conserva- tion strategy for this endangered species, and serves as a starting point for further research. More broadly, the information gained here may be relevant to the conservation of other rare annual plants in other Mediterranean-type grass- land ecosystems, which are under increasing pressure from the same non-native invaders. Acknowledgments We thank Paul Wilson, California State University, Northridge, for his generous assistance and input on this project. Funding for this research was contributed by the National Park Service, Santa Monica Mountains National Recreation Area, and Western National Parks Association. These experiments were carried out in compliance with the laws of the United States of America. Literature Cited Brigham, C. A. 2007. Managing rare plants at the wildland-urban interface: an example from the Santa Monica Mountains and Simi Hills. Pp. 109- 130 in D. A. Knapp (ed.), Flora and Ecology of the Santa Monica Mountains: proceedings of the 32nd Annual Southern California Botanists Symposium. Southern California Botanists Special Publication No. 4, Fullerton, CA. Brooks, M. L. 2000. Competition between alien annual grasses and native annual plants in the Mojave Desert. American Midland Naturalist 144:92-108. — , C. M. D’Antonio, D. M. Richardson, J. B. Grace, J. E. Keeley, J. M. DiTomaso, R. J. Hobbs, M. Pellant, and D. Pyke. 2004. Effects of invasive alien plants on fire regimes. BioScience 54:677-688. California Invasive Plant Council (Cal-IPC). 2008. Statewide risk mapping. Centaurea meli- tensis. California Invasive Plant Coucil, Berkeley, CA. Website: http://www.cal-ipc.org/ip/mapping/ statewide_maps/Centaurea_melitensis.php [ac- cessed 3 November 2008]. Coleman, H. M. and J. M. Levine. 2007. Mecha- nisms underlying the impacts of exotic annual grasses in a coastal California meadow. Biological Invasions 9:65-71. Corbin, J. D. and C. M. D’Antonio. 2004. Compe- tition between native perennial and exotic annual 2011] MORONEY ET AL.: EUROPEAN ANNUALS IMPACT A RARE ENDEMIC 77 grasses: implications for an historical invasion. Ecology 85:1273-1283. , A. R. Dyer, and E. W. Seabloom. 2007. Competitive interactions. Pp. 156-168 in M. R. Stromberg, J. D. Corbin, and C. M. D’Antonio (eds.), California grasslands: ecology and manage- ment. University of California Press, Berkeley, CA. Dukes, J. S. 2001. Biodiversity and invasibility in grassland microcosms. Oecologia 126:563-568. — . 2002. Species composition and diversity affect grassland susceptibility and response to invasion. Ecological Applications 12:602-617. Dyer, A. R. and K. J. Rice. 1997. Intraspecific and diffuse competition: the response of Nassella pulchra in a California grassland. Ecological Applications 7:484=492. and . 1999. Effects of competition on resource availability and growth of a California bunchgrass. Ecology 80:2697-2710. Eliason, S. A. and E. B. Allen. 1997. Exotic grass competition in suppressing native shrubland re- establishment. Restoration Ecology 5:245-255. Evans, R. D., R. Rimer, L. Sperry, and J. Belnap. 2001. Exotic plant invasion alters nitrogen dynam- ics in an arid grassland. Ecological Applications 11:1301-1310. Fotheringham, C. J. and J. E. Keeley. 1998. Ecology and distribution of Braunton’s milkvetch {Astragalus brauntonii) and Lyon’s pentachaeta {Pentachaeta lyonii). California Department of Fish and Game, Region 5. Website: http://nrm.dfg.ca. gov/FileHandler.ashx?DocumentID = 3154 [ac- cessed 30 June 2011], Gurevitch, J., S. M. Scheiner, and G. Fox. 2006. The ecology of plants. Sinauer Associates, Inc., Sunderland, MA. Heady, H. F. 1988. Valley grassland. Pp. 491-514 in M. G. Barbour and J. Major (eds.). Terrestrial vegetation of California. California Native Plant Society, Sacramento, CA. Hickman, J. C. (ed.) 1993, The Jepson manual: higher plants of California. University of California Press, Berkeley, CA. Hobbs, R. J. and H. A. Mooney. 1998. Broadening the extinction debate: population deletions and additions in California and western Australia. Conservation Biology 12:271-283. Hooper, D. U. and J. S. Dukes. 2010. Functional composition controls invasion success in a Califor- nia serpentine grassland. Journal of Ecology 98:764-777. Huenneke, L. F. and J. K. Thomson. 1995. Potential interference between a threatened endemic thistle and an invasive nonnative plant. Conservation Biology 9:416-425. Keeley, J. E. 1990. The California valley grassland. Pp. 2-23 in A. Schoenherr (ed.), Endangered plant communities of southern California. Southern California Botanists, Claremont, CA. Kimball, S. and P. M. Schiffman. 2003. Differing effects of cattle grazing on native and alien plants. Conservation Biology 17:1681-1693. Kingston, N., S. Waldren, and N. Smyth. 2004. Conservation genetics and ecology of Angiopteris chauliodonta Copel. (Marattiaceae), a critically endangered fern from Pitcairn Island, south central Pacific Ocean. Biological Conservation 117:309- 319. Lande, R. 1993. Risks of population extinction from demographic and environmental stochasticity and random catastrophes. American Naturalist 142:911-927. Miller, A. L. and R. P. Duncan. 2003. Extrinsic and intrinsic controls on the distribution of the critically endangered cress, Ischnocarpus exilis (Brassicaceae). Biological Conservation 110: 153-160. Ogden, J. A. E. and M. Rejmanek. 2005. Recovery of native plant communities after the control of a dominant invasive plant species, Foeniculum vul- gare: implications for management. Biological Conservation 125:427^139. Primack, R. B. 2010. Essentials of conservation biology. Sinauer Associates, Sunderland, MA. Schiffman, P. M. 2000. Mammal burrowing, erratic rainfall and the annual lifestyle in the California prairie: is it time for a paradigm shift? Pp. 153-160 in J. E. Keeley, M. Baer- Keeley, and C. J. Fotheringham (eds.), 2nd Interface between ecolo- gy and land development in California. U.S. Geological Survey Open-File Report 00-62. Web- site: http://pubs.usgs.gov/of/2000/of00-062/ [ac- cessed 30 June 2011]. — . 2007. Pleistocene and pre-European grassland ecosystems: species composition at the time of first European settlement. Pp. 52-56 in M. R. Strom- berg, J. D. Corbin, and C. M. D’Antonio (eds.), California grasslands: ecology and management. University of California Press, Berkeley, CA. Seabloom, E. W., E. T. Borer, V. L. Boucher, R. S. Burton, K. L. Cottingham, L. Goldwasser, W. Gram, B. E. Kendall, and F. Micheli. 2003. Competition, seed limitation, disturbance, and reestablishment of California native annual forbs. Ecological Applications 13:575-592. SlMBERLOFF, D. 2005. Non-native species do threaten the natural environment! Journal of Agricultural and Environmental Ethics 18:595-607. Stromberg, M. R., C. M. D’Antonio, T. P. Young, J. Wirka, and P. R. Kephart. 2007. California grassland restoration. Pp. 254-280 in M. R. Stromberg, J. D. Corbin, and C. M. D’Antonio (eds.), California grasslands: ecology and manage- ment. University of California Press, Berkeley, CA. Thomson, D. 2005. Measuring the effects of invasive species on the demography of a rare endemic plant. Biological Invasions 7:615-624. United States Fish and Wildlife Service. 1999. Recovery plan for six plants from the mountains surrounding the Los Angeles Basin. U.S. Depart- ment of the Interior, Fish and Wildlife Service, Region 1, Portland, OR. Website: http://www.fws. gov/ecos/ajax/docs/recovery_plan/990930a.pdf [ac- cessed 30 June 2011]. Walck, J. L., J. M. Baskin, and C. C. Baskin. 1999. Effects of competition from introduced plants on establishment, survival, growth and reproduction of the rare plant Solidago shortii (Asteraceae). Biological Conservation 88:213-219. Zavaleta, E. S., R. J. Hobbs, and H. A. Mooney. 2001. Viewing invasive species removal in a whole- ecosystem context. Trends in Ecology & Evolution 16:454=4-59. Madrono, Vol. 58, No. 2, pp. 78-85, 2011 POLLEN SIRING SUCCESS IN THE CALIFORNIA WILDFLOWER CLARKIA UNGUICULATA (ONAGRACEAE) Nancy L. Smith-Huerta Department of Botany, Miami University, Oxford, OH 45056 smithhn@muohio.edu Frank C. Vasek12 Department of Botany and Plant Sciences, University of California, Riverside, CA 92521 Abstract Cryptic self-incompatibility (CSI) is a type of non-random mating observed in self-compatible plants in which outcross pollen sires proportionally more seeds than self pollen when both pollen types are available on the stigma. Levels of CSI are known to vary among individuals and populations. We conducted competitive pollinations consisting of mixtures of self and outcross pollen to investigate reports of CSI in certain populations of Clarkia unguiculata Lindl. We also investigated how the order of self and outcross pollen deposition on the stigma influences the degree of nonrandom mating. Finally, we looked at whether the source population of outcross pollen affected the ability of outcross pollen to outcompete self pollen. We utilized recessive (white-petaled) maternal (and self pollen donor) plants from a locality near Morro Bay, California, and dominant (pink-petaled) outcross pollen donor plants from 17 localities widespread through the species range in California. Progenies from pollinations made with equal mixtures of self and cross pollen included significantly more outcross- pollinated than self-pollinated offspring in 10 of the 17 cross-pollen donor populations. However six populations showed no significant difference between self- and outcross-pollinated offspring, and one population yielded a majority of progeny sired by self pollen. Progenies from sequential self pollen followed by outcross pollinations included significantly more self offspring in 12 of the donor populations, no significant difference between outcross and self offspring in four donor populations and significantly more outcross offspring in one donor population. Progenies from sequential outcross pollen followed by self pollinations included significantly more outcross offspring in 15 donor populations and no significant