UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Biosystematics of the annual species of the Mimulus guttatus species complex in British Columbia, Canada Benedict, Beverly G. 1993

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

Item Metadata


831-ubc_1993_fall_benedict_beverly.pdf [ 12.9MB ]
JSON: 831-1.0086202.json
JSON-LD: 831-1.0086202-ld.json
RDF/XML (Pretty): 831-1.0086202-rdf.xml
RDF/JSON: 831-1.0086202-rdf.json
Turtle: 831-1.0086202-turtle.txt
N-Triples: 831-1.0086202-rdf-ntriples.txt
Original Record: 831-1.0086202-source.json
Full Text

Full Text

BIOSYSTEMATICS OF THE ANNUAL SPECIES OF THEMimulus guttatus SPECIES COMPLEXIN BRITISH COLUMBIA, CANADAbyBEVERLY G. BENEDICTB.Sc., University of Alberta, 1986A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS OF THE DEGREE OFMASTERS OF SCIENCEinTHE FACULTY OF GRADUATE STUDIES(Department of Botany)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAMay 1993© Beverly G. Benedict, 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(SignatuDepartment of ^The University of British ColumbiaVancouver, CanadaDate  Vi7e.)t ^7 /93DE-6 (2/88)iiAbstractThe M. guttatus species complex consists of several morphologically andecologically distinct taxa, which, for the most part, are able to intercross; this makesit both an evolutionarily fascinating group to study and a taxonomic nightmare. Inthis study I found that on one ecologically diverse hillside on Vancouver Island,British Columbia (B. C.), Canada, three distinct species of the M. guttatus speciescomplex grow sympatrically. These three species included: 1) M. guttatus, a largeflowered, facultatively annual diploid 2) M. nasutus, a small flowered, annual diploid,and 3) a small flowered, annual tetraploid to which a previous name could not beapplied. This taxon has been named M. queue . In addition to the B. C.populations, M. nasutus and M. guttatus populations from California and M. queuepopulations from Oregon were also studied.Pre-zygotic crossing barriers reduce gene exchange between M. nasutusand M. guttatus. Both pre- and post-zygotic crossing barriers have effectivelyisolated M. queue from M. guttatus and M. nasutus. Under most growingconditions, the three taxa can be distinguished morphologically. However, whencollected from dry habitats the two small flowered species are difficult to separate.Nineteen different measurements of the flower and calyx were taken. The ranges ofall 19 measurements were overlapping for M. nasutus and M. queue. Medianswere significantly different for calyx measurements, the difference between calyxlength and pistil length and the difference in length of the upper and lower corollalobes. Boxplot comparisons of the B. C. and California populations of M. nasutusrevealed no significant differences in the medians of the floral measurements. Onlyone floral measurement was found to be significantly different when the Oregonpopulations of M. queue were compared to the B. C. populations of M. queue.Discriminant and principle component analysis grouped individuals moderately wellinto the proper species. Discrimination was poorest when very small flowers wereused.iiiEnzyme electrophoresis indicated distinct differences in genetic diversity and inthe compartmentalization of the diversity between the primarily autogamous M. queueand M. nasutus, and M. guttatus with its mixed mating system. The southernpopulations of M. nasutus and M. queue exhibited somewhat greater gene diversitythan the northern populations. The presence of fixed heterozygosity in all populationsof M. queue indicated that it is likely an allopolyploid. Whether the B. C. and Oregonpopulations are directly related or independently derived could not be determined fromthe present data. More collecting is required to determine the geographic distribution ofM. queue and to shed light on its evolutionary past.ivTable of ContentspageAbstract ^  iiList of Tables vList of Figures^ viAcknowledgement viii1. Introduction1.1 Taxonomy and Distribution ^ 11.2 Chromosomal Patterns 31.3 Breeding Systems^ 51.4 Enzyme Electrophoresis 61.5 Objectives^ 72. Materials and Methods2.1 Collection Sites ^ 82.2 Crossing Studies 92.3 Morphological and Developmental Data^ 102.4 Chromosomal Studies^  122.5 Enzyme Electrophoresis  123. Results3.1 Crossing Studies^ 263.2 Morphological Analysis 343.3 Field Observations and Phenology^ 603.4 Chromosome Number^ 693.5 Enzyme Electrophoresis 714. Discussion^ 965. Literature Cited 116Appendix I^  121Appendix II  123VList of Tablespage2.1 Collection Sites in B.C. ^ 152.2 Collection Sites in California and Oregon^ 162.3 Buffer Systems Used^ 242.4 Description of Enzymes and Staining Procedures^ 253.1 Within Species Crossing Results^ 313.2 Between Species Crossing Results 313.3 - 3.4 Component Loadings for PCA Analysis^ 523.5 Canonical Loadings for Discriminant Analysis^ 543.6 - 3.8 Allelic Summaries by Population - 6PGD and IDH ^ 80-813.9 - 3.11 Allelic Summaries by Population - DIA, TPI and PGI ^ 82-833.12 - 3.14 Allelic Summaries by Population - AAT^ 84-853.15 Summary of Alleles in M. queue Populations^ 873.16 Alleles Found in M. alsinoides^ 923.17 Genetic Variation by Population 943.18 Genetic Variation by Region and Species^ 95viList of Figurespage2.1 Nanoose Hill Habitat^ 172.2 Map of Species Distribution and California/OregonPopulations^ 182.3 Map of B.C. Populations^ 192.4 Floral Measurements Illustration^ 202.5 Calyx Measurements Illustration 212.6 Difference in Corolla Lobe Length Illustration^ 222.7 Notched Boxplot Interpretation Key^ 233.1 Crossing Triangle^ 303.2 Pollen Viability of B.C. Plants^ 323.3 Pollen Viability of Crosses including M. queue^ 333.4A Species Comparisons: Growth Chamber Grown^ 413.4B Species Comparisons: Field Collected ^ 413.5 - 3.7 Boxplot Comparisons of Flower and Calyx MeasurementsM. guttatus, M. queue, and M. nasutus^ 42-443.8 - 3.10 Boxplot Comparisons of Flower and Calyx MeasurementsRegional Differences - M. queue, and M. nasutus^ 45-473.11 - 3.12 Boxplot Comparisons of Flower Part Ratios -M. queue, and M. nasutus^ 48-493.13 PCA of Flower and Calyx Measurements - M. guttatus,M. queue, and M. nasutus^ 503.14 PCA of Flower and Calyx Measurements -M. queue, and M. nasutus^ 513.15 Photo Illustration of Relative Pistil and Calyx Lengths -M. queue, and M. nasutus^ 533.16 Discriminant Analysis 54viipage^3.17^Boxplot and BargraphComparisons of Corolla Lobe Differences^ 553.18^Corolla Shape Differences - M. queue, and M. nasutus^ 563.19^Young Plants of M. queue, and M. nasutus^ 573.20^Examples of M. guttatus, M. nasutus and a Hybrid^ 583.21^Boxplot Comparisons of Flower and Calyx MeasurementsM. guttatus, M. nasutus and their Hybrids^ 593.22^M. guttatus and M. queue Flowering in the Field^ 643.23^Boxplot and Bargraph of Days to Germination 653.24^Boxplot and Bargraph of Days to Flower^ 663.25^Boxplot of Germination Percentages 673.26^M. queue Prior to Anthesis^ 683.27^Chromosome Counts 703.28^PGI Electrophoretic Banding Patterns^ 863.29^AAT Electrophoretic Banding Patterns 883.30^6PGD Electrophoretic Banding Patterns^ 893.31^DIA Electrophoretic Banding Patterns 903.32^IDH Electrophoretic Banding Patterns^ 913.33^TPI Electrophoretic Banding Patterns 913.34^MDH Electrophoretic Banding Patterns^ 933.35^PGM Electrophoretic Banding Patterns 93VIIIAcknowledgmentsI would like to thank my advisory committee, Gerald B. Straley, W. B. Schofieldand particularly my major advisor, Fred R. Ganders for their advice, and insights duringthe course of this study. For generously allowing me the use their equipment I wouldlike to thank Bruce A. Bohm, the UBC Botanical Gardens and Michael Weis. Fred R.Ganders, Helen Kennedy, Alan Reid and Ken Marr provided expertise and support inthe field while Stewart Schultz was very helpful in the lab. I would like to thank JackMaze, Stewart Schultz and Tom Wells for sharing with me their statistical insights, P. E.Harding for assistance with the Latin and Dion Durnford for his help throughout theproject. My thanks also to the Royal British Columbia Museum Herbarium staff for theirassistance and the herbarium at the University of Utah for their loan of Mimulusspecimens. Finally, funding for this project was provided through a University of BritishColumbia Graduate Fellowship and a Natural Sciences and Engineering ResearchCouncil of Canada grant to Fred R. Ganders.1Chapter 1 - Introduction1.1 Taxonomy and DistributionConsiderable research has been conducted on Mimulus guttatus Fischer exDC. (Scrophulariaceae). It is typically a fugitive species and although native towestern North America, has become naturalized in temperate regions around theworld from Britain (McArthur 1974) to New Zealand (Grant 1924). Traits whichmake it a good colonizer, such as abundant seed production, mixed mating systemand short life cycle coupled with the ease with which it is cultivated also make M.guttatus a good research tool. Mimulus guttatus is part of a species complexconsisting of a group of about 6-12 closely related species, subspecies or varieties.It therefore lends itself to studies of the early stages of speciation and the forcespromoting and limiting speciation.At least 74 different name combinations have been applied to taxa in thegroup since it was initially described in 1813 (Grant 1924, Campbell 1950, Pennell1951). Most of the names have been applied to California material indicating thatCalifornia is the center of diversity for the M. guttatus complex. This also reflectsthe proportionally large amount of work and collecting which has occurred inCalifornia.The number of taxa recognized ranges from four in the most recent treatment(Thompson 1993) to twenty-three (Pennell 1951). This discrepancy reflects, in part,the lack of complete crossing barriers between the species as shown by Vickery's(1974B) extensive crossing experiments. Natural hybridization between severaldifferent pairs of species has been observed in the field (Vickery 1978, Dole 1991).Natural hybrids also have been observed between M. guttatus and other membersof section Simio/us such as Mimulus tilingii, M. glabratus var. utahensis (Lindsayand Vickery 1967) and in Britain, M. luteus and M. cupreus (McArthur 1974).Another factor contributing to the taxonomic confusion is the high degree ofmorphological plasticity exhibited by M. guttatus and its allied species. Vickery2(1974A) was able to demonstrate significant environmental influence on both leafsize and leaf shape of M. guttatus.The M. guttatus species complex is distributed from the Aleutians in Alaskato northern Mexico and from the Pacific Ocean to the Rocky Mountains. Thedistribution of individual populations is patchy as they are confined to sunny, wet orseasonally wet sites. If these requirements are met M. guttatus has shown aremarkable capacity to adapt to different and sometimes adverse growingconditions. It has been found growing from sea level in the salt spray zone to subalpine habitats. Ecotypes have been collected on serpentine soils, copper tailingsand around hotsprings (Pennell 1951, Macnair 1983, Pinder-Moss and Pojar143223 (UBC), Szczawinski 93713 (UBC), Delmer 1972). In addition, populationsmay be either perennial, facultatively annual or annual. Vickery (1978) hassuggested that M. guttatus is a typical fugitive species, a good colonizer with smallisolated populations which undergo periodic extinction and recolonization. Hesuggests that the gene flow caused by these frequent and often long distancerecolonizations may act as a binding force, limiting more complete differentiationand speciation within the complex.Species complexes in Mimu/us seem to be the rule rather than the exception,particularly in section Simiolus. Similar taxonomic problems also exist in thespecies complexes of Mimulus glabratus (eastern North America), M. tilingii(mountain monkeyflower) and M. luteus (South America) (Vickery 1974B). Othersections of Mimu/us such as Diplacus (e.g., M. calycinus) and Erythranthe (e.g., M.cardinalis) also exhibit similar patterns (Beeks 1962, Vickery 1974B).1.2 Chromosomal PatternsBoth aneuploidy and polyploidy are common in the genus Mimulus and bothhave been found in the M. guttatus species complex. The basic haploidchromosome number for the complex is n=14. This may potentially represent an3ancient tetraploid as chromosome counts of n=8 have been found in 3 differentsections of Mimu/us and n=7 characterizes section Mimulastrum (Vickery 1978). Inthe M. guttatus complex, M. platycalyx Pennell is distinguished by counts of n=15although there is speculation that the extra chromosome is simply an accessory orB chromosome. Some populations of M. nasutus Greene have n=13 chromosomeswhile others have the typical n=14 counts. Aneuploidy does not seem to representa barrier to crossing. Vickery (1974B) reported that one of the n=13 M. nasutuspopulations was almost completely genetically isolated from the rest of the complex,however other n=13 populations did not differ from the n=14 populations in theircrossability with M. guttatus .Several populations of tetraploids have been discovered, none of which arefrom California. Vickery located populations in Arizona (2), Colorado (2), NewMexico (3), the Columbia River area of Oregon (1) and Alaska (1). None of thesouthern tetraploid populations have been reported to be morphologically distinctfrom the typical diploid M. guttatus ( McArthur et al 1972, Vickery 1978). Calder andTaylor (1965) collected and named a tetraploid from the Queen Charlotte Islands(M. guttatus ssp. haidensis) which differed morphologically from adjacent diploids inthat it was puberulous, nonglandular and grew at higher elevations than the localdiploid M. guttatus. No further work has been done on M. guttatus ssp. haidensis.McArthur et al. (1972) suggest that this name may be applicable to the northerntetraploids in their study as they were also somewhat distinct morphologically.Vickery et al. (1968) suggest that De Candolle's type for the M. guttatus specieswhich was collected in Unalaska, Alaska, may have been a tetraploid.The chromosomes of M. guttatus are very small, from 0.5 to 3.0 pm(McArthur 1974), and pairing is difficult to visualize. Nevertheless, several studieshave been done on meiotic chromosome pairing behavior of M. guttatus hybridsand of M. guttatus tetraploids (Mukherjee and Vickery 1962, Vickery et al. 1968,McArthur et al. 1972). McArthur et al. (1972) suggest that the southern tetraploids4are of autopolyploid origin based on an average of 6 (range 3-9) quadrivalents atmeiosis, their morphological similarity to diploid M. guttatus, and the fact that two ofthese populations had mixtures of tetraploid and diploid plants. An earlier study,however, discovered no evidence of quadrivalent configurations in an Arizonatetraploid population (Mia et al. 1964). Given the location of the southerntetraploids, it is possible that they may be the result of an interspecific hybridizationand subsequent polyploidization between M. guttatus and a member of either the M.glabratus or M. tilingii species complexes. Two of the New Mexico tetraploidpopulations are reported to be from locations which must be very close to diploid(n=15) populations of M. glabratus var. fremontii (Benth.) Grant (McArthur 1972,Vickery, 1978). In addition, a tetraploid (n=30) population of M. glabratus var.fremontii from Texas was found by Vickery (1974B) to exhibit some morphologicalsimilarities with M. guttatus as well as enhanced crossability with it. However, nonatural hybridization has been observed by Vickery (1978) in Chihuahua, Mexicowhere the two diploid species are frequently sympatric. Another interestingpolyploid species, M. tilingii var. corallinus (n=24), is suspected to be anallopolyploid derivative of M. guttatus and a member of the M. tilingii speciescomplex (n=14, 28) (Vickery et al. 1968).It is speculated that both auto and allopolyploidy may allow species tobecome established beyond the diploid's boundaries through the maintenance ofgreater heterozygosity (Roose and Gottlieb 1976). Vickery (1978) found that boththe tetraploid and aneuploid populations he studied had a wider range ofgermination temperature tolerance than the diploids. The range of the tetraploidpopulations of M. guttatus, near the periphery of the species range, also indicatessuch a scenario.51.3 Breeding SystemsThe typical mixed mating form of M. guttatus has large yellow corollas withred anthocyanic spots on the lower lip which may function as nectar guides. Kiang(1972) observed that for two California populations of M. guttatus the most commonpollinators were Bombus spp. (bumblebees). The stigma of M. guttatus is sensitiveand the two lobes will close together upon being brushed or touched. Dudash andRitland (1991) speculate that this may be a mechanism to reduce selfing aspollinators back out of the flower. Stigma closure was found to be highly variableamong different flowers on the same plant and also varied with the type ofstimulation (Ritland and Ritland 1989).Although M. guttatus has many features to promote outcrossing it is alsocapable of substantial self pollination. Southern B. C. and northern Washingtonpopulations have estimated selfing rates ranging from 24 to 59% (Ritland andGanders 1987). Self pollination in large flowered M. guttatus has been shown tooccur as the lower stigmatic lobe curls down bringing it in contact with the anthersand/or during corolla abscission as the anthers are dragged over the stigma (Dole1992). Self fertilization is frequently associated with colonizing species which 1) aregenerally annuals and therefore need to produce many seeds each generation and2) are favored if they are able to form a population from a single individual (Baker1955).At least three different taxa, M. nasutus, M. cupriphilus and M. platycalyx, inthe M. guttatus species complex are predominantly selfers (Kiang 1972, Macnair1989, Dole 1991). Ornduff (1969) documented several morphological changes thatcan accompany the evolution from a mixed or outcrossed mating system to apredominantly selfing one. The most frequently reported are 1) the anthers becomeadjacent to the stigma 2) the pistil becomes shorter 3) the style is included in thecorolla, 4) stigma receptivity and anther dehiscence become synchronous and 5)the petals are smaller (Wyatt 1988).6Ritland and Ritland (1989) found that the two most highly selfed taxa in theirstudy, M. platycalyx and M. micranthus, had the shortest stigma, ovary and anthersac lengths, and were the only taxa with styles shorter than stamens. They alsohad the least proportion of their biomass allocated to what was deemed the malecomponent (stamens and corolla) and the smallest pollen to ovule ratios. However,other taxa such as the large flowered M. nasutus and M. laciniatus also had highselfing rates although the stigma/anther separation was quite large. Estimates ofselfing rates for M. nasutus were 68% for a large flowered form and 84% for whatwas identified as M. micranthus Heller (Ritland and Ritland 1989) but which greatlyresembles a small flowered M. nasutus. Ritland and Ritland (1989) estimatedselfing rates for a single population of M. platycalyx at 70%.1.4 Enzyme electrophoresisEnzyme electrophoresis has become a common and important tool in planttaxonomy. It is a relatively simple method by which different alleles for a specificprotein can be detected by their charge differences. It must be recognized thatenzyme electrophoresis underestimates the number of allelic differences as onlyabout 30% of the nucleotide replacements in a gene will result in the substitution ofan amino acid with a different net charge (Gottlieb 1977). The proteins used arelargely those acting in pathways of primary metabolism and are usually conservedfor the number of loci expressed and in their subunit structure. Allozymes makeconvenient genetic markers because they are inherited in a codominant manner,represent single loci and are for the most part free of pleiotropic and epistaticinteractions (Weeden and Wendel 1989). In M. guttatus, starch gel electrophoresishas been used to examine various parameters of the mixed mating system (Ritlandand Ganders 1987, Ritland 1989, Dudash and Ritland 1991, Dole 1991, Fensterand Ritland 1992). Very little allozyme work has been done on M. guttatus withregard to taxonomic questions. McClure (1973) studied enzyme variability of7populations from the North Yuba River drainage area in California and Vickery et al.