Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Genetic variation and ecological differentiation between two southern Utah endemics : U.S. federally… Jennings, Linda M. 2005

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

Item Metadata


831-ubc_2005-0229.pdf [ 5.75MB ]
JSON: 831-1.0092079.json
JSON-LD: 831-1.0092079-ld.json
RDF/XML (Pretty): 831-1.0092079-rdf.xml
RDF/JSON: 831-1.0092079-rdf.json
Turtle: 831-1.0092079-turtle.txt
N-Triples: 831-1.0092079-rdf-ntriples.txt
Original Record: 831-1.0092079-source.json
Full Text

Full Text

Genetic variation and ecological differentiation between two southern Utah endemics: U.S. federally threatened Townsendia aprica and a closely related congener, T. jonesii var. lutea (Asteraceae: Astereae) by Linda Jennings B.Sc, The University of British Columbia, 2000 A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Science in The Faculty of Graduate Studies Botany . . ^ i v r The University of British Columbia March 2005 © Linda Jennings, 2005 Abst rac t Townsendia aprica Welsh & Reveal (Asteraceae) is one of 17 rare plants in Utah receiving protection under the U. S. Federal Endangered Species Act. Townsendia aprica (last chance Townsendia) is an herbaceous perennial, narrow endemic, occurring along the Colorado Plateau in southeastern Utah. Recent discoveries have increased its known range from roughly 301 square kilometers at the time it was first listed to 4186 square kilometers. Many of these populations comprise fewer than 100 individuals. It was presumed that small population sizes together with geographical isolation would lead to low levels of genetic variation and inbreeding within the different populations, as a result of genetic drift and restricted gene dispersal. With the use of enzyme electrophoresis, nine informative loci were resolved to gain insight into the level and pattern of genetic variation within T. apnea's range. These patterns were compared with those of another narrow endemic that is a close congener, Townsendia jonesii var. lutea. Analysis of inter- and intra-taxon genetic distances for T. aprica and T. jonesii var. lutea provides important insights into the consequences of rarity, as well as into the genetic and ecological distinctness of the two narrowly endemic taxa. The two taxa have very similar, relatively high values estimated for genetic variation parameters based on isozyme analysis of nine loci (%P > 70, A > 2.0, and H e > 0.30), with much of the genetic variation being found within populations (FST = 0.060 for T. jonesii var. lutea and FST = 0.100 for T. aprica). No significant inbreeding found was in most populations of either taxon. Forty percent of the variation in genetic distance between populations of T. jonesii var. lutea could be explained by geographical distance, where as only about 10% of genetic variation in T. aprica's could be explained by physical distance. Bottlenecks were detected in more than half of each taxon's populations. The results also reveal that populations of each taxon are more similar to one another (FPT = 0.14), than to populations of the other taxon (FST = 0.21). Significant differences were found in characteristics of the soils on which the two taxa occur, though so far, all measurable morphological characters are overlapping in these two taxa. T A B L E O F C O N T E N T S ABSTRACT II TABLE OF CONTENTS IV LIST OF FIGURES ; V LIST OF TABLES VI ACKNOWLEDGEMENTS VII CHAPTER 1 INTRODUCTION 1 1.1 RARITY, ENDEMISM AND U T A H 1 1.2 CONSERVATION GENETICS , 4 1.3 TOWNSENDIA : 7 1.4 TOWNSENDIA APRICA AND T. JONESII WAR. LUTEA 9 1.5 OBJECTIVES 15 CHAPTER 2 MATERIALS AND METHODS 16 2.1 F IELD SAMPLING 16 2.2 L A B METHODS 21 2.2.1 Pollen size andstainability 21 2.2.2 Electrophoretic procedures 21 2.2.3 Soil procedure 23 2.2.4 Pappus measurement procedure 23 2.3 ANALYSIS 24 2.3.1 Isozyme analysis 24 2.3.2 Genetic analysis between T. aprica and T.jonesii var. lutea 28 2.3.3 Soil analysis 28 CHAPTER 3 RESULTS 28 3.1 BREEDING SYSTEM IN TOWNSENDIA APRICA 2 8 3.2 GENETIC DIVERSITY IN TOWNSENDIA APRICA AND T. JONESII VAR. LUTEA 3 0 3.3 RELATEDNESS OF TOWNSENDIA APRICA AND T. JONESII VAR. LUTEA 3 9 CHAPTER 4 DISCUSSION 44 4.1 BREEDING SYSTEM 45 4.2 LEVELS AND PATTERNS OF GENETIC DIVERSITY 48 4.3 RELATEDNESS CONCLUSIONS 55 4.4 FUTURE DIRECTIONS 57 4.5 CONSERVATION CONSIDERATIONS 58 LITERATURE CITED 60 APPENDIX I TOWNSENDIA APRICA HERBARIUM VOUCHERS EXAMINED 67 APPENDIX H GEOGRAPHICAL DISTANCE 69 APPENDIX III ISOZYME BUFFER RECD7ES 70 APPENDIX IV ALLELE FREQUENCIES 71 APPENDIX V REPRESENTATIVE PHOTOS OF EACH TAXON 73 iv L I S T O F F I G U R E S Figure 1.1. Phylogenetic relationships of the species of Townsendia and a summary of habit and the presence of apomixis in the species. From Beaman, 1957 8 Figure 1.2. Map of Utah showing the geographic range of T. aprica and T.jonesii var. lutea in relation to the northwest boundary of the Colorado Plateau 10 Figure 2.1. Townsendia aprica and T.jonesii var. lutea sites collected in Utah in 2001 and 2002. 17 Figure 3.1. Neighbor-joining tree using Cavalli-Sforza and Edwards chord distances for T. aprica 36 Figure 3.2. Neighbor-joining tree using Cavalli-Sforza and Edwards chord distances for T. jonesii var. lutea 37 Figure 3.3. Graph of regression analysis between pairwise FST and geographical distance between populations for T. aprica 38 Figure 3.4. Graph of regression analysis between pairwise FST and geographical distance between populations for T.jonesii var. lutea 38 Figure 3.5. Neighbor-joining tree using Nei's distances (1978) between T. aprica (1-13) and T. jonesii var. lutea (14-20) 41 v L I S T O F T A B L E S Table 2.1. Townsendia aprica localities and collection information for this study, along with additional information available for each population 19 Table 2.2. Townsendia jonesii var. lutea localities and collection information for this study, along with additional information available for each population 20 Table 2.3. Electrode buffer systems including modifications in p H and run time 22 Table 2.4. Resolved enzymes with each buffer system used 22 Table 3.1. Estimated pollen stainability of Townsendia aprica individuals 29 Table 3.2. Enzymes and loci number resolved 30 Table 3.3. Genetic variability parameters for populations o f Townsendia aprica 32 Table 3.4. Genetic variability parameters for populations of Townsendia jonesii var. lutea 32 Table 3.5. Wright's F-statistics (Fis, FIT, FST) for all isozyme loci examined and their means for T. aprica and T. jonesii var. lutea, respectively 33 Table 3.6. Cavalli-Sforza and Edwards (1967) chord distance pairwise comparisons o f 13 populations o f 71 aprica 35 Table 3.7. Cavalli-Sforza and Edwards (1967) chord distance pairwise comparisons of 7 populations o f T. jonesii var. lutea 35 Table 3.8. Nei ' s genetic distance (1972) (below diagonal) and identity (above diagonal) for all pairwise comparisons between T. aprica (1-13) and T. jonesii var. lutea (14-20) 40 Table 3.9. Soil characteristics comparing T. aprica ( N = 23) and T. jonesii var. lutea ( N = 7). ..42 Table 3.10. Ray and disk pappus measurements of T. aprica and T. jonesii var. lutea. 43 Table 1.1. A l l known herbarium vouchers for Townsendia aprica 67 Table 1.2. List populations, sites and herbaria and their acronyms 68 Table II. 1. Geographic distances (km) between the different sampled populations o f T. aprica (1-13) and T. jonesii var. lutea (14-20) 69 Table II.2. Geographic distances (km) between the different sampled populations T. jonesii var. lutea (14-20) 69 Table III. 1. Running buffer recipes used for both T. aprica and T. jonesii var. lutea 70 Table III.2. Extraction buffer recipe used for leaf tissue o f T. aprica and T. jonesii var. lutea.... 70 Table IV .1 . Allele frequencies of Townsendia aprica 71 Table IV.2 . Allele frequencies of Townsendia jonesii var. lutea 71 v i A C K N O W L E D G E M E N T S Without a doubt the first person who I am totally indebted to is my supervisor, Jeannette Whitton (a.k.a. The Gambler). She is someone who took a big chance with me, to allow me to utilize my drive to the fullest and gave me the support (money and emotional) needed to make this project and my experiences the best I could possibly hope for. She is a one of a kind "rare" supervisor and I will be forever thankful for this experience with her. I also want to thank my committee members, Sally Aitken and Fred Ganders for taking such an interest in this project. I want to thank Sally for giving her time to look over my analysis and results with a fine tooth comb, asking the critical questions, and to pushing me those extra steps to better understand rare species and population genetics. I want to thank Fred Ganders for taking time to look over my data early on and for helping me to understand how closely linked breeding systems and population genetics really are, and for encouraging me to put all this data into the larger context of speciation. I also want to thank the Whitton lab crew: Gina Choe for her crazy sense of humor, and support over dinner and beer and Chris Sears for taking an interest in T. aprica by continuing on-going studies of this special little plant. I would also like to thank Debbie Clark for spending time with me in the field showing me localities and being the "go to gal" when I needed information from the different US federal agencies about T. aprica. I would also like to add to this list Tom Clark (NPS), Bob Campbell (USFS), Lori Armstrong (BLM), and Larry England (USF&W) for giving me the permits to allow this project to happen. My personal thanks goes out to my mom and sister (Theresa), who call me and write me on a weekly basis to see how I'm doing, if I need anything, and to remind me of how proud they are of me. I also must include my immediate family: Timmie and Ophilia, my cats, without whom I might need an alarm clock and Beckett and Woofie, my dogs, who force me to take "walkies" or pay the consequences. I also must thank my two best friends, Bahzad and Netza for not allowing me to take myself too seriously and reminding me how important laughing really is in life. And most of all, with all of my heart, my biggest thanks goes to Michael Lipsen, a person who has impacted my life so much in the past eleven years, that I can sometimes barely recognize myself. Not only is he a great scientist, editor, and field assistant, but his love and encouragement of this project and my life have made the past eleven years the best eleven years. I can't wait to see what is to come. C h a p t e r 1 In t roduct ion 1.1 Rarity, Endemism and Utah Roughly 250,000 plant species are thought to have existed in historical times, with almost 1000 becoming extinct in the last century, and there are estimates that sixty times that many could go extinct in the next fifty years (Raven, 1987). The International Union for Conservation for Nature and Natural Resources (IUCN) has been assessing the conservation status of species, subspecies and varieties on a global scale for the past four decades. The IUCN Red List of Threatened Plants has documented a 40% increase in the past decade of threatened plant species in North America (Walter and Gillett, 1998; Species, 2004). Utah's flora ranks as one of the top five among continental states in abundance of rare plant species and has one of the highest levels of endemism in North America (Welsh et al., 1975). It has been estimated that 225 to 240 plant species are endemic in Utah, with 200 considered threatened (Welsh et al., 1975; Shultz, 1993). Townsendia aprica Welsh & Reveal is one of 17 rare plant species in Utah that has been listed and given protection under the U. S. Federal Endangered Species Act. More than 10% of the Utah's flora is endemic to the state, with the southeastern part of Utah, occupied by the Colorado Plateau, containing 50% of Utah's endemic flora (Shultz, 1993). It is speculated that the majority of these rare and endemic species are concentrated in this region of Utah because species have diverged and speciated in response to environmental stresses in this arid desert habitat (Shultz, 1993). Stebbins (1952) concluded that topography and soil features will have a greater effect on the character of vegetation where moisture is limiting, when compared to where water is in adequate supply. Shultz (1993) also pointed to similar factors to explain much of the diversity and the high occurrence of endemics found in Utah. Specifically, she points to habitat diversity of the Colorado Plateau region, with its islands of unique soil types, the severely arid climate, averaging 115-230 mm rain per year, and what is described as 1 the occurrence of contact zones between six phytogeographical regions, including that between the Great Basin and the Colorado Plateau which runs directly through the middle of the state. There are many ways to define rarity. Specifically, rarity can describe a situation where few individuals are found, where a species is found in a very localized, area, or where the species only occurs in a rare habitat (Rabinowitz et al., 1986). Species may be rare either because they have a historically small range, or rare species may be remnants of much more widespread species (Welsh et al., 1975). In either case, threatened species are somehow not able to expand their numbers establishing new populations beyond current boundaries with possibly even declining numbers. Failure to expand beyond the current range could be due to limited seed dispersal, or lack of appropriate genetic variation to allow adaptation to new environments, perhaps because of a recent bottleneck. Holsinger and Gottlieb report that the average rare and endangered species is usually defined as having few (1000-5000) total individuals, fewer than five populations, with many populations reduced by habitat loss (Fiedler, 1987), and inability to inhabit neighboring ecosystems (Drury, 1974). Drury (1974) gives the definition of a rare species as having small populations that are widely separated with interbreeding being seriously reduced between the subpopulations, or restricted to one population. Drury also explains that there are two types of extreme adaptive strategies seen commonly among plants and animals, either having a competitive or physiological means of advantage. The competitive strategy can be defined as the capacity for successful competition against other species at the sacrifice of the ability to handle extreme habitats. In the physiological strategy, individuals are able to withstand extreme habitat conditions at the expense of a competitive edge with other species under more favorable conditions. Because of small range size and low numbers of populations and individuals, rare species are potentially at greater risk of extinction should they become threatened. The International 2 Union for the Conservation of Nature and Natural Resources (IUCN) Red Book (Nature, 1966) defines a threatened species as still abundant in parts of its range, but in threat due to declining numbers. Threats may include new competition (e.g. with non-native species), new pathogen or pest, habitat disturbance or loss (natural or anthropogenic), loss of pollinators or seed dispersers, small population size (<100), and low or fixed genetic diversity. Some of these factors have been identified as possible threats to the persistence of T. aprica populations. Because endemic species already have the physiological mechanisms in place to survive in Utah's desert conditions, most do not have to deal with competition of non-native species, which do not tend to persist under these conditions (Welsh et al., 1975; Shultz, 1993). Although there have been no studies of new pathogens or pests in T. aprica, there have also been no observations to suggest that these currently pose a threat to T. aprica. Habitat disturbance, both natural and anthropogenic, has been a noted factor in different reports about T. aprica (Welsh, 1978; Service, 1985; U. S. Fish and Wildlife Service, 1993; College, 1994; Robinson, 2002, 2003) and many of the other endemic species of Utah (Welsh et al., 1975). Specifically, industrial (mining), agricultural (cattle), and recreational activities have been noted as being a threat to T. aprica's long term survival. Fluctuations in population size have been observed in this species, with some populations extirpated, have been attributed to environmental fluctuations, mainly in precipitation (Armstrong and Thorne, 1991; Robinson, 2002, 2003). And though no data have been collected, it has been suggested that T. aprica may have difficulty attracting density-dependent, nutrient-demanding pollinators, like bees (Tepedino et al., 2004). Townsendia aprica is known to have many populations with fewer than 100 individuals, which fluctuate in size. The effects of small population size on genetic variation have not been specifically been examined in this species. 