difference between outcross and self offspring in two donor populations. Our results confirm the occurrence of non-random mating in C. unguiculata, and demonstrate that the degree of non-random mating can depend on the order of self vs. outcross pollen deposition and the source population of outcross pollen. This non-random mating can influence the proportion of self and outcross progeny in sequential pollinations. Key Words: Clarkia unguiculata, cryptic self-incompatibility, geitonogamy, non-random mating, sexual conflict. (Marshall 1998). This fact suggests that the source of outcross pollen is likely to affect the degree of non-random mating due to CSI. Self-incompatibility may evolve in outcrossing populations in response to inbreeding depression. If a normally outcrossing population carries a genetic load, self-incompatibility could confer a fitness advantage to maternal plants if they produce more outcross progeny than selfed progeny in mixed pollinations. In contrast to “complete” self-incompatibility, plants exhibiting CSI maintain the ability to produce progeny through self-fertilization. This ability should confer the fitness advantage of reproductive assurance to annual plants that grow where access to pollinators or mates is limited. During growing seasons when pollinators and other resources are plentiful, non-random mating by favoring outcross pollen allows maternal plants to increase fitness by increasing the proportion of outcross progeny they produce. During less optimal growing seasons, when resources and Cryptic self-incompatibility is essentially a case of non-random mating in which normal seed set occurs upon self pollination in the absence of competitive outcross pollen. However, with mixtures of self and outcross pollen, weak self rejection reactions promote preferential fertiliza- tion by outcross pollen (Bateman 1956; Weller and Ornduff 1977; Eckert and Allen 1997; Kruszewski and Galloway 2006). Nonrandom mating in plants is not limited to examples of CSI but also occurs in mixtures of outcross pollen from several donors, as in Raphanus sativus L. (Marshall and Ellstrand 1986; Marshall 1991, 1998). Of particular interest in these cases is the observation that pollen donors may differ significantly in their ability to sire seed, and the rank order of their siring ability can be consistent across several maternal plants 1 Retired. 2 Present address: 3418 Mono PI., Davis, CA 95618. 2011] SMITH-HUERTA AND VASEK: POLLEN SIRING SUCCESS IN CLARKIA 79 pollinators are limiting, maternal fitness is enhanced through self-pollination. Another important consideration is the role that non-random mating by favoring outcross pollen might play in promoting outcross pollina- tion in plants with varied patterns of pollen deposition. The number of pollen grains deposit- ed on the stigma in natural populations of Clarkia unguiculata Lindl. increases with time (Nemeth and Smith-Huerta 2003), indicating that pollina- tors make multiple visits to flowers, depositing loads of pollen sequentially over time. These loads of pollen are likely to vary in composition from mixtures of outcross pollen to loads of self pollen. Geitonogamous self pollination is possible since C. unguiculata is strongly protandrous (Vasek 1968, 1977; Vasek et al. 1987). A pollinator could visit a flower in the male phase and then visit another flower on the same plant in the female phase resulting in self pollination. It is possible that CSI could promote outcross pollination in cases where the stigma of a flower first receives a load of self pollen and soon after receives a load of outcross pollen. In fact, although geitonogamy is possible (and probable) in C. unguiculata , out- crossing rates in a natural population have been measured at greater than 90% (Vasek 1965). Non- random mating by favoring outcross pollen (CSI) may well have contributed to these high outcross- ing rates. In fact, CSI is known to occur in at least one population of C. unguiculata (Bowman 1987) and in one population of C. gracilis A. Nelson and J. F.Macbr.(Jones 1994). Interestingly, CSI was reported to be completely absent from another population of C. unguiculata (Travers and Mazer 2000). This suggests that the strength and nature and of non-random mating could vary from population to population and also with the source of outcross pollen. The purpose of our investigation was to examine a single population of C. unguiculata maternal plants to determine whether the order that self and outcross pollen is deposited on the stigma affects non-random mating. Furthermore, we examined if the source population of outcross pollen affects levels of CSI. This closer examina- tion of non-random mating in C. unguiculata is important given the possibility of sequential and geitonogamous pollination in C. unguiculata (Vasek 1968, 1977; Vasek et al. 1987; Nemeth and Smith-Huerta 2003), the variation in levels of CSI documented previously in two different populations of C. unguiculata (Bowman 1987; Travers and Mazer 2000), and that pollen donors may differ significantly in their ability to sire seed (Marshall 1998). Materials and Methods Mixed-donor pollinations were employed to determine the relative frequency of fertilization by self vs. cross pollen. Test plants were genetically marked for petal color. Flower color in C. unguiculata varies from white to red or purple but most populations are characterized by pink or lavender-pink flowers. White flowers are conditioned by a single recessive allele and pink or red flowers are conditioned by a dominant allele at the same locus (Rasmuson 1920; Vasek 1965, plus extensive unpublished data; Bowman 1984). Plants with pink flowers contain malvidin 3, 5-diglucoside in their flowers, leaves, stems and seedlings (Bowman 1987) whereas plants with white flowers lack this and other anthocyanins. Consequently, flower color can be determined in the seedling stage merely by scoring seedlings as green (non-anthocyanous) or red (anthocyanous) because green seedlings grow up to produce white flowers and red seedlings grow up to produce pink flowers (Bowman 1987; Vasek 1965). Organization of Experiments Ten white-flowered families were developed from available stocks from a natural population near Morro Bay, California because they were available and we knew they were fully self-fertile (Vasek 1986). Each family consisted of approx- imately 6—8 siblings (homozygous for white petal) and was used as a line of maternal test plants, and simultaneously as self-pollen donors. Seventeen pink-flowered families were devel- oped from available stocks from Morro Bay and 16 widespread localities in California (Table 1). Each family consisted of about 6-8 siblings and was used as a line of cross pollen donors. These plants were homozygous for pink flowers because they were grown from selected, available stocks known to have produced only pink-flowered progeny. Seeds from the source localities were sown on vermiculite in December, 1987 in a University of California greenhouse in Riverside, California. Subsequently, seedlings were transplanted to 6 inch pots of standard UC soil mix, irrigated as needed, fertilized weekly with half strength Hoagland’s solution and grown to maturity in the same greenhouse. Experimental pollinations were made during a period of about two months in the spring of 1988. Mixed-Donor Pollinations All experimental pollinations utilized pollen from a pink-flowered cross-pollen donor and pollen from a white-flowered self pollen donor. The self pollen donor was also the seed parent. Pollen was always applied to stigmas which were 1-2 days old as judged by the degree of stigma expansion not greater than 1 80 degrees (see figure 2 in Smith-Huerta and Vasek 1984). 80 MADRONO [Vol. 58 Table 1. Seed Collection Localities (All in California) for Cross-Pollen Donor Plants. DIR = direction from Morro Bay: S = Coast Ranges to the south; N = Coast Ranges to the north; E = Sierra Nevada to the East; DIS = approximate linear distance in km from Morro Bay. Population Locality County DIR DIS 15 Morro Bay - Atascadero Rd San Luis Obispo — 00 4 Santa Maria Santa Barbara S 58 5 Pinnacles Natl. Monument San Benito N 126 11 Rancheria Rd Kem E 185 6 Kern River Kern E 188 13 Caliente Hills, Low Canyon, Kern E 189 17 Caliente Hills, High Canyon Kern E 190 8 Santa Paula Ventura S 193 7 Sequoia Tulare E 220 9 Bouquet Canyon Los Angeles S 238 14 Laurel Canyon Los Angeles S 260 2 Jackson Amador E 330 3 Riverside -Fairmount Park Riverside S 350 10 San Luis Rey San Diego S 398 16 Old Castle Rd San Diego s 406 1 Clear Lake Lake N 430 12 Feather River, North Fork Butte E 485 One protocol (MIX) followed the methodology used by Bowman (1987). Cross pollen and self pollen were mixed together in equal amounts and then applied to the stigma of the maternal test plant, which was also the self-pollen donor. In the second protocol (S/X), self pollen was applied first followed immediately by cross pollen. In the third protocol (X/S), cross pollen was applied first, followed immediately by self pollen. We pollinated 246 flowers in 82 competitive pollinations. Each pollination included one pol- lination by each of the three protocols described above. Several siblings of each cross pollen donor population were used, for a total of 47 pollen donors from the 17 cross-pollen donor lines. Thus, the actual competitive comparisons in- volved closely related siblings rather than identi- cal individuals. Following competitive pollinations, the result- ing capsules were harvested at maturity, the seeds were then separated, counted, weighed and stored over the summer. Data Harvest Beginning in October 1988, seeds were sown on vermiculite in small plastic pots and put in a temperature controlled chamber. Temperatures were approximately 20 degrees C with a 12 h day 12 h night lighting schedule. Two weeks after germination, seedlings were moved to a green- house for continued development because space in the temperature controlled