(1989) used M. guttatus as an outgroup for allozyme studies in the sectionErythranthe.1.5 ObjectivesThe first goal of this project was to determine the mechanism(s) that allowed twoflower size morphs of the M. guttatus species complex to be maintained sympatrically,without evidence of intermediates, in certain populations on southern Vancouver Islandand the Gulf Islands in southwestern B.C., Canada. Seeds collected from populationsin this area were germinated and the resulting plants crossed. It was subsequentlyfound that three taxa of the M. guttatus species complex were growing in this area,sometimes sympatrically. These included the large flowered, facultatively annual M.guttatus, and two small flowered, predominantly selfing taxa. One of the small selfershas been identified as M. nasutus and the other is a novel species which has beennamed (M. queue) and described in this paper. The perennial coastal populations ofM. guttatus were not used in this study because their distribution did not overlap that ofeither the facultative or obligate annual members of the M. guttatus species complex.The second part of this study was directed at: 1) documenting the morphologicalchanges that have accompanied the shift from a mixed mating system to anautogamous one in both M. queue and M. nasutus, 2) describing M. queue anddifferentiating it from M. nasutus using morphology, crossing behavior, chromosomes,phenology and allozymes 3) using allelic diversity parameters to compare localpopulations of M. guttatus, M. queue and M. nasutus with California and Oregonpopulations of the same species and outcrossing taxa with selfing taxa and, 4)shedding some light on the evolutionary history of M. queue.8Chapter 2 - Materials and Methods2.1 Collection SitesTables 2.1 and 2.2 provide exact localities for the populations used in this study.All B. C. populations were either from open, seepy hillsides that dried up during thesummer months (as in Figure 2.1), or from sites in sand or among rocks near theocean. The California/Oregon populations were situated either by ephemeral streamsand waterfalls or on seepy slopes. All of the plants used in this study were collectedfrom sites where they grew as annuals or winter annuals. Mimulus nasutus grew mixedwith M. guttatus at three out of the four sites where it was collected. Population 90-08was the only site where M. nasutus was the only species collected.Figure 2.2 depicts the natural range of the M. guttatus species complex and theapproximate locations of the Oregon, California and B. C. populations. A closer look atthe locations of the B. C. populations is provided in Figure 2.3. Although it has beenincluded with the B. C. group of populations, population 90-13 is actually in northwestern Washington State.The plants used as the parental generation in experimental hybridizations weregrown either from field collected seed (B. C. populations) or from selfs of transplants(California/Oregon populations). Field and growth chamber voucher specimens aredeposited in the UBC herbarium.Individuals used in pollen analysis, morphological measurements and for thecrossing and enzymatic studies were grown in a growth chamber under a regime of 16hour days at 18°C and 8 hour nights at 12°C. During the summer of 1991 hybridsurvivorship data were collected for plants grown under natural light conditions in agreenhouse. Seeds were germinated by sprinkling them on the surface of a moistenedblend of one part SunshineTM potting mix to one part vermiculite. Containers werecovered with plastic wrap prior to germination to prevent seeds from being blownaround and to maintain high humidity. If plants were kept longer than two months they9were fertilized weekly with 20:20:20 fertilizer. A 2% w/v agar (Difco) solution was usedto germinate some partially shriveled seeds which had failed to germinate on pottingmix. They were transplanted to potting mix approximately five days after germination.2.2 Crossing StudiesEmasculation was required for M. queue and M. nasutus when they were thefemale parents of a cross. The first four flowers and older flowers were avoided in bothM. queue and M. nasutus as they were often cleistogamous, very small and difficult toemasculate. Emasculation was done before the flower opened, about 1-2 days beforeanther dehiscence and 2-3 days before the stigma was receptive. The large floweredM. guttatus did not need to be emasculated prior to pollination as long as newly openedflowers in which anther dehiscence had not yet occurred were used. Plants of M.guttatus were also selfed. A limited number of crosses were performed with M.platycalyx . Seeds required about a month to ripen and were harvested promptlythereafter to avoid seed loss.Crosses were then divided into the following classes according to the seed theyproduced: 1) good seed - plump obviously viable seed or slightly shriveled seed thatgerminated producing a flowering Fl, 2) shriveled seed - highly shriveled seed whichfailed to germinate or moderately shriveled seed that germinated but the germinantsdid not survive to adulthood and 3) no seed.Approximately ten seeds from each cross were planted and 3-5 individualsgrown to flowering. Pollen viability of the Fl hybrids was then tested. Preliminaryinvestigation indicated that lactophenol cotton blue staining gave comparable results tothe fluorochromatic (FCR) test procedure described by Heslop-Harrison et al. (1984).Lactophenol cotton blue was consequently chosen to test pollen viability because of itssimplicity and widespread use. Pollen grains which were severely shrunken or did nottake up the stain were scored as inviable. Pollen viability was tested only for F1s and10for the parental generation from the B.C. populations. Plants from the F1 generationwere self pollinated and the seed collected but not germinated.A limited amount of information was collected on pollen diameter.Measurements were made utilizing a compound microscope with an ocular scale whichwas calibrated with a micrometer slide. Calculations for each group were based on asample of three plants and at least six normally shaped randomly sampled pollengrains from each plant.2.3 Morphological and Developmental DataFlower morphology details were collected from both early and late flowers asflower size is somewhat plastic even for the outcrossing M. guttatus . No vegetativecharacters were scored because they were highly modified by the environment andalthough plants were grown in a growth chamber, conditions were not uniform, a resultof overcrowding and insect infestations. Fifteen floral and calyx measurements weretaken from fresh material and four other distances calculated from these. Lengthdifferences were made only for valid distances, i.e., only if the lengths of the twovariables were approximately parallel and the same initial reference point was used fortheir measurements. Although both sets of stamens are attatched to the corolla tubeslightly above the base, their measurement was taken from the base of the corolla sothat their position in the flower was reflected relative to other flower parts.Some individuals were sampled twice, once at an early stage of flowering andagain at a later stage. Every individual used was the offspring of a different fieldcollected plant. The boxplots, PCA and discriminant analysis were calculated usingcombined data from B. C. and California/Oregon populations. The measurementsmade are depicted schematically in Figures 2.4 and 2.5. Another measurement, thedifference in lengths of the upper and lower corolla lobes, was taken only for asubgroup of individuals and is depicted in Figure 2.6. Three additional measurementstaken are not depicted here. These are pedicel length, gynoecium stipe length and the11width of the calyx as measured by the distance between the two lateral calyx lobes ona dorsiventrally flattened calyx.Floral measurements were analyzed initially using boxplots. Parametricstatistics such as t tests were not used for the flower measurement data because veryfew of the variables were normally distributed and transformations of variables werenever successful for all three taxa. Boxplots indicate the median of a group ofmeasurements. If the notches of the boxplot are not overlapping then there is a 95%probability that the medians of two groups of data are significantly different. A key tofurther interpretation of box plots is provided in Figure 2.7. Three of the characterswere also analyzed by subtracting their logs to approximate a ratio. Thus theadvantages of ratios as a means of eliminating size differences and allowingdifferences in shape to be elucidated has been maintained while eliminating theirassociated statistical difficulties (Atchley et al 1976, Hills 1978).All the calculations as well as the notched boxplots were performed withSYSTAT and SYGRAPH for the Macintosh (Wilkinson 1990). The remaining graphswere generated by KaleidaGraph 2.1.3. Error bars in the bargraphs are the standarderror values for each category. Individuals from population 91- 05 that were suspectedto be part of a M. guttatus/M. nasutus hybrid swarm were eliminated from the PCA anddiscriminant analysis. All variables were standardized for the PCA and the discriminantanalysis. Several of the variables were dropped from the multivariant analysisbecause the information they provided was redundant as indicated in PCA by the firstthree principal component loadings and in discriminant analysis by the canonicalloadings. Other indications of redundancy among variables in PCA were negativeeigenvalues.Germination and flowering time observations were taken on average every twodays after planting. Plants from which seeds did not germinate after one month wereeliminated from the study. Where possible, ten seeds were planted for each field12collected plant. Flowering data were collected from the first germinant to flower.Thinning was performed at an early stage to prevent crowding.2.4 Chromosomal StudiesChromosome counts were taken from meiotic cells. Buds were collected 6-9hours after the lights came on in the growth chamber, and were fixed in 3:1 95%ethanol:glacial acetic acid for four hours. They were then rinsed and stored in 70%ethanol at 3°C until squashes were done. Anthers were dissected out of the buds inacetic acid and squashed with a rusty nail in aceto-carmine until they were wellmacerated and the solution had taken on a purplish tinge. Equal parts of Hoyer'smounting solution was added, and a coverslip placed upon it. The slide was warmedover a flame for a few seconds and then the preperation was squashed under a papertowel. Pictures were taken with a Zeiss Photomicroscope II using T-max 100 film.2.5 Enzyme ElectrophoresisVigorously growing leaf tissue from both seedling and adult plants was used forthe starch gel electrophoresis. Approximately 0.75 cm 2 of leaf tissue was ground with3-4 drops of grinding buffer #1 (Wendel and Weeden 1989) with a pestle in a 12 welledceramic dish. Several wicks (aproximately 3.5 by 13 mm) of 3 mm Whatman paperwere added to each well. Grinding and wicking were both done on ice. The wickswere blotted and were quickly loaded into chilled gels, approximately three cm fromone end. Every sixth wick was a known standard and 30-36 wicks were loaded per gel.Gels were placed immediately in the cooled electrophoresis chamber and a currentapplied so the proteins migrated across the gel. At least two wicks from every samplewere saved by wrapping in plastic and aluminum foil and placing in the -90 freezer forreuse if necessary.Table 2.3 provides a summary of buffers, run conditions and stains used. Gelswere made with 11.5% starch (Sigma) and 4% sucrose. The current was removed13when the red food dye markers had run past the end of the gel. Immediately thereafterthe gel was manually sliced using 3.6 kg fishline and plastic guides. The top slice wasnot used and the bottom slice was used only if there had been a problem with one ofthe 4 middle slices. The photoreactive ingredients and agar were added to the stainsolutions at this time, poured onto the slice and incubated in the dark. The referencesfor the stain recipes used in this study and their incubating conditions are provided inTable 2.4.Allele frequencies were used to determine the individual locus diversity (dl) bythe following formula 1-1 x2 where x represents the allele frequency as a proportion ofthe total for the locus. DI represents the average for dl over all ten loci for a population.Total (species) gene diversity(Ht) was calculated individually for both the local andCalifornia/Oregon species groups and for each species as a whole by the formula1-1, R2 . Average within population gene diversity (Hs) was calculated by averaging theDi for each population over all the loci. The average gene diversity among populationsDst was calculated by the formula Ht = Hs + Dst. Finally, to allow for bettercomparisons of between population gene diversity between species with very differentHt values, the coefficient of gene differentiation (Gst) was calculated by dividing the Dstby the Ht. To calculate gene diversity in the tetraploid M. queue, any locus where fixedheterozygosity was found in any population was deemed to be duplicated for allpopulations. Therefore 15 loci were analyzed in M. queue while only 10 were analyzedin M. guttatus and M. nasutus.For the most part each sample represents a separate field collected individual ora single representative from its offspring. Sib data were used infrequently. Inheritancepatterns for the allozymes studied had been analyzed previously by Ritland andGanders (1987) and by McClure (1973). Analysis of progeny arrays contributed furtherto the assignation of alleles to the proper loci.The number of individuals surveyed are not constant over all loci as someenzyme systems (i.e., PGI, DIA and TPI) stained more dependably than others (i.e.146PGD and AAT). A limited number of individuals were available for the M. guttatuspopulations 90-5, 90-4 and 90-3 so they were grouped together for the analysis of genediversity as they were located close to one another and had similar enzyme profiles.Two other enzyme systems (PGM and MDH) were found to be informative. They werenot used in the calculation of the genetic diversity parameters because for MDH thebanding patterns were not completely understood and for PGM staining wasinconsistent and too few individuals were scored.Table 2.1 - Collection sites in B. C. and Washington StateColl. # Location Habitat Taxa Collected90-07 B.C., Vancouver Isl. (V.I.), Nanoose Hill -seepy, open slope Mimulus queue, M. nasutus, M. guttatus90-08 B.C., Gabriola Isl., Drumbeg Park -seepy, beach area M. nasutus90-09 B.C., V.I., 5km N. of the Cowichan Laketurnoff on Highway 1-roadcut M. queue, M. guttatus90-10 B.C., V.I., along Finlayson Arm Road,near Goldstream Provincial Park-rocky outcropping M. queue90-11 B.C., Saltspring Island, end of IsabelleRoad-runoff path of culvert,open woodlandsM. queue, M. guttatus90-12 B.C., V.I., Saanich, Observatory Hill -south facing slope M. queue90-13 Washington, Skagit Co., Anacortes,Washington Park-rocks along oceanheadlandsM. guttatus90-14 B.C., Horseshoe Bay, cliffs above -west facing slope M. guttatusHighway 99Table 2.2 - Collection sites in California and OregonColl. # Location Habitat Taxa Collected91-02 Calif., Tuolumne Co., Yosemite Junction -bank of road cut M. guttatus91-03 Calif., Calaveras Co., Copperopolis, west of store -vacant lot M. guttatus91-04 Calif., Calaveras Co., Copperopolis, behind store -near ephemeral stream M. guttatus91-05 Calif., Calaveras Co., Copperopolis area,intersection of Little John Dr. and Coppercove Dr.-streambank M. guttatus, M. nasutus91-06 Calif., Tuolumne Co., Peoria Flats -ditch near culvert M. guttatus91-07 Calif., Tuolumne Co., Yosemite Jnct. -ephemeral stream bed M. guttatus91-08 Calif., Tuolumne Co., Red Hills Recreation Area -in streambed and water courses M. guttatus91-09 Calif., Tuolumne Co., Wildcat Creek -culverts where the streamcrosses the highwayM. guttatus, M. nasutus91-13 Ore., Josephine Co., Highway 5, north of Grants -seepy roadcut M.queuePass, Smith Hill Summit (1730 ft.)91-14 Ore., Douglas Co., Jnct. of Tyee Road and -west facing roadcut M. platycalyxHighway 13891-15 Ore., Douglas Co., Umpqua River Valley near -south facing waterfall M. platycalyxKellogg Springs turnoff91-17 Ore., Douglas Co., Umpqua River Valley,Highway 138 near Ed Jon Lane-south facing waterfall M. queue17Figure 2.1 - Typical M. guttatus habitat at collection site 90-07, Nanoose Hill, B. C.18Figure 2.2 - Map of Western North America. The shading indicates the range of the M.guttatus species complex and the * symbol indicates the locations of theOregon and California populations used in this study. The box indicatesthe portion of the map depicted in Figure 2.3 showing the B. C. sites.70KmHorseshoe BayNanoose Bay•Anacorte ri.oV•19Figure 2.3 - Inset from Figure 2.2 of lower Vancouver Island, some of the surroundingGulf Islands and part of the mainland showing the locations of the B. C.populations of the M. guttatus complex used in this study.20Figure 2.4 - Schematic representation of a Mimulus flower section showing some ofthe measurements taken. A - Corolla length, B - Corolla tube length, C -Corolla width, D - Corolla lip width, E - Corolla lip length, F - Long stamenlength, G - Short stamen length.21Figure 2.5 - A schematic diagram of a longitudinal calyx cross section of Mimulusshowing the calyx and gynoecium measurements taken: A - Pistil length,B - Calyx length at the upper lobe*, C - Ovary length, D - Calyx length atthe trough, E - Calyx length at the lateral lobe. *The length of the calyx atthe upper lobe is used to designate calyx length.22Figure 2.6 - A corolla of M. queue showing the measurement for length differences ofcorolla lobes. Scale markings are 1 mm apart.