UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Patterns of edaphic and phenotypic variation in the Lasthenia californica species complex (Asteraceae) Choe, Gina 2007

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2007-0361.pdf [ 6.04MB ]
Metadata
JSON: 831-1.0100864.json
JSON-LD: 831-1.0100864-ld.json
RDF/XML (Pretty): 831-1.0100864-rdf.xml
RDF/JSON: 831-1.0100864-rdf.json
Turtle: 831-1.0100864-turtle.txt
N-Triples: 831-1.0100864-rdf-ntriples.txt
Original Record: 831-1.0100864-source.json
Full Text
831-1.0100864-fulltext.txt
Citation
831-1.0100864.ris

Full Text

P A T T E R N S O F E D A P H I C A N D P H E N O T Y P I C V A R I A T I O N I N T H E L A S T H E N I A C A L I F O R N I C A SPECIES C O M P L E X ( A S T E R A C E A E ) by Gina Choe A T H E S I S S U B M I T T E D IN P A R T I A L F U L F I L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F S C I E N C E in T H E F A C U L T Y O F G R A D U A T E S T U D I E S (Botany) T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A September 2007 © Gina Choe, 2007 A B S T R A C T Members of the Lasthenia californica species complex show extensive ecological and phenotypic diversity. This thesis investigates patterns of edaphic, physiological, flavonoid, and morphological differentiation within the Lasthenia californica species complex in two distinct studies. In the first study, patterns of ecological and physiological variation along with flavonoid polymorphism and ITS1 species type distribution were characterized. The hypothesis of parallel evolution of edaphically differentiated flavonoid races across lineages of L. californica sensu lato was also examined in this study. We found evidence refuting the predictions of this hypothesis, particularly that flavonoid type and edaphic environments are correlated across the complex. However, when major axes of ecological and physiological variation were characterized within the complex, the greatest amount of edaphic/tissue variation occurred along lines of toxicity, reaffirming prior characterization. In the second study, the basis of the phenotypic differentiation observed between inland and coastal ecotypes of Lasthenia californica sensu stricto was investigated by growing populations of each ecotype in controlled conditions under varying levels of salinity (an important environmental factor differentiating ecotypic environments). Reaction norms of populations were compared under saline conditions to test for evidence of differential fitness between ecotypes (differential salt tolerance) and differences in intrinsic and induced levels of trait expression (e.g. sodium accumulation and succulence). The hypothesis of trade-offs between salt tolerance and growth potentials was also investigated. Under experimental conditions, coastal populations demonstrated uniformly greater salt tolerance than inland populations. We found that most traits showed both intrinsic differentiation and differential responses to salinity, where the extent and direction of trait responses were ecotype specific with little within ecotype differentiation. When comparing growth rate trends between salt tolerant and non-tolerant populations across salinity treatment levels, we found evidence of trade-offs between maximum intrinsic growth rates and investments in salt tolerance. T A B L E O F C O N T E N T S Abstract i i Table of Contents iv List of Tables v List of Figures.. .'• vi i Acknowledgements x Chapter 1: Introduction •• 1 General introduction 1 The Lasthenia californica species complex 3 Chapter 2: Patterns of edaphic, physiological, ITS lineage and flavonoid variation across the Lasthenia californica species complex: re-evaluating the hypothesis of parallel evolution of edaphic races 12 Introduction 12 Materials and Methods 16 Results 20 Discussion 26 Chapter 3: Differentiation in morphological and physiological responses to salinity between coastal and inland populations of Lasthenia californica sensu stricto 53 Introduction 53 Materials and Methods. 58 Results ; 64 Discussion... 70 Chapter 4: Conclusion 88 Significant findings and implications 88 Hypotheses requiring confirmation and possible future studies. 89 Literature Cited 93 I V LIST O F T A B L E S Table 1.1 Table 2.1 Summary of the taxonomic history of the Lasthenia californica species complex from 1966 to present. Occurrences of flavonoid races in the members of the complex are included. The parentheses denote the flavonoid profile most common to the lineage 11 Site ID, ITS species identity, locality, habitat, and flavonoid race determinations from this and previous surveys are summarized. Data from the current study is abbreviated as 2007; data from Rajakaruna and Bohm (1999) and Rajakaruna et al., 2003 surveys are abbreviated as 1999&2003; data from Desrochers and Bohm 1993 is abbreviated as 1993. Sites are roughly ordered geographically, with the northernmost sites listed at the top of the table 38 Table 2.2 Transformations used in statistical analyses. A ) Transformations used in A N O V A and P C analyses for soil variables. B) Transformations used in A N O V A ' s for tissue variables 40 Table 2.3 Table 2.4 Table 2.5 A N O V A results for soil variables. Variables showing significant differences (P < 0.05) between A and C populations are bolded. Means that are significantly higher are bolded as well. Only differences between A and C were tested; mixed populations of both A and C were not included in the A N O V A . Mineral concentrations are reported in ppm; C E C in meq/lOOg 41 Summary of number of A , C , or mixed A and C profile populations present on edaphically extreme sites surveyed in 2005 41 P C A results of soil data. Eigenvalues, % of variance explained, cumulative % of variance explained, and Pearson's correlations (r) to axes are included 42 Table 2.6 Means and standard error of soil variables for serpentine, inland benign pastures and coastal sites. Variables showing significant differences (P < 0.05) are bolded. Significant pair-wise differences (after a Tukey H S D correction) between habitat groupings are denoted with letters where shared letters represent non-significant differences. Mineral concentrations are reported in ppm; C E C in meq/lOOg Table 2.7 Means and standard error of tissue variables for A , C , and mixed A and C populations. Variables showing significant differences (P < 0.05) between A and C populations are bolded and means that are significantly higher are bolded as well. Only differences between A 43 V and C were tested; mixed populations of both A and C were not included in the A N O V A . A l l mineral concentrations are reported in ppm 44 Table 2.8 P C A results of tissue data. Eigenvalues, % of variance explained, cumulative % of variance explained, and Pearson's correlations (r) to axes are included 45 Table 2.9 Means and standard error of tissue variables for serpentine, inland benign pastures and coastal sites. Variables showing significant differences (P < 0.05) are bolded. Significant pair-wise differences (after a Tukey H S D correction) between habitat groupings are denoted with letters. Shared letters represent non-significant differences. Mineral concentrations are reported in ppm; C E C in meq/100 46 Table 2.10 Results for linear regressions of soil and tissue concentrations of P, K , Ca , S, Na, Z n , Fe, Cu , B and N (ppm) 47 Table 3.1 Summary of ecotype, population identification, localities, soil and tissue data from field sites used in this study 79 Table 3.2 Summary of A N O V A results for biomass (g), relative growth rates (In mg days"1), total leaf number, L D M C (g g"1), S L A c m 2 g"1, and succulence (g H 2 0 cm"2) 80 Table 3.3 Summary of A N O V A results for ion traits: Na (ppm), K (ppm), and N a / K , 81 Table 3.4 Values and significance of Pearson product-moment correlations between leaf values of sodium and potassium, sodium and calcium, sodium and magnesium, and sodium and succulence 82 Table 3.5 Summary of A N C O V A results for sodium tissue contents and succulence 82 vi LIST O F F I G U R E S Figure 2.1 P C A ordination of axes 1 and 2 for soil variables, (a) Habitat group symbols correspond to serpentine (A), coastal (X) , inland benign pastures (+), and inland alkaline flats (o). Confidence ellipses represent one standard deviation from group centroids (excluding alkaline flats; N=2). (b) Symbols correspond to uniformly A profile populations (A) , uniformly C profile populations (C), or populations with mixed A and C individuals (M) 48 Figure 2.2 P C A ordination of axes 1 and 2 for tissue variables, (a) Habitat group symbols correspond to serpentine (A), coastal (X) , inland benign pastures (+), and inland alkaline flats (o). Confidence ellipses represent one standard deviation from group centroids (excluding alkaline flats; N=2). (b) Symbols correspond to uniformly A profile populations (A), uniformly C profile populations (C), or populations with mixed A and C individuals (M) 49 Figure 2.3 Bivariate distribution of soil (x) and tissue (y) concentrations of Ca (a) and B (b). Symbols correspond to serpentine sites (S), coastal sites (C), inland benign pastures (P), and inland alkaline flats (A). Linear regressions were fitted for the complex 50 Figure 2.4 Bivariate distribution of log soil (x) and log tissue (y) concentrations of phosphorus (a), sulfur (b), sodium (c), and zinc (d). Symbols correspond to serpentine sites (S), coastal sites (C), inland benign pastures (P), and inland alkaline flats (A). Linear regressions were fitted for the complex 51 Figure 2.5 Bivariate distribution of soil (x) and tissue (y) concentrations of magnesium (a) and potassium (b). Symbols correspond to serpentine sites (S), coastal sites (C), inland benign pastures (P), and inland alkaline flats (A). Confidence ellipses represent one standard deviation from group centroids 52 Figure 3.1 Graphical results for growth traits. Reaction norms with means and standard errors for biomass (g) and relative growth rates (In mg days" ') are shown in figure (a) and (b) respectively. Reaction norms for each population are represented by a unique line described in the legend. Coastal and inland populations are labeled with a C or I, respectively. Control, medium and high stress treatments are denoted with a C , M and H , respectively, (c) Population means and standard errors for R G R values at the control treatment. Significant pair-wise differences between populations are denoted with letters, where shared letters represent non-significant differences. Inland V l l populations and coastal populations are represented by empty bars or shaded bars, respectively 83 Figure 3.2 Reaction norm plots including means and standard errors for (a) leaf number and (b) L D M C (g g"1). Reaction norms for each population are represented by unique lines described in the legend. Coastal and inland populations are labeled with a C or I, respectively. Control, medium and high stress treatments are denoted with a C, M and H , respectively, (c) L D M C (g g"1) population means and standard errors. Significant pair-wise differences between populations are denoted with letters, where shared letters represent non-significant differences. Inland populations and coastal populations are represented by empty bars or shaded bars, respectively 84 Figure 3.3 Reaction norm plots including means and standard errors for (a) S L A (cm 2 g"1) and (b) succulence (g H 2 0 cm"2), Reaction norms for each population are represented by unique lines described in the legend. Coastal and inland populations are labeled with a C or I, respectively. Control, medium and high stress treatments are denoted with a C , M and H , respectively. Population means and standard errors for (c) S L A (cm 2 g"') and (d) and succulence (g H 2 0 cm"2). Pair-wise differences between populations are denoted with letters, where shared letters represent non-significant differences. Inland populations and coastal populations are represented by empty bars or shaded bars, respectively 85 Figure 3.4 Reaction norms with means and standard errors for leaf ion concentrations of (a) sodium (ppm) and (b) potassium (ppm). Reaction norms for each population are represented by a unique line described in the legend. Coastal and inland populations are labeled with a C or I, respectively. Control, medium and high stress treatments are denoted with a C, M and H , respectively, (c) Population mean and standard errors for leaf ion concentrations of sodium (ppm) at control treatment. Significant pair-wise differences between populations are denoted with letters, where shared letters represent non-significant differences. Inland populations and coastal populations are represented by empty bars or shaded bars, respectively 86 viii Figure 3.5 Population means and standard errors for N a / K tissue ratios. Pair-wise differences between populations at each treatment level are denoted with letters, where shared letters represent non-significant differences. Note that Pair-wise tests were performed at each treatment level separately. Population 49, 54, 67, 72 are respectively represented by empty bars, lightly shaded bars, darkly shaded bars, and by black bars..'. 87 Figure 3.6 Bivariate distribution of tissue sodium contents and succulence. Linear slopes are fitted for each population 87 I X A C K N O W L E D G E M E N T S I fondly thank my supervisor, Jeannette Whitton, for her guidance, encouragement and support throughout my academic pursuits at U B C . Thanks also to my committee for their helpfulness: Joerg Bohlmann, Aine Plant (SFU), and Keith Adams. I also wish to thank my lab mates Chris Sears and Chris Lee for their help in the lab and for providing a fun and supportive work environment. Thanks to the staff and researchers at Jasper Ridge Biological Preserve for their hospitality r In particular, thanks to Nona Chiariello for her help. Thanks to Veva Stansel. I 'd like to thank Sarah M c K i m for her comments and Wi l l i am Hay for his grammatical edits. Thanks to my external examiners for their comments: Sally Aitken and Mary Berbee. Thanks to the U B C Herbarium, and the U B C Botany Dept. administration and staff, especially Jin Meng for providing technical support for the growth chambers. A very special thank you to my unfailingly supportive mother, Chan Sook Choe, who has always encouraged me to follow my interests. I gratefully acknowledge the financial assistance provided to me: The University of British Colombia, Department of Botany (Entrance Scholarship) as well from the Natural Sciences and Engineering Research Council ( P G S - M award). C H A P T E R 1 Introduction General introduction Environmental gradients and discontinuities profoundly affect community structure, niche differentiation, species diversity, and population and trait differentiation. Ecologists and evolutionary biologists have long recognized the importance of environmental heterogeneity in promoting and maintaining diversity from the scale of communities to genotypes (Kassen 2002; Jasmin and Kassen 2007; Weeks and Hoffman 1998; Freestone and Inouye 2006; Reynolds et al. 1997; Schulter 1993, 1996; Coyne and Orr 2004; Tilman 1988; Wellborn et al. 1996). Patterns of biological discontinuities associated with environmental heterogeneity remains one of the most important systems of study in ecology and evolution. Edaphic mosaics can produce the most dramatic changes in plant distributions over both large and small spatial scales (Anderson et al. 1999) and provide model systems for the study of differentiation along discontinuous environmental features. Edaphic gradients or discontinuities that include harsh substrates such as serpentine or salinized soils are likely to exert disruptive selection pressures on neighboring populations that inhabit contrasting sites. The hypothesis that this form of edaphically mediated disruptive selection promotes evolutionary diversification and differentiation (Kruckeberg 1984, 1986, 2002; Raven 1964) is well supported empirically by the numerous cases of edaphically differentiated populations (e.g. Berglund et al. 2004; Antfinger 1981; Bradshaw 1952), ecotypes (e.g. Nagy 1995; Sambatti and Rice 2006; 1 Reimann and Breckle 1995), and species (e.g. Kruckeberg 1984; El l is et al. 2006; Robichaux et al. 1990; Baldwin 2005). Parallel edaphic selection regimes on geographically isolated populations have the potential to repeatedly generate entities with shared adaptations. Edaphically derived polyphyletic taxa are not uncommon (Levin 2001; Pepper and Norwood 2001) because individuals growing on harsh substrates usually acquire similar sets of adaptations to tolerate their stressful conditions (e.g. Flowers et al. 1986; Kruckeberg 1984). For example, salt tolerant taxa are commonly reported as showing increased sodium accumulation and compartmentalization as well as succulence (Flowers et al. 1986). The similarity of morphology shown by coastal populations of Lasthenia californica have historically lead to the recognition of the polyphyletic species L. macrantha sensu latd (Chan et al. 2002; Desrochers and Bohm 2003; Rajakaruna 2002). Interestingly, the genus Lasthenia, and members of the Lasthenia californica species complex in particular, have a disproportionately high number of edaphically differentiated taxa and races, somej of which are thought to have arisen in parallel in closely related lineages (Rajakaruna et al. 2003a,b,c). Lasthenia has approximately 21 species (according to Chan et al. 2001), many with narrow ecological ranges. Members of the genus are reported to inhabit coastal bluffs, guano deposits, vernal pools, alkaline flats, deserts, oak woodlands, serpentine outcrops, grasslands, meadows, chaparral and deserts. Almost half of the species of Lasthenia are ecological specialists, endemic to harsh or unusual habitats. A s examples, L. playcarpha and L. maritima are specialists on alkaline flats and guano laden seabird roosting sites (respectively). In contrast, members of the Lasthenia californica complex 2 exhibit a wide range of ecological tolerances. Populations of the Lasthenia californica complex have been found in all of the habitats listed for the genus except for seabird guano deposits. The Lasthenia californica system has complex phenotypic and genetic associations between its members, where morphological diversity is either cryptic or divergent, and is not always congruent with levels of genetic similarity or degrees of reproductive isolation. The complexity of genetic, reproductive, and phenotypic patterns within the system alongside the diversity of edaphically differentiated populations have made the Lasthenia californica complex a rich system for studying patterns and mechanisms of adaptation and differentiation to divergent ecological pressures. The Lasthenia californica complex The Lasthenia californica complex is the most widely distributed taxon in the genus. Members of the complex are found in abundance at low elevations throughout the California Floristic Province. The range extends from southern Oregon to Baja California, between the coast (including the Channel Islands) and the foothills of the Sierra Nevada, and continues eastward into Arizona. Most of this range is Mediterranean in climate, characterized by cycles of mild, moist winters and hot dry summers; however, the coastal range is more moderate, with cooler temperatures in the summer and with rainfall that is more consistent throughout the year. Although this obligate outcrossing herbaceous plant is primarily a spring annual, year-long growth and perenniality have been recorded in coastal populations of the complex. 3 The Lasthenia californica complex shows extensive cytological, morphological, biochemical, and physiological variability. A l l of these characteristics have had a long history of study by various U B C researchers and those at other institutions. Three different groups of researchers have identified and confirmed the existence of diploid (n = 8), tetraploid, and (rarely) hexaploid populations throughout the range of the complex (Ornduff 1966; Desrochers and Bohm 1995; Chan et al. 2002). Interestingly, Desrochers and Bohm (1995) did not observe a geographical pattern to ploidy level variation. L. californica (from here on referred to as L. californica sensu lato) was described by Ornduff in his 1966 monograph as a morphologically variable species (note that the name L. chrysostoma was originally used in the 1966 monograph but was later renamed by Ornduff (1993) as L. californica). Although he attributed some of this variation to phenotypic plasticity, he concluded that this species was composed of several genetically distinct races that followed coastal, inland non-desert, and inland desert boundaries. The coastal race is the most divergent of these, distinguished from the inland races primarily by its succulent morphology. A s an added complication, succulence can sometimes also be found to a limited degree in some desert alkali sites. Desrochers and Bohm 1993 agreed with the inland versus coastal distinction among populations of Lasthenia californica sensu lato but found that the extent of overlap in morphological variation across the inland populations did not warrant recognition of inland desert versus inland non-desert races. Ornduff (1966) had also recognized a distinct coastal species, Lasthenia macrantha (referred to here as Lasthenia macrantha sensu lato) which has been described by other researchers as "exceedingly difficult to distinguish from L. californica [sensu 4 lato]" p. 1103 Chan et al. 2002. L. macrantha sensu lato was primarily distributed along the northern Californian and southern Oregon coast. This taxon was composed of three subspecies, subsp. macrantha, subsp. bakeri, and subsp. prisca. Typically L. macrantha sensu lato is described as an annual to perennial succulent herb with wide, strapped shaped leathery leaves. Apart from perenniality, there are no reproductive or vegetative characters that easily distinguish L. macrantha sensu lato from the coastal race of L. californica sensu lato. Even the life history traits used to distinguish the two taxa gradate. When grown in common conditions in a growth chamber, populations of coastal L. californica sensu lato have substantially longer generation times in comparison to their inland counterparts (e.g. L. californica sensu lato individuals from coastal bluffs in Salt Point State Park: G . Choe, personal observation). Furthermore, L. macrantha can behave as an annual, depending on water availability (Ornduff 1966; Chan 2001, Chan et al. 2002). The morphological overlap between these two formerly distinct species is reflected in their high degree of interfertility. Ornduff (1966) conducted inter- and intraspecific crosses and found that populations of L. macrantha sensu lato readily produced viable offspring in crosses with populations of inland L. californica sensu lato. He also found that interspecific crosses tended to be as or more successful than intraspecific L. californica sensu lato crosses and that L. californica sensu lato populations with significant geographical separation had very low crossability (especially between northern and southern Californian populations). The magnitude of variation within the Lasthenia californica complex has resulted in a complex taxonomic history. Throughout this history, its various members have been assigned to different species and subspecies across three genera and seven species. 