3 1.2 Conservation Genetics It is widely accepted that demographic instability is the most important factor affecting the long term survivorship of a rare species (Lande and Shannon, 1996). Because of the limited resources available for protecting federally endangered species, the value of conservation genetic data has been questioned in the past, but more recent reviews have shown how conservation genetics can be very useful in understanding rare species and improving management strategies for conserving a species (Ellstrand and Elam, 1993; Holsinger et al., 1999; Soltis and Gitzendanner, 1999; Amos and Balmford, 2001). Holsinger et al: (1999) explained how genetic assimilation via hybridization can pose a threat to a species' long-term persistence if there has been a critical loss of self-incompatibility alleles. Soltis and Gitzendanner (1999) agree that genetics can be used to detect the effects of hybridization on a rare species, but also argue that molecular systematics and population genetics can be used for clarification of species boundaries and identification of important lineages. Ellstrand and Elam (1993) show that small populations of rare plant species are at genetic risk from the effects of genetic drift and inbreeding. The risks of lower genetic variability, both through lowering fitness by inbreeding depression and through the loss of adaptive potential in the event of a severe bottleneck, have also been noted (Amos and Balmford, 2001). Population size, the degree of isolation by pollination and seed dispersal mechanisms, geographic isolation, and bottlenecks are all traits that can affect the genetic diversity found within populations of a species (Richardson, 1986). Gene flow limitations can occur if a pollinator declines or is lost, due to human activities such as pesticide use or even a loss of secondary plant species facilitating the pollinator interaction (Thomson, 1978; Tepedino, 1979; Rathcke, 1983; Sipes and Tepedino, 1993). As Wright (1931) showed, the effect of random genetic drift and fixation of alleles occurs more quickly in smaller populations (less than 100 4 individuals). Stebbins (1942) also discussed the correlation between genetic variation within populations and the rarity or commonness of a plant species. He suggested that rare plants are genetically depauperate, though he later qualified this idea in his 1980 paper, citing examples o f highly genetically variable rare species, and widespread species with low genetic variation (Stebbins, 1980; Gitzendanner and Soltis, 2000). Even so, many papers comparing average genetic variability in rare and widespread species have shown lower genetic variation in rare or endemic plants when compared with widespread species (Hamrick and Godt, 1989; Hamrick et al., 1991; Gitzendanner and Soltis, 2000; Cole, 2003). To better understand patterns of genetic variation in widespread and rare taxa, the current emphasis is on comparing close relatives, so that factors such as breeding system, dispersal mechanisms and phylogeny can be taken into account when comparing the two types of species (Karron, 1987; Gitzendanner and Soltis, 2000). Some more recent studies have concentrated on comparing genetic diversity o f closely related endemics (Lopez-Pujol et al., 2001; McCraken, 2001) as I have done here. Conservation studies are increasingly using genetics to better understand patterns of variation in endemic, rare or endangered plant species, which has helped inform government agency resource managers so that limited resources are used wisely. Two recent examples are studies o f the narrow endemic species, Artemisia molinieri (Asteraceae) in southwestern Europe (Torrell et al., 1999) and one of the only known population o f Eriogonum ovalifolium var. williamsiae (Polygonaceae) in Nevada (Archibald et al., 2001). In both studies, the species were found to have high genetic variability when compared to typical values for endemic species presented in Hamrick and Godt (1989). Because genetic diversity is often associated with long term fitness and adaptability of a species, finding high genetic variability in an endemic, rare or federally listed species w i l l lead researchers to conclude that the species is not at genetic risk, as concluded in the above two cases. This enables management to concentrate resources and efforts 5 toward protecting the species in other ways such as limiting habitat destruction and increasing population sizes with translocations. As pointed out by Holsinger and Gottlieb (1991), a key priority and first step in conservation of a rare species is knowledge of its breeding system. The first step towards understanding variation within T. aprica was to determine whether all populations are diploid, and have an outcrossing breeding system. This outcome would affect the assumptions of Hardy-Weinberg equilibrium (HWE), and would potentially affect the conservation status of populations because apomictic species are not necessarily eligible for federal protection under the US Endangered Species Act. With the use of isozyme electrophoresis, it is possible to detect recent polyploidy using observations of duplicate loci and different staining intensities of banding patterns (Gottlieb, 1982; Wendel and Weeden, 1989; Acquaah, 1992). It is also possible to detect an apomictic breeding system by observing fixed differences between homologous loci, so called "fixed heterozygosity" (Soltis and Rieseberg, 1986; Noyes and Soltis, 1996; Richards, 1997; Brown and Young, 2000). In Erigeron compositus (Noyes and Soltis, 1996) many individuals displayed identical multilocus genotypes, where a single heterozygous enzyme phenotype was found in most or all members of an apomictic population, while in diploid populations both alleles segregated and were found in heterozygous or homozygous states. With the use of isozymes, I conducted an investigation into the distribution and patterns of genetic variation found within the geographical range of T. aprica. The results of this study will enable the United States Fish and Wildlife Service to manage the individual populations more carefully by taking into account which, if any, populations are apomictic, which populations are exchanging genetic information, and with so many small T. aprica populations, which ones would need to be monitored more closely if inbreeding is observed. This genetic information will also be useful in the implementation of any recovery plan for the translocation of T. aprica individuals into existing populations and for establishment of any new populations. 6 1.3 Townsendia Townsendia was first described with one species, T. sericea, by William Jackson Hooker in 1834 (Larsen, 1927). The genus currently consists of about 30 taxa that occur in western North America, from the southern Yukon to central Mexico (Beaman, 1957; Welsh and Reveal, 1968). Utah represents an area of very high diversity for Townsendia, with sixteen native taxa, four of which are endemic to the state (Welsh, 1983; Welsh et al., 1987). John Beaman published a monograph of Townsendia (1957) that established much of the current taxonomic framework for the genus, and in which he shared his plethora of knowledge of Townsendia's breeding system, distribution, morphology, and proposed relationships of the taxa. Two features of the genus are particularly relevant to this thesis. First, Townsendia includes both sexual and asexual populations, with no apomictic diploids. Asexual reproduction in Townsendia occurs by diplosporous apomixis, in which the embryo develops parthenogenetically from an unreduced egg cell. Sexual diploids and asexual polyploids have been found in different populations of nearly half of the species of Townsendia, with the remaining species only exhibiting sexual diploids. Diploid Townsendia have 2n=2x=18 and apomictic plants are mostly tetraploids but some triploids and aneuploids are also found (2n = 27-36) (Beaman, 1954, 1957); Windham, unpublished). Based on morphology and geographic distribution patterns, Beaman constructed an informal phylogenetic tree for Townsendia (Figure 1.1) and pointed out that apomictic taxa tend to be acaulescent (lacking a well developed stem, i.e. growing as a dense tuft of leaves), long-lived perennials. He also states that "without exception, the apomicts of Townsendia are concentrated at the higher elevations and higher latitudes relative to the total range of the species in which they occur" (Beaman 1957, p. 51). 7 Z APQMtJCiS WFB£9V£.'tT 0. APQViXiS WItHQWN Figure 1.1. Phylogenetic relationships of the species of Townsendia and a summary of habit and the presence of apomixis in the species. From Beaman, 1957. The second feature worth noting is that Beaman and others (Welsh et al., 1975; Welsh et al., 1987; Shultz, 1993) have stated that many taxa within Townsendia appear to be restricted to specific soil types. For example, T. smithii and T. gypsophila are restricted to soils that contain substantial amounts of gypsum (calcium sulfate). Beaman speculated that patterns of speciation in the genus have likely been influenced by edaphic factors, as well as by geographic barriers and climatic stresses. Narrow endemism is a common feature of species within Townsendia, and Beaman considered only three species to be widespread geographically (T excapa, T.florifer, T. incana). Beaman also postulated a relatively recent divergence of the species, influenced by the retreat of Pleistocene glaciers, which lead to areas becoming available for rapid migration and the evolution of new types with the addition of isolation and time. 8 1.4 Townsendia aprica and T. jonesii var. lutea Townsendia aprica was listed as a federally threatened species in 1985. This was due to T. aprica's narrow geographical range of 301 km2, from as few as 15 populations ranging from a reported 6-2000 individuals per population (very rough estimate), and compounded by the potential mining hazard and cattle activity threatening 95% of T. aprica populations (U. S. Fish and Wildlife Service, 1993). Townsendia jonesii (Beaman) Reveal var. lutea Welsh, which is also a narrow endemic in Utah is not eligible for federal protection because it does not have full species status. This is a grave concern because the taxon has a more restricted distribution, fewer populations and individuals, and has many populations that are threatened by active mining operations. As shown below, because of the shared feature of narrow endemism and the evidence of a close relationship between the taxa, they form a logical pair for inclusion in a joint study of conservation genetics. Townsendia aprica had not been discovered when John Beaman conducted his studies. This inconspicuous, early flowering taxon appears to have been first collected by Welsh and Reveal in 1966. Townsendia aprica was later described by Welsh and Reveal (Welsh and Reveal, 1968) after the generic placement of the taxon was verified by Arthur Cronquist, John Beaman, and Arthur Holmgren. These experts were consulted in part because this new taxon had yellow ray ligules, a condition not previously known in Townsendia. Townsendia aprica (last chance townsendia) is an herbaceous perennial, endemic to the Colorado Plateau region of southeastern Utah and has a narrow distribution of 4,186 km which spans three counties in Utah (Sevier, Emery and Wayne; Figure 1.2). Townsendia aprica is acaulescent and grows woody stems with age, which may indicate that it is a long-lived perennial (Hamrick et al., 1991) although no lifespan studies have been done. No published chromosome counts exist for T. aprica, although a recent breeding system study of three populations determined that these populations are sexual (Tepedino et al., 2004) and thus likely diploid. Most 9 populations of T. aprica occur at middle elevations (1800m - 2350m), with one single, isolated, high elevation population (2500m) whose breeding system has not been characterized. Figure 1.2. Map of Utah showing the geographic range of T. aprica and T. jonesii var. lutea in relation to the northwest boundary of the Colorado Plateau. In their 1968 paper, Welsh and Reveal proposed that two taxa, T. montana specifically var. minima (recognized now as T. minima Eastw.) and T. mensana var. jonesii (classified now as T. jonesii var. lutea, T. jonesii var. tumulosa, and T. jonesii var. jonesii), were most closely allied to T. aprica. Townsendia montana differs from T. aprica by its larger size and less pulvinate (cushion-like) habit, longer and wider leaves, ray color (white or purple), and longer and less densely pubescent achenes. Ecologically, T. minima was described as being found only on red sandy soils, whereas T. aprica was thought to be restricted to heavy clay soils. Townsendia jonesii differs from T. aprica in a less densely pulvinate habit, with a more open and spreading growth form, with somewhat larger heads having peduncles (stalk of inflorescences), and longer leaves, with less densely strigose (apprised hairs all in the same Townsendia jonesii var. lutea northwest edge of the Colorado Plateau Townsendia aprica 10 direction) leaf bases. Interestingly, what was supposed to be the key character defining T. aprica, its unique yellow ligules, was later found to occur in the newly discovered T. jonesii var. lutea (Welsh, 1983b). This new variety was described in part from material originally assigned to T. aprica, but the long pappus of the ray florets (2-4.5 mm) suggested a closer affinity to the remainder of T. jonesii. This new taxon has an even more narrow distribution (311 km2) than T. aprica and although the two taxa occur in nearby regions, their distributions are non-overlapping (Figure 1.2). Townsendia jonesii var. lutea is known from Sevier, Piute and possibly Juab counties, where it occurs most commonly on Arapian shale containing commercial deposits of gypsum. As more surveys of both T. aprica and T. jonesii var. lutea have uncovered more populations of the two taxa, it has become apparent that the ligule colors in both taxa can range from white, to cream to yellow (Armstrong and Thorne, 1991; Clark, 1999, 2000, 2001). Welsh (1987; Welsh et al., 1993) uses yellow ray color in combination with ray pappus length to split these two taxa, even though both color and pappus length seem to overlap. The ray pappus length measurement for T. aprica has been recorded in the range of 0.7-1mm (Welsh et al., 1987; Welsh et al., 1993), with the latest reported measurements, based on many more individuals, given as between 0.5-2.5mm (Armstrong and Thorne, 1991). Pappus length in T.jonesii var. lutea reportedly ranges from 2-4.5mm (Welsh and Reveal, 1968; Welsh et al., 1987; Armstrong and Thorne, 1991; Welsh et al., 1993). Despite the overlap in these features, even the most recent Townsendia keys (Welsh et al., 1987; Welsh et al., 1993) use yellow ray color in combination with ray pappus length to distinguish T. aprica and T.jonesii. A close relationship between T. aprica and T. jonesii var. lutea has more recently been confirmed through molecular phylogenetic studies of Townsendia based on ITS and ETS sequencing (Thompson et. al., unpublished). Based on sequences from single individuals of T. aprica and the three varieties of T. jonesii, it is clear that the four taxa are very closely related. Although there are substitutions that distinguish the ITS sequences of T. aprica from those of T. jonesii, whether these are shared among populations of the individual taxa is not currently known. It does, however, appear that Townsendia aprica and the three T.jonesii varieties form a moderately well-supported monophyletic sister group to one other Utah endemic, the more northern T. mensana (formally T. mensana var. mensana). Surprisingly, the other Utah endemic thought to be closely related to T. aprica, T. minima, does not appear to be closely related to this group. Interestingly, Beaman (1957) shows unknown apomixis in T. mensana var. mensana and T. mensana var. jonesii, but does show infrequent apomixis in T. minima (formally T. montana var. minima) (Figure 1.1). It is hoped that sequencing of additional regions for multiple populations of the T. aprica-T. jonesii clade, currently underway, will further resolve species boundaries and the relationships in this group. With so much evidence of a close relationship between T. aprica and T. jonesii var. lutea, their overlapping flowering times and close geographical ranges, it was decided that T. jonesii var. lutea would be the best taxon to use for a genetic comparison with T. aprica. In addition, comparisons of T. aprica and T. jonesii var. lutea might aid in clarifying whether these two taxa are genetically distinct. Previous Studies of Townsendia aprica Owing to its designation as a threatened species, T. aprica has been the subject of a number of surveys and studies aimed at better understanding the biology of this species. Townsendia aprica grows on land administered by four different U.S. government agencies [(Bureau of Land Management (BLM), National Forest Service (NFS), National Park Service (NPS), and State Lands (SL)] who have carried out population monitoring studies and abiotic impact studies aimed at providing better protection for T. aprica. A breeding system study was carried out by the USDA Forest Service on T. aprica (Tepedino et al , 2004). They experimentally determined the breeding system of three T. aprica 12 populations by testing for autogamy and apomixis by excluding pollinators, for geitonogamy (selfing) by pollinating receptive stigmas with fresh pollen from the same plant, and for xenogamy (outcrossing) by pollinating receptive stigmas with fresh pollen from a different plant. Based on the lack of seed when pollinators are excluded or florets are self-pollinated, they concluded that the individuals tested are neither apomictic nor self compatible but show an obligate outcrossing breeding system. The authors of this study report seeing mostly native solitary bees, such as Osmia, and ground-nesting species like Synhalonia fulvitarsis, nesting among T. aprica plants. Over a 5 year period, the Utah Bureau of Land Management (BLM) monitored one plot of T. aprica at the type locality population (in this study classified as 9-TL) to observe fluctuations in survivorship of existing individuals and recruitment of seedlings. The populations were considered stable from 1996-1999, with a sharp decline in seedling and adult survivorship during 2000-2001, which was attributed to drought followed by an extended winter (Armstrong, pers. com.). During this period, soil samples were also collected and soil composition and levels of micro and macronutrients were described from 17 sites and compared with soils data from five T. jonesii var. lutea sites (Armstrong and Thorne, 1991). Armstrong found that sites of the two taxa share many overlapping soil characteristics, with some large differences in %sand, %clay, %CaCo3, and %gypsum (Armstrong and Thorne, 1991). My project included a continuation of this study with additional data gathered from the southern T. aprica populations that were not included in the first study because they were unknown at the time or not on B L M land. Another population monitoring project took place in 2002 and 2003 (Robinson, 2002, 2003) in which previously known northern sites of T. aprica were revisited over a two year period to see if they were extant and to look for evidence of population size fluctuations from year to year. The study found a 40% decrease in the number of standing populations from 13 previously documented sites (many at the edges of the northern range), with many of the known sites containing fewer individuals in 2003 than in the previous year. The conclusion, though not explicitly tested, was that heavy drought conditions during the two year study contributed to population decline. Site visit reports were done for three years by CARE (Capitol Reef National Park Survey Team) (Clark, 1999, 2000, 2001) and during this time, four new populations (1-4 in this study) were found, extending T. apnea's southern distribution limit by 20 km. The team raised concerns that these new populations showed different morphological and edaphic characteristics than the northern populations, and therefore sought out expertise to help confirm their identity. With many new populations being found since its federal listing, representatives of Federal Fish and Wildlife were proposing to delist T. aprica if these new discoveries helped fulfill the criteria for minimal viable population size set out in T. apnea's recovery plan (U. S. Fish and Wildlife Service, 1993). With the discovery of the southern populations, and the proposal for delisting, representatives of the various management agencies were eager to see a genetic study of T. aprica undertaken, and this was one of the reasons that management agencies cooperated by providing support and granting permits for this thesis. Even if there were enough populations and individuals to propose delisting of T. aprica, conservation genetics can inform the managers about other factors that may cause species rarity to turn into species extinction. Genetic information will also be useful in the implementation of any recovery plan for the population augmentation of T. aprica individuals into existing populations and for any new establishment of populations. Knowledge of the pattern of among population variation will also inform strategies based on levels of genotypic and allelic diversity shared among the populations and will show how much those populations contribute to the species' diversity as a whole. 