chamber was limited. Red pigments in stems and along leaf veins develop well with cool temperatures and bright daylight. Our greenhouse conditions were not always optimum for red pigment develop- ment (e.g., warm weather). Consequently, if any seedlings were not clearly red or not clearly green they were grown to maturity for direct scoring of flower color. Analysis We expect the progeny from competitive (MIX) pollinations to include half pink-flowered outcrosses if mating is non-random. However, preferential functioning of either cross or self pollen will yield ratios of white to pink progeny significantly different from 1:1. The frequency of outcrosses in each resulting progeny was multi- plied by 100 and expressed as a percent. The progeny of (S/X) pollinations should include significantly more self pollinated white than pink progeny if mating is random. We expect the first pollen grains on a receptive stigma should interfere with the normal germination of pollen arriving later. The progeny of (X/S) pollinations should include significantly more pink than white progeny. Each of the 246 progenies from MIX, S/X and X/S pollinations was tested for significant depar- ture from an expected 1:1 ratio of outcrosses to seifs by a G-test (Zar 1984). A Bonferroni P-value correction was applied to avoid inflated type-I error (avoid high risk of false positive results) across the multiple tests (Zar 1984). Progenies from all of the competitive pollinations within the 17 populations were also tested for significant departure from an expected 1 : 1 ratio of outcross to seifs by a G-test (Zar 1984). Results Both the order of pollen deposition and source of outcross pollen had variable affects on the degree of non-random mating due to CSI in our competitive pollinations (Table 2, Fig. 1). In the 2011] SMITH-HUERTA AND VASEK: POLLEN SIRING SUCCESS IN CLARKIA 81 Table 2. G-Test Values for all Progeny Produced in Competitive Pollinations. The symbols “*,” “***,” and “ns” indicate significant difference from a 1:1 ratio of self to outcross progeny at the 5%, 1%, and 0.1% levels, and not significantly different, respectively. Population MIX S/X X/S 1 68.219*** 48.753*** 61.926*** 2 16.934*** 0.134 ns 9.150** 3 34.464*** 1.590 ns 51.451*** 4 17.357*** 0.0055 ns 71.653*** 5 32.142*** 19.534*** 55.477*** 6 12.659*** 29.500*** 46.019*** 7 9.419** 11.027*** 16.933*** 8 14.710*** 37.016*** 42.659*** 9 7.905** 80.327*** 2.987 ns 10 7.266** 12.162*** 14.009*** 11 1.697 ns 29 354*** 30.372*** 12 2.337 ns 7.703** 24.856*** 13 0.032 ns 26.640*** 23.774*** 14 0.643 ns 9.550** 25.641*** 15 0.0303 ns 144.658*** 3.412 ns 16 0.024 ns 0.468 ns 13.832*** 17 11.512*** 106.878*** 9.662*** MIX competitive pollinations, 10 of the 17 cross pollen donor populations exhibited CSI, yielding progenies that significantly favored outcross pollen (Table 2, Fig. 1). However six of the 17 pollen donor populations did not exhibit CSI, and one population yielded a majority of progeny sired by self pollen (Table 2, Fig. 1). In the S/X competitive pollinations, one of the 17 cross pollen donor populations yielded prog- enies that significantly favored outcross pollen, 12 pollen donor populations significantly favored self pollen, and four of the donor populations did not differ significantly from a 1:1 ratio (Table 2, Fig. 1). Fifteen of the pollen donor populations in X/S competitive pollinations significantly favored outcross pollen, and two donor populations did not differ significantly from a 1:1 ratio (Table 2, Fig. 1). Discussion Overall, the results of our MIX pollinations confirm the occurrence of CSI and non-random mating in Clarkia unguiculata. More than half of the competitive MIX pollinations yielded proge- nies that significantly favored outcross pollen, and approximately 20% of competitive pollina- tions yielded progenies with significantly more offspring produced by self pollen. Only about 23% of competitive pollinations yielded progenies that did not differ from a 1:1 ratio. Our investigation differs in detail and scope from previous studies of CSI and non-random mating in Clarkia (Bowman 1987; Jones 1994; Travers and Mazer 2000), and may help to explain their variable results. Previous studies were conducted within single populations of plants (Jones 1994; Travers and Mazer 2000) or used a commercial seed source (Bowman 1987). Our study included maternal plants derived from one population and outcross pollen plants derived from 17 popula- tions distributed throughout California. In the present study, competitive pollinations made with pollen from 10 of the 17 populations yielded results similar to those found by Bowman (1987) and Jones (1994) with a majority of pollinations favoring outcross pollen. Competitive pollina- tions with outcross pollen derived from six of the populations yielded results similar to those found by Travers and Mazer (2000), with outcross pollen favored in fewer than half the pollinations. In contrast to all of the previous studies, competitive pollinations made with outcross pollen derived from 1 of our study populations yielded a majority of progenies in which self pollen was favored significantly over outcross pollen. Non-random mating when outcross pollen is favored over self pollen has the potential to reduce the negative effects of inbreeding depres- sion in populations through the increased pro- duction of outcross progeny, at the same time preserving the ability of individuals to produce offspring by selfing. Inbreeding depression has been documented in several populations of C. tembloriensis Vasek (Holtsford and Ellstrand 1990) and the magnitude of this inbreeding depression was shown to vary between popula- tions. It is possible that populations of C. unguiculata could vary with respect to genetic load, and selective pressure promoting non- random mating favoring self pollen might vary between populations. This could account for the variation observed in non-random mating in previous studies (Bowman 1987; Jones 1994; Travers and Mazer 2000). As stated above, non-random mating by favoring outcross pollen not only promotes the production of outcross progeny, but also pre- serves the ability of a plant to produce offspring through self pollination. This provides reproduc- tive assurance to annual plants that grow where access to mates may be limiting. Numbers of individuals in populations of C. unguiculata can vary from only a few plants to thousands of individuals (Lewis and Lewis 1955; Vasek 1964) and those growing in more marginal areas of the species range may experience large seasonal fluctuations in population size (Lewis and Lewis 1955). During seasons when plant populations are small, pollinators are rare, and access to mates limited, fitness might be enhanced by the production of self progeny. Our sequential pollinations address the ques- tion of whether non-random mating by favoring outcross pollen can somehow mitigate the effects of geitonogamy when self pollen arrives first on 82 MADRONO [Vol. 58 Outcross Progeny sig. < 50% Outcross Progeny not sig. diff. from 50% Outcross Progeny sig. > 50% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Outcross Pollen Source Population Fig. 1 . Percent of total competitive pollinations within each of 1 7 cross pollen donor populations that yielded outcross progenies either significantly more or significantly less than expected by chance, or yielded outcross progeny not significantly different from chance expectations. For the MIX treatment, cross and self pollen were mixed together in equal amounts and then applied to the stigma of the maternal test plant, which was also the self- pollen donor. For S/X, self pollen was applied first followed immediately by cross pollen. For X/S, cross pollen was applied first, followed immediately by self pollen. the stigma followed by outcross pollen. Rates of geitonogamy are influenced by the number of flowers that are open simultaneously on a plant (Karron et al. 2004). Clarkia unguiculata is strongly protandrous (anthers shed pollen before the stigma becomes receptive), and may have as many as 12 open flowers per spike (Vasek 1968, 1977; Vasek et al. 1987). Anthers of individual flowers shed pollen for up to 11 days before the stigma becomes receptive (Vasek 1968, 1977; Vasek et al. 1987). If pollinators visit more than one flower per plant, geitonogamous self polli- nation is likely. Geitonogamy does not provide reproductive assurance and is considered to be “an unavoidable by-product of selection for outcrossing success” in plants that have numer- ous flowers that are open simultaneously in an inflorescence (Goodwillie et al. 2005). It is possible that non-random mating by favoring outcross pollen could mitigate the unavoidable selfing that occurs when pollinators visit more that one flower per plant in C. unguiculata, and our S/X sequential pollinations were designed to explore this idea. If mating is random, S/X 2011] SMITH-HUERTA AND VASEK: POLLEN SIRING SUCCESS IN CLARKIA 83 pollinations (outcross pollen grains deposited immediately after self pollen grains) should yield progenies with significantly more self than outcross offspring. Twelve of the donor popula- tions did yield progenies that significantly fa- vored self pollen, even though six of these populations displayed non-random mating in MIX pollinations. In these cases, the documented presence of non-random mating did not appear to mitigate the effects of geitonogamy. In contrast, S/X sequential pollinations in five of the donor populations did not favor self progeny. Of these, four of the populations produced progeny that did not differ from a 1:1 ratio of self to outcross, and one population significantly favored outcross pollen. With one exception (population 16), these were the donor popula- tions that displayed the strongest non-random mating by favoring outcross pollen in MIX pollinations. It appears that strong non-random mating by favoring outcross pollen has the potential to mitigate the