*Upper limit of range95% Confidence Upper —interval values^Lower —Figure 2.7 - Notched boxplots are used to compare the medians of flower measurements for the three different species. Thepoint at which the box returns to its full width indicates the 95 percent upper and lower confidence interval limits.If the confidence intervals of two species do not overlap there is a 95 percent certainty that the medians of thetwo populations are different.Lower limit of rangeMedian of upperhalfryMedian of lower^ whalf* The upper and lower limits of the range indicate the range of numbers which fall within 1.5 Hspreads of the medians of theupper and lower halves (known as hinges). Hspread is the absolute value of the difference between the medians of the upperand lower halfs. Outliers (not shown) indicated by * are outside 1.5 Hspreads of the hinges and those indicated by • areoutside 3 Hspreads of the hinges.Electrode Buffer^Gel Buffer^Enzymes^Run ConditionsStainedA1 Boric acid (.2 M),LiOH (.04M),H2O, pH 8.3B2 Boric acid (.263M),LiOH (.039M)H2O, pH 7.8C3 L-Histidine free base(.065M)Citric acid monohydrate(.019M)H2O, pH 5.7DL-Histidine free base(.065M)Citric acid monohydrate(.009M)H2O, pH 6.4ACitric acidanhydrous(.03M),Tris-HCL (.2M),H2O, pH 8.3BTris (.042M),Citric acid anhydrous (.007),Boric acid (.025),LiOH (.004),H2O, pH 7.8CL-Histidine free base (.065M)citric acid monohydrate(.006M)H2O, pH 5.7DL-Histidine free base (.065M)citric acid monohydrate(.003M)H2O, pH 6.4AAT^10 hours -50 ampsPGITPIAAT 9.5 hours - 50 ampsDIAPGITPIIDH 6 hours - 40 ampsMDH6PGDPGMIDH 7.5 hours - 40 ampsMDH6PGDPGMTable 2.3 - Buffer systems used in enzyme electrophoresis1 Ritland, pers. comm.2 Soltis et al, 19833 Wendel and Weeden, 198925Table 2.4 - Description of enzymes and staining proceduresStain Symbol Reference # of LociScoredQuaternaryProteinStructureAspartate aminotransferase AAT 1 3 dimericDiaphorase DIA 1 Method I 1 monomericIsocitrate dehydrogenase IDH 2pH 8.5 1 dimericMalate dehydrogenase MDH 2 ? dimericPhosphoglucoisomerase PGI 2 1 dimericPhosphoglucomutase PGM 2 Agarose 3 monomeric6-Phosphogluconatedehydrogenase6PGD 2 2 dimericTriose-phosphateisomeraseTPI 2 Agarose 2 dimeric1 Wendel and Weeden, 19792 Soltis et al, 198326Chapter 3 - Results3.1 Crossing StudiesOver 400 crosses and controlled emasculations were performed using all ofthe B. C. populations and most of the California/Oregon populations listed in Tables2.1 and 2.2. The bulk of the data was collected for crosses using the plants from B.C. populations. Crossability measured in terms of percentage of successful fruit setwas high for crosses within a single taxon and for crosses between M. guttatus andM. nasutus (Figure 3.1). Although the number of crosses attempted between B. C.and California/Oregon material was not large, the number of successful crosseswas enough to indicate definite crossability within species and between M. guttatusand M. nasutus from the two geographically widely separated groups of populations(Tables 3.1 and 3.2). In addition, all eight crosses attempted between M. guttatusand M. platycalyx were successful as were two out of the three crosses attemptedbetween M. nasutus and M. platycalyx. Pollen viability ranged from a low of 60% toa high of 97% with an average of 81% in Fl hybrids from crosses between M.guttatus and M. nasutus or between different populations of M. guttatus. Both thehigh and low values were from crosses between M. guttatus populations.Natural hybridization between M. guttatus and M. nasutus is most certainlyoccurring in population 91-5 as indicated by the intermediacy of some of the plantscollected. Putative natural hybrids between M. guttatus and M. nasutus were alsofound in population 90-7. The hybrids were intermediate in flower size and werewilting during the hot part of the day whereas the surrounding M. guttatus plantswere not.As indicated in Table 3.2, crossability of both the B. C. and Oregonpopulations of M. queue with M. guttatus and M. nasutus is very low. There weretwo categories of unsuccessful crosses. These were 1) crosses that produced noseed and 2) crosses that produced shriveled seed. It is unclear what proportion, ifany, of the crosses generating no seed are actually products of crossing barriers27and how many are result from technique problems such as mechanical damage tothe pistil during the emasculation/pollination process or use of immature pollen.The percentage of crosses resulting in no seed production was probably greater inboth of the small flowered taxa because their flower size made emasculation atricky procedure (Table 3.1). They also produce significantly less pollen whichappeared to be at the appropriate stage (i.e., mature but not yet dehisced) for ashorter period of time than that of M. guttatus. It may therefore be inappropriate toassociate this type of unsuccessful cross with crossing barriers. However, lack ofseed production was one of the crossing barriers in Mimu/us identified by Vickery(1974B). It was therefore decided to include this category in the totals from whichthe percentage of successful crosses was calculated because there was insufficientevidence to exclude them as crossing barriers.Conversely, the production of shriveled seed was a repeatable result of somecrosses and can be associated more confidently with a genetically controlledcrossing barrier. One must be careful when making this association becauseenvironmental stress during seed formation can also result in shriveled seeds.Environmental stress rather than breeding barriers may be the factor that causedshriveled seed production in five M. guttatus X M. guttatus crosses. Periodically M.guttatus X M. guttatus crosses would yield mixtures of shriveled and normal seed,these crosses were scored as successful because two of these crosses wererepeated and resulted in normal seed. No pattern was observed for thisphenomenon.The shriveled seeds produced in the intertaxon crosses with M. queue arelikely to represent a barrier to crossing as they were harvested from healthy plantsand occurred repeatedly. The degree of shriveling varied from severe to slight. Theslightly shriveled seed were pear shaped rather than oval and sometimes could beencouraged to germinate if planted on agar. A few of these plants have survived toflowering. They tended to resemble M. queue more than M. guttatus and28observations indicated that they were self pollinating. However no seed was everformed. Only 8% of all the intertaxon crosses with M. queue produced seed whichcould be germinated. The amount of shriveling was typically equal for all seedsfrom the same cross. Not enough crosses were attempted to determine if crossesusing the same parents would yield equivalent amounts of shriveling. Most of theless severely shriveled seed was the result of crosses between M. queue and M.nasutus. Occasionally one or two normal seeds were found among the shriveledseed when M. queue was crossed with the diploid species. To date the ones thatproduced mature plants exhibited no morphological hybrid traits and wereconsidered to be the result of accidental contamination or self fertilizations.Pollen viability data for the parental generation, progeny of the within speciescrosses, and hybrids of the M. guttatus and M. nasutus crosses is presented inFigure 3.2. More then 25 individuals were used for each group except for the M.queue X M. queue crosses and the M. nasutus parental generation where 6 and 11individuals were used respectively. In all cases average pollen viability was greaterthan 70 percent.Figure 3.3 presents pollen viability of the Fls of crosses involving either M.queue or plants from the California or Oregon population. While pollen viability isaround 75% for both intraspecific and interspecific crosses not including M. queue,it is very low for the interspecific crosses with M. queue. Very few of these lattercrosses have been achieved and these data are from eight different plantsrepresenting seven different crosses. Most of the pollen grains scored as inviablewere highly shrunken. They did, however, become stained. The pollen grainsscored as viable were similar to those of parental plants except that in 4 out of the 6plants that showed greater than 1% pollen viability, normal shaped pollen grainsaveraged approximately 15% larger in diameter than in M. queue. On a limitedsample size of 3 plants for each group, it was found that the crosses between M.queue and M. guttatus had pollen grains ranging in size from 35 pm to 58 p.m while29M. queue plants had pollen grains ranging from 35 um to 46 pm in diameter. Pollengrain diameters for M. guttatus plants ranged from 29 um to a maximum of 38 um.30Figure 3.1 - Percentage of successful crosses between M. guttatus, M. nasutus andM. queue from B.C. populations. The direction of the arrow indicates thedirection travelled by the pollen. Numbers above or below the ellipsesindicate the percentage of successful crosses within the species.31Table 3.1 - Results of experimental hybridizations within the same speciesSpecies Group NormalSeedShriveled^NoSeed^SeedTotal PercentSuccessfulM. guttatus B.C.XB.C. 143 5 3 151 95(g) B.C.XCA. 18 18 100CA.XCA. 4 1 5 80Totals 165 5 4 174 95M. nasutus B.C.XB.C. 9 1 10 90(n) B.C.XCA. 1 3 4 25CA.XCA. 1 1 0Totals 10 0 5 15 67M. queue B.C.XB.C. 24 5 29 83(q) B.C.XOre. 4 4 8 50Ore.XOre.Totals 28 0 9 37 76Table 3.2 - Results of between species crossing experimentsSpecies Group Normal Shriveled No Total PercentSeed Seed Seed SuccesfulgXn B.C.XB.C. 58 8 66 88CA.XCA. 2 2 100B.C.XCA. 7 1 8 87Totals 67 0 9 76 88gXq B.C.XB.C. 2 73 20 95 2CA.XCA. 3 3 3 9 33B.C.XCA. 2 8 2 12 18Totals 7 80 25 116 6nXq B.C.XB.C. 2 30 9 41 5CA.XCA. 2 2 0B.C.XCA. 1 1 5 7 17Totals 3 28 14 50 6100080604020IIIIIII1, 1T91.1 T86.991.5T T71.8^71.132Figure 3.2 - Pollen viability of the parental generation from B. C. populations of M.guttatus (G) and M. nasutus (N) and of the Fl genertion of crossesbetween them.G^GXG^GXN^NXG^NCategory33Figure 3.3 - Pollen viability of the Fl s from crosses where plants from Californiapopulations are used. Catagory 1 is Fl hybrids between M. guttatusand M. queue Catagory 2 consists of Fl hybrids from crosses withinthe same species but between different populations. Over 80% of thesecrosses are between California and B.C. material. Catagory 3 is madeup of interspecific crosses between M. guttatus, M. nasutus and M.platycalyx.343.2 Morphological AnalysisReliable morphologicial differentiation among the three species was notalways possible prior to flowering. Once flowering had begun, M. guttatus waseasily distinguished from both M. queue and M. nasutus by its larger flower size andby the fact that the pistil of M. guttatus extends much farther than the upper calyxlobe. However, differentiation between the two small flowered species, M. queueand M. nasutus, requires closer analysis. When they are grown under favorableconditions, such as the growth chamber, their distinctions become more obvious(Figure 3.4A). The following three general categories of morphological charactershave been found to be useful in separating the B. C. populations of M. queue andM. nasutus: 1) relative and absolute sizes of floral parts, 2) overall flower shape and3) plant height and leaf shape under favorable growth conditions. These arepresented in more detail later in this section.Growth chamber problems made conditions less than optimum when theCalifornia and Oregon populations were grown and it was not possible to determinewhether plant height and leaf shape are also distinct in the southern populations.Problems in morphological differentiation between the two small floweredspecies, M. nasutus and M. queue, arose when they were collected from dry oradverse habitats where plants did not reach their maximum potential size (Figure3.4B). A series of morphological measurements were therefore undertaken toidentify whether any simple floral or calyx measurements could be found whichcould quickly and accurately place a small flowered sample into the proper species.Of the 19 floral measurements scored, the following nine displayed the samebasic pattern as the corolla length measurements summarized by boxplot A, Figure3.5. The lengths of: the pistil, both sets of stamens, the corolla tube, corolla lip andpedicel, the widths of the corolla and corolla lip and the distance between thestigma and the upper stamen were all significantly smaller for the two smallflowered taxa than they were for M. guttatus. In addition there were no significant35differences in these measurements, (as indicated by overlapping boxplot notches)between the two small flowered taxa, M. queue and M. nasutus.Calyx measurements (Figure 3.6) for M. queue were significantly smallerthan for M. nasutus and M. guttatus. Two of the calculated distances, thedifferences in length between the calyx and pistil and between the pistil and thecorolla tube, showed significant differences in all groups (Figure 3.7). However themedians for the two small flowered taxa were still nearer to one another than theywere to M. guttatus.Figures 3.8, 3.9 and 3.10 show how plants from the Oregon populations ofM. queue compare to those from B. C. and how M. nasutus plants from theCalifornia populations compare to M. nasutus plants from B. C. in eleven of thefloral and calyx measurements. For the most part, M. queue plants from Oregonshowed the same trends in flower and calyx sizes as plants from B. C. The onlysignificant difference was in the difference between the calyx length and pistil length(Figure 3.8B). Similarily, there were no significant differences in the medians of M.nasutus from B. C. and California populations.Selfing in the two small flowered species has been promoted in part becausethe anther - stigma separation has been reduced (Figure 3.11A). In order toascertain whether this is entirely a result of flower reduction or if there has alsobeen nonproportional reduction in stamen and pistil length their values relative toeach other were calculated by subtracting the logs of pistil length and stamen lengthfrom the logs of corolla length (Figure 3.11B and C). Mimulus queue and M.nasutus are compared to M. guttatus, which is the most likely large floweredancestor. It appears that in M. queue both the stamen and the pistil haveundergone greater reduction than the corolla. Stamen length in M. nasutus hasbeen reduced to a greater degree than corolla length. It is borderline whether pistillength has been reduced to a greater extent than corolla length in M. nasutus.Figure 3.11 also indicates that when M. nasutus and M. queue are compared,36stamen, pistil and corolla reduction has been proportional. To answer the questionas to whether the anthers and pistil in M. nasutus and M. queue were reducedequally, the logs of their lengths were subtracted. Again using M. guttatus as thereference, pistil and stamen lengths have not been reduced proportionally. It is notpossible with these data to determine whether the pistil has been reduced more orthe stamens have been reduced less, or some of both.One of the criteria that has been used to distinguish M. nasutus from othermembers of the M. guttatus complex is the long upper calyx lobe as compared tothe other calyx lobes (Grant 1924). Figure 3.12A indicates that in absolute termsthis is not a valid assumption. The same is true when size is factored out of thecomparisons using subtraction of logs (Figure 3.12B)It is evident from Figures 3.5-3.7 that although some of the variables for thetwo small flowered species were significantly different, there was still considerableoverlap in their ranges and no single flower or calyx measurement can reliablydistinguish the two. Multivariate statistics provides a means by which these smalland overlapping differences in variables can be combined and summarized bylinear equations. Individuals can then be plotted in two dimensional space andgroupings visualized.Principal Components Analysis (PCA) was used to explain in as fewdimensions as possible the variance among individuals. Although PCA is mostfrequently used where there are no a priori assumptions of relationships within thedata, it is used in this study to confirm that the three species are naturally grouping.Plots of the first two principal components axis from the PCA of floralcharacters tended to group members of each taxon together although notcompletely (Figure 3.13). The greatest overlap was between M. nasutus and M.queue and this overlap was decreased somewhat when M. guttatus individualswere removed from the analysis (Figure 3.14). All the M. nasutus individuals thatfailed to clump with the rest of the M. nasutus grouping, in both of the analyses,37were ones where measurements had been taken from very early flowers (1-4),which were extremely small and verging on being cleistogamous.Component loadings for the PCA analysis are provided in Tables 3.3 and3.4. When all three taxa were included in the analysis, all measurements figuredimportantly in the loading of the first axis with pistil length and corolla length havingthe highest component loadings (Table 3.3). This component separated M. guttatusfrom the two small flowered taxa. The second component separated out M. queuefrom M. nasutus and M. guttatus using calyx length and the difference between thepistil length and the corolla tube length as the most important criteria.When PCA was performed on only the two small flowered species, M.nasutus and M. queue, once again corolla length and corolla tube length were veryimportant in separating them. As indicated in Table 3.4, the difference in thelengths of the calyx and pistil was relatively unimportant in the first componentloading but was the most important variable in the loading of component two.Figure 3.15 demonstrates how M. nasutus and M. queue differ in the lattercharacter. Note also in this photo that although their stigmas are full of pollen theyare not completely closed. The touch sensitive stigmatic response has beenreduced in M. nasutus and M. queue as compared to M. guttatus. Pedicel lengthhad consistently low loading values.In discriminant analysis, a priori assumptions such as the groupings ofindividuals allow linear discriminant functions, which maximize the differencesbetween groups, to be calculated. When assumptions of multivariate normality arenot met, as was the case in this study, discriminant analysis is best viewd asproviding an approximation to the total sample information (Johnson and Wichern1982).For the most part the discriminant analysis mirrored the PCA but as indicatedin Figure 3.16 species tended to cluster together better. Again the first factortended to discriminate between the large flowered species (M. guttatus) and the38small flowered species (M. queue and M. nasutus), therefore corolla measurementsand the difference in lengths of the calyx and pistil were most important (Table 3.5).The second factor tended to discriminate between the two small flowered speciesand calyx measurements became important as well as ovary and stipe lengths. Asin the PCA, the greatest amount of overlap between groups was with M. nasutusindividuals where measurements were taken on the early, highly reduced flowers.These individuals tended to be located near the margins of M. queue clusters.There are a few other floral characters that are likely to be meaningful indifferentiating M. queue from M. nasutus. Preliminary observations (Figure 3.17)indicate that on a dorsiventrally flattened flower, the difference between the lengthof the upper corolla lobes and the bottom corolla lobe is greater for M. nasutus thanfor M. queue. However, like the other characters studied there is considerableoverlap in this character partially because of flower size variability. Thismeasurement was made for only a third of the plants studied so was not included inthe PCA or discriminant analysis. In the B. C. populations particularly, there alsoappears to be a distinct (but difficult to measure), difference in the overall shape ofthe floral tube (Figure 3.18A). M. nasutus tends to have a somewhat wider corollabase that flares out more gradually than in M. queue. Mimulus queue populationsfrom Oregon tended to be somewhat intermediate in this character (Figure 3.18B).Growth chamber observations of material collected from B. C. suggested thatthe maximum height of the plants and the size of the largest leaves were greater forM. nasutus than for either M. queue or M. guttatus (Figure 3.4A). In the field, similarlarge M. nasutus plants, as well as some very small ones were found in population90-8. However on the relatively dry slopes of Nanoose Hill where population 90-7was collected, all plants of M. nasutus, M. guttatus and M. queue were quite smalland about the same size (approximately 10-15 cm tall) (Figure 3.4B).In addition to flower morphology and plant size, the small flowered speciesfrom B. C. could be separated on the basis of anthocyanin deposition. Mimulus39queue is fixed for anthocyanic spotting of the calyx which was not found to bepresent in the two studied populations of M. nasutus B. C. This is not a reliablecharacter for seperating the two species over their entire range because the two M.nasutus populations from California had calyx spots. Both populations ofM.platycalyx and some of the M. guttatus individuals also exhibited calyx spotting.Both B. C. M. nasutus populations were characterized by a large, heartshaped splotch on the lower corolla lip. This splotch is prominent in the drawing ofM. nasutus in Pennell's (1951) treatment of the species. Similar corolla splotcheswere seen occasionally on M. guttatus plants from population 90-07 and on M.nasutus from California population 91-09. Crosses between M. nasutus and M.guttatus also exhibited this splotch suggesting it is controlled by one or fewdominant genes.No obvious patterns were observed for the deposition of anthocyanin in otherparts of the plant apart from being very common and distinct at the base of the leafblade in all populations of M. nasutus. Similar though less intense markings werealso a feature of M. queue plants. Conversely, leaf spotting was also found in bothM. queue and M. nasutus, however it was most intense in M. queue. Somepopulations of M. guttatus were characterized by a large amount of leaf anthocyanin(population 91-06) while others had relatively little. This character variedconsiderably within M. guttatus populations. A single plant (from population 90-14)was found which lacked anthocyanin. This plant appeared to be particularlysensitive to attack by powdery mildew.Further differences can be seen in the young individuals of M. nasutus andM. queue grown from seed collected in B. C. (Figure 3.19). Mimulus nasutus tendsto have leaves that are more bullate, less regularly toothed, less rounded and morepubescent than those of M. queue. Whether these differences hold in the Californiaand Oregon populations of M. nasutus and M. queue has yet to be determined.40Oregon populations of M. queue are somewhat more pubescent than their B. C.counterparts.Flower measurement data were also collected for the hybrids produced fromcrosses between M. guttatus and M. nasutus plants. Figure 3.20A shows examplesof the two parental species and Figure 3.20B shows a hybrid beside a M. guttatusplant. For all 10 measurements taken, the means of the hybrid measurements werebetween the averages for the parental taxa. This was true no matter which specieswas the female parent. The same pattern is found for the medians as demonstratedby the boxplots in Figure 3.20. Note that in Figure 3.21 B the hybrid is taller and hassmaller flowers than the M. guttatus plant. These data were only obtained fromplants grown from B. C. collections. All the preceding data were collected from adifferent set of plants that also included California populations, which is why themedians for certain measurements of M. guttatus and M. nasutus may besomewhat different.A description has been provided for M. queue in Appendix I followed by a key tothe species in the M. guttatus species complex in Appendix II.A.41Figure 3.4 - Morphological and growth habit differences in M. guttatus (G), M. nasutus(N) and M. queue (Q) when A. grown in the growth chamber and B.collected from Nanoose Hill (population 91-07) on a dry year.N1 I^aB.