14 1.5 Objectives The main objective of this study, as with many studies of rare endemics, was to estimate the level and pattern of genetic variation and inbreeding in T. aprica. Isozyme electrophoresis was employed to determine the spatial distribution of genetic variation among and within populations sampled throughout T. apnea's known range. The genetic parameters obtained for T. aprica were compared to one sub-specific variant of its closest known congener, T. jonesii var. lutea, as well as to average levels found in other studies of rare endemic plants (Hamrick and Godt, 1989; Karron, 1991; Bayer, 1992; Gitzendanner and Soltis, 2000). With fewer than 15 known populations of T. aprica, many isolated from one another and containing fewer than 100 individuals, the expectation was of finding low overall genetic variation and inbreeding within populations of both taxa. A number of Townsendia species have variable breeding systems, including both sexual diploid and apomictic polyploid populations, but the U.S. Endangered Species Act does not extend protection to apomictic species (United States, 1972) and would have to reconsider protection if this listed species was found to be exclusively apomictic. One of the first goals of this study was to confirm that the known populations of T. aprica are sexual outcrossers. By including experimentally confirmed outcrossing populations of T. aprica (Tepedino et al., 2004) as well as uncharacterized populations in this electrophoretic study, I will be able to make inferences about the breeding system of the uncharacterized populations. The third goal of this study is to better understand any distinguishing differences and the degree of relatedness between T. aprica and T. jonesii var. lutea by comparing inter- and intra-taxon genetic distances, comparing their soil types, as well as determining whether pappus length is really overlapping in the two taxa. 15 Chapter 2 Materials and Methods 2.1 Field Sampling Population localities for T. aprica were released through herbarium records from Utah's Natural Heritage Program that are deposited at B R Y , N Y B G , SJC, C R N P , O F S H (Table 1.1-Appendix I). First-hand knowledge of T. aprica population locations, based on field surveys by different government agencies, was obtained from Debbie Clark, partner of Capitol Reef National Park. Populations o f Townsendia aprica have patchy distributions and are isolated from one another throughout its range, and individuals within populations have clustered distributions. These populations were defined by D . Clark based on the assumption that the individuals at a given site are in close enough proximity to form an interbreeding group. A l l T. aprica herbarium vouchers, as well as some representatives of T. jonesii var. lutea, were loaned to The University of British Columbia ( U B C ) Herbarium. These herbarium vouchers were used to examine morphological differences, as well as for pollen counts and pappus measurements. In A p r i l o f 2001 and M a y o f 2002, samples were collected from 13 populations in three counties (Sevier, Wayne, Emery) representing the known range o f T. aprica (Figure 2.1). We collected 25-30 individuals per sampled population, except for two populations, where too few individuals were found (Table 2.1). From each individual plant, 7-10 leaves and ten seeds were collected, along with measurements of whole plant diameter, number of rosettes, number of buds, number of flowering heads, number of seed heads, and ray ligule color, i f available. Under the terms o f the permit, I was not allowed to k i l l any T. aprica in the course o f this study. This limitation included collection of herbarium vouchers, as well as progeny that might be raised from seed for a breeding system study or chromosome counts. Photographs of the habit and habitat for each taxa were taken, along with GPS coordinates. 16 • Townsendia aprica (1-13) Townsendia jonesii var lutea (14-20) Piute « 7 * 6 #1 * 5 3 2 Wayne 40 km Figure 2.1. Townsendia aprica and T. jonesii var. lutea sites collected in Utah in 2001 and 2002. During the 2001 and 2002 field seasons, T. jonesii var. lutea was also collected at eight sites, all in Sevier County (Table 2.2). Although T jonesii var. lutea cannot be listed for protection due to its infraspecific taxonomic status, the same collecting guidelines used for T. aprica were followed for this rare, endemic variety to minimize any impact on these populations. When larger plants were found, extra leaves were collected to use for screening and optimizing isozymes protocols. The morphological measurements and descriptive characters noted for T. aprica were not documented for Tjonesii var. lutea due to time constraints. It should be noted that the geographic ranges of the two taxa are markedly different. The geographic range of Townsendia aprica is 82km from north to south and 43km from east to west with the closest pair of sampled populations being 3km apart (11-FJ & 12-FJR) and the furthest, 82km apart (2-MMD & 13-MCR) (Table 11.1-Appendix II). The range for Townsendia jonesii var. lutea is 33km from north to south and 13km from east to west. The closest neighboring 17 populations sampled were <1 km apart (18-TME & 19-TMW) while the most distant were 33km apart (14-GH & 20-WC) (Table II.2-Appendix II). Once collected, leaves were placed in individual cryogenic vials and held on ice for no less than one hour, until transferred to a portable liquid nitrogen tank while in the field. Upon return to UBC, the vials where transferred into a -80°C freezer until they were analyzed. Population density estimates were obtained for T. aprica by counting the number of individuals within a randomly chosen 10m xlOm string quadrat (divided into four 5m x 5m plots) and then averaging the four counts from each sampled site (Table 2.1) (Kent and Ciocker, 1994). In 2001, soil collections were made at six T. aprica sites and two T.jonesii var. lutea sites (Table 2.1 and Table 2.2). This was done to further an earlier survey by Armstrong and Thorne (1991), aimed at determining whether the two taxa occur on distinct soil types. Roughly two hundred grams of soil was collected from two spots at each site and deposited into a clean plastic bag with non-metal closures, sealed to avoid contamination and brought back to UBC for storage at room temperature in the Whitton lab. 18 Table 2.1. Townsendia aprica localities and collection information for this study, along with additional information available for each Pop. number Pop. acronym Population name Latitude & Longitude Location County Elevation (m) Collection year Sample size Soil samples PC/BS/ C O DNA estimated density (100 m 2) estimated size 1 MCC Mini-Cockscomb 38°15.165'N 111 °24556'W S.onHwy 12,turnW.onroadto Teasdale,Cockscomb Wayne 2166 2001 21 J PC 46 <100. 2 M M D Miners Mt.Dbae 38° 11399'N 111 0 16.568' W S.on Hwy 12,turn Eon to dirt road (Miners Mt.Rd.), drive S. 45 air miles to population on h i Wayne 2330 2001 27 J 15 <100 3 WHF Wide Hollow Fork 38°13208'N 111 0 19257'W S.on Hwy 12, turn Eon to dirt road (Miners Mt.Road.),drive S.for 2 air miles thenN.for2airmiles Wayne 2354 2001 27 J 22 MOO 4 JM Johnson's Mesa 38°15.012'N 111 0 17386'W S.onHwy 12,tumEontodirtroad (Miners Mt.Rd.),drive S.for2air mites then N.for 6 air mites Wayne 2167 2002 27 PC 13 >100 5 HNE Hartnet East 38°27237'N 111 0 16.005'W E.on Hwy 24,turn N.on dirt road,cross riverinto Hartnet,drive for30 miles, population on mound Wayne 1824 2001 25 J 17 <100 6 CV CathedralValley 38°28354'N 111 °21.736'W E.on Hwy 24,turn N.on to dirt road, cross river into Hartnet,drive to CathedralValley Campground Wayne 2113 2002 27 48 MOO 7 BCP Bristlecone Pine 38°26385'N 111 °25585'W FromCathdral Valley Campground drive Paradise Flats Campgroud, hike 3 mites to Billings Pass Wayne 2749 2002 27 PC 36 MOO 8 JHO Johns Hole Overtook 38°38.408'N 111 °02213'W E.on 1-70, exit 97,headS. for 25 air miles, population on side of road Emery 2030 2002 27 2 <100 9 TL Type Locality 38°40544'N 111 °24.743'W l-70,exit72,drive for 15 miles, populatcn on E,side of road Stever 2265 2001 15 A & T BS 22 <100 10 FJ Fremont Junction 38°43.968'N 111 °23.844'W I-70,est 72,drive for3mites,population onW.side of road Siever 2200 2001 27 A & T BS, DNA 43 MOO 11 FJR Fremont Junction Rest Stop 38°45.191'N 111 °25210'W I-70,e?flt at Fremont Juction Rest Stop, just beyond fence behind rest stop Siever 2194 2002 27 A & T BS 18 <100 12 DVM Dog Valley Mine 38°46311'N 111 0 16043'W E.I-70,e?dt on to dirt road in Canyon,NE of Windy Peak, drive E.for4 miles, population on old mine Emery 1967 2002 27 A & T 14 MOO 13 M C R M oore Cut Off Road 38°51419'N 110 °55.473'W Eon I-70,ext off Moore Cut Off Road, drive for 1 mite sand pit Emery 1994 2002 27 23 MOO. Soil samples taken at each site are as follows: J - Jennings, (this thesis); A&T - Armstrong & Thorne (1999). Information on breeding system (BS) is from Tepedino et. al. (2004). DNA samples extracted (Thompson, et al., unpublished). Pollen count (PC) (Jennings, this thesis) are from herbarium specimens (Appendix I- Table 1.1). Table 2.2. Townsendia jonesii var. lutea localities and collection information for this study, along with additional information availab e for eac i population. Pop. number Pop. acronym Population name Latitude & Longitude Location County Deration (m) Year Collected Sample size Soil samples PC/BS/ CC/DNA 14 GH Gleenwood Hills 38°45.558'N 111 055.278'W E. on Hwy 119, NEof Gfenwood in Ra inbow HiDs jus t be fore Hwy. 24 Sevier 1950 2001 27 J 15 SFS Sage Flat South 38°49.112'N 111 "56.01 l'W S. on Hwy 24.headEofSigurd.turn N. on Sage Flat Road,population on S. side of road Sevier 1683 2002 27 16 SFN Sage Flat North 38°53.291'N 111 °51.638'W S. on Hwy 24, he ad E of Sigurd, turn N. on Sage Flat Road, go 7.5 air miles, population on S. s ide of road Sevier 1655 2002 27 A & T 17 GBR Goos e be rry Roa d 38°49.380'N 111 °43.010'W Eonl-70,take SokherCanyonRoadS.9 air miles,population on S.side of dirt road Sevier 2167 2002 27 IS T M E Triangle Mountain East 38°55.634'N 111 °48.754'W Eon 1-70,exit on SoldierCanyonRd.,on the nwside of Triangle Mountain Sevier 1659 2001 27 J . A & T cc 19 T M W Triangle Mountain West 38°55.675'N 111 °48.889'W Eon 1-70,exit S.onSoUierCanyon Road,W.across roadfromTriangle Mountain Sevier 1659 2001 27 20 WC WilbwCreek 39°01.247'N 111 °47 .791'W N.on Hwy. 89,E.of Redmond,head towards WilbwCreek Reservoir, population on S. s ide of road Sevier 1690 2002 27 A & T Soil samples taken at each site are as follows: J - Jennings, (this thesis); A & T - Armstrong & Thorne (1999). Chromosome counts (CC) are from (Windham, M . , University of Utah, unpublished). 2.2 Lab Methods 2.2.1 Pollen size and stainability Herbarium specimens were used to determine whether there was any evidence that T. aprica is apomictic by estimating pollen viability and size variation using aniline (cotton) blue staining (Beaman, 1957). This method is standard for determining polyploidy and apomixis in Townsendia (Beaman, 1957) and other groups (Hauser and Morrison, 1964; Torrell et al., 1999). Only those T. aprica populations whose breeding system were more in question (southern populations 1-4 and the highest elevation population 7-BCP), were sampled using pollen from herbarium vouchers (Table 1.1-Appendix I). From each voucher, one floret from each individual plant was macerated on a slide with a drop of cotton blue stain, which viable pollen nuclei will take up within a few minutes of exposure. After staining, the slides were observed with a light microscope and pollen counts were done within ten minutes of dye exposure. Counts of fully stained (presumed viable) and unstained (not viable) pollen grains were made while also noting any size differences that might indicate polyploidy. Sample sizes per individual ranged from 79 to 268 pollen grains per floret. 2.2.2 Electrophoretic procedures Designing and optimizing the isozyme protocol including extraction buffer, running buffers and enzymes was done with the extra leaf material of T. jonesii var. lutea. Looking for clear banding patterns and variability within each isozyme, I systematically tested six different extraction buffers with 2-5 leaves on three different running buffers and screened 21 enzymes to come up with the final isozyme protocols. Electrophoresis was performed with two electrode buffer systems, morpholine and lithium borate (Table 2.3). 21 Table 2.3. Electrode buffer systems including modifications in pH and run time. Buffer system Electrode buffer pH Gel buffer pH Typical run time and mA Reference Morpholine 8.0 8.4 7 hrs / 40 mA Clayton & Tretiak, 1972 Lithium borate 8.0 8.0 4 hrs / 45 mA Ridgeway et al, 1970 Buffer recipes are listed in Appendix III (Table III. 1). Approximately 2-3 frozen leaves from each individual sample were ground to a slurry in individual vials on ice with 2-3 drops of germanium extraction buffer pH 7.1 (Table III.2-Appendix III), using a hand held drill fitted with a custom made plastic bit. Wicks (3mm x 15mm) made of Whatman chromatography paper were used to absorb the supernatant and were then briefly blotted on a paper towel to remove any excess material. The wicks were then inserted into a cut 11.5% starch gel chilling on ice. The gels were then placed onto eletrophoretic rigs, put into a refrigerator and run at a low voltage (160/35 V/mA) for the first twenty minutes. Once the dye indicator had moved into the starch matrix, the wicks were removed, and plastic spacers were inserted into the end of the tray to ensure good contact across the cut completing the circuit. The trays were placed back into the refrigerator with ice packs laid over the gels, and run for the appropriate time for adequate separation of alleles (Table 2.3). Once the running time had elapsed, gels were removed from the fridge, sliced into seven layers, and one slice was set into the appropriate staining tray for each enzyme. Trays were set in an incubator at 37° for approximately 30 minutes to complete the reaction. Seven enzymes resolved consistently for both taxa (Table 2.4). Staining protocols for resolving enzymes L A P and M D H followed Soltis et al. (1996) and A A T , P G M , IDH, 6PGDH followed Murphy et al. (Murphy et al., 1996). Table 2.4. Resolved enzymes with each bu ffer system used. Locus Enzyme name E.C. number Buffer 6-PGDH 6-Phosphogluconate dehydrogenase Morpholine A A T Aspartate aminotransferase Morpholine G6PGD Glucose-6-phosphate dehydrogenase Ridgeway IDH Isocitrate dehydrogenase Ridgeway L A P Leucine aminopeptidase Ridgeway M D H Malate dehydrogenase Ridgeway P G M Phosphoghicomutase Morpholine E.C. number = Enzyme Commission number 22 As soon as bands were visible, each gel was scored by eye and documented with a digital image. The phenotypes were interpreted as genotypes with the most anodal (top of gel) isozyme numbered locus '1 ' and the fastest migrating allozyme designated allele 'a' with slower bands receiving the appropriate letter based on gel position. 