effects of geitonogamy when self pollen arrives first on the stigma followed by outcross pollen. Our X/S pollinations explored further the interactions of non-random mating and geitono- gamy in Clarkia. The X/S pollinations were expected to produce significantly more outcross than self progeny in all pollinations, since the advantage of outcross pollen arriving first on the stigma should be enhanced by non-random mating. This proved to be the case in all but two of the donor populations which showed no significant difference between the number of self and outcross progeny produced. Interestingly, these two donor populations were the only populations in which self pollen was significantly favored in all of their sequential S/X pollinations. The physical mechanism responsible for differ- ential siring success in non-random mating and CSI has been investigated in other studies (Hessing 1989; Weller and Ornduff 1989; Aizen et al. 1990; Eckert and Allen 1997; Kruszewski and Galloway 2006; Figueroa-Castro and Holts- ford 2009). In all of the above investigations, with one exception, outcross pollen germinated faster on the stigma and grew faster through the style than self pollen. The single exception occurred in Campanulas trum americanum Small in which pollen tube growth rates did not differ between self and outcross pollen (Kruszewski and Gallo- way 2006). Although we did not measure pollen germination and tube growth rates in the present study, this was the focus of previous studies in both C. unguiculata (Nemeth and Smith-Huerta 2002; Smith-Huerta et al. 2007) and C. temblor- iensis (Smith-Huerta 1996; Kerwin and Smith- Huerta 2000). Similar to Campanulastrum amer- icanum (Kruszewski and Galloway 2006), no difference in percent germination or rate of pollen tube growth was observed between self and outcross pollen in single donor pollinations in these two Clarkia species (Smith-Huerta 1996; Kerwin and Smith-Huerta 2000; Nemeth and Smith-Huerta 2002). It is possible that self and outcross pollen must be present together on the stigma and in the style for the non-random mating of CSI to occur. This appears to be the case in Clarkia. Pollen germination was signifi- cantly reduced in two donor pollinations of self + outcross pollen and in outcross + outcross pollen (from two different donors) in C. unguiculata (Nemeth and Smith-Huerta 2002). Further, germination of pollen decreased with increasing contact between pollen grains (Nemeth and Smith-Huerta 2002). Pollen-pollen interactions, mediated by the stigma, might provide a possible mechanism to explain the differences in the relative success of self and outcross pollen observed in reported non-random mating and CSI in Clarkia. The present investigation goes beyond an examination of self vs. outcross pollen perfor- mance within a single population, and examines the performance of outcross pollen derived from foreign populations. These outcross pollen source populations occur from 58 to 485 km from the maternal (self pollen) population. Clearly, the maternal plants in our study would not normally encounter outcross pollen from these popula- tions. Interestingly, the Morro Bay population, which provided all of our maternal plants, did not show CSI in MIX crosses when Morro Bay plants were the source of outcross pollen. In contrast, levels of CSI could be very high when pollen from foreign populations was used. This great difference in outcross pollen donor success suggests that pollen-pistil interaction may evolve differently in each population of C. unguiculata, as a result of sexual conflict between male and female function. In plants, sexual conflict occurs when optimal reproductive fitness strategies for pollen differ from those of the maternal plant. The trait of CSI has the potential to enhance the fitness of ovules but not of pollen, thus creating a male-female sexual conflict. Although sexual conflict has been studied mostly in animals, several recent studies have examined the occur- rence of sexual conflict in plants (Prasad and Bedhomme 2006; Lankinen and Larsson 2009; Madjidian and Lankinen 2009). In one instance, similar to the present study, male-female interac- tions were investigated in cross pollinations between plants derived from 4 different popula- tions of Collinsia heterophylla Graham (Madji- dian and Lankinen 2009). In this plant, the onset of stigma receptivity may be affected by both the source and recipient of the pollen, with early receptivity and fertilization resulting in the production of fewer seeds than late receptivity. In experimental cross pollinations, it was found that pollen donors from foreign populations were 84 MADRONO more successful at inducing stigma receptivity than pollen donors derived from the same population (Madjidian and Lankinen 2009). These results may be interpreted to suggest that there is “sexually antagonistic coevolution” between maternal plants and pollen within populations of C. heterophylla (Madjidian and Lankinen 2009). It appears that when maternal plants receive pollen from foreign populations they are somehow “released from the cost of local pollen” (Madjidian and Lankinen 2009). In the present study, CSI and non-random mating varied extensively depending on the population source of outcross pollen. It is possible that this reflects a similar “release from the cost of local pollen” in our experiments. In sum, we document the occurrence of non- random mating and CSI in crosses between different populations of Clarkia unguiculata, and demonstrate that this non-random mating can also influence the proportion of self and outcross progeny in sequential pollinations. The non-random mating observed in CSI may pro- mote outcrossing in protandrous plants subject to geitonogamous pollination and contribute to reproductive assurance when access to mates is limited. Finally, the fact that levels of CSI vary between sources of pollen donors suggests that sexual conflict between pollen and maternal plants may result in coevolution unique to each population of C. unguiculata. Acknowledgments We are grateful to the University of California Agricultural Experiment Station for support, to Mi- chael Hughes for help with the statistical analysis, to Vincent Weng for his technical assistance, and to Carla Smith for assistance with seed collection. We also thank two anonymous reviewers for their very helpful suggestions for improvement of the manuscript. This work is dedicated to the memory of our mentor Professor Harlan Lewis. Literature Cited Aizen, M. A., K. B. Searcy, and D. L. Mulcahy. 1990. Among- and within-flower comparisons of pollen tube growth following self- and cross- pollinations in Dianthus chinensis (Caryophylla- ceae). American Journal of Botany 77:671-676. Bateman, A. J. 1956. Cryptic self-incompatibility in the wallflower. Heredity 10:257-261. Bowman, R. N. 1984. Experimental non-stigmatic pollinations in Clarkia unguiculata (Onagraceae). American Journal of Botany 71:1338-1346. — . 1987. Cryptic self-incompatibility and the breeding system of Clarkia unguiculata (Onagra- ceae). American Journal of Botany 74:471^-76. Eckert, C. G. and M. Allen. 1997. Cryptic self- incompatibility in tristylous Decodon verticillatus (Lythraceae). American Journal of Botany 84: 1391-1397. Figueroa-Castro, D. M. and T. P. Holtzford. 2009. Post-pollination mechanisms in Nicotiana [Vol. 58 longiflora and N. plumb aginifolia : pollen tube growth rate, offspring paternity and hybridization. Sexual Plant Reproduction 22:187-196. Goodwillie, C., S. Kalisz, and C. G. Eckert. 2005. The evolutionary enigma of mixed mating systems in plants: occurrence, theoretical explanations, and empirical evidence. Annual Review of Ecology, Evolution, and Systematics 36:47-79. Hessing, M. B. 1989. Differential pollen tube success in Geranium caespitosum. Botanical Gazzette 150:404^110. Holtsford, T. P. and N. C. Ellstrand. 1990. Inbreeding effects in Clarkia tembloriensis (Ona- graceae) populations with different natural out- crossing rates. Evolution 44:2031-2046. Jones, K. N. 1994. Nonrandom mating in Clarkia gracilis (Onagraceae): a case of cryptic self-incom- patibility. American Journal of Botany 81:195-198. Karron, J. D., R. J. Mitchell, K. G. Holmquist, J. M. Bell, and B. Funk. 2004. The influence of floral display size on selfing rates in Mimulus r ingens. Heredity 92:242-248. Kerwin, M. A. and N. L. Smith-Huerta. 2000. Pollen and pistil effects on pollen germination and tube growth in selfing and outcrossing populations of Clarkia tembloriensis. International Journal of Plant Science 161:895-902. Kruszewski, L. J. and L. F. Galloway. 2006. Explaining outcrossing rate in Campanulas trum americanum (Campanulaceae): geitonogamy and cryptic self-incompatibility. International Journal of Plant Sciences 167:455-461. Lankinen, A. and M. C. Larsson. 2009. Conflicting selection pressures on reproductive functions and speciation in plants. Evolutionary Ecology 23: 147-157. Lewis, H. and M. E. Lewis. 1955. The genus Clarkia. University of California Publications in Botany 20:241-392. Madjidian, J. A. and A. Lankinen. 2009. Sexual conflict and sexually antagonistic coevolution in an annual plant. PLoS ONE 4:e5477. Marshall, D. L. 1991. Nonrandom mating in wild radish: variation in pollen donor success and effects of multiple paternity among one- to six-donor pollinations. American Journal of Botany 78:1404-1418. — . 1998. Pollen donor performance can be consistent across maternal plants in wild radish (Raphanus sativus, Brassicaceae): a necessary condition for the action of sexual selection. American Journal of Botany 85:1389-1397. — and N. C. Ellstrand. 1986. Sexual selection in Raphanus sativus : experimental data on the nonrandom fertilization, maternal choice, and consequences of multiple paternity. American Naturalist 131:736-759. Nemeth, M. B. and N. L. Smith-Huerta. 2002. Effects of pollen load composition and deposition pattern on pollen performance in Clarkia unguicu- lata. International Journal of Plant Science 163:795-802. AND . 