5 Ornduff s (1966) revision of Lasthenia californica had combined previously distinct taxa into this one species and this circumscription was unchallenged until recently when phylogenetic analysis revealed that the species outlined by Ornduff actually spanned two separate phylogenetic lineages: L. gracilis and L. californica subsp. californica (Chan et al. 2002, Desrochers and Dodge 2003). Lasthenia californica sensu lato, therefore has now been recircumscribed to reflect phylogenetic relationships (Chan et al. 2002, Chan 2001). The two cryptic phylogenetic species now recognized generally follow geographic lines: L. californica is distributed predominately in southern Oregon and northern California, while L. gracilis is found primarily in southern California, Arizona and Baja California (Chan et al. 2002; Desrochers and Dodge 2003). The taxonomic changes between Ornduff's (1966, 1993) and Chan's (2001) circumscriptions are summarized in Table 1.1. The new circumscription of L. californica sensu stricto now represents a monophyletic lineage that not surprisingly includes subspecies of L. macrantha sensu lato (=L. californica subsp. macrantha and L. californica subsp. bakeri) and the northern populations of L. californica sensu lato (=L. californica subsp. californica). The third subspecies of L. macrantha sensu lato, subspecies prisca (a coastal Southern Oregon endemic) has been elevated to the status of species (=L. orndujfii; Table 1.1). Both D N A sequence (Chan et al. 2001, 2002; Desrochers and Dodge 2003) and R A P D marker data (Rajakaruna 2002) identify L. macrantha sensu lato as a polyphyletic lineage. Populations with the 'macrantha' phenotype nest within geographic groups that include neighboring inland populations of L. californica subsp. californica. 6 The biochemical complexity of this system has also had a long history of study and predates the phylogenetic research. The existence of flavonoid polymorphisms was first noted in two cryptic species of L. californica sensu lato distributed in parapatry along a serpentine grassland at Jasper Ridge Biological Preserve. Four pigment profiles were determined, of which two predominate, profiles A and C / B o h m 1987; Bohm et al. 1989). Profile A was characterized as the more complex profile of the two and is differentiated from C by the presence of flavonol diglycoside sulfates and a flavanone eriodictyol 7-O-glucoside. Bohm et al. (1989) also determined that these flavonoid profiles were genetically determined as opposed to environmentally induced. He grew offspring of both A and C plants and found that A individuals produced exclusively A progeny and that C plants produced progeny that were virtually all C (Bohm et al. 1989). Desrochers and Bohm (1993, 1995) later found that these A and C profiles are distributed throughout the range of the complex. They also determined that individuals with flavonoid profiles A and C generally grouped into southern (recently recognized by Chan 2001 as L. gracilis) and northern clades (recently recognized by Chan 2001 as L. californica subsp. californica), but were not restricted to these geographical zones. Numerous polymorphic A / C populations throughout the range of the complex were also recorded, but they found that these populations were more common in the central region of California, where northern and southern clades broadly overlap. Table 1.1 summarizes A / C flavonoid occurrences in each phylogenetic lineage. Subsequent ecological and flavonoid surveys of this complex conducted by Rajakaruna and colleagues (Rajakaruna and Bohm 1999; Rajakaruna et al. 2003a) observed a strong correlation between flavonoid profile and soil type. Populations 7 growing on salty soils typically show flavonoid profile A (referred to as race A ) ; while populations growing in drier, but ionically benign soils tended to show flavonoid profile C (referred to as race C ; Rajakaruna and Bohm 1999). Rajakaruna and Bohm (1999) also found that race A populations growing on salty soils tended to accumulate certain ions such as N a + and M g 2 + t o a greater degree than race C populations. However, unlike Desrochers and Bohm (1993), Rajakaruna and Bohm (1999) found that populations were fixed for either A or C flavonoid profiles. These observations from field studies suggested that the two flavonoid races were also distinct edaphic races. Rajakaruna and colleagues (1999, 2003a,b,c) hypothesized that race A individuals had evolved greater stress tolerance and that race C individuals had evolved greater drought tolerance. To test the hypothesized differentiation of water stress tolerance between races they investigated differential responses of A and C ' s to water stress in a system where A and C 's grow in parapatry along a gradient of increasing water stress (Rajakaruna et al. 2003b). They grew A and C individuals in controlled conditions with differing levels of water availability (Rajakaruna et al. 2003b). A s predicted, they found that that C 's were less affected by the negative impacts of drought stress (as indicated by number of flower heads produced). These data supported their hypothesis that race C individuals demonstrate greater drought tolerance. To test the ionic stress tolerance component of the edaphic/flavonoid race hypothesis, Rajakaruna and colleagues (Rajakaruna et al. 2003c) grew two populations of each race from each phylogenetic lineage (four populations in total) in controlled hydroponic conditions to determine whether the ion accumulation differences observed in the field reflected genetic differences. Specific ion uptake, accumulation, and tolerance 8 indices were measured for N a + and M g 2 + . They found that race A , regardless of species membership, had higher uptake rates, accumulation, and tolerance to these ions (Rajakaruna et al. 2003c). Rajakaruna and colleagues also proposed a potential mechanistic link between flavonoid chemistry and ionic stress resistance. They had realized that sulfated flavonoids (one of the major flavonoid fractions distinguishing A from C) were found in greater levels in salt marsh species, and from this they hypothesized that ionic stress tolerance may be conferred by or linked to the presence of sulfated flavonoids. Rajakaruna et al. (2003a) determined phylogenetic species identities and flavonoid profiles of populations in a complex wide survey and confirmed that while the majority of L. gracilis populations were A , and that the majority of L. californica subsp. californica populations were C , both A and C ' s indeed occurred in both of the recently recognized phylogenetic taxa. In the survey they published in 2003, of the 16 populations of L. gracilis that they sampled, 13 were A , and 3 were C ; of the 17 populations of L. californica subsp. californica sampled it was determined that 11 were race C and 6 were race A (Rajakaruna et al. 2003a). Also , note that they found only fixed A or C populations. Field and experimental data had led Rajakaruna and colleagues (Rajakaruna et al., 2003a,b,c) to conclude that these two races, distinct in edaphic tolerances, ion physiology, drought tolerance, and flavonoid chemistry had evolved in parallel in both phylogenetic lineages. Objectives: This thesis examines patterns of edaphic, physiological, and morphological 9 differentiation within and among members of the Lasthenia californica species complex, in two distinct studies: Chapter 2 has two primary objectives: 1) To describe patterns of edaphic and physiological variation in populations across the species complex, as indicated by soils and tissue data. 2) To re-examine the flavonoid/edaphic race hypothesis proposed by Rajakaruna et al. (2003a,b,c) by comparing changes in flavonoid race distribution and the edaphic associations of these races between various surveys published between 1993 and this current survey. Chapter 3 investigates the basis of the phenotypic differentiation observed between coastal and inland ecotypes (=subspecies) of L. californica sensu stricto. The hypothesis of adaptive ecotypic differentiation in salt tolerance and associated traits was examined by comparing norms of reaction between populations of each ecotype grown in common conditions with varying levels of salt. Upon detection of differential reactions to salt stress, evidence of trade-offs between salt tolerance and growth potentials was then investigated. 10 Table 1.1: Summary of the taxonomic history of the Lasthenia californica species complex from 1966 to present. Occurrences of flavonoid races in the members of the complex are included. The parentheses denote the flavonoid profile most common to the lineage. Ornduff's circumscriptions Presence of flavonoid races' Chan's circumscriptions L. californica (A)and C L. gracilis2 A a n d ( C ) L. californica subsp. californica L. macrantha subsp. macrantha A and (C) 3 L. californica subsp. macrantha L. macrantha subsp. bakeri N . A . L. californica subsp. bakeri L. macrantha subsp. prisca N . A . ' L. ornduffii Notes: 1) Races B and D are not included in this table. 2) L. gracilis was previously recognized as a part of L. californica sensu Ornduff (1966, 1993). 3) Determined in the current study. 11 C H A P T E R 2 Patterns of edaphic, physiological, ITS lineage and flavonoid variation across the Lasthenia californica species complex: re-evaluating the hypothesis of parallel evolution of edaphic races. I N T R O D U C T I O N The distribution of plant species is influenced by a combination of climatic, biotic and edaphic variables. O f these factors, soil characteristics can produce the most dramatic changes in plant distributions over small spatial scales (Anderson et al. 1999). Substrates that produce the most noticeable discontinuities in plant distributions are typically those that are unfavorable to the majority of plant species; edaphic mosaics incorporating substrates such as ultramafic rock, gypsum, limestone, and alkali flats have the potential to promote and maintain patches of divergently adapted ecotypes by means of disruptive selection. The suite of morphological and physiological adaptations that differentiate plants inhabiting these harsh soils suggests that edaphically mediated selection can be an important source of evolutionary diversification. Numerous instances of edaphically driven differentiation have been reported (Ellis et al. 2006; Robichaux et al. 1990; Kruckeberg 1986; Kruckeberg 2002) with particularly interesting examples from serpentine habitats (Sambatti and Rice 2006; Gardner and MacNair 2000; Gottlieb et al. 1985; Baldwin 2005). Ultramafic soils occupy less than 1% of the land surface of the earth (Brooks 1987) but are associated with disproportionately high levels of endemic species (Kruckeberg 1984, 2002; Brooks 1987). These soils are particularly extreme because they impose both chemical (harsh Mg/Ca ratios, low macronutrient concentrations, and high heavy metal concentrations) and physical stressors (coarsely textured with low organic matter; Brooks 1987). 12 Tolerant taxa often show similar adaptive responses; for instance, several serpentine tolerant plants hyperaccumulate heavy metals (Reeves et al. 1999; Boyd and Martens 1998). Tolerances to toxic substrates such as serpentine and mine tailings have been reported as occurring repeatedly between genetically isolated populations (Berglund et al. 2004; Al -Hiya ly et al. 1988; W u et al. 1975; Gregory and Bradshaw 1965). For instance, Schat et al. (1999) showed that populations of Silene vulgaris from geographically isolated sites across Europe evolved heavy metal tolerance independently and in parallel. Another example of recurrent evolution of stress tolerant races may be found in the Lasthenia californica complex (Rajakaruna et al. 2003a,c; Rajakaruna 2002). Two races with distinct flavonoid and edaphic tolerances are hypothesized to have evolved in parallel in each of the two phylogenetic lineages that compose Lasthenia californica sensu lato. Exploring the strength of this phenomenon in the Lasthenia californica complex is one of the major objectives of this study. Members of Lasthenia californica grow in a wide range of habitats including serpentine outcrops, chaparrals, grazed and ungrazed pastures, grasslands, oak woodlands, coastal bluffs, saline and alkaline flats as well as vernal pools. This vast ecological diversity is mirrored by extensive morphological, cytological and biochemical complexity. Not surprisingly, the taxonomic history of the system has been equally complex. Prior to Ornduff's monograph (1966), members of L. californica sensu lato were assigned to different species and subspecies across three genera. Ornduff's (1966) revision of Lasthenia californica combined previously separated taxa into one species 13 and this circumscription was unchallenged until recently. Phylogenetic analysis revealed that the species outlined by Ornduff actually spans two separate phylogenetic lineages: L. gracilis and L. californica subsp. californica (Chan et al. 2002, Desrochers and Dodge 2003). L. californica sensu lato therefore has been split to reflect phylogenetic lineages (Chan et al. 2002; Chan 2001). These two cryptic phylogenetic species follow geographic lines: L. californica is distributed predominately in southern Oregon and northern California while L. gracilis is found primarily in southern California, Arizona and Baja California (Chan et al. 2002, Desrochers and Dodge 2003). O f noteworthy mention, this newest circumscription of L. californica now includes subsp. macrantha and subsp. bakeri, which were considered distinct species prior to Chan's phylogenetic work (Chan et al 2002; Chan et al. 2001). The biochemical complexity of this system has had a long history of study, and predates the phylogenetic research done on this system. The existence of flavonoid polymorphisms was first noted in L. californica sensu lato into two populations distributed in parapatry along a serpentine grassland at Jasper Ridge Biological Preserve. Two main pigment types, A and C, were noted (Bohm 1987; Bohm et al. 1989); where A was characterized as the more complex profile of the two. Profile A is differentiated from C by the presence of flavonol diglycoside sulfates and a flavanone eriodictyol 7-0-glucoside. Desrochers and Bohm (1993, 1995) later found flavonoid profiles A and C were distributed throughout the range of the complex. They also determined that flavonoid profiles A and C generally grouped into southern (L. gracilis) and northern clades (L. californica subsp. californica), but were not restricted to these geographical zones. Subsequent ecological and flavonoid surveys of this complex conducted by 14 Rajakaruna and colleagues (Rajakaruna and Bohm 1999; Rajakaruna et al. 2003a) observed a strong correlation between flavonoid profile and soil type. Populations growing on salty soils typically show flavonoid profile A (referred to as race A ) ; while populations growing in dry, but ionically benign soils have flavonoid profile C (referred to as race C ; Rajakaruna and Bohm 1999; Rajakaruna et al. 2003a). Rajakaruna and Bohm (1999) found that race A populations growing on salty soils tended to accumulate certain ions such as N a + and M g 2 + to a greater degree than race C populations. They also confirmed that race A and C populations occurred in both of the recently recognized phylogenetic taxa (L. gracilis and L. californica subsp. californica) that once comprised L. californica sensu lato (Rajakaruna et al. 2003a). The observations made in the field suggested that the two flavonoid races were also distinct edaphic races, a hypothesis that I wi l l refer to from this point onwards as the flavonoid=edaphic race hypothesis. To test this hypothesis, Rajakaruna and colleagues grew two populations of each race in each phylogenetic lineage (four populations in total) in controlled hydroponic conditions to determine whether the ion accumulation differences observed in the field reflected genetic differences (Rajakaruna et al. 2003c). Specific ion uptake, accumulation, and tolerance indices were measured for N a + and M g 2 + . They found that race A , regardless of species membership, had higher uptake rates, accumulation, and tolerance to these ions (Rajakaruna et al. 2003c). Both field and experimental data led Rajakaruna and colleagues (Rajakaruna et al. 2003a,c) to conclude that these two races, distinct in edaphic tolerances, ion physiology, and flavonoid chemistry had evolved in parallel in both phylogenetic taxa. 15 In this study we re-examine the flavonoid=edaphic race hypothesis with a more in-depth sampling of the complex, increasing both the number of populations and the numbers of individual flavonoid profiles per population sampled. We also describe changes in flavonoid race distribution and frequency over time by comparing surveys published in 1993 by Desrochers and Bohm (this survey wi l l be referred to as "1993"), in 1999 and 2003 by Rajakaruna and Bohm (1999) and Rajakaruna et al. 2003a (which wil l be referred to as "1999&2003") and the current work from field surveys conducted in 2005 (referred to as r"2007"). The final objective of this study is to characterize the major axes of ecological and physiological variation found across the species complex. M A T E R I A L S A N D M E T H O D S Field sites Flower heads, plant tissue and soil samples were collected from sites previously surveyed by Rajakaruna and Bohm (1999), Rajakaruna et al. 2003, and Desrochers and Bohm (1993), as well as additional sites not previously sampled by those researchers. Although we attempted to relocate and resample all sites, this was not always possible. Locality and habitat information are summarized in Table 2.1. Flavonoid identification Flavonoid profiles were determined as described in Desrochers and Bohm (1993) except that Machery-Nagel T L C plates were used instead of "home-made" plates (p 449, Desrochers and Bohm 1993). Ray florets were soaked in M e O H for several days. The extracts were spotted on 20X20cm Polyamide 6 thin layer chromatography plates (0.1 16 mm Polyamide T L C 6 U V 2 5 4 ; Machery-Nagel, Gutenberg, Germany). Plates were developed one-dimensionally in a water, n-butanol, acetone, and dioxane (70:15:10:5) solution. Plates were air dried, then sprayed with aminoethyl-diphenylborate, visualized under U V light (Alpha-Imager 1200 gel documentation system; Alphalnanotech Corporation, C A , U S A ) , and scored. A total of 670 individual flower heads was tested for flavonoid profiles. 15 to 30 individuals per population were initially sampled. When initial surveys showed fixed flavonoid profiles, no further sampling for these populations was done. Populations showing mixtures of race profiles, or differences between surveys, were sampled in greater depth. In addition, flowers from Desrochers' and Rajakaruna's herbarium vouchers with known flavonoid status were re-sampled in this study to check for consistency in flavonoid profile interpretation. Questionable samples were re-run twice and excluded if unreliable. ITS species identity Phylogenetic species L. gracilis and L. californica were identified for populations not previously determined by Rajakaruna et al. 2003a. These two phylogenetic lineages are reliably separated by a l l b p indel in the r D N A internal transcribed spacerl (ITS1) region. ITS1 size differences were determined following Rajakaruna et al. 2003 except that D N A was extracted with a modified C T A B (Doyle and Doyle, 1987) extraction using QIAquick purification columns (Qiagen, Mississauga, Ontario, Canada) to purify D N A extracts. The ITS1 region was amplified using primers ITS2 and ITS5 (White et al. i990). 