UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Seed ecophysiology and plant population ecology of Cynoglossum officinale L. and Tragopogon spp. Qi, Meiqin 1993

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

Item Metadata


831-ubc_1993_spring_phd_qi_meiqin.pdf [ 8.45MB ]
JSON: 831-1.0086387.json
JSON-LD: 831-1.0086387-ld.json
RDF/XML (Pretty): 831-1.0086387-rdf.xml
RDF/JSON: 831-1.0086387-rdf.json
Turtle: 831-1.0086387-turtle.txt
N-Triples: 831-1.0086387-rdf-ntriples.txt
Original Record: 831-1.0086387-source.json
Full Text

Full Text

Seed Ecophysiology and Plant Population Ecology ofCynoglossum officinale L. and Tragopogon spp.byMEIQIN QIB.Ag., Inner Mongolia Agriculture College, P.R. China, 1982M.Sc., University of Saskatchewan, 1988A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIES(Department of Plant Science)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAAPRIL^1993© Meiqin Qi, 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.Department of I L The University of British ColumbiaVancouver, CanadaDate  Acirl 2_6 193DE-6 (2/88)iiAbstractIn order to understand persistence strategies of Cynoglossumofficinale L., Tragopogon pratensis L., and T. dubius Scop. inthe rangelands of British Columbia, ecophysiologicalcharacteristics of seed germination in C. officinale andTragopogon spp. and dynamics of seed and plant populations in T.pratensis were investigated.A deep innate (primary) dormancy was found in C. officinale seeds, and the seed coat of this species played an important rolein the regulation of seed germination by controlling 0 2availability to the embryo. Removal of the seed coat stimulated02 uptake which was due to both an increase in seed respirationand phenolic oxidation; seeds of C. officinale contained highlevels of phenolic substances and seed extracts showed highpolyphenol oxidase activity. The seed coat also preventedleaching of phenolic substances, which however, were notnecessary for seed germination. Rosmarinic acid, the mostprominent phenolic substance present in the seeds of C.officinale, did not inhibit germination of decoated seeds atconcentrations up to 3 mM. Analysis of the pattern ofmethanol-extractable phenolic substances showed no significantquantitative or qualitative correlation between changes inspecific phenolic compounds and seed germination induced bystratification or seed coat removal. These results suggest thatiiiphenolic substances in C. officinale seeds do not inhibitgermination of this weed, but seed coat regulates germination bycontrolling 0 2 availability to the embryo.Ecophysiological characteristics of T. pratensis and T.dubius seed germination were studied under controlled conditions.Existence of secondary dormancy in these species was demonstratedby inducing dormancy by anaerobiosis treatment. This dormancy mayplay an important role in the persistence of Tragopogon spp.Secondary dormancy could be released by stratification andafter-ripening in T. pratensis, but not by light. Seedsgerminated over a range of temperatures with 15 C as the optimumin both weeds. A seed burial study showed that increased seedingdepth had no effect on seed germination, but reduced seedlingemergence. These results suggest that seeds in Tragopogon spp.can germinate in darkness and the reservoir of non-dormant seedswill become completely depleted under conditions conducive togermination regardless of burial depth.Patterns of plant mortality and reproduction in populationsof T. pratensis and dynamics of its seed banks were investigated.Seedlings were marked in 1990 and 1991 and their fates monitored.High density-independent seedling mortality was observed in bothcohorts. None of the plants flowered in the first or secondgrowing season in the 1990 cohort. These results indicate that T.pratensis is not a biennial species. Size of age-specific plantsvaried significantly suggesting variations in growth rate amongivindividuals. Flowering and seed production in non-age-specificpopulations were correlated with the root crown diameter (RCD).The variation of minimum RCD for plants that flowered in twoyears (0.2 and 0.6 cm) suggested that T. pratensis does not havea critical size requirement for flowering. The dynamics of seedbanks were studied in space and time. Seed burial depth had asignificant effect on seed dormancy after 2 to 3 months ofburial. The seed populations were almost completely depletedafter 9 to 10 months of burial. Since T. pratensis only has atransient seed bank, high seed production and long vegetativesurvival may play important roles in the persistence of thisweed.Table of ContentsPageAbstract^ iiTable of Contents^List of Tables viiList of Figures viiiAcknowledgments xiGeneral Introduction^  1Chapter 1. Literature review^  61. Ecophysiology of seed germination and seed dormancy^ 6A. External factors controlling seed germination^ 8B. Effects of seed coat on germination^  102. Dynamics of seed populations^  13A. The seed bank and its implications on weeds^ 13B. Population dynamics of buried seeds^  14C. Seasonal changes of germination and dormancy  163. Dynamics of plant populations^  17A. Life history^  17B. Reproductive behaviour in relation toweed persistence  194. Weed biology^  21A. Cvnoglossum officinale^  21B. Tragopogon pratensis and T. dubius^  25Chapter 2. Mechanisms of seed dormancyin Cynoglossum officinale^  28Abstract^  28Introduction  29Materials and methods^  30A. Seed source  30B. Germination studies  30C. Permeability of seed coat to water^  31D. Respiration measurements^  31E. Enzyme extraction and assay  32F. Leaching of Folin reagent-active substances^ 32G. Extraction and assay of phenolic substances  32Results^  33Discussion  48Chapter 3. Seed germination ecophysiologyof Tragopogon pratensis and T. dubius^  54viAbstract^  54Introduction  55Materials and methods^  57A. Seed source  57B. Germination study  57C. Induction of secondary dormancy^  57D. Effect of temperature on seed germination^ 58E. Effect of light on seed germination  59F. Effect of seeding depth on seedling emergence^ 60G. Statistical analysis^  60Results^  61Discussion  69Chapter 4. Seasonal changes of seed dormancyof buried Tragopogon pratensis seeds^  78Abstract^  78Introduction  79Materials and methods^  82A. Description of study sites^  82B. Field data collection  84C. Data analysis  86Results^  88A. Effect of burial depth on seed dormancy^ 88B. Seasonal change in seed dormancy^  91C. Comparison of survivorship curves between years^ 94Discussion^  94Chapter 5. Population ecologyof Tragopogon pratensis plants^  107Abstract^  107Introduction  108Materials and methods^  110A. Study sites  110B. Definitions  111C. Life history study  111D. Flowering behavior and seed production^ 112E. Data analysis^  113Results^  114A. Pattern of age-specific plant survivorship^ 114B. Size distribution of plant populations  118C. Reproduction of T. pratensis plants^  121Discussion^  133General conclusions^  144A. Cynoglossum officinale^  144B. Tragopogon pratensis and T. dubius^  145References^  150vi iList of TablesPageChapter 2Table 1. Effects of pricking (P) and GA, (GA) treatmentson Cynoglossum officinale seed germingtion ^  37Table 2. Effect of rosmarinic acid (RA) and coumarin onCynoglossum officinale seed germination  43Chapter 3Table 1. Induction of secondary dormancy in Tragopogonpratensis and T. dubius seeds by anaerobiosis treatment^ 62Chapter 4Table 1. Effect of burial depth on populationdynamics of Tragopogon pratensis seedsin 1989 cohort ^  89Table 2. Effect of burial depth on populationdynamics of Tragopogon pratensis seedsin 1990 cohort^  90Table 3. Life table of Traqopogon pratensis seedpopulations at Riske creek, B.C. for 1989 and 1990cohorts^  96Chapter 5Table 1. Life table of Tragopogon pratensis at RiskeCreek, B.C. from 1990 to 1991^  116Table 2. Summary statistics of size (RCD) distributionof age-specific plant population of Tragopogon pratensis in 1990 cohort ^  122Table 3. Summary statistics of size (RCD) distribution ofnon-age-specific plant populations of Tragopogon pratensis in 1990 and 1991 cohorts^  124viiiList of FiguresPageChapter 2Fig. 1. Effect of removal of seed coat on germinationand on water uptake of Cynoglossum officinale seeds^ 34Fig. 2. Effect of washing to remove inhibitors on seedgermination ^  36Fig. 3. Effect of seed coat on 02 uptake and CO 2release  38Fig. 4. Effect of pricking on 02 uptake andseed germination^  39Fig. 5. Polyphenol oxidase activity in Cynoglossumofficinale seed extracts^  41Fig. 6. Germination inhibition of decoated seeds byphenolic compounds  42Fig. 7. Leaching of Folin reagent-active substances bydecoated and coated Cvnoglossum officinale seeds ^ 44Fig. 8. Time-course of germination of decoated seeds andlevel of phenolic compounds ^  46Fig. 9. Effect of stratification on seed germination andthe level of phenolic compounds  47Chapter 3Fig. 1. Effect of anaerobiosis treatment on Tragopogonpratensis and T. dubius seed viability ^  63Fig. 2. Effect of temperature on germinationof Tragopogon pratensis and T. dubius seeds  65Fig. 3. Germination response of dormant Tragopogonpratensis seeds to duration of stratification (5 C)and the temperature during germination assay^ 66Fig. 4. Effect of temperature during after-ripening ofair-dried Tragopogon pratensis seeds on seed dormancy^ 68Fig. 5. Effect of seeding depth on emergenceof Tragopogon pratensis and T. dubius seedlings ^ 70ixChapter 4Fig. 1. The mean temperature and precipitation in RiskeCreek, Williams Lake of B.C. between August 1989and August 1991 ^  83Fig. 2. The distribution of Tragopogon pratensis seedsin soil profile  85Fig. 3. The fate of Tragopogon pratensis seeds buriedin 1989 at 0, 2, and 5 cm depths in the soil^ 92Fig. 4. The fate of Tragopogon pratensis seeds buriedin 1990 at 0, 2, and 5 cm depths in the soil  93Fig. 5. Survivorship curves of buried Tragopogon pratensis seeds in 1989 and 1990 cohorts ^  95Fig. 6. Possible within-year germination schedules ^ 103Chapter 5Fig. 1. The survivorship curves of Tragopogon pratensis^ 115Fig. 2. The relationship between plant density and mortalityof Tragopogon pratensis populations ^  117Fig. 3. The effect of temperature and precipitation on theage-specific mortality rates and killing power ofTragopogon pratensis population ^  119Fig. 4. The size (RCD) distributions of age-specific plantpopulation over two growing seasons  120Fig. 5. The size (RCD) distribution of non-age-specificplant populations in two years ^  123Fig. 6. Proportions of flowered, non-flowered, anddead plants of Tragopogon pratensis in two cohorts^ 125Fig. 7. The relationship between the flowering percentage ofTragopogon pratensis plants and the RCD ^  127Fig. 8. Secondary flowering of Tragopogon pratensis after removal of the primary flowering  129Fig. 9. The frequency distributions of flower heads perplant and florets per plant in Tragopogon pratensis ^ 130Fig. 10. Relationship between RCD and floret and flowerxhead in Tragopogon pratensis, and the correlationbetween floret number and flower heads per plant in 1990.... 