2003. Pollen deposition, pollen tube growth, seed production, and seedling perfor- mance in natural populations of Clarkia unguicu- lata (Onagraceae). International Journal of Plant Science 164:153-164. 2011] SMITH-HUERTA AND VASEK: POLLEN SIRING SUCCESS IN CLARKIA 85 Prasad, N. G. and S. Bedhomme. 2006. Sexual conflicts in plants. Journal of Genetics 85:161-164. Rasmuson, H. 1920. Die Hauptergebnisse von einigen genetischen Versuchen mit verschiedenen von Tropaeolum, Clarkia und Impatiens. Hereditas 1:270-276. Smith-Huerta, N. L. 1996. Pollen germination and tube growth in selfing and outcrossing populations of Clarkia tembloriensis (Onagraceae). Internation- al Journal of Plant Science 157:228-233. and F. C. Vasek. 1984. Pollen longevity and stigma pre-emption in Clarkia. American Journal of Botany 71:1183-1191. , S. R. Carrino-Kyker, and A. J. Huerta. 2007. The effects of maternal and paternal nutrient status on pollen performance in the wildflower Clarkia unguiculata (Onagraceae). Journal of the Torrey Botanical Society 134:451-457. Travers, S. E. and S. J. Mazer. 2000. The absence of crypic self-incompatibility in Clarkia unguiculata (Onagraceae). American Journal of Botany 87:191-196. Vasek, F. C. 1964. The evolution of Clarkia derivatives adapted to relatively xeric environments. Evolution 18:26-12. . 1965. Outcrossing in natural populations. II. Clarkia unquiculata. Evolution 19:152- 156. — . 1968. The relationships of two ecologically marginal, sympatric Clarkia populations. Ameri- can Naturalist 102:25-40. — . 1977. Phenotypic variation and adaptation in Clarkia section Phaeostoma. Systematic Botany 2:251-279. — . 1986. Evolutionary modification in Clarkia. I. Selection for hairlessness. American Journal of Botany 73:517-523. — , V. Weng, R. J. Beaver, and C. K. Huzar. 1987. Effects of mineral nutrition on components of reproduction in Clarkia unguiculata. Aliso 11:599-618. Weller, S. G. and R. Ornduff. 1977. Cryptic self- incompatibility in Amsinckia grandiflora. Evolution 31:47-51. — AND . 1989. Incompatibility in Amsinckia grandiflora (Boraginaceae). American Journal of Botany 78:801-804. Zar, J. H. 1984. Biostatistical analysis, 2nd ed. Prentice-Hall, Inc., Englewood Cliffs, NJ. Madrono, Vol. 58, No. 2, pp. 86-91, 2011 SMALLER OLEA EUROPAEA FRUITS HAVE MORE POTENTIAL DISPERSERS: IMPLICATIONS FOR OLIVE INVASIVENESS IN CALIFORNIA Clare E. Aslan and Marcel RejmAnek Department of Evolution and Ecology, University of California, Davis, One Shields Avenue, Davis, CA 95616 ceaslan@ucdavis.edu Abstract Olea europaea (European olive) is invasive in Australia and widely planted in California. Vertebrates, particularly birds, mediate Olea seed dispersal. Fruits are large, but their sizes range widely. We measured fruit widths from 12 study stands in California and constructed esophageal probes in diameters spanning the resulting size range. We then obtained whole-bird frugivorous bird carcasses and used the probes to determine the fruit sizes that each bird species would be anatomically capable of swallowing. This allowed us to develop lists of potential dispersers for each study stand. Even stands with the largest measured fruits had several potential disperser species, but the list of species expanded greatly as fruit sizes decreased. Feral Olea stands with remarkably small fruits have been observed in California and Australia. An increase in the incidence of such stands might augment the regional spread rate for the species. Key Words: Bird-mediated dispersal, European olive, gape width, invasion, museum specimens, Olea europaea. A variety of invasive, or potentially invasive, woody species are dispersed by birds (Richardson et al. 2000; Aslan and Rejmanek 2010). Among these is Olea europaea L. (Oleaceae) (European olive, hereafter, Olea), an upland species native to the Mediterranean. Olea is widely cultivated in California, as well as in other regions similar in climate, as a crop and landscaping species. Olea has become invasive in natural areas in Australia, where its fruits are bird-dispersed (Spennemann and Allen 2000b). Field observations have confirmed that birds consume Olea fruits in California, as well (Aslan 2010). Since (a) the climate is appropriate for Olea establishment, and (b) the species demonstrates dispersal mutu- alisms with resident birds, a potential incipient invasion by Olea in California appears possible. Effective dispersal of Olea (and, by extension, the likelihood of invasion) may be dictated by fruit sizes. Olea fruit sizes range enormously (e.g., from a minimum of 5.6 mm to a maximum of 28.2 mm in width for fruits we collected from studied Californian populations). In general, however, the species is large-fruited and seed mass is positively correlated with fruit size (Alcantara 1995), making vertebrate-mediated dispersal necessary for removal of seeds from parental locations. Since the species is cultivated primarily for human food, artificial selection for large fruits has been imposed on Olea strains (Rey and Gutierrez 1996). However, fruit size may constrain seed dispersal. Particularly large fruits of Olea and other species are often pecked rather than swallowed whole by birds (Rey and Gutierrez 1996). This strategy is necessitated by gape width limitations, but appears energetically costly for birds, which prefer to swallow fruits whole when they are able to do so (Rey and Gutierrez 1996). Fruit pecking transforms a mutualistic relationship into seed predation by rarely promoting effective dispersal beyond the stand canopy (Rey et al. 1997). Among wild olives, therefore, gape width limitations of potential dispersers appear to counteract selective pressures favoring large seeds (e.g., enhanced germination success and greater seedling vigor) and likely dictate seed size maxima (Alcantara and Rey 2003). Olea was introduced to Australia in 1800 and planted in many locations over the next two centuries, but its profitability as a crop species was generally low, and most olive stands were abandoned (Hartmann 1962; Spennemann and Allen 2000b). Fruit and seed sizes in feral Olea populations in Australia are reportedly smaller than those of cultivated orchards (Spennemann and Allen 2000a). Size reduction in self-seeded populations that are geographically separated from any parent stand may result from one of two mechanisms: lack of anthropogenic cultiva- tion and care in the early stages of stand establishment, or selection of smaller fruits by birds. Care (in the form of irrigation, pruning, fruit harvesting, or fertilization) is no longer provided in abandoned olive stands in Australia nor in our study stands in California, all of which are old and established as hedgerows rather than orchards; however, at least some irrigation was likely performed at initial planting in both countries. The phenomenon of fruit size reduc- tion following disperser-mediated selection has been observed for palm seeds following toucan 201 1] ASLAN AND REJMANEK: EUROPEAN OLIVE FRUIT SIZE IMPLICATIONS 87 Table 1 . Olea europaea Study Stands and Relevant Characteristics. Study stand (ranked by minimum fruit size) Location Land-use type Varietal Feral 1 Chico Semi-natural Mission Yes 2 Chico Semi-natural Mission No 3 Chico Urban Mission No 4 BCCER Semi-natural Mission No 5 PC Agricultural Manzanillo No 6 Davis Agricultural Manzanillo No 7 PC Agricultural Manzanillo No 8 Davis Agricultural Manzanillo No 9 BCCER Semi-natural Mission No 10 Davis Urban Manzanillo No 11 PC Agricultural Manzanillo No 12 Davis Urban Manzanillo No exclusion (Galetti et al. 2010), while birds in a Spanish woodland selected Prunus mahaleb L. fruits that were significantly smaller on average than the mean size of available fruits (Jordano 1995). Furthermore, minimum fruit size was found to be more predictive of dispersal interac- tions than was average fruit size in a network of frugivores and large-seeded fruit in New Zealand (Kelly et al. 2010), implying that fruits small enough for ingestion were attractive even when they were surrounded by larger seeds. Logically, fruit size is likely to decrease following these selective pressures, and the suites of vertebrate species able to potentially disperse olive seeds may grow as a result (Spennemann and Allen 2000a). Theoretically, then, rates of dispersal and invasion may escalate as feralization continues. In California, spontaneous spread of Olea is presently rare; we know of only two completely feral and reproductive populations of Olea at this time. Since Olea stands in California exhibit a range of fruit and seed sizes, we expect that the bird-mediated dispersal potential of different stands varies according to those sizes. Most Olea individuals, cultivated in orchards, have large fruits and receive irrigation, fertilizer, and prun- ing. Outside of orchards, many trees in old hedgerows that receive little care today but were likely irrigated at planting also exhibit large fruits. We hypothesize that dispersal of seeds from such trees is possible for only a limited number of bird species and is therefore relatively rare. We have, however, identified several Olea stands (generally planted for landscaping rather than crop purposes) with fruits that are quite small. We hypothesize that a wider diversity of bird species are capable of dispersing these fruits. As a general rule, decreasing bird body size shows a slight but significant correlation with increasing population density across species (Juanes 1986). As olive fruits shrink and birds of smaller body size are capable of consuming them, the number of individual birds capable of dispersing fruits should increase at a greater rate than the number of disperser species. Corre- spondingly, dispersal potential itself will grow at an accelerating rate. Due to both of these increases (increased diversity and increased population sizes) among disperser birds, we consider it likely that expand- ed suites of potential