25 pi reactions of 10 ng template D N A , 30 m M Tris -Hcl , 50 pM KC1, 2 m M 17 M g C l 2 , 0-1 m M each dNTPs, 10 pmoles primer, and 1.5 units D N A polymerase were amplified in a PT-100 thermal cycler (MJ Research, Waltham, M A , U S A ) with an initial 3 minute denaturation at 94°C, followed by 35 cycles of (1 minute 94°C, 1 minute 50°C, 1 minute 72°G), and a final extension of 7 minutes at 72°C. 2 ]A aliquots of each P C R sample were separated by electrophoresis in 2% SeaPlaque agarose ( B M A , Rockland, M E , U S A ) with 0.5X T B E buffer. The gels were then visualized by U V light after an ethidium bromide stain and scored against samples of known ITS size as standards. Soil and tissue analyses Soil samples were collected from rooting zone depths (5-15cm) using a hand trowel. Subsamples were taken haphazardly within each site and pooled into one sample per population. A l l soil samples were air-dried in open plastic bags then rocks and gravel were removed. Major mineral concentrations of soil samples were determined by ICP spectrometry following a Mehlich III extraction. Heavy metals were extracted by HN03:HC1 total digestions. Extractable phosphorus and NO 3 " concentrations were determined following Bray PI and colorometric C a C l 2 Cd reduction procedures, respectively. Organic material concentrations were determined by loss on ignition at 360°C. P H was determined in 1:1 soil to H 2 0 mixtures. Individuals collected within a population were pooled into one representative tissue sample. These samples were first air-dried at room temperature, and processed by removing adhering environmental particles and by separating organ types (roots, shoots, flower heads). Shoot tissue was analyzed for mineral concentrations by ICP 18 spectrometric analysis after a total acid digestion. A l l soil and tissue mineral quantification were conducted at A & L laboratories inc., London, Ontario, Canada. Data analysis Flavonoid profiles were summarized as simple tallies and are shown in Table 2.1. Univariate A N O V A s were used to determine whether soil and plant tissue characteristics between A and C populations differed significantly for any single mineral. The residuals of these analyses were inspected for normality and homogeneity of variance. Variables not adhering to the assumptions of A N O V A were transformed. Transformed variables are listed in Table 2.2. Note that populations with mixtures of A and C were excluded from the analysis. We decided that excluding these populations would be more prudent than including these points because inclusion would require points to be entered twice in both A and C categories (resulting in pseudoreplication). A n important caveat is that differences between groups using tests excluding mixed populations may exaggerate biological differences between races. Multivariate patterns in soils and tissue data were explored with principle component analyses (PCA) . P C A is an eigenvector-based analysis that summarizes data from multiple variables by reorganizing their variance components into new orthogonal axes. The axes are extracted so that the first axis explains the greatest amount of the variance in the original variables, the second axis explains much of the remaining variance, and so on. The assumptions of P C A (linear association between variables and multivariate normality) were checked and deviations were remedied by transformations. For soil analyses, transformations used were those listed in Table 2.2a. For plant tissue 19 analyses, variables were square root transformed. Note that strict adherence to the assumptions of P C A is not as critical as these are only conditions that improve the explanatory power of the analysis (Tabachnick and Fidell 2001). P C analyses were performed using correlation matrices to standardize the distribution of variance between variables. From the P C A ordinations, we confirmed the distinctiveness of a priori ecological groupings of coastal, serpentine, alkaline flat, and inland grasslands/pastures: We then determined whether there were significant differences between ecological groupings for soil and tissue variables using univariate A N O V A s . Pair-wise differences between groups were determined for significant A N O V A tests. Family-wise type I error inflation rates were controlled for each pairwise analysis by a Bonferroni adjustment. Relationships between soil and tissue characteristics were explored with linear regressions. Residuals were inspected for normality. Significant regressions from misleading relationships (driven by the presence of numerous outliers) were not included. Single outliers with extraordinary leverage unmanageable by transformations were also excluded from a few analyses. R E S U L T S Flavonoid and ITS analyses The results of flavonoid analyses (Table 2.1) show that the distribution of A and C profiles usually adhered to geographical patterns. The A profile was most common in southern California, while C profile individuals were found mostly in northern California and southern Oregon. These results are largely consistent with the patterns described by 20 Desrochers and Bohm (summarized under the heading "1993" in Table 2.1) and Rajakaruna and Bohm (summarized under the heading "1999&2003" in Table 2.1). We confirmed that flavonoid profiles A and C were most prevalent but not restricted to ITS species L. gracilis and L. californica, respectively. Flavonoid composition in central California populations showed the most variation between surveys. Whereas Rajakaruna and colleagues (1999&2003) reported the coastal bluff site SPS (Table 2.1) as homogenously A , we found that this same population consisted mostly of C ' s with only a few A individuals present. The 1993 survey also found a mixed population but with a greater proportion of A ' s to C 's than was shown by our survey. A similar trend was observed at Mt . Tamalpais (21-41, Table 2.1). A t this serpentine site, mixtures of mostly A and some C ' s were reported in 1993 and this was also the case in our survey. Only individuals with A profiles were reported at this site by Rajakaruna and colleagues (1999&2003). In contrast, similarities between 1999&2003 and the current survey (2007) were found in the pasture lands of interior central California. Both 1999&2003 and 2007 survey authors found homogenous populations of C ' s where as Desrochers and Bohm (1993) reported mixed populations of A and C 's (e.g. site 26, Table 2.1). A n alkali flat in Byron (22, Table 2.1) showed a striking difference between flavonoid composition between the 1999&2003 and 2007 surveys. While Rajakaruna (1999&2003) reported a uniformly A population, this same site was found to be completely C in our survey. The populations that changed most in flavonoid composition between the current survey and 1999&2003 were those central California populations inhabiting the most extreme habitats, such as alkali flats and serpentine sites. 21 Soil analyses Race A and C differences Univariate A N O V A ' s indicate that organic matter, p H , P, N a / K , K , Na, S, Fe, and A l concentrations are significantly different between C and A sites (Table 2.3). In comparison to Rajakaruna and Bohm (1999), we found greater numbers of significantly different soil variables between A and C sites in our survey. A n important difference between years is that sodium and sulfur levels are significantly higher in C than A sites in our survey. After verifying increased toxicity of soils found in serpentine, alkali flat, and coastal areas, we tallied the number of A , C , and mixed flavonoid race populations observed in edaphically harsh habitats in our sample year. A s shown in Table 2.4, no homogenously A populations grew on salinized soils. Also , while some fixed A populations were found on serpentine soils, both C and mixed populations were well represented in these habitats as well. Patterns of edaphic diversity over the complex Much of the soil chemistry variation found across the Lasthenia californica complex follows habitat differences. Principal component analyses ( P C A ) show that over half of the variation between sites lies along axes that can be interpreted as serpentine and coastal. Table 2.5 describes the correlations found between the first two P C axes and the original soil variables. The first axis is correlated to variables contributing to the "serpentine syndrome": concentrations of magnesium and heavy 22 metals, chromium, cobalt and nickel. A x i s 2 represents a gradient of saline factors, indicated by the strong correlation to Na and S. P C axis 3 does not account for much of the original variation, but showed the greatest amount of separation between pasture sites (data not shown), suggesting that it represents an axis of ionically benign inland pasture site variation. There are no strong correlations between soil characteristics and this axis (Table 2.5). A P C A ordination of axes 1 and 2 (Figure 2.1a) shows that sites from similar habitats cluster into discrete groupings: serpentine sites cluster in the upper right quadrant, the saline sites in the lower regions, and inland pastures in the upper left quadrant. When populations with profile A , C and mixed A and C populations were used as symbols to visualize the distribution of flavonoid races in this ordination, it then appeared that much of the variation within A and C groupings lie on axes 1 and 2, respectively (Figure 2.1b). Figure 2.1b also makes it apparent that the edaphic extremes (increasingly positive x axis values and increasingly negative y axis values) are not associated with profile A , as predicted under the flavonoid=edaphic race hypothesis. Univariate A N O V A ' s on the following habitat groupings: coastal, inland pastures, and serpentine, corroborate the differences in soil characteristics illustrated by the P C A (Table 2.6). Coastal habitats are significantly higher in sodium and sulfur. Serpentine sites are significantly higher in heavy metal concentrations (nickel, chromium and cobalt), and magnesium (Table 2.6). Inland pastures and grasslands are typically lower in toxic minerals and higher in beneficial elements like potassium (marginally significant), phosphorus and calcium than the edaphically extreme habitats (Table 2.6). Contrary to the literature, nutrient concentrations of potassium, calcium, nitrogen and organic matter 23 content were not significantly lower in serpentine areas (Table 2.6). Plant tissue analyses Differences between A and C Univariate A N O V A s found few significant differences in tissue variables between A and C populations (Table 2.7). The only significant differences found between A and C populations were for phosphorus, sodium, and manganese. Differences in sodium between A and C most likely reflect the greater number of coastal C representatives present in the data set than coastal A representatives. Patterns of tissue differentiation between populations Similar to the soil trends, differences in tissue characteristics between populations of the Lasthenia californica complex follow habitat differences. P C analyses suggest that tissue ion characteristics of populations have the greatest variation along lines of coastal and serpentine gradients. P C axis 1 represents a coastal gradient and is strongly correlated to sodium concentrations in tissues (Table 2.8). P C axis 2 is strongly correlated to the serpentine variable nickel, but not to other serpentine variables such as magnesium (Table 2.8). Magnesium shows a split loading between axis 1 and 2, and thus is not strongly correlated to either axis (Table 2.8). Despite this, P C axis 2 separates populations based on serpentine and non-serpentine characteristics (Figure 2.2a). A P C A ordination of axes 1 and 2 (Figure 2.2a) clearly show that populations cluster based on habitat origin. When flavonoid information was added to the P C A ordination, it became apparent that much of the variation within A and C groupings lay 24 on separate axes (Figure 2.2b). Tissue characteristics suggesting increased ionic stress tolerance (increasingly positive P C 1 values) are not associated with .profile A , as earlier hypothesized by Rajakaruna and colleagues (Rajakaruna and Bohm 1999; Rajakaruna et al. 2003a,c). Univariate A N O V A s using habitat types as groups found significant differences between habitats for many of the elements analyzed (Table 2.9). Coastal sites showed significantly higher levels of sodium tissue accumulation than in other habitat groups (pastures and serpentine; see Table 2.9). Also sulfur levels were higher in coastal groups (Table 2.9). Interestingly, plants inhabiting coastal sites are significantly higher in most other ions, such as copper, zinc, aluminum, and manganese (Table 2.9). Plants in serpentine habitats are differentiated by their significantly higher tissue levels of magnesium and nickel (Table 2.9). Tissues of plants from inland pastures are higher in macronutrients such as calcium and potassium (Table 2.9). Relationships between edaphic and tissue characteristics across the complex Relationships between soil and tissue ion concentrations across the Lasthenia californica system are complex and rarely linear without transformation. Exceptions to this trend include the relationships between soil and tissue levels of calcium and boron (Table 2.10 and Figure 2.3a,b). A s shown in Table 2.10, many scatter plots were cloud-like, lacking any obvious relationship between soil and tissue element concentrations (e.g. potassium, iron, copper, nitrogen). The relationships between tissue and soil concentrations of other minerals such as phosphorus, sulfur, sodium and zinc are moderately or weakly defined by a linear relationship on logarithmic scales (Table 2.10; 25 Figure 2.4a,b,c,d). Relationships between tissue and soil concentrations of cations across the L. californica complex were not always homogenous between habitat groupings. This caused difficulties when modeling the relationship between the two axes. For example, the relationship between magnesium levels in tissues and soils seemed to be split between coastal groupings and the other habitat types (Figure 2.5a). This finding is consistent with the split loading of magnesium in the P C analysis of tissues. Although it would have been more accurate to model the soil-tissue relationship of certain ions separately for each habitat type, all groups other than "inland pastures" lacked sufficient N . Besides a few elements such as potassium, seemingly random associations (in regards to linear correlations) between soil and tissue levels showed discernable patterns of differentiation when grouped by habitat (Figure 2.5b). D I S C U S S I O N Re-evaluating the flavonoid=edaphic race hypothesis. The flavonoid=edaphic race hypothesis proposes that salt tolerance is partly conferred by or is at least linked to the presence of sulfated flavonoids. Individuals with flavonoid profile A are differentiated from profile C in part by the presence of sulfated flavonoids, and these A ' s are thought to be salt tolerant whereas individuals with flavonoid profile C are thought to lack salt tolerance. If this is true, one would expect that ionically extreme habitats across the range of the complex would be occupied most often by race A individuals. We have shown that this is not the case in the current survey by the following results: simple tallies show that the number of edaphically extreme sites 26 colonized by C is higher than the number of sites inhabited by A ; soil A N O V A s indicate significantly higher levels of sodium in C than A sites; P C A ordinations show that sites on the ionically harsher ends of P C 1 and P C 2 are not associated with profile A populations. These data do not support predictions made under the flavonoid=edaphic race hypothesis; Another clear refutation of the flavonoid=edaphic race hypothesis is shown by the fact that individuals with profile C increased in frequency in the majority of edaphically extreme sites initially reported as polymorphic in 1993. Under the predictions of the hypothesis, we should find that selection increases the frequency of A individuals over time in edaphically extreme sites with initial A / C polymorphism. Instead, we found that the two serpentine sites on Mt . Tamalpais (21-41,42), and the coastal bluff site (SPS) in Sonoma Co. increased substantially in C ' s relative to A ' s between 1993 and our current study. Interestingly, the frequencies of A and C changed even more dramatically between the Rajakaruna and colleagues surveys of 1999&2003 and this current survey. In the 1999&2003 surveys, the authors report fixed A populations on the Mt. Tamalpais site and the coastal bluff site SPS. The geographic location of the occurrence of A / C frequency shifts also weakens the flavonoid=edaphic race hypothesis. We found that only populations in central California (the interface between the mostly southern A and northern C) shifted in A / C frequency between 1999&2003 and our survey. This suggests that gene flow affects flavonoid frequency as much as or more than selection for flavonoid/ionic tolerance trait complexes. If individuals with favorable flavonoid/ionic tolerance trait complexes were sufficiently fitter than maladapted migrants in ionically harsh environments, selection 27 should sustain increases in A frequency despite gene flow to those sites. The marked increases of C ' s in polymorphic populations in central California inhabiting ionically extreme sites as well as the complete shift to all C 's from all A ' s at the alkali site at Byron (also in central California) suggest that this is not the case. In light of recent evidence suggesting that levels of introgression and gene flow are thought to occur at higher rates and distances than traditionally expected (Ellstrand 2003; Morjan and Rieseberg 2004), the spread and maintenance of flavonoid polymorphisms by gene flow is a plausible hypothesis. Although the above results provide strong evidence against the flavonoid=edaphic race hypothesis, differences between surveys may also be due to differences in methodology. For example, differences in sampling procedures within sites can produce apparent shifts in A / C frequencies i f between years the different researchers collect from different locations within sites. This would then overlook potential within-site covariation in flavonoid/substrate patterns. Another obvious methodological difference that may contribute to the differences found between years lies in the number of individuals sampled. Surveys in 1999&2003 only tested 3-6 individuals per population compared to the larger number of individuals tested in 1993 and in this survey. Because larger sample sizes are more likely to incorporate low levels of flavonoid diversity within populations, we believe that the most obvious explanation for the result that only the recent and 1993 surveys had reported mixed populations (with the exception of Jasper ridge) is due to the depth of sampling. Other than methodological differences, temporal heterogeneity in opposing selection pressures can also drive the observed changes in flavonoid frequencies. 28 Theoretical and empirical evidence support the hypothesis that polymorphisms can be maintained by temporal variation in antagonistic selection pressures (Reimchen and Nosil 2002; Borash et al. 1998; Schemske and Bierzychudek 2001). It is possible that an opposing selection pressure favoring C 's exists and fluctuates in strength between years. If the opposing selection pressure overcomes the ionic stress pressure imposed in edaphically extreme sites, C individuals could then rise in frequency in spite of their hypothesized inefficient ionic stress tolerance physiology. It is enticing to present drought stress as a counter selection pressure favoring C individuals over A ' s because another facet of the flavonoid=edaphic race hypothesis proposes that race C individuals are greater drought tolerators than A ' s . In light of this, it is worth noting that drought stress was unlikely to have occurred in the 2005 growth season, as much of California had recorded higher than average rainfall that year (Western U.S . Climate Historical Summaries, http://www.wrcc.dri.edu/climsum.html). Evaluating the adaptiveness of sulfated flavonoids in saline environments The link between increased sulfate conjugation with phenolic compounds such as flavonoids in plants subjected to waterlogging and high salinity has been well documented (Harborne 1975, 1993; Barron et al. 1988). The toxicity of elevated sulfate concentrations in these environments may be ameliorated by the conjugation of inorganic sulfates with flavonoids (Harborne 1975, 1993; Barron et al. 1988). From the results of Tomas-Barberan et al.'s 1987 study, Rajakaruna et al. 2003 c suggested that sulfated flavonoids may provide a counter ion to ameliorate the negative effects of increasing salt accumulation in plants growing in saline areas. Although often cited in the literature, the 29 adaptiveness of sulfated flavonoids in waterlogged and saline environments is largely untested and remains mostly conjectural. While sulfated flavonoids have been identified in several halophytes such as Suaeda maritima (Harborne 1993), and Helianthus paradoxus (Karrenburg et al. 2006) they are absent in others such as Plantago maritima (Harborne 1993). Furthermore, the presence of sulfated flavonoids is not correlated to salt tolerance across the genus Lasthenia. While many species of Lasthenia inhabit saline areas (eg. L. chrysantha, L.ferrisiae, L. glabrata, L. maritima, L. playcarpha) none of these taxa produce sulfated flavonoids. The two species other than L. californica sensu lato that produce sulfated flavonoids are L. conjugens and L. fremontii and they inhabit waterlogged but ionically benign environments (vernal pools and wet meadows; Ornduff 1993). To conclude, the results from our survey show that the correlation between ionic tolerance and sulfated flavonoids is not as strong as previously hypothesized by Rajakaruna et al. 2003a,c. However, the evidence provided in this study is correlative, and experimental evidence providing a mechanistic link between sulfated flavonoid production and enhanced salt tolerance is necessary to fully address this issue. Patterns of ecological and physiological differentiation across the complex Results from P C A and univariate A N O V A s of soil and tissue characteristics indicate that the major axes of edaphic variation in the complex occur along lines of increasing edaphic toxicity. Although the majority of sampled sites were ionically benign inland sites (e.g. grazed pastures), our P C analyses failed to detect strong 30 correlations between original tissue/soil variables and the major axis of variation that described this inland group. The patterns of variation in soil and tissue characteristics found between serpentine, coastal, and pasture habitats are fairly congruent with findings reported in the literature. The following sections wi l l assess the similarities and differences of the edaphic and physiological differentiation found across the Lasthenia californica complex with other findings reported in the literature. Patterns of edaphic variation Edaphic patterns for serpentine sites sampled P C analyses and A N O V A s show that levels of magnesium, nickel, chromium and cobalt are quite divergent between, serpentine sites and the other habitats groups analyzed. In our survey, mean magnesium concentrations from serpentine sites were almost 5 times higher than soils from non-serpentine sites. Elevated magnesium concentrations are common to serpentine soils (Brooks 1987). Concentrations of up to 36% M g O in soils from Mt . Tamalpais have been reported, whereas the percentage composition o f M g O in non-ultramafic soils are generally less than 1% (Robinson et al. 1935, in Brooks 1987). High concentrations of heavy metals such as chromium, nickel and cobalt have been reported in soils derived from ultramafic rocks across the globe, including Mt . Tamalpais (Robinson et al. 1935, in Brooks 1987). Although A N O V A ' s show significant increases in chromium and cobalt concentrations in serpentine soils, the actual availability of these heavy metals can be quite low (Brooks 1987), and so may not exert a stress on plant growth (Halstead 1968; Crooke and Inkson 1955). This point wi l l be re-examined below in light of tissue data findings. 