131Fig. 11. Relationship between RCD and floret production inTragopogon pratensis in 1991 ^  132xiAcknowledgmentsI am grateful to the Grassland Research Institute, ChineseAcademy of Agriculture Science for giving me this opportunity tostudy in Canada.My sincere thanks go to my advisor Dr. Mahesh, K. Upadhyayawho supported my work generously with his time and advice, notto mention his research grant.I would also like to express my thanks to my supervisorycommittee, Drs. Roy A. Turkington, Michael D. Pitt, and JudithH. Myers for their valuable advice.My sincere thanks also go to Fred Knezevich and DonBlumenauer. It would have been impossible for me to work in thefield without their generous help.I would also thank my colleagues and friends in theDepartment of Plant Science for their help and friendship duringlast few years. In particular I want to thank Ted Herrington,Dave Konesky, Nancy Furness, Jana Moziskova, Monika MacHutchon,Kate Sircome, Leroy Scrubb, and Juan Hoyos. My thanks also go tomy dearest friends Yuncai Gao, Guangxi Wu, Shewfong Yu, LinYang, and Jin Zhao who comforted me when I was down, andencouraged me when I was hesitating.Finally, but most importantly, I would like to thank myhusband Hongsheng, for his understanding, patience,encouragement, and most of all for his love. All his love duringthe last six years made our long distance relationshipsuccessful.1General IntroductionCynoglossum officinale L. (hound's-tongue), Tragopogonpratensis L. (meadow goat's-beard, meadow salsify) and T. dubiusScop. (goat's-beard, western salsify), all of which are nativeto Europe, have become common rangeland weeds in BritishColumbia (Alex et al., 1980; Upadhyaya et al., 1988).Cynoglossum officinale was listed as a noxious weed in B.C. in1986 due to its toxicity to livestock and large scaleinfestation of rangelands. Why these weeds are so dominant andpersistent in the rangelands is not known. Information on thephysiological and ecological basis of persistence of these weedsand on mechanisms controlling their densities is required beforesuccessful control measures can be developed.Persistence of many weeds depends on their seed dormancy andability to maintain seed banks (Harper, 1959; Harper and White,1974; Simpson et al., 1989). Genetic variability in duration ofseed dormancy and the influence of the environment distributeseed germination over time, and delayed germination leads to thedevelopment and maintenance of seed banks (Karssen, 1982;Bouwmeester and Karssen, 1989; Parker et al., 1989). The extentof germination after a seed separates from the mother plantdepends on whether innate dormancy is present and whether thetemperature, moisture and light requirements of the otherwisenon-dormant seeds are met (Roberts, 1986; Parker et al., 1989).Determining the reservoir of weed seeds in the soil, the rate of2seedling recruitment, and seedling mortality can allowprediction of potential weed infestations (Radosevich and Holt,1984). While C. officinale and Traqopogon spp. reproduce only byseeds (Alex et al., 1980), ecological and evolutionary rolesplayed by seeds in persistence of these weeds are poorlyunderstood.Persistence of some weeds is also affected by delay in seedproduction for several years (Kelly, 1989b; De Jong et al.,1989). Many biennial species take three to five years to produceseeds because a critical rosette size (diameter) is required forflowering (Harper and Ogden, 1970; Werner, 1975; Oxley, 1977;Baskin and Baskin, 1979; Van der Meijden and Van der Waals-Kooi,1979; Gross, 1981; Kachi and Hirose, 1983; De Jong et al., 1986;Klinkhamer et al., 1987; Powell, 1988). If seed production isdelayed, a large number of vegetative plants may form a bud bankwhich can buffer the population against low seedling recruitmentor high seedling mortality in some years (De Jong andKlinkhamer, 1988). The life history characteristics such assurvivorship, growth rate, and reproductive output of a plantpopulation are components of its evolutionary fitness (Venable,1984). Understanding of the characteristics of each growth stagein a plant's life cycle can provide an insight into thepersistence strategy of a species.The processes controlling weed population densities arecomplex in agroecosystems. Effective weed management depends oninformation provided by ecological theories and studies.3Although ecological studies have been emphasized in weed scienceduring the last two decades, objectives have usually beenfocused either on seed population or plant population dynamicsper se. Since such studies draw conclusions from only one ofthese two aspects to predict entire plant population, theconclusions drawn can be misleading. In addition, physiologicalstudies are essential to the understanding of the mechanismsregulating the persistence of a weed. Therefore, bothphysiological and ecological information on factors thatregulate population density should be obtained in order tounderstand abundance of a weed in nature.Research objectivesCynoglossum officinale seeds showed a deep primary dormancyin preliminary studies. The mechanism(s) regulating thisdormancy is (are) not known. This thesis focuses on themechanism of primary seed dormancy in C. officinale, because itcould determine the potential of this weed to form a persistentsoil seed bank. Persistence strategies of Tragopogon spp. areexplored by investigating the main regulating factors ofpopulation density at certain growth stages, particularly theirseed germination behaviour, seasonal changes in seed dormancy,and dynamics of plant population.The overall objective of this thesis is to understand somefactors that affect the persistence of C. officinale and4Tragopogon spp. Ecophysiological characteristics of seeds of C.officinale and Tragopogon spp., and population ecology of T.pratensis plants were investigated to determine:1) the mechanism(s) involved in the regulation of primarydormancy of C. officinale seeds,2) the interaction of Tragopogon spp. seed germinationregulation with environmental conditions (e.g. temperature,light, and seeding depth) and whether these species rely onsecondary dormancy for their persistence,3^if T. pratensis relies on a persistent soil seed bank tomaintain its population,4 roles of prolificity and delayed reproduction in thepersistence of T. pratensis populations in rangelands.Answers to these questions will help understand thepersistence strategies of C. officinale and Tragopogon spp. inrangelands and develop effective control measures for theseweeds.Organization of this thesisChapter 1 reviews the main factors that affect persistenceof other species in nature. Roles played by seed dormancy,5dynamics of seed banks, and reproductive characteristics in someweeds in relation to their persistence are specificallyemphasized. Some biological characteristics of C. officinale, T.pratensis, and T. dubius found in other studies are alsosummarized. Four objectives outlined above are studied inchapters 2, 3, 4, and 5 respectively. The specific objectivesare given following a brief introduction in each individualchapter. Finally, general conclusion is followed.6Chapter 1. Literature review1. Ecophysiology of seed germination and seed dormancyGermination, the transition from seed to seedling, isperhaps the most vulnerable stage in the survival of plants(Radosevich and Holt, 1984; Fenner, 1985). Viable seeds that donot germinate, under conditions otherwise favorable forgermination, are considered dormant (Vegis, 1964). Dormancy is amechanism of enhancing survival at a time when the environmentis unfavorable for germination (Kozolowski, 1972; Harper, 1977;Egley and Duke, 1985).Dormancy has been classified in various ways. Oneclassification divides dormancy into primary and secondarydormancy (Bewley and Black, 1982). Seeds that are dormant at thetime of release from the parent plant are called primarydormant. Seeds that acquire dormancy sometime after theirdetachment from the parent plants are termed secondary dormant.Whether physiological mechanisms involved in the regulation ofprimary and secondary dormancy are similar is not known.However, both primary and secondary dormant seeds in manyspecies respond to various dormancy-breaking treatmentssuggesting a physiological similarity in the two types ofdormancy (Bewley and Black, 1982).Another widely-used system suggested by Harper (1959)divides dormancy into innate, induced, or enforced dormancy.Innate dormancy is the same as primary dormancy, i.e., the7dormancy with which a seed is shed. Induced dormancy is the sameas secondary dormancy. This is induced in seeds by particularenvironmental conditions and will persist even after theinductive conditions have passed and environment has becomefavorable for germination. Enforced dormancy occurs when theseed is prevented from germination by the ambient conditions,such as during a drought period or in cold winters. Thisdormancy does not persist and the seed germinates whensurrounding conditions become favorable.When primary dormancy is not significant, secondary dormancymay be an important factor in the development of a persistentseed bank. Induction of secondary dormancy by anaerobiosis hasbeen reported in many species such as apple and wild oats (Comeand Tissaour, 1973; Tilsner and Upadhyaya, 1985). Seed dormancycan also be induced by burial (Wesson and Wareing, 1969; Weaverand Cavers, 1979; Chancellor, 1982). In rangelands, seeds getburied by grazing animals or deposition of eroded soil. Soilcompaction caused by trampling animals (Thomas, 1960) could forman anaerobic environment, which in turn could induce secondaryseed dormancy. Since secondary dormancy can be induced byenvironmental factors such as temperature, light, and oxygenunder laboratory conditions, it is a useful system to study theregulation of seed dormancy (Chancellor, 1982; Tilsner andUpadhyaya, 1985).Harper (1959) stressed the importance of dormancy to weedseed survival and recognized the problem that dormancy poses in8weed control. Conventional weed control practices, such as useof chemicals and burning, kill weed seedlings, but dormant seedspersist in the soil and maintain the weed population. Thedevelopment of integrated systems of weed control partly dependson the predictability of germination of buried seeds. Therefore,an understanding of factors affecting seed dormancy andgermination may help predict the potential of forming apersistent seed bank and develop effective control methods.A. External factors controlling seed germinationSeveral factors have been shown to affect weed seed dormancyand different species have been shown to differ in theirresponse to these factors. The following is a brief descriptionof how temperature, light, and depth of seed burial affect seedgermination. Little is known about how environmental factorsaffect seed germination in C. officinale and Tragopogon spp.This thesis will partly focus on this aspect.Temperature: Temperature often determines the time of weed seedgermination in nature and is particularly important in temperateclimates where it is the most variable environmental factor(Koller, 1964; Hegarty, 1973; Stoller and Wax, 1973; Bewley andBlack, 1985; Taylorson, 1987; Roberts, 1988; Hermanutz andWeaver, 1991). Many weed seeds with innate or secondary dormancybegin to lose the dormancy upon storage at high temperatures orduring moist-chilling (4 to 5 C) treatment (stratification)9(Roberts, 1988). This after-ripening process is very common inmany weeds (e.g., Polygonum pennvlvanicum and Chenopodium album)(Baskin and Baskin, 1987). Secondary dormant seeds of Avena fatua after-ripened rapidly after a period of dry storage(Tilsner and Upadhyaya, 1985). The requirements ofstratification by dormant seeds ensures that all seeds do notgerminate immediately before inclement weather. This encouragesseedling establishment in the spring when there is minimalcompetition with other plants (Grime et al., 1981;Froud-Williams et al., 1984).Light: Light plays an important role in the regulation of seedgermination in many weedy species (Steinbauer and Grigsby, 1957;Wesson and Wareing, 1969; Toole, 1973; Pons, 1991). A lightrequirement for seed germination may act as a predictor ofdisturbed areas suitable for weed colonization. Light-requiringseeds that fail to germinate while buried in the soil build aseed bank which helps in the persistence of the species (Egley,1986). Seed germination is also regulated by the quality oflight (Egley, 1986). Chlorophyllous tissue filters out much ofthe red (R) portion of the spectrum, which stimulatesgermination of light sensitive seeds, and allows the far-redlight (FR), which inhibits seed germination, to pass through(Cresswell and Grime, 1981; Egley and Duke, 1985; Taylorson,1987). Light-sensitive dormant seeds of knapweed (Centaureadiffusa L. and C. maculosa L.) germinated in response to red10light; light-insensitive dormant seeds failed to germinate evenafter a 5-day continuous exposure to red light (Nolan andUpadhyaya, 1988).Depth of seed burial: Depth of seed burial both delays anddecreases seed germination and seedling emergence percentage(Froud-Williams et al., 1984; Blackshaw, 1990). Soiltemperature, soil moisture content, and the composition of airin the soil can affect seed germination (Taylorson, 1972, 1987).Seedling emergence of Datura spp., for example, has been shownto decrease with increasing sowing depth; no emergence occurredfrom the depths below 10 cm (Reisman-Berman et al., 1991). Inrangeland situation where cultivation is not common, seedsgenerally stay at or near the soil surface (Roberts, 1981).These seeds can, however, get buried by the movement of grazinganimals and deposition of wind-eroded soil particles. McRill andSagar (1973) reported that some seeds become part of the soilseed bank by burial beneath litter and by the activity of soilanimals.B. Effects of seed coat on germinationThe seed coat represents the interface between the embryoand its environment (Ballard, 1973). Seed germination in C.officinale seems to be inhibited by the seed coat. The mechanismof seed dormancy in this weed has not been thoroughly studied.The seed coat or pericarp can inhibit seed germination by11various mechanisms such as disallowing gas exchange, wateruptake, light penetration, or escape of inhibitors from theembryo (Taylorson and Hendricks, 1977; Werker, 1980/81). Reliefof these restrictions may explain an increase in germinationfollowing seed coat removal.Restriction of water passage: Water is an essential factor forseed germination. However, even when climatic conditions arefavorable for germination, seeds of some species cannot imbibewater and remain dormant (Ballard, 1973). The seed coat in somespecies is resistant to abrasion, covered with a wax-like layer,and appears to be entirely impermeable to water (Ballard, 1973).Seed coat impermeability to water has been particularly