^*A.5 -4 -3 -2 .1 .0 0.1I^I^■ G^N^QTaxonG^N^QTaxon*I*1G^NTaxonG^N^QTaxonC.0.30.0D.0.20.1***I054321042Figure 3.5 - Boxplot comparisons of flower measurements from M. guttatus (G), M.nasutus (N) and M. queue (Q).*.*•1 II43Figure 3.6 - Boxplot comparisons of calyx measurements from M. guttatus (M), M.nasutus (N), and M. queue (Q).A.^ B.G^N^QTaxonD.E 0.5c.)G^N^Q^mTaxonI^I G^N^0Taxon.2.0E 1.5c.)=MC 1.0a)_1x0.5co00.0C.1.50.0G^NTaxonI I^IMIImNM1.50.01 .00.5*0-1-2**I1 IE 1c.)B..csca)..."a)I-caID,_ Eo u0-Ea)ca)...1=Y.44Figure 3.7 - Boxplot comparisons of flower/calyx measurements from M. guttatus(G), M. nasutus, (N), and M. queue (Q).A.G^N^QTaxon1.0*0.50.0 1 .**-0.5 lo . •-1.0 I^I aG^N QTaxon45Figures 3.8 - Boxplot comparisons among B.C., California and Oregon populations ofM.nasutus and M. queue for flower and calyx measurements.1.41.3 IN1.2 IN1.1 IN1. MI0.60.5 F0.4 1^m^Ici^:-: Li^2):m^c.)o^co^oZ^i^6^6B.A.c.imiI^m^■.."-:ToLii6^clico 86 6t:*0.60.5D.E0=0.4 Th.Ccu_r0.3 cEas0.2 MIC7)0.1 .^I I^ILi^..zTom^c.)I^iLi^aiDi 0e,^6CategoryC.^.:._d^Li^dim^c.) COZ i 6 6Category46Figure 3.9 - Boxplot comparisons among B.C., California and Oregon populations ofM.nasutus and M. queue for flower and calyx measurements.Category Category1.0 =,.L..., 0.9 =Eo 0.8 ="F. 0.7:5_ 0.6 I=5.0.J0.50.4 NMct1 0.3(7)16 0.2C.) 0.1 NI0.0 t .^.6^Ta ci^2..;°?^0z i °?^9o a-E.-'032 •.c .C)Ca)FR—J •CosT)18C?0 I^. ...:6. 6m^U I"?Z^i 0.47Figure 3.10 - Boxplot comparisons among B. C., California and Oregon populations ofM.nasutus and M. queue for flower and calyx measurements.A. BCategory CategoryC.0.7 —E  0.6:F.^0.5C)cpa0.4_1p. 0.3-Jco 0.22O o.U .10.0N.=m= II=1 1 I6 7t1 ^U^?-;m^0 ai^0Z i 6 6Categorytr)0.7.J0.60-cr)0.50 0.2C- J• 0.1O0.08 -0.1ty)OG^N^QTaxon*G N^QTaxon0. -0.1Q.-al -0.2O- I -0.3*G N^QTaxon- JCCOcoa)O-J *MNCCOcornOCO-616C)****Taxon48Figure 3.11 - Two character comparisons using subtraction of absolute lengths andsubtaction of logs for M. guttatus (G), M. nasutus (N) and M. queueA.^(Q).^ 2 B.• 0.7a)• 0.6-J0.5• 0.44C2 E▪ 0.1C• 0.0ci)".=^-0.1b^-0.2C.^ D. •I I^I0.^N^Q49Figure 3.12 - Two character comparisons using subtraction of absolute lengths andlogs for M. guttatus (G), M. nasutus (N) and M. queue (Q).A.G^N^QTaxonTaxon321Nctc.)a0-1-2-3o Mimulus guttatuso M. queue• M. nasutus50Figure 3.13 - PCA on standardized flower measurements of M. guttatus, M. nasutus and M.queue. The first component accounts for 69% of the variation and the secondcomponent accounts for 12% of the variation.-3^-2^-1^0^1^2^3PCA 151Figure 3.14 - PCA on standardized flower measurements of M. queue and M. nasutus.The first component accounts for 69% of the variation and the secondcomponent accounts for 18% of the variation.432N 1<c)a.-1-2-3• M. nasutusj M. queue-3^-2^-1^0^1^2^3PCA 152Table 3.3 - Component loadings for the variables used in the PCA of M. guttatus, M.nasutus and M. queueVariable Component Loadings1^2Pistil length 0.942 -0.249Corolla length 0.942 -0.153Ovary length 0.905 0.177Corolla lip width 0.905 -0.254Calyx length at the upper trough 0.899 0.325Calyx length 0.831 0.482Calyx width at the lateral lobes 0.817 0.313Stipe length 0.784 0.092Pedicel length 0.672 -0.339Pistil length - corolla tube length 0.524 -0.658Table 3.4 - Component loadings for the variables used in the PCA of M. nasutus andM. queueVariable Component Loadings1^2Corolla length 0.930 -0.238Corolla tube length 0.897 -0.016Corolla lip length 0.863 -0.366Calyx length at the upper trough 0.857 0.386Ovary length 0.847 0.194Corolla lip width 0.768 -0.532Stipe length 0.646 0.359Calyx length - pistil length 0.210 0.882Pedicel length 0.333 0.14453Figure 3.15 - Calyces of chasmogamous flowers after the removal of the corolla. A. M.nasutus (8x) B. M. queue (8x).54Figure 3.16 - Scatterplot of factor scores obtained from discriminant analysis of flowermeasurements from M. guttatus, M. nasutus and M. queue.Mimulus guttatusM. queueM. nasutus-4^-2^0^2^4^6Factor 1Table 3.5 - Canonical Loadings of the variables used in the discriminant analysis of M.guttatus, M. nasutus and M. queue.Variable Canonical LoadingsFactor 1^Factor 2Corolla length 0.631 0.462Corolla lip length 0.615 0.350Corolla lip width 0.573 0.162Ovary length 0.284 0.637Calyx length 0.179 0.815Calyx width 0.215 0.584Pedicel length 0.293 0.235Stipe length 0.257 0.522Calyx length - pistil length -0.830 0.331Pistil length - corolla tube length 0.314 -0.073I55Figure 3.17 - The distance by which the lower corolla lobes excedes the upper corollalobes in length in M. guttatus (G), M. nasutus (N) and M. queue (Q). A)Bargraph of the means with standard errors. B) Notched boxplots.A.0.m0.1 + MN1 1G N^QCategoryB.a)la^0.60-I-,E 0.5co 013^0.4...o IS'0 c 0.3L. E)a) w'.4.-.^0.2o —_In^0.1at 05 CT)^0.0CL. a)a) -1 -0.10.o.m^-0.2G N^QCategory56Figure 3.18 - Comparisons of corolla shapes among A. M. queue (0) and M. nasutus(N) from B. C. (12X) and B. M. queue from B. C. and Oregon and M.nasutus (3.3X).B.57Figure 3.19 - Young plants at the 3-4 leaf stage A. M. nasutus (1x) and B. M. queuex).B.58Figure 3.20 - A. Examples of M. guttatus (G) and M. nasutus (N) (0.4x). B. A hybrid(GXN) next to a M. guttatus plant (0.64x).G^ NGXN^ G3_021INMII^I^ I2.0 -*•I*0.5t*I59Figure 3.21 - Boxplot comparisons of flower measurements of M. guttatus (G), M.nastutus (N) and their hybrids.A.^ B.G GXG GXN N NXG^G GXG GXN N NXGCategory^ CategoryC.^ D.540:IS7:3363 2Z8U 10G GXG GXN N NXG^G GXG GXN N GXNCategory^ Category603.3 Field Observations and PhenologyAll three species acted as annuals or winter annuals in the populationsstudied. Observations from plants grown in growth chambers indicate that M.queue is probably an obligate annual as the plants were replaced in the pots byhundreds of seedlings after about four months growth. It was not determinedwhether the M. guttatus and M. nasutus populations used in this study were obligateannuals. All previous reports of M. nasutus indicate that it is an annual (Kiang andHamrick 1978, Dole 1992, Vickery 1964). Most annual populations of M. guttatuspreviously studied have been deemed to be facultatively perennial (Macnair andCumbes 1989, Kiang and Hamrick 1978, Dole 1991) except for an obligate annualpopulation of M. guttatus identified by Vickery (1959). Previous observations of M.guttatus plants from some of the B. C. populations indicated that they were capableof indefinite growth (Ganders, pers. comm.). Individuals of one of the coastalpopulations of M. guttatus, 90-13, had a greater tendency to produce stolons thanthe other M. guttatus populations. The largest M. guttatus flowers observed in thisstudy were produced by plants from this population as well as from the Nanoose Hillpopulation (90-07). Population 90-13 also tended to be more pubescent thanaverage. Isozyme patterns, however, were not distinct for this population.Field observations of B. C. populations suggest that M. queue is able tocomplete its life cycle earlier than M. guttatus. By May 18/19, 1990, M. queueplants in populations 90-9 and 90-11 had finished flowering and fruit was ready tobe harvested while M. guttatus plants had just begun to flower or were in mid-flowerand no fruit was mature. M. guttatus and M. nasutus plants had mature seed whenvisited on June 20. Observations in 1991 on May 1 indicated that M. queue and M.guttatus were flowering at the same time at Sooke Potholes Park on VancouverIsland (Figure 3.22), and all three taxa were flowering together at Nanoose Hill (90-07). They could sometimes be found growing right beside one another however itdid appear that M. nasutus was not found in the very driest areas. The two61populations 90-12 and 90-10, where M. queue was the only taxon collected, werevery dry south facing hillsides which had already dried up by May 18, 1990. Theremaining B. C. sites were either west or east facing or more complex south facingslopes with moist microhabitats (90-07 and 90-11).The California populations were visited between April 9 and 11, 1991. At thistime they were either not flowering or just beginning to flower. One of the M.nasutus/M. guttatus populations (91-05) had not yet begun to flower while atanother, (91-09) some of the M. nasutus had begun to flower while none of the M.guttatus were yet in flower. Another nearby population of M. guttatus (91-08) was infull flower.The Oregon collections were made on April 13, 1991. Both M. queuepopulations had begun to flower as had the two populations of M. platycalyx. Thefruit was still immature at this time. Several sites of the perennial M. guttatus werevisited along the Oregon coast and none of these had started to flower.Germination records were not started until the fifth day after planting bywhich time many of the M. guttatus seeds had already germinated and thus theboxplot for M. guttatus is a line (Figure 3.23A). Some of these appear to havegerminated by day four. Both boxplots and bargraphs are provided so that meansand medians can be compared (Figures 3.23A and 3.23B). The slow germinationrate of M. queue, as compared to M. guttatus and M. nasutus, carries oversomewhat into the measurement of days to first flower. Records of blooming timesfor B. C. populations of the three taxa when grown in the growth chamber indicatethat although the earliest blooming dates were the same for all three taxa, the meanand median initial blooming dates were somewhat earlier for M. nasutus than for M.queue and M. guttatus (Figures 3.24A and 3.24B). Although the same trends arevisible in the means and medians as indicated in the bargraph and boxplots theyappear more significant in the boxplots. This is a result of the nonnormality of thedata.62There also seemed to be substantial differences in germination percentage(#germinated after 29 days/#planted) between M. guttatus, M. nasutus and M.queue (Figure 3.25). Almost all seeds planted of M. guttatus germinated. Althoughthere was germination in most plantings of M. nasutus and M. queue , rarely did allthe seeds germinate. The situation was intermediate for the hybrids.As Kiang (1972) observed previously, the first 2-4 flowers produced by M.nasutus were very small and essentially cleistogamous. The first few flowersproduced by M. queue were also smaller, however they generally opened. Flowersize increased for both M. queue and M. nasutus until it peaked at about the 16-30th , flower depending on growing conditions. This was followed by a decline inflower size until the flowers were again cleistogamous. As soon as both M. queueand M. nasutus plants were environmentally stressed they also began to producecleistogamous flowers such as those shown in Figure 3.26A.The two small flowered taxa were also similar in that, by the time thechasmogamous flowers opened, self pollination had already occurred.Observations during emasculations suggest that pollination occurred about 2 to 3days prior to opening when the stigma was curled down directly on top of theanthers. Figure 3.26B shows the positions of floral parts in a M. queue flower budjust prior to anthesis. This is the stage when emasculations were performed forexperimental crosses. Occasionally flowers could be found in both taxa where theupper anthers or the stigma had been extruded out of the flower prior to its opening.When M. guttatus is severely stressed, flowers that are only slightly largerthan the largest M. queue flowers are produced. The pistils of these stunted M.guttatus flowers did not appear to be reduced in proportion to the other flower parts.Typically each subsequent flower produced by M. guttatus is progressively largeruntil a peak flower size is attained. Unlike M. queue and M. guttatus, however,flower size does not begin to decrease at this point but stays relatively constant asthe plant ages.63M. guttatus is probably pollinated by bumble bees (Bombus) in the field.Pollinator activity was observed in only one of the populations, 90-14. Variouspollinators were active in the other populations, however they were ignoring M. guttatusfor other species such as Plectritus congesta. (Valerianaceae). Contrary toobservations by Grant (1924), who concluded that Mimulus guttatus lacked nectarglands, a small amount of nectar was observed at the base of the corolla in all threetaxa. Nectar was very reduced and often absent in the two small flowered species.64Figure 3.22 - Sympatric populations of Mimulus queue on the left and M. guttatus onthe right at Sooke Potholes Park, Vancouver Island, B. C.30G N Q 0XCIaI^I**•o 25ilCE 2016(..5o 15co>,aco 105 Z aX x(.5^z••A..65Figure 3.23 - A. Boxplot and B. Bargraph analysis of the days required for germinationof M. guttatus (G), M. nasutus (N), M. queue (Q) and the Fl progeny ofcrosses between M. guttatus and M. nasutus.A.Category14B. 12^471^—I-10.2 T^10.2 "1089.46 7.1" 6.2 II420G^N^0^GXG^GXN^NXGCategoryA.B.50043.311II 42.4111111IIII40302010....mis,40.1.......42.9^42.6 42.466Figure 3.24 - A. Boxplot and B. Bargraph analysis of days to first flower for M. guttatus(G), M. nasutus (N), M. queue (Q) and the Fl progeny of crosses betweenM. guttatus and M. nasutus.7060INI**** *5040*MI30 IIIIG^N^Q^0^ZX^X^Xa^0^ZCategoryG^N^Q^GXG^GXN^NXGCategory67Figure 3.25 - Boxplot of germination percentages of M. guttatus (G), M. nasutus (N),M. queue (Q) and the Fl progeny of crosses between M. guttatus and M.nasutus..150wEncoE'wo 100iii0.C0IPaS 50CEi)0G N 0 (5 Z 0X X X(.5 (.5 ZCategoryFigure 3.26 - A. A chasmogamous flower surrounded by two cleistogamous flowers ofM. queue (1 scale unit = 1 mm) B. A dissection of a M. queue bud justbefore anthesis (1 scale unit = 1mm).693.4 Chromosome numberCounts for three M. guttatus plants from B. C. populations at meioticprophase I (diakinesis) indicate that they are diploid (n=14) (Figure 3.27A). This isthe typical chromosome number of the species. No definitive count has been madefor M. queue. However partial counts for M. queue from both B. C. and Oregonpopulations at meiotic prophase II indicate that they have more than n=14chromosomes and are probably tetraploids (Figure 3.27B).Indirect evidence for polyploidy is found in the enzyme electrophoresisanalysis. The B. C. populations of M. queue have fixed heterozygosity in one locuseach, controlling the enzymes PGI and AAT and at two loci, controlling PGMenzymes. Oregon populations of M. queue have fixed heterozygosity in loci of PGI,6PGD and PGM enzymes.Chromosome counts were not obtained for M. nasutus. However, lack of fixedheterozygosity in the isozyme data indicates that it is probably not polyploid. Previouscounts of n=13 were obtained by Mukherjee and Vickery (1960) for M. nasutus plantsfrom a population collected along the same creek as population 91-09. Counts of n=14have been made for other California M. nasutus populations (Mukherjee and Vickery1960, Vickery 1964)70Figure 3.27 - Squashes of meiotic cells of: A. M. guttatus in prophase I (with 14chromosomes) and B. M. queue in prophase II (polyploid cell, withpossibly 28 chromosomes).713.5 Enzyme ElectrophoresisTen loci from six enzyme systems were found which could be scoredsatisfactorily. Tables 3.6 - 3.14 summarize the frequency in which alleles were found ineach population and the proportion of the total represented by each allele for eachlocus. All ten enzyme loci examined were found to be polymorphic although the fastlocus of TPI was monomorphic except for a single individual from a Californiapopulation of M. guttatus, 91-02. All other rare alleles were found in at least threeindividuals.PGIThe PGI-2 locus had the greatest number of different alleles (5). All five allelesare represented in lanes 1 - 6 in Figure 3.28. Allele 2c is very close to 2b and 2dmaking it quite possible that some heterodimers may have been scored ashomozygotes. This is a conservative estimate of the number of alleles as there ispossibly an additional allele between 2a and 2b as indicated by the heterodimer in lane5. Not surprisingly from the number of alleles identified, PGI also had consistantly thegreatest gene diversity (Tables 3.8-3.10).Almost all individuals from B. C. populations of M. queue were fixed for alleles2a and 2b (Table 3.11). This is assumed to reflect gene duplication of PGI locus 2 as aresult of tetraploidy. One locus is fixed for the a allele and the other for the b allele sothey are completely homozygous while maintaining functional heterozygosity. Twoindividuals from population 90-07 had only the 2b allele.The Oregon populations of M. queue also exhibited a considerable amount offixed heterozygosity. One individual from M. queue population 91-13 had only allele 2a.Except for this individual, which also differed at TPI-2, AAT-3 and 6PGD-2, the enzymeprofiles for all the 91-13 plants screened would have been identical. It is possible thatthis may be a contaminant as the plant was sampled and destroyed at the cotyledonstage before it could be determined whether it was morphologically similar to the rest ofpopulation 91-13. Enzyme data for this plant, therefore, were not included in the72summaries. All other individuals from population 91-13 were fixed for the 2b and 2dalleles of the duplicated PGI-2 locus.When the enzyme data from all loci were summarized (Table 3.15) it wasapparant that there were four distinct genotypes of M. queue represented in this studyby: 1) all B. C. populations 2) Oregon population 91-13, 3) Oregon 91-17A and 4)Oregon 91-17B. Genotype 91-17A had banding patterns identical to that of the B. C.populations except at the 6PGD-2 locus. The PGI-2 enzyme staining patterns found forM. queue in this study (except for the two plants which had only allele 2b) arepresented in lanes 13-24 Figure 3.28. Note the two different heterodimers in M. queuepopulation 91-17 in lanes 19 -23.Overall, allele 2b was by far the most common at the PGI-2 locus and was foundin every population surveyed. Ninety-seven percent of the M. nasutus plants surveyedfrom both the B. C. and the California populations were fixed for this locus (Figure 3.28,lanes 28 and 29) and generally this locus alone could be used to differentiate M. queuefrom M. nasutus . The slowest allele, 2e, was found in only one population, 91-08.A relatively monomorphic fast locus was also observed for PGI. It was moreapparent when system B (Figure 3.28, lanes 1-12 and 25-36) was used instead ofsystem A (Figure 3.28 lanes 13-24). At least two and possibly three closely migratingalleles were observed. Most B. C. plants of all three species appeared to behomozygous for the PGI-1 b allele with the possible exception of a M. guttatus plantfrom population 90-13 and four plants from 90-11, which may have had allele la. Thisallele was found at a higher frequency (still lower than allele 1 b) in four Californiapopulations (91-2, 3, 4 and 5) and was the most common allele in the one M. platycalyxpopulation surveyed (91-15) (lanes 10 and 11, Figure 3.28). The difficulty indifferentiating alleles at this locus and in picking out heterodimers in all but the bestgels has precluded its use in further analysis.Staining of another enzyme was sometimes found near the bottom of gelsstained for PGI. Similar looking bands were also observed near the top of TPI stained73gels. Since TPI and PGI are stained on the same gel these must be two differentenzymes. Superoxide dismutase (SOD) is probably responsible for one of thesegroups of bands (Wendel and Weeden, 1989). The bands were visualized as atranslucent area in the gel and appeared to be highly monomorphic.AATUp to four loci have been identified for AAT from studies of other