2.2.3 Soil procedure Once back in the lab, soil samples were air-dried, then sifted through a 2mm sieve. Paired samples from each site were then combined and sent to the Brigham Young University Soils Lab where the same procedures used in the 1991 survey (Armstrong and Thorne, 1991) were followed. Samples were analyzed for extractable forms of potassium, calcium, magnesium, phosphorus, sodium, phosphate, all in ppm (ug/ g of dry soil), pH, %sand, %clay, %silt, %calcium carbonate (free lime - CaCOs), %calcium sulfate (gypsum - CaS04), sodium absorption ratio (SAR), %Moisture 1/3 bar, and %Moisture 15 bar, also known as soil moisture tension. This measurement indicates at lower (1/3 bar) and higher (15 bar) tensions to what degree the water adheres to the soil and the difference between the two should be an indication of available water to a plant. 2.2.4 Pappus measurement procedure From each taxon, three to four ray and disk pappus length measurements were taken from at least one individual from each population to see if their length measurements were really non-overlapping as the flora of Utah suggests (Welsh et al., 1987; Welsh et al., 1993), or if they are overlapping, as the B L M report suggest (Armstrong and Thorne, 1991). Some of the herbarium vouchers were used for this part of the study (Table 1.1-Appendix I) while the other pappus measurements were from seed collected in my 2002 field season. 23 2.3 Analysis 2.3.1 Isozyme analysis The following genetic variability parameters were analyzed in P O P G E N E (Yeh, 1997): percent polymorphic loci (%P), mean number of alleles per locus (A), effective number of alleles (A e ) , expected heterozygosity (H e ) and observed heterozygosity (H 0 ) . Estimates of allele and expected genotype frequencies were generated in P O P G E N E from the sampled populations of Townsendia aprica and T. jonesii var. lutea, and used to test the null hypothesis that populations are in Hardy-Weinberg equilibrium (HWE) where H 0 = H e . In this study, loci were considered polymorphic i f more than one allele was observed, regardless o f its frequency. Mean number of alleles per locus (A) (sometimes called allelic richness) is simply the total number of alleles observed at all loci divided by the total number o f loci investigated. However, A does not take into consideration the frequency of the different alleles and is dependent on sample size, and therefore is only a very general indicator o f diversity. The mean number of effective alleles per locus (A e ) (Kimura and Crow, 1964) as described by Berg and Hamrick (1997), is a measure of the evenness of the most common alleles, developed to account for the observation that rare alleles have a minor effect on genetic diversity (H e ) compared with the most common alleles. It is defined as Expected heterozygosity (He) is the expected proportion of heterozygous loci per individual assuming random union o f gametes and for multiple alleles is defined as where n is the number of alleles, and p] is the expected genotypic frequency o f homozygotes for the ith allele (Hedrick, 2000). Since H 0 is observed, it is a biased diversity indicator because it A e = 1/(1-H e) 1 < A e < OO n 24 can be affected by any evolutionary processes that violates the assumptions of HWE such as non-diploid individuals, small populations size, overlapping generations, asexual reproduction, non-random mating, natural selection, migration, and mutation (Hartl, 1987; Hedrick, 2000). A significance test (t-test, p = 0.05) between H 0 and He was carried out using SYSTAT (version 10.0) (Wilkinson et al., 1996) to see if Ho and He are significantly different. If dofferemces are significant, this could indicate that one or more of the HWE assumptions have been violated. Expected and observed heterozygosity often differ due to violations of HWE. These deviations can be more closely measured with Wright's (1951) three F-statistics (FST, FIS, FIT). These are also known as fixation indices which describe the distribution of genetic diversity within and among sub-populations. Fis and FIT are defined as the probability of two gametes uniting (individual-I) relative to their frequency in the subpopulation (S) and total population (T), respectively. FST measures the degree of differentiation of these subpopulations by correlation of two randomly drawn gametes in each subpopulation (Nei, 1987). Their relationship is described as follows l - F / r = ( 1 - ^ X 1 - ^ ) F-statistics for both taxa and all loci were estimated in GENEPOP (Raymond and Rousset, 1995) using the modified coefficients of Weir and Cockerham (1984) which are considered better suited to small data sets than Nei's, because they do not make assumptions about the number of populations, number of samples, or heterozygote frequencies. A significance test for linkage disequilibrium was also run in GENEPOP, using a Markov chain method. This tests the hypothesis that the genotypes at each locus are independent of other loci by looking for evidence of physical linkage or segregation distortion. I also analyzed all populations for evidence of bottleneck effects using the program Bottleneck (version 1.2.02) (Cornuet and Luikart, 1997). The theory behind this test assumes that if a population has experienced a recent bottleneck, allele numbers will be reduced more than 25 expected heterozygosity ( H e ) (Maruyama and Fuerst, 1985; Cornuet and Luikart, 1997; Piry et al., 1999). This is because the loss of low frequency alleles during a bottleneck event will have little initial impact on H e . The program can detect a population bottleneck by comparing He with what is expected if the population was at mutation-drift equilibrium ( H e q ) given the numbers of alleles. They defined this as heterozygosity excess, which is not to be confused with an observed excess of heterozygotes ( H G > H e ) . The raw allele frequency data was inputed and analyzed following the Infinite Allele Model (IAM) of mutation, and significance was based on the Wilcoxon sign-rank test with 1000 iterations. Nei's (1973) genetic diversity statistics (Hr, H s , DST, GST) were assessed in FSTAT (Goudet, 2001) to allow comparison with other population genetic studies. The total genetic diversity (Hp) is estimated from average allele frequencies over all populations, H s is the genetic diversity based on allele frequencies within populations, and DST is the genetic diversity found among populations with their relationship described as: HT =HS+ DST GST is described by Nei as the measure among-population differentiation, relative to total diversity, and is calculated as: Genetic distances can reflect the relatedness of the sampled populations under the assumption that populations that are more closely related will share alleles at similar frequencies. A variety of algorithms exist to estimate genetic distances between populations, each with its own set of assumptions. Cavalli-Sforza and Edwards (1967) chord distances were used as the basis for tree construction using PHYLIP (version 3.63) (Felsenstein 1995). Cavalli-Sforza and Edwards (1967) chord distances were considered more appropriate when looking at relationships within taxa because differences found between the populations are more likely to be due to drift 26 alone rather than drift and mutation, as Nei's genetic distance assumes. Allele frequencies were input into GENEDIST program, and the distance matrix was placed into the tree joining program 'NEIGHBOR' in PHYLIP (Felsenstein 1995) where a tree was generated for populations of each taxon separately. The tree output from Phylip was input into TREEVIEW (Page, 1996) to generate a printable tree with a scale bar for genetic distance. Bootstrapping was used in PHYLIP to evaluate the statistical support of the branching patterns. The allele frequency matrices were re-sampled 200 times in SEQBOOT and the re-sampled bootstrap data was then inserted into GENEDIST, again choosing Cavalli-Sforza and Edwards. The set of matrices was then entered into NEIGHBOR-joining to construct 200 unrooted trees which where then entered into CONSENSE (Margush and McMorris, 1981) to calculate the final bootstrap values based on the number of times that each branch was recovered in the bootstrap trees. Correlation between genetic distance and geographic distance within T. aprica and T. jonesii var. lutea were examined using a Mantel's test (Mantel, 1967) in GENEPOP with 1000 permutations used to determine significance. The program ISOLDE follows Rousset (1997) who suggested that the use of FST instead of genetic identity is a more appropriate parameter to use for evaluating isolation by distance. Rousset argues that FST is relatively independent of mutation rate and mutational,processes and is remarkably accurate at short distances. Since samples at small geographic distances are not expected to follow the general theory of isolation by distance, the ISOLDE program allows for input of the minimum geographical distance between populations and then only pairwise comparisons of larger distances are used to estimate the regression coefficient. 27 2.3.2 Genetic analysis between T. aprica and T. jonesii var. lutea A genetic distance tree showing the relationship between the two taxa was constructed in PHYLIP and evaluated using bootstrapping following the methods described above. Nei's genetic distance (Nei, 1972) was used in comparisons of the two taxa because its assumption of mutation-drift equilibrium is more appropriate for inter-taxon comparisons (K. Ritland, personal communication). 2.3.3 Soil analysis Available cations of Ca, Mg, K and Na were analyzed by the ammonium acetate method at pH 8.5 (Normandin et al., 1998). Phosphorus measurements, which is normally performed by extraction using NaHCC»3, instead followed Harper's protocol (University of Utah, unpublished), which involves extraction in 0.2N acetic acid. This protocol was used for samples analyzed in 1991, and we followed this protocol for the additional samples collected in 2003. Acid neutralizing potential or % calcium carbonate (CaCCb) extraction followed the method of Allison and Moode (1965), while analysis of %calcium sulfate (CaS04) followed the method of Nelson (1982). Measurement of percent moisture at 1/3 bar and 15 bars, soil moisture tension, along with percent sand, clay and silt followed Day (1965). Soil data were averaged and significance for each taxon was assessed using Kruskal-Wallis non-parametric test in SYSTAT (Wilkinson et al., 1996). Chapter 3 Results 3.1 Breeding system in Townsendia aprica Pollen grains were counted and percent pollen stainability estimated from one floret per flower head from a single herbarium voucher from each of three populations (Table 3.1). Counts were carried out on two (1-MCC & 2-MMD) of the four southern populations, which local field biologists have noted are of questionable identity (note that no voucher exist for 3-WHF and 28 there were no flowers on the voucher for 4-JM). I also looked at pollen from the one high elevation population (7-BCP) because apomixis and polyploidy tend to be found at higher elevations in other Townsendia species with narrow ranges. Table 3.1. Estimated pollen stainability of Townsendia aprica individuals. Total Pollen Viable % Viable Population Name Collection Flower Grains Pollen Pollen K . Heil/D. Schles 1-MCC 7559 A 169 146 86.0 (southern) B no flowers C 204 190 93.0 population mean 89.5 F lemming/ Romme 2-MMD 603 A 106 101 95.0 (southern) B 268 260 97.0 C 225 200 89.0 population mean 93.7 7-BCP Mark Porter 3855 A 229 201 88.0 (high elevation) B no flowers C 226 217 96.0 D 79 35 44.0 E 144 126 87.0 population mean 78.8 7-BCP K.Heil 4376 A 148 78 53.0 (high elevation) B 140 65 47.0 C 231 164 71.0 population mean 57.0 All vouchers are from San Juan College (SJC). Al l counts revealed pollen stainability above 20%, which is considered to be the cut off for low viability in Townsendia (Beaman, 1957). Also, there was no variation in pollen grain size observed in any of the samples. While pollen for the two southern populations is around 90%, viable, the stainability of pollen from population 7-BCP was significantly lower that that of the southern populations (t-test; p < 0.05). Most of the enzymes resolved in this study show no indication of duplicate banding patterns or fixed differences between homologous loci across any of the populations. There were two loci resolved for the enzyme IDH, which has previously been documented in other genera to have only one locus (Wendel and Weeden, 1989), but as these authors noted, this is not consistent across all species. 29 3.2 Genetic diversity in Townsendia aprica and T. jonesii var. lutea After screening 21 different enzymes, seven enzymes were chosen that resolved nine loci consistently for both taxa (Table 3.2). Other enzymes were also resolved during the screening process, but either had too many overlapping loci that made it too complex to score (EST), or did not stain consistently across all individuals in each population (ME, PGI), and therefore were excluded from this study. The preliminary screening showed that EST and PGI are polymorphic, while ME was monomorphic. Locus Enzyme name Resolved loci Subunit structure 6-PGDH 6-Phosphogluconate dehydrogenase 1 diner A A T Aspartate Aminotransferase 2 dimer G6PGD Glucose-6-phosphate dehydrogenase 1 dimer IDH Isocitrate dehydrogenase 2 dimer L A P Leucine arninopeptidase 1 monomer M D H Malate dehydrogenase 1 dimer P G M Phosphogtucomutase 1 monomer inzyme a )breviation, full name of the enzyme with the number of loci resolved and structure. Seven of the nine isozymes were polymorphic across most populations, and AAT-1 and IDH-1 were monomorphic across all individuals. The known quaternary structures of PGM and LAP are monomeric, while the other five enzymes in this study are known to be dimeric. No significant disequilibrium was detected between any of the enzymes analyzed, showing that all genotypes are independent observations. Both taxa share all alleles scored for in this study, but not all alleles are present within each population of T. aprica, nor are the alleles necessarily at similar frequencies in all populations (Table IV. 1 and Table IV.2-Appendix IV). Genotypes inferred from isozyme banding patterns were used to estimate genetic parameters for individual populations of T. aprica and T. jonesii var. lutea, as well as averages 30 for each taxon (Table 3.3 and Table 3.4). In this study, 7 of the 9 loci were considered polymorphic across all populations of T. jonesii var. lutea, and most populations of T. aprica, with percent polymorphic loci (%P) of 77.8 and 74.4, respectively. The mean number of alleles (A) for T. aprica is 2.09 with only slight fluctuations among populations, while the mean for T. jonesii var. lutea is 2.22, with no variation among populations as every allele scored was found in every population. The effective number of alleles (Ae) varied among populations of both taxa, with a mean of 1.71 for T. aprica and mean of 1.65 for T. jonesii var. lutea. There were no significant differences found in these genetic parameters between the two taxa. The mean observed and expected heterozygosity values for T. aprica are 0.31 and 0.33, respectively, while T.jonesii var. lutea means are overall slightly lower at 0.25 and 0.30. There was no significant difference between H 0 and H e for either taxon. The fixation index (Fis) has a theoretical range between -1 and 1, where a significantly negative value reflects excess heterozygosity, and a significantly positive value reflects heterozygote deficiency, or inbreeding (Table 3.3 and Table 3.4). Although both excess and deficiency of heterozygotes relative to Hardy Weinberg expectations were found in T. aprica populations, in only one case was the deviation significant, with an excess of heterozygosity found in population 10-FJ (Fis = -0.20). The majority of observations of excess heterozygosity were found in the northern populations (8 - 13). The highest positive Fis values in T. aprica occurred in isolated, southern populations 2-MMD & 4-JM (0.14 and 0.17, respectively), although these were not significantly different from zero. While all but one T.jonesii var. lutea population had positive Fis values, only two populations (18-TME & 19-TMW) had significant deficiency of heterozygotes (Fis = 0.25 and 0.30, respectively), indicating possible inbreeding within these two geographically close, isolated populations. 31 Table 3.3. Genetic variability parameters for populations of Townsendia aprica Population name Sample # of 1 "1 • 1 and number size loci #P % P mean A mean As mean Het 0 mean Het e Fis 1 IVICC 21 9 6 66.7 2.00 (0.19) 1.45 (0.13) 0.22 (0.06) 0.23 (0.05) 0.06 2 M M D 27 9 7 77.8 2.00 (0.14) 1.60 (0.13) 0.26 (0.05) 0.30 (0.05) 0.14 3 w i n 27 9 7 77.8 2.11 (0.15) 1.60 (0.13) 0.33 (0.06) 0.30 (0.05) -0.09 4 J M 27 9 6 66.7 2.00 (0.17) 1.62 (0.13) 0.25 (0.05) 0.30 (0.05) 0.17 5HNE 25 ' 9 7 77.8 2.22 (0.19) 1.69 (0.15) 0.30 (0.06) 0.32 (0.05) 0.04 6 CV 27 9 7 77.8 2.11 (0.15) 1.50 (0.11) 0.30 (0.06) 0.27 (0.04) -0.13 7 BCP 27 9 7 77.8 2.22 (0.19) 1.51 (0.10) 0.27 (0.05) 0.28 (0.04) 0.03 8 J H O 27 9 6 66.7 1.90 (0.15) 1.42 (0.12) 0.23 (0.06) 0.21 (0.05) -0.12 9 T L 15 9 7 77.8 2.22 (0.25) 1.90 (0.20) 0.44 (0.09) 0.40 (0.07) -0.14 10 FJ 27 9 7 77.8 2.11 (0.15) 1.75 (0.13) 0.43 (0.07) 0.36 (0.05) -0.20* 11 FJR 27 9 7 77.8 2.11 (0.15) 1.75 (0.13) 0.38 (0.07) 0.36 (0.05) -0.05 12 D V M 27 9 7 77.8 2.22 (0.19) 1.71 (0.12) 0.36 (0.06) 0.35 (0.05) -0.02 1 3 M C R 27 9 7 66.7 1.90(0.15) 1.60 (0.12) 0.30 (0.06) 0.29 (0.05) -0.04 mean 25 9 7 '74.4 2.09 (0.05) 1.71 (0.04) 0.31 (0.02) 0.33 (0.01) -0.03 Percent polymorphic loci (%P), number of alleles per locus (A), effective number of alleles (A e), observed heterozygosity (HQ), expected heterozygosity (Hg), and the inbreeding coefficient (Fis). Standard errors are in parentheses. * Significance level of (p < 0.05) Table 3.4. Genetic variability parameters for po pulations of Townsendia jonesii var. lutea. Population name Sample #of and number size loci # P % P mean A mean Ae mean Het„ mean Het e Fis 1GH (14) 27 9 7 77.8 2.22 (0.19) 1.57 (0.12) 0.33 (0.06) 0.30 (0.04) -0.13 2 SFS (15) 27 9 7 77.8 2.22 (0.19) 1.58 (0.14) 0.27 (0.06) 0.28 (0.05) 0.03 3 SFN (16) 27 9 7 77.8 2.22 (0.19) 1.50 (0.13) 0.24 (0.06) 0.25 (0.05) 0.02 4 GBR (17) 27 9 7 . 