reportedin families such as Leguminosae, Convolvulaceae, Geraniaceae,and Malvaceae (Ballard, 1973; Taylorson, 1987).Restriction of oxygen passage: Limitation of oxygen movementthrough the seed coat may reduce respiratory activity of theembryo (Taylorson, 1987). Reduction of seed germination due torestriction of oxygen uptake by the seed coat has been reportedin Xanthium spp., apple and sugar beet seeds (Ballard, 1973;Richard et al., 1989). Pesis and Ng (1986) attributed increasedgerminability of muskmelon (Cucumis melo) seed to an increase inavailable oxygen following removal of the seed coat. Decoatingalso increased CO2 evolution in muskmelon seeds. Reisman-Bermanet al. (1989) indicated that during soaking for a short period,12water is trapped in the space between the seed coat and nucellusof Datura ferox L. and D. stramonium L. and in the intracellularspaces of the spongy tissue within the hilum. This water 'plug'may limit the diffusion of gases to and from the embryo throughthe hilum, thus inhibiting seed germination.Edwards (1968) and Come and Tissaour (1973) suggested thatthe oxidation of phenolic substances in seed coats may restrictmovement of oxygen to the embryo and thus cause dormancy. Thishypothesis has been further supported by several studies. Forexample, dormancy of freshly harvested barley and oat caryopsesand sugarbeet seeds resulted mainly from the polyphenol oxidasemediated oxidation of phenolic compounds present in high amountsin the glumellae and seed coats (Coumans et al., 1976; Lenoir etal., 1986; Corbineau et al., 1986). Marbach and Mayer (1974,1975) reported that dormant Pisum seeds which had a high contentof phenolics and a high activity of polyphenol oxidase resultedfrom impermeable seed coats, but the non-dormant Pisum seedswhich lacked polyphenol oxidase had permeable seed coats.Germination inhibitors: Phenolic compounds of various kinds,present in seeds of a number of species, have been implicated inthe regulation of seed germination (Hamilton and Carpenter,1975, 1976; Naqvi and Hanson, 1982; Williams and Hoagland, 1982;Sreeramulu, 1983; Khan and Ungar, 1986; Mayer andPoljakoff-Mayber, 1989). Seeds of these species germinate whenthe phenolic inhibitors are metabolized or leached out of the13seed, or when the level of germination promoting substancesincreases during after-ripening or following exposure todormancy-breaking conditions (Flentje and Saksena, 1964;McDonough and Chadwick, 1970; Hamilton and Carpenter, 1975,1976; Enu-Kwesi and Dumbroff, 1980; Wright et al., 1980;Thapliyal and Nautiyal, 1989).Little is known how specific phenolic substance regulateseed dormancy. Qualitative and quantitative analyses of phenolicsubstances during seed germination should be investigated todetermine their involvement in seed germination regulation. Inaddition, the relationship between the change of phenolicsubstances and the effect of environmental factors such as lowtemperature over time should be monitored to understandecophysiology of seed germination regulation.2. Dynamics of seed populationsA. The seed bank and its implications on weedsThe seed bank is composed of dormant seeds present on thesoil surface or in the soil (Harper, 1977). Two major types ofseed banks are defined in the literature: the aerial seed bank(seeds attached to mother plants) and the soil seed bank (Sagerand Mortimer, 1976; Weaver and Cavers, 1979; Parker et al.,1989). Thompson and Grime (1979) classified soil seed banks into"transient" (persisting in the soil for less than one year) and"persistent" (longer than one year) seed banks.14Compared to plant population dynamics, seed bank dynamicshas been a much neglected topic in ecological studies (Cook,1980; Roberts, 1981; Cavers, 1983). Studies of viable andnon-viable, and dormant and non-dormant seeds in the soilrequires physical separation of seeds from the soil. This isoften laborious and extremely time-consuming. Nevertheless, theseed bank data are valuable to monitor the success of anylong-term weed management strategy. Determining the reservoir ofweed seeds in the soil, the rate of seedling recruitment, andseedling mortality will help the prediction of potential weedinfestations (Radosevich and Holt, 1984).Little information is available on longevity of rangelandweed seeds. Most studies on the longevity of buried seeds areconducted under cultivated conditions which involve regulardisturbance. The classical experiments of longevity tests forburied seeds were Beal's started in 1879 (Darlington, 1931) andDuvel's started in 1902 (Duvel, 1902). However, seeds in thoseexperiments were buried under artificial conditions of enclosurein sealed jars and in sand rather than in the natural soilsystem. Knowledge of seed transience or persistence, germinationcues, and environmental conditions suitable for establishment isof value for effective weed management (Froud-Williams et al.,1984; Parker et al., 1989).B. Population dynamics of buried seedsBecause habitats are spatially diverse, both horizontally15and vertically, the fate of seed banks can be greatly influencedby the diversed habitats (Parker et al., 1989). Burial inhibitsseed germination or induces secondary dormancy in many weedyspecies, and the seeds persist in the soil for periods rangingfrom a few months to several decades (Kivilaan and Bandurski,1973; Egley and Chandler, 1983; Roberts and Boddrell, 1984;Taylorson, 1987). Seed dormancy, induced by burial, allows weedto escape the effects of direct control measures (e.g.,herbicides, cultivation) and provides a mechanism for prolongedsurvival in the soil (Harper and White, 1970; Watkinson, 1978;Van Baalen, 1982). For some habitats, such as those where fireis an important factor, vertical distribution of seeds withinsoil is critical for survival.While burial inhibits seed germination in some species,little is known about how burial depth affects seed bankdynamics. For example, in Beal's experiment, seeds of Rumex crispus and Brassica nigra were buried at only one depth andsome seeds were viable after 50 years burial. Odum (1965)reported that the seeds of Chenopodium album and Spergulaarvensis could remain viable in the soil for 1600 years from oneburial depth. Over 50% of Centaurea maculosa L. seed populationremained viable after 5 years of burial in the rangeland (Davis,1990). Seed bank dynamics could be influenced by burial depthsince factors such as temperature, moisture, and light, requiredfor seed germination and seedling emergence vary between soildepths (Stoller and Wax, 1973).16C. Seasonal changes of germination and dormancyThe seasonal patterns of dormancy are of high survival valueto a species. Periodicity of seed germination at the soilsurface and in the soil is governed by changing environmentalfactors such as temperature and precipitation (Karssen, 1982).The extent of germination of a mature seed depends on thepresence of innate dormancy and whether the temperature,moisture and light requirement of germination are met (Roberts,1986; Parker et al., 1989). The presence of a persistent seedbank for many weed species has become a great barrier toeffective weed management.The fate of buried seeds depends on parameters such asgerminated, dormant and viable, and non-viable fractions(Sarukhan, 1974). The change among the relative abundance ofthese fractions within a seed population over time can helppredict the fate of the plant population. Germination timingtypically depends on seasonal variation in environmentalconditions that are relatively predictable, such as winter intemperate forests, or seasonal droughts in temperate grasslands(Karssen, 1982; Baskin and Baskin, 1983; Bouwmeester andKarssen, 1989; Venable, 1989; Baskin and Baskin, 1990). Dormancyof winter annuals is broken by high temperature, and of summerannuals by low temperature (Baskin and Baskin, 1985). Forexample, buried seeds of Sisymbrium officinale and Polygonumpersicaria lose primary dormancy during winter, and germinate in17early spring when conditions are conducive to seed germination(Bouwmeester and Karssen, 1989). These seasonal patterns ofgermination vary among species. The timing and extent of seedgermination within a year determines how many seeds will be inthe soil at different times of the year and how many seeds willbe carried over between years (Roberts and Boddrell, 1983;Venable, 1989).3. Dynamics of plant populationsA. Life historyLife history characteristics such as survivorship, growthrate, and reproductive output of a plant population arecomponents of its evolutionary fitness. Understanding of thecharacteristics of each growth stage in a plant life cycle canprovide insight into persistence strategy of weedy species. Manybiennials were considered as opportunistic species which exploitthe rare occasions suiting their special requirements forestablishment and growth; they were, therefore, characteristicof early successional habitats in which they remain for only asingle generation (Holt, 1972; Harper, 1977; Silvertown, 1983)."It is a curious feature of the dynamics of these populationsthat, when they are conspicuous enough to be chosen as an objectof study, they are already in a condition of decline (Harper,1977)." Harper (1977) also described biennials as a speciesgroup with good dispersal in time (large persistent seed bank)18and poor dispersal in space. However, some biennial speciesbehave in a considerably different manner from the generalpicture presented above. For example, Daucus carota andTragopogon dubius were able to establish in the presence oflitter and vegetation and can behave as short-lived perennials(Gross and Werner, 1982; Gross, 1984). Tragopogon pratensis alsois able to establish in the presence of other plants ingrassland. Whether T. pratensis is a biennial remains to beinvestigated.Age structure of a population has been used as a predictorfor the fate of a natural plant population (Pianka and Parker,1975). Harper and White (1974) and Werner (1975) argued,however, that age can be a poor predictor of vegetative growth,especially for some species for which reproductive output isdelayed. This is because the size or growth stage of a plant canvary independently of its chronological age. Growth stage playsan important role in plant demographic behaviour of some species(Werner, 1975). Although differential growth has been observedin natural plant populations, its demographic implications haveseldom been documented (Sarukhan et al., 1984).Four major factors can control the number of individuals inweed populations. These are: reproduction, mortality,immigration, and emigration (Law, 1981). Mortality andreproductive output can be interpreted as components of fitnessso that the strength of natural selection on life history can bemeasured (Venable, 1984). The rate of mortality among juvenile19plants is very high in some species, and the seedling shortlyafter germination is, therefore, the most susceptible phase inthe ontogeny of the individual (Cook, 1979). A significantnumber of demographic studies have been attempted to explain andpredict major patterns of life history of some species (Schaal,1984), but the factors limiting weed population dynamics aregenerally not well understood. The factors affecting T.pratensis population dynamics, therefore, need to beinvestigated before any control measure is implemented.B. Reproductive behaviour in relation to weed persistenceThe persistence strategies of some weedy species can bealtered by the delay of seed production under naturalconditions. For example, some biennial plants may flower intheir second growing season as strict biennials, or take longerin the field (Kelly, 1985; Silvertown, 1984). Whether theyflower in the second year or later depends on soil fertility andplant size (Boorman and Fuller, 1984; De Jong et al., 1989). Ifthe seed production is delayed, a large number of vegetativeplants may form a bud bank and can buffer the population againstlow seedling recruitment or high seedling mortality in someyears (De Jong and Klinkhamer, 1988).Growth rate has been shown to be positively correlated withthe diameter of the root crown in many species. The criticalsize thresholds at which plants switch to reproductivedevelopment have been described for many species such as Senecio 20vulgaris (Harper and Ogden, 1970), Dipsacus fullonum (Werner,1975), Digitalis purpurea (Oxley, 1977), Grindelia lanceolata (Baskin and Baskin, 1979), Senecio lacobaea (Van der Meijden andVan der Waals-Kooi, 1979), Verbacum thapsus (Gross, 1981), andOenothera erythrosepala (Kachi and Hirose, 1983), etc. Fastgrowing plants reach the critical size and flower the followingyear while slow growing plants persist as vegetative plantsuntil they reach a critical size regardless of the age (Gross,1981; Powell, 1988). Silvertown (1984) suggested that the phraseIsemelparous perennial' should be used to describe species onceclassified as biennials. However, Kelly (1985) has indicatedthat some species, like