31 The chemistry of serpentine soils challenges plant growth with a combination of increased magnesium and heavy metal levels and decreased nutrient concentrations, in particular, that of calcium (Walker 1954). Although some researchers have shown that calcium availability is not limiting in serpentine environments (eg. Proctor et al. 1971; Proctor et al. 1981), most researchers support the hypothesis that the negative effects of lowered calcium levels is exacerbated by the high levels of magnesium also found in these soils (Kruckeberg 1954, 1989; Proctor and Woodell 1975) because plants generally require magnesium/calcium ratios of less than one for optimal growth. O f the habitat groupings we sampled, mean calcium levels were lowest in serpentine soil; however, the difference between serpentine and coastal groupings was not significant. The overall magnesium/calcium ratio did prove to be significantly higher in serpentine soils. It is suspected that other nutrients such as nitrogen have lowered concentrations in serpentine soils due to the sparse vegetation associated with these habitats (Kruckeberg 1984; Brooks 1987; Nagy and Proctor 1997; Proctor 1971a). We found that on average, nitrogen concentrations were not significantly lower in serpentine sites. Potassium deficiencies are not commonly associated with serpentine soils, but deficiencies have been reported in some sites in Western North America because of low mineralogical concentrations in ultramafic parent material (Burt et al. 2001). We found no significant differences in potassium concentrations in serpentine soils. These results when considered together suggest that the sampled serpentine populations of Lasthenia californica sensu Ornduff are generally not suffering from nutrient deficiencies. The chemical adversities of serpentine soils are often coupled with unfavorable physical properties such as low organic matter percentage, low water retention, and 32 coarse texture (Brooks 1987). These characteristics can then induce drought stress in serpentine dwelling plants. This stress is hypothesized to be an important component to the low productivity shown by serpentine vegetation in dry climates such as California and South Africa (Kruckeberg 1954; Gardner and Macnair 2000; Hughes et al. 2001). Although we did not directly measure particle size and water content, we did measure levels of organic matter content, which affects soil texture and water retention capacity (Brooks 1987). We found that the sampled serpentine sites did not have lower than average organic matter content. This result is consistent with the moderate levels of nutrients found in our serpentine sites. The overall conclusion from our results suggests that the stress imposed on populations of Lasthenia californica sensu Orduff from serpentine sites is due primarily to chemical toxicity (elevated magnesium and heavy metal concentrations) and not due to nutrient deficiency or physical harshness. Edaphic patterns for coastal sites sampled A s mentioned, the P C 2 axis was tightly correlated to concentrations of sodium and to a lesser extent sulfur, which agrees with other surveys showing increased substrate salinities found in coastal sea spray communities (Boyce 1954). The levels of substrate salinity can be variable at coastal sites, and is dependent on the region and distance from the sea (Boyce 1954; Barbour 1978). M g C l is another salt in sea spray that may exert a potential stress on plants growing in coastal communities (Wu 1981; Ashraf et al 1981; Hodson et al. 1981). However, we did not find a significantly increased level of magnesium in the coastal sites surveyed compared to ionically benign pastures. 33 Relationships between ecological and physiological variation Serpentine environments impose a diverse ensemble of chemical and physical stresses. Hypothesized physiological adaptations are also diverse and include drought tolerance (Gardner and Macnair 2000; Hughes et al. 2001), resilience to low nutrient levels (Proctor 1971a), mechanisms of tolerating adverse magnesium/calcium ratios (O 'Del l et al. 2006; O ' D e l l and Claassen 2006; Main 1974; Madhok and walker 1969; Marrs and Proctor 1976) and resistance to heavy metal toxicity (Reeves et al. 1999; Boyd and Martens 1998). A s mentioned, the most likely stress imposed by serpentine soils onto the sampled L. californica populations is one of excess magnesium and perhaps heavy metals. A N O V A ' s performed on tissue samples from these sites show that magnesium levels are significantly higher in tissues from serpentine sites than compared to benign inland pasture habitats. These results suggest that serpentine plants have acquired tolerances to high shoot concentrations of magnesium. It may also suggest that these plants cope with high magnesium by accumulating this ion in the shoot (Marrs and Proctor 1976; Main 1981). Soils data show elevated heavy metal concentrations of nickel, chromium, and cobalt from serpentine soils, suggesting that these ions may impose a stress in serpentine sites. Interestingly, plant tissues from serpentine sites do not show elevated concentrations of cobalt and chromium. Low concentrations of chromium and cobalt in the tissues of plants from serpentine habitats suggests two possibilities: that these populations have adapted efficient mechanisms of heavy metal exclusion; or more likely, that these heavy metals are not present in an available form. Amelioration of the potential toxicity of heavy metals occurs when serpentine soils are sufficiently high in pH 34 and organic matter content (Brooks 1987). In particular, chromium becomes increasingly insoluble at higher pH, a condition typical of serpentine soils. For example, it has been shown that chromium concentrations in plant-available forms are very low in serpentine soils with a pH over 6.8 (Brooks 1987). A N O V A ' s of soil data confirm that serpentine soils from the sites we sampled have a mean pH (6.8) that is significantly higher than in other habitats. A s previously stated, increased organic matter content has been shown to reduce plant available heavy metal concentrations by chelating the ionic forms of these metals (Crooke 1956; Halstead et al. 1969). Given that the organic composition of these soils is not lower than in other habitats, there is likely sufficient organic mass to reduce available heavy metals in these serpentine soils. Finally, competition between heavy metal ions and other abundant divalent cations during uptake may contribute to heavy metal stress resistance (Proctor 1971b). For instance, Proctor et al. 1971b and Robertson et al. 1985 demonstrated that increasing divalent ions in soil solutions with high N i 2 + or Cr 2 * decreased the toxicity of these heavy metals. Thus, for the serpentine populations of Lasthenia californica studied, high magnesium in these soils may mitigate the toxicity of heavy metals. Hyperaccumulation is another commonly reported adaptation acquired by serpentine adapted taxa (Boyd and Martens 1998). Tissue data show that hyperaccumulation is not occurring in serpentine tolerant plants in Lasthenia californica sensu lato. Hyperaccumulators such as Thlaspi montanum var. siskiyouense and Streptanthus polygaloides have been shown to accumulate as much as 10,000 ppm of nickel (Reeves et al. 1983). Our tissue data show that nickel hyperaccumulation is not occurring in serpentine tolerant plants tested; the mean nickel concentration in tissues 35 from serpentine tolerant Lasthenia californica sensu lato sites tested is 17 ppm. This finding is corroborated by another tissue analysis performed on individuals at a serpentine grassland at Jasper Ridge which showed similarly low levels of nickel accumulation (Rajakaruna and Bohm, 1999). Other than the inconspicuous physiological mechanisms acquired to cope with the challenges of serpentine, serpentine ecotypes also evolved putatively adaptive morphological traits such as xeromorphy or increased root to shoot ratios (Kruckeberg 1984). Obvious above-ground morphological differentiation by serpentine tolerant populations of Lasthenia californica sensu lato is noticeably absent. In contrast, individuals from saline areas, in particular coastal areas, have noticeably divergent morphologies. For example, specimens from salt-deserts from southern California are noticeably stunted and succulent. Moreover, coastal populations of Lasthenia californica are very morphologically diverse. Although usually succulent, this trait varies in magnitude. Plants from the coast can be annual or perennial and large or stunted. Some of these morphological changes are analogous to adaptations reported in the majority of halophytes. These adaptations include: succulence, increased cell volume, increased epidermal thickness, and increased root to shoot ratios (Flowers et al. 1986). Morphological changes such as succulence are correlated with physiological adaptations such as sodium accumulation, vacuolar compartmentalization of sodium, and the ability to replace sodium for potassium as a vacuolar osmoticum (Reimann and Breckle 1995; Greenway and Munns 1980; Flowers et al. 1986). Tissue data suggest that populations of saline inhabiting populations of Lasthenia californica have adapted to salt stress by increased sodium accumulation. This may be evident in the significantly higher 36 tissue levels of sodium in coastal areas. Potassium tissue levels did not correlate significantly to potassium soil levels across the complex, but rather, appeared to group based on ecology. Although soil potassium means were almost identical between coastal and inland sites, concentrations in tissues were significantly lower in coastal versus inland populations as determined by pairwise contrasts. This may mean that coastal individuals differentially accumulate sodium and use it to replace potassium as a vacuolar osmoticum. While potassium tissue levels seem differentiated based on ecological origins, levels of sodium accumulation generally respond to levels of substrate sodium across sites fairly consistently (indicated by the significant regression). This suggests that across the complex, plants respond to increased sodium levels by increasing sodium uptake and accumulation. Members of the Lasthenia californica complex display a wide range of ecological, morphological, and physiological diversity. Soi l , tissue and morphological data collected based on field sampling all suggest that salinity is associated with the most striking patterns of variation in the complex. Although the morphological distinctiveness of L. californica coastal populations has sparked the curiosity of several botanists (Ornduff 1966, Rajakaruna et al. 2003a), the putative adaptiveness of coastal traits has not been tested. In Chapter 3, we explore whether differentiation in coastal traits is genetically determined and whether these traits are responsive to the major environmental variable differentiating its environment: salinity. 37 Table 2.1: Site ID, ITS species identity, locality, habitat, and flavonoid race determinations from this and previous surveys are summarized. Data from the current study is abbreviated as 2007; data from Rajakaruna and Bohm (1999) and Rajakaruna et al., 2003 surveys are abbreviated as 1999&2003; data from Desrochers and Bohm 1993 is abbreviated as 1993. Sites are roughly ordered geographically, with the northernmost sites listed at the top of the table. Site ID ITS species 19 A 93 C 1999&2003* 20 A 07 C Location and habitat State County 55 L. californica C 15 Medford, across road from a water treatment plant, grassy pasture OR Jackson 56 L. californica C 15 Medford, near Kodac plant, grassy pasture OR Jackson 54 L. californica .30 C 20 Table rock summit, meadow and vernal pool. OR Jackson 73-75 L. orduffii**** Otter Point, coastal bluff OR Curry 74 L. orduffii**** Cape Sebastian, coastal bluff OR Curry 76 L. orduffii**** Cape Blanco, coastal OR Curry 27 L. californica 33 C 15 Near Red Bluff, pasture C A Tehama 72 L. californica 32 Pudding, coastal bluff C A Mendocino * 67 L. californica 20 Near Corlevaro way and Grill way, coastal bluff C A Sonoma SPS L. californica 25 5 A 8 37 Salt Point State Park, coastal bluff C A Sonoma 26 L. californica 22 5*** C 20 Near Clay, pasture C A Sacramento 25-36 L. californica C 30 Near Clay, pasture C A Sacramento 24-35 L. californica 25 Outside Clay, pasture C A Amador 51 L. californica v 15 Just past Lake Co. and Colusa Co. boundary, pasture C A Colusa 48 L. californica 20 Near Middletown, pasture C A Lake 49 L. californica 20 Junction of Rte. 53 and Rte. 20, pasture C A Lake 43 L. californica 30** C 20 Point Reyes, near coastal lighthouse C A Marin 58 L. californica 18 Steep Ravine Campground, serpetine till C A Marin 62 L. californica 15 Dillion Beach, coastal C A Marin 63 L. californica 30 18 Near Dillion Beach, rocky roadside C A Marin 42 L. californica 11 15 Mt. Tamalpais, near O'roukes, serpentine C A Marin 21-41 L. californica 54 6 A 18 5 Mt. Tamalpais, near old mine trail, serpentine C A Marin 59-44 L. californica 21 Chimney Rock, coastal C A Marin JRA L. californica mostly A 12 Jasper ridge lower reaches of serpentine ridge C A San Mateo JRC L. gracilis mostly C 2 10 Jasper ridge upper reaches of serpentine ridge CA San Mateo 40 L. californica A 13 Kirby, serpentine C A Santa Clara 22 L. gracilis A 24 Byron, alkali flat C A Contra Costa 3 L. gracilis A 15 Coalinga, sloping pasture C A Fresno 4 L. gracilis A 15 - Coalinga, pasture C A Fresno 30 L. californica A 15 Arroyo Seco, sloping meadow C A Monterey P 1 L. gracilis C 2 18 Front gate at Palmer Ranch, salinized pasture C A Monterey p2 L. gracilis A 8 33 Palmer Ranch near a swampy area, salinized pasture C A Monterey 17-34 L. gracilis 36 A 20 Priest Valley, pasture C A Monterey 5 L. gracilis 65 A 15 Near Tehachipi, road side embankment C A Kern 10 L. gracilis A 20 Outside Motte Rimrock, grassy yard near houses C A Riverside 7 L. gracilis 32 A 15 Palmdale, meadow C A Los Angeles Notes: * Rajakaruna and colleagues (Rajakaruna and Bohm 1999; Rajakaruna et al. 2003) reported only uniformly A or C populations (except at Jasper ridge). The number of plants surveyed per site was between 3-6. ** Originally reported as "type B" and described as a modified profile of C (Desrochers and Bohm 1993), this profile waslater interpreted as C by Rajakaruna et al. 2003. This interpretation is followed here (2007 data). *** One of these C's was originally reported as profile B (Desrochers and Bohm 1993). * * * * u s fype similar to L. californica. This ecotype is now recognized as L. Ornduffii. \ Table 2.2: Transformations used in statistical analyses. A ) Transformations used in A N O V A and P C analyses for soil variables. B) Transformations used in A N O V A ' s for tissue variables. Soil variables Variable Transformation Organic matter Square root C E C Square root pH None N L o g l f / P Fourth root K Square root Ca none M g L o g l O S Log lO Na Log lO Z n Log lO M n Log lO Fe None Cu Log lO B Square root A l Log lO Cr Log lO Co L o g l O N i L o g l O Tissue variables Variable Transformation N Log lO S Square root P Log lO M g None Ca None K None Na Fourth root Fe Fourth root M n Log lO B Log lO C u Log lO Z n Square root A l Fourth root Cr Log lO (1 +Cr) N i Log lO (1 + Ni) 40 Table 2.3: A N O V A results for soil variables. Variables showing significant differences (P < 0.05) between A and C populations are bolded. Means that are significantly higher are bolded as well . Only differences between A and C were tested; mixed populations of both A and C were not included in the A N O V A . Mineral concentrations are reported in ppm; C E C in meq/lOOg. C Mean Variable A Mean S.E. S.E. A C mixed Mean S.E. Organic matter (%) C E C pH Nitrate P Na/K Mg/Ca K C a M g S Na Zn M n Fe Cu B Al Cr C o N i 3.0 14.6 6.7 16.5 37.7 0.112 1.864 193.5 937.3 913.6 6.7 14.6 3.4 44.5 64.8 1.0 0.5 471.6 204.5 44.7 689.3 0.5 2.5 0.1 9.0 7.4 0.046 0.881 26.6 143.1 312.0 0.6 5.1 0.5 4.5 5.9 0.2 0.1 47.2 97.9 18.6 351.6 4.5 13.5 5.9 10.4 15.5 0.902 0.761 128.8 747.1 458.3 10.3 106.4 3.3 51.4 85.3 1.4 0.7 753.9 117.8 25.0 248.2 0.4 0.8 0.1 3.4 2.3 0.296 0.218 12.3 77.4 89.4 I. 1 29.0 0.4 I I . 0 4.3 0.2 0.3 83.4 53.1 9.3 189.8 5.0 14.3 6.3 8.9 18.9 0.823 1.718 114.1 518.8 746.3 19.5 119.9 2.4 27.3 76.4 0.8 0.5 749.4 557.0 61.6 940.3 1.8 1.8 0.3 1.7 7.5 0.294 0.650 35.2 110.5 195.7 9.2 59.6 0.3 6.3 9.0 0.2 0.1 301.6 294.8 29.3 522.5 Table 2.4: Summary of number of A , C , or mixed A and C profile populations present on Edaphic type A C Mixed Salinized (coastal bluffs and alkaline flats) 0 . 5 2 Serpentine 3 1 2 41 Table 2.5: P C A results of soil data. Eigenvalues, % of variance explained, cumulative % of variance explained, and Pearson's correlations (r) to axes are included. P C axis 1 2 3 Eigenvalues 5.494 4.132 2.041 % of variance 28.918 21.75 10.742 Cumulative % of variance 28.918 50.668 61.41 Variables Organic matter (%) 0.311 -0.711 0.451 C E C 0.655 -0.282 0.608 pH 0.4 0.617 0.177 Nirate 0.146 0.002 0.124 P -0.761 0.059 0.413 K -0.529 -0.025 0.627 Ca -0.436 0.557 0.474 M g 0.875 0.134 0.378 S -0.11 -0.829 0.252 Na 0.119 -0.845 -0.052 Fe 0.26 -0.377 -0.24 Zn -0.411 0.273 0.549 M n -0.023 0.76 -0.031 Cu 0.298 0.395 0.051 B 0 -0.348 -0.102 A l -0.281 -0.513 0.154 Cr 0.929 -0.044 0.108 Co 0.894 0.302 0.108 N i . 0.938 0.065 0.048 42 Table 2.6: Means and standard error of soil variables for serpentine, inland benign pastures and coastal sites. Variables showing significant differences (P < 0.05) are bolded. Significant pair-wise differences (after a Tukey H S D correction) between habitat groupings are denoted with letters where shared letters represent non-significant differences. Mineral concentrations are reported in ppm; C E C in meq/lOOg. Coastal bluffs Inland pastures ** Serpentine Variables Mean S.E. Mean S.E. Mean S.E. Organic matter (%) 6.5 0.8 a 3.0 0.3 b 4.1 0.5 ab C E C 16.0 1.2 a 11.4 0.8 b 19.9 2.9 a PH 5.8 0.2 a 6.2 0.1 a 6.9 0.1 b Nitrate 8.1 3.0 14.5 5.8 11.2 2.8 P 20.2 4.3 * 28.4 4.9 * 7.2 1.0 N a / K 1.444 0.468 a 0.143 0.028 b 0.319 0.049 b M g / C a 0.653 0.044 b 0.418 0.099 b 4.178 0.949 a K t 159.8 22.2 160.0 17.7 88.2 17.3 C a 525.6 67.5 a 1049.0 79.7 b 406.7 30.5 a M g 335.0 40.1 a 403.8 77.6 a 1850.8 302.8 b S 16.1 1.8 b 6.9 0.3 a 6.5 0.9 a N a 180.8 32.2 a 17:5 3.8 b 27.2 6.8 c Zn 2.9 0.5 3.6 0.5 2.2 0.1 M n 16.0 3.0 a 64.1 10.0 b 35.8 6.7 b Fe 78.0 5.4 72.8 4.6 77.8 7.9 Cu 1.0 0.3 1.5 0.2 1.1 0.1 B 0.5 0:1 a 0.3 0.0 b 0.6 0.1 a Al 1105.2 153.7 a 593.8 37.3 b 376.3 45.8 c C r 69.9 12.4 a 53.3 11.3 a 866.7 141.6 b C o 9.9 1.2 a 1.6.8 1.8 a 139.6 18.3 b .Ni 48.7 8.2 a 58.7 19.8 a 2480.2 404.4 b * Pair-wise differences are not shown because nonsensical differences were found due to insufficient power after Tukey H S D corrections. fP=0.053. ** The saline/alkaline pools were excluded from the inland pasture grouping, and not included as a separate habitat category because of the low N of 2. 43 Table 2.7: Means and standard error of tissue variables for A , C , and mixed A and C populations. Variables showing significant differences (P < 0.05) between A and C populations are bolded and means that are significantly higher are bolded as well. Only differences between A and C were tested; mixed populations of both A and C were not included in the A N O V A . A l l mineral concentrations are reported in ppm. A C A C mixed Variable Mean S.E. Mean S.E. Mean S;E. N 12262.0 3319.3 10182.6 666.9 9477.5 2356.9 S 1355.7 219.4 2204.6 299.1 2628.6 778.5 P 3258.4 568.7 1685.8 120.6 1872.8 382.2 Mg/Ca 3.599 0.139 0.373 0.052 0.635 0.137 Na /K 0.091 0.078 0.455 0.154 0.378 0.190 K 20055.5 1821.0 20688.6 1339.2 15159.3 1527.0 Na 854.5 612.5 6616.7 2000.6 5487.7 2669.8 M g 2878.8 672.3 3559.5 362.3 5401.3 624.5 Ca 11978.9 1308.3 10749.1 556.5 9426.4 1874.2 Fe 225.2 36.3 288.1 53.9 150.9 39.5 Mn 54.3 9.5 154.3 20.7 97.7 48.4 B 24.0 1.6 29.9 4.7 25.4 3.4 Cu 5.9 0.3 7.2 0.5 5.2 L 0 Zn 48.6 5.0 52.0 4.1 33.5 2.1 A l 37.5 8.8 107.2 30.7 71.8 30.5 Cr 1.8 0.4 2.5 0.4 1.9 0.2 N i 5.1 2.3 3.1 1.2 8.2 3.8 44 Table 2.8: P C A results of tissue data. Eigenvalues, % of variance explained, cumulative % of variance explained, and Pearson's correlations (r) to axes are included. P C axis 1 2 3 Eigenvalues 5.264 3.394 1.22 % variance 35.094 22.624 8.132 Cumulative % variance ; 35.094 57.718 65.85 Variables N 0.28 0.546 0.027 • S 0.877 0.183 0.108 P -0.21 0.534 -0.316 M g 0.705 -0.57 0.1.28 Ca -0.215 0.82 0.02 K -0.58 0.451 0.444 Na 0.95 -0.021 -0.186 Fe 0.692 0.045 0.313 M n 0.479 0.56 0.327 B 0.397 0.251 -0.656 Cu 0.649 0.535 -0.094 Zn 0.131 0.247 0.446 A l 0.857 0.223 0.018 Cr 0.721 -0.222 0.046 N i 0.175 -0.866 0.115 45 Table 2.9: Means and standard error of tissue variables for serpentine, inland benign pastures and coastal sites. Variables showing significant differences (P < 0.05) are bolded. Significant pair-wise differences (after a Tukey H S D correction) between habitat groupings are denoted with letters. Shared letters represent non-significant differences. Mineral concentrations are reported in ppm; C E C in meq/lOOg. Variable Coastal Mean S.E. Inland pastures** Mean S.E. Serpentine Mean S.E. •N 13140.0 694.1 a 10781.1 1867.9 ab 6964.0 1267.9 b S 38.82.3 217.2 a 1269.5 127.9 b 1555.8 267.9 b P 1799.0 188.2 * 2593.0 361.8 * 1212.3 210.0 * Mg/Ca 0.484 0.039 a 0.196 0.027 b 1.048 0.089 c Na/K 1.106 0.224 a 0.006 0.001 b 0.269 0.135 b K 15609.2 1281.1 a 23707.0 925.6 b 13767.0 1761.2 a Na 15952.0 2183.7 a 148.1 12.7 b 2813.6 940.0 c Mg 5236.9 411.9 a 2265.1 173.7 b 6032.1 369.7 a Ca 10991.4 740.0 a 12588.1 575.3 a 5917.0 570.2 b Fe 444.3 91.9 a 196.1 29.6 b 188.4 51.4 ab Al 245.7 35.9 a 27.8 5.4 b 24.1 10.7 b Mn 197.4 22.1 a 107.0 20.9 b 36.1 3.9 c B 31.9 1.4 a 22.7 1.0 b 21.0 1.2 b Cu 8.9 0.9 a 6.0 0.2 b 4.7 0.5 c Zn 54.5 9.3 48.8 3.0 45.1 6.2 Cr 3.5 0.6 a 1.5 0.2 b 3.0 0.7 a Ni 2.4 0.9 a 1.8 0.6 a 17.5 1-9 b * Pair-wise differences are not shown. Nonsensical differences were found due to insufficient power after Tukey H S D corrections. ** The saline/alkaline pools were excluded from the inland pasture grouping, and not included as a separate habitat category because of the low N of 2. 46 Table 2.10: Results for linear regressions of soil and tissue concentrations of P, K , Ca, S, Na , Zn, Fe, Cu , B and N (ppm). Element X Transformation Y Transformation Constant Slope R 2 P P log(base 10) log(base 10) 2.655 0.506 0.642 0.0001 K none none 0.289 C a * none none 3.227 0.067 0.449 0.0001 S * log(base 10) log(base 10) 2.285 1.021 0.584 0.0001 Na log(base 10) log(base 10) 1.041 1.247 0.712 0.0001 Zn log(base 10) log(base 10) 1.518 0.335 0.185 0.012 Fe none none 0.496 C u log(base 10) log(base 10) 0.1058 B * none none 17.291 18.556 0.348 0.0001 N * . >. none none 0.441 *in all these cases, 1 outlier of exceptional leverage or influence was removed from the analysis 47 a) 3 2 1 0 CN o -1 C L -2 -3 -4 -5 + I I I - 4- 4- X . 