species(Weeden and Wendel, 1989). McClure (1973) concluded from genetic andelectrophoretic analysis that there were two and possibly three AAT loci in M. guttatus.Genetic analysis of the enzyme loci was not conducted in this study however sibs andcrosses were analyzed and indicated that there were likely three distinct loci, with amiddle locus which overlapped both the fast and the slow one. Thus, the assignmentof bands to a particular locus was sometimes very difficult. The relative positions of thealleles for each locus are shown in Figure 3.29 along with the various staining patternsscored. Note the occurrence of three banded patterns 3 and 5 in Figure 3.29. Thesedo not approximate dimers so it was concluded that at least three loci were beingobserved.It is apparent from Figure 3.29 that locus 2 is problematic in that allele 2a isslightly below and stains with allele 1 b. Additionally, allele 2b is slightly above andstains with allele 3a. In some situations, such as patterns 2 and 4, it is necessary tolook at band intensities to determine which allele is present at locus 2.Patterns 2 and 6 (Figure 3.29) are typical of some individuals of M. queue. Inthese patterns, locus AAT-2 is represented by a band juxtaposed over either AAT-1 orAAT-3. The intensity of the fastest bands does appear to be greater relative to M.guttatus plants with pattern 1 and M. nasutus plants with pattern 3 indicating that locusAAT-2 is represented in patterns 2 and 6 by allele 2a rather than 2b. Two M. queueplants from population 90-9 appear to have allele 2b rather than 2a (pattern notshown). These were side by side on the same gel and there is a possibility that thiswas a gel artifact.74All M. guttatus populations had both alleles of the AAT-2 locus and all M.nasutus populations were fixed for allele 2a which, as mentioned above, is also by farthe most common allele in the four M. queue populations.Locus 1 of AAT was monomorphic for allele lb in all M. queue populations andin all the California M. guttatus and M. nasutus populations. McClure (1973) found thislocus to be invariant in the California populations she studied. A very rare allele (la) isfound at a low frequency in some of the British Columbia populations of M. guttatus andat a high frequency in the M. nasutus from population 90-07. From the bandingpatterns of these plants it was initially unclear whether this band belonged to AAT-1 or2 however they were assigned to AAT-1 on the basis of M. guttatus individuals withpatterns 5 and 8 (Figure 3.29). In addition, crosses of these distinct M. nasutusindividuals with pattern 3 with M. guttatus plants with pattern 1 yielded plants withpattern 9.Although allele AAT-3a overlaps allele AAT-2b, it was readily identified by a lackof other bands present (pattern 7, Figure 3.29). The AAT-3 locus was relativelypolymorphic for M. guttatus and one of the California populations of M. nasutus. It wasalso polymorphic in half of the M. queue populations and within population 91-17 inthat some individuals exhibited fixed heterozygosity for alleles a and b (pattern 6) whileothers were monomorphic for allele a (pattern 2).6PGDLess reliable staining was obtained for 6PGD. Staining, particularly at the slowlocus, was sometimes too light or too smeared to be reliably scored. Therefore resultswere available for fewer individuals than most of the other enzymes.Table 3.8 indicates that a unique allele (1 a) was found in one of the enzymegenotypes in the California M. queue population 91-17 (Figure 3.30, lanes 5, 8 and11). It is quite close to allele 1 b and it is possible that allele la may have been missedin other populations on gels where this locus resolved poorly. Undoubtedly though,75allele 1 a is still very rare. Ritland (1989) identified three alleles in M. guttatus for thislocus which correspond to alleles 1 b, c and d in this study.Gene diversity for the 6PGD-1 locus was higher for the California than for theB.C. M. guttatus populations. This locus was fixed for all M. nasutus populations.However, one of the California populations was fixed for a different allele, (1 c) than theother three. Mimulus guttatus population 91-8 is geographically closest to 91-09 andalso had a greater frequency of allele lc than it did of allele 1 b, which in all otherpopulations of both M. nasutus and M. guttatus was the most common.Both groups of M. guttatus had moderately high gene diversity for the slowlocus, 6PGD-2. Once again for M. nasutus populations, diversity was 0 or very low.Likewise M. queue populations were all fixed. British Columbia populations have onlythe most common allele 2b (lanes 12-18, Figure 3.30). Oregon populations 91-13(lanes 22, 23, 26-29), and the one genotype of 91-17 (lanes 7, 9 and 10) have fixedheterozygosity for alleles 2b and 2c.DIADiaphorase has not been as well studied and tends to be more complex incharacter than the other enzymes so far discussed. It has been found to bemonomeric, dimeric and tetrameric in activity depending on the plant being studied(Weeden and Wendel, 1989). In addition, from one to four DIA loci have beendocumented. Mimulus guttatus appears to have at least three zones of stainingactivity, which are apparent in Figure 3.31A. Ritland (pers. comm.) using a differentbuffer system and corolla or cotyledon tissue rather than older leaf tissue was able toscore the middle locus, which he found to be tetrameric. In the present study, thefastest migrating locus was the only one for which adequate resolution was achieved.It is apparent from Figure 3.30A that there is variation in both of the slower loci.The upper locus was found to be monomeric with two common alleles and athird rare allele. Lanes 16-19 (Figure 3.31 A) show the only four banding patterns,which were found for the upper locus. Interestingly, the allele that was most common76in the M. guttatus populations (lanes 8 and 9) was rarely represented in the M. queuepopulations and three out of the four M. nasutus populations (lanes 1-6). Allpopulations of M. queue sampled were fixed for allele 1 a as shown in Figure 3.31 B.IDHIsocitrate dehydrogenase was a rather peculiar staining enzyme in that no lessthan two bands are ever seen although it is probable that only one locus is beingstained (Figure 3.32). The bands were separated by about one mm and the fastestband was usually darker. Ritland (pers. comm.) also observed this phenomenon on amorpholine buffer system that differs from the one used to detect IDH in this study. Heconcluded through genetic analysis that this two banded pattern represented thehomozygous condition and that the second band probably represented a degradationproduct. IDH is one of the enzyme systems where "ghost" banding is sometimesencountered (Kephart, 1990). This trait is apparent for all alleles (Figure 3.32 lanes 8and 9). Dimers, however, are represented by the typical 1:2:1, three banded pattern(Figure 3.32 lane 5). Like 6PGD, IDH tended to be very streaky and sometimes as inFigure 3.32, a much fainter, slower migrating group of bands can be seen. Thisprobably represents the NAD form of IDH.All populations of M. queue and three out of four populations of M. nasutus arefixed for the most common allele, 1 b. Fenster and Ritland (1992) also observed 4alleles, and in agreement with this study, the slowest two were very rare.TPIVery good resolution was obtained for TPI. Unfortunately it was one of the leastpolymorphic of the enzymes studied. All plants were fixed for the same allele at theTPI-1 locus except for one individual from population 91-2 which was homozygous for aslower allele. Two individuals of Mimulus alsinoides collected from population 90-12were screened and they also had the common allele at TPI locus 1 indicating that thisallele is relatively conserved. Table 3.16 provides a synopsis of the alleles observed atall loci for the two Mimulus alsinoides plants assayed.77Locus TPI-2 was somewhat more polymorphic for M. guttatus and the rarer 2ballele was detected in three B. C. populations and one California population (Figure3.33 lane 11). Mimulus queue and M. nasutus were completely fixed for both TPI loci.Two other enzyme systems were analyzed, MDH and PGM. MDH was not usedin genetic diversity calculations because it has at least three overlapping loci, and itappears that there is either a gene duplication or interaction between loci. There is aband which appears to be in all plants of M. queue analyzed (Figure 3.34 lanes 1-12and 16-20), some M. guttatus, but none of the M. nasutus plants (lanes 13-15) and thusfurther electrophoretically differentiates the two small flowered species. The uppermostgroup of bands were usually very faint and the only thing that can be said about them isthat there is variation. All M. queue plants appear to have the same patterns for themiddle locus except for the one genotype of population 91-17 (lanes 7 and 10, Figure3.34) which has a longer middle smear. McClure (1973) was not able to interpret thebanding pattern for MDH but concluded that all loci appeared to have a high genediversity.Good staining for PGM was not achieved until late in the study and therefore alimited number of plants were analyzed. In agreement with Vickery et al (1989), therewas an extra PGM locus in addition to the two that are habitually found for PGM. AsPGM is a monomeric enzyme this duplication means that the minimum number ofbands observed in an individual plant, if all loci are homozygous, are three. Vickery etal (1989) suggested that this gene duplication was likely an ancient event as it was alsofound in M. rupestris Greene which is in Section Erythranthe.All M. queue stained so far have a five banded pattern (Figure 3.35, lanes 5, 11-13 and 17) indicating gene duplication at at least 2 loci. All of the M. nasutusindividuals surveyed were homozygous for the commonest alleles at each locus (lanes1-5). Thus the PGM banding profiles further differentiate M. queue and M. nasutus.As there is not enough evidence to assign positively the fastest four alleles totheir proper loci they are referred to as PGM-1 a,b,c and d. Analysis of band intensities78of M. guttatus individuals in which three out of four of these alleles are found indicatethat alleles la and lc likely occupy one locus and alleles lb and 1 d the other locus.Further evidence for this allele assignment is shown if Figure 3.35. Although this gelhas been overstained and some of the loci have begun to blur together it is stillapparent that in lane 15 there is an individual with alleles 1a and 1 b, and in lane 6, anindividual with alleles lb and 1 c. This, coupled with the fact that many individuals arefound with alleles lc and 1 d allows alleles to be assigned by the process of elimination.The slowest locus, PGM-3, appears to have one very common allele and a slowermuch rarer allele.Staining was also achieved for the following enzymes however adequateresolution was lacking and they were dropped from the study: esterase (EST),aconitase (ACO), malic enzyme (ME) and shikimate dehydrogenase (SKD). Inaddition, glyceraldehyde-3-phosphate dehydrogenase (G3PDH), fructose-bisphosphatase (F1,6DP), aldolase (ALD), leucine aminopeptidase (LAP) and the NADform of isocitrate dehydrogenase, were also assayed but staining was very light ornonexistent.Overall allelic diversity for B. C. and California M. guttatus populations seem tobe quite similar although three of the B. C. populations are at the low end of the range(Table 3.17). Both selfing species have significantly lower diversity than M. guttatus(Table 3.18). The relatively high GST (the coefficient of gene differentiation) value forthe self ing species indicates that, compared to M. guttatus, a greater proportion of theirgene diversity is between populations.It is informative to compare gene diversity for M. nasutus and M. guttatuscollected from the same sites as in populations 90-07 and 91-05. Population 91-09was also a mixed population however only one of the M. guttatus plants survived.Greater gene diversity was found in M. guttatus from population 90-07 than in the M.nasutus plants in seven out of the ten loci surveyed while the reverse was true for twoout of ten loci. While no more than eight M. guttatus plants were scored from79population 91-05, they still exhibited greater gene diversity than the fifteen M. nasutusplants from the same population in half of the loci screened. The remaining five lociwere monomorphic for both species.Table 3.6 - Allele frequencies and gene diversities of the 6PGD-1 and 2 loci for M. guttatus populations. Alleles arerepresented by letters a-e, with 'a' representing the fastest allele.region pop. a^b6PGD-1c^d di n a b6PGD-2c^d di a bIDHc di nB. C.^90-07 -^0.91 0.03 0.06 0.17 22 - 0.86 0.14 - 0.24 18 0.89 0.09 0.02 0.20 3390-09 -^1 - - 0 30 - 0.72 0.24 0.04 0.42 22 0.97 0.03 - 0.06 4690-11 -^1 - - 0 14 - 0.50 0.34 0.16 0.61 16 0.84 0.16 - 0.28 1990-13 -^0.81 - 0.19 0.31 18 - 0.74 0.12 0.15 0.41 17 0.90 0.03 0.07 0.18 3090-14 -^1 - - 0 19 0.14 0.60 0.26 - 0.55 18 0.80 0.18 0.02 0.33 22Calif.^91-002 -^0.74 0.10 0.16 0.42 19 - 0.65 0.35 - 0.46 17 0.50 0.37 0.13 0.60 1991-3 -^0.91 - 0.09 0.17 12 0.04 0.64 0.11 0.21 0.53 14 0.61 0.17 0.22 0.55 1891-04 -^0.58 0.33 0.08 0.55 12 - 0.54 0.46 - 0.50 12 0.53 0.30 0.17 0.60 1591-05, -^1 - - 0 7 - 0.25 0.75 - 0.37 8 0.27 0.09 0.64 0.51 1106,0791-08 -^0.20 0.50 0.30 0.62 10 - 0.38 0.61 - 0.48 9 0.94 - - 0.11 8* An allele faster than IDH 'a' was found only in this population and at a frequency of .06di = the population gene diversity at a single locusn = the sample sizeTable 3.7 - Allele frequencies and gene diversities of the 6PGD-1, 6PGD-2 and IDH loci for M. nasutus populationsregionpop.a b6PGD-1c^d di n a b6PGD-2c^d^dineIDHb^c di nB. C. 90-07 -^1 -^- 0 9 - 1 - -^0 7 1^-^- 0 1090-08 -^1 -^- 0 28 0.1 0.9 - -^0.18 27 1^-^- 0 27Calif. 91-05 -^1 -^- 0 11 - 0.83 0.17 -^0.28 12 1^-^- 0 1591-09 -^- 1^- 0 13 - 1 - -^0 13 1^-^- 0 13Table 3.8 - Allele frequencies and gene diversities of the 6PGD-1 6PGD-2 and IDH loci for M. queue populationsregion pop. a b6PGD-1*c^d^di n_ a^b6PGD-2*c^d^di nIDHa^b^c di nB. C.^90-07 - 1,1 -^-^0,0 17 -^1,1 - -^0,0 14 1^-^- 0 1590-09 - 1,1 -^-^0,0 18 -^1,1 - -^0,0 18 1^-^- 0 1790-10 - 1,1 -^-^0,0 18 -^1,1 - -^0,0 18 1^-^- 0 2890-11 - 1,1 -^-^0,0 15 -^1,1 - -^0,0 14 1^-^- 0 2590-12 - 1,1 -^-^0,0 11 -^1,1 - -^0,0 11 1^-^- 0 17Ore.^91-13 - 1,1 -^-^0,0 17 -^1 1 -^0,0 17 1^-^- 0 1291-17 0.15 1,0.85 -^-^0,0.26 10 -^1,0.25 0.75 -^0,0.38 12 1^-^- 0 10*The assumption was made that because fixed heterozygosity was found in at least one of the genotypes that the locus isduplicated due to tetraploidy in all genotypes. Gene diversities are calculated for both loci.Table 3.9 - Allele frequencies and gene diversities of the DIA-1, TPI-2 and PGI-2 loci for M. guttatus populationsregion pop. a bDIA-1c di n aTPI-2b^di n a b cPGI-2d e di nB. C.^90-07 0.05 0.95 - 0.09 41 0.92 0.08 0.15 39 0.08 0.56 0.04 0.32 - 0.58 4490-09 0.57 0.43 - 0.49 37 1 - 0 37 0.26 0.58 - 0.17 - 0.57 4590-11 0.50 0.50 - 0.50 36 1 - 0 20 0.05 0.61 0.14 0.20 - 0.57 2290-13 0.13 0.87 - 0.19 16 0.79 0.21 0.33 14 0.42 0.24 0.05 0.29 0.68 1990-14 0.12 0.80 0.08 0.34 25 0.32 0.68 0.44 33 0.02 0.83 0.15 - - 0.29 33Calif.^91-02 0.14 0.79 0.07 0.35 14 1 - 0 12 0.02 0.32 0.16 0.50 - 0.62 2591-03 - 1 - 0 3 1 - 0 9 0.24 0.45 - 0.32 - 0.64 1991-04 0.12 0.88 - 0.21 7 1 - 0 8 0.23 0.45 - 0.33 - 0.64 1191-05, 0.19 0.81 - 0.31 8 1 - 0 9 - 0.65 0.19 0.16 - 0.52 1306, 0791-08 0.50 0.5 - 0.50 5 0.94 0.06 0.11 8 0.10 0.35 - 0.30 0.25 0.71 10Table 3.10 - Allele frequencies and gene diversities of the DIA-1, TPI-2 and PGI-2 loci for M. nasutus populationsregion pop. a bDIA-1c^di nTPI-2a^b^di n a bPGI-2c^d e^di nB. C.^90-07 0.90 0.10 -^0.18 10 1^- 0 8 - 1 -^- -^0 1190-08 1 - -^0 29 1^- 0 26 - 0.97 0.03^- -^0.06 32Calif.^91-05 1 - -^0 11 1^- 0 14 - 1 -^- -^0 1491-09 0.25 0.75 -^0.38 8 1^- 0 9 0.05 0.95 -^- -^0.10 10Table 3.11 - Allele frequencies and gene diversities of the DIA-1, TPI-2 and PGI-2 loci for M. queue populationsregion pop.DIA-1a^b^cclinTPI-2a^b^di n a b^cPGI-2*de di nB. C.^90-07 1^-^- 0 14 1^- 0 16 0.8 0.2,1^- -^- 0.32,0 1890-09 1^-^- 0 17 1^- 0 23 1 1^- -^- 0,0 2390-10 1^-^- 0 21 1^- 0 28 1 1^- -^- 0,0 2690-11 1^-^- 0 16 1^- 0 24 1 1^- -^- 0,0 3290-12 1^-^- 0 16 1^- 0 15 1 1^- -^- 0,0 25Ore.^91-13 1^-^- 0 9 1^- 0 5 - 1^- 1^- 0,0 1491-17 1^-^- 0 6 1^- 0 8 0.67 1^- 0.33^- 0.44,0 6*The assumption was made that because fixed heterozygosity was found in at least one of the genotypes that the locus isduplicated due to tetraploidy in all genotypes. Gene diversities are calculated for both loci.Table 3.12 - Allele frequencies and gene diversities for the AAT-1,2 and 3 loci for M. guttatus populationsregion pop. aAAT-1b^dineAAT-2b^di n aAAT-3b^c^di nB. C.^90-07 0.03 0.97 .06 32 0.42 0.58 0.49 33 0.27 0.67 0.07 0.47 3090-09 - 1 0 28 0.77 0.23 0.35 28 0.13 0.82 0.05 0.31 2890-11 0.03 0.97 .06 16 0.90 0.09 0.18 16 - 0.97 0.03 0.06 1690-13 0.04 0.96 .08 25 0.55 0.35 0.45 28 0.04 0.60 0.37 0.50 2890-14 - 1 0 27 0.73 0.27 0.39 33 - 0.80 0.20 0.32 27Calif.^91-02 - 1 0 22 0.27 0.73 0.39 44 0.20 0.80 - 0.32 2291-03 - 1 0 15 0.68 0.32 0.44 28 - 0.79 0.21 0.33 1491-04 - 1 0 15 0.67 0.33 0.44 15 0.07 0.93 - 0.06 1591-05, 06, 07 - 1 0 7 0.35 0.65 0.45 7 - 0.64 0.36 0.46 791-08 - 1 0 6 0.25 0.75 0.38 6 - 1 - 0 6Table 3.13 - Allele frequencies and gene diversities at the AAT-1,2 and 3 loci for M. nasutus populationsregion pop. aAAT-1b^dinAAT-2a^b^di n aAAT-3b^c^di nB. C.^90-7 0.87 0.13 0.23 15 1^- 0 17 - 1^- 0 1790-8 - 1 0 40 1^- 0 40 - 1^- 0 40Calif.^91-5 - 1 0 15 .^1^- 0 15 - 1^- 0 1591-9 - 1 0 7 1^- 0 7 0.07 0.93^- 0.13 7Table 3.14 - Allele frequencies and gene diversities at the AAT-1,2 and 3 loci for M. queue populationsregion pop.AAT-1a^b^di n aAAT-2*b di n aAAT-3*b^c di nB. C.^90-7 -^1 0 14 1 - 0 14 1 1^- 0,0 1490-9 -^1 0 22 0.91,^.91 0.09,0.09 0.16,0.16 22 1,0.23 0.77^- 0,0.34 2290-10 -^1 0 18 1,1 - 0,0 13 1,0.06 0.94^- 0, 0.10 1890-11 -^1 0 22 1,1 - 0,0 20 1 1^- 0,0 2290-12 -^1 0 20 1,1 - 0,0 21 1,0.68 0.32^- 0,0.44 19Ore.^91-13 -^1 0 13 - 0,0 14 1,1 -^- 0,0 1391-17 -^1 0 11 1,0.91 0.09 0,0.17 11 1,0.20 0.8^- 0,0.32 10*The assumption was made that because fixed heterozygosity was found in at least one of the genotypes that the locus isduplicated due to tetraploidy in all genotypes. Gene diversities are calculated for both loci.Figure 3.28 - Enzyme electrophoresis banding patterns for PGI. Lanes 6, 12, 18, 24, 30 and 36 are all M. queue - B. C.Lanes 1-5 are M. guttatus - 91-08 Calif. Lanes 7, 8, 19-23 are M. queue- 91-17 Ore. Lanes 9-11 are M.platycalyx- 91-15 Ore. Lanes 13-17 are M. queue - 91-13 Ore. Lanes 25, 28, 29 and 31-35 areM. guttatus - 90-09 B. C. Lanes 26 and 27 are M. queue - 90-09 B. C.1-6^7-12r■In rwimmin13-18 19-24 25-30^31-36►Olio • irk_whe°1•87Table 3.15 - Comparisons of allelic profiles of the four genotypes of M. queuesampled in this study.Enzyme B.C.All Pops. 91-13Oregon91-17a 91-17b6PGD-1 bb,bb bb,bb bb,bb aa,bb6PGD-2 bb,bb bb,cc bb,cc bb,bbAAT-1 bb bb bb bbAAT-2 *aa,aa orbb,bbaa,aa aa,aa aa,bbAAT-3 aa,bb oraa,aaaa,aa aa,bb aa,aaPGI-2 aa,bb orbb,bbbb,dd aa,bb bb,ddTPI-1 aa aa aa aaTP1-2 aa aa aa aaDIA aa aa aa aaIDH aa aa aa aa*When two allele profiles are shown, the first one was the most common.Figure 3.29 - Diagrammatic representation of the assignment of alleles to AAT loci and the interpretation of some commonbanding patterns.Allele^ab a•MEMMIYMI.abcLocus^1^2^3^1 b,b 1 b,b 1 a,a^1 b,b^1 a,a^1 b,b 1 b,b^1 a.b^1 a,b 1 b,b^coco2 b,b 2 a,a 2 a,a 2 a,a^2 b,b 2 a,a 2 b,b 2 b,b 2 a,b 2 a,b3 b,b 3 a,a 3 b,b 3 b,b^3 b,b 3 a,b 3 a,a^3 b,b 3 b,b 3 b,cPattern^ 1^2^3^4^5^6^7^8^9^10bFigure 3.30 Enzyme electrophoresis banding patterns for 6PGD. Lanes 1-4 are M. guttatus - 91-05 and 91-02 Calif. Lanes5, 7-11 are M. queue - 91-17 Ore. Lanes 6 and 12-18 are M. queue - 90-10, 90-11, 90-07, 90-09 B. C. Lanes19-21 are M. guttatus - 91-08 Calif. Lanes 22, 23, 26-29 are M. queue - 91-13 Ore. Lanes 24 and 30 are M.nasutus - B. C.1-6^7-12^13-18^19-24^25-30rommommitri■I■ii- ^40 ID • • • mi, %AD • 100 eel.V14 go • il•i Ir.. b to A.B. 1-6^7-1290Figure 3.31 - Enzyme electrophoresis banding patterns for DIA. A. Lanes 1-3 are M.nasutus - 90-07 B. C. Lanes 4-7, 11 and 13 are M. queue-- 90-07, 90-11B. C. Lanes 8-10, 12, 14 and 15 are M. guttatus - 90-07, 90-11, 90-10 B.C. B. Lanes 1-6 and 12 are M. queue - 90-11, 90-10, 90-12 B. C. Lanes7-10 are M. queue - 91-13 and 91-17 Ore.91Figure 3.32 - Enzyme electrophoresis banding patterns for IDH. Lanes 1-4, 6, 10-12are M. queue - B. C. Lanes 5 and 8 are M. guttatus - B. C. Lane 8 is M.guttatus - 91-08 Calif. Lane 9 is M. nasutus - 91-05 Calif.1-6^7-12Figure 3.33 - Enzyme electrophoretic banding patterns for TPI. Lanes 1-5 and 7-11 areM. guttatus - 91-04 and 91-03 Calif. Lanes 6, 12 and 13 are M. guttatus -B. C.1-6^7-1292Table 3.16 - Alleles found in the two individuals of M. alsinoides which were sampledgrowing with M. queue in population 90-12. All alleles were in thehomozygous state.Enzyme^Locus^AlleleTPI^1 aTPI 2^bPGI^1 aPGI 2^bAAT^1 aAAT^2^bAAT^3 cIDH 1^cDIA^1 d93Figure 3.34 - Enzyme electrophoresis patterns for MDH. Lanes 1-5 are M. queue - 91-13 Ore. Lanes 6, 12, 16-20 and 24 are M. queue - 90-10, 90-09 B. C.Lanes 7-11 are M. queue - 91-17, Ore. Lanes 13-15 are M. nasutus - 91-05, Calif. Lanes 21-23 are M. guttatus -B. C.Figure 3.35 - Enzyme electrophoresis patterns for PGM. Lanes 1-3 are M. nasutus -91-05 Calif. Lane 4 is M. nasutus - 90-07 B. C. Lanes 6, 12 and 18 areM. queue - 90-12 B. C. Lanes 7-11 are M. guttatus - 91-08 Calif. Lanes13 and 14 are M. queue - 91-17 Ore. Lanes 15-17 are M. platycalyx - 91-15 Ore.94Table 3.17 - Comparisons of genetic variation between populations. The gene diversityaveraged over all loci is represented by Dt. Ten loci were scored for M.guttatus and M. nasutus and 15 for M. queue.Type Region Population Dt Alleles/Locus ProportionPolymorphiclociM. guttatus B. C. 90-7 0.24 2.4 0.9090-9 0.22 1.9 0.6090-11 0.23 2.0 0.7090-13 0.32 2.4 0.9090-14 0.27 2.1 0.70Calif. 