77.8 2.22 (0.19) 1.48 (0.11) 0.20 (0.05) 0.25 (0.05) 0.20 5 T M E (18) 27 9 7 77.8 2.22 (0.19) 1.69 (0.14) 0.24 (0.05) 0.32 (0.05) 0.25* 6TMW(19) 27 9 7 77.8 2.22 (0.19) 1.62 (0.12) 0.21 (0.05) 0.30 (0.05) 0.30* 7 WC (20) 27 9 7 77.8 2.22 (0.19) 1.69 (0.12) 0.28 (0.05) 0.33 (0.05) 0.15 mean 27 9 7 77.8 2.22 (0.07) 1.65 (0.05) 0.25 (0.02) 0.30 (0.02) 0.12 * Significance level of (p < 0.05) Wright's F-statistics for each isozyme locus (Table 3.5) show that T. aprica has a slight excess of heterozygotes overall, while the mean Frs values for T. jonesii var. lutes indicating 12% fewer heterozygotes than expected if populations are randomly mating. Table 3.5. Wright's F-statistics (Fis, FIT, FST) for all isozyme loci examined and their means for T. aprica and T. jonesii var. lutea, respectively. Locus Fis FIT Locus Fis F S T FIT P G M 0.487 0.274 0.628 P G M 0.461 0.184 0.560 6PGDH -0.210 0.050 -0.150 6PGDH 0.098 0.022 0.118 AAT-1 0.000 0.000 0.000 A A T 0.000 0.000 0.000 AAT-2 0.440 0.013 0.447 AAT-2 0.417 0.000 -0.282 MDH1 -0.272 0.033 -0.230 MDH1 -0.327 -0.002 -0.329 IDH1 0.000 0.000 0.000 IDH1 0.000 0.000 0.000 IDH2 0.153 0.150 0.281 IDH2 -0.124 0.022 -0.099 G6PDH -0.274 0.128 -0.111 G6PDH -0.043 0.001 -0.041 L A P -0.217 0.051 -0.155 LAP 0.418 0.105 0.479 mean -0.027 0.100 0.075 mean 0.120 0.060 0.173 The fixation index (FST) which gives an indication of how much divergence exists among populations, ranges between 0 (no differentiation among populations) and 1 (all genetic variation among populations, with no within-population variation). An FST < 0.05 is considered negligible, while FST > 0.25 shows great differentiation between the populations analyzed. The mean FST value for T. aprica (0.100) and T. jonesii var. lutea (0.060) indicates that most of the isozyme variation is found within populations and little structuring exists among the populations. Neither value is significantly different from zero. The mean Cavalli-Sforza and Edwards chord distance (1967) between populations of T. aprica was 0.118 (range: 0.022-0.334) while the mean chord distance between T. jonesii var. lutea populations was 0.061 (range: 0.006-0.195), with a significant difference found between values in the two taxa (p < 0.01; Table 3.6 and Table 3.7). The neighbor-joining trees produced using the genetic distances within taxa reveal little support for structuring of groups of populations within taxa, with few exceptions. The two most southern populations in T. apnea's range, 1-MCC and 2-MMD, clustered together and were somewhat distant from the rest of the 33 populations, with moderate bootstrap support of 63% (Figure 3.1). In T.jonesii var. lutea the neighbor-joining tree (Figure 3.2) showed three of the four branches with bootstrap values above 60%. The bottleneck analysis revealed that the majority of populations of both taxa have gone through recent bottleneck events, with excess heterozygosity found (H e > H e q). Specifically, in T. aprica, the more southern populations (2-5) showed significance (p < 0.05) and the most northern populations (9-13) showed greater significance (p < 0.01) with more heterozygosity than expected under mutation-drift equilibrium (Heq) (Table IV. 1-Appendix IV). In T.jonesii var. lutea populations 1, 2, 5 and 7 also showed significant evidence of bottlenecks (p < 0.05) (Table IV.2-Appendix IV). The Mantel test revealed a significant relationship between genetic and geographic distances within both taxa, suggesting that the populations show evidence of isolation by distance. Townsendia aprica showed a significant regression between genetic and geographic distance that explained roughly 11% of the variation in genetic distances (r2 = 0.112; p = 0.03) (Figure 3.3), while geographic distances explain much more of the variation in genetic distances among T.jonesii var. lutea populations (r2 - 0.4125; p = 0.004) (Figure 3.4), with the latter being sampled over a smaller geographical area. 34 Table 3.6. Cavalli-Sforza and Edwards (1967) chord distance pairwise comparisons of 13 populations of T. aprica 1 M C C 2 M M D 3 W H F 4 J M 5 H N E 6 C V 7 BCP 8 J H O 9 T L 10 FJ 11 FJR 12 D V M 13 M C R 1 M C C ***** 2 M M D 0.045 ***** 3 WH F 0.135 0.097 ***** 4 J M 0.130 0.131 0.101 ***** 5 H N E 0.145 0.141 •0.071 0.075 ***** 6 C V 0.142 0.124 0.031 0.066 0.045 ***** 7 B C P 0.166 0.136 0.061 0.105 0.025 0.022 ***** 8 J H O 0.217 0.209 0.073 0.167 0.156 0.079 0.133 ***** 9 T L 0.181 0.130 0.153 0.134 0.065 0.142 0.102 0.334 ***** 10 FJ 0.209 0.150 0.040. 0.095 0.076 0.036 0.066 0.118 0.122 ***** 11 FJR 0.223 0.201 0.073 0.083 0.066 0.064 0.097 0.161 0.129 0.033 ***** 12 D V M 0.224 0.197 0.049 0.117 0.037 0.061 0.062 0.139 0.112 0.045 0.031 ***** 13 M C R 0.317 0.219 0.088 0.128 0.146 0.089 0.118 0.177 0.202 0.075 0.118 0.109 ***** Mean Cavalli-Sforza and Edwards chord distance for T. aprica = 0.118. Table 3.7. Cavalli-Sforza and Edwards (1967) chord distance pairwise comparisons of 7 populations of T. jonesii var. lutea. 14 G H 15 SFS 16 S F N 17 G B R 18 T M E 19 T M W 20 W C 14 G H ***** 15 SFS 0.067 ***** 16 S F N 0.099 0.021 ***** 17 G B R 0.195 0.070 0.048 ***** 18 T M E 0.118 0.051 0.037 0.020 ***** 19 T M W 0.146 0.062 0.039 0.017 0.008 ***** 20 W C 0.112 0.066 0.055 0.031 0.006 0.012 ***** Vlean Caval i-Sforza and Edwards chord distance for T. jonesii var. lutea = 0.061. 9-TL 1-MGC Figure 3.1. Neighbor-joining tree using Cavalli-Sforza and Edwards chord distances for T. aprica. Bootstrap value above 60% is indicated along the branch. 36 1-GH 0 .1 Figure 3.2. Neighbor-joining tree using Cavalli-Sforza and Edwards chord distances for T. jonesii var. lutea. Bootstrap values above 60% are indicated along their branches. 37 Townsendia aprica 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 y = 0.001 l x+ 0.0594 2 r2 = 0.1123 p = 0.03 10 20 30 40 50 Distance (km) 60 70 80 90 Figure 3.3. Graph of regression analysis between pairwise FST and geographical distance between populations for T. aprica. Townsendia jonesii var. lutea 0.25 -, 0.2 0.15 J FST 0 1 0.05 -0.05 y = 0.0054x-0.0304 . r2 = 0.4125 p = 0.004 10 •• 15 20 25 30 35 40 Distance (km ) Figure 3.4. Graph of regression analysis between pairwise FST and geographical distance between populations for T. jonesii var. lutea. 38 3.3 Relatedness of Townsendia aprica and T. jonesii var. lutea Nei's (1972) mean intra-taxon genetic distances within T. aprica and T. jonesii var. lutea are 0.055 and 0.031, respectively (Table 3.8). The mean inter-taxon genetic distance was significant (p < 0.01) and estimated as 0.119. Nei's genetic distance (1972) was used to construct a tree to study the relationship between T. aprica and T. jonesii var. lutea. The neighbor-joining tree revealed an interesting pattern (Table 3.5), with populations of T. jonesii var. lutea forming a moderately well-supported group (69% bootstrap). Support for two of the significant groupings obtained in analyses within taxa was again recovered in this combined tree. Populations 1 and 2 in T. aprica form a group with 68% bootstrap support and populations 17-20 of Townsendia jonesii var. lutea have 69% bootstrap support. Results from the hierarchical analysis confirm the distance results indicating structuring of genetic variation between T. aprica and T. jonesii var. lutea with more genetic differentiation being found between the two taxa (FST = 0.211) than within them (FPT = 0.137). There were three soil characters that were found to differ dignificantly between the two taxa, %sand, %clay, and %CaCo3 (Table 3.9). Townsendia aprica soils have significantly greater %sand and lower %clay than soils collected at T. jonesii var. lutea populations. Townsendia jonesii var. lutea soils have significantly higher average %CaCo3 at 47.8% than T. aprica with 13.7%, a difference that is significant (p < 0.001). 39 Table 3.8. Nei's genetic distance (1972) (below diagonal) and identity (above diagonal) for all pairwise comparisons between T. 1 MCC 2 MMD 3 WHF 4 JM 5 HNE 6 CV 7 BCP 8 JHO 9 TL 10 FJ 11 FJR 12 DVM 13 MCR 14 GH 15 SFS 16 SFN 17 GBR 18 TME 19TMW 20 WC 1 MOC **** 0.992 0.920 0.930 0.919 0.917 0.905 0.901 0.917 0.888 0.880 0.889 0.847 0.829 0.830 0.808 0.791 0.822 0.831 0.825 2MMD 0.009 **** 0.939 0.937 0.928 0.928 0.919 0.918 0.934 0.916 0.895 0.905 0.877 0.833 0.848 0.824 0.817 0.844 0.846 0.848 3 WHF 0.083 0.063 **** 0.960 0.968 0.987 0.974 0.984 0.903 0.979 0.953 0.982 0.966 0.828 0.892 0.869 0.878 0.879 0.883 0.875 4 JM 0.073 0.065 0.041 **** 0.983 0.976 0.965 0.957 0.947 0.967 0.978 0.968 0.950 0.890 0.939 0.928 0.902 0.917 0.922 0.905 5 HNE 0.085 0.074 0.032 0.017 **** 0.987 0.994 0.954 0.964 0.969 0.980 0.987 0.945. 0.905 0.964 0.952 0.932 0.942 0.947 0.929 6CV 0.086 0.075 0.013 0.024 0.013 **** 0.996 0.983 0.921 0.984 0.963 0.976 0.963 0.837 0.908- 0.894 0.877 0.882 0.889 0.867 7 BCP 0.100 0.084 0.026 0.036 0.007 0.004 **** 0.963 0.940 0.973 0.957 0.976 0.950 1 0.846 0.926' 0.918 0.903 0.905 0.914 0.889 8 JHO 0.104 0.086 0.016 0.044 0.047 0.017 0.037 **** 0.861 0.968 0.937 0.956 0.935 • 0.780 0.858 0.830 0.829 0.830 0.833 0.814 9 TL 0.087 0.068 0.102 0.055 0.037 0.082 0.062 0.149 **** 0.932 0.945 0.937 0.905 0.927 0.948 0.945 0.916 0.944 0.945 0.940 10 FJ 0.118 0.088 0.021 0.034 0.032 0.016 0.027- 0.032 0.071 **** 0.977 0.980 0.976 0.845 0.901 0.886 0.876 0.888 0.886 0.875 11 FJR 0.127 0.111 0.049 0.022 0.020 0.038 0.044 0.065 0.057 0.023 **** 0.989 0.944 0.924 0.960 0.947 0.934 0.951 0.945 0.938 12DVM 0.118 0.100 0.018 0.033 0.013 0.024 0.025 0.045 0.066 0.020 0.011 **** 0.958 0.895 0.952 0.937 0.942 0.947 0.948 0.939 13MCR 0.166 0.131 0.035 0.052 0.057 0.038 0.051 0.067 0.100 0.024 0.058 0.043 **** 0.825 0.899 0.890 0.873 0.869 0.873 0.863 14 GH 0.188 0.182 0.189 0.117 0.100 0.178 0.168 0.248 0.076 0.169 0.079 0.111 0.192 **** 0.958 0.940 0.896 0.929 0.917 0.933 15SFS 0.187 0.165 0.114 0.063 0.037 0.097 0.077 0.153 0.054 0.104 0.041 0:049 0.107 0.043 **** 0.997 0.976 0.981 0.977 0.971 16SFN 0.213 0.193 0.140 0.075 0.049 0.112 0.086 0.186 0.057 0.121 0.054 0.065 0.117 0.061 0.003 **** 0.981 0.984 0.984 0.973 17 GBR 0.234 0.202 0.130 0.103 0.070 0.132 0.103 0.187 0.088 0.133 0.069 0.060 0.135 0.110 0.024 0.019 **** 0.996 0.995 0.990 18 TME 0.196 0.170 0.129 0.087 0.060 0.126 0.100 0.186 0.058 0.119 0.050 0.055 0.141 0.074 0.020 0.016 0.004 **** 1.001 1.000 19TMW 0.186 0.168 0.124 0.081 0.054 0.118 0.090 0.182 0.057 0.121 0.057 0.054 0.136 0.087 0.024 0.016 0.006 -0.001 **** 0.997 20 WC 0.192 0.165 0.133 0.100 0.074 0.143 0.117 0.206 0.062 0.133 0.064 0.063 0.148 0.070 0.029 0.027 0.010 0.000 0.003 **** Nei's mean intra-taxon genetic distance for T. aprica (0.055) and T.jonesii var. lutea (0.031), and inter-taxon genetic distance 0.119. / 10-FJ 8-JHO 1 13-MGR Figure 3.5. Neighbor-joining tree using Nei's distances (1978) between T. aprica (1-13) and T. jonesii var. lutea (14-20). Bootstrap values above 60% are indicated along their branches. 41 Table 3.9. Soil characteristics comparing T. aprica (N = 23) and T. jonesii var. lutea (N = 7) Feature Taxa Mean (SE) Ranges pH in CaCh T. aprica 7.50 (0.06) 6.9 - 8.06 Tjonesii var. lutea 7.45 (0.08) 7 .12-7 .83 %Sand T. aprica 46.0 (3.19)* 18.15 - 64.49 Tjonesii var. lutea 25.6 (5.58)* 2.2 - 39.11 %Clay T. aprica 20.3 (1.55)** 8.24 - 35.08 Tjonesii var. lutea 35.4 (3.56)** 22.76 - 45.28 %sut T. aprica 33.4 (2.37) 19.37- 57.16 Tjonesii var. lutea 38.9 (2.99) 28.46 -54.23 ppm K T. aprica 6.8 (0.97) 7.36 -294.40 Tjonesii var. lutea 9.5 (2.32) 10.88 - 256.00 ppm Ca T. aprica 19.2 (4.51) 25.28 - 669.50 Tjonesii var. lutea 22.8 (10.69) 19.20 - 512.00 ppm Mg T. aprica 4.0 (0.66) 2.88 - 75.20 Tjonesii var. lutea 4.3 (1.26) 6.24 - 51.20 ppm Na T. aprica 1.9 (0.69) 3.50 - 37.76 Tjonesii var. lutea 1.3 (0.22) 5.50 - 20.00 ppmP T. aprica • . 2.7 (0.45) 1.95 - 89.15 Tjonesii var. lutea 1.9(0.40) 12.34 - 51.58 %CaCo3 T. aprica 13.7(2.20)*** 1.38 - 48.75 Tjonesii var. lutea 47.8 (4.04)*** 32.94 - 59.94 %Gypsum T. aprica 4.0 (0.46) 0.93 - 10.65 Tjonesii var. lutea 5.8 (0.86) 3.81 - 10.60 SAR T. aprica 0.25 (0.03) 0 .09-0 .58 Tjonesii var. lutea 0.37 (0.08) 0.06 - 0.64 %Moisture 1/3 T. aprica 13.7 (1.03) 7.25 - 25.00 Tjonesii var. lutea 17.7 (2.40) 9.04 - 25.88 %Moisture 15 T. aprica 6.9(0.51) 0 .96- 11.40 Tjonesii var. lutea 9.0 (0.77) 4.76 - 10.64 EC (mS/cm) T. aprica 0.95 (0.15) 0 .37 -2 .80 Tjonesii var. lutea 1.0(0.34) 0.46 - 2.40 E C = electrical conductivity, S A R = sodium adsorption ratio, ppm (ng/g soil) p-values are as fol lows; p < 0.05*, p < 0.01 * * , p < 0 .001** 42 The results from measurements of ray and disk pappus show that pappus length characters overlap between both taxa (Table 3.10). However, the overlap is found in the few extremes of the measurements (in parenthesis), while the majority of individuals in each taxa, has pappus length in both disk and ray that does not overlap (in bold). Since overlap was found between the two taxa pappus length, there was no need to count more then just the one individual from each population. If non-overlapping pappus length would have been found, then counts would have continued to get statistical strength for the difference. Table 3.10. Ray and disk pappus measurements of T. aprica and T. jonesii var. lutea. N N ray ray (mm) Ndisk disk (mm) T. aprica 17 52 (0.7) 1 - 2 (3) 26 (3) 4 - 5 (6) Tjonesii var. lutea 9 39 2-3(3.5) 24 (4) 5-6 Majority of measurements for pappus length taken are in bold, and the few extremes of the pappus length found are in parenthesis. 43 Chapter 4 Discussion This study was motivated in part by the fact that Townsendia aprica is being considered for delisting by the U.S. Federal Fish and Wildlife, yet many aspects of the biology of this rare, endemic species are still unknown. Although many new populations of T. aprica have been discovered since its original listing in 1980, many populations are small in size (<100 individuals) with estimates reporting as few as six individuals in some populations. The four more recently discovered southern populations (1-4) are disjunct from the rest of T. aprica's known populations and also geographically isolated from one another. The identity of the southern populations has also been questioned by different government agencies who manage the species, because of difficulty in identifying these plants using the Flora of Utah (Welsh et al., 1987; Welsh et al., 1993). Notably, the ray pappus length appears to be shorter than originally described for T. aprica. Therefore, one reason to initiate a population genetic study was to clarify whether the southern populations are distinct from the remaining populations of T. aprica. Endemic species are likely to be genetically depauperate (Stebbins, 1942), at least relative to their widespread congeners (Gitzendanner and Soltis, 2000; Cole, 2003), and are therefore theoretically less able to adapt to environmental fluctuations (Beardmore, 1983). It was important that patterns of genetic variation in T. aprica be examined before a decision on loss of federal protection occurred. Using isozyme electrophoresis, many populations throughout the range of T. aprica were sampled and I was able to more closely examine some of the agency's concerns and also test the hypothesis that this rare, endemic plant is genetically depauperate. This study also adds to our general understanding of patterns of genetic variation in rare taxa. Karron (1987) and Gitzendanner and Soltis (2000) have pointed out that patterns of genetic variation in rare taxa are best considered in comparison with levels in a more widespread congener. In the absence of a suitable widespread taxon, genetic variation in T. aprica was 44 compared with that of a closely related endemic congener, T. jonesii var. lutea. Molecular phylogenetic data (ETS and ITS sequences) reveal that the three varieties of T. jonesii and T. aprica form a clade. I selected Tjonesii var. lutea for my comparison with T. aprica because both taxa share many characteristics, such as narrow endemism in Utah, along with similar ecology and mostly overlapping morphological features. In addition to comparing the patterns and levels of genetic variation between the two taxa, genetic relatedness was also assessed, along with discernable differences in soil types (Armstrong and Thorne, 1991) and a key morphological difference in ray pappus length (Welsh and Reveal, 1968; Welsh, 1983; Welsh et al., 1987; Armstrong and Thorne, 1991). 