4-4- 4- * ] + ' A A A ^ V-4-\ A 4^  — X 4-/o — /xr * J — O / x / — I •xf / I I I -2 0 PC 1 3 2 I- A A A 14-.0 CN o -1 C L -2 -3 -4 -5 _L — I — c — I — A C c ~ l 1 A H M M A C A -4* M c c c M I I I -6 -4 -2 0 2 4 PC 1 Figure 2.1: PCA ordination of axes 1 and 2 for soil variables, (a) Habitat group symbols correspond to serpentine (A), coastal (X), inland benign pastures (+)> and inland alkaline flats (o). Confidence ellipses represent one standard deviation from group centroids (excluding alkaline flats; N=2). (b) Symbols correspond to uniformly A profile populations (A), uniformly C profile populations (C), or populations with mixed A and C individuals (M). a) -5 0 5 10 PC 1 I A I A c c c A & M . c c — A C ^ M C -A C - -~ . A -M M c A I I -5 0 5 10 PC 1 Figure 2.2: PCA ordination of axes 1 and 2 for tissue variables, (a) Habitat group symbols correspond to serpentine (A), coastal (X), inland benign pastures (+), and inland alkaline flats (o). Confidence ellipses represent one standard deviation from group centroids (excluding alkaline flats; N=2). (b) Symbols correspond to uniformly A profile populations (A), uniformly C profile populations (C), or populations with mixed A and C individuals (M). V Figure 2.3. Bivariate distribution of soil (x-axis) and tissue (y-axis) concentrations of Ca (a) and B (b). Symbols correspond to serpentine sites (S), coastal sites (C), inland benign pastures (P), and inland alkaline flats (A). Linear regressions were fitted for the complex. a) b) E CL -S 00 ID O 1.0 1,5 ' Log [Soil P (ppm)], 0.6 0.7 0.8 0:9 1.0 1.1 1:2 1.3 1.4 Log [Soil S (ppm)] . c) d) 1 2 3 Log [Soil Na (ppm)] J I J I I I J L 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Log [Soil Zn (ppm)] Figure 2.4. Bivariate distribution of log soil (x-axis) and log tissue (y-axis) concentrations of phosphorus (a), sulfur (b), sodium (c), and zinc (d). Symbols correspond to serpentine sites (S), coastal sites (C), inland benign pastures (P), and inland alkaline flats (A). Linear regressions were fitted for the complex. S1 [Soil Mg] ppm [K soil] (ppm) Figure 2.5. Bivariate distribution of soil (x-axis) and tissue (y-axis) concentrations o f magnesium (a) and potassium (b). Symbols correspond to serpentine sites (S), coastal sites (C), inland benign pastures (P), and inland alkaline flats (A) . Confidence ellipses represent one standard deviation from group centroids. C H A P T E R 3 Differentiation in morphological and physiological responses to salinity between coastal and inland populations of Lasthenia californica sensu stricto I N T R O D U C T I O N Plants can exhibit a large amount of phenotypic variation within species. Patterns of ecotypic differentiation that are associated with environmental gradients demonstrate the importance of selection in generating morphological diversity and have inspired a rich history of botanical inquiry (e.g. Turesson 1922a,b; Clausen et al. 1948; Grant and Hunter 1962). Divergent selection pressures such as those imposed by coastal versus inland environments produce particularly interesting lines of phenotypic - ecological covariation and remain relevant systems for studying patterns and mechanisms of local adaptation and trait differentiation (Antfinger 1981; Nagy 1995). Coastal salt spray communities are subject to the stresses imposed by salt exposure (Oosting 1945; Boyce 1954; Barbour 1978; Cartica and Quinn, 1980; Sykes and Wilson 1988, 1999) along with other stresses that may include sand accretion (Van der Valk 1974; Sykes and Wilson 1990, 1999), wind exposure, and low soil water-holding capacity due to coarse, sandy substrates (Boyce, 1954). Salt loads from sea spray induce stress on vegetation by both increasing substrate salinity (Barbour 1978; Sykes and Wilson 1999) and causing leaf necrosis because of salt accumulation on leaf surfaces (Griffiths and Orians 2004). While coastal habitats, including those in California, are under constant levels of salinity stress, the climate along the coast is generally milder than in associated inland areas, which receive sufficient year-round precipitation. In contrast, the interior of California alternates between moist winters and hot summers which exert the predictable droughts that characterize Mediterranean climates (Barbour et 53 al. 2000; Tenhunen et al. 1985). In habitats such as disturbed roadsides and grasslands, the success of many annuals depends on rapid and competitive growth during the wet season before the onset of summer droughts (Grime 1979; Aronson et al. 1992; Fitter and Hay 2002). Although the extent of sodium stress varies between and within coastal sites - salt loads change with proximity to the source of exposure and with the amount of wind force available to propel sea spray (Barbour et al. 1985) - sodium has been suspected or indicated as being the primary factor affecting species distributions or morphological variation of plants in these areas (Wells and Shunk 1938; Sykes and Wilson 1999; Griffiths and Orians 2004; Goldstein et al. 1996). For example, an ecological survey at Point Reyes, in central California, has shown that salt amounts in collection traps correlates with plant species distribution (Barbour 1978). The importance of salt as a selection pressure in coastal environments is also demonstrated by detection of genetic differentiation in salt tolerance in coastal populations of a number of unrelated taxa that span inland and coastal habitats (Reimann and Breckle 1995; Kohl 1997; Tiku and Snyadon 1971; Rozema et al. 1978; Ab-shukor 1988; Watt 1983). Studies of coastal ecotypes further reveal that populations growing in salt exposed environments have independently acquired similar sets of morphological and physiological adaptations to cope with the stresses of salinity. Most of these adaptations tackle aspects of the ionic and osmotic stresses induced by increased salinity (Flowers et al. 1986; Greenway and Munns 1980). Enhanced ionic stress tolerance through increased sodium accumulation and vacuolar compartmentalization of sodium has been reported in many coastal taxa (Flowers et al. 1986; Tester and Davenport 2003). Osmotic regulatory 54 adaptations such as vacuolar sodium/potassium osmoticum replacement and the production of organic osmotic regulating compounds (eg. proline and mannitol) are also commonly reported adaptations in coastal plants (Maathuis and Amtmann 1999; Munns 2002; Harborne 1993; Y e o 1998; Harborne 1993). These physiological and morphological investments in stress regulatory mechanisms are empirically and theoretically linked to trade-offs in lowered intrinsic growth potentials (RGRmax) when these stresses are relaxed (Grime and Hunt 1975; Arendt 1997; Chapin 1980, 1991; Orians and Solbrig 1977; Fernandez and Reynolds 2000). While broadly advantageous alleles can spread rapidly across populations under low levels of gene flow and promote cohesiveness within species (Rieseberg and Burke 2001; Morjan and Rieseberg 2004), tradeoffs have the potential to maintain variation between neighboring populations by limiting the spread of alleles underlying traits that differ in net fitness in contrasting environments. In this study we focus on patterns of differentiation in stress tolerance and growth as well as potential trade-offs between these in coastal and inland ecotypes of Lasthenia californica (Asteraceae). The taxonomic history of this species has been complicated by complex patterns of cryptic genetic and morphologically divergent variation within members of the L. californica species complex. The new circumscription of L. californica sensu stricto (henceforth referred to as L. californica) now reflects a monophyletic lineage that includes two formerly distinct species: L. macrantha and the northern populations of L. californica sensu lato. L. macrantha (as it was formerly known) was a mostly coastal assemblage of three subspecies all showing distinct morphology from interior populations of L. californica sensu lato. Lasthenia macrantha 55 was distinguished by larger strap shaped, succulent leaves and facultative perenniality (Ornduff 1966; Chan et al. 2002). Both D N A sequence (Chan et al. 2001, 2002) and R A P D marker data (Rajakaruna 2002) have shown that L. macrantha is polyphyletic in origin, with populations tending to nest within geographically neighboring inland populations of L. californica sensu lato. While easily distinguishable in overall shape and form, there is little reproductive differentiation between the two formerly distinct species (Ornduff 1966; Chan et al. 2002). The high level of interfertility between these taxa (Ornduff 1966) reflects the general genetic cohesiveness of L. californica shown by molecular data. Recent field surveys of populations of L. California have confirmed the ecological, physiological and morphological distinctiveness of coastal versus inland populations. Soil samples show that coastal sites have markedly higher levels substrate salinity (>25 fold) than inland sites. These environmental differences are amplified in levels of tissue sodium (>130 fold higher sodium in coastal tissues). Substrate potassium levels were also higher in coastal sites (almost doubled), however, coastal tissue potassium contents was only 2/3 that of inland sties (a result generally shown by halophytes; Flowers et al. 1986). Observations of succulence found that while differentiation is greatest between inland and coastal sites, significant variation can also be found within coastal populations. A t coastal sites, leaves of individuals growing on edges of sea exposed bluffs tended to be the most succulent with succulence decreasing with distance from the sea spray exposure line (G. Choe, personal observation). The morphological and physiological observations in the field suggest that phenotypic 56 divergence between populations of Lasthenia californica may be driven by divergent responses to environmental sodium. Here we investigate the basis of the phenotypic differentiation observed between ecotypes of Lasthenia californica by growing two populations of each ecotype in common conditions with varying levels of the environmental factor we hypothesize as the most likely to promote the phenotypic divergence shown between ecotypes: salinity. Phenotypic divergence in ecologically relevant traits can be attributed to: 1) adaptive genetic differentiation; 2) phenotypic plasticity, in which different phenotypes are induced by environmental differences; or 3) by a combination of these components where the environment induces genetically differentiated responses between ecotypes. Given the body of evidence indicating similar trait differentiation in a number of coastal taxa, it is unlikely that the patterns under study reflect random genetic differentiation due to drift. We take the classical approach of comparing norms of reaction between populations across salinities to investigate the possibility that coastal populations are adaptively differentiated in salt tolerance and associated tolerance traits. We first tested differentiation in performance (i.e. salt tolerance) between populations under increasing salinity. Adaptive differentiation was inferred when coastal populations showed uniformly higher fitness under high salinity. We then tested for differences in intrinsic and induced stress tolerance expressed in morphological and physiological traits. Upon detection of differentiation in both salt tolerance and related traits, we looked for evidence of ecotypic differences in trade-offs between stress tolerance and growth potentials. 57 M A T E R I A L S A N D M E T H O D S Seed collection The populations used in this study were selected from sites widely distributed within the California floristic province. This sampling strategy was implemented to minimize possible trait similarities between populations due to shared population histories or gene flow. Four populations, two inland and two coastal, were tested (Table 3.1). Between Apr i l and July of 2005 seeds were collected from these sites from one mature flower head for each of >30 individuals sampled haphazardly at each site. Each > head contains approximately 50-200 full or half-siblings. These sites were also part of a broader survey of soil and plant tissue variables (Chapter 2). Inland and coastal sites differ in numerous environmental variables. Inland habitats are Mediterranean in climate, with cool, wet winters and hot, dry summers. The mean annual precipitation at the two sites inland sites ranges between 48 - 63 cm (National Climatic Data Center, 1948-2006; http://www.ncolc.noaa.gov). Coastal sites are cooler and consistently moist year round, with mean annual precipitation at these two sites between 96 - 104 cm (National Climatic Data Center, 1948-2006; http://www.ncolc.noaa.gov). Sites also differ in various soil characteristics. Most notably, sodium concentrations in the soil are substantially higher in coastal populations than inland populations (Table 3.1). Field data also show that the average level of tissue sodium and potassium concentrations of individuals from these sites differ dramatically in inland versus coastal populations (Table 3.1). 58 Germination and experimental design Seeds were ripened at room temperature for several months and then pooled within populations. These seeds were sown in Petri-dishes filled with nutrient medium composed of these following ingredients: macronutrients: 0 .25mM K H 2 P 0 4 , 0 .5mM K N 0 3 , 0 .5mM C a ( N 0 3 ) 2 . 4 H 2 0 , 0 .5mM M g S 0 4 . 7 H 2 0 , O . l m M N H 3 N 0 3 ; micronutrients: 20/<M F e ( E D T A ) . 2 H 2 0 , 5 0 ^ M H 3 B 0 3 , 12/ /M M n S 0 4 . H 2 0 , lpM Z n S 0 4 , 7 H 2 0 , l ^ M C u S 0 4 . 5 H 2 0 , 0 . 2 ^ M N a 2 M o 0 4 . 2 H 2 0 , 5 0 ^ M N a C l . The nutrient solution was neutralized to a final pH of 6 using Ca(OH) 2 . 0.7% agar was added to the solution to create a semi-solid medium. The sown seeds were placed in a growth chamber with initial conditions set at 17°C days and 14°C nights, cycling at 12 hour photoperiods. Germination and initial seedling growth occurred over 7 days. A total of 512 pots (2 by 7 inch Cone-tainers, Stuewe & Sons, Inc., Corvallis, OR, U S A ) were filled with 6 parts soil-less potting medium (Terra Lite Soil M i x , W . R. Grace, Ajax, Ontario, Canada) and 1 part inert fritted clay (Turface, Profile Products L L C , Buffalo Grove, IL, U S A ) . Two seedlings were planted in each pot. These pots were watered with a nutrient solution similar in composition to the agar medium described above, with the two exceptions that macronutrient concentrations were doubled and no agar was added. Pots were randomly arranged in a growth chamber set to provide a 14 hour photoperiod with 22°C day and 18°C night cycles. Photosynthetic photon flux of approximately 290/<mol m"V1 was measured at plant level using an Apogee Q M S W - S S quantum meter (Apogee,.Logan, Utah, U S A ) . After 14 days of growth, seedlings were inspected for vigor, and then pots with healthy seedlings were grouped into fours. Groups were arranged by size. One 59 member of each group was used for an initial harvest to obtain initial dry weights for relative growth rate calculations (Evans 1972). The remaining unharvested pots were re-randomized within the chamber where they grew until the final harvest. N a C l treatment commenced after the initial harvest. N a C l treatments were applied in a randomized arrangement to the pots. Given both population and treatment application were both randomized without restriction, the experiment followed a completely randomized design. Treatment solutions were composed of the aforementioned nutrient solution plus either a OmM, 7 0 m M , 120mM addition of N a C l for the control, medium stress, and high stress treatment, respectively. Medium and high N a C l treatment concentrations were determined based on trial experiments which suggested that N a C l concentration causing 50% decline in biomass for coastal population was within the range of 120 m M N a C l , which was not overly toxic to inland populations. NaCl stress was increased Incrementally with each watering. This slow ramping of stress was implemented in order to avoid sudden N a C l shock. The initial treatment for both medium and high stress received only 3 5 m M N a C l . For the second treatment, 7 0 m M N a C l was applied for both the medium and high stress treatments. The final 120mM N a C l concentration for the high NaCl stress treatment was applied on the fourth watering after an intermediary 90mM NaCl watering step. Watering treatments were continued as needed until the final harvest. To ensure that tissue analyses reflected uptake amounts, all plants were sprinkled from above with distilled water several days prior to the final harvest. This was done to rinse off salts that may have accumulated on the leaves during treatment applications. 60 The plants were harvested over two days, 54-55 days after the initial planting. Harvest began before the onset of flowering to ensure that most individuals were at similar developmental stages (pre-bolting). Plants were cut at the root-shoot transition zone. Previous attempts to reliably record root mass were unsuccessful as plants could not grow robustly in media where their roots could be easily separated from the matrix. Leaves were separated and their fresh weight was determined. The leaves were arranged in a flatbed scanner; leaf number and leaf area of the scanned images were then determined using Sigmascan Pro (SPSS 1999). Leaves were then dried in a forced draft oven at 60°C for 2-3 days until a constant weight was reached. Above ground dry mass (biomass) was then determined. This was used as a measure of fitness. The concentrations of Na + , K + , M g 2 + and C a 2 + in dried leaf tissues were determined by ICP detection. This procedure involved an initial V H P closed vessel microwave acid digestion and a subsequent ICP spectrometer analysis of digested solutions (Analytical Chemistry Laboratories, Ministry of Forestry, Victoria, B C , Canada). The dry matter of three pots was randomly pooled to generate sufficient mass required for the procedure. In total, 96 groups were created (8 groups of 3 pots per population X 4 populations X 3treatments). Due to equipment malfunction, the leaf ion data of only 93 samples are reported here. The data set was trimmed to 288 pots (24 per population X 4 populations X 3 treatments) to maintain a balanced design (all crossed levels having 24 cases) in spite of unequal mortality and additional losses due to harvesting errors. Analyses that explored the relationship between ion content and growth and morphology traits used averaged growth and morphology values within pooled ion samples. 61 From the traits measured during the two harvests, the following derived traits were determined: Relative growth rate (RGR) , leaf dry matter content ( L D M C ) , specific leaf area ( S L A ) and succulence. Relative growth rates were calculated from the following formula (Evans 1972): [In dryweight (final harvest) - In dryweight (initial harvest)] / number of days. Succulence was calculated with the following equation (Debez et al. 2004; Karrenberg et al. 2006): (fresh weight - dry weight) / surface area. Specific leaf area ( S L A ) and leaf dry matter content ( L D M C ) were calculated as (leaf dry weight / area) and (fresh weight /dry weight), respectively. Data analysis Differences in growth, morphology and leaf ion traits were analyzed with an analysis of variance using the G L M packages in S Y S T A T 11 ( S Y S T A T , 2005) and J M P IN 4.04 (SAS Institute, 2000). Significant interaction terms were first tested for each trait. We interpreted significant interaction terms for biomass as evidence for differentiation in salt resistance, and significant population by salt treatment interactions of R G R (specifically, crossed reaction norms) as evidence of trade-offs in growth rates and salt tolerance. When significant interaction terms were detected, the nature of the interaction was further explored by decomposing the interaction sum of squares into the following treatment X population contrasts testing 1) whether the response to salinity differed between inland versus coastal ecotypes 2) whether the response to salinity was consistent between populations within each ecotype (i.e: inland 49 vs. 54 or coastal 67 vs. 72). These essentially test for differences of reaction norm slopes between assigned groups. These contrasts were tested as a single df contrast between the linear trend of 1) 62 \ averaged response of coastal populations versus inland populations and 2) populations from the same habitat. Before linear contrasts were investigated, an initial null of no linear trend was tested to confirm that reaction trends could be sufficiently captured by a linear model. Sets of orthogonal quadratic polynomials were examined, but in all tests, linear trends provided a better fit (data not shown) and so quadratic models were not included. Because contrasts were pre-planned and orthogonal, their significance was tested against an alpha of 0.05 without adjustments (Sokal and Rohlf 1995). Along with interaction terms, significant main effects were determined. Significant simple effects of populations within treatment levels were determined for a few traits; of particular importance, simple population effects for R G R ' s at the control treatment were determined to test for differences in RGRmax. Pair-wise tests between means were tested when significant terms were determined for a few traits of interest (such as R G R ) . Post-hoc, pair-wise contrasts were controlled for inflated family-wise type I error rates with a Bonferroni adjustment. The residuals of A N O V A tests were visually inspected and tested with Wilk-Shapiro 's test (a test of normality) and Levene's test (a test of unequal variances) to verify their adherence to the assumptions of A N O V A . N a + leaf concentrations and N a + / K + content of leaves were log or log(l + x) transformed to improve variance differences between groups. In general, non-extreme deviations from A N O V A assumptions were generally ignored as A N O V A s with balanced cells tend to be fairly robust against minor deviations (Quinn and Keough 2002). We used A N C O V A s to determine whether the relationship between sodium accumulation and succulence differs between ecotypes. The A N C O V A tests whether the 63 linear relationship between sodium leaf content and succulence is similar or divergent between ecotypes by testing whether the ecotype slopes are homogenous. The relationships between leaf ion accumulation between N a + and K + and C a 2 + and M g 2 + were determined by calculating Pearson product-moment correlations. For these multiple pair-wise correlations, a false discovery rate was controlled using a Bonferroni adjustment on the family of tests. R E S U L T S Growth measures: biomass and relative growth rates Both population and treatment main effects were significant for biomass (Table 3.2). Populations strongly differed in biomass production in response to salinity (indicated by the highly significant population by treatment interaction; Table 3.2 and Figure 3.1a), which provided evidence of population differentiation in salt tolerance. When this interaction was further explored, we found that the linear trend between inland and coastal ecotypes showed a significant difference but the trend between the populations within each ecotype did not differ (Table 3.2). This suggests that while on the ecotypic level there is clear differentiation in salt tolerance, there is little difference in tolerance to salinity between populations from similar habitats. Coastal populations were uniformly less hindered by the negative effects of salinity than were inland populations. This is illustrated by the decreased steepness of the reaction norms displayed by coastal populations versus those of inland populations (Figure 3.1a). Interestingly, the population least affected by salinity stress was also from the saltiest coastal site (population 67, Table 3.1 and Figure 3.1a); but as mentioned earlier, the difference in 64 reaction norms between populations of the same ecotype did not prove to be significant. We found that biomass decreased in fairly equal increments across salt treatments for coastal populations. In contrast, the decline in biomass between the control and medium stress treatments for inland populations was noticeably more severe than the decrease in biomass between medium and high stress treatments. This could be due to population differences in the response to the slower ramping of stress levels in the high versus medium stress treatment. There was a significant population by treatment interaction term for relative growth rates (Table 3.2; see Figure 3.1b). The significant crossing of reaction norms (Figure 3.1b) suggests trade-offs in R G R s ' and salt tolerance. The decline of R G R s ' for the inland ecotype under increasing salinity was more severe than the decline of R G R s ' for the coastal ecotype, a result supported by the significant ecotype X treatment contrast (Table 3.2). A s found for biomass, populations from the same habitat were not differentiated in their response to salinity; the linear trends between populations within ecotype groupings were not significantly different from each other (Table 3.2; Figure 3.1b). Both population and treatment main effects were significant (Table 3.2). We tested for difference in RGRmax (simple population effects at control treatment level), and found significant differences. Pair-wise tests between populations at the control treatment showed that inland populations have significantly higher RGRmax than coastal populations (Figure 3.1c). 