Linum catharticum and Gentianella amarella, are strict biennials (they take two years to completetheir life cycles) in natural populations. Therefore, in thisthesis term "biennial" still will be used to describe anyspecies that takes two years to finish its life cycle.Why some species require the critical size for flowering isnot known. The availability of nutrients may affect the onset offlowering (Harper and Ogden, 1970; Werner, 1975). Law (1979)suggested that when the supply of resources is restricted, anincreased emphasis is placed on non-reproductive activities.However, Kachi and Hirose (1983) argued that the amount of foodreserve did not contribute to the size-dependent flowering ofOenothera erytheosepala. Powell (1988) suggested thatsize-dependent switching maximizes the fitness of Centaurea diffusa due to 1) high juvenile mortality, 2) extensive21variation in individual growth rates, and 3) seasonal climaticchanges. Therefore, a long-term study is necessary in order todetermine the mechanism which regulates the plant reproduction.Many species require vernalization and/or a long photoperiodfor induction of bolting and flowering (Schwabe, 1954; Gross,1981; Kachi and Hirose, 1983). Rosettes which were smaller thanthe critical size (<9 cm) of Oenothera ervthrosepala did notrespond to inductive vernalization and photoperiod treatments(Kachi and Hirose, 1983). Baskin and Baskin (1979) suggestedthat Grindella lanceolata, a monocarpic perennial, may need toreach a critical size before responding to vernalization. In astudy involving four biennial species, Gross (1981) reportedthat few Daucus carota rosettes flowered in the greenhouse;Oenothera, Verbascum and Tragopogon dubius rosettes, which wereabove the critical size for flowering, did not bolt undergreenhouse conditions suggesting a vernalization requirement forbolting.4. Weed BiologyA. Cynoglossum officinale (hound's-tongue)Cynoglossum officinale L. (fam. Boraginaceae) is a biennialor short-lived perennial which reproduces only by seeds (Alexand Switzer, 1976; Alex et al., 1980) (Plate I). It formsrosettes in the first year and flower stems in the second year.Each flower produces four large bristly nutlets (Frankton and22Plate I. a) young Cynoglossum officinale plant, b) rosettes (inforeground) and dried plants that matured in previous years(in background), c) nutlets on mature plants, d) nutletsattached to cow's hair.23Mulligan, 1970). The seed production per plant varies from 50 to800 (Van Leeuwen and Van Breeman, 1980) to more than 2,000(Boorman and Fuller, 1984).Cynoglossum officinale occurs in Russia, throughout Europeand in the United States (see Upadhyaya et al., 1988). It wasintroduced to North America from Eurasia as a seed contaminantin cereals (Knight et al., 1984). In Canada, it appears to bemost abundant in southern British Columbia and Ontario(Upadhyaya, et al., 1988).In B.C., C. officinale occurs on disturbed sites of theInterior Douglas-fir and ponderosa pine-bunchgrassbiogeoclimatic zones. It is also associated with Eutric andDystric Brunisolic, Brown and Dark Brown Chernozemic andLuvisolic soils (Cranston and Pethybridge, 1986).The biology of C. officinale has been reviewed by Upadhyayaet al. (1988) and De Jong et al. (1990). Over 2,000 ha of forestrangeland, grassland and roadsides were infested by C.officinale in B.C. in 1986 (Cranston and Pethybrige, 1986). Themost detrimental aspect of this species is its poisonous nature(Greatorex, 1966; Mandryka, 1979; Bartik and Pistac, 1981;Knight et al. 1984). Plants in C. officinale containpyrrolizidine alkaloids (Knight et al., 1984) which arepotentially toxic to animals. Burred seeds disseminate byattaching to animal fur and reduce the quality of animal fur.Cynoglossum officinale can behave as a biennial or ashort-lived perennial. Under fertile soil conditions, plants24show 100% flowering in their second year (De Jong et al., 1987);under infertile soil conditions it can take several years beforeplants have reached a sufficient size for vernalization (Boormanand Fuller, 1984; De Jong et al., 1987). Plants have an equalchance of flowering at root crown diameter (RCD) of 0.8-0.9 cmcorresponding with a taproot dry weight of 0.8 g (De Jong etal., 1986; De Jong and Klinkhamer, 1989). Boorman and Fuller(1984) found that C. officinale was not strictly monocarpicbecause approximately 15% of the plants flowered again in theirthird or fourth year.Studies have suggested that the C. officinale seed bank inthe soil is short-lived because most seeds germinate in thespring (Roberts, 1986; Roberts and Boddrell, 1984; Van Leeuwenand Van Breeman, 1980; Boorman and Fuller, 1984). A persistentseed bank may, however, result from seeds over-wintering eitheron the plant or at the soil surface (Van Breeman and VanLeeuwen, 1983). Van Breeman (1984) suggested the burrs on C.officinale fruit coat keep its seeds on the soil surface.The mechanisms involved in regulation of seed germinationand seed dormancy in C. officinale are poorly understood.Roberts and Boddrell (1984) postulated a moist-chillingrequirement to break seed dormancy of C. officinale as seedremained dormant throughout the fall and winter but readilygerminated in the spring. Moist-chilling (6 to 12 weeks) at 0 to10 C has been shown to release seed dormancy in C. officinale (Van Breemen, 1984), but the total germination percentage was25low. Seed placement in the soil affected germination andemergence of C. officinale seeds (Van Breeman, 1984). At a depthof 1 cm, 48% of the seeds germinated; seeds buried at a 5-cmsoil depth germinated but did not emerge; seeds on the soilsurface desiccated and did not germinate.Newly ripened seeds of C. officinale have been reported toexhibit primary dormancy (Lhotska, 1982; Boorman and Fuller,1984; Roberts and Boddrell, 1984; Van Breeman, 1984), but littleis known about its regulation. The pericarps of these seeds havebeen implicated in the regulation of seed dormancy (Lhotska,1982; Dickerson and Fay, 1982). Under laboratory conditions,peak germination percentage was achieved when seeds with damagedtesta were placed on a germination bed with the ruptured sidefacing down - suggesting movement of germination inhibitors outof the seed (Lhotska, 1982). A detailed physiological study isnecessary to investigate the mechanism involved in regulation ofseed dormancy in C. officinale.B. Tragopogon pratensis (meadow goat's-beard) and T. dubius (goat's-beard)Tragopogon pratensis L. and T. dubius Scop. (fam.Compositae) have been reported as biennial species (Alex andSwitzer, 1976) (Plate II). Although very similar in appearance,Tragopogon pratensis and T. dubius can be distinguished by thefollowing characteristics. The leaves were curled at the apexand crisped at the margin in T. pratensis, the bracts of T.Tragopogondubius pratensib'Plate II. A) a flowering Tragopogon plant, B) rangeland infestedby Tragopogon spp., C) seeds, leaves, and flower heads of T.dubius and T. pratensis.27dubius are longer than its flowers, and the flower stalk belowthe flower head is thick and hollow in T. dubius (Plate II)(Frankton and Mulligan, 1970; Alex and Switzer, 1976). Seed ofT. dubius are generally larger and heavier than seeds in T.pratensis.Little information is available about the biologicalcharacteristics of these two weeds. They flower between late Mayand September in U.K., but flowering usually occurs in June andJuly in Canada (Frankton and Mulligan, 1970). Numerous seeds areproduced from each flower head (Alex and Switzer, 1976). Theseeds which have long, slender beaks and an umbrella-like pappusare dispersed by the wind. Tragopogon pratensis produces anextensive root system which competes for water and nutrientswith grasses (personal communication, Don Blumenauer, WilliamsLake, B.C., 1988). Interference by T. dubius has been shown toreduce the leaf area and the shoot/root ratio of Agropyronspicatum, an important component of B.C. grasslands (Mcllvride,unpublished result). Roberts (1986) reported that in GreatBritain an average of 11.8% of seeds of T. pratensis sown inJuly produced seedlings in August, but the main emergenceoccurred the following February and March; very few seedlingsemerged after this. Cresswell and Grime (1981) reported thatgermination of T. pratensis was strongly inhibited in the dark;when mixed into the soil, dormancy was enforced. Emergence of T.dubius seedlings was not decreased significantly in the presenceof vegetation (Gross, 1984).28Chapter 2. Mechanisms of seed dormancyin Cynoglossum officinale L.AbstractThe roles of the seed coat and methanol-extractable phenolicsubstances, which have been shown to regulate seed dormancy inother species, in the regulation of C. officinale seed dormancywere studied. 02 uptake of seeds increased approximatelysix-fold upon removal of the seed coat. The increase in 0 2uptake induced by seed coat removal was due to both an increasein seed respiration (measured by CO2 evolution) and a high levelof non-respiratory 02 consumption. The seed coat also preventedthe leaching of phenolic substances. This leaching, however, wasnot necessary for germination. Seeds of C. officinale containedhigh levels of phenolic substances and seed extracts showed highpolyphenol oxidase activity. Rosmarinic acid, the most prominentphenolic of C. officinale seeds, did not inhibit the germinationof the decoated seeds at concentrations up to 3 mM. Analysis ofthe pattern of methanol-extractable phenolic substances,however, showed no significant quantitative or qualitativecorrelation between changes in specific phenolic compounds andseed germination induced by stratification or seed coat removal.It is concluded from this study that C. officinale seed coatsinhibit seed germination by controlling 02 availability to theembryo.29IntroductionSeeds of Cynoglossum officinale L. (hound's-tongue), anoxious rangeland weed (Upadhyaya et al., 1988), show a deep,innate dormancy. This dormancy apparently contributes topersistence of this species by distributing seed germinationover time, which in turn results in the formation of seed banksin the soil. The physiological basis of seed dormancy in C.officinale is not known.In other species, the seed coat or pericarp can inhibit seedgermination by various mechanisms such as preventing gasexchange, water uptake, light penetration, or escape ofinhibitors from the embryo (Taylorson and Hendricks, 1977;Werker, 1980/81). Seed coats may also mechanically restrainembryo growth or contain germination inhibitors. Free orconjugated phenolic substances, which occur in seeds of a numberof species, have been implicated in the regulation of seedgermination (Hamilton and Carpenter, 1975, 1976; Naqvi andHanson, 1982; Williams and Hoagland, 1982; Sreeramulu, 1983;Khan and Ungar, 1986). Seeds of these species germinate when thephenolic inhibitors are metabolized or leached out of the seed,or when the level of germination promoting substances increasesduring after-ripening or following exposure to dormancy breakingconditions (Flentje and Saksena, 1964; McDonough and Chadwick,1970; Hamilton and Carpenter, 1975, 1976; Enu-Kwesi andDumbroff, 1980).30The preliminary studies showed that C. officinale seedscontain a high level of phenolic substances in the embryo andthat germination occurred when the seed coats were damaged orremoved. The overall objective of this study was to investigatethe roles of seed coat and methanol-extractable phenolicsubstances present in the seeds in the regulation of C.officinale seed dormancy. The specific objectives were todetermine: 1) effect of C. officinale seed coat on 0 2 uptake; 2)if seed coats contain any water-soluble germination inhibitors;3) if seed coats prevent leaching of inhibitors from the embryo;and 4) effects of methanol-extractable phenolic substancespresent in the seeds on seed germination in C. officinale.Materials and methodsA. Seed sourceFruits (nutlets) of C. officinale were collected in Octoberof 1987, 1988, and 1989 from Kamloops, B.C. and stored at -20 C.Prior to use, they were imbibed for 10 min in water, and seedswere separated from the nutlet coverings using a scalpel.Decoated seeds were prepared by carefully removing the brittle,black seed coat.B. Germination studiesSeeds were incubated in darkness in 9-cm Petri dishes at25 C on two Whatman No.1 filter disks wetted with 5 ml distilled31water. Seed germination was assayed by monitoring protrusion ofthe radicle. Seeds with 5 mm or more radicle protrusion wereconsidered to be germinated. Fruits and seeds (fruit coatremoved) of C. officinale were treated with GA 3 (0.5 mM) and/orpin-pricked and the effect on germination was studied at 15 and25 C. Effects of coumarin, R-coumaric acid, gallic acid (Sigma,U.S.A.) on germination of decoated C. officinale seeds werestudied by incubating seeds in 1 to 7 mM solutions of thesesubstances. In a separate experiment, the effects of coumarinand rosmarinic acid (RA), purified from Anchusa officinalis cellculture, on