91-2 0.33 2.2 0.8091-3 0.27 2.0 0.6091-4 0.30 2.0 0.7091-5,6,7 0.26 1.7 0.6091-8 0.29 2.0 0.70M. nasutus B. C. 90-7 0.04 1.2 0.2090-8 0.02 1.2 0.20Calif. 91-5 0.03 1.1 0.1091-9 0.06 1.3 0.30M. queue B. C. 90-7 0.02 1.1 0.0790-9 0.04 1.2 0.2090-10 0.01 1.1 0.0790-11 0 1.0 090-12 0.03 1.1 0.07Ore. 91-13 0 1.0 091-17 0.10 1.5 0.4095Table 3.18 - Comparisons of genetic variation between regions and species.Type Group AllelesperProportionPolymorphicHt Hs Dst GstLocus LociM. guttatus B. C. 2.7 0.90 0.29 0.26 0.03 0.10California 2.7 0.80 0.35 0.29 0.06 0.17Combined 3.0 1 0.29 0.27 0.02 0.07M. nasutus B. C. 1.4 0.40 0.06 0.03 0.03 0.45California 1.5 0.50 0.12 0.06 0.06 0.50Combined 1.8 0.60 0.11 0.04 0.07 0.64M.queue B. C. 1.3 0.27 0.04 0.02 0.02 0.50Oregon 1.5 0.40 0.10 0.05 0.05 0.50Combined 1.5 0.47 0.08 0.03 0.04 0.50Ht = Total gene diversityHs = Gene diversity within populationsDst = Between population diversityGst = Coefficient of gene differentiation96Chapter 4 - DiscussionTaxonomyThe M. guttatus complex has been subject to extraordinarily divergent taxonomictreatments by different authors. Pennell (1951) recognized 23 taxa and Thompson(1993) recognized 4 taxa. Despite the large number of biosystematic and evolutionarystudies, there is still no consensus on the recognition of the taxa. It is also well knownthat the M. guttatus complex can exhibit extensive phenotypic plasticity (Vickery1974A). In British Columbia, the variation in flower sizes of plants of the M. guttatuscomplex found near seeps on hillsides of Vancouver Island and the Gulf Islands hasfrequently been attributed to environmental effects. I confirmed in this study that thedifferences in flower size have a genetic basis. Furthermore, I determined that therewere two distinct small flowered species in addition to the facultatively annual, largeflowered M. guttatus. One of the small flowered species, M. nasutus, has beenstudied quite extensively in California (Vickery 1974B, Kiang 1972, 1983, Kiang andHamrick 1978). It appears to be near its northern limit in B. C. and is not common.The other small flowered species, M. queue, is newly described in this study. It is morecommon than M. nasutus in the study area and frequently grows sympatrically with M.guttatus. Populations of M. queue were also located in Oregon. Several differentexperimental approaches have been used to confirm that M. queue and M. nasutus aredistinct from M. guttatus and from one another.Hybridization StudiesThrough extensive experimental crosses between members of the M.guttatus species complex, Vickery (1974a, 1978) identified several incompletebarriers to hybridization. These barriers, however, did not appear to be correlatedwith taxonomic groupings within the M. guttatus species complex. The finding thatM. queue was essentially genetically isolated from M. nasutus and M. guttatus wastherefore unexpected. Further investigation indicated that M. queue is a tetraploid.97Ninety two percent of the crosses between M. queue and the two diploids, M.guttatus and M. nasutus, resulted in either no seed or shrivelled seed which failed togerminate. Crosses that result in no seed may simply be a function of the technicaldifficulties associated with manipulating the small flowered taxa for crossing. Theproduction of shrivelled seed, however, was a repeatable result that indicates thatsome basic incompatibility exists that prevents proper seed development. Theremaining 8 % of the crosses also produced shrivelled seed which could beencouraged to germinate on agar and occasionally on soil. Germination wasunusually slow, possibly because the shrivelled seed lacked adequate seedreserves.The resulting Fl generation was sterile and had pollen where most pollengrains were shrivelled while the rest were extremely large. These hybrids wereprobably triploids as they tended to resemble M. queue more than the diploidparent. Backcrosses were not attempted in this study so gene flow from triploidsback to the parents such as documented by Stebbins and Zohary (1959) forDactylis glomerata cannot be eliminated as a possibility in crosses between M.queue and the diploids.Mia et al. (1964) also obtained sterile hybrids from crosses between largeflowered tetraploid and diploid M. guttatus. Mia et al. (1964) did not, however,report seed shriveling analogous to that documented in this study for crossesbetween tetraploids and diploids. Shriveled seed may therefore be a problem ofgenome incompatibility rather than a function of crosses between ploidy levels.Previous researchers (Kiang 1972, Vickery 1964, 1978) have found that M.nasutus crosses readily with M. guttatus and such was the case in this study. Thehybrids were fertile although pollen fertility was reduced from a mean of 91 % in theparents to a mean of 71 % in the Fl hybrids. Kiang (1972) also noted reducedpollen fertility in M. nasutus X M. guttatus hybrids. Chromosome counts of n=13and n=14 have been found for M. nasutus (Mukherjee and Vickery 1960, McArthur98et al. 1972, Mia et al. 1964). However, Vickery (1964, 1978) found that aneuploidydid not seem to be associated with isolating mechanisms. This does not eliminatethe possibility that, in nature, hybrids are selected against. Dole (1991) reportedsuch findings for 2nd and 3rd generations of crosses between M. platycalyx and M.guttatus and attributed it to the disruption of co-adapted gene complexes byrecombination.Although no significant post-zygotic isolating mechanisms have beenidentified for crosses between M. nasutus and M. guttatus, there are two pre-zygoticisolating mechanisms that appear to decrease cross fertilization of both M. nasutusand M. queue with M. guttatus and with one another.The most obvious way in which M. queue and M. nasutus are isolated is bytheir predominantly selfing mode of reproduction. Late flowers of M. queue andboth early and late flowers of M. nasutus are often cleistogamous. In addition, bothspecies may produce cleistogamous flowers when the plants are stressed. Thereis effectively no opportunity for cross pollination to occur in these small closedflowers, except occasionally when either the stigma or anthers are exerted.Cleistogamous flowers were observed in the field on M. nasutus plants. However,cleistogamous flowers on M. queue were not observed in the field. Perhapsbecause M. queue grows on habitats that dry out early, plants die before latecleistogamous flowers are produced.Even when chasmogamous flowers are produced, the potential forinterspecific outcrossing is low because self pollination occurs before the flowersare open. Furthermore, the flowers of M. nasutus and M. queue are consistentlysmaller than those of M. guttatus with little if any size overlap. Therefore they maybe better suited to different pollinators than M. guttatus. Mimulus nasutus and M.queue are also likely to be less attractive to pollinators as they have a smalleramount of pollen and very little if any nectar. Selfing rates were not estimatedbecause both M. queue and M. nasutus were essentially monomorphic at isozyme99loci. A selfing rate of 84 % has been estimated for M. micranthus (=M. nasutus)(Ritland and Ritland 1989).In some environmentally diverse habitats such as at Nanoose Hill, all threespecies grow and flower sympatrically. They appear to be ideally suited to slightlydifferent sites and flowering times. In the field M. queue usually flowers earlier inthe year than M. guttatus and is thus able to occupy certain seepy slopes which dryout too early in the spring for M. guttatus to complete its life cycle. When M. queueand M. guttatus grow sympatrically, their flowering times often overlap. Kiang andHamrick (1978) found that California populations of M. nasutus flowered earlier thannearby facultatively annual populations of M. guttatus. In the growth chamber, M.nasutus initiates flowering before the other taxa, which supports their findings.Mimulus nasutus is near the northern limits of its range in the Gulf Islandsand on southern Vancouver Island. It was found only twice and both times near themoderating influence of the ocean. In contrast, M. queue is more common andmany of the populations are inland and at elevations of up to 400 m. Thus the twoprobably do not grow sympatrically very frequently in this area. Not enoughfieldwork was conducted in Oregon to ascertain whether sympatric populations ofthe two species are common there.Suspected natural hybrids between M. guttatus and M. nasutus occurred inpopulations 90-07, 91-05 and 91-09 and have been documented in other studies(Kiang and Hamrick 1978, Munz 1959, Vickery 1964). In addition, field studies ofmixed plantings have also indicated a limited amount of natural hybridization betweenM. guttatus and M. nasutus (Kiang and Hamrick 1978). These pre-zygotic isolatingmechanisms are therefore incomplete between M. nasutus and M. guttatus.It is also likely that pre-zygotic isolating mechanisms do not completelyeliminate fertilization between M. queue and other members of the M. guttatuscomplex. However, gene exchange between M. queue and the two diploids appearsto be an unlikely event when these two pre-zygotic isolating mechanisms are100combined with the two very strong post-zygotic barriers. Because the parentspecies are facultative or obligate annuals it is unlikely that a sterile triploid hybridwould be capable of vegetative reproduction.MorphologyMimulus guttatus is easily distinguished from M. queue and M. nasutus onceflowering has begun. Although flowers of M. guttatus are reduced in size inresponse to stress, they are never as small as the largest flowers of M. queue andM. nasutus. In addition, the pistil always protrudes from the calyx in M. guttatus to amuch greater extent than in the two small flowered taxa. This character isparticularly helpful when annotating herbarium specimens collected after flowering.Leaf characters are extremely plastic. When grown in the growth chamber, M.queue tends to have more rounded, evenly toothed leaves than M. nasutus and M.guttatus. Additionally, M. nasutus often had large, lyrate leaves with an anthocyanicsplotch at the base. However, when collected together on Nanoose Hill, they allhad small, very similar leaves and only their flowers and calyces weredistinguishing.Both M. queue and M. nasutus have small flowers and the only significantdifferences in flower measurements were in the calyx. Unfortunately, there isconsiderable overlap in all measurement characters so comparisons of the lengthsof two variables are more useful in distinguishing the species. In British Columbia,M. queue usually can be distinguished from M. nasutus by: 1) a slightly shortercalyx compared to the rest of the flower resulting in the pistil being slightly longerthan the calyx whereas in M. nasutus the pistil was almost never longer than thecalyx, 2) in M. nasutus the lower corolla lobe is 3 mm longer than the upper corollalobe while in M. queue this distance is approximately 1.5 mm, 3) the corolla tube ofM. queue is funnel shaped while that of M. nasutus tendis towards tubular, 4) M.queue was not able to attain the height, stem width and leaf size of M. nasutus.101Fewer individuals of M. queue and M. nasutus were studied from Oregon andCalifornia, so it is not clear if all these differences hold for the species throughouttheir ranges. Corolla tube shape was less distinctive in Oregon M. queuepopulations (Figure 3.17B). Separate analysis of the flower measurement data forthe B. C. and Oregon populations of M. queue did not provide any indication thatthe two groups of populations were not morphologically homogeneous. The sameis true for the B. C. and California populations of M. nasutus. In the PCA anddiscriminant analysis, M. queue from Oregon tended to group with M. queue from B.C and M. nasutus from California tended to group with M. nasutus from B. C.Problems in distinguishing M. queue and M. nasutus can arise when they aregrowing under marginal conditions because of the extreme morphological plasticitythat characterizes all members of the M. guttatus complex (Vickery 1974B). Thecleistogamous flowers are so reduced that differences between those of M. queueand M. nasutus become unperceivable. Additionally, although M. nasutus has thecapacity to become significantly larger than M. queue, under dry conditions theyboth remain very stunted. Finally, many of the differences involve the flowers,which are small and usually not adequately pressed in herbarium specimens.Enzyme ElectrophoresisEnzyme electrophoretic banding patterns further differentiated the three taxa.Although M. queue and M. nasutus were deficient in unique alleles that could beuseful as species markers, the distribution of the alleles was distinct for eachspecies. As expected from a species with a mixed mating system, M. guttatusexhibited significant heterozygosity. In contrast, M. queue and M. nasutus werealmost completely homozygous at every locus; this would be consistent with verylow rates of outcrossing. The compartmentalization of genetic variation asmeasured by the GST statistic also differed in the two small flowered species as102compared to M. guttatus. Much more of the genetic differentiation in M. queue andM. nasutus was between populations rather than within populations.Although individual M. queue plants were almost completely homozygous atenzyme loci, tetraploidy resulted in significant fixed heterozygosity and in generalmuch more complex banding patterns than was observed for M. nasutus.Except for DIA-1, the most common alleles in M. queue and M. nasutus werealso the most common alleles in M. guttatus. Although gene frequencies variedconsiderably among populations, the alleles that tended to be most common in B.C. populations of M. guttatus were also the most common in the Californiapopulations. This low amount of enzyme differentiation suggests that M. queue, M.guttatus and M. nasutus are very closely related. It also suggests a recent origin forM. queue and also M. nasutus. Plant speciation without divergence at genesspecifying soluble enzymes has also been noted by Gottlieb for Stephanomeria(1973) and Clarkia lingulata (1974).Taxonomic ConclusionsConsiderable research has been conducted on several taxa of the M.guttatus species complex, specifically, M. nasutus, M. platycalyx and M. micranthus(Pennell 1951, Kiang 1972, Kiang and Hamrick 1978, Vickery 1978, Ritland andRitland 1989, Dole 1991, 1992, Fenster and Ritland 1992), which are notrecognized in either the "The Jepson Manual Higher Plants of California"(Thompson 1993) or "Vascular Plants of the Pacific Northwest" (Hitchcock et al.1964) the two prevailing floras covering the range of the M. guttatus speciescomplex. This lack of species recognition is understandable given the complexity ofthe M. guttatus species complex and the ability of these species to cross both in thelab and in nature. However after completion of this project and an extensive surveyof the literature it is apparent to me that it is counterproductive to combine M.platycalyx, M. nasutus and M. queue with M. guttatus as the first two have103consistently been recognized as distinct by researchers of the M. guttatus speciescomplex and all three are obviously on separate evolutionary courses.Although M. platycalyxwas not recognized until 1947 by Pennell, it hasseveral characters that set it apart from other commonly recognized members of theM. guttatus species complex and is worthy of recognition. Chromosome counts forM. platycalyx indicate that it is an aneuploid with n=15 chromosomes (Mukherjeeand Vickery 1962). It is also distinct in that it is more highly selfing than M. guttatus(Ritland and Ritland 1979, Dole 1991). Morphologically it is easily distinguishedfrom M. nasutus and M. guttatus by its inflated, rather blunt calyx.Taxonomic difficulties are encountered with M. nasutus in that the name hasnot been applied consistently throughout the literature. Two recent papers (Ritlandand Ritland 1989, Fenster and Ritland 1992) have dealt with M. micranthus, a taxonGrant (1924) considered to be a variety of M. nasutus. In neither Ritland andRitland (1989), Pennell (1951) nor the original treatment (Heller 1912) is it clear howM. micranthus differs from M. nasutus and should be considered synonymous withM. nasutus.These two papers co-authored by Ritland also use M. nasutus to describe alarge flowered taxa. Grant (1924) recognized a small flowered M. nasutus specieswith a large flowered subspecies, M. nasutus var. insignis. It is not clear how thelarge flowered M. nasutus species described by Ritland and Grant differs fromfacultatively annual M. guttatus used in this study. Grant (1924) separated M.nasutus from M. guttatus by its longer upper calyx lobe however she then goes onto say that the large flowered variety of M. nasutus has a shorter upper calyx lobethan the typical M. nasutus species which would place it back in with M. guttatus inher key. The common perception that in M. nasutus the upper calyx lobe is longerrelative to the other calyx lobes than it was in M. guttatus (Grant 1924) was notborne out in this study. Previous workers were possibly measuring the calyx lobesduring fruiting rather than during flowering. At fruiting, the bottom lobes upturn104towards the upper lobes and would thus appear shorter. The use of calyx lobelength does not appear to be a character useful for the separation of M. nasutusfrom M. guttatus.In the original description of M. nasutus, Greene (1884) specified that thecalyx was distinctive in that the upper tooth (calyx lobe) almost equals the tube inlength and that the lower ones enfolded around the upper tooth so it had a verysnout like appearance in profile. The bottom lobes of M. guttatus also curve upwardas the fruit matures but not to such a degree as in M. nasutus. From myobservations, however, the differences in the fruiting calyces hinges on how thelateral lobes are folded. In M. nasutus and to a lesser degree in M. queue, they arefolded almost parallel to the upper lobe while in M. guttatus they are folded at moreof a right angle. Difficulties in separating M. nasutus from M. guttatus arise from anoveremphasis on calyx characters and are easily eliminated if flower size, wingedstem and plant habit are also stressed.Likewise M. queue is deserving of specific status. It differs from M. guttatusand M. nasutus in chromosome number, crossing ability, electrophoretic bandingpatterns, somewhat in habitat preference and can be morphologically distinguishedreliably from M. guttatus and in most cases from M. nasutus. To date I havecollected M. queue from five locations on Vancouver Island, one on SaltspringIsland and two locations in the south western area of Oregon. Examination of theM. guttatus collections at the University of British Columbia (UBC) and the RoyalBritish Columbia Museum (V) indicates that they are confined to this tiny corner ofthe province. Locally, however, M. queue does not appear to be particularly rare asspecimens from 16 new sites were identified from herbarium