dispersers will result in greater dispersal, possibly creating small-seeded and small-fruited feral populations as has likely occurred in Australia (Spennemann and Allen 2000b). The appearance and multiplication of such feral populations in California might signify the beginning of a large-scale invasion by this species. Here, we used esophageal probes and Califor- nian bird carcasses to identify the suites of bird species that are likely capable of swallowing Olea fruits of varying sizes. We compare the diversity of birds comprising each suite and discuss the probable implications of fruit size reduction for Olea invasiveness. Methods Study Species: Olea europaea In California, Olea occurs in at least 27 counties (Calflora 2010) and is valued at more than $86 million as an agricultural commodity (CDFA 2008). For this study, we examined fruits from Olea stands in Yolo and Butte Counties in north-central California. Among the 12 focal stands, three were located in urban areas (Chico 3, Davis 3, and Davis 4), five in agricultural areas (Davis 1, Davis 2, PC 1, PC 2, and PC 3), and four in semi-natural areas (an abandoned home- stead (BCCER 1 and 2), a municipal protected area (Chico 2) and one completely feral popula- tion in a previously-grazed, protected canyon (Chico 1) (Table 1). None of the study stands currently receive irrigation or fertilization. Esophageal Fruit Passage Assessment In January, 2009, we collected 10 ripe olives from each of up to 12 trees per study stand, for a 88 MADRONO [Vol. 58 total of 60-120 olives from each of the 12 study stands. We used digimatic (Mitutoyo America Corporation, Aurora, IL) calipers to measure the width of each olive to the tenth of a millimeter. We calculated the mean, minimum, and maxi- mum fruit widths for each population. Trans- verse diameter (width) was utilized because it is likely to be the limiting dimension determining the maximum size of fruits that can be swallowed whole (Wheelwright 1985). We examined the relationship between fruit sizes and land use type using ANOVA. We then constructed gape width probes to determine which bird species are likely physically capable of swallowing average- and minimum- sized olives from each stand. Each probe consisted of an ellipsoid made of Teflon plumb- ing tape attached to one end of a wooden dowel. A total of 9 probes were created. The probes were measured to the tenth of a millimeter using digimatic calipers and created to incrementally span the range of average fruit sizes from the study stands. Because we used Teflon tape instead of a rigid or brittle substance, the probes had a smooth and slightly flexible surface, which to the touch resembled the surface of ripe Olea fruits and yielded similarly (slightly) when squeezed in the hand. We wrapped the Teflon tape as tightly as possible to create surface tension, mimicking the firmness of a full, ripe fruit. We obtained previously-frozen, whole frugivo- rous bird carcasses from the University of California, Davis, Wildlife Museum. Once they had thawed, all carcasses were in good condition; we used none that had dried or otherwise stiffened, and all had mouths and esophagi that were open and unobstructed. The use of such carcasses instead of dried, prepared skins with closed mouths enabled us to both (a) reduce the risk of skin shrinkage that would alter measure- ments and mislead our conclusions, and (b) verify the ability of each bird to pass olives all the way through the esophagus and into the stomach. For each carcass, we measured the external bill gape width (bill width at the commissures) and the external maximum bill height (from the top to the bottom of the bill at its base). We then inserted each of the nine probes into each carcass’s mouth to determine whether it could pass through the full esophagus to the stomach. We repeated this for up to 10 carcasses per bird species (depending upon availability). These trials generated a list of the stands containing fruits that could morpho- logically be dispersed by each tested bird species. When all available carcasses (representing 18 species) had been tested, a few bird species remained that we believed to be candidates for olive dispersal (i.e., frugivorous in habit and occurring in olive stand sites) but that were not available as whole carcasses. These species included American crow ( Corvus brachyr- hynchos), California quail ( Callipepla calif ornica), northern mockingbird ( Mimus polyglottos ), spot- ted towhee ( Pipilo maculatus), western meadow- lark ( Sturnella neglecta ), and wild turkey ( Melea - gris gallopavo). Through logistic regression, we ascertained that external bill gape width was predictive of successful esophageal passage of all probes (P < 0.0001). We thus measured external gape width of prepared museum specimens of the remaining candidate bird species to determine which stands they are capable of dispersing. We used 10 museum specimens each for these species except when fewer than 10 were available (the museum owned only four turkey specimens, and all were used). All measurements were taken at the commissures and therefore relevant to the hard bill dimensions rather than skin dimensions; nevertheless, some drying and shrinkage could occur on study skins during their preparation. In light of our hypotheses (that a greater number of bird species should feasibly disperse fruits of smaller size), any shrinkage that may have led us to falsely conclude a negative (that a given bird species could not disperse a given fruit size) would lead us to underestimate the number of species capable of dispersal, making our ap- proach conservative. In winter, 2008-2009, we conducted bird counts at 70 points in each of the four study sites. Each count lasted seven minutes. At each point, all birds detected by sight or sound were recorded by species and number of individuals. To compare population densities among birds of smaller and larger body size, we examined point count results truncated at 25 m from the point (i.e., only birds detected within a 25-m radius of the point, or an area of 1963.5 m2, were included in analyses). We calculated the density of detections (number of individuals per hectare) for each of the 22 potential disperser species identified by the esophageal probes and logistic regression. We then performed two linear regres- sions; first, we regressed the number of potential dispersers on the minimum fruit width to determine whether there was a significant in- crease in the size of potential disperser suites as minimum olive size decreased across stands. Second, we regressed the total density of detec- tions of potential disperser species on the minimum fruit width to determine whether the density of total potential dispersers, across all species, increased more rapidly than did the diversity of dispersers as minimum olive size decreased. To compare slopes between the two regressions, we used proportional values, angu- lar-transformed to meet regression assumptions, and graphed them together in the same space. All statistical analyses were performed in JMP 5.0 (SAS Institute Inc., Cary, NC). Significance was accepted at P < 0.05. 2011] ASLAN AND REJMANEK: EUROPEAN OLIVE FRUIT SIZE IMPLICATIONS W G %-B «? « H 73 S -d ‘ G G W (G P on G a x) e g < s or pj X 0 -d < • G H >< Q is u X < CO S g II II II fli llfeiSs? I|38*4 •G . . . r . r oo ■ x •a .22 > o g f JH h £ o o 9 baX o 5-aMlt St g " e •2 S 11 11 « = < O X >h O g X O > .G ^ ( . ^ . r. C/3 £3 <2 S s ^ 2 o ^ 73 * .3 a e P .tJ _ fl ^ o ^ C Jh (U C/5 5^ I °£ 3 = * ^ r ^ ° > H 3 G G < o 3 o „ go < 'S .3 II I II g i'ODSS^ 0 & a2 b £ CD G d- X > CD i> gj x < !> X ^ ^ r„ Ui^ b 2 8 " tI O - B d x a .5 (B x o _Tx x •'->x • Crt O W V3 > s S -2 C 2 x f ^ £ G n d G G « O-h -d £ Ja p + 1 C/3 5 B X G a* §2 “j s 7? (U 3 s O o 0 p M o 5 j a 3 Q (X w O > W d « 0 S O < < U d M £ a< u r u Q CL, 7 CO £ go X S3 SE II $ 1 1 < O < i- i 5 " ®.fe| » » O •§ .. | S s 8-0 52 ^ > X -d G d X X r g G G W * o a g Ji S 6 g-g B o b h x X ^ X g ° „ o p ^ g II a ^ G O.Hl II II U x b H 11 UJ X o <£ ffi H pq g g . r x o ^ || ^ CO > ■S£<* fc-2-S §3y ass p ^ S 73 w 5 2 ^ o g x w o 2 G X 2 ^ s v s | X „ 'd d p !> x O 2 2^ II 'd H x ”o o x « -r-d 11 H S bg X X Q ^ oS«$ig ■§Sg°g t: Stf 1.^-1 u C^o X a X n ;X g c O oo«8,S? ?S3°« (3< -ox W G £ SI >P d3 X G 5 G °J«0gD0° ao w ag (w)X'^x>0/P>0/x^O/x^oo udru^aubauP^uPDw X X X P S S s X X X X X X p p s H X X X < < < > > > > < 1—3 < ^3 < ^-5 p X H X X < oo CO GO > 0 0 0 S *> S Z ?.S 3 ” s " 2 " S> S uoygoSgoSgow^ogS < < X X >H JH DO H X H H oo oo D D X < x < x < _r < _r < < W W ^ q2 H H o4 H pi H J? p2 < c c c < e£b gj is *K X> ^ ffi ^ b w ^ p< i a^> 2 X S X X H < > H GO 6 H X CO ax * 56>8 O £ < O ! X X CO D< PQ <0 U r_ X CO X QQ 0*,-<:_r ^ X .X shp^h n wOH H oo X X o H co D X §1 2? 10% and four of them have molar Ca/Mg ratios <0.7 (Fig. 2). One sample with Ca/Mg = 1.14 seems out of place plotted between 0.3 and 0.4 (Fig. 2), and maybe it is, but it has nearly as much hornblende as olivine. Springer (1980) showed analyses of 29 samples of gabbro. Only two of them had molar Ca/Mg < 1.0 and one had Ca/Mg < 0.7. He arranged the gabbro into four groups, based on the calcium concentration in the plagioclase: group A, An>85; B, An70_85; C, An55_69; and D, An<55. The means of the four groups are shown in Fig. 2. There might have been appreciable additions of aeolian dust, or loess, to the soils developed on gabbro in the Pine Hill area and the Peninsular Ranges. Very fine sand and coarse silt are the major components of aeolian dust (Pye 1995). The main sources of aeolian dust have been the Sacramento Valley, about 40 km from the Pine Hill soil sampling sites, and the coast of the Pacific Ocean, about 60 km from the Peninsular Ranges soil sampling sites. No appreciable amounts of coarse sand are likely to have been carried these distances, but appreciable accumu- lations of very fine sand are possible in the soils. Soil Parent Material Influences on Plant Species Distributions The chemical composition of gabbro is inter- mediate between ultramafic and granitic rocks. Ultramafic rocks are represented by the base of the triangle in Fig. 2 and granitic rocks (quartz diorite to granite) are more silicic than rocks represented by the left edge of the triangle. Anothosite, represented by the apex of the triangle, is an ancient rock found in California only in the San Gabriel and Orocopia Moun- tains; it is generally much less calcic than anorthite, which is the plagioclase feldspar with more than 90% Ca and less than 10% Na. Ultramafic rocks and soils are well known for having unique plant communities and many 116 MADRONO [Vol. 58 Fig. 3. Hornblende-bearing gabbro. Molar Ca/Mg ratios (0. 