germination of decoated C. officinale seeds werestudied by incubating seeds in 1 to 3 mM solutions of thesesubstances.C. Permeability of seed coat to waterWater uptake by intact and decoated seeds was compared bymonitoring the increase in seed weight during imbibition at25 C. Seeds were incubated in Petri dishes as described aboveand weighed periodically.D. Respiration measurements02 uptake by decoated and intact seeds (imbibed for 0.5 to24 hours at 25 C) was measured at 25 C using a Clark-type oxygenelectrode and YSI Mode 30 oxygen meter. CO2 evolution bydecoated and intact seeds (imbibed for 0.5 to 24 hours at 25 C)was measured using a Li-CoR LI-6200 Primer Portable32Photosynthesis System. In a separate study, 0 2 uptake bypin-pricked and intact seeds (imbibed up to 7 days) was measuredat 25 C and germination of these seeds also was monitored.E. Enzyme extraction and assayDecoated seeds were imbibed in distilled water for 4 hoursand homogenized in 7 ml of 0.01 M sodium phosphate buffer (pH6.8) at 4 C. The homogenate was centrifuged at 15,000 g for 20min and polyphenol oxidase (PPO) activity in the supernatant wasassayed using an 0 2 electrode. Gallic acid, caffeic acid, andchlorogenic acid (4.2 mM) were used as substrates. Lipoxygenaseactivity was also assayed using an 02 electrode and linoleicacid as the substrate.F. Leaching of Folin reagent-active substancesTen seeds (with or without seed coat) or seed coatsseparated from ten seeds were incubated in 2 ml water in 21-mlvials (imbibed for 0.5, 1, 2, 4, 6, 8, and 24 hoursrespectively, three replicates per treatment) and release ofFolin reagent-active substances in the incubation medium wasmonitored for up to 24 hours.G. Extraction and assay of phenolic substancesSeeds that had been stratified (at 5 C for 0 to 7 weeks), ordecoated seeds that had been imbibed (at 25 C for 0 to 48hours), were homogenized in 1 ml of 50% methanol using a Caframo33(Canlab, Canada) homogenizer. The homogenate was boiled for 15min, centrifuged at 18,000 g for 10 minute, and the supernatantwas removed. The pellet was extracted again as above (10 minboiling) and the two supernatants were pooled. Total solublephenolics were measured using Folin reagent (Ferraris et al.,1987), and chlorogenic acid as a standard. Extracts werefractioned by reversed phase gradient HPLC on an AlltechHypersil MOS (5 um) column using 2% aqueous acetic acid assolvents A and 2% acetic acid in acetonitrile as solvents B(gradient conditions: 0 to 20% B over 10 min, 20% B for 10 min,20 to 50% B over 20 min, 50 to 100% B over 5 min, 100% B for 5min, 100 to 0% B over 1 min). The eluent stream (1 ml/min) wasmonitored by UV absorption at 320 nm to measure total solublephenolic substances.ResultsA. Effect of seed coat on seed germinationRemoval of the seed coat resulted in 100% seed germinationwithin 48 hours of imbibition, whereas seeds with intact seedcoats showed less than 10% germination after 20 days ofincubation (Fig. 1A). No additional germination occurred whenthe seeds were incubated for up to two months (data not shown).This was not due to restrictions to imbibition, because completeremoval of the seed coat or "pin-pricking" of the coat did notsignificantly improve water uptake (P > 0.05) (Fig. 1B).0 4^9  ^ElA-e- + seed coat-e- - seed coat8^12Time (d)0^016^204 8 12^16 20 24-e- + c oa t -pricking-9- +coat +pricking-rAs" -coat -pricking60504030201034Time (h)Fig. 1. A. Effect of removal of seed coat on germination ofCynoqlossum officinale seeds. B. Effect of seed coat on wateruptake by seeds. Values are the means of four replicates of10 seeds each.35Washing of seeds with intact seed coats in 100-m1 distilledwater on a rotary shaker for 24 hours prior to the germinationassay did not significantly improve (P > 0.05) seed germination(at four weeks) compared to the "unwashed" control (Fig. 2)."Pricking" treatment with or without "washing" improved seedgermination significantly (P < 0.05). However, the germinationof pricked seeds was much slower than the decoated seeds(compare Fig. lA with Fig. 2).GA3 and pricking treatments stimulated germination of C.officinale seeds; an interaction of the two treatments, GA 3 andpricking, was most significant (P < 0.05) (Table 1). Thesetreatments, when given to intact fruits were, however, ineffec-tive at stimulating seed germination.B. Effect of seed coat on oxygen uptakeRemoval of the seed coat significantly increased seed 0 2uptake (Fig. 3a) as well as CO2 release (Fig. 3b) compared tothat of intact seeds (P < 0.05). However, the rate of 0 2 uptakeby the decoated seeds at 0.5 hour increased by 602% whereas thatof CO2 release increased by only 71%. At 24 hours of incubation,the rates of 02 uptake and of CO 2 release increased by 606% and130% over the values for seeds with intact coats, respectively(Fig. 3). 02 consumption by the pricked seeds increased withimbibition time, but was not significantly different from theunpricked seeds during the initial 72 hours imbibition (Fig. 4);then oxygen uptake and germination of the pricked seeds80( 060CC50F0 40CD2028244^8^12^16^20-8-- +washing +pricking,L +washing -pricking-0- -washing +prickingE -washing -pricking10036Time (d)Fig. 2. Effect of washing to remove inhibitors on seedgermination. Values are the means of eight replicates of 10seeds each.37Table 1. Effects of pricking (P) and GA 3 (GA) treatments onCynoglossum officinale seed germination at four weeks.Temp.(C)Treat-mentPergent germinationSeed^Fruits15 C +GA+P 100b 13 + 6.4+GA-P 45 + 9.9 0-GA+P 58 + 9.5 6 + 3.2-GA-P 11 + 2.1 025 C +GA+P 100 12 + 2.5+GA-P 96 + 1.8 0-GA+P 72 + 9.2 5 + 2.8-GA-P 37 + 12 0a Either fruits or seeds (fruit coat removed) were treated.b Mean (+SE) of eight replicates of 10 seeds each.00.5 -1 2 8 2425205BLS D0.053840_.0 300).__0 20EFN0 1000.5^1^2^8^24Imbibition time (h)Fig. 3. Effect of seed coat on: A, 0 uptake and B, CO, release.Seeds were incubated and 0 uptake and CO„ release`werelmonitored for 0 to 24 hr. Values are the eans of threereplicates of 1 g (±27 mg) seeds each.1002039—0-- -pricking (02 uptake), pricking(02 uptake)_^-pricking(germination)-El- *pricking( germination),-150(13C1 10C0C)X 5C)06^24^72^120^168Time of Imbibition (h)60 0CC40 EN20Fig. 4. Effect of pin-pricking on oxygen uptake and germinationof Cynoglossum officinale seeds. Values are the means ofthree replicates of 10 seeds each.40increased significantly with imbibition time.C. Effect of phenolic substances on seed germinationExtracts prepared from decoated seeds showed high levels ofPPO activity using chlorogenic acid and caffeic acid assubstrates (Fig. 5); no PPO activity was observed when gallicacid was used as a substrate. No lipoxygenase activity wasdetected in the extracts of C. officinale seeds using linoleicacid as the substrate (data not shown).The phenolic substances coumarin and p-coumaric acidinhibited germination of decoated seeds at concentrations up to3 mM, but gallic acid did not (Fig. 6). However, the rosmarinicacid (RA), the phenolic substance detected in C. officinale seedextracts, purified from Anchusa officinalis, did not inhibitgermination of C. officinale seeds at concentrations up to 3 mM(Table 2).Decoated seeds released significant amounts of Folinreagent-active substances into the incubation medium (Fig. 7A).The amount of active substances accumulated in the incubationmedium increased linearly up to eight hours of incubation andthen began to level off (Fig. 7A). Decoated seeds released asignificantly higher amount of Folin reagent-active substancesthan did the seeds with coats (black seeds) (Fig. 7B). Leachateobtained from seeds with intact coats was quantitatively similarto that obtained from the seed coat alone (Fig. 7B). The AriA'01-e- gallic acid-8- chlorogenic acidA caffeic acidah^■■^AI^AL^IM^ALN.^Mr LI11.1. r410.6OF 0.50 0,40_F0.3ci00c 0.2EDOD0 012^4^6^8^10Time (min)Fig. 5. Polyphenol oxidase activity in Cynocilossum officinale seed extracts in the presence of 4.2 mM gallic acid,chlorogenic acid, and caffeic acid. Values are the means oftwo replicates of 20 (.-0.3g) seeds each.LSD0.05- -19- coumarin--A- p-coumaric a id--e- gallic acid1 42805 50(0EE 400)CD202^4^6^8Concentration (mV)Fig. 6. Germination inhibition of decoated seeds by phenoliccompounds coumarin, p-coumaric acid, and gallic acid. Valuesare the means of three replicates of 10 seeds each.43Table 2. Effect of rosmarinic acid (RA) and coumarin onCynoglossum officinale seed germination.Treatment (mM)Time (hour)48 56 144Control 53.3 + 2.7 93.3 + 2.7 100RA 1 50.0 + 4.8 93.3 + 2.7 1002 46.7 + 2.7 83.3 + 2.7 96.7 + 2.73 50.0 + 9.4 96.7 + 2.7 100Coumarin 1 0 56.7 + 7.2 80.0 + 4.72 0 0 3.3 + 2.73 0 0 0Decoated seeds were incubated at 25 C in RA and coumarinsolutions in Petri dishes lined with two Whatman No. 1 filterdiscs. Values are means of germination percentages (±SE) ofthree replicates of ten seeds each.25205I0--)Te15mbibition ti10442.52Loc\i1.5<10.50Fig. 7. A. Time course of leaching of Folin reagent-activesubstances by decoated Cynocrlossum officinale seeds. B.Effect of seed coat on leaching of Folin reagent-activesubstances at 2 and 6 hr of incubation. Values are the meansof three replicates of 10 seeds each.45leachate tested positively both for protein(s) as measured byBradford's reagent (Bradford, 1976) and for phenolic substancesthat were not identified.There was no significant change (P > 0.05) in the level ofphenolic substances prior to or during protrusion of the radiclein decoated seeds (Fig. 8A). RP-HPLC profiles of methanolextracts of decoated seeds incubated for 0 and 48 hours (Fig.8B,C, respectively) showed no significant quantitative orqualitative differences in the chromatographic profiles; similarresults were obtained for seeds incubated for 4, 8, 12, 24, and36 hours (profiles not shown). Five peaks were observed in theRP-HPLC profiles of the methanol extracts (Fig. 8B,C). The mostprominent peak eluted at the position corresponding torosmarinic acid (RA), a widespread phenolic component ofLamiaceae and Boraginaceae species. No attempt was made toidentify the other phenolic compounds, as their profiles did notchange in a fashion that could be correlated with seedgermination.Stratification also was able to break seed dormancy in C.officinale, but again there was no significant change (P > 0.05)in the level of phenolic substances prior to or duringstratification-induced radicle protrusion (Fig. 9A). RP-HPLCprofiles of methanol extracts of seeds stratified for 0 and 7weeks showed no significant quantitative or qualitativedifferences (Fig. 9B,C, respectively); similar results wereobtained at 1, 2, 3, 4, 5, and 6 weeks of incubation.800-,--,u ^60_Ecu 40ED20————Ow0 A-4.--,— 1 =(1.)-4---,0U—0.5 C.)=' 0CD0 a_4880 —RA 0 h100B0 60cN<E)40 —20 —10^2030^40^50Time (min)80 48 h100CRA60402000^10^20^3Q^, 40^50Time (min)c \ vq----146111- germination (%).mg phenolics seed-1i^i^i 4 8 12^24^36hnni bibition time (h)120100Fig. 8. A. Time-course of germination of decoated seeds and levelof phenolic compounds. Values are the means of threereplicates of 10 seeds each. RP-HPLC elusion profiles ofmethanol extracts of decoated seeds imbibed for 0 (B) and 48(C) hours.2.521.510.504735302520151050I01. mg phenolics seedgermination (%)AAA^A^A^A^A1 2 3 4 5Duration of stratification100^BRA7^8weeks)80^7 weeksRA60400 week10080060402000^10^20^30^40^50Time (min)10^20^30^40^50Time (min)20Fig. 9. A. Effect of stratification on seed germination and thelevel of phenolic compounds. Values are the means of threereplicates of 10 seeds each. RP-HPLC elusion profiles ofmethanol extracts of decoated seeds stratified for 0 (B) and7 (C) weeks.48DiscussionA. Role of seed coat on seed germination and 0 2 uptakeSince C. officinale embryos germinate readily following seedcoat removal, seed dormancy in this species cannot be due to aninherent embryo dormancy. Low permeability of the seed coat towater and gases has been shown to inhibit seed germination inother species (Taylorson and Hendricks, 1977; Werker, 1980/81).The seed coat in musk melon (Cucumis melo), for example,inhibits seed germination by preventing 02 uptake (Pesis and Ng,1986). An increase in seed coat permeability for 02 maystimulate seed germination by increasing the rate of respirationand/or by inactivating endogenous germination inhibitors throughoxidation reaction(s). The results in this study show that seedcoats of C. officinale do not prevent water uptake by theembryo, but the intact seeds had a very low rate of 0 2 uptake.The rate of 02 uptake, however, increased by approximatelysix-fold following seed coat removal. This suggests that lowpermeability of the seed coat for 0 2 may be responsible for seeddormancy in C. officinale. In nature, these seeds may germinateafter their coats are physically altered due to alternatewetting and drying, freezing and thawing, temperaturefluctuations, or microbial degradation.49B. Seed coat removal and phenolic oxidationAlthough removal of the seed coat increased the rate of 0 