4.1 Breeding system The first step towards a better understanding of T. aprica was to determine that all populations are diploid, and have an outcrossing breeding system. This outcome would not only affect the analysis of genetic variation in T. aprica, but the contrary outcome could result in the loss of its federal protection under the U.S. Endangered Species Act (United States, 1972). For example, if T. aprica were found to be an apomictic form of T. jonesii, then its taxonomic status could be called into question, which could result in delisting. When this study was initiated, no information was available about the ploidy level or breeding system of this taxon. Previous evidence from a recent study of the breeding system of T. aprica indicates that populations studied to date are outcrossing diploids (Tepedino et al., 2004). I used two additional lines of evidence: pollen features and isozyme banding patterns, to assess whether there was evidence for variation in ploidy or breeding system (sexual versus apomictic) in T. aprica. Pollen stainability and size, as well as isozyme numbers and banding patterns were consistent with the populations being predominantly sexual and diploid. Furthermore, by extension of previous 45 studies of Townsendia (Beaman, 1957), I concluded that both T. aprica and T jonesii var. lutea are most likely to have an outcrossing breeding system. Apomictic plants produce asexual seed through a variety of developmental pathways. In Townsendia, apomixis is diplosporous (embryos develop from unreduced eggs) and autonomous (pollen is not needed to stimulate endosperm formation). In Townsendia, as in most cases, diplosporous apomixis occurs only in polyploid individuals. Apomicts of this sort tend to have high levels of pollen sterility, perhaps because of relaxed selection for viable pollen. Because of this, levels of pollen viability are taken as an indirect indication of apomixis. Beaman (1957) found that apomictic individuals in Townsendia typically had less than 20% stainable pollen (stainability in cotton blue is considered an indirect measure of viability), along with heteromorphic pollen sizes, with some large viable and some small in-viable grains. None of the twelve individuals from four populations that I examined were found to have less than 20% viability and all pollen was uniform in size. Therefore, these would be considered sexual diploids according to Beaman's criteria. The mean pollen stainability that I observed was 80%, with the lowest levels (57 %) found at the Bristle Cone Pines population (7-BCP; Table 3.1). The reason for lower viability at 7-BCP is unclear, but the population showed no significant inbreeding ( F i s = 0.03; Table 3.3): Jones (1976) attributes differences in percent viable pollen m Aster to environmental conditions, such as cold temperatures or nutrient shortages. Considering that 7-BCP is the highest elevation population known for T. aprica (2750 m), and sets seed later in the season, it is possible that one of the vouchers examined (K. Heil 4376), was collected too early in the season for that year, and therefore all pollen had not fully developed. However, the other voucher examined (M. Porter 3855) was collected in mid June, which is typically a prime time to collect T. aprica seed at 7-BCP, and this collection had even lower stainability than the previous voucher. Given that at 46 least some of the florets from this population had high stainability, developmental timing seems like the most likely explanation for low viability found. The second line of evidence used to describe T. aprica'?, breeding system came from isozyme electrophoresis. By observing duplicate loci and different staining intensities of banding patterns, polyploidy can be detected, and an apomictic breeding system can be inferred if fixed differences between homoelogous loci (fixed genotypes) are found. This inference is strengthened when comparisons are made with known diploid and/ or outcrossing populations (Weeden and Wendel, 1989). The subunit structure (monomer, dimer, tetramer) of each functional enzyme used in this survey is known to be highly conserved, and can be assumed to be constant among species under study (Weeden and Wendel, 1989). On the other hand, the number of isozymes (loci) and their subcellular distribution can vary among species, tissues or developmental stages, and can not be relied on to be consistent even with a species (Wendel and Weeden, 1989; Allphin, 2002). I compared banding patterns within both taxa using a single source of tissue (leaves), collected from tissues of similar age. Therefore, the assumption was made that isozyme number and subcellular distributions should not vary in either taxon if there is no variation in ploidy. In all sampled populations and all enzymes, only the enzyme IDH showed a different number of isozymes than typically found in diploid plants. Two isozymes were resolved in Townsendia, where typically only one cytosolic locus is inferred (Wolf, 1997; Archibald et al., 2001), though other studies (Richardson, 1986) have also resolved two IDH loci for diploid taxa. The remaining enzymes resolved in this study showed no sign of gene duplication and no variation in the number of loci detected among populations. These comparisons included the three known outcrossing populations of T. aprica (9-11; Table TV. 1-Appendix). In addition, none of the populations showed evidence of apomixis. None displayed patterns of "fixed genotypes" across loci or individuals, and no significant linkage disequilibrium was detected. Although the focus of 47 this section was to look for variation in breeding system and ploidy level within T. aprica, it is worth noting that T. jonesii var. lutea also appears to be a diploid outcrosser. All populations share the same number of isozyme loci found in T. aprica, and the single reported chromosome count for T. jonesii var. lutea is 2n=2x=18 from population 18-TME (Table 2.2), which was also surveyed in this study. Thus, with all this evidence taken together, it appears that both taxa are diploid outcrossers. 4.2 Levels and Patterns of Genetic Diversity Overall findings Townsendia aprica and T. jonesii var. lutea both showed higher levels of genetic diversity (as reflected in %P, A, Ae, H 0 and He; Table 3.3 and Table 3.4) than expected on average in endemic taxa (Hamrick and Godt, 1989). Both taxa were found to have genotype frequencies in HWE, as expected for outcrossing species (Hedrick, 2000) and no significant inbreeding was detected in most populations of both taxa (Figure 3.5). The majority of genetic variation was found within populations in both taxa, with only weak evidence of divergence among populations (Table 3.5). Geographic distance between populations accounts for 11% of the among population genetic diversity found in T. aprica, and 40% in T. jonesii var. lutea (Figure 3.3 and Figure 3.4). Al l the loci and alleles screened in this study are shared between these two taxa, with many of the T. aprica populations lacking one or two low frequency alleles (Appendix IV- Table IV. 1). This could be due to factors such as founder or bottleneck events or genetic drift. Indeed, results of the bottleneck test suggest that many of the populations in both taxa have gone through a bottleneck, although there was no obvious correlation between detection of a population bottleneck event and loss of alleles in those populations. Therefore, the overall pattern suggests that each taxon contains reasonable levels of genetic variation likely maintained by gene flow between populations, reflected by low levels of differentiation among populations. 48 Some of the limitations in the genetic part of this study have been encountered in many studies of federally listed species; few populations, few individuals sampled per population, and low numbers of total loci resolved. Berg and Hamrick (1997) state that for population genetic studies to have high statistical power (i.e. for high confidence in estimates of genetic diversity parameters), they should include sampling at least 30 individuals per population for at least 10 polymorphic loci. In this study, 13 T. aprica populations and all 7 known T.jonesii var. lutea populations were sampled, the average sample size per population was 27, and nine enzymes were resolved, with seven consistently polymorphic and scorable. The most significant deviation from the guidelines of Berg and Hamrick is in the number of populations sampled, although my sampling design did cover the known range of the two taxa, including the extreme edges of their ranges. Genetic diversity in T. aprica and T. jonesii var. lutea All populations of both taxa show similar levels genetic variability as reflected by the percent polymorphic loci (%P), mean alleles per locus (A) and mean effective alleles per locus (Ae), and expected heterozygosity (He) (Table 3.3,Table 3.4). For both taxa, levels of genetic diversity were found to be much higher than the means reported in Hamrick and Godt (1989) for locally endemic plant species (P = 40.0%, A = 1.80, A e = 1.15, He = 0.01). Surprisingly, both taxa showed even higher levels of most genetic diversity parameters than reported for widespread species in Hamrick and Godt (1989) (widespread species, P = 58.9%, A = 2.29, A e = 1.31, H e = 0.20 ). Both taxa were also found to have higher total genetic diversity than endemic species (HT = 0.26; Hamrick and Godt„ 1989) but interestingly, the total genetic diversity for T. jonesii var. lutea is Hr = 0.30, which is the same as the mean estimates for narrow species (Hr = 0.30), while T. aprica total genetic diversity (H T = 0.35), falls more in line with mean estimates for widespread species (H T = 0.35; Hamrick and Godt, (1989). 49 Because of the limited availability of leaf material and low enzyme activity in the extreme outer layers of the gels, I was unable to include additional non-polymorphic loci and %P can be taken to be an overestimate of variability because I targeted polymorphic isozymes. Also, %P, A, and A e are considered strongly affected by population sample size because larger sample sizes will tend to pick up more low frequency alleles. Within this study, populations 1-MCC (21 individuals) and 9-TL (15 individuals) had the fewest individuals collected, but neither shows evidence of an effect on these three parameters when compared to all other populations (Table 3.3). Therefore, it may be more meaningful to compare the values of these parameters between the taxa and their populations within this study rather than drawing conclusions based on comparisons of different studies. Townsendia jonesii var. lutea has slightly higher %P and A than T aprica, while T. aprica shows a bit higher A e and expected heterozygosity than T.jonesii var. lutea. Since all alleles were found in all populations of Townsendia jonesii var. lutea but not in all T. aprica populations, %P and A are higher for Townsendia jonesii var. lutea, but this is somewhat surprising since Townsendia jonesii var. lutea has a much more narrow range than T. aprica. The higher values of Ag are expected when higher heterozygosity is found, as in the case of T. aprica. Surprisingly for a rare, endemic plant there was no significant deficiency of heterozygotes or inbreeding found in any populations of T. aprica (Table 3.3), while most of the T. jonesii var. lutea populations did have a deficiency of heterozygotes with the two most geographically near to each other but geographically isolated from the other sampled T. jonesii var. lutea populations having significant inbreeding (Table 3.4). The northern populations of T. aprica (9-12) show the greatest genetic diversity, showing slight excess of heterozygosity, on average with one population (10-FJ) showing significant excess heterozygosity (Table 3.3). The southern populations (1,2, and 4) show the greatest deficiency of heterozygotes in T. aprica, indicating a tendency toward inbreeding. This pattern of genetic diversity and sub-structuring 50 can not be explained just by the physical distance between the three southern populations (13km), when compared to the similar physical distances in four northern populations (17km) (Table II.l-Appendix II), nor can it be explained by small population size (see below). One effect that could explain this pattern might be the greater degree of geographic isolation of southern populations from one another. The landscape in this region appears to be more fragmented, perhaps leading to stronger barriers to gene exchange between populations. Although large populations are expected to maintain higher levels of genetic variability than small populations (Wright, 1931; Kimura and Crow, 1964), this was not the case with T. aprica (Table 3.3). There was no evidence found that random genetic drift had a stronger effect in small populations of T. aprica (<100) (Table 2.1) which did not show an overall tendency towards lower genetic diversity (Table 3.3) or more fixed alleles (Table IV.l-Appendix IV). By comparing the genetic diversity and allele frequencies of three small, geographically isolated populations (1-MCC, 2-MMD, and 8-JHO) to three other small populations which are not as isolated (5-HNE, 9-TL, and 11-FJR), it becomes apparent, that geographic isolation has a greater effect on genetic diversity patterns and allele fixation in T. aprica then the size of a population. Another explanation could be a recent reduction in population size or isolation which would not reflect the historic size of the population or effective population size (Ne) and not enough time has passed for a drop to occur in genetic diversity. Similar high levels of genetic diversity have also been reported in narrow endemics of other genera such as Dedeckera (Polygonaceae) (Nickrent, 1989), Abronia (Nyctaginaceae) (Williamson and Werth, 1999), Daviesia (Brassicaceae) (Young and Brown, 1995), and endemic Asteraceae such as Artemisia molinieri (Torrell et al., 1999), and Achillea millefolium ssp. megacephala (Purdy and Bayer, 1996). Nickrent (1989) points to a strong correlation of breeding system with genetic variation, with the highest levels of polymorphism found in long-lived, outcrossing species. Purdy and Bayer (1996) also speculate that high levels of polymorphism in 51 Achillea millefolium ssp. megacephala relative to the widespread ssp. lanulosa may have to do with the greater reliance of ssp. megacephala on sexual reproduction. Williamson and Werth (1999) points to high genetic variation being found in the endemic Abronia macrocarpa as possibly the results of a recent speciation event involving divergence from a widespread species through long distance dispersal, or possibly an isolation event, where populations were once more widespread, with continuous distributions that have now been broken up. The source of high genetic diversity in both taxa is not entirely clear, although a number of factors could be contributing. First, in the case of T. jonesii var. lutea, as its taxonomic rank assumes, is may still be capable of interbreeding with the other two T. jonesii varieties, most likely T. jonesii var. jonesii, whose distribution extends into the same counties as T. jonesii var. lutea. Collectively, these three varieties could be considered just one widespread species, or possibly part of a larger species complex, as suggested by John Strother (New York Botanical Garden, per. comm.). As for T. aprica, it is also possible that genetic diversity estimates are inflated as a result of a past hybridization event, or present-day hybridization occurring with one or both of the two physically close taxa, T. jonesii var. lutea or T. incana, which both have overlapping flowering times with T. aprica. Townsendia incana grows very close to many populations of T. aprica and was shown by Beaman to hybridize readily in the lab and in natural sympatric situations with other taxa, but preliminary screening of cpDNA has found no indication of shared maternal haplotypes between the two species (Chris Sears, per. comm.). Although T. jonesii var. lutea is thought to have a non-overlapping range with T. aprica, some of the mountainous terrain that separates the two could be hiding some unknown, sympatric or neighboring populations providing a bridge for the exchange of genetic material between the two taxa and contributing to maintenance of the high genetic variability. Another possible reason for the high genetic variation found could be due to a somewhat recent divergence of T. aprica from the T. jonesii group. There are very close similarities found in morphological features and 52 ecological attributes, as reflected by very similar ITS/ETS sequences. Townsendia jonesii var. lutea and T. aprica also share similar levels of genetic variability and patterning with little genetic distance found between the two taxa, but importantly, they do form distinct clusters in the neighbor-joining tree. A preliminary screening of cpDNA also has yet to detect differences between the two taxa. Distribution of genetic variation The fixation index (FST) (Table 3.5) shows evidence of modest levels of differentiation among populations within each taxon. The majority of genetic variation was found within populations, with only slight divergence found among populations of T. aprica (FST = 0.10) and T. jonesii var. lutea (FST = 0.06). The greater divergence found in T. aprica could be due to its larger range, with greater geographical and elevation distances found between many of the populations, including strong geological features that could drive isolation between different sets of populations, but there is not enough data to say. The results of the Mantel's tests support this notion, with a stronger relationship found between genetic diversity and geographic distances within the more restricted T.jonesii var. lutea, (r2 = 0.41 ; Figure 3.4) than in the wider-ranging T. aprica (r2 = 0.11; Figure 3.3). This more obvious geographic structuring of differentiation within T. jonesii var. lutea, may explain why relationships among populations within this taxon are better supported (Figure 3.2) than those within T. aprica, despite lower mean genetic distance among its T.jonesii var. lutea populations (0.061 versus 0.118 for T. aprica). Pollinators appear to be mainly native, solitary bees (Tepedino et al., 2004), which are unlikely to contribute significantly to long distance pollen flow. However, the pappus of Townsendia is barbed, which may contribute to possible animal dispersal of fruits. Hamrick and Godt (1996) found that animal dispersal is associated with lower levels of genetic differentiation than many other dispersal mechanisms. While pollen and seed dispersal distances have not been characterized in these or other townsendias, they appear to have some adaptations for long distance achene dispersal. 