65 Leaf morphology Most morphological traits showed significant population by treatment interaction terms. Specifically, leaf number, L D M C , and succulence showed significant interaction terms. S L A is the only trait lacking significant interactions between factors. Generally, population differences in morphological responses to environmental sodium were muted compared to differences in responses shown by growth traits (biomass and R G R ) . This result can be intuitively appreciated by the lack of crossed reaction norms for morphological traits, as compared to biomass and R G R , which both showed crossed trends between ecotypes. Most traits also showed significant main effects of both treatment and population. These results suggest that the majority of morphological traits are genetically differentiated between populations and that environmental sodium levels have an effect on morphology. Leaf number decreased across sodium treatments for all populations tested (significant treatment effect; Table 3.2 and Figure 3.2a). However, the extent of the decline in leaf number is dependent on ecotype. When interaction terms were partitioned into a contrast between ecotype, the contrast in ecotype trends proved significant (Table 3.2). Populations within ecotypes showed similar reactions to one another as indicated by non-significant contrasts of trends between populations 49 and 54, as well as between 67 and 72. Generally, coastal populations had more leaves across all sodium treatments (indicated by the significant population effect; Table 3.2 and Figure 3.2a). Population by salt treatment interaction for L D M C was significant (Table 3.2). Like most other traits, the reaction norms were significantly different between ecotypes but were not significantly different between populations of the same ecotype (tested by 66 linear contrasts; see Table 2 and Figure 3.2b). In general, coastal leaves had a higher water component to their leaves (indicated by the lower ratio of dry over fresh weight) regardless of salt treatment. This result is demonstrated by the pair-wise contrasts of marginal (pooled across salt treatments) population means (Figure 3.2c). Figure 3.3a illustrates that while S L A is both affected by salt treatments and is intrinsically different between populations, the way in which S L A changes in response to environmental sodium is fairly consistent between populations. A N O V A results show that while treatment and population main effects were highly significant (P < 0.0001 in both cases), S L A did not show a significant population X treatment interaction (Table 3.2). Pair-wise tests between populations show that coastal populations have significantly thicker/denser leaves than inland populations as indicated by their lower S L A values (Figure 3.3c). Succulence can be interpreted as a measure that combines the water content component of L D M C and the leaf thickness component of S L A . Whereas the responses of S L A and L D M C to salt treatments are either non-significant ( S L A ) or similar in response between populations ( L D M C ) , succulence has a clear difference in the direction of response to salinity between populations (population X treatment interaction with P <0.0001; Table 3.2). When main effects were investigated, a significant population and a non-significant treatment term were found. Note that the significance of the salt treatment effect was masked by the nature of the population X treatment interaction. The salt treatment main effect tests for significant differences by comparing the pooled means (across all populations) at each treatment level. Because salt affected populations in a way that either increased or decreased succulence values fairly symmetrically, the 67 resulting means were therefore fairly equal across treatment levels. We tested for differences in succulence values between the control and high stress treatment for each population separately (a test of simple treatment effects within populations) and found that succulence differed significantly between sodium treatments for each population tested (tested at an alpha of 0.05, consistent with contrasts using sets of orthogonal tests). This confirms that environmental salt levels do indeed impose a biological effect on succulence. Figure 3.3b graphically illustrates this finding showing that while coastal populations tend to gain in succulence under increasing salinity, inland populations tend to lose succulence under saltier conditions. In general, much of the variation in response to salinity can be attributed to ecotypes. Whereas the reaction norms between ecotypes differ significantly (P<0.0001), the responses between populations of the same ecotype do not differ significantly (Table 3.2). Coastal populations had significantly higher values of succulence compared to inland populations across all salt treatments (as indicated by pairwise comparisons of means; Figure 3.3d). Leaf ion content A l l populations dramatically increased in sodium tissue content under increasing salt stress. However, the magnitude of the response differs between ecotypes (indicated by the significant treatment X population interaction and the significant inland versus coastal trend contrast; Table 3.3, Figure 3.4a). The responses between populations of the same ecotype did not differ significantly (Table 3.3). Coastal populations tended to accumulate greater levels of sodium over all treatments including the control (population effect < 0.0001; pair-wise differences at control are shown in Figure 3.4c). Populations 68 thus appear both genetically differentiated in constitutive levels of sodium accumulation as well as in their ability to increase sodium accumulation with increasing environmental salinity. Potassium levels in tissues also responded to salt exposure. Potassium levels decreased demonstrably with increasing salinity (Figure 3.4b) and showed significant treatment by population interactions (Table 3.3). Leaf potassium concentration responses showed a greater differentiation between populations and a greater complexity in the shape of the response than did sodium concentrations. This is demonstrated by the following results: 1) leaf potassium trends were significantly different between the two inland populations, which is indicated by the 49 vs. 54 X treatment contrasts (Table 3.3) and 2) the reaction norms of 67 crossed both 49 and 54. Generally, coastal populations tended to have lower potassium contents in their leaves in conditions of high salinity (see pair-wise differences at medium and high NaCl stress, Figure 3.4b); but interestingly, coastal population 67 showed the highest level of leaf potassium concentration at the control (Figure 3.4b). Results for sodium/potassium ratios of leaves logically followed the trends shown by leaf sodium and potassium concentration results already described. Sodium/potassium ratios increase greatly under increasing salinity for all populations (indicated by the highly significant treatment effect, Table 3.3), but the magnitude of sodium/potassium replacement differed between populations (significant interaction term, Table 3), ecotypes (significant ecotype trend contrast, Table 3.3), and populations of each ecotype (significant population within ecotype trend contrast, Table 3.3). Coastal populations tended to have the highest levels of sodium to potassium (Figure 3.5). 69 Correlations between traits Correlations between N a + and K + leaf contents are very strong for all populations, with P<0.0001 (Table 3.4). In comparison, correlations between N a + and divalent cations, M g 2 + and C a 2 + , are mostly non-significant (Table 3.4). This suggests that the accumulation of N a + and M g 2 + and C a 2 + are generally independent. Coastal population 72 is the one exception to this trend. N a + accumulation is negatively correlated to both M g 2 + and C a 2 + levels in leaf tissues (P<0.0001; Table 3.4). Although all populations showed significant correlations between sodium accumulation and leaf succulence (Table 3.4), the direction of the correlation differed between ecotypes. Inland populations showed a negative correlation between succulence and sodium accumulation while coastal populations showed a positive correlation between sodium content and succulence (Figure 3.6). When homogeneity of slopes was tested between ecotypes using an A N C O V A , a significant sodium slope X ecotype interaction confirmed that the relationship between sodium accumulation and succulence did indeed differ between populations (Table 3.5). D I S C U S S I O N Local adaptation to an environmental pressure is interpreted from norms of reaction data when two criteria are met (Murren et al. 2006; Schmitt 1993; Silander and Antonovics 1979): 1) when significant habitat origin X treatment interactions in plant performance is present 2) when the populations showing highest performance in experimental conditions are from similar historical conditions in the field. When we examined the performance (biomass) of Lasthenia californica populations in differing 70 levels of salinity we found a significant population by treatment interaction; secondly, we found that coastal populations were more fit under increasing salinity. The result that both populations from habitats with greater substrate salinity were less impacted by the negative effects of salt (a result supported by the ecotype trend contrast) strengthens the argument that coastal populations have evolved greater salt tolerance due to historical environmental pressures. But note that because the seeds used in this experiment were field collected, differences shown between populations could be a combined result of genetic and maternal effects (Rossiter 1996; Roach and Wulff 1987). However, maternal effects have been shown to diminish with time (Schmitt and Antonovics 1986) and so have less influence in mature plants, such as those used in this experiment. In general, our results are supported by a large body of literature reporting genetic differentiation in salt tolerance between coastal and non-coastal populations (Reimann and Breckle 1995; Kohl 1997; Tiku and Snyadon 1971; Rozema et al. 1978; Ab-shukor 1988; Watt 1983; K i k 1989) and imply that accompanying morphological and physiological adaptations that accommodate greater salt tolerance are likely to be found in coastal populations of Lasthenia californica. Field observations suggest that increased succulence, sodium accumulation and compartmentalization, and vacuolar osmoticum replacement of potassium with sodium are all possible morphological and physiological investments made to enhance salt tolerance in coastal environments. Our results suggest that the phenotypic divergence in these traits in the field reflects both genetic differentiation in traits and in trait responses to salt. When we analyzed the response of population trait means across sodium levels, we found a clear differentiation between populations in the direction or magnitude of trait responses to environmental sodium. Like the results for biomass, population differences in trait expression under experimental conditions followed predictions made from observed differentiation in the field. Succulence, sodium accumulation and sodium/potassium ratios were expressed at higher levels in coastal sites, and for the most part, these coastal individuals showed greater intrinsic and induced levels of expression in controlled conditions. We also found surprisingly convergent intrinsic and induced levels of expression from populations of the same habitat type, which further supports the argument that salinity pressure, in part, shaped the divergent morphology of coastal versus inland ecotypes of Lasthenia californica. The following paragraphs wi l l outline in greater detail our hypothesis for the ecological relevance of the morphological and physiological trait differences shown in controlled conditions. Field collected coastal plants contain over 140 times the sodium as those in non-saline sites. The results presented here suggest that these differences reflect both genetic differentiation in constitutive levels of sodium accumulation (as indicated by the differences at the control) and differences in their ability to increase sodium accumulation with increasing environmental salinity. In general, increased levels of sodium accumulation indicate increased efficiencies in sodium compartmentalization (Flowers et al. 1986). Efficient sodium compartmentalization in the vacuole is thought to be adaptive because it alleviates both the ionic and osmotic stress components of sodium toxicity (Chinnusamy et al. 2005). Vacuolar compartmentalization of N a + decreases ionic stress by reducing cytosolic Na + , where it competes with K + in binding sites of important enzymatic reactions (Yeo 1998; Tester and Davenport 2003). Osmotic stress is partly 72 relieved by vacuolar accumulation of sodium by lowering solute potential in the vacuole, which in turn promotes turgor (Chinnusamy et al. 2005). Typically most glycophytic plants use potassium, and to a lesser degree, calcium and magnesium as vacuolar osmotica (Chinnusamy et al. 2005; Greenway and Munns 1980; Tester and Davenport 2003; N iu et al. 1995). Salt tolerant species are commonly found to replace sodium for potassium (Flowers et al. 1986, Chinnusamy et al. 2005). This is adaptive because it allows the remaining potassium to localize in the cytosol, and removes toxic sodium from the cytosol to the vacuole, where it is then used as a 'cheap osmoticum' (Tester and Davenport 2003; Chinnusamy et al. 2005). A s mentioned, field data suggest that this type of osmoticum switching is employed by coastal individuals as a way of achieving salinity resistance; potassium levels in coastal tissues were found to be two-thirds the levels of inland tissues despite higher substrate potassium levels in coastal areas. Our experimental results support the hypothesis of enhanced osmoticum replacement by coastal individuals. When grown under common conditions, potassium levels tended to decrease more dramatically in coastal populations with increasing environmental sodium. The significant ecotype trend contrast for leaf sodium/potassium reinforces the result that potassium is replaced by sodium under increasing salinity to a higher degree in coastal than inland ecotypes. Osmoticum replacement not only occurs between sodium and potassium, but sodium can also replace magnesium and calcium to a limited degree in the vacuoles of some halophytes (Greenway and Munns 1980; Karrenberg et al. 2006). Only Coastal population 72 showed a significant negative correlation between tissue levels of sodium and magnesium or sodium and calcium. Because these correlations were absent in an 73 equally salt tolerant population (coastal 67), it suggests that salinity resistance in L. californica can be achieved by different pathways. Succulence is the most obvious morphological difference between these two ecotypes of L. californica in the field (Ornduff 1966). In coastal areas, succulence levels can differ between and within populations. In contrast, the level of succulence in inland populations (excluding alkali flats) shows little variation. Our experimental results indicate genetic differentiation in intrinsic and salt induced levels of succulence. A t every treatment level, coastal individuals are at least twice as succulent as inland individuals. Furthermore, coastal populations show a slight but significant increase in succulence under increasing salinity while inland populations decrease in succulence with increasing sodium stress. The changes in succulence shown by coastal populations in experimental conditions partly explains the observed differences in coastal succulence in the field. A t coastal sites, individuals tend to become more succulent with increasing salt spray exposure, our experimental results suggest that these differences are in part due to plastic responses to sodium. The minor decreases in succulence shown by inland individuals under increasing experimental sodium levels most likely reflects losses of turgor, presumably due to an inability to overcome the osmotic stress (water potential) exerted by high salt concentrations. The suggestion that water content is higher for coastal versus inland leaves is supported by the results that show that L D M C (leaf dry mass/leaf fresh mass) values are lowest for coastal populations in all treatment levels. Succulence is associated with increases in cell and vacuolar volume (Flowers et al. 1986; Yeo and Flowers 1980; Hajibagheri et al. 1984). Increased vacuole volume can lead to increased sodium accumulation, and in this way, succulence is thought to be 74 adaptive in saline environments (Greenway and Munns 1980). Since succulence is correlated to a greater capacity to regulate sodium, the amount of sodium in leaves should then be positively correlated with succulence in salt tolerant genotypes. Under experimental conditions, coastal populations show a positive correlation between succulence and sodium tissue levels whereas inland populations show a negative correlation between succulence and leaf sodium levels. The difference in trait association between ecotypes is supported by the significant sodium by ecotype interaction in an A N C O V A test. This result reinforces the argument that coastal populations attain greater sodium tolerance by a suite of composite adaptations involved in enhancing sodium regulation via sodium accumulation and compartmentalization. Correlations between traits can arise by linkage disequilibrium (nonrandomly associated alleles tied by the influence of evolutionary action) or pleiotropy (where one locus contributes to the expression of both traits). Under linkage disequilibrium, strong selection can maintain trait associations, where the fitness benefit of joint trait expression overcomes recombination (Lynch and Walsh 1998). A n association between sodium accumulation and succulence is broadly reported across taxa (Flowers et al. 1986), and our results add to this pattern, as these two traits are correlated in coastal ecotypes in L. californica. Given that only coastal ecotypes show a positive correlation between these two traits, selection rather than pleiotropy is the likely cause of the correlation. f Negative genetic correlations in the form of constraints have important impacts on the direction of response to selection, where the response of one trait to selection is limited by the fitness of the correlated trait (Lynch and Walsh 1998; Maynard Smith et al. 75 1985; Mitchell-Olds 1996). When selection favors alternate traits in neighboring populations, trade-offs have the potential to enhance or maintain trait variation between these populations (Fry 2003). We asked whether there is evidence of trade-offs between increased salt tolerance and lowered intrinsic growth potentials for populations of Lasthenia californica. To test for evidence trade-offs between growth potentials and stress tolerance, we first established that coastal populations are more salt tolerant. We then compared relative growth rates of populations under control levels of salt, as well as their changing growth rates under increasing salinity. Our results show that growth rates are uniformly lower for coastal types than inland populations when grown under non-stressful conditions (RGRmax), and that with increasing salinity, growth rate trends cross between ecotypes so that coastal populations show higher relative growth rates than inland populations under conditions of high salinity stress. These results support the existence of trade-offs between stress tolerance and relative growth rates in coastal versus inland populations of L. californica. The hypothesis of trade-offs between growth potential and stress tolerance has a wealth of empirical support from studies of various harsh environments, in particular nutrient stressed habitats (Arendt 1997; Chapin et al. 1993; Chapin 1980, 1991; Orians and Solbrig 1977; Fernandez and Reynolds 2000; Grime and Hunt 1975). The trade-off between tolerance and growth is thought to arise because of genetically determined constraints on resource allocation, where investments to stress, tolerance adaptations (such as succulence) come at the expense of resource expenditure on growth (Arendt 1997; Chapin et al. 1993). Correlations between leaf investment indices such as S L A (a 76 measure of leaf density) and growth potentials have been studied across numerous taxa. It is hypothesized that leaves of high R G R species maximize growth potentials by increasing their photosynthetic capacity (leaf area) while minimizing their investments in structural or protective investments (leaf area/mass), which then results in higher S L A values (Lambers and Poorter 1992; Chapin et al. 1993; Fitter and Hay 2002). These studies have generally shown that S L A and growth rates are widely correlated over numerous taxa (Poorter and Remkes 1990; Lambers and Poorter 1992), with lower values of S L A associated with slow growing stress tolerant genotypes, and with higher S L A values associated with ruderal or highly competitive plants with rapid growth rates (Fitter and Hay 2002; Poorter and Remkes 1990; Lambers and Poorter 1992). Our results are consistent with overall trends in S L A values shown by other stress tolerant genotypes and non-stress tolerant genotypes. S L A values were found to be uniformly lower in coastal than inland populations of L. californica. In life-history strategy theories proposed by Grime (1977, 1979), environments with high abiotic stress and low disturbance are predicted to favor genotypes that allocate more resources to stress tolerance adaptations and less resources to rapid growth (i.e. 