8-3.0) of gabbro composed of plagioclase (PI), hornblende (Hb), and pyroxene (Py). Plagioclase is represented by labradorite (Cao.6Nao.4Al! 6Si2.408), hornblende by Ca2Mg3AlFeAl2Si6022(OH)2, and pyroxene by augite (Cao^Mgo^Feo^SiCb). Symbols as in Fig. 2: o, Los Pinos complex; # Santa Ana Mountains. Rock samples from the three Guatay-Cuyamaca soils were plotted by the percentages of plagioclase, clinopyroxenes, and hornblende (symbols CGI, CG2, and CG3). endemic species. The absence of many species that are common on soils of more silicic rocks has been attributed to low exchangeable Ca/Mg ratios (Alexander et al. 2007). Some gabbro has low Ca/Mg ratios, but the ratios were not particularly low in the gabbro soils sampled in the Pine Hill area and Peninsular Ranges. Vegetation differences from gabbro to quartz diorite soils are evident, but not as dramatic as the differences from gabbro to ultramafic soils (Whittaker 1960). More perfuse vegetation on quartz diorite than on gabbro soils (Whittaker 1960) might be explained by greater fertility of quartz diorite soils related to greater amounts of K and P in diorite than in gabbro (Table 1). On a mafic to silicic scale, fertility may be greatest on diorite soils, which generally have a more favorable balance of plant nutrients than either mafic or granitic soils. An anorthosite soil in the San Gabriel Moun- tains was sampled and characterized by Graham et al. (1988). Although the parent rock had more than 90% plagioclase, the feldspar was andesine that has slightly more Na than Ca. Nevertheless the soil had much more exchangeable Ca than Na. It had low exhangeable K and presumably low P, although the analyses did not include P. Vlamis et al. (1954) grew lettuce in 13 soils sampled on gabbro, anorthosite, diorite, and granodiorite. The lettuce responded to N addi- tions on all of the soils and to P on most of them, except on an infertile anorthosite soil where there was no response to a complete N + P + K fertilizer either. Only on an anorthosite soil did the lettuce respond to K addition. Although plant cover appeared to be related to soil productivity, there were no distinct differences in plant species distributions among the soils. Other than sparse trees ( Pinus coulteri D. Don and Quercus chrysolepis Leibm.), only shrubs and Yucca sp. were reported from the soil sampling sites; the shrubs were Adenostoma fasciculatum Hook. & Arn., Arctostaphylos spp., Ceanothus spp., Cer- cocarpus sp., Eriodictyon sp., Lotus scoparius (Nutt.) Ottley, Quercus spp., Malosma laurina (Nutt.) Abrams, and Salvia spp. Gabbro soils and their parent rocks were sampled at two sites with unique vegetation and one without unique vegetation in both Research Natural Areas (RNAs) in the Peninsular Ranges and the Pine Hill Preserve in the foothills of the Sierra Nevada. Altitudes are 460 to 590 m at the Pine Hill sites and 1270 to 1385 m at the Guatay and King Creek RNA sites in the Peninsular Ranges. The mean annual precipitation at these sites is in the 750 to 800 mm range. At a site on the upper one-half of a moderately steep (21-36% 2011] ALEXANDER: GABBRO SOILS AND PLANTS f gradient) convex slightly slope in each plant community, three sets of soil and rock samples were taken from subsites three to nine meters apart. The soils were sampled at the 0-15 cm and 30-45 cm depths. Each subsite soil sample was a composite of three subsamples taken from points one to three meters apart. Plants were identified by reference to The Jepson Manual (Hickman 1993) and cover areas were estimated visually. Plant species in a list for the Pine Hill sites were verified by Graciela Hinshaw (Bureau of Land Management, Eldorado Hills, CA) and those in a list for the Peninsular Ranges sites were verified by Todd Keeler-Wolf (California Department of Fish and Game, Sacramento). Soil samples were dried and sieved to separate the fine-earth (particles <2 mm) for sand and chemical analyses. Texture and consistence of wetted samples were estimated by feel and identified by USD A nomenclature (Soil Survey Staff 1993). Soil family classification follows Soil Taxonomy of the USDA (Soil Survey Staff 1999). Minerals were estimated in 36 fields of view in thin-sections of rocks-one rock from each sample site. Sand was separated from fine-earth samples following treatments overnight in household bleach, decantation of clay, and overnight with Na dithionite in Na citrate solution. Coarse silt (30-62 pm) was obtained from the Pine Hill samples by decantation. Sand separates were dry sieved to obtain five size fractions, and the fine sand (0.125-0.25 mm) fraction was separated into light and heavy grains with bromoform (SG = 2.85). Magnetic grains were separated from the fine sand, very fine sand, and coarse silt fractions with a hand magnet. Percentages of very fine sand and coarse silt with low refractive indices (RI < 1.56) were ascertained by observing 300 grains from each Pine Hill soil. Soil pH was ascertained with a glass electrode in 1:1 water:soil suspensions. Calcium and magnesium were extracted with molar KC1 and measured by atomic absorption (AA) spectrom- etry. Exchangeable acidity was extracted with 0.5 molar KCl-triethanolamine pH 8.2 buffer and back-titrated with 0.08 molar HC1 to a methyl red endpoint (Peech 1965). Soil organic matter was approximated by loss-on-ignition (LOI) at 360°C. Results of chemical analyses are displayed as means of three samples from each site with standard deviations in parenthesis. One-way ANOVA was run for the three Pine Hill sites and the three Peninsular Ranges sites; that is three sites with three replications at each site in each run of ANOVA. Differences between means in each set of three sites were evaluated by the least significant differences (Snedecor and Co- chrane 1967). Where the differences are signifi- cant at the 95% level of confidence (a < 0.05), the values are designated a and b if the means are on two levels, or a, b, and c if they were on three p P 0 03 CJ P a p >> P U o U o bfl .* +_> O O p p U W £ T3 § u p ft ii 1 g o „ P P P H X kJ P ' P -G 03 X ™ ft cj 03 ft 00 S3 X W x X o o £ g CJ O t/5 X W X o CJ CJ G 03 S o £ <5 M X ^ o 23 (D HH 03 ^ 0 o % CJ X S-2 c/3 ft r P 2 X G CJ X ; O « X CJ rX X ^ CO cj ;L’ft o a OJ X £ -2 CO ft P a X -O o =3 13 X o x u ■ T 2 8 ap P - O 0 o CJ w 8* 00 in t> in 00 in Tt m O bO ft a ft ^ bfl (U T3 £ 2J X X X o ft ft u 118 MADRONO [Vol. 58 Table 3. Minerals (Percent by Volume) in the Rocks. Sites PH are from Pine Hill, El Dorado Co., sites CG are from Cuyamaca-Guataty, San Diego Co. aBlack, opaque minerals, mostly magnetite. bSome alteration of feldspars, apparently to clinozoisite. Site Pyroxene Hornblende Olivine Feldspar Opaquea Other PH4 64 2 10 11 13 Biotite, clinozoisiteb PH5 58 trace 3 24 15 Green spinel PH9 56 1 15 18 10 Biotite, green spinel CGI 29 17 1 52 1 CG2 3 34 4 55 4 CG3 30 12 3 54 1 levels of magnitude. Where the differences are significant at the 99% level of confidence (a < 0.01), A, B, and C are the level designations. Results Pine Hill Preserve Sites Soils at the three sites in the Pine Hill Preserve, two in the Pine Hill unit and one in the Penny Lane unit, are all Alfisols. One site in the Pine Hill unit has a very stony variant of Rescue soil (site PH 5) with a mixed chamise chaparral plant community and the other has a clayey variant of Boomer soil (site PH4) with a black oak/toyon- Lemon ceanothus plant community (Table 2). The site in the Penny Lane unit has a clayey variant of Rescue soil (site PH9) with a chamise- manzanita chaparral plant community. Rock samples representing the soil parent materials of these sites are composed of predominantly clinopyroxene, with lesser amounts of plagioclase feldspar and olivine (Table 3). The compositions of these rocks plotted on Fig. 2 indicate that they are expected to have Ca/Mg ratios about 0.5 to 1.0 mol/mol. Sand separates from the soils showed medium (0.25-0.5 mm) and coarse (0.5- 1 .0 mm) sand modes. Light separates were about 0.2 (20%) of fine sand fractions, comparable to the feldspar estimates in thin-sections (Table 3). About 1/3, or more, of the heavy fine sand separates from sites PH4 and PH5 and about 1/5 to 1/4 from site PH9 were black opaque grains attracted to a hand magnet, suggesting that the black opaque minerals were mostly magnetite. The means of fine sand, very fine sand, and coarse silt from the 0-15 cm depth in soils at the Pine Hill sites were 93, 71, and 33 g/kg and the masses of magnetic grains were 34, 12, and 2 g/kg. Grains with low refractive indices (RI < 1.56) averaged 5, 8, and 1 1% in these size fractions and about 1% of the grains in the coarse silt fraction were isotropic (presumably volcanic glass) and the isotropic grains were subrounded to rounded, in contrast to other grains which were mostly angular to subangular. Most, or practically all, of the grains with low refractive indices could be feldspars and vein quartz from the gabbro parent rock. Although the glass is presumed to be from an aeolian source, its mass is <0.1% of the mass of the surface (0-15 cm) soil. The clayey Rescue variant at site PH9 had vegetation that is common on Rescue soils and the soils at sites PH4 and PH5 had unique vegetation (Table 4). Mature Jeffrey pine trees on the north side of Pine