specimens. A verylimited amount of field work was done in Oregon and California and none was donein Washington so the extent of the distribution of M. queue is unknown although itseems probable it will also be found in Washington and northern Oregon.105A very early name, M. microphyllus Benth. exists which may be applicable towhat has been described as M. queue in this study. It was applied in 1846 tomaterial collected by Douglas along the Columbia River. Alternatively it may besynonymous with M. guttatus or even with M. nasutus in which case it wouldprecede the name M. nasutus. In the original treatment (Bentham 1846) flower sizewas not specified, although the calyx appears to be quite small. Mimulusmicrophyllus has not been interpreted consistently in succeeding treatments. InPennell (1951), M. microphyllus appears closest to M. guttatus. Gray (1876)renamed it M. luteus var. depauperatus, which was subsequently recombined intoM. guttatus var. depauperatus by Grant (1924). In Grant's description it is anannual distributed in six states from Idaho to Washington to California. The size ofthe corollas range from very small to large. Campbell (1950), synonymized M.microphyllus with M. guttatus var. typicus. Since nothing definitive can bedetermined from the treatments, the type specimen for M. microphyllus has beenrequested from Kew Gardens.In addition to recognizing M. queue, M. platycalyx and M. nasutus as distinctspecies, taxonomic treatments should also recognize that the latter two are bothcapable of hybridization with M. guttatus in the field. It should also be stressed thatunder particularly dry conditions, M. nasutus and M. queue may be difficult toseparate. Only when armed with this information can one make sense out of thecomplex situations which can be found in nature. It is biologically incorrect todisregard these three species because they are difficult to deal with.EvolutionPolyploidyHaving determined that three distinct species of the M. guttatus species complexare capable of growing sympatrically, in at least one site in B. C., the next logicalquestion is what are their evolutionary histories?106The discovery that M. queue is a tetraploid is not surprising when it isconsidered that polyploidy is a common process in plant evolution It has beenestimated that approximately 47-52 % of angiosperm species are polyploid (Grant1981). Although polyploidy is less common in annuals than in herbaceous or woodyperennials (Levin and Wilson 1976), within the polyploid annuals, there is a definitecorrelation between polyploidy and autogamy. Indeed, Grant (1981) suggests thatautogamy is a polyploidy promoting factor.Polyploids up to 6x have been found in section Simiolus of Mimulus(McArthur et al. 1972). In the M. guttatus species complex, tetraploidy appears tohave evolved on at least three separate occasions. The southern tetraploids arefound in New Mexico, Arizona, Colorado and Oregon (Mia et al. 1964, Vickery et al.1968 and McArthur 1972). All but the Oregon population are found at relatively highelevations, they are all large flowered and have stolons suggesting perenniality.Vickery (1968) indicated that they were not distinct from the surrounding diploidpopulations and were probably autopolyploids.A second group of tetraploids is found at the northern end of the speciesrange in the Queen Charlotte Islands (Calder and Taylor 1965) and have beencalled M. guttatus ssp. haidensis. These tetraploids can be distinguishedmorphologically from the surrounding diploids by leaf and pubescence characters.They also appear to grow at higher elevations. Like the previous group oftetraploids, M. guttatus ssp. haidensis is large flowered and perennial.Mimulus queue represents a third distinct group of tetraploids, whichgeographically do not appear to overlap the northern and southern tetraploids exceptfor one population identified from near the Columbia River in Oregon (McArthur et al.1972). Unlike the other two groups they are small flowered annuals which so far havebeen identified only from elevations of lower than 540m. In addition, M. queue isprobably of allopolyploid origin.107While M. queue does resemble M. nasutus, it is morphologically distinct inseveral ways that are probably not associated with that fact that it is a polyploid. It alsoexhibits fixed heterozygosity, which would not be expected in a true autopolyploidwhere any of the four homologous chromosomes can be inherited together. In a selfingspecies in particular, such tetrasomic inheritance would be expected to result in genesegregation and very quickly complete homozygosity at all loci. Whereas inallopolyploids, chromosomes from the two contributing genomes rarely pair and suchsegregation is unusual. The inability of chromosomes from the different genomes topair commonly has been attributed to genetic divergence and possibly structuraldifferences. However, according to Jackson (1976) a small number of genes that affectthe placement of chromosome attachment to the nuclear membrane can also decreasemultivalent pairing during meiosis in autotetraploids. Finally, autopolyploidy is foundalmost exclusively in outcrossing species (Bingham 1979) and clear cut examples ofautopolyploidy are not common (Soltis and Rieseberg 1986).All B C. populations of M. queue have essentially the same enzyme profiles. Intwo of the three loci where differences were detected, most of individuals exhibitedfixed heterozygosity while the remaining individuals had only a single allele. Suchresults could be caused by a loss of activity of the protein produced by the onegenome. There were also no detectable morphological differences between the B. C.populations. It is therefore probable that all the B. C. populations analyzed werederived from the same ancestor.Polyploid species can be polyphyletic (Soltis et al. 1992) and the alleledifferentiation found between the B. C. and Oregon populations and even withinOregon population 91-17 was significant. It was not, however, great enough tocompletely rule out a common origin for B. C. and Oregon M. queue if the originalparents were extremely heterozygous. It has been pointed out the M. queue plantsfrom Oregon differed somewhat morphologically from the B. C. plants. Some of thedifferences such as a greater amount of pubescence may simply reflect the different108selection pressures they have endured. Other features such as the slight differences inflower shape may be more fundamental and indicate different parentage. BritishColumbia and Oregon populations of M. queue crossed readily with each otherproducing fertile progeny. Conversely, Vickery et al. (1968) were not able to cross thenorthern tetraploids with the southern tetraploids of M. guttatus suggesting that theyoriginated independently.Several of the factors commonly associated with polyploidy and summarizedby Grant (1981) are found in M. queue. Like many polyploids, M. queue isherbaceous, it is found at relatively high latitudes and has a selfing mode ofreproduction. One of the major correlations, that of perenniality, does not hold forM. queue. In addition, unlike many diploid/polyploid groups M. queue does nothave a larger geographical range than the closely related diploids in the M. guttatusspecies complex. Other correlations have been identified when polyploids arecompared to their suspected progenitors. These are difficult to speculate on in thisstudy as it is not clear which living species if any in the M. guttatus complex are theprogenitors of M. queue. I will therefore compare M. queue to the two annualmembers of the M. guttatus complex with which it can grow sympatrically,particularly M. nasutus with which it shares its breeding behavior, certain life historytraits and a similar morphology. It should be noted also that M. platycalyx wasfound growing in Oregon within a mile of and on a very similar site to the M. queuepopulation 91-17. Mimulus platycalyx also tends towards earlier flowering,increased autogamy and is an annual (Dole 1992).It has been speculated that polyploidy can lead to slower cell division as aconsequence of the increase in the amount of DNA in each cell (Bennett 1972).Lewis (1977) demonstrated that this resulted in polyploids of Claytonia virginicawith slower growth and consequently later flowering and fruiting time than thediploids. If the approach is taken that M. nasutus is a progenitor of M. queue thensuch a scenario would provide a convenient explanation why M. nasutus appears to109be able to flower, on average, earlier than M. queue in the growth chamber. Thismay indicate that because of its polyploidy, M. queue is inherently slower growingthan M. nasutus. However, because only a single light and temperature regime wasused to grow the plants it could also be argued that these conditions were simplynot as optimal for germination and flowering in M. queue as they were for M.nasutus. In addition, conflicting studies summarized by Tal (1979) indicate thatcaution should be used when correlating an increase in ploidy level with an increasein the time required for mitosis.Field observations indicate that M. queue occupies sites which dry out veryearly in the year and thus must complete its life cycle very quickly and earlier in theyear than M. guttatus. However, as indicated by the growth chamber studies, this islikely not because M. queue has a faster rate of growth but because it is able togerminate or initiate growth earlier in the season, under cooler temperatures and/orshorter days. Similarily, Vickery (1978) found that tetraploid and aneuploidpopulations of the M. guttatus complex had a wider range of germinationtemperature tolerance than diploids.Such observations appear to support deWet's theory (1979) that mostsuccessful polyploids are allopatric with their closest relatives, or if sympatric, theytend to differ in habitat preference. Thus direct competition is avoided. In this studyit was shown that in B. C. M. queue grew successfully on certain hillsides where M.guttatus and M. nasutus were unable to colonize. Not enough is known about theOregon members of the M. guttatus species complex to speculate on habitatpartitioning in this area of its range.It could be argued that a shift in optimal environmental requirements hadoccurred in M. queue. However, such findings have often been attributed toanother commonly reported consequences of allopolyploidization, the maintenanceof increased heterozygosity in the polyploid as compared to the parents (Barrett andShore 1989). Stebbins (1971) among others has speculated that the enhanced110genetic diversity of polyploids allows them to exploit different habitats better thantheir diploid progenitors. Furthermore it is frequently suggested that there is apositive correlation between polyploidy and colonizing ability (Clegg and Brown1983). Indeed all 18 of the world's worst weeds are polyploid (Holm et al. 1977).While it is true that M. queue appears to be an effective colonizer and four out of theseven sites where it had been collected were roadcuts, M. guttatus and M. nasutusare also proficient colonizers and were also collected from roadcuts and disturbedsites.Increased heterozygosity would seem to be particularly important if theparental species were tending towards autogamy. In this study both of the selfersM. queue and M. nasutus had very low total gene diversities in the 0.04 to 0.12range. Comparable gene diversities were calculated by Fenster and Ritland (1992)for M. micranthus and for the 16 annual and biennial selfers summarized in Hamricket al. (1979). Because M. queue has fixed heterozygosity, its effective genediversity would be significantly higher. From 20-30 percent of the loci surveyed in M.queue were characterized by fixed heterozygosity (more than this if PGM isincluded) and four different enzyme genotypes were identified. The percentage ofloci exhibiting fixed heterozygosity in four selfing, polyploid weed species studied byWarwick (1990) ranged from 14% to 54% and the number of genotypes identifiedfor each species ranged from 2-10. It should be recognized, though, that Warwick'sstudy was based on weeds which have been introduced only recently into NorthAmerica and the number of populations sampled was much greater.Surprisingly, given the importance of polyploidy in plant evolution, very little workhas been conducted on the mechanism of polyploid formation. Most of the modelsdeveloped have been for perennials where there is ample time for somatic doubling tooccur. Although this and other methods have been suggested for polyploid formation inannuals (deWet 1979), Clausen et al. (1945) suggest that non-reduction of gametes innatural hybrids is the most common. Grant (1956), pointed out that fusion between a111pair of unreduced gametes is more likely in annuals if the diploid hybrid is also selfpollinating. The chances of such an event occurring therefore increases with theamount of selfing and it can therefore be speculated that the pseudocleistogamousbreeding system likely preceded the polyploidy event. Additionally, the greater thehomozygosity of the parents resulting from selfing, the more significant is the selectiveadvantage of increased functional heterozygosity brought about by allopolyploidy. Theimpact of this theory is somewhat reduced in this case when one considers that in theM. guttatus complex, polyploidy have been formed on at least two separate occasionsfrom the large flowered mixed mating forms (Mia et al., 1964). These large floweredtetraploids are perennials, however, and may have been produced in a differentmanner.AutogamyMorphology, phenology and enzyme electrophoresis all indicate that both M.queue and M. nasutus are highly selfing species. In agreement with Grant's (1981)conclusions that autogamy is usually not so complete as to exclude someoutcrossing, variation in the enzyme data as well as the presence in naturalpopulations of interspecific hybrids of M. nasutus and M. guttatus suggests that asmall amount of outcrossing is occurring in M. queue and M. nasutus.It is believed that selfers generally arise from outcrossers and are anevolutionary dead end (Stebbins 1974) although more recently the reversibility ofthe process has been demonstrated (Barrett and Shore, 1987). Greaterreproductive assurance may have helped promote the formation of the selfingspecies in the M. guttatus complex by increasing colonizing ability and, assuggested by Dole (1992), in allowing colonization of drier sites which are suitableonly for occupation early in the year when pollinators may be limiting.Like polyploidy, the evolution towards greater autogamy appears to haveoccurred in the M. guttatus complex several times, in M. nasutus, in M. platycalyx112(Dole 1992) and in M. cupriphilus, a newly described and probably recently evolvedform which is endemic to copper mine tailings in California (Macnair 1989). It istherefore not improbable that selfing also arose independently in M. queue from anoutcrossing tetraploid relative. A single tetraploid, large flowered population hasbeen identified from the Columbia Gorge area of Oregon (McArthur et al. 1972)however these were perennials. Postulating such an ancestor for M. queue wouldrequire that it evolved both an annual habitat and small flowers.Selfing has been promoted by several changes in both M. queue and M.nasutus as compared to M. guttatus . Firstly, they are both capable of producingcleistogamous as well as chasmogamous flowers. Cleistogamous flowers werenever produced by M. guttatus.The second change which serves to promote selfing in M. queue and M.nasutus is the timing of anther dehiscence. In both the cleistogamous andchasmogamous flowers of M. queue and M. nasutus anther dehiscence occurs inthe bud when the stigma is still curled down and directly adjacent to the anthers.Thus selfed pollen has at least a days head start on any that may be deposited afterthe flower opens if it does so. Occasionally cleistogamous flowers were foundwhere the anthers or the pistil had been extruded out of the flower. So although theflower never opened there was the potential for outcrossing. This was uncommonand usually occurred only when the plants were drought stressed.The third change has been the reduction in overall flower size. As has beenfound frequentlyin selfers (Ornduff 1969), most components of the flower weresmaller in M. queue and M. nasutus. The one exception is ovary length which maynot have been significantly smaller in M. nasutus as compared to M. guttatus. It isreasonable that less reduction would have occurred in the ovary length as ovarysize and seed number would be correlated and it is less likely that there would beselection for lower fecundity. It thus appears the reduction in size has not beenproportional for all flower parts. By using the log approximation of ratios it is113apparent that both pistil length and stamen length have undergone somewhatgreater reduction than the corolla for both autogamous species. Comparisons ofpistil and stamen lengths indicate that the pistil has undergone relatively greaterreduction than the stamen thus bringing the anthers and stigma into closerproximity. These findings suggest that several genes were involved in determiningflower size. Macnair and Cumbes (1989) concluded from analysis of M. guttatus, itssmall flowered derivative M. cupriphilus and their hybrids, that 3-7 genes wereresponsible for controlling floral size. However they determined that in M.cupriphilus the stigma and anthers were simply brought closer together by overallreduction in flower size.Accompanying decreases in pedicel lengths for both M. queue and M.nasutus and a reduction in calyx size in M. queue have also occurred. Thereappears to have been little change in calyx size in M. nasutus as compared to M.guttatus except possibly in calyx width. In addition the proportions of calyx lobelengths in M. nasutus do not seem to be altered as compared to M. guttatus. If it isassumed that M. nasutus is one of the progenitors of M. queue then the decrease incalyx size in M. queue as compared to M. nasutus could be explained if a shortcalyx species such as M. platycalyx was the other parent.The above changes which contribute to enhanced self pollination allcorrespond to the second category of changes documented in Leavenworthia byLloyd (1965) for selection towards autogamy. His first group of changes involvedthe loss of features associated with the attraction of pollinators. Features of M.guttatus which are believed to be important in outcrossing, such as nectarproduction and the stigmatic closure response, have been greatly reduced in M.queue and M. nasutus. Ritland and Ritland (1989) reported a complete loss of thestigmatic closure response in M. micranthus (=M. nasutus).The documented lack of gene diversity in autogamous species (Hamrick etal. 1979, Hamrick 1989) has been borne out for M. queue and M. nasutus. Fenster114and Ritland (1992) obtained similar results when comparing M. guttatus to theautogamous species M. micranthus (=M. nasutus). They questioned whether lowgenetic diversity is a function of selfing or if M. micranthus (=M. nasutus) isinherently less genetically diverse because it is a derivative species.The number of alleles per locus found for M. queue and M. nasutus in thisstudy were somewhat higher than the average calculated by Hamrick et al. (1979)for selfing species. The proportion of polymorphic loci was much larger (1.5-3.3x)for the selfers in this study as compared to those summarized by Hamrick et al(1979) and Warwick (1990). However, measurements by Fenster and Ritland(1992) for M. micranthus (=M. nasutus) are almost identical to the results obtainedfor M. nasutus in this study when the B C. and California were combined. Theunusually high diversity values obtained in this study may be an artifact of theenzymes surveyed as they can differ considerably in variability (Crawford 1990).Many of the same loci were used in both studies. In addition, the Hamrick et al.study (1979) provides data that indicate that high fecundity (>10 4) is positivelyassociated with high genetic diversity. Stebbins (1958) suggests that greaterdiversity is maintained in highly fecund populations resulting from increasedselection pressure for heterozygotes. Mimulus nasutus and M. queue certainlyhave the capacity for such high seed production. In addition, high diversity is afrequently reported characteristic of widely distributed species (Hamrick 1979).High gene diversity is expected to be advantageous to widely distributed specieswhich would be expected to experience a wider range of environments.Every measure of gene diversity for M. queue was slightly lower than for M.nasutus. However, because of its polyploidy its effective gene diversity wouldprobably be greater. The GST statistic is high for both species and indicates thatgene diversity has largely been compartmentalized between populations rather thanwithin populations. Such a finding is typical for selfing species (Hamrick 1989).Fenster and Ritland (1992) calculated a somewhat smaller GST for M. micranthus115than what was obtained in this study for M. nasutus. Possibly they found greaterdiversity within populations because their sample sizes were larger.Overall, genetic variation appears to be greater for the California and Oregonpopulations of all three species than for the B C. populations. This is particularlyapparent for M. queue and M. nasutus where Ht is twice as great for the California andOregon groups of populations as it is for the B C populations. This may suggest thatthese southern populations are nearer to the species' centers of origins than are the BC. populations. However, a more probable explanation is that because the B. C. siteswere under ice 13,000 years ago while the California and Oregon sites were notglaciated, founder effects following long range dispersal similar to those described forcolonizing weeds (as summarized in Warwick 1990) have resulted in the paucity ofenzyme diversity in B C. for M. queue and M. nasutus.Living up to a reputation earned in California studies for being complicated, theM. guttatus species complex in the Gulf Islands and southern Vancouver Island of B. Chas turned out to be an interesting group. The seepy, rocky, sunny hillsides which areinterspersed with the coniferous forest in this area, are perfect habitat for M. guttatus,and within the flora can be found three distinct species of the M. guttatus speciescomplex. They all differ in significant ways. Mimulus guttatus is a large flowered,facultatively annual, diploid with a mixed mating system. Mimulus nasutus is apredominantly selfing small flowered, annual, diploid confined in B. C. to coastallocations. Mimulus queue is a heretofore unknown tetraploid that is also a smallflowered selfer. Mimulus queue, is capable of blooming very early in the spring, ofgrowing inland and is further differentiated by its inability to cross readily with thediploids. Only when these taxa are recognized as distinctive can there be meaningfulinterpretation of research on the M. guttatus complex in this and other areas.116LITERATURE CITEDAtchley, W.R., C.T. Gaskins and D. Anderson. 