'stress tolerators'). In environments with low stress but high disturbance or in environments of overall low stress, genotypes with rapid and/or competitive growth are thought to be favored (i.e. 'ruderals' or 'competitors'; Grime 1977, 1979). Inland populations of Lasthenia californica grow in areas that are either disturbed (eg. roadsides) or relatively competitive (eg. grasslands; MacDougall and Turkington 2004) and are likely to exert selection favoring ruderal or competitive strategies (Grime 1979). Moreover, the climate of interior California may select for an accelerated life-history 77 strategy to avoid the predictable droughts of the hot and dry Mediterranean summers (Aronson et al. 1992; Fitter and Hay 2002). In comparison, coastal ecotypes grow in relatively moist and ocean-moderated climates that can support slower, year-round growth (evident in the later flowering times or perenniality of coastal L. californica) provided that the plants allocate sufficient resources to tolerate the constant abiotic stress imposed by their environment. In these contrasting environments, the relative importance of abiotic versus biotic pressures on fitness are thought to shift (Weldon and Slauson 1986; Greiner L a Peyre 2001; Crain et al. 2004), where traits related to tolerance are more intensely selected for in stressed environments while traits related to rapid growth or competitive ability are favored in non-abiotically stressed environments (Chapin et al. 1993). Because trade-offs act to magnify divergence in traits when contrasting environments favor opposing traits (Futuyma and Moreno 1988; Rice and Salt 1990), potential trade-offs between allocation to growth and stress tolerance may help to explain why the morphologies of coastal versus inland populations of Lasthenia californica are so divergent in the field. 78 Table 3.1: Summary of ecotype, population identification, localities, soil and tissue data from field sites used in this study. Ecotype Site ID Locality Soil N a + (ppm) Soil K + (ppm) Tissue N a + (ppm) Tissue K + (ppm) Inland 49 Junction of Rte. 53 and Rte. 20, grassy pasture. Elevation: 501m. Lake Co. C A 10 120 112 22879 Inland 54 Table rock summit, andesite deposit. A long meadows near vernal pools. Elevation: 610m. Jackson Co. O R 17 80 142 26264 Coastal 67 Near Corlevaro way and Gr i l l way, coastal bluff. Elevation 15m. Sonoma Co. C A 237 269 17688 16975 Coastal 72 Pudding, coastal bluff. Elevation 4m. Mendocino Co. C A 159 136 16750 13911 79 Table 3.2. Summary of A N O V A results for biomass (g), relative growth rates (In mg days"1), total leaf number, L D M C (g g ' ) , S L A c m 2 g \ and succulence (g H 2 0 cm' 2). Traits Source of variation df Biomass RGR Leaf number L D M C S L A Succulence Population 3 SS 0.1636969 0.0005638 87982 0.021361 1643900 0.0709201 P <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 Treatment 2 SS 0.2854945 0.05135924 74395 0.010429 103559 0.0000074 P <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 N.S Population X Treatment 6 SS 0.0452703 0.01302682 6831 0.001091 17276 0.0012794 P <0.0001 <0.0001 0.0003 0.0001 N .S <0.000L *(49,S2) vs (67,72) X Treatment 1 SS 0.0331827 0.0115853 4681 0.000378 P <0.0001 <0.0001 <0.0001 0.0017 12044 0.0010371 <0.0001 *49 vs 54 X Treatment 1 SS P 0.0007889 N.S 0.0000045 N.S 284 N.S 0.000001 N.S 2 0.0000045 N.S *67 vs 72 X Treatment SS P 0.0014797 N.S 0.0000252 N.S 1 0.000072 N.S N.S 2494 0.0000339 N.S Error 276 SS 0.1168866- 0.00321575 71133 0.010386 436727 0.0078214 *above are contrasts of linear trends / Table 3.3. Summary of A N O V A results for ion traits: Na (ppm), K (ppm), and Na/K. Ion trait Log (Na) K L o g ( l + N a / K ) 3.37257 . 7134348375 0.625194 <o!oooi <0.0001 <0.0001 63.27856 35150000000 1.701708 <0.0001 <0.0001 <0.0001 0.50520 1500246175 0.294249 <0.0001 <0.0001 <0.0001 0.46119 925502722 0.201967 <0.0001 <0.0001 <0.0001 Source of variation df Population 3 SS P : Treatment 2 SS P Population X Treatment 6 SS P *(49,52) vs (67,72) X Treatment 1 SS P *49 vs 54 X Treatment 1 SS P *67 vs 72 X Treatment 1 SS P Error 81 above are contrasts of linear trends 0.00014 99341598 0.009832 N.S 0.0133 0.0123 0.00027 38852113 0.019300 N.S N.S 0.0005 0.26854 1255021012 0.126200 81 Table 3.4. Values and significance of Pearson product-moment correlations between leaf values of sodium and potassium, sodium and calcium, sodium and magnesium, and sodium and succulence. Population N a + - K + N a + - M g 2 + N a + - C a 2 + N a + - Succulence 49 r (-) .0.940 - - (-) 0.701 P <0.0001 N.S. N .S . <0.0001 54 r (-) 0.941 _ _ (-) 0.759 P <0.0001 N.S . N.S . <0.0001 67 r (-) 0.986 _ _ 0.436 P <0.0001 N.S . N.S . 0.033 72 r (-) 0.992 (-) 0.875 (-) 0.936 0.745 P <0.0001 <0.0001 <0.0001 <0.0001 Table 3.5. Summary of A N C O V A results for sodium tissue contents and succulence. Source of variation df SS Ecotype Na ppm Ecotype X Na ppm Error 1 1 1 89 0.01925744 0.00002102 0.00035386 0.00073502 <0.0001 N.S . <0.0001 82 a) b) 0.25 0.14 C M Treatment C M Treatment c) Legend 49 54 67 72 Legend - 49 - 54 - 67 72 trt CD E c 0.14 0.13 -0.12 0.11 0.10 I-n 0.09 k O 1 1 0.08 | -0.07 0.06 49 54 67 72 Population Figure 3.1. Graphical results for growth traits. Reaction norms with means and standard errors for biomass (g) and relative growth rates (In mg days"1) are shown in figure (a) and (b) respectively. Reaction norms for each population are represented by a unique line described in the legend. Coastal and inland populations are labeled with a C or I, respectively. Control, medium and high stress treatments are denoted with a C, M and H , respectively, (c) Population means and standard errors for R G R values at the control treatment. Significant pair-wise differences between populations are denoted with letters, where shared letters represent non-significant differences. Inland populations and coastal populations are represented by empty bars or shaded bars, respectively. 83 a) b) 150 100 M Treatment Legend 49 54 67 72 C) 0.130 0.110 0.095 y 0.080 Q _ i 0.065 0.050 I i I I i 1 ..-IC ....-I--""' „ „ I C I ' -I i I C M H Treatment 49 Legend — 54 67 72 0.060 49 54 67 72 Population Figure 3.2. Reaction norm plots including means and standard errors for (a) leaf number and (b) L D M C (g g '). Reaction norms for each population are represented by unique lines described in the legend. Coastal and inland populations are labeled with a C or I, respectively. Control, medium and high stress treatments are denoted with a C, M and H , respectively, (c) L D M C (g g"1) population means and standard errors. Significant pair-wise differences between populations are denoted with letters, where shared letters represent non-significant differences. Inland populations and coastal populations are represented by empty bars or shaded bars, respectively. 84 a) b) 550 0.080 C M H Treatment 0.015 M Treatment Legend 49 54 67 72 Legend 49 54 67 72 C) d) 0.080 49 54 67 72 Population 0.010 54 67 72 Population Figure 3.3. Reaction norm plots including means and standard errors for (a) S L A (cm 2 g ') and (b) succulence (g H 2 0 cm"2). Reaction norms for each population are represented by unique lines described in the legend. Coastal and inland populations are labeled with a C or I, respectively. Control, medium and high stress treatments are denoted with a C, M and H , respectively. Population means and standard errors for (c) S L A (cm 2 g"1) and (d) and succulence (g H 2 0 cm"2). Pair-wise differences between populations are denoted with letters, where shared letters represent non-significant differences. Inland populations and coastal populations are represented by empty bars or shaded bars, respectively. 85 a) b) Legend 49 54 67 72 Legend 49 54 67 72 C) 2000 1500 ? Q. £ 1000 1 F Z 500 Ql I I « I MMII I 49 54 67 72 Population Figure 3.4. Reaction norms with means and standard errors for leaf ion concentrations of (a) sodium (ppm) and (b) potassium (ppm). Reaction norms for each population are represented by a unique line described in the legend. Coastal and inland populations are labeled with a C or I, respectively. Control, medium and high stress treatments are denoted with a C , M and H , respectively, (c) Population mean and standard errors for leaf ion concentrations of sodium (ppm) at control treatment. Significant pair-wise differences between populations are denoted with letters, where shared letters represent non-significant differences. Inland populations and coastal populations are represented by empty bars or shaded bars, respectively. 86 C M H Treatment Figure 3.5. Population means and standard errors for N a / K tissue ratios. Pair-wise differences between populations at each treatment level are denoted with letters, where shared letters represent non-significant differences. Note that Pair-wise tests were performed at each treatment level separately. Population 49, 54 ,67 ,72 are respectively represented by empty bars, lightly shaded bars, darkly shaded bars, and by black bars. 0.07 E o X CD (f) Legend o x + A 49 54 67 72 0 • X a? J> it [Na Tissue] (ppm) Figure 3.6. Bivariate distribution of tissue sodium contents and succulence. Linear slopes are fitted for each population. 87 C H A P T E R 4 Conclusion Significant findings and implications The complex patterns of ecological, morphological, physiological, biochemical and genetic variability shown by members of the Lasthenia californica complex offer both an interesting and challenging study system regarding questions related to ecological differentiation and adaptation. This complexity has contributed to the recognition of polyphyletic taxa (particularly L. macrantha sensu lato) and to the hypothesis of parallel evolution of flavonoid/edaphic races in closely related but phylogenetically distinct lineages. While in the first example, the polyphyletic assemblage of L. macrantha sensu lato has been confirmed independently by different researchers (Chan et al. 2002; Desrochers and Dodge 2003), the results presented in this thesis suggest that the second case of recurrent evolution proposed for L. californica should be re-examined. Our results suggest that the association between edaphic tolerances and flavonoid profile is not as strong as previously hypothesized by Rajakaruna and colleagues (Rajakaruna and Bohm 1999; Rajakaruna 2002, Rajakaruna et al. 2003a,b,c). We found that populations of A ' s did not show a greater affinity to ionically stressed habitats than did C 's . It is also likely that the presence of A / C flavonoid polymorphisms did not arise in these phylogenetic lineages by means of parallel evolution. The localized distribution of populations showing polymorphisms and/or shifts in A / C frequency suggests that gene flow is a plausible mechanism of introducing and/or maintaining this polymorphism. 88 In chapter 3, our results suggest that coastal populations of L. californica sensu stricto are more salt tolerant than the inland populations and that these two ecotypes are genetically differentiated in a number of ecologically relevant traits. One such trait, relative growth rate, was found to be significantly lower in coastal populations under non-stressful growth conditions, but then showed greater growth rates than inland populations under stressful conditions. These results support the hypothesis of trade-offs between growth potential and tolerance, as predicted by Grime (1977). Trade-offs between growth potentials and stress tolerance are most widely studied in nutrient gradients (Grime and Hunt 1975; Chapin 1980, 1991; Arendt 1997) and to a lesser degree in aridity gradients (Orians and Solbrig 1977; Fernandez and Reynolds 2000). This hypothesis has relatively less empirical support for saline systems although it is a natural extension of the theory and has been generally hypothesized as occurring (e.g. Chapin et al. 1993). It is generally thought that investigations of traits in intraspecific systems are particularly useful (e.g. Freemon and Herron 2001) because the confounding effects of phylogeny are not a concern as they are in cross species comparisons commonly used for the study of trade-offs in the ecological literature (e.g. Fernandez and Reynolds 2000). Hypotheses requiring confirmation and possible future studies A s mentioned, the results presented in Chapter 2 provide evidence refuting the hypothesis of parallel evolution of edaphic/flavonoid races across L. californica sensu ( lato. Even in light of these findings, the possible link between flavonoid production and ionic stress physiology has not been definitively answered. Specifically, the strength of genetic linkage between flavonoid production and physiological traits conferring ionic 89 stress tolerance should be determined. Flavonoid profiles A and C are differentiated in a number of compounds, and it could be that only the genes producing sulfated flavonoids co-segregate with tolerance conferring traits. In this way, the sulfated flavonoid/tolerance affinity aspect of the flavonoid=edaphic race hypothesis could be confirmed in the lab, even though in the field, associations between A and ionically stressed habitats seem not to exist. The flavonoid polymorphism found within edaphically harsh sites (e.g. serpentine populations at Mt. Tamalpais and the coastal bluff site at Salt Point State) hold interesting possibilities for future study. In the Lasthenia californica system, environmental heterogeneity along a serpentine grassland at Jasper Ridge has been shown to support the coexistence of two parapatric populations of L. californica subsp. californcia and L. gracilis that also differ in flavonoid composition. Generally, environmental heterogeneity has the potential to promote and maintain diversity in genotypes and species across landscapes (Kassen 2002; Jasmin and Kassen 2007; Weeks and Hoffman 1998; Freestone and Inouye 2006). It is possible that sites such as SPS and Mt. Tamalpais have established similar covarying distributions of soil' chemistry and flavonoid type. There is some observational data from our current work that suggest that increased salt exposure is linked to colonization of A flavonoid types at Salt Point State coastal bluffs (SPS). We have found that in one area of the SPS population, locations closer to the bluff edge had higher proportions of A than C individuals than did the less spray exposed areas further from the bluff's edge. If microsite co-variation between flavonoid profile and soil variation is found at these polymorphic sites, this would then 90 support the link between edaphic profiles and stress physiology (but only if A types tend to inhabit saltier areas). The polyphyletic assemblage L. macrantha sensu lato, provides an interesting system to the study of salinity adaptation. We provided evidence supporting differentiation of salt tolerance traits between populations. However, to confirm that the traits examined in our experiment are correlated to salt tolerance it is necessary to analyze the correlation between tolerance traits and salt resistance within populations. This could be done using standard quantitative genetic methods with replicated families of full or half sibling. We have also provided evidence of local adaptation of coastal populations to salt exposure. To confirm the hypothesis of local adaptation, it would be ideal to analyze the fitness of these populations in the field, in the form of a reciprocal transplant study. It would then also be possible to analyze the correlation of the traits that differentiate ecotypes to their fitness in the field. A s mentioned, Lasthenia macrantha sensu lato provides an interesting system for the study of salinity adaptation, but it also has the potential to provide insights into the study of stress adaptation in general. In our study we provided evidence of potential trade-offs between increased stress tolerance and lowered intrinsic growth rates. To explore this hypothesis in greater depth, the extent of heritable variation for alternative combinations of traits, particularly between tolerance traits and growth, should be determined. A l so in Chapter 3, we proposed a potential counter selection pressure that could be associated with inland sites (namely competition) that would favor increased growth rates. For a trade-off to occur, competitive interactions should be at a greater 91 intensity in grasslands and pastures sites compared to the intensity of biotic interactions at coastal sites. A l so , competition at coastal sites should be relatively less important than . abiotic selection pressures. These proposed counter selection pressures should be verified in the field by measuring the relative intensity of biotic stress on fitness at these contrasting sites. In closing, the Lasthenia macrantha sensu lato system can also be useful to study the mechanisms of recurrent phenotypic evolution. The 'macrantha' phenotype is expressed in other phylogenetically distinct lineages in addition to the L. californica lineage, namely, L. gracilis (the southerly distributed member of L. californica sensu lato) and L. ornduffii (which was a former subspecies of L. macrantha sensu lato). It would be of interest to examine whether the recurrent evolution of this phenotype follows lines of greatest genetic variation and covariance in quantitative traits (e.g. Schluter 1996a) and whether the biased use of a few key genes universally present in these closely related lineages produced the similarity in phenotypes shown (e.g. Schluter et al. 2004). 92 LITERATURE CITED Ab-Shukor, N.A., Kay, Q.O.N., Stevens, D.R, and D.O.F. Skibinski. 1988. Salt tolerance in natural populations of Trifolium repens L. New Phytologist 109: 483-490. Ahmad, I., and S.J. Wainwright. 1977. Tolerance to salt, partial anaerobiosis and osmotic stress in Agrostis stolonifera. New Phytologist 79: 605-612 Al-Hiyaly, S.E.K., McNeilly, T., and A.D. Bradshaw. 1988. The effect of zinc contamination fromelectricity pylons - evolution in a replicated situation. New Phytologist 110: 571-580. Anderson, R.C., Fralish, J.S., and J.M. Baskin. 1999. Savannas, barrens and rock outcrop plant communities of North America. Cambridge University Press, Cambridge. Antlfinger, A.E. 1981. The genetic basis of microdifferentiation in natural and experimental populations of Borrichia frutescens in relation to salinity. Evolution 35: 1056-1068. Aronson, J., Kigel, J., and A. Shmida. 1992. Adaptive phenology of desert and Mediterranean populations of annual plants grown with or without water stress. Oecologia 89: 17-26. Arendt, J.D. 1997. Adaptive intrinsic growth rates: an integration across taxa. The Quarterly Review of Biology 72: 149-177. Ashraf, M., McNeilly, T., and A.D. Bradshaw. 1989. The potential for evolution of tolerance to sodium chloride, calcium chloride, magnesium chloride and seawater in four grass species. New Phytologist 112: 245-254. Baldwin, B.G. 2005. Origin of the serpentine-endemic herb Layia discoidea from the widespread L. glandulosa (compositae). Evolution 59: 2473-2479. Barbour, M.G. 1978. Salt spray as a microenvironmental factor in the distribution of beach plants at Point Reyes, California. Oecologia 32: 213-224. Barbour, M.G., and W.D. Billings. 2000. North American Terrestrial Vegetation, Second Edition. Cambridge University Press, Cambridge. Barbour, M.G., De Jong, T.M., and B.M. Pavlik. 1985. Marine beach and dune plant communities. In The physiological ecology of North American plant communities. Eds. Chabot, B.F., and H.A. Mooney. Chapman & Hall, London. Barron, D., Varin, L., Ibrahim R.K., Harborne, J.B, and CA. Williams. 1988. Sulfated flavonoids - an update. Phytochemistry 27: 2375-2395. 93 Berglund, A - B . N . , Nyberg, B . , Dahlgren, S., and A . Westerbergh. 2004. Evidence for parallel evolution and site-specific selection of serpentine tolerance in Cerastium alpinum during the colonization of Scandinavia. New Phytologist 161: 199-209. Bohm, B . A . 1987. Intraspecific flavonoid variation. Botanical Review 53: 197-279. Bohm, B . A . Herring, A . , Nicholls, K . W . , Bohm, L .R . , and R. Ornduff. 1989. A six-year study of flavonoid distribution in a population of Lasthenia californica (Asteraceae). American Journal of Botany 76: 157-162. Bohm, B . A . , Saleh, N . A . M . , and R. Ornduff. 1974. The flavonoid chemistry of Lasthenia (Compositae). American Journal of Botany 61: 551-561. Bohm, B . A . and T.F. Stuessy. 2001. Flavonoids of the sunflower family (Asteraceae). Springer-Verlag/Wien, Austria. Borash, D.J. , Gibbs, A . G . , Johi, A . , and L . D . Mueller. 1998. Genetic polymorphism maintained by natural selection in a temporally varying environment. The American Naturalist: 148-156. Boyce, S.G. 1954. The salt spray community. Ecological Monographs 24: 29-67. Boyd, R.S., and S.N. Martens. 