Hill are about 38 m tall, indicating moderate site productivity even with the relatively low precipitation on Pine Hill. There are no Jeffrey pine trees on the Rescue soils. The plants that are local endemics are Pine Hill ceanothus ( Ceanothus roderickii W. Knight), Pine Table 4. Vascular Plants on the Gabbro Soils. Botanical authorities are those given in Hickman (1993). PH9 site: Pinus sabiniana and Quercus xvizlizenii occur on adjacent slopes. Abundance symbols: ++++, 30-100%; +++, 10-30%, ++, 3-10%, +, 1-3%; -, trace; 0, none. PH4 PH5 PH9 A. Pine hill Trees Pinus sabiniana Quercus kelloggii Quercus wislizenii Cercis occidentalis Shrubs Adenostoma fasciculatum Arctostaphylos viscida Ceanothus lemmonii 0 +++ 0 0 +++ ++ + 0 0 0 ++ ++H-T +++ ++ ALEXANDER: GABBRO SOILS AND PLANTS 119 2011] Table 4. Continued PH4 PH5 PH9 Ceanothus roderickii 0 + 0 Heteromeles arbutifolia ++ ++ 0 Quercus durata - 0 0 Rhamnus ilicifolia + — + Rhamnus tomentella 0 ++ + Salvia sonomensis Toxicodendron diversilobum 0 ++ 0 ++++ 0 Eriodictyon californicum 0 - - Forbs Calochortus albus 0 0 + Chlorogalum grandiflorum 0 - - Chlorogalum pomeridianum 0 - - Dichelostemma multiflorum 0 0 + Dichelostemma volubile — 0 0 Fritilaria micrantha - 0 0 Galium californicum ssp. sierrae - 0 0 Galium spp. + + + Geranium molle - 0 0 Sanicula bipinnatifida 0 - - Senecio (Packera) layneae 0 0 - Triteleia bridgesii - 0 0 Wyethia angustifolia - 0 0 Wyethia reticulata — — 0 Grasses Avena fatua - 0 0 Bromus diandrus — 0 — Bromus laevipes + 0 0 Bromus madritensis ssp. rubens — 0 0 Cynosurus echinatus - 0 0 Elymus glaucus + 0 0 Elymus multisetus — 0 0 Fritilaria micrantha - 0 0 Gastridium ventricosum 0 0 - Melica sp. + 0 - Nassella lepida - 0 ++ B. Cuyamaca-Guatay CGI CG2 CG3 Trees Callitropis stephensonii +++ 0 0 Callitropis forbesii 0 0 +++ Shrubs Adenostoma fasciculatum +++ ++ ++ Arctostaphylos glandulosa + +++ - Ceanothus foliosus 0 + 0 Ceanothus greggii var. perplexens - 0 ++++ Cercocarpus betuloides 0 0 - Ericameria parishii +++ ++ 0 Heteromeles arbutifolia + 0 0 Rhamnus crocea — + 0 Quercus berberidifolia ++ 0 ++++ Salvia sonomensis + 0 0 Subshrubs and succulents Eriophyllum confer tifolium ++ 0 0 Trichostema parishii + - 0 Hesperoyucca whipplei - 0 0 Forbs Calystegia collina - + 0 Dichelostemma capitatum + 0 0 Grasses Calamagrostis koeleroides 0 + 0 120 MADRONO [Vol. 58 Hill flannelbush ( Fremontodendron californicum (Torr.) Coville ssp. decumbens (R. M. Lloyd) Munz), El Dorado bedstraw ( Galium californicum Hook. & Arn. ssp. sierrae Dempster & Stebbins), and El Dorado mule-ears ( Wyethia reticulata Greene). All four of these species are present in the vicinity of sites PH4 or PH5 in the Pine Hill unit, but only El Dorado bedstraw and El Dorado mule-ears are present in the Penny Lane unit (G. Hinshaw, U.S. Bureau of Land Man- agement personal communication). Surface soil Ca concentrations are relatively high on the very stony Rescue variant (PH 5) and the Mg relatively high on the clayey Rescue variant (PH9). The Ca/Mg ratios are significantly higher in the very stony Rescue variant than in the other soils, and the organic matter (LOI, Table 5) and exchangeable acidity are higher, also. The relatively high soil organic matter and high CEC (mostly exchangeable Ca and acidity) at site PH5 may be related to a thick O-horizon of predominantly live oak, toyon, and coffeeber- ry leaves. The O-horizon was not analyzed chemically. Aqua regia digestion recovered rela- tively low amounts of Ca for the soil at site PH5, raising doubts about the usefulness of results from the digestion. Aqua regia is an aggressive solvent, but it does not recovery the total amounts of all elements in soils. Peninsular Ranges, Cuyamaca-Guatay Sites Soils at two sites sampled in the Cuyamaca area (CGI and CG2) are Alfisols and the one on Guatay Mountain (CG3) is a Mollisol. Site CGI site is on an extremely stony variant of Las Posas soil with a Cuyamaca cypress/chamise-golden- yarrow plant community; site CG2 is on a yellowish red variant of the dark reddish brown Las Posas soil with an Eastwood manzanita plant community; and site CG3 on a soft, or friable, black soil (a Mollisol) with a Tecate cypress/scrub oak-cupleaf lilac plant community (Table 2). Rock samples representing the soil parent materi- als of these sites are predominantly plagioclase feldspar, with substantial amounts of clinopyrox- ene and hornblende, and only minor olivine (Table 3). The compositions of these rocks plotted on Fig. 3 indicate that they are expected to have Ca/Mg ratios about 1.5 mol/mol. Sand separates from the soils showed fine sand (0.125-0.25 mm) modes. Light separates are about 0.6 (60%) of fine sand fractions, comparable to the feldspar esti- mates in thin-sections (Table 3). About 0.3 (30%) of the heavy fine sand separates were black opaque grains attracted to a hand magnet, suggesting that the black opaque minerals are mostly magnetite. The yellowish red Las Posas variant at site CG2 has vegetation that is common on Las Posas soils and the soils at sites CGI and CG3 have unique vegetation (Table 4). Exchangeable Ca and Mg in the surface and K in the subsoil are relatively high in the Mollisol (site CG3, Table 5). As at site PH5 on Pine Hill, the Mollisol at site CG3 has a relatively thick O- horizon, but the surface soil organic matter content (LOI, Table 5) is higher in the extremely stony Las Posas variant at site CGI that had been burned a year or two prior to sampling. Differences in element recovery by aqua regia digestion, such as relatively low P content in the yellowish Las Posas variant, may not be related to vegetation distributions. Potentially toxic elements (Cr, Co, and Ni) are as high in the soil at site CG2 lacking cypress as they are in the soil at site CG3 with Tecate cypress (Table 6). Although P is lower in the soil at site CG2, that might be a result of the denser vegetation and cypress trees at sites CGI and CG3 recycling more P rather than an indication of lower inherent soil fertility at site CG2. Discussion and Conclusions The chemistry of gabbro soils is sufficiently different from others that some plants that grow on them do not grow on soils of either granitic or ultramafic rocks. In both the Pine Hill area and the Peninsular Ranges, chemical differences between the gabbro soils with common vegetation and those with unique vegetation do not appear to explain the differences in vegetation. The soil differences, other than stoniness, appear to be related to differences in the plant communities on them, rather than the reverse. The hypothesis that the exchangeable CaMg ratios of gabbro soils are important in the distributions of unique species is not verified from the chosen sites. Perhaps the unique plants would grow on all of the gabbro soils sampled, but they have not all been colonized by those plants. The sites lacking unique species are the hottest and driest sites, which may have limited colonization of them. Site PH9 is lower in elevation than sites PH4 and PH5 that have more unique plants, and site GC2 lacking cypress is on a steep south-facing slope at a lower elevation than site GC 1 which is also on a south-facing slope. Some aeolian dust, or loess, may have been added to the coarse silt fractions of the gabbro soils, but these fractions are small. Only 3.3% of the Pine Hill surface soils were coarse silt, not much more than the 1.0% coarse silt at the 15- 30 cm depth. The small amounts of alkali (K and Na > Ca) feldspars and sparse quartz in the coarse silt and other particle-size fractions suggests that the contributions of loess have been minor. Any possible aeolian inputs do not alter the fact that four plants are endemic on the Pine Hill gabbro (Wilson et al. 2010) and that in San Diego Co. Cupressus arizonica Greene ssp. stephensonii (C.B. Wolf) R. M. Beauch (Cuya- 2011] ALEXANDER: GABBRO SOILS AND PLANTS 121 X > Oh • - >> c/3 rj 2 o 15 3 & £ | c/3 5 ^ fN G °° *X .O Oh 172 G - G OG .a ‘5 x i? G £ ^ £ >, £ 5 T3 ’ G G O 00 p $ -2 G ’ a G • g G so <*H G 0 2 <» S ' G G G G 2 ^ g G Su- c/3 O G © C/3 0 qj §•§ 1 s t/3 <2 Ih d .2 °2 II (72 g .2 ,Q G U o ro G (U o 2.2 a.s ^ o II! C*h ^ O X ° 43 •£ ro G O X C72 (D G G o O . s « ^ a £5 0 +_ • 2 G G o a is 1 G O bo D ’c/2 00 , ^ /<_v a ON q 00 ON fN NO 'oo o 1 ^ , o | d i a 1 d ■ To U-? 1 2 d; o O) q q q a d ON a a a NO # i-H t> fN NO (N) ^H NO fN fN da 0 2 2 0 0 do 0 m G" r-H 00^1 fN r- 00 o ON a o r- no r- NO NO NO NO r-' !> NO oa o PT ^ fd (n dr ar G" ^H rT — ^H i — 1 s a a OH d o o o o d d o d ~G 2 NO ca xi 00 o X r- oo in ON3 X q 'd On oo do q u a a a no a a a a - a o a a o ^ r- +2 m NO fN G 00 fG in fN 00 2 w w o « 0 2: do 'd M 2 in no in a f^r G >n NO ON NO a < in ■G- r-H >n NO O 00 m in o G' o a £ /— s s fN OO 00 -H fN -H r- i™h m fN in rG fN G o o a o d o d o o o o o a 0 0 0 0 0 0 di; do do X in no in r"~ G- fN o o ON G a a o a o a o a a a o a a c/T G O / N H— * NO ^ 00 m 00 G 00 in — m 1-H NO ^H NO — NO NO o G G s da N— f N— ^ s-^ Tx ds v-^ O di r- no m OO ON m a5 o CG fN G^ 2 m Tf fN m m r- G 00 m G- NO NO G G 00 G o no o ^ G- 00 s hG Ca fN t-H fG 1-H G ^H fG t-- On X W ca nC ~2 To — m 0 si 0 NO G- NO 2 HCj- cS o CC X NO G d! G^ O oo in m O i-H a ON t-' NO r- O X o ON 00 oo —H (N q 00 q fN q a NO a a a NO NO a a a NO NO a (D sp ds d a sp P sp vp sp vp sp a C/3 0 on a a a a a c/f c/T a a c/T a o C/3 > C/3 > C/3 > C/D > C/3 > C/3 > U X a y a a y a X) u X) a a a y G H a go on on a U U X) X on on on > > > a u > > U 2 > > m fN z 2 G; G- NO m rn G NO a (G 2 m ro m fG m ^ m a CG G a a 0 5 a a C^ a a oi a a a a a 2 a a d >< ^ ^ > in P P m in G in in in >n in >n in a m in a a 2 G G NO G- G1 G a G fG in Dry >n G- NO 2 a G a G no a G a ^ G1 G1 ^ ^ a G a in a a a a oi Pi a 2 a d a a a in a d >< J>h in in a m a in in a in in in in a a in a in a a G 2h < CQ < CQ < a < CQ < a < CP a G" G- >n in ON ON i-H fN fN cn (G X X X X X X o 0 00 0 0 in a a a a a a U U u u u u G G 0 G G G a G >. G u a o — < r<2 OO OO On (N (N OO \D tf O 00 G" fN —i on — i g- o m on oo