1976. Statistical properties of ratios. I.Empirical results. Syst. Zool. 25: 137-148.Baker, H.G. 1955. Self-compatibility and establishment after "long-distance" dispersal.Evolution 9: 347-348.Barrett, S.C.H. and J.S. Shore. 1987. Variation and evolution of breeding systems inthe Turnera ulmifolia complex (Turneraceae). Evolution 41: 340-354.Barrett, S.C.H. and J.S. Shore. 1989. Isozyme variation in colonizing plants. In D. E.Soltis and P. S. Soltis [eds.] lsozymes in plant biology, 106-126. DioscoridesPress, Portland.Beeks, R.M. 1962. Variation and hybridization in California populations of Diplacus(Scrophulariaceae). El Aliso 5: 83-122.Bennett, M.D. 1972. Nuclear DNA content and minimum generation time in herbaceousplants. Proc. Roy. Soc. London (B) 181: 109-135.Bentham, G. 1846. Scrophulariaceae. In deCandolle [ed.] Prod. Syst. Nat. Rebn. Veg.X., Victoris Masson, Paris.Bingham, E.T. 1979. Maximizing heterozygosity in autopolyploids. In W. H. Lewis [ed.]Polyploidy, biological relevance, 471-490. Plenum Press, New York.Calder, J.A. and R.L. Taylor. 1965. New taxa and nomenclatural changes with respectto the flora of the Queen Charlotte Islands, British Columbia. Can. J. Bot. 43:1387-1399.Campbell, G.R. 1950. Mimulus guttatus and related species. El Aliso 2: 319-335.Clausen, J., D.D. Keck and W.M. Hiesy. 1945. Experimental studies on the nature ofspecies. II. Plant evolution through amphiploidy and autoploidy with examplesfrom the Madiinae. Carnegie Inst. Wash. Publ. 564: 1-174.Clegg, M.T. and A.D.H. Brown. 1983. The founding of plant populations. In C. M.Schonewald-Cox, S. M. Chambers, B. MacBryde and W. L. Thomas [eds.]Conservation Genetics, 216-228. Benjamin/Cummins, CA.Crawford, D.J. 1990. Plant Molecular Systematics - Macromolecular approaches. JohnWiley & Sons, New York.Delmer, D.P. 1972. Studies on the nature of the adaptations of the monkey flower,Mimulus guttatus to a thermophilic environment. Can. J. Bot. 52: 1509-1514.deWet, J.M.M. 1979. Origins of polyploids. In W. H. Lewis [ed.] Polyploidy - BiologicalRelevance. Plenum Press, New York.Dole, J.A. 1991. Evolution of mating systems in the Mimulus guttatus complex. Ph.Ddissertation University of California, Davis, CA.117Dole, J.A. 1992. Reproductive assurance mechanisms in three taxa of the Mimulusguttatus complex (Scrophulariaceae). Amer. J. Bot. 79: 650-659.Dudash, M.R. and K. Ritland. 1991. Multiple paternity and self-fertilization in relation tofloral age in Mimulus guttatus (Scrophulariaceae). Amer. J. Bot. 78: 1746-1753.Fenster, C.B. and K. Ritland. 1992. Chloroplast DNA and isozyme diversity in twoMimulus species (Scrophulariaceae) with contrasting mating systems. Amer. J.Bot. 79: 1440-1447.Gottlieb, L.D. 1973. Genetic differentiation, sympatric speciation, and the origin of adiploid species of Stephanomeria. Amer. J. Bot. 60: 545-553.Gottlieb, L.D. 1974. Genetic confirmation of the origin of Clarkia lingulata. Evolution 28:244-250.Gottlieb, L. 1977. Electrophoretic evidence and plant systematics. Ann. Mo. Bot. Gard.64: 161-180.Grant, A.D. 1924. A monograph of the genus Mimulus. Ann. Mo. Bot. Gard. 11: 99-388.Grant, V. 1956. The influence of breeding habit on the outcome of natural hybridizationin plants. Amer. Nat. 90: 319-322.Grant, V. 1981. Plant Speciation. Columbia University Press, New York.Gray, A. 1876. Bot. Calif. 1: 567 (As quoted in Grant, 1924).Greene, E.L. 1884. Studies in the botany of California I. Bulletin of the Calif. Acad. ofSciences 1: 66-176.Hamrick, J.L., Y.G. Linhart and J.B. Mitton. 1979. Relationships between life historycharacteristics and electrophoretically detectable genetic variation in plants. Ann.Rev. Ecol. Sys. 10: 173-200.Hamrick, J.L. 1989. Isozymes and the analysis of genetic structure in plant populations.In D. E. Soltis and P. S. Soltis [eds.] lsozymes in plant biology, 87-105.Dioscorides Press, Portland.Heller, A.A. 1912. A small flowered Mimulus. Muhlenbergia 8: 132.Heslop-Harrison, J., Y. Heslop-Harrison and K.R. Shivanna. 1984. The evaluation ofpollen quality, and a further appraisal of the fluorochromatic (FCR) test procedure.Theor. Appl. Genet. 67: 367-375.Hills, M. 1978. On ratios - a response to Atchley, Gaskins, and Anderson. Syst. Zool.27: 61-62.Hitchcock, C.L., C. Cronquist, M. Owenbey and J.W. Thompson. 1964. Vascular plantsof the pacific northwest. Part 4. University of Washington, Seattle.Holm, L.G., D.L. Plucknett, J.V. Pancho and J.P. Herberger. 1977. The world's worstweeds: distribution and biology. University Press of Hawaii, Honolulu.118Jackson, R.C. 1976. Evolution and systematic significance of polyploidy. Ann. Rev.Ecol. Syst. 7: 209-234.Johnson, R.A. and D.W. Wichern. 1982. Applied multivariate statistical analysis.Prentice-Hall Inc., Toronto.Kephart, S.R. 1990. Starch gel electrophoresis of plant isozymes: a comparativeanalysis of techniques. Amer. J. Bot. 75: 693-712.Kiang, Y.T. 1972. Pollination study in a natural population of Mimulus guttatus.Evolution 26: 308-310.Kiang, Y.T. and J.L. Hamrick. 1978. Reproductive isolation in the Mimulus guttatus-M.nasutus complex. Amer. Midl. Nat. 100: 269-276.Kiang, Y.T. 1983. Floral structure, hybridization and evolutionary relationship of twospecies of Mimu/us. Rhodora 75: 225-238.Levin, D.A. and A.C. Wilson. 1976. Rates of evolution in seed plants. Net increase indiversity of chromosome numbers and species number through time. Proc. Nat.Acad. Sci. USA 73: 2086-2090.Lewis, W.H. 1977. Temporal adaptation correlation with ploidy in Claytonia virginica.Sys. Bot. 1: 340-347.Lindsay, D.W. and R.K. Vickery Jr. 1967. Comparative evolution in Mimulus guttatus ofthe Bonneville Basin. Evolution 21: 439-456.Lloyd, D.G. 1965. Evolution of self-compatibility and racial differentiation inLeavenworthia (Cruciferae). Contr. Gray Her. 195: 3-134.Macnair, M.R. 1983. The genetic control of copper tolerance in the yellow monkeyflower, Mimulus guttatus. Heredity 50: 283-293.Macnair, M.R. 1989. A new species of Mimulus endemic to copper mines in California.Bot. J. Linn. Soc. 100: 1-14.Macnair, M.R. and Q.J. Cumbes. 1989. The genetic architecture of interspecificvariation in Mimulus. Genetics 122: 211-222.McArthur, E.D., H.T. Alum, F.A. Eldredge, W. Tail and R.K. Vickery Jr. 1972.Chromosome counts in section Simiolus of the genus Mimu/us(Scrophulariaceae). IX. Polyploid and aneuploid patterns of evolution. Madrono21: 417-420.McArthur, E.D. 1974. The cytotaxonomy of naturalized British Mimu/us. Watsonia 10:155-158.McClure, S. 1973. Allozyme variability in natural populations of the yellow monkey-flower, Mimulus guttatus, located in the north Yuba River drainage. Ph.D.dissertation Univ. California, Berkeley.119Mia, M.M., B.B. Mukherjee and R.K. Vickery Jr. 1964. Chromosome counts in thesection simiolus of the genus Mimulus (Scrophulariaceae). VI. New numbers in M.guttatus, M. tigrinus, and M. glabratus. Madrono 17: 156-160.Mukherjee, B.B. and R.K. Vickery Jr. 1960. Chromosome counts in the sectionSimiolus of the genus Mimulus (Scrophulariaceae) IV. Madrono 15: 239-245.Mukherjee, B.B. and J. R.K. Vickery. 1962. Chromosome counts in the section Simiolusof the genus Mimulus (Scrophulariaceae). V. The chromosomal homologies of M.guttatus and its allied species and varieties. Madrono 16: 141-172.Munz, P.A. 1959. A California flora. University of California Press, Berkeley.Ornduff, R. 1969. Reproductive biology in relation to systematics. Taxon 18: 121-133.Pennell, F.W. 1951. Mimulus. In L. Abrams [ed.] Illustrated Flora of the Pacific States,688-731. Stanford University Press, Stanford.Ritland, K. 1989. Correlated matings in the partial selfer Mimulus guttatus. Evolution43: 848-859.Ritland, K. and F.R. Ganders. 1987. Covariation of selfing rates with parental genefixation indices within populations of Mimulus guttatus. Evolution 41: 760-771.Ritland, C. and K. Ritland. 1989. Variation of sex allocation among eight taxa of theMimulus guttatus species complex (Scrophulariaceae). Amer. J. Bot. 76: 1731-1739.Roose, M.L. and L.D. Gottlieb. 1976. Genetic and biochemical consequences ofpolyploidy in Tragopogon. Evolution 30: 818-830.Soltis, D.E., D.H. Haufler, D.C. Darrow and G.J. Gastony. 1983. Starch gelelectrophoresis of ferns: a compilation of grinding buffers, gel and electrodebuffers, and staining schedules. Amer. Fern J. 73: 9-27.Soltis, D.E. and L.H. Rieseberg. 1986. Autopolyploidy in Tolmiea menziesii(Saxifragaceae): genetic insights from enzyme electrophoresis. Amer. J. Bot. 73:310-318.Soltis, P.S., J.J. Doyle and D.E. Soltis. 1992. Molecular data and polyploid evolution inplants. In P. S. Soltis, D. E. Soltis and J. J. Doyle [ed.] Molecular systematics ofplants, 177-201. Chapman and Hall, New York.Stebbins, G.L. 1958. Longevity, habitat and release of genetic variability in higherplants. Cold Spring Harbor Symp. Quant. Biol. 23: 365-378.Stebbins, G.L. and D. Zohary. 1959. Cytogenetic and evolutionary studies in the genusDactylis. VII. I. Morphology, distribution, and interrelationships of the diploidsubspecies. Univ. Calif. Publ. Bot. 31: 1-40.Stebbins, G.L. 1971. Chromosomal evolution in higher plants. Edward Arnold(Publishers) Ltd., London.120Stebbins, G.L. 1974. Flowering Plants: Evolution Above the Species Level. BelknapPress, Cambridge, MA.Tal, M. 1979. Physiology of polyploids. In W. H. Lewis [ed.] Polyploidy, BiologicalRelvance, 61-75. Plenum Press, New York.Thompson, D.M. 1993. Scrophulariaceae. In J. C. Hickman [ed.] The Jepson Manual-Higher Plants of California, University of California Press, Berkeley.Vickery, R.K., Jr. 1959. Barriers to gene exchange within Mimulus guttatus(Scrophulariaceae). Evolution 13: 300-310.Vickery, R.K., Jr. 1964. Barriers to gene exchange between members of the Mimulusguttatus complex (Scrophulariaceae). Evolution 18: 52-69.Vickery, R.K., Jr. 1974A. Growth in artificial climates-an indication of Mimulus' ability toinvade new habitats. Ecology 55: 796-807.Vickery, R.K., Jr. 1974B. Crossing barriers in the yellow monkey flowers of the genusMimulus (Scrophulariaceae). Genet. Lect. 3: 33-82.Vickery, R.K., Jr. 1978. Case studies in the evolution of species complexes in Mimulus.Evol. Biol. 11: 405-507.Vickery, R.K., Jr., K.W. Crook, D.W. Lindsay, M.M. Mia and W. Tai. 1968.Chromosome counts in section Simiolus of the genus Mimulus(Scrophulariaceae). VII. New numbers for M. guttatus, M. cupreus, and M. tilingii.Madrono 19: 211-218.Vickery, R.K., Jr., J.W. Ajioka, E.S.C. Lee and K.D. Johnson. 1989. Allozyme-basedrelationships of the populations and taxa of section Erythranthe (Mimulus). Am.Midl. Nat. 121: 232-244.Warwick, S.I. 1990. Allozyme and life history variation in five northwardly colonizingNorth American weed species. Pl. Syst. Evol. 169: 41-54.Weeden, N.F. and J.F. Wendel. 1989. Genetics of plant isozymes. In D. E. Soltis andP. S. Soltis [eds.] Isozymes in Plant Biology, 46-72. Dioscorides Press, Portland,Ore.Wendel, J.F. and N.F. Weeden. 1989. Visualization and Interpretation of PlantIsozymes. In D. E. Soltis and P. E. Soltis [eds.] Isozymes in Plant Biology, 5-45.Dioscorides Press, Portand, Ore.Wilkinson, L. 1990. SYGRAPH: The system for graphics. SYSTAT, Inc., Evanston, IL.Wyatt, R. 1988. Phylogenetic aspects of the evolution of self-pollination. In L. D.Gottlieb and S. K. Jain [eds.] Plant Evolutionary Biology, Chapman & Hall Ltd.,New York.121Appendix IMimulus queue Benedict, sp. nov.Herba annua obligata, a Mimulus guttatus Fischer ex DC. pistillo 5-13 mmlongo, corolla 6-20 mm longa et pistillo calycem aequante vel paulo longiore differt; aMimulus nasutus Greene calyce 5-13 mm longo, tubi apica tubi base diplo latiore,foliis non bullatus, et caulis non alatis differt; planta tetraploidea.Annual or winter annual herb, bearing opposite pedicilate basal leavesgraduating into sessile cauline leaves, 3-25 cm high, glabrous to minutely pubescent .Roots fibrous.Leaf blade palmately veined, regularly denticulate, widely ovate, apex obtuseto acute, 0.5-3 X 0.5-2.5 cm becoming gradually reduced up the stem; leaf bladeabove green frequently with anthocyanic spotting, glabrous to minutely pubescent,veins often purplish red near leafbase; leaf surface below silver-green to purple,glabrous, veins green. Petiole 0-2 cm long, green-white to red-white, glabrous. Stemtending to quadrangular, <2 mm.Inflorescence, few flowered to racemose, terminal; 1 primary raceme,occaisional secondary racemes arising from leaf axils, flowers opposite in leaf axils.Pedicel 3-22 mm long, red, glabrous.Calyx 5-13 mm long, upper calyx lobe longer than other four, green, often withanthocyanic spotting, white hairs on margin. Corolla bilabiate or sometimescleistogamous, 5-22 X 2-13 mm, yellow, corolla lobes subequal, palate densely hairy,red spotted, extending into tube as two ridges; tube 4-13 mm long. Stamensdidynamous, two shorter stamens exterior; long stamen 4-12 mm. Pistil 5-13 mm; stylewhite, minutely pubescent; stigma yellow; ovary, 2-5 mm, green; stipe 0-1 mm. Stigmalobes may be thigmotropic.Capsule dehiscing by longitudinal slits, persistent style, crowned by apersistent calyx; lower calyx lobes curved upwards toward upper calyx lobe. Seedsup to 300 per capsule, oval, brown, 0.5 X 0.2 mm.Found on wet, sunny, hillsides and cutbanks on Vancouver Island and the GulfIslands, British Columbia and in the southwestern part of Oregon, from sea level to600 m. Flowers from late March to May.Tetraploid.TYPE: CANADA. Prov. British Columbia: on a southwest facing, open, wet hillside inSooke Potholes Park beside the Sooke River, 75 m, 48°26"N, 123°43"W, 1 May 1991,Benedict 91-16 (holotype, UBC); south slope of Nanoose Hill above Nanoose Bay, 20- 50 m, 49°16"N, 124°12"W, 1 May 1991 Benedict 91-17 (paratype, DAO).Additional specimens examined (if only one number is given it is the accessionnumber of the herbarium where the specimen is located): CANADA: B. C.: LasquetiIsland, Trematon Mountain, 19 May1985, 144698 (V); N. Pender Island, Oak Bluffs, 4Ap1983, 133335 (V); Saltspring Island, 5 1/2 km southwest of Ganges, Lot 34, 18Ap1976, 136977 (V); Mayne Isl., Heck Hill, open bluff, 13 March 1980, 107521 and 6Ap1979, 98035 (V); Galiano Island, 12 May 1975, 97333 (V); V.I., Gonzales Hill nearVictoria, Ap 1916, 42590 (V); V.I., Cedar Hill district near Victoria, Ap 1916, 52967 (V);V.I., Maculey Point near Victoria, 1 June 1907 (V); V.I., Alberta Head, 42592 (V);Denman Island, wet cliffs facing Hornby Island, 7 Jul 1952, Brink 68843 (UBC); V.I.,Durrance Lake drainage on rock outcrop, 9 May 1963, Young 108599 (UBC); V.I.,Ucluelet, rocky ledges, 23 May 1975, Rose 177970 (UBC); V.I., Anderson Hill inVictoria, 17 May 1950, Krajina and Spilsbury 55012 (UBC); V.I., Mount Wells, 8 miles122W of Victoria on moist, rocky cliffs, 12 May 1957 (Calder and Taylor 20776) 80960(UBC); V.I., Esquimalt, 17 Ap 1917, Darling45840 (UBC); V.I., Victoria, 4 March 1912,Henry80455 (UBC); V.I., Finlayson Arm Hills, under power line, 22 Ap 1978 (AC); V.I.Maple Mountain, 9 Ap 1978, (AC); Saltspring Island, south slope of Mount Maxwell, 5May 1978 (AC); V.I., Koksilah Park, 2 May 1978 (AC); V.I., Sooke Hills, RuggedMountain north of Glinz Lake, 21 May 1981 (AC); V.I., Mount Jeffrey, Malahat Ridge, 3May 1975 (AC); V.I., Maple Bay, Arbutus Point, 5 May 1974 (AC)Relationships: Very similar to M. nasutus. All characters overlap to a degreewith M. nasutus but, under favorable growth conditions, the following structures tendto be more enlarged in M. nasutus (M. queue measurements are in brackets): stemwidth <4 mm (<1), calyx length 6-16.5 mm (5-13), leaves 0.5-10 X 0.5-7.5 cm (0.5-3 X0.5-2.5), height 5-50 cm (3-25), pedicel length 4-26 mm (3-22), stipe length 0.5-2 mm(0-1). Mimulus nasutus tends to have a shorter pistil as compared to its calyx and thedifference in calyx and pistil lengths range from 0-6 mm(-2.5-3.5). The ratio of thewidth of the top of the floral to to the base in M. nasutus is usually <2 (>2). Mimulusnasutus tends to have a more sharply four angled and winged stem and the leavesare often bullate.Origin of name: Initially it was questionable as to the taxonomic status of thetaxon. Later in the study it was identified as a distinct line (queue).V = Royal British Columbia Museum HerbariumUBC = University of British Columbia HerbariumDAO = Ottawa Biosystematics Research CentreAC = Adolf Ceska collection, not yet accessioned123Appendix IIKey to the species in the Mimulus guttatus complex*1a Bracts at nodes subtending flowers completely fused around stem forming a ±circular disk, glaucous^ M. glaucescans1b Bracts at nodes subtending flowers petioled or fused around stem only at theirbases, not forming a circular disk, not glaucous2a Leaves or at least some ± pinnately lobed or dissected into narrow segmentsM. laciniatus2b Leaves ± entire or ± crenate, not pinnately lobed or dissected (but base oftenirregularly dissected or small lobed)3a Bract or leaf pairs at nodes subtending flowers linear to lanceolate, notfused at base^ M. nudatus3b Bract or leaf pairs at nodes subtending flowers ovate to cordate or round,sometimes fused at base around stem4a Calyx much inflated, cup shaped and blunt (i.e. the upper calyx tooth isonly slightly longer than others)^ M. platycalyx4b Calyx somewat inflated but not cup shaped, upper calyx toothprominently longer than others5a Corolla > 20 mm long, pistil exserted from the calyx at least 5 mmM. guttatus5b Corolla < 20 mm long, pistil scarcely or not exserted from the calyx6a Pistil included within or equal to calyx, corolla tube nearlycylindrical, plants from 5 - 50 cm tall, large ones withquadrangular winged stem, diploid^M. nasutus6b Pistil usually exserted from calyx (up to 3 mm), corolla tubenarrowly funnel-shaped (infundibular), plants from 5-25 cm tall,stems tending to quadrangular but not winged, tetraploidM. queue*Couplets 1-3 were taken from Thompson (1993)


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