1998. Nickel hyperaccumulation by Thalspi montanum var. montanum (Brassicaceae): a constitutive trait. American Journal of Botany 85(2): 259-265. 1998. Bradshaw, A . D. 1952. Populations of Agrostis tenuis resistant to lead and zinc poisoning. Nature 169: 1098. Burt, R., Fillmore, M . , Wilson, M . A . , Gross, E.R., Langridge, R .W. , and D . A . Lammers. 2001. Soil properties of selected pedons on ultramafic rocks in Klamath Mountains, Oregon. Communications in Soil Science and Plant Analysis 32: 2145-2175. Cartica, R.J . , and J .A. Quinn. 1980. Responses of populations of Solidago sempervirens (Compositae) to salt spray across a barrier beach. American Journal of Botany 67: 1236-1242 Chan, R. 2001a. Taxonomic changes and a new species in Lasthenia sect. Amphiachaenia (Compositae: Heliantheae sensu lato). Madrono 48: 205-210. Chan, R., Baldwin, B . G . and R. Ornduff. 2001b. Goldfields revisited: a molecular phylogenetic perspective on the evolution of Lasthenia (Compositae: Heliantheae sensu lato). International Journal of Plant Sciences 162: 1347-1360. 94 Chan, R., Baldwin, B . G . and R. Ornduff. 2002. Cryptic goldfields: A molecular phylogenetic reinvestigation of Lasthenia californica sensu lato and close relatives (Compositae: Heliantheae sensu lato). American Journal of Botany 89: 1103-1112. Chapin, F.S. 1980. The mineral nutrition of wild plants. Annual Review of Ecology and Systematics 11: 233-260. Chapin, F.S. 1991. Integrated responses of plants to stress. Bioscience 41: 29-36. I Chapin, F.S., Autumn, K . , and F Pugnaire. 1993. Evolution of suites of traits in response to environmental stress. The American Naturalist 142: S78-S92. Chinnusamy, V . , Jagendorf A . , and J -K. Zhu. 2005. Understanding and Improving Salt Tolerance in Plants. Crop Science 45: 437-448 Coyne, J .A. and J .A. Orr. 2004. Speciation. Sinauer Associates, Sunderland. Crain, C M . , Sill iman, B .R. , Bertness, S.L., and M . D . Bertness. 2004. Physical and biotic drivers of plant distribution across estuarine salinity gradients. Ecology 85: 2539-2549. Craine, J . M . 2005. Reconciling plant strategy theories of Grime and Tilman. Journal of Ecology 92,: 1041-1052. Crooke, W . M . 1956. Effect of soil reaction on uptake of nickel from a serpentine soil. Soil Science 81: 269-276. Crooke, W . M . , and R . H . E . Inkson. 1955. The relation between nickel toxicity and major nutrient supply. Plant and Soil 6: 1-15. Debez, A . , Hamed, K . B . , Grignon, C , and C. Abdelly. 2004. Salinity effects on germination, growth, and seed production of the halophyte Cakile maritime. Plant and Soil 262: 179-189. Desrochers, A . M . and B . A . Bohm. 1993. Flavonoid variation in Lasthenia californica. Biochemical Systematics and Ecology 21: 449-453. Desrochers, A . M . and B . A . Bohm. 1995. A biosystematic study of Lasthenia californica (Asteraceae). Systematic Botany 20: 65-84 Desrochers A . M , , and B . Dodge. 2003. Phylogenetic relationships in Lasthenia^ (Heliantheae: Asteraceae) based on nuclear r D N A Internal Transcribed Spacer (ITS) sequence data. Systematic Botany 28: 208-215. 95 Ellis , A . G . , Weis, A . E . , and B.S . Gaut. 2006. Evolutionary radiation of "Stone Plants" in the genus Argyroderma (Aizoaceae): unraveling the effects of landscape, habitat, and flowering time. Evolution 60: 39-55 Ellstrand, N . C . 2003. Current knowledge of gene flow in plants: implications for transgene flow. Philosophical Transactions of the Royal Society of London B 358: 1163-1170. Evans, G .C . 1972. The Quantitative Analysis of Plant Growth. University of California Press, Berkeley. Fernandez, R.J . , and J.F. Reynolds. 2000. Potential growth and drought tolerance of eight desert grasses: lack of a trade-off? Oecologia 123: 90-98. Flowers, T.J. , Hajibagheri, M . A . , and N.J .W. Clipson. 1986. Halophytes. The Quarterly Review of Biology 3: 313-337. Fitter, A . H . and R . K . M . Hay. 2002. Environmental Physiology of Plants. Third Edition. Academic Press Limited, London. Futuyma, D. J., and G . Moreno. 1988. The evolution of ecological specialization. Annual Review of Ecology and Systematics 19: 201-233. Fry, J.D. 2003. Detecting ecological trade-offs using selection experiments. Ecology 84: 1672-1678. Freeman, S. and J.C. Herron. 2001. Evolutionary Analysis. Second Edition. Prentice-Hal l , New Jersey. Freestone, A . L . , and B . D . Inouye. 2006. Dispersal limitation and environmental heterogeneity shape scale-dependent diversity patterns in plant communities. Ecology 87: 2425-2432. Gardner, M . , and M . MacNair . 2000. Factors affecting the co-existence of the serpentine endemic Mimulus nudatus Curran and its presumed projenitor Mimulus guttatus Fisher ex D C . Biological Journal of the Linnean Society 69: 443-459. Goldstein, G . , Drake, D.R. , Alpha, C , Melcher, P., Heraux, J., and A . Azocar. 1996. Growth and photosynthetic responses of Scaevola sericea, A Hawaiian coastal shrub, to substrate salinity and salt spray. International Journal of Plant Sciences 157: 171-179. Gottlieb, L . D . , Warwick, S.I., and V . S . Ford. 1985. Morphological and electrophoretic divergence between Layia discoidea and L. glandulosa. Systematic Botany 10: 484—495. Grant, S.A., and R.F . Hunter. 1962. Ecotypic differentiation of Calluna vulgaris (L.) in relation to altitude. New Phytologist 61: 44-55. 96 Greiner, L a Peyre, M . K . , Grace, J .B. , Hahn, E , and I.A. Mendelssohn. 2001. The importance of competition in regulating plant species abundance along a salinity gradient. Ecology 82: 62-69. Gregory, R .P .G . , and A . D . Bradshaw. 1965. Heavy metal tolerance in populations of Agrostis tenuis Sibth. and other grasses. New Phytologist 64: 131-143. Greenway, H „ and R. Munns. 1980. Mechanisms of salt tolerance in non-halophytes. Annual Review of Plant Physiology 31: 149-190. Griffiths, M . E . , and C M . Orians. 2003. Salt spray differentially affects water status, necrosis, and growth in coastal sandplain heathland species. American Journal of Botany 90: 1188-1196. Grime, P. 1977. Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. American Naturalist 11: 1169-1194. Grime, J.P. 1979. Plant Strategies and Vegetation Processes. John Wiley, New York. Grime, J.P, and R. Hunt. 1975. Relative growth rate: its range and adaptive significance in a local flora. Journal of Ecology 63: 393-422. Hajibagheri, M . A . * Hal l , J .L. , and T.J. Flowers. 1984. Stereological analysis of leaf cells of the halophyte Suaeda maritima (L.) Dum. Journal of Experimental Botany 35: 1547-1557. Halstead, R. L . , 1968. Effect of different amendments on yield and composition of oats and its concentration in plants. Canadian Journal of Soil Science 49: 335-342. Halstead, R . L . , Finn, B .J . , and A . J . MacLean. 1969. Extractability of nickel added to soils and its concentration in plants. Canadian Journal of Soil Science 49: 335-342. Harborne, J .B. 1975. Review. Flavonoid Sulfates: a new class of sulfur compounds in higher plants. Phytochemistry 14: 1147-1155. Harborne, J .B. 1993. Introduction to Ecological Biochemistry, Fourth edition. Academic Press Inc, San Diego. Hodson, M . J . , Smith, M . M . , Wainwright, S.J., and H . Opik. 1981. Cation cotolerance in a salt-tolerant clone of Agrostis stolonifera L . New Phytologist 90: 253-261. Hughes, R., Bachmann, K . , Smirnoff, N . , Macnair, M . R . 2001. The role of drought tolerance in serpentine tolerance in the Mimulus guttatus Fisher ex. D C complex. South African Journal of Science 97: 581-586. 97 Jean-Nicolas, J. and R. Kassen. 2007. On the experimental evolution of specialization and diversity in heterogeneous environments. Ecology Letters 10: 272-281. Karrenberg, S., Edelist, C , Lexer, C , and L . Rieseberg. 2006. Response to salinity in the homoploid hybrid species Helianthus paradoxus and its progenitors H. annuus and H. petiolaris. New Phytologist 170: 615-629. Kassen, R. 2002. The experimental evolution of specialists, generalists, and the maintenance of diversity. Journal of Evolutionary Biology \5: 173-190. K i k , C. 1989. Ecological genetic of salt resistance in the clonal perennial, Agrostis stolonifera L . New Phytologist 113: 453-458. Kohl , K . I . 1997. The effect of N a C l on growth, dry matter allocation and ion uptake in salt marsh and inland populations of Armeria maritima. New Phytologist 135: 213-225. Kruckeberg, A . R . 2002. Geology and plant life: the effects oflandforms and rock types on plants. University of Washington Press, Seattle. Kruckeberg, A . R . 1986. A n essay: the stimulus of unusual geologies for plant speciation. Systematic Botany 11:455-463. Kruckeberg, A . R . 1984. California Serpentines: Flora, Vegetation, Geology, Soils and Management Problems. University of California Press, Berkeley. Kruckeberg, A . R . 1954. Plant species in relation to serpentine soils. Ecology 35: 267-274. Lambers H . and H . Poorter. 1992. Inherent variation in growth rate between higher plants: a search for physiological causes and ecological consequences. Advances in Ecological Research 23: 187-261. Levin, D . A . 2001. The recurrent origin of Plant Races and Species. Systematic Botany 26: 197-204. Lynch, M . and B . Walsh. 1998. Genetics and Analysis of Quantitative Traits. Sinauer Associates, Sunderland. Maathuis, F.J. , and A . Amtmann. 1999. K + nutrition and N a + toxicity: the basis of cellular K + / N a + ratios. Annals of Botany 84: 123-133. MacDougall , A . S . and R. Turkington. 2004. Relative importance of suppression-based and tolerance-based competition in an invaded oak savanna. Journal of Ecology 92: 422-434. 98 Madhok, O P . , and R . B . , Walker. 1969. Magnesium nutrition of two species of sunflower. Plant Physiology 44: 1016-1022. Main, J .L. 1974. Different responses to magnesium and calcium by native populations of Agropyron spicatum. American Journal of Botany 61: 931-937. Marrs, R . H . and J. Proctor. 1976. The response of serpentine and non-serpentine Agrostis stolonifera L . to magnesium and calcium. Journal of Ecology 64: 953-964. Maynard Smith, J. , Burian, R, Kauffman, S., Alberch, P., Campbell, J., Goodwin, B . , Lande R., Raup, D. , and L . Wolpert. 1985. Developmental constraints and evolution. The Quarterly Review of Biology 60: 265-287 Mitchell-Olds, T. 1996. Pleiotropy causes long-term genetic constraints on life-history evolution in Brassica rapa. Evolution 50: 1849-1858. Morjan, C. L . and L . H . Rieseberg. 2004. How species evolve collectively: implications of gene flow and selection for the spread of advantageous alleles. Molecular Ecologyl3: 1341-1356. Munns, R. 2002. Comparative physiology of salt and water stress. Plant, Cell and Environment 25: 239-250. Murren, C.J . , Douglass, L . , Gibson, A . , and M . R . Dudash. 2006. Individual and combined effects of C a / M g ratio and water on trait expression in Mimulus guttatus. Ecology 87: 2591-2602. Nagy, E.S., and K . J . Rice. 1997. Local adaptation in two subspecies of an annual plant: Implications for migration and gene flow. Evolution 51: 1079-1089. Nagy, L . , and J. Proctor. 1997. Plant growth and reproduction on a toxic alpine ultramafic soil: adaptation to nutrient limitation. New Phytologist 137: 267-274. Niu , X . , Bressan, R . A . , Hasegawa, P . M . , Pardo, J . M . 1995. Ion homeostasis in NaCl stress environments. Plant Physiology 109: 735-742. O'Dell , R .E . , and V . P . Claassen. 2006. Serpentine and nonserpentine Achillea millefolium accessions differ in serpentine substrate tolerance and response to organic and inorganic amendments. Plant and Soil 279:253-269. O'Dell , R., James, J.J., and J .H . Richards. 2006. Congeneric serpentine and nonserpentine shrubs differ more in leaf C a : M g than in tolerance of low N , low P, or heavy metals. Plant and Soil 280:49-64. Orians, G . H . , and O.T. Solbrig. 1977. A cost-income model of leaves and roots with special refrence to arid and semiarid areas. American Naturalist 111: 677-690 99 Oosting, H.J . 1945. Tolerance to salt spray of plants of coastal dunes. Ecology 26: 85-89. Ornduff, R. 1966. A biosystmatic survey of the Goldfield genus Lasthenia (Compositae: Helenieae). University of California Publications in Botany 40: 1-92. Ornduff, R. 1971. A new tetraploid subspecies of Lasthenia (Compositae) from Oregon. Madrono 21: 96-98. Ornduff, R. 1993. Lasthenia. In: Hickman, J.C., ed. The Jepson Manual: Higher plants of California. University of California Press, Berkeley, California. 298-300. Poorter, H . , and C . Remkes. 1990. Leaf area ratio and net assimilation rate of 24 species differing in relative growth rate. Oecologia 83: 553-559. Pepper A E , Norwood L E . 2001. Evolution of Caulanthus amplexicaulis yar. barbarae (Brassicaceae), a rare serpentine endemic plant: a molecular phylogenetic perspective. American Journal of Botany 88: 1479-1489 Proctor, J. 1971a. The plant ecology of serpentine. II Plant response to serpentine soils. Journal of Ecology 59: 397-410. Proctor J. 1971b. The plant ecology of serpentine. III. The influence of a high calcium/magnesium ratio and high nickel and chromium levels in some British and Swedish serpentine soils. Journal of Ecology 59: 827-842 Proctor, J., and S.R.J. Woodell . 1975. The ecology of serpentine soil. Advances in Ecological Research 9: 255-366. Proctor, J., Johnston, W.R. , Cottam, D . A . , and A . B . Wilson. 1981. Field capacity water extracts from serpentine soils. Nature 294: 245-248. Quinn, G.P. and M . J . Keough. 2002. Experimental Design and Analysis for Biologists, Cambridge. Rajakaruna, N . 2002. Evolution and differentiation of edaphic races in the Lasthenia californica complex (Asteraceae: Heliantheae). PhD thesis. University of British Columbia. Vancouver, B . C . , Canada. Rajakaruna, N . and B . A . Bohm. 1999. The edaphic factor and patterns of variation in Lasthenia californica (Asteraceae). American Journal of Botany 86:1576-1596 100 Rajakaruna, N . , Baldwin, B . G . , Chan, R., Desrochers, A . M . , Bohm, B . A . and J. Whitton. 2003a. Edaphic races and phylogenetic taxa in the Lasthenia californica complex (Asteraceae: Heliantheae): an hypothesis of parallel evolution. Molecular Ecology 12: 1675-1679. Rajakaruna, N . , Bradfield, G .E . , Bohm, B . A . and J. Whitton. 2003b. Adaptive differentiation in response to water stress by edaphic races of Lasthenia californica (Asteraceae). International Journal of Plant Sciences 164: 371-76. Rajakaruna, N . , Siddiqi, M . Y . , Whitton, J., Bohm, B . A . and A . D . M . Glass. 2003c. Differential responses to N a + / K + and C a 2 + / M g 2 + in two edaphic races of the Lasthenia californica complex (Asteraceae): A case for parallel evolution of physiological traits. New Phytologist 157: 93-103. Raven, R H . 1964. Catastrophic selection and edaphic endemism. Evolution 18: 336-338. Reeves, R .D. , McFarlane, R . M - , and R.R. Brooks. 1983. Accumulation of nickel and zinc by western North America genera containing serpentine-tolerant species. American JournalofBotany 70: 1297-1303. Reeves, R .D . , Baker, A . J . M . , Borhidi, A . , and R. Berazain. 1999. Nickel hyperaccumulation in the serpentine flora of Cuba. Annals of Botany 83: 29-38. Reimchen, T .E . , and P. Nosi l i . Temporal variation in divergent selection on spine number in threespine stickleback. 2002. Evolution 56: 2471-2483. Reimann, C , and S.W. Breckle. 1995. Salt tolerance and ion relations of Salsola kali L . - differences between ssp. tragus (L.) Nyman and ssp. ruthenica (Iljin) Soo. New Phytologist 130: 37-45. Reynolds, H . L . , Hungate, B . A . , Chapin, F.S., and C M . DAntonio . 1997. Soil heterogeneity and plant competition in an annual grassland. Ecology 78: 2076-2090. Rieseberg, L . H . , and J . M . Burke. 2001. The biological reality ofspecies: gene flow, selection, and collective evolution. Taxon 50:47-67. Rice, W.R. , and G . Salt. 1990. The evolution of reproductive isolation as a correlated character under sympatric conditions: experimental evidence. Evolution AA: 1140-1152. Roach, D . A . , and R . D . Wulff. 1987. Maternal effects in plants. Annual Review of Ecology and Systematics 18: 209-235. Robertson, A . I . 1985. The poisoning of roots of Zea mays by nickel ions, and the protection afforded by magnesium and calcium. New Phytologist 100: 173-189. 101 Robichaux, R . H . , Carr, G .D . , Liebman, M . , and R .W. Pearcy. 1990. Adaptive radiation of the Hawaiian Silversword Alliance (Compositae- Madiinae): ecological, morphological, and physiological diversity. Annals of the Missouri Botanical Garden 77: 64-72. Rossiter, M . C . 1996. Incidence and consequences of inherited environmental effects. Annual Review of Ecology and.Systematics 27: 451-476. Rozema, J., Rozema-Dijst, E . , Friesen, A . H . J , and J.J.L. Huber. 1978. Population differentiation within Festuca rubra L . with regard to soil salinity and soil water. Oecologia 34: 329-341. Sambatti, J . B . M . , and K . J . Rice. 2006. Local adaptation, patterns of selection, and gene flow in the Californian serpentine sunflower (Helianthus exilis). Evolution 60: 696-710. S A S Institute. 2000. J M P I N Version 4.04. Cary, N C , U S A : S A S Institute Inc Schat, H . , Vooijs, R., and E. Kuiper. 1996. Identical major gene loci for heavy metal tolerances that have independently evolved in different local populations and subspecies of Silene vulgaris. Evolution 50: 1888-1895. Schemske, D .W. , and P. Bierzychudek. 2001. Perspective: evolution of flower color in the desert annual Linanthus parryae: Wright revisted. Evolution 55: 1269-1282. Schemske, D .W. , and H . D . Bradshaw. 1999. Pollinator preference and the evolution of floral traits in monkeyflowers (Mimulus). Proceedings of the National Academy of Sciences (USA) 96: 11910-11915 Schluter, D . 1993. Adaptive radiation in sticklebacks: size, shape and habitat use efficiency. Ecology 74: 699-709. Schluter, D . , Clifford, E . A . , Nemethy, M . , and J.S. McKinnon . 2004. Parallel evolution and inheritance of quantitative traits. The American Naturalist 163: 809-822. Schluter, D . 1996a. Adaptive radiation along genetic lines of least resistance. Evolution 50: 1766-1774. Schluter, D . 1996b. The ecological causes of adaptive radiation. The American Naturalist 148: S40-S64 Schmitt, J. 1993. Reaction norms of morphological and life history traits to light availability in Impatiens capensis. Evolution 47:1654—1668. Seddon, G . 1974. Xerophytes, xeromorphs and sclerophylls: the history of some concepts in ecology. Biological Journal of the Linnean Society 6: 65-87. 1 0 2 Silander, J. A . , and J. Antonovics. 1979. The genetic basis of the ecological amplitude of Spartinapatens. I. Morphometric and physiological traits. Evolution 33: 1114-1127. Sokal, R.R. and F.J. Rohlf. 1995. Biometry. Third edition. W . H . Freeman, New York. SPSS. 1999. Sigmascan Pro Image Analysis, Version 5.0.0. Chicago, IL: SPSS. Sykes, M . T . and J .B. Wilson. 1988. A n experimental investigation into the response of some New Zealand sand dune species to salt spray. Annals of Botany 62: 159-166. Sykes, M . T . and J .B. Wilson. 1990. A n experimental investigation into the response of New Zealand sand dune species to different depths of burial by sand. Acta Botanica Neerlandica 39: 171-181. Sykes, M . T . and J .B. Wilson. 1999. Is zonation on coastal sand dunes determined primarily by sand burial or by salt spray? A test in New Zealand dunes. Ecology Letters 2: 233-236. S Y S T A T , 2001. SYSTAT version 10. SPSS Science. Chicago, IL. Tabachnick, B . and L . Fidell . 2001. Using Multivariate Statistics, fourth edition. A l l y n and Bacon, Needham Heights Tenhunen, J.D., Catarino, F . M . , Lange, O.L. , and W . C . Oechel. 1985. Plant Response to Stress - Functional Analysis in Mediterranean Ecosystems. Series G: Ecological Sciences Vol. 15. Springer-Verlag, New York. Tester, M . and R. Davenport. 2003. N a + tolerance and N a + transport in higher plants. Annals of Botany 91: 503-527. Tilman, D. 1988. Plant Strategies and The Dynamics and Structure of Plant Communities. Princeton University Press, Princeton. Tiku B . L . and R . W . Snaydon. 1971. Salinity tolerance within the grass species Agrostis stolonifera L . Plant and Soil 35: 421-431. Tomas-Barberan, F . A . , Harborne, J .B. , and R. Self. 1987. Twelve 6-oxygenated flavone sulfates from Lippia nodiflora and L. canescens. Phytochemistry 26: 2281-2284. Turesson, G . 1922a. The genotypical response of the plant species to the habitat. Hereditas 3:211-350. Turesson, G . 1922b. The species and the variety as ecological units. Hereditas 3: 100-113. Van Der Valk , A . G . 1974. Environmental factors controlling the distribution of forbs 103 on coastal foredunes in Cape Hatteras National Seashore. Canadian Journal of Botany 52: 1057-1073 Walker, R . B . 1954. Factors affecting plant growth on serpentine soils. Ecology 35: 259-266. Watt, W . A . 1983. The effects of salt water and soil type upon the germination, establishment and vegetative growth of Holcus lanatus L . and Lolium perenne L . New Phytologist 942: 275-291. Weeks, A . , and A . A . Hoffmann. 1998. Intense Selection of Mite Clones in a Heterogeneous Environment. Evolution 52: 1325-1333. Welden, C . W . and W . L . Slauson. 1986. The intensity of competition versus its importance: an overlooked distinction and some implications. The Quarterly Review of Biology 61: 23-44. Wellborn, G . A . , Skelly, D . K . , and E .E . Werner. 1996. Mechanisms creating community structure across a freshwater habitat gradient. Annual Review of Ecology and Systematics 27: 337-63. Wells, B .W. , and I .V. Shunk. 1938. Salt spray: an important factor in coastal ecology. Bulletin of the Torrey Botanical Club 65: 485-492 White, T.J. , Bruns, R., Lee, S., and J. Taylor. 1990. Amplification and direct sequencing of fungal ribosomal genes for phylogenetics. In PCR Protocols: A Guide to Methods and Applications, Volume 4. Eds. Innis, M . , Gefand, D. , Sninsky, J. , and T. White. Academic Press, San Diego. Wu, L . 1981. The potential for evolution of salinity tolerance in Agrostis stolonifera L . and Agrostis tenuis Sibth. New Phytologist 89: 471-486. Wu, L . , Thurman, D . A . , and A . D . Bradshaw. 1975. The potential for evolution of heavy metal tolerance in plants. III. The rapid evolution of copper tolerance in Agrostis stolonifera. Heredity 34: 165-187. Yeo, A . R . 1998. Molecular biology of salt tolerance in the context of whole-plant physiology. Journal of Experimental Botany 49: 915-929. Yeo, A . R . , and T.J . Flowers. 2004. Salt tolerance in the halophyte Suaeda maritime L . Dum: Evaluation of the effect of salinity upon growth. Journal of Experimental Botany 31:1171-1183. 104 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0100864/manifest

Comment

Related Items