2uptake six-fold, the rate of CO 2 release increased by only 71%.This suggests that some of the increase in 0 2 uptake followingdecoating is not due to seed respiration. Oxidation of phenolicsubstances present in the seed may explain the higher rate of 02consumption. Extensive phenolic oxidation following decoating isalso suggested by the observation that some decoated seedsshowed some browning upon imbibition, and released brownpigments into their surroundings. The increase of 0 2 uptake inthe pricked seeds after 72 hours imbibition indicates thatpin-pricking may slowly increase the rate of 0 2 uptake duringthe imbibition and lead to seed germination.C. Effect of seed coat on leakage of germination inhibitorsThe imbibition of seeds in 100 ml water on a rotary shakerfor 24 hours ("washing" treatment) did not improve C. officinaleseed germination. This suggests that the seed coats of thisspecies do not contain any water soluble germination inhibitor.When seed coats were "pin-pricked" using a fine pin andincubated with the pricked side away from the wet filter paper(to disallow leaching of any germination inhibitor), the seedsshowed a high level of germination. This argues against thepresence of either water soluble or insoluble germinationinhibitors in the seed coat.The testa can prevent leakage of intracellular contents from50the embryo and may protect embryo cells from rupturing (Larson,1968; Powell and Matthews, 1978; Simon, 1978; Duke and Kakefuda,1981). Physical damage to the testa has been reported toincrease the leaching of a variety of substances from the embryo(Flentje and Saksena, 1964; Scroth and Cook, 1964; McDonough andChadwick, 1970; Abdel-Samad and Pearce, 1978). These substancesmay include germination inhibitors, which when leached allowgermination. The presence of brown circles around decoated C.officinale seeds on moist filter papers, as described above,suggests that phenolic substances present are released followingdecoating. Decoated seeds incubated in one milliliter of waterwere found to release substances that reacted with Folinreagent; the presence of a seed coat, which prevented seedgermination, also prevented the leaching of these substances.However, this leaching was not necessary for seed germinationbecause decoated seeds incubated in humid air, having no contactwith free water, also germinated readily (data not shown). Noattempt was made, therefore, to identify specific substances inthe leachate. Preliminary analysis using RP-HPLC, Sephadex G-25chromatography, as well as Bradford and Folin reagents(Bradford, 1976; Ferraris et al., 1987), however, detected thepresence of protein(s) and phenolic compounds in the leachate(data not shown).D. Effect of phenolic substances on seed germinationPhenolic inhibitors have been implicated in the regulation51of seed germination in a number of species (Hamilton andCarpenter, 1976; Simon, 1978; Enu-Kwesi and Dumbroff, 1980;Naqvi and Hanson, 1982; Williams and Hoagland, 1982; Kim andKil, 1989). Coumarin and p-coumaric acid were found to inhibitC. officinale seed germination, whereas, gallic acid had littleor no effect. The lower activity of gallic acid has also beenreported in other species (Khan and Ungar, 1986). Enu-Kwesi andDumbroff (1980) reported a rapid decline in p-coumaric acidlevels preceding stratification-induced germination in Acersaccharum seeds. The rosmarinic acid (RA), which was the mostprominent phenolic in RP-HPLC profile of methanol extract of C.officinale seeds, however, did not inhibit the germination ofdecoated seeds at concentrations up to 3 mM (Table 2). Thisindicated that at least this phenolic substance is not involvedin the regulation of seed germination in C. officinale.In order to determine if the level of methanol-extractablephenolic substances in C. officinale seeds was related to theirgerminability, the levels of methanol-extractable phenolicsprior to and during protrusion of the radicle (germination)induced by decoating or stratification were studied. Sincedecoated C. officinale seeds germinate within 48 hours, theyprovide an excellent system to study changes in the level ofendogenous phenolic substances in relation to seed germination.The results of this study showed no significant change in thelevel of methanol-extractable phenolic substances prior togermination induced by either decoating or by stratification.52RP-HPLC elution profiles of seed extracts showed no changebetween time zero and the time when the seeds germinated. Thissuggests that this pool of substances is not likely to beinvolved in the regulation of seed germination in C. officinale.Gibberellic acid and kinetin have also been reported toreverse the germination-inhibiting effect of highly active,exogenously applied phenols in Atriplex triangularis (Khan andUngar, 1986). GA 3 also stimulates the germination of "pricked"C. officinale seeds in this study (Table 1). It is, therefore,possible that the stratification treatment may stimulate C.officinale seed germination not by causing a decline in thelevel of phenolic substances but by increasing the level ofendogenous germination promoter(s) such as gibberellins, as hasbeen reported in other species (Frankland and Wareing, 1962;Webb et al., 1973; Hamilton and Carpenter, 1975, 1976).Water-soluble phenolic substances leached from plants intothe rhizosphere have been implicated in allelopathicinteractions between plants (Einhellig and Rasmussen, 1979; Kimand Kil, 1989). These substances exert their influence byaffecting a number of plant processes (Demos et al., 1975;Patterson, 1981), and they can also protect seeds from insectpests and pathogens (Webb and Agnihotri, 1970). Oxidation andpolymerization of phenolics may also provide a sealing mechanismto repair any physical damage of the seed coat. Such damagecould increase 02 permeability thereby causing untimelygermination of seeds in nature. Whether the high levels of53phenolics present in C. officinale seeds play any such rolesremains to be determined.In conclusion, a systematic investigation of the role ofseed coat in regulation of C. officinale seed germination showsthat seed coat does not interfere with water uptake but it doescontrol 02 availability to the embryo, which in turn, mayinhibit seed germination. The seed coat also prevents theleaching of phenolic substances. This leaching, however, is notrequired for germination. Seeds of C. officinale contain highlevels of phenolic substances and seed extracts showed highpolyphenol oxidase activity. While C. officinale seeds contain ahigh level of phenolic compounds, no relationship between thesecompounds and germination could be established.54Chapter 3. Seed germination ecophysiologyof Tragopogon pratensis L. and T. dubius Scop.AbstractTo understand persistence strategies of Tragopogon pratensisL. and T. dubius Scop., ecophysiological characteristics oftheir seed germination were studied. Anaerobiosis (immersion indeoxygenated water) induced secondary dormancy in seeds of bothspecies. Dormancy could be induced in 86% of T. pratensis seedsand in 65% of T. dubius seeds by a one-day anaerobiosistreatment. Induced dormancy was gradually released during thestorage of air dried secondary dormant seeds and rate of thisrelease was significantly influenced by the storage temperature;30 C was more effective than 10 and 20 C in releasing secondarydormancy. These results suggest that the two species, whichexhibit only a short term innate dormancy, may rely on theinduced dormancy as an option in their persistence strategy.The optimum temperature for germination of non-dormant seedsof both species was 15 C. Maximum germination percentages forboth species were established within 4 to 6 days of incubationat 15 C and within 14 to 28 days of incubation at 25 C.Non-dormant seeds did not germinate below 10 C or above 30 C.Stratification (at 5 C for 2 to 10 weeks) significantlystimulated germination of secondary dormant T. pratensis seeds.This stratification requirement is expected to prevent55germination of dormant Tragopogon seeds in the fall therebyavoiding high seedling mortality in the winter. Light (red andfar-red) had no effect on the germination of secondary dormantseeds.Seeds planted in pots at depths of 2 to 14 cm had maximumseedling emergence from seeds buried 2 cm deep. Seeds planted at8 cm or deeper germinated but did not emerge. These resultssuggest that Tragopogon spp. seed reservoirs will becomecompletely depleted under conditions conducive for germinationregardless of the depth of seeds in the soil profile.IntroductionPersistence of many weed species depends on their ability tomaintain aerial (seeds attached to mother plants) and/or soilseed banks (Roberts, 1981; Simpson et al., 1989). The dynamicsof these banks is affected by environmental factors which varyin space and time (Karssen, 1982; Bouwmeester and Karssen, 1989;Parker et al., 1989). Genetic variability in the duration ofseed dormancy and the influence of the environment (on theexpression of this trait) distribute seed germination over time,which in turn leads to the development and maintenance of seedbanks. Knowledge of the ecophysiology of seed germination,genetic variability, and environmental influences is, therefore,valuable for understanding the persistence strategies of weedspecies that reproduce by seeds. In particular, we need to know56dormancy and germination characteristics of a given species andhow factors that vary seasonally as well as in space in soilprofile, viz. temperature, moisture, and oxygen, influence thegermination characteristics of a given species.Tragopogon pratensis L. (meadow goat's-beard) and T. dubius Scop. (goat's-beard) are rapidly becoming dominant rangelandweeds in British Columbia. Their success may be attributed totheir large seed production, effective seed dissemination bywind, and their ability to distribute germination over time.While both weeds reproduce only by seeds, little information isavailable on their dormancy characteristics and ecophysiology ofgermination regulation. Seeds buried at different soil depthsexperience different environments and are expected to differ intheir germination behaviour (Wesson and Wareing, 1969;Froud-Williams et al., 1984). The depths from which the seeds ofthe two Tragopogon spp. can emerge, and whether the non-dormantseeds buried deep in soil become dormant are not known. Theobjectives of this study were to determine: 1) if anaerobiosiscan induce secondary dormancy in T. pratensis and T. dubius andif the two species differ in this regard, 2) the influence oftemperature and light on seed dormancy and seed germination, 3)the stratification requirement for germination of secondarydormant seeds, and 4) the influence of seeding depth on seedgermination and seedling emergence. These characteristicsdetermine the ecophysiological bases of seed bank dynamics ofTragopogon species.57Materials and MethodsA. Seed sourcesTragopogon pratensis seeds were collected from Riske Creek,B.C. in summer of 1986 and 1988, and T. dubius seeds from Canoein 1986 and Williams Lake in 1988, B.C. The seeds were stored at-20 C before use.B. Germination studySeeds of T. pratensis and T. dubius were treated with thefungicide Captan (N-[(trichloromethyl)thio]-4-cyclohexene-1,2-dicarboximide), 50% wettable powder, Later Chemicals Ltd.,Richmond, B.C.) and incubated in 9 cm Petri dishes lined withtwo Whatman No. 1 filter discs wetted with 5 ml of distilledwater. Unless stated otherwise, each treatment consisted of fourreplicates of 10 seeds each. Petri dishes were sealed in plasticcontainers lined with two paper towels. To exclude light,containers were covered with opaque plastic bags with analuminum foil lining in between. Germination was monitoredperiodically and, unless stated otherwise, germination values atfour weeks of incubation are given in the Results.C. Induction of secondary dormancyFreshly harvested seeds of T. pratensis and T. dubius showedonly a short-term primary dormancy (Chapter 4). Non-dormantseeds of T. pratensis and T. dubius were obtained by58after-ripening seeds at room temperature (-25 C). Secondarydormancy was induced by immersing non-dormant seeds of T.pratensis and T. dubius in deoxygenated water (approximately 50uM dissolved 0 2 ) in sealed Erlenmeyer flasks (125 ml) at 25 C(in dark) for 1 to 4 days. Deaeration was accomplished byboiling distilled water and allowing it to cool at 25 C in anair-sealed system, permitting intake of boiled water to fill thespace produced on cooling. Induction of secondary dormancy wasstudied by removing seeds from the flasks and assaying theirgermination in Petri dishes (at 20 C in darkness) under aerobicconditions. Seed viability after each treatment was determinedby tetrazolium staining (Moore, 1972). To study the effect oftemperature on induction of secondary dormancy, T. pratensis seeds were immersed in deoxygenated water at differenttemperatures (15, 20, 25, 30 C) for 4 to 24 hours, and thentransferred