53 Patterns of differentiation and isolation by distance are not as clear within T. aprica, perhaps because differentiation between southern (populations 1-4) and northern (populations 8-13) populations obscures finer substructure (note that populations 5-7 have the largest range of elevation in T. aprica (Table 2.1), and it is unclear whether they belong to the northern or southern group). Mean genetic distance between southern populations is 0.107, compared with a mean of 0.127 between northern populations. In contrast, the mean genetic distance between northern and southern populations is 0.153, suggesting that northern and southern groups are somewhat differentiated. Although the northern and southern groups of populations show some evidence of differentiation, these trends do not show up clearly in the neighbor-joining tree (Figure 3.1). Evidence of bottleneck events was detected in most T. aprica populations (2-5, p< 0.05 and 9-13, p< 0.01; Table IV.l-Appendix VI) and Tjonesii var. lutea populations (14, 15, 18, 20, p < 0.05; Table IV.2-Appendix VI). The interpretation of these findings is not straightforward, because all populations had very similar patterns and levels of genetic variation. The bottleneck test used here (Cornuet and Luikart, 1997) compares the expected heterozygosity (based on observed allele frequencies) with the heterozygosity expected if populations are at mutation-drift equilibrium. These patterns could be driven by periodic extinction and re-colonization dynamics, which have been reported to occur within T. aprica. M . Robinson (USFS) documented population sizes of T. aprica over two years (2002-2003; USFS, unpublished). She found both decreases in population sizes as well as some population extinctions within those two years. In fact, 43% of T. aprica locations showed a decrease between 2002 and 2003 with 18 of the 66 sites absent in 2003 and assumed to have died off. It would be helpful to have seed bank information about T. aprica, in order to understand how this might contribute to the observed patterns, though none is available at this time. It is possible that seed banks, as well as dispersal, contribute to re-colonization dynamics in this species. 54 In summary, I found that T. aprica and T.jonesii var. lutea both have overall levels of genetic variation considered high for rare, endemic taxa. The majority of the genetic variation was found within populations but results show relatively low levels of divergence among the populations. Much of the pattern of genetic variation between populations of T. jonesii var. lutea can be accounted for by geographic distances between populations. Patterns of differentiation within T. aprica are more complex, perhaps driven in part by differentiation between northern and southern groups of populations. As Gitzendanner and Soltis (2000) point out, rare or endemic species may have a tendency towards low genetic diversity, but this does not mean that all rare species have low genetic variation. 4.3 Relatedness conclusions Townsendia aprica and T. jonesii var. lutea have shown very slight morphological (Welsh et al., 1987; Welsh et al., 1993) (Appendix V), ecological (Armstrong and Thorne, 1991) and genetic (Thompson et al., unpublished) differences in this and previous studies. However, I have found modest yet consistent genetic differentiation between these two taxa, along with further evidence of consistent ecological differences. The neighbor-joining analysis based on Nei's genetic distance reveals a tendency for populations of T. jonesii var. lutea to cluster together, with weak bootstrap support of 69% (Figure 3.5). Stronger bootstrap support (73%) for grouping ofT.jonesii var. lutea populations was obtained when I used Cavalli-Sforza chord distance (not shown). Looking closely at the branching pattern (Table 3.5) and the allele frequencies (Table IV. 1, IV.2-Appendix I) it becomes apparent that the presence of the MDH-lof allele drives this attraction between the two taxa. The MDH-lcf allele was found in all populations of T.jonesii var. lutea and in populations 5, 9, 11, and 12 of T. aprica. The effect of the MDH-ld allele was examined in more detail by removing M D H - U and adding its frequency to the most common MDH allele. Constructing a 55 tree with Nei's genetic distance based on these new allele frequencies for MDH resulted in a tree in which T. aprica and Tjonesii var. lutea form two separate clusters with a bootstrap support of 72% along the branch that separates them. Therefore, the presence of this allele appears to emphasize the similarity between some populations of T. aprica and those of T. jonesii var. lutea, again underlining their close relationship, but still showing some distinct differences. These differences were also substantiated by the hierarchical analysis of genetic differentiation which revealed greater differentiation between the two taxa (FST = 0.211) than within each taxon ( F P T = 0.137). Soil features provide additional evidence for a distinction between the two taxa. In general, % sand, % clay and % CaCo3 distinguished the soil habitats of the two taxa, with larger ranges of most soil attributes found in T. aprica. It should also be noted, that T. aprica has been repeatedly described as being found on clay or clay-silt soils (Welsh and Reveal, 1968; Welsh et al., 1987; Welsh et al., 1993), where this and earlier studies (Armstrong and Thorne, 1991) show that T. aprica has a higher affinity for sandy soils (Table 3.9). As T. aprica was found to have higher genetic diversity and effective alleles than T. jonesii var. lutea, this could indicate that T. aprica has been able to adapt to a broader set of soil niches than T. jonesii var. lutea. The fact that some T aprica populations lack two alleles present in Tjonesii var. lutea populations (MDH-1J and PGM-lc) raises the possibility that Tjonesii var. lutea could be the progenitor of T. aprica, because progenitors will have a tendency to loose alleles during a founder event. Another explanation could be a buffering against loss of alleles in a bottleneck event in populations of T. jonesii var. lutea or that migrants replenish allele. However, additional lines of evidence, such as from DNA sequence data, would be needed to look more closely at the ancestral hypothesis (Gottlieb, 2004). While the levels of differentiation detected could be used to support the continued recognition of distinct taxa, the genetic data are possibly biased toward higher levels of 56 differentiation because of the decision to focus on scoring polymorphic loci. If we considered these two distinct taxa, then we can be fairly certain that it has been a recent divergence, as postulated by Beaman for all of townsendias. Though the importance of the line of distinction between T. aprica and T. jonesii var. lutea remains open to debate, evidence has been presented of the existence of detectable and significant genetic and soil differences between the two taxa. 4.4 Future Directions With such high genetic variation found, it is important to try to rule out hybridization as a possible contributing factor to the observed patterns. One step to ruling out hybridization is to continue studies of the chloroplast region of the genome, but possibly on a finer scale (more individuals sampled) or increasing the number of loci included in the screening. By using this maternally inherited genome, we can look for detectable differences between T. aprica and T. jonesii var. lutea and T. incana to see if there has been any recent or past hybrid interactions, or to look for evidence that T. aprica arose though a hybridization event. A hybrid origin for T. aprica may seem unlikely, given the absence of unique alleles in T. aprica that do not occur in T. jonesii var. lutea, but how strong this evidence is depends on whether T. incana also shares these alleles with either taxa. In addition to looking in greater detail at phylogenetic data, I would suggest that crossing studies should be done between T. aprica and T. jonesii var. lutea and T. incana to see if they do readily hybridize and if they produce viable and fertile offspring. This would be valuable for a number of reasons. First, it would tell us whether breeding barriers exist between the various taxa, which is important for managers to know, especially in the case of T. incana which co-occurs with T. aprica in a number of localities and overlaps in flowering times. Second, this would be relevant to checking for the possibility that T. aprica originated via hybridization. 5 7 Another future priority would be to sample the very disjunct, and most southern population of Tjonesii var. lutea at Piute Reservoir. The distribution of Townsendia jonesii var. lutea was extended south in 1981 Piute County, (D. Atwood & E. Neese, 7794 NYBG). The original species identification of this disjunct population was as T. aprica. The herbarium voucher collection (NYBG 32226) loaned to the UBC herbarium from New York Botanical Garden is apparently being utilized for the Flora of North America project to write the taxonomic key for T. aprica. There is a note about the ray pappus length (1.5mm) and disk pappus length (4.5mm) on the herbarium specimen, but according to Welsh's description (Welsh et al., 1987; Welsh et al., 1993) that would mean that the ray pappus fits neither taxon, with the disk pappus suggesting it is T. aprica. According to Armstrong and Throne's measurements of Tjonesii var. lutea this specimen would be T. aprica, but they used these sites in their study in 1991 and had the herbarium specimen identified as T. jonesii var. luteal During my field season, we searched this region extensively but we were unable to locate and collect from this important Townsendia population. Obviously, if this population could be found, and sampled, then electrophoresis data could help in its identification. Also, with conflicting results in the bottleneck test, more information on pollen and seed movement, and a study to determine the existence and importance of a seed bank would help to fill in some of the missing pieces to this puzzle. In particular, it is important to understand how and when populations disappear and reappear with environmental fluctuations. The addition of chromosome counts from each population of T. aprica could also be done to confirm and extend my results and those of Tepedino et al.'s (2004) breeding system study. 4.5 Conservat ion considerations In T. apnea's recovery plan, there were three criteria would have to be met for delisting; maintaining 30,000 individuals total for five consecutive years, maintaining 20 populations with 58 a minimum of 5,000 individuals, and ensuring the long-term maintenance of undisturbed habitat. At this stage, I would estimate a maximum total of 4,000- 5,000 individuals, with no more than 10 stable populations and fewer than 500 individuals each. If T. aprica is a diploid, obligate outcrossing species, then describing six individuals or even 50 as a population could give the impression that more long-term stability is present than actually exists. Careful attention to how populations are defined could dramatically change the understanding and the long term recovery plan of this threatened species. As demonstrated in this study, the overall genetic diversity of this species is not in immidiate peril, but careful attention to threats could affect the probability of survival of certain populations. As shown, it is not just the size of the population that affects overall genetic differentiation between all the populations of T. aprica, but the compound factor of geographic isolation. Many of the southern populations (1-MCC, 2-MMD, 4-JM) show signs of lower genetic variation and a deficiency of heterozygotes compared with all other T. aprica populations and should be monitored more closely for size fluctuations (bottlenecks) and more carefully protected from potential threats. Of special concern is 1-MCC because it is easily accessible from the roadside with indications of human threat with fires and camping near this site. It also seems to be the only population found on the unique substrate of Navajo Sandstone. If a re-introduction program were initiated for T. aprica, my suggestion to management agencies at this stage of our understanding would be to use the highly genetically variable northern populations (9-12) for use in establishing new populations, although an initial transplant experiment should be carried out prior any transplanting projects to see if the northern individuals have the same amount of fecundity or vigor when transplanted to any southern population sites. 59 Literature Cited ACQUAAH, G. 1992. Practical Protein Electrophoresis for Genetic Research. Dioscorides Press, Portland, OR. ALLISON, L. E., AND C. C. MOODE. 1965. Carbonate. In C. A. Black [ed.], Methods of Soil analysis Part 2, 1387-1388. American Society of Agronomy, Inc., Madison, WI. ALLPHIN, L. 2002. Morphological and genetic variation among populations of the rare Kachina daisy (Erigeron kachinensis) from southeastern Utah. Western North American Naturalist 62: 423-436. AMOS, W., AND A. BALMFORD. 2001. When does conservation genetics matter? Heredity 82: 257-265. ARCHIBALD, J. K., P. G. WOLF, V. J. TEPEDINO, AND J. BAIR. 2001. Genetic relationships and population structure of the endangered steamboat buckwheat, Eriogonum ovalifolium var. williamsiae (Polygonaceae). American Journal of Botany 88: 608-615. ARMSTRONG, L., AND K. H. THORNE. 1991. Challenge cost share report, target species: Townsendia aprica Welsh and Reveal. Bureau of Land Management, Salt Lake City. BAYER, R. J. 1992. Allozyme variation, genecology, and phytogeography of Antennaria arcuata (Asteraceae), a rare species from the Great Basin and Red Desert with small disjunct populations. American Journal of Botany 79: 872-881. B E A M A N , J. H. 1954. Chromosome numbers, apomixis, and interspecific hybridization in the genus Townsendia. Madrono 12: 169-180. . 1957. The Systematics and Evolution of Townsendia (Compositae). The Gray Herbarium of Harvard University, Cambridge, Mass. BEARDMORE, J. A. 1983. Extinction, survival and genetic variation. In C. M . Schoenwald-Cox, S. M . Chambers, B. MacBryde, and L. Thomas [eds.], Genetics and Conservation, 125-151. Benjamin-Cummings, Menlo Park, California. BERG, E. E., AND J. L. HAMRICK. 1997. Quantification of genetic diversity at allozyme loci. Canadian Journal of Forestry Research 27: 415-424. BROWN, A. H. D., AND A. G. YOUNG. 2000. Genetic diversity in tetraploid populations of the endangered daisy Rutidosis leptorrhynchoides and implications for its conservation. Heredity 85: 122-129. CAVALLI-SFORZA, L. L., AND W. F . EDWARDS. 1967. Phylogenetic analysis: Models and estimation procedures. Evolution 21: 550-570. CLARK, D. J. 1999. Survey Results for Last Chance Townsendia (Townsendia aprica). Capitol Reef National Park, Torry, Utah. 60 . 2000. Survey Results for Last Chance Townsendia (Townsendia aprica). Capitol Reef National Park, Torry, Utah. . 2001. Survey Results for Last Chance Townsendia (Townsendia aprica). Capitol Reef National Park, Torry, Utah. COLE, C. T. 2003. Genetic variation in rare and common plants. Annual Review of Ecology and Systematics 34:213-237. COLLEGE, S. J. 1994. The Effects of Grazing on Threatened/ Endangered Plant Species in the Hartnet and Sandy III Grazing Allotments, Capitol Reef National Park, Utah, Farmington, New Mexico. CORNUET, J. M. , AND G. LUIKART. 1997. Description and power analysis of two test for detecting recent population bottlenecks from allele frequency data. Genetics 144: 2001-2014. D A Y , P. R. 1965. Particle fractionation and particle-size analysis. In C. A. Black [ed.], Methods of Soil Analysis Part 1, 562-566. American Society of Agronomy, Madison, WI. DRURY, W. H. 1974. Rare Species. Biological Conservation 6: 162-169. ELLSTRAND, N . C , AND D. R. E L A M . 1993. Population genetic consequences of small population size: implication for plant conservation. Annual Review of Ecology and Systematics 24: 217-242. FIEDLER, P. L. 1987. Concepts of rarity in vascular plant species, with special reference to the genus Calochortus Pursh (Liliaceae). Journal of Ecology 75: 977-995. GITZENDANNER, M . A., AND P. S. SOLTIS. 2000. Patterns of genetic variation in rare and widespread plant congeners. American Journal of Botany 87: 783-792. GOTTLIEB, L. D. 1982. Conservation and Duplication of Isozymes in Plants. Science 216: 373-380. . 2004. Rethinking classic examples of recent speciation in plants. New Phycologist 161: 71-82. GOUDET, J. 2001. FSTAT, a program to estimate and test gene diversities and fixation indices (version 2.9.3). HAMRICK, J. L., AND M . J . W. GODT. 1989. Allozyme diversity in plant species. In A. H. D. Brown, M . T. Clegg, A. H. Kahler, and B. S. Weir [eds.], Plant Population Genetics, Breeding and Genetic Resources, 43-63. Sinauer Associates, Sunderland, MA. . 1996. Conservation genetics of endemic plant species. Chapman & Hall, New York. HAMRICK, J. L., M . J. W. GODT, D. A. MURAWSKI, AND M . D. LOVELESS. 1991. Correlations between Species Traits and Allozyme Diversity: Implications for Conservation Biology. Oxford University Press, New York. 61 HARTL, D. L . 1987. A Primer of Population Genetics. Sinauer Associates, Sunderland, Mass. HAUSER, J. P., AND J. H. MORRISON. 1964. The cytochemical reduction of nitro blue, tetrazolium as an index of pollen viability. American Journal oj Botany 51: 748-752. HEDRICK, P. 2000. Genetics of Populations. Jones and Bartlett, Sudbury, MA. HOLSINGER, K. E., AND L. D. GOTTLIEB. 1991. Conservation of Rare and Endangered Plants: Priciples and Prospects. In D. A. Falk and K. E. Holsinger [eds.], Genetics and conservation of rare plants. Oxford University Press, New York. HOLSINGER, K. E., R. J. MASON-GAMER, AND J. WHITTON. 1999. Genes, Demes, and Plant Conservation. In L. F. Landweber and A. P. Dobson [eds.], DNA and the Conservation of Biodiversity. Princeton University Press, Chiehester, West Sussex. JONES, A. G. 1976. Environmental effects on the percentage of stainable and presumed normal pollen m Aster (Compositae). American Journal of Botany 63: 657-663. KARRON, J. D. 1987. A comparison of levels of genetic polymorphism and self-compatibility in geographically restricted and widespread congeners. Evolutionary Ecology 1: 47-58. . 1991. Patterns of genetic variation and breeding systems in rare plant species. In D. A. Falk and K. E. Holsinger [eds], Genetics and Conservation of Rare Plants, 87-98. Oxford University Press, Oxford. KENT, M. , AND P. CIOCKER. 1994. Vegetation Description and Analysis A Practical Approach. John Wiley and Sons, New York. KIMURA, M. , AND J. F. CROW. 1964. The number of alleles that can be maintained in a finite population. Genetics 49: 725-738. LANDE, R., AND S. SHANNON. 