into Petri dishes for the germination assay asdescribed above. Interaction between temperature and duration ofanaerobiosis on the induction of secondary dormancy was studied.D. Effect of temperature on seed germinationSeeds of T. pratensis and T. dubius were incubated at 5, 10,15, 20, 25, 30, 35, or 40 C for four weeks in darkness, andtheir germination was monitored.Secondary dormant seeds of T. pratensis were imbibed totheir maximum capacity, incubated at 4 to 5 C in darkness for 2to 10 weeks (stratification), and then incubated at 10, 20, and5930 C for the germination assay.To study the effect of temperature on after-ripening, airdried secondary dormant (dormancy induced by 24 hoursanaerobiosis) T. pratensis seeds were incubated at 10, 20 and 30C. After 1 to 12 months of dry storage, seeds were imbibed at 25C in Petri dishes in darkness for the germination assay.E. Effect of light on seed germinationRed (R) and far-red (FR) lights were obtained as describedby Nolan and Upadhyaya (1988). Red light was obtained byfiltering light from five cool-white fluorescent tubes (40W,General Electric, F4OCW) through a 3-mm-thick red filter (Rohmand Hass Plexiglas No. 2423). Far-red light was produced byfiltering light from eight incandescent bulbs (100W,Westinghouse) through single layers of red Plexiglas andRoscolux No. 95 medium blue green plastic filters. Both R and FRsources were determined using an International Light IL700radiometer and IL785A photomultiplier. Secondary dormancy innon-dormant seeds of T. pratensis was induced by a 24 hoursanaerobiosis. Secondary dormant seeds, which had been imbibed indarkness at 25 C for 8 hours, were exposed to R or FR light for5 min and their germination was assayed at 25 C in darkness. A5-min exposure to red or far-red light has been shown to beeffective on seeds of other members of Compositae, e.g. lettuceand knapweed (Centaurea diffusa L.) (Bewley and Black, 1985;Nolan, 1989). Since the response of primary dormant lettuce60seeds to red and far-red lights has been characterized (Bewleyand Black, 1985), these seeds were also used only to check theeffectiveness of red and far-red lights in this study.F. Effect of seeding depth on seed germination and seedlingemergenceSeeds of T. pratensis and T. dubius (10 seeds in each 6"pot) were planted at depths of 2, 5, 8, 11, and 14 cm (sixreplicates each) in sterilized sandy loam greenhouse soil mix(74.9% sand, 16.1% silt, 9% clay, and 11.4% organic matter;Cation Exchange Capacity 36.6 meg/100g, electrical conductivity2.1 mmhos/cm) in pots under greenhouse conditions. Seedlingemergence was periodically recorded. At the end of theexperiments, six pots in each species were randomly selected todetermine the fate of the seeds that failed to send shoots abovethe soil surface.G. Statistical analysisA randomized complete block design with a factorialarrangement, was used in all experiments. All experiments wererepeated at one or more times with similar results. Germinationpercentages were arcsin frCtransformed and subjected to analysesof variance (ANOVA) by the MGLH procedure of SYSTAT (Wilkinson,1990a). Results of stratification experiment were subjected toResponse Surface Analysis which allows a visual interpretationof the response function and interaction between independent61variables. The response surface graph was generated using theSYGRAPH package (Wilkinson, 1990b). For some experiments, meanswere separated using Fisher's Protected LSD test (P < 0.05).ResultsA. Induction of secondary dormancyAnaerobiosis induced secondary dormancy in both T. pratensis and T. dubius seeds (Table 1). Dormancy could be induced in 86%(determined based on the control) of T. pratensis seeds and in65% (determined based on the control) of T. dubius seeds by aone-day anaerobiosis treatment. Seeds of T. pratensis were moreresponsive to anaerobiosis; about 90% of seeds of this speciescould be induced into secondary dormancy by as little as a4-hour of anaerobiosis in some experiments (data not shown).Temperature during anaerobiosis did not influence the inductionof secondary dormancy (data not shown).The seed viability of both Tractopogon spp. was similarwithin 3 days of anaerobiosis treatment (according to errorbars) (Fig. 1). However, the seed viability declined rapidlywhen anaerobiosis durations increased beyond 3 days. After 12days of anaerobiosis, all seeds of both species were no longerviable as indicated by the tetrazolium test.B. Effect of temperature on seed germinationNon-dormant T. pratensis and T. dubius seeds germinated62Table 1. Induction of secondary dormancy in Tragopogon pratensisand T. dubius seeds by anaerobiosis treatment.Anaerobio-sis duration Percent germinationT. pratensis^T. dubius(days)0 93 801 13 282 10 323 0 04 5 10LSD(0.05)^9.31^9.31Percent germination values are the means of four replicates often seeds each. LSD is to compare means for differentanaerobiosis durations within each species.100080604020630^2^4^6^8^10^12Anaerobiosis duration (days)Fig. 1. Effect of anaerobiosis treatment on Tragopogon pratensisand T. dubius seed viability. Each value is the mean of fourreplicates of 10 seeds each. Vertical bars indicate standarderror of the mean.64under a wide range of temperatures (Fig. 2). Maximum germinationpercentages were established within 4-6 days at 15 C and within14-28 days at 25 C for the two Tragopogon species (Figs. 2 A,B).The optimum germination temperature for both weeds was 15 C(Fig. 2C). Some seeds germinated abnormally (i.e. no radicalelongation) above 30 C.In a separate study, temperature (5 to 35 C; up to 28 daysof exposure) did not stimulate the germination of secondarydormant seeds of both Tragopogon species (data not shown).Germination percentages of secondary dormant seeds of both weedswere below 20% when incubated at temperatures from 5 to 35 C; nogermination occurred at 40 C.C. Effect of stratification on secondary dormant seedsWhen secondary dormant seeds of T. pratensis were stratifiedat 4 to 5 C for 2 to 10 weeks and transferred to highertemperatures (10 to 30 C) for germination assay, germinationgenerally increased with the duration of stratification (Fig.3). The duration of stratification had a linear effect andtemperature during germination assay a quadratic effect on T.pratensis seed germination. There was a significant interaction(P < 0.05) between the duration of stratification and thetemperature during germination assay. At 10 and 20 C germinationassay temperature, germination increased with increasingstratification duration with total germination percentagereaching over 70% after 10 weeks. At 30 C, the germination100806040200A^A^1612100 -80 -60 -40 -2000B 4,--A A • A A/IrlOI/.4^8 20 2 824655 C0 10 cA 15 C• 20 C0 25 C■ 30 CIncubation time (days)C0-0CEQ.)05^10^15^20^25^30Temperature (C)Fig. 2. Effect of temperature on the germination of Tragopogonpratensis and T. dubius seeds. Each value is the mean of fourreplicates of 10 seeds each. A. time-course of germination ofT. pratensis, B. time-course of germination of T. dubius,C. Maximum germination percentage at various temperature in T.pratensis and T. dubius. Vertical bars indicate standard errorof the mean.C00CEQ)C0•_C•_EQ)066Fig. 3. Germination response of dormant Tragopogon pratensis seedsto duration of stratification (5 C) and temperature duringgermination assay. The regression equation is Z = -36.409 +11.262D + 4.929T - 0.115T - 0.298DT, where Z = germinationpercentage, D = guration of stratification (weeks), T =temperature (C). R = 0.751.67percentage remained less than 20% during 2 to 4 weeks ofstratification and then increased to approximately 40% at 10weeks. Seeds that did not germinate were confirmed as viable atthe end of the experiment using the tetrazolium stainingprocedure (Moore, 1972).D. The effect of after-ripening on seed germinationThe duration of after-ripening as well as temperature duringafter-ripening had significant effects on the release ofsecondary seed dormancy in T. pratensis (Fig. 4). Theinteraction between the effects of duration of after-ripeningand the temperatures during after-ripening was also significant(P < 0.05). The germination percentage increased significantlywith the duration of dry storage at 30 C with the maximum valuereaching to almost 80% after 12 months of dry storage. Seedgermination of T. pratensis gradually improved with increasingduration of after-ripening at 10 and 20 C, however improvementwas slow and the total germination percentages reached only 50%and 53%, respectively after 12 months. The difference of seedgermination at 10 C and 20 C was not significant (P < 0.05). Theviability of secondary dormant seeds after 12 months ofafter-ripening remained over 90%.E. Effect of light on seed germinationRed or far-red light had no effect on the germination of T.pratensis seeds with secondary dormancy (data not shown). These20 C10 C3 0 CLSD 0.05680^2^4^6^8^10^12^14Duration of after-ripening (months)Fig. 4. Effect of temperature during after-ripening of air-driedTragopogon pratensis seeds on seed dormancy. Each value is themean of eight replicates (from two experiments) of 10 seedseach. LSD is for comparing different temperature levels for thesame duration of after-ripening.69treatments were found to influence germination of lettuce seedsas has been reported in the literature (Bewley and Black, 1985).F. Effect of seeding depth on seed germination and seedlingemergenceSeeding depth had a significant effect on seedling emergenceat P < 0.05. Both T. pratensis and T. dubius seeds planted at 2cm deep in the soil germinated 100% and almost 80% of theseedling emerged (Fig. 5). The emergence percentage decreasedwith the increase of the seeding depth. Buried seeds wereexhumed for germination observations at the end of theexperiment. Seeds planted 8 cm and deeper germinated but theseedlings did not emerge.DiscussionA. Induction of secondary dormancySeed dormancy is a widespread mechanism in the persistenceof weeds (Harper, 1959; Chancellor, 1982; Bewley and Black,1985; Baskin and Baskin, 1989a). While primary (innate) dormancyhas received much attention, secondary (induced) dormancy canalso provide non-dormant seeds with an additional mechanism toavoid germination under conditions unfavorable for seedlinggrowth (i.e. water logging, low winter temperatures) (Come andTissaoui, 1973; Tilsner and Upadhyaya, 1985; Mayer andPoljakoff-Mayber, 1989). For example, seeds may experience80°Q 60.......-- ,-,■.,000C 40E 20Lil0)0/^/^/^/ \k- C\ 1 05 8Seeding depth (cm)280"----'60®0C400)0)4- 20EWI5Seeding depth (cm)270Fig. 5. Effect of seeding depth on the emergence of Traqopogonpratensis (A) and T. dubius (B) seedlings. Percent emergencevalues are the means of six replicates of 10 seeds each. LSD isfor comparing the means for different seeding depths.71anaerobic conditions when buried deep in soil, or as a result ofwater logging caused by rain or melting snow in the spring.Anaerobiosis using deaerated water is a convenient technique toinvestigate induction of secondary dormancy under conditions oflow oxygen availability. Using this technique, high levels ofsecondary dormancy in seeds of both Tragopogon spp. could beinduced by a one day anaerobiosis treatment (Table 1). Thisobservation is significant because T. pratensis, which does notexhibit long-term innate dormancy (Chapter 4), may rely onsecondary dormancy to delay its germination under conditions oflow oxygen availability. Secondary dormancy delays Tragopogon spp. seed germination until environmental conditions arefavorable for germination and seedling growth. While secondarydormancy was induced in the seeds of both Tragopogon spp., T.pratensis was more responsive than T. dubius; a high level ofdormancy was induced in T. pratensis by a 4-hour anaerobiosistreatment which had no effect on T. dubius.Higher anaerobiosis temperatures (25 and 30 C vs. 15 and 20C) have been shown to be more effective at inducing secondaryseed dormancy in Avena fatua (Tilsner and Upadhyaya, 1985).However, in this study temperature during anaerobiosis did notinfluence the induction of secondary dormancy in Tragopogon spp.