1996. The role of genetic variation in adaptation and population persistence in a changing environment. Evolution 50: 434-437. LARSEN, E. L . 1927. A revision of the genus Townsendia. Annals of the Missouri Botanical Garden 14: 44. LOPEZ-PUJOL, J., M . BOSCH, J. SIMON, AND C BLANCHE. 2001. Allozyme diversity of two endemic Petrocoptis species: P. montsicciana and its close relative P. pardoi (Caryophyllaceae). Canadian Journal of Botany 79: 1379-1389. MANTEL, N . 1967. The detection of disease clustering and a generalized regression approach. Cancer Research 27: 209-220. M A R U Y A M A , T., AND P. A. FUERST. 1985. Population bottlenecks and non-equilibrium models in population genetics. II. Number of alleles in a small population that was formed by a recent bottleneck. Genetics 111: 675-689. M C C R A K E N , C. L. 2001. Genetic relationships between two rare plant species, Aliciella caespitosa and A. tenuis, and their putative progenitor, Utah State University, Logan, Utah. M.Sc.Thesis. 62 MURPHY, R. W., J . W. SITES, B. D. G., AND C. H. HAUFLER. 1996. Proteins: Isozyme electrophoresis. In D. M . Hill, C. Moritz, and B. K. Mable [eds.], Molecular Systematics, 51-120. Sinauer Associates, Sunderland, MA. NATURE, I. U. F. C. O. 1966. Red Data Book. IUCN, Morges, Switzerland. NEI, M . 1972. Genetic distance between populations. American Naturalist 106: 283-292. . 1973. Analysis of gene diversity in subdivided populations. Proc. National Academy of Science 70: 3321-3323. . 1987. Molecular Evolutionary Genetics. Columbia University Press, New York. NELSON, R. E. 1982. Carbonate and gypsum. In C. A. Black [ed.], Methods of Soil Analysis Part 2. American Society of Agronomy, Inc., Madison, WI. NICKRENT, D. L. 1989. Genetic Diversity in the Rare California Shrub Dedeckera eurekensis (Polygonaceae). Systematic Botany 14: 245-253. NORMANDIN, V., J. KOTUBY-AMACHER, AND R. O. MILLER. 1998. Modification of the ammonium acetate extractant for the determination of exchangeable cations in calcareous soils. Soil Science Plant Anals 29: 1785-1791. NOYES, R. D., AND D. E. SOLTIS. 1996. Genotypic variation in agamospermous Erigeron compositus (Asteraceae). American Journal of Botany 83: 1292-1303. PAGE, R. D. M . 1996. TREVIEW: An application to display phylogenetic trees on personal computers. Computer Applications in the Biosciences 12: 357-358. PIRY, S. G., G. LUIKART, AND J. M . CORNUET. 1999. Bottleneck: a computer program for detecting recent reductions in the erective population size using allele frequency data. Journal of Heredity 90: 502-503. PURDY, B. G., AND R. J . BAYER. 1996. Genetic variation in populations of the endemic Achillea millefolium ssp. megacephala from the Athabasca sand dues and the widespread spp. lanulosa in western-North America. Canadian Journal of Botany 74: 1138-1146. RABINOWITZ, D., S. CAIRNS, AND T. DILLON. 1986. Seven forms of rarity and their frequency in the flora of the British Isles. In M . E. Soule [ed.], Conservation Biology: The Science of Scarcity and Diversity. Sinauer Associates, INC., Sunderland, Mass., USA. RATHCKE, B. 1983. Competition and facilitation among plants for pollination. In L . Real [ed.], Pollination Biology. Academic Press, New York. RAVEN, P. H. 1987. The scope of the plant conservation problem world-wide. In D. Bramwell, O. Hamann, V. Heywood, and H. Synge [eds.], Botanic Gardens and the World Conservation Strategy, 19-29. Academic Press, London. RAYMOND, M. , AND F . ROUSSET. 1995. GENEPOP (version 1.2): Population genetics software for exact test and ecumenicism. Journal of Heredity 86: 248-249. 63 RICHARDS, A. J. 1997. Plant Breeding Systems. Chapman & Hall, London. RICHARDSON, B. J. 1986. Allozyme electrophoresis. Academic Press Australia, North Ryde, N.S.W. ROBINSON, M . 2002. Challenge Cost Share Townsendia aprica Status Report. Robinson Research, Monroe, UT. . 2003. Challenge Cost Share Townsendia aprica Status Report. Robinson Research, Monroe, UT. ROUSSET, F. 1997. Genetic differentiation and estimation of gene flow from F-statistics under isolation by distance. Genetics 145: 1219-1228. SERVICE, U. S. F. A. W. 1985. Biological opinion for the affects of the Hogan Pass Road on Townsendia aprica, the Utah prairie dog, and the bald eagle. Federal Highway Administration, Salt Lake City, Utah. SHULTZ, L. M . 1993. Patterns of endemism in the Utah Flora. Proceedings of the Southwestern rare and endangered plant conference, Santa Fe, New Mexico, 2: 249-263. SIPES, S. D., AND V. J. TEPEDINO. 1993. Pollinator Lost? Reproduction by the Enigmatic Jones Cycladenia, Cycladenia humilis war.joneii (Apocynaceae). SOLTIS, D. E., AND L. H. RIESEBERG. 1986. Autopolyploidy in Tolmiea menziesii (Saxifragaceae): genetic insights from enzyme electrophoresis. American Journal of Botany 73: 310-318. SOLTIS, P. S., AND M . A. GITZENDANNER. 1999. Molecular systematics and the conservation of rare species. Conservation Biology 13: 471-483. SPECIES, I. R. L. o. T. 2004. 2004IUCN Red List of Threatened Species. A Global Species Assessment. IUCN, Cambridge, UK. STEBBINS, G. L. 1942. The genetic approach to problems of rare and endemic species. Madrono 6: 241-272. . 1952. Aridity as a stimulus to plant evolution. Americal Naturalist 84: 401-418. . 1980. Rarity of plant species: a synthetic viewpoint. Rhodora 82: 77-86. TEPEDINO, V. J. 1979. Bee visitation of Phlox bryoides (Polmoniaceae). Great Basin Naturalist 39:197-198. TEPEDINO, V. J., D. D. SIPES, AND T. L. GRISWOLD. 2004. Reproduction and demography of Townsendia aprica (Asteraceae), a rare endemic of the southern Utah plateau. Western North American Naturalist 64: 465-470. THOMSON, J. D. 1978. Effect of stand composition on insect visitation in two-species mixtures of Hieracium. American Midland Naturalist 100: 431-440. 64 TORRELL, M . , M . BOSCH, J. MARTIN, AND J. V A L L E S . 1999. Cytogentic and isozymic characterization of the narrow endemic species Artemisia molinieri (Asteraceae, Anthemideae): implications for its systematics and conservation. Canadian Journal of Botany 77: 51-60. U. S. FISH AND WILDLIFE SERVICE, R. 1993. Last Chance Townsendia Recovery Plan. U. S. Fish and Wildlife, Denver, Colorado. UNITED STATES, C , SENATE, COMMITTEE ON COMMERCE,. 1972. Endangered species conservation act of 1972, Subcommittee on the Environment, 285. U. S. Govt. Print Office, Washington. WALTER, K . S., AND H. L. GILLETT. 1998. 1997 IUCN Red List of Threatened Plants. International Union for the Conservation of Nature and Natural Resources- The World Conservation Union, Cambridge. WEEDEN, N . F., AND J. F. WENDEL. 1989. Genetics of plant isozymes. In D. E. Soltis and P. S. Soltis [eds.], Isozymes in Plant Biology, 46-72. Dioscorides Press, Portland, OR. WEIR, B. S., AND C. C. COCKERHAM. 1984. Estimating F-statistics for the analysis of population structure. Evolution 38: 1358-1370. WELSH, S. L. 1978. Status report: Townsendia aprica. U. S. Fish and Wildlife Service, Denver, Colorado. . 1983. Utah flora: Compositae (Asteraceae). Great Basin Naturalist 43: 179-357. WELSH, S. L . , AND J. L. REVEAL. 1968. A new species of Townsendia (Compositae) from Utah. Brittonia 20: 375-377. WELSH, S. L., N . D. ATWOOD, AND J. L. REVEAL. 1975. Endangered, threatened, extinct, endemic and rare or restricted Utah vascular plants. The Great Basin Naturalist 35: 327-338. WELSH, S. L., N . D. ATWOOD, S. GOODRICH, AND L. C. HIGGINS. 1987. A Utah flora. Great Basin Naturalist Mem. No. 9: 894. . 1993. A Utah Flora. Brigham Young university, Provo, Utah. WENDEL, J. F., AND N. F. WEEDEN. 1989. Visualization and interpretation of plant isozymes. In D. E. Soltis and P. S. Soltis [eds.], Isozymes in Plant Biology, 5-45. Dioscorides Press, Portland, OR. WILKINSON, LELAND, B L A N K , GRANT, GRUBER, AND CHRISTIAN. 1996. Desktop Data Analysis with SYSTAT. Prentice Hall, Upper Saddle River, NJ. WILLIAMSON, P. S., AND C R. WERTH. 1999. Levels and patterns of genetic variation in the endangered species Abronia macrocarpa (Nyctaginaceae). American Journal of Botany 86: 293-301. 65 WOLF, P. G. 1997. Highly differentiated populations of narrow endemic plant maguire primrose (Primula maguirei). Conservation Biology 11: 375-381. WRIGHT, S. 1931. Evolution in Mendilian populations. Genetics 16: 97-159. . 1951. The genetic structure of populations. Annal Eugneics 15: 323-354. Y E H , F. C , Y A N G , R - C , BOYLE, TIMOTHY, B.J., Y E , Z-H., AND M A O , J U D Y X . 1997. POPGENE, the user-friendly shareware for population genetic analysis. Molecular Biology and Biotechnology Centre, University of Alberta, Canada. YOUNG, A. G., AND A. H. D. BROWN. 1995. Comparative population genetic structure of the rare Woodland Shrub Daviesia suaveolens and its common congener D. mimosoides. Conservation Biology 10: 1220-1228. 66 Appendix I Townsendia aprica herbarium vouchers examined Table LI. All known herbarium vouchers for Townsendia aprica. Pop Site Voucher number Herbarium Township, Range, Section Year Pollen Count Ray Pappus Disk Pappus North Q C R 346172 B R Y 22S-5E-31 S W Q 1991 1 4 M C M C O R 330860 B R Y 22S-9E-33 NEQ 1989 1 n/a DV DVM 199193 B R Y 23S-6E-19 N W Q 1979 1-2.5 4 DV E S C L 196940 B R Y 23S-6E-19 S E Q / S E Q 1979 1 4.5 DV DVM 328952 B R Y 23S-6E-33 S W Q 1989 1 4 DV S R S 212698 B R Y 24S-6E-4 1980 1 n/a DV E S C L 199196 B R Y 24S-6E-5 NEQ 1979 1 4 DV RC 196931 B R Y 24S-6E-6 1979 1 5 F J W P 346181 B R Y 23S-5E-35 S E Q / S W Q 1991 1.5-2.5 5.5 F J F J 345948 B R Y 24S-5E-5 1991 F J E F J R 346167 B R Y 24S-5E-6 N W Q 1991 1.5 5.5 F J S F J 346168 B R Y 24S-5E-7&8 1991 1.5-2 5 F J J F R 345949 B R Y 24S-5E-8 1991 F J NHJT 345950 B R Y 24S-5E-16 N E Q / S W Q 1991 F J 28.3 229536/ 119969 M S U /UTC botype 1966 F J 28.3 67350 B R Y Hobtype 1966 F J 28.3 84379/ 32228 B R Y / N Y B G 28.3 miles North of Fremont 1969 2.5 5-Apr F J 28.3 76255/ 32227 LIT/ N Y B G 28.3 miles North of Fremont 1969 F J F J 12401 S J C Fremont Junction 1988 F J F J 284769/ 32225 B R Y / N Y B G 25S-5E-1 1985 1.5 5 TL TL 281778 B R Y 24S-4E-34 W-half 1985 1 5 TL TL 346169 B R Y 24S-5E-29 S W Q 1991 1.2 5 TL TL 168719 B R Y 24S-5E-30 1977 <1 4-5 TL TL 346171 B R Y 24S-5E-31 1991 TL S F J 346170 B R Y 24S-5E-31 S W Q 1991 T L L C C 346180 B R Y 25S-5E-5 S E Q / N W Q 1991 TL L C C 346176 B R Y 25S-5E-6 N W Q 1991 TL TL 187456 UTC S. of Fremont Junction 1984 L C G F O W B 328909 B R Y 25S-7E-18 S E Q 1989 1-1.5 n/a J H JHO 330849 B R Y 25S-8E-16 S E Q 1989 Welsh Welsh 95687 B R Y 24S-7S-30 1971 DC B C P 1906 C R N P 27S-4E-25 1981 DC B C P 9585 S J C 27S-4E-25 1988 X DC B C P 6131 S J C 27S-4E-25 E-half 1986 X HN cvc 1907 C R N P 27S-5E-9 1981 HN cv 8032 S J C 27S-5E-10 S W Q 1987 HN cv 10801 S J C 27S-5E-10 S W Q 1987 HN cv 24361 S J C 27S-5E-10 1988 HN C V R 1905 C R N P 27S-6E-15 1981 HN S D O 3107 S J C South Desert Overlook 1984 1 4 HN S D O 269077 BRY 1 South Desert Overlook 1984 HN S D O 11569 S J C South Desert Overlook 1987 HN S D O 9564 S J C 27S-6E-13 1988 J M J M 8027 S J C 29S-6E-29 S W Q 1987 M M M M / NP 9633 S J C 30S-6E-11 S E Q 1988 X cc M C C 19958 S J C 29S-5E-31 N W Q 1993 X C C M C C 20190 S J C (29S-5E-31 NWQ) 1994 P R P R 32226/ - N Y B G / O F S H 29S-3W-3 1981 2 5 For full names of populations, sites, and herbaria please see Table 1.2 Appendix I continued Table 1.2. List populations, sites and herbaria and their acronyms. Herbarium Site Name BRY Brigharn Young University BCP Bristle Comb Pine CRNP Capitol Reef National Park CV Cathedral Valley MSU Mi±gan State University cvc Cathedral Valley Campground NYBG New York Botanical Garden CVR Cathedral Valley Road OFSH OfFtal Forest Service Herbarium DVM Dog Valley Mine SJC San Juan CoEe^ EFJR East of Fremont Junctbn Reststop UTC Utah State University ESCL East of Sevier County line UT TJrriversity of Utah FJ Fremont Junctbn IFR Jim's Farm Road Pop Name JHO John's Hole Overbok CC Cocks Comb JM Johnsoris Mesa DC Deep Creek LCC Last Chance Creek DV Dog Valley MCC Mini-Cocks comb FJ Fremont Junction MCOR Moore Cut Off Road HN Hartnet MM/NP Miners Moutain / Natbnal Park JH John's Hob NHJT North Holbw Jeep Trail JM Johnson's Mesa OWB Oil Wefl Bench LCGF Last Chance Gas Field PR Pile Reservoir MC Moore Cut Off Road QCR Quitchupah Creek Road M M Miners Mountain RC Rock Canyon TL Type Locality SDO South Desert Overbok SFJ South of Fremont Junctbn SRS San Rafeel Swell TL Type IjocaSty WSW Wilbw Springs Wash WP Windy Peak 68 Appendix II Geographical distance Table II. 1. Geographic distances (km) between the different sampled populations of T. aprica (1-13) and T. jonesii var. lutea (14-20). 1- MCC 2-MMD 3-WHF 4-JM 5-HNE 6-CV 7-BCP 8-JHO 9-TL 10-FJ 11-FJR 12-DVM 13- MCR 1- MCC 2- M M D 8 3- WHF 13 - 5 * * * * -4- J M 10 10 5 **** 5-HNE 27 31 27 23 **** 6-CV 26 35 30 26 10 7-BCP 23 33 28 25 15 6 8-JHO 55 56 54 49 30 35 42 * * * * 9-TL 50 58 54 50 30 26 29 34 **** 10-FJ 55 63 60 55 34 31 34 33 8 * * ** 11-FJR 57 65 62 58 37 33 36 36 10 3 12- D V M 61 67 64 59 38 36 41 27 17 13 15 * ** * 13-MCR 81 82 80 75 55 59 65 27 46 44 45 31 14-GH 76 83 88 82 69 64 59 82 48 49 46 60 94 15-SFS 80 86 86 86 71 64 63 82 49 48 46 59 90 16-SFN 83 96 90 87 72 65 64 78 46 45 42 53 81 17- GBR 70 78 81 74 57 51 51 64 31 29 27 40 67 18- T M E 85 96 91 90 72 64 65 75 45 43 39 51 76 19- T M W 85 95 91 90 71 64 64 75 45 43 39 50 76 20-WC 93 100 105 97 75 72 72 80 51 48 44 54 77 Table II.2. Geographic distances (km) between the different sampled populations T. jonesii var. lutea (14-20). 14-GH 15- SFS 16- SFN 17- GBR 18- T M E 19- TMW 20-WC 14-GH **** 15- SFS 9 **** 16- SFN 18 12 **** 17- GBR 23 18 11 **** 18- T M E 24 18 9 12 **** 19- TMW 24 17 9 13 0 20-WC 33 26 17 21 11 12 **** On Appendix III Isozyme buffer recipes Table III.l. Running buffer recipes used for bot l T. aprica and T. jonesii var. lutea. Lithium-bo rate/ Tris-citrate pH 8.1/ 8.4 45mA / 175-300V / 7hrs Discontinuous Mol. Chemical Electrode buffer 0.188M Boric Acid (H3B03) titrate to pH 8.1 with 0.038M Lithium hydroxide (LiOH) Gel buffer 0.045M Trizma Base (Tris) titrate to pH 8.4 with 0.007M Citric Acid anhydrous 1:9 mixture of EB:GB to make gel with 45ml EB with 405 ml GB Amine-Citrate fMoroholine) pH 8.0 50-60 mA /160-175 V / 4hrs Continuous Mol. Chemical Electrode buffer 0.04M Citric Acid monohydrate titrate to pH 8.0 with N-(3-aminopropyl)-morpholine Gel buffer 1:19 mixture of 22.5ml EB: 427.5 ml of ddH20 ' Table 111.2. Extraction buffer recipe used for leaf tissue of T. aprica and T. jonesii var. lutea. Germanium Chemical Amount Molarity Heat up in micro wav dH20 25 ml Add whenH20 is hot Germanium 0.028 g 0.005 M Add once H20 has cooled Diethydicarb acid 0.091 g 0.021 M f i PVP 40 1.306 g 0.001 M I I Borax 0.37 g 0.039 M I I Sodium bisulfate (meta) 0.1 g 0.039 M 11 Asorbic Acid 1.3 g 0.262 M i i Phophate buffer 2.67 ml 0.160 M Add day of grinding DMSO 2.67 ml 0.136 M T! 2-phemoxyethonol 0.17ml 0.062 M It 2-mercaptoethanol 2 drops 0.010 M Appendix IV Allele Frequencies Table IV. 1. Allele frequencies of Townsendia aprica. Populations 1 2 3 4 5 6 7 8 9 10 11 12 13 Locus/ allele MCC M M D WHF J M HNE CV BCP JHO TL FJ FJR DVM M CK Bottle neck significance P<0.05 P< 0.05 P<0.05 P<0.05 P<0.01 P<0.01 P<0.01 P<0.01 P<0.05 N 21 27 27 27 25 27 27 27 . 15 27 27 27 27 PGM-1 a 0.81 0.72 0.11 0.26 0.14 0.07 0.09 0.50 0.06 0.04 0.02 b 0.12 0.28 0.80 0.65 0.70 0.83 0.87 0.91 0.40 0.78 0.61 0.74 1.00 c 0.07 0.09 0.09 0.16 0.09 0.04 0.09 0.10 0.17 0.35 0.24 6PGDH-1 a 0.48 0.52 0.28 0.26 0.44 0.44 0.57 0.39 0.53 0.37 0.19 0.22 0.20 b 0.29 0.15 0.37 0.44 0.34 0.44 0.35 0.46 0.17 0.44 0.50 0.43 0.32 c 0.24 0.33 0.35 0.30 0.22 0.11 0.07 0.15 0.30 0.19 0.32 0.35 0.48 AAT-l a 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 b AAT-2 a 0.93 0.83 0.85 0.80 0.88 0.87 0.87 0.96 0.73 0.78 0.85 0.87 0.69 b 0.07 0.17 0.15 0.20 0.12 0.13 0.13 0.04 0.27 0.22 0.15 0.13 0.32 MDH-1 a 0.24 0.33 0.19 .0.46 0.16 0.19 0.09 0.46 0.17 0.35 0.37 0.19 0.19 b 0.62 0.43 0.59 0.48 0.52 0.56 0.57 0.44 0.43 0.41 0.39 0.57 0.59 c 0.14 0.24 0.22 0.06 0.18 0.26 0.28 0.09 0.20 0.24 0.24 0.17 0.22 d 0.14 0.06 0.20 0.07 IDH-1 a 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 b IDH-2 a 0.91 0.89 0.91 0.59 0.52 0.78 0.63 1.00 0.33 0.72 0.46 0.57 0.74 b 0.10 0.11 0.09 0.41 0.48 0.22 0.37 0.00 0.67 0.28 0.54 0.43 0.26 G6PDH-1 a 1.00 0.93 0.87 0.94 0.96 0.91 0.93 0.98 0.70 0.63 0.80 0.83 0.59 b 0.07 0.13 0.06 0.04 0.09 0.07 0.02 0.30 0.37 0.20 0.17 0.41 LAP-1 a 0.88 0.78 0.70 1.00 0.86 0.91 0.83 0.80 0.83 0.78 0.82 0.67 0.96 b 0.12 0.22 0.30 0.14 0.09 0.17 0.20 0.17 0.22 0.19 0.33 0.04 Table IV.2. Allele frequencies of Townsendia jonesii var. lutea. 71 Populations 14(1) 15(2) 16 (3) 17(4) 18(5) 19(6) 20(7) Locus/ allele G H SFS SFN GBR T M E T M W WC1 Bottleneck significance P<0.05 P<0.05 P< 0.05 P<0.05 N 27 27 27 27 27 27 27 IX, M l a 0.13 0.04 0.09 0.15 0.24 0.28 0.30 b 0.13 0.69 0.74 0.83 0.61 0.65 0.54 c 0.74 0.28 0.17 0.02 0.15 0.07 0.17 6PGDH-1 a 0.28 0.28 0.26 0.09 0.13 0.15 0.09 b 0.06 0.13 0.20 0.28 0.33 0.37 0.28 c 0.67 0.59 0.54 0.63 0.54 0.48 0.63 AAT-1 a 1.00 1.00 1.00 1.00 1.00 1.00 1.00 b AAT-2 a 0.74 0.87 0.83 0.91 0.83 0.80 0.74 b 0.26 0.13 0.17 0.09 0.17 0.20 0.26 MDH-1 a 0.20 0.19 0.13 0.13 0.17 0.11 0.13 b 0.48 0.46 0.54 0.52 0.44 0.57 0.50 c 0.20 0.20 0.26 0.32 0.32 0.24 0.32 d 0.11 0.15 0.07 0.04 0.07 0.07 0.06 IDH-1 a 1.00 1.00 1.00 1.00 1.00 1.00 1.00 b IDH-2 a 0.19 0.15 0.02 0.06 0.07 0.07 0.11 b 0.82 0.85 0.98 0.94 0.93 0.93 0.89 G6PDH-1 a 0.91 0.98 0.96 0.98 0.94 0.98 0.94 b 0.09 0.02 0.04 0.02 0.06 0.02 0.06 LAP-1 a 0.91 0.87 .0.91 0.59. 0.61 0.63 0.56 b 0.09 0.13 0.09 0.41 0.39 0.37 0.44 Appendix V Representative photos of each taxon T. aprica at 10-FJ yellow flowered. T. aprica at 2-MMD cream flowered. T. lutea var. lutea 14-GH cream flowered. T. lutea var. lutea 18-TME pink flowered. T. lutea var. lutea 18-TME cream flowered. 71 lutea var. lutea 20-WC disk pappus. 7 3 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items