(data not shown).Long durations of anaerobiosis can influence the viabilityof non-dormant seeds. It is, therefore, important to separatethe influence of anaerobiosis on seed dormancy and seed72viability. Many researchers often do not make the distinctionbetween dormant and dead seeds and assume all seeds that fail togerminate to be dormant. This research shows that Tragopogonseeds subjected to one day anaerobiosis do not germinate becauseof induction of dormancy and not due to the loss of seedviability. On the other hand, low germination of seeds receivinglonger anaerobic treatments is in part due to induction ofdormancy and in part due to loss of viability. Anaerobiosis for12 days or longer resulted in complete loss of viability. Thefact that seed viability of both weeds declined with increasingexposure to anaerobic conditions (Fig. 1) suggests that theseweeds are adapted to well-drained soils (e.g. loam or sandy loamsoil). Waterlogging for periods longer than three days wouldincrease the proportion of dormant seeds in seed banks;waterlogging for 12 or more days would result in completeelimination of all non-dormant seeds from the seed banks.B. Effect of temperature on germinationA seed's response to temperature can restrict itsgermination to specific times of the year and is particularlyimportant in temperate zones where temperature variessignificantly during the year (Koller, 1964; Stoller and Wax,1973; Bewley and Black, 1985; Roberts, 1988; Hermanutz andWeaver, 1991). This response can play a significant role in theadaptation of species to a new habitat. The characteristicbell-shaped distributions of germination versus incubation73temperature for T. pratensis and T. dubius are similar to twoother asteraceous weeds Centaurea diffusa and C. maculosa foundin B.C. rangelands (Nolan, 1989). The maximum germination isattained at 20 C in both C. maculosa and C. diffusa seeds.Accordingly, it would not be surprising if both Tragopogonspecies also infest areas infested by knapweeds. The observedoptimum germination temperature of 15 C for both Tragopogon spp.suggests that their seeds can germinate and seedlings establishearly in the spring. The seeds that did not germinate at 5 Cwere still viable. When seeds that failed to germinate at 30 Cor higher temperature were dissected, a high percentage of seeddecay was observed.C. Effect of stratification on seed dormancyGermination of Tragopogon seeds in autumn would result inhigh seedling mortality due to competition from establishedplants as well as from freezing injury in the winter. This studyshowed a stratification (at 4 to 5 C) requirement forgermination of secondary dormant Tragopogon seeds. Thestratification requirement would not allow dormant seeds togerminate immediately before the inclement winter, therebyencouraging establishment of seedlings early in the spring whenthere is minimal competition with other plants and a reducedfreezing injury.74D. Effect of after-ripening on secondary dormancyRecently matured, innately dormant weed seeds lose theirdormancy as they after-ripen during dry storage (Tilsner andUpadhyaya, 1985; Baskin and Baskin, 1987; Roberts, 1988).Although induced dormancy provides a good system to studyafter-ripening, little is known about the changes in dormancylevel of secondary dormant seeds during after-ripening. Theresults in this study showed loss of secondary dormancy in T.pratensis seeds during dry storage. There also was a significantinteraction between temperature and duration of after-ripening.The ability of Tragopogon seeds to become secondary dormant andgradual release of this dormancy suggest that these species mayrely on induced dormancy as an option in their persistencestrategy. The interaction between the durations ofafter-ripening and temperature (Fig. 4) may also be important ininfluencing the timing of germination in nature.E. Effect of light on seed germinationCresswell and Grime (1981) reported a strong inhibition ofT. pratensis seed germination in darkness. The results in thisstudy, however, show that light does not stimulate germinationof T. pratensis seeds with secondary dormancy. Completegermination of seeds buried 14 cm deep (Fig. 5) further supportsthe light-insensitivity of seed germination in these species.The underlying basis of the difference in the results of twostudies could be due to differences in genotype, environment,75and/or the nature of seed dormancy (i.e. primary or secondarydormancy) in these studies. The results in this study, however,suggest that Tragopogon seeds can germinate on the soil surface,while buried in soil, or under a leaf canopy when other factorsare conducive to germination.F. Effect of seeding depth on seed germination and seedlingemergenceIn rangelands, unlike cultivated fields, seeds generallystay at or near the soil surface (Roberts, 1981). In apreliminary field study, about 86% of T. pratensis seeds werefound to be present at or near the soil surface (0-1 cm)(Chapter 4). Weed seeds present on the soil surface inrangelands can, however, become buried by the movement ofgrazing animals (Thomas, 1960), earthworms (McRill and Sagar(1973), or deposition of wind eroded soil particles (Sheldon,1974). Burial of seeds to a deeper soil profile can induce seeddormancy (Chancellor, 1982), which can delay or decrease seedgermination, and may influence seedling emergence(Froud-Williams et al, 1984; Blackshaw, 1990). Froud-Williams etal. (1984) reported that the critical depth for seedlingemergence was at 5 cm for most species investigated in theexperiment except species with large seed size. Germination anddecay of achenes of Crupina vulgaris prior to removal from thefield were greater at 0-10 cm burial depth than at 20 cm burialdepth (Thill et al., 1985). Blackshaw (1990) reported that76seedling emergence of Malva pusilla was optimal at depths of 0.5to 2 cm but progressively declined as depth increased. Noemergence was recorded at 8 or 10 cm depth. Seedling emergenceof Datura spp. decreased with increasing sowing depth and noemergence occurred from the depths below 10 cm (Reisman-Bermanet al., 1991). To understand the ecophysiology of persistence ofT. pratensis and T. dubius, the fate of their buried seeds needsto be investigated. In particular, we need to know whether theseeds buried deep in the soil become (secondary) dormant, andfrom what depth a seed can produce a seedling which emergesabove the ground.Results of the greenhouse study show that the emergence ofTragopogon seedlings decreased with increasing seeding depth andno seedlings emerged when the seeds were planted deeper than 5cm (Fig. 5). The fact that seeds buried deep in soil did notbecome dormant suggests that the reservoir of non-dormantTragopogon seeds would become completely depleted underconditions conducive to germination regardless of theirdistribution (up to 14 cm) in the soil profile. Seasonal changesin dormancy behaviour of seeds under natural conditions, and theability of Tragopogon spp. to maintain a persistent seed bank inthe soil will be discussed later (Chapter 4).In summary, secondary dormancy could be a mechanismresponsible for the persistence of T. pratensis and T. dubius.These weed seeds germinate in autumn or spring when the77temperature rises above 10 C. Stratification and/orafter-ripening are essential for breaking the seed dormancy.Deep burial of seeds decreases the emergence but does notinhibit the seed germination because light is not required forgermination. These results could help develop an effective weedcontrol strategy for these weeds. Seasonal changes of seeddormancy under natural conditions should be investigated todetermine the persistence of seeds of these weeds.78Chapter 4. Seasonal changes of seed dormancyof buried Tragopogon pratensis L. seedsAbstractThe dynamics of seed populations of Tragopogon pratensis were investigated under natural conditions to monitor seasonalchanges in dormancy behaviour of buried seeds. Freshly maturedseeds were mixed with sterilized soil, the mixture was placedinto nylon bags, and the bags were buried at three soil depthsin a T. pratensis infested rangeland. The burial depth had asignificant effect on innate seed dormancy at 2 to 3 months ofburial. The seed populations were, however, nearly completelydepleted after 9 to 10 months of burial. Less than 3% of theburied seeds remained viable after 13 months. Two germinationpeaks in T. pratensis seeds were observed: one in autumn afterburial and the other in the following spring. The survivorshipcurves between 1989 and 1990 cohorts were not significantlydifferent as indicated by the Logrank test. These resultsindicate that T. pratensis seeds are short-lived in the soil andsize of its persistent seed bank is very small. The fact thatseed banks of T. pratensis are short-lived also suggests thatpersistence of this weed in rangelands may depend on its highseed production.79IntroductionSeed dormancy plays an important role in survival of manyannual and perennial weeds (Roberts, 1970; Sarukhan, 1974;Weaver and Cavers, 1979; Roberts and Chancellor, 1979; Robertsand Nelson, 1981; Roberts and Boddrell, 1983; Colosi et al.,1988; and Kalisz, 1991). Seed dormancy distributes seedgermination over time, which contributes to the formation ofseed banks in soil. For example, Centaurea maculosa L., anoxious rangeland weed of British Columbia, has a persistentseed bank with over 50% of its seed population still viableafter 5 years of burial (Davis, 1990). A persistent seed bank isa great barrier to effective weed management. Knowledge of seedbank dynamics, germination cues, and environmental conditionssuitable for seedling establishment help develop bettereffective weed control (Frond-Williams et al., 1984; Parker etal., 1989).Seed bank dynamics of individual populations are primarilyinfluenced by germination, dormancy and mortality (Sarukhan,1974; Watkinson, 1978; Weaver and Cavers, 1979; Hutchings,1986a; Parker et al., 1989). Changes among these fractions varyover spatial and temporal scales. Seed banks can be eithertransient, with seeds germinating within one year of initialdispersal, or persistent, with seeds remaining viable in thesoil for more than one year (Thompson and Grime, 1979).Tragopogon pratensis L., like many other weeds of the Compositae80family, reproduces only by seeds and produces numerous seeds perplant. However, the role of seed banks in its persistence hasnot been investigated. The fate of buried seeds, seasonalpatterns of dormancy, and the relationship between emergence andenvironmental conditions for this weed are poorly understood.Because habitats are spatially diverse, heterogeneity ofseed banks, both horizontal and vertical, can affect theestablishment of a species (Parker et al., 1989). In nature,seeds are present at different depths in the soil profile, andvariation in the soil profile climate can affect seedgermination and longevity. Burial of seeds in some species candelay seed germination either by preventing it or by inducingseed dormancy (Wesson and Wareing, 1969; Watkinson, 1978; Weaverand Cavers, 1979; Roberts and Boddrell, 1983; Froud-Williams etal., 1984; Eberlein, 1987), which in turn could result information of seed banks in the soil. Roberts and co-workers(Roberts and Nelson, 1981; Roberts and Boddrell, 1983, 1984;Roberts, 1986) investigated seed survival and seedling emergencein a wide range of arable weeds under simulated cultivationconditions. They, however, did not determine the effect ofburial depth on the fate of weed seeds. Whether characteristicsof T. pratensis seed banks are affected by burial depth hasnever been explored.Seasonal patterns of seed dormancy vary among species.Environmental factors such as temperature and moisture alsoinfluence seed germination and seed dormancy (Karssen, 1982;81Baskin and Baskin, 1983, 1984; Bouwmeester and Karssen, 1989;Venable, 1989; Baskin and Baskin, 1990). For example, buriedseeds of Sisymbrium officinale and Polygonum persicaria loseprimary dormancy due to stratification in winter and germinateearly in spring when conditions are favorable for seedgermination (Bouwmeester and Karssen, 1989). The timing of seedgermination within one year determines seasonal fluctuations inthe size of seed banks and how many seeds will persist to thenext year (Roberts and Boddrell, 1983; Venable, 1989).Artificially established seed banks are particularly useful insuch studies because burial depths can be controlled andquantitative information about seed germination and survivalduring the burial treatment can be obtained (Roberts, 1986;Karssen, 1982).Differences in survival of seeds among cohorts orpopulations are not often analyzed in ecological studies (Weaverand Cavers, 1979; Bouwmeester and Karssen, 1989). This issue hasrecently been addressed by several scientists (Pyke andThompson, 1986; Hutchings et al., 1991). The application ofstatistical analyses in demographic studies has also beenreviewed (Hutchings et al., 1991). Statistical analyses helpecologists interpret the causes of seed mortality through time.In this study, survivorship curves will be statisticallycompared to determine the similarity in survival rates of T.pratensis seed cohorts.The specific objectives of this study were to answer the82following questions: 1) does burial depth affect the dormancyand survival of buried seeds, 2) does seed dormancy change withseasonal changes in environment, 3) do survival rates of buriedseeds differ between experimental years (cohorts), and 4) is thepersistence of T. pratensis determined by a long-lived seedbank. Answers to these questions are of value in understandingthe persistence strategies of this weed and in the developmentof its control measures.Materials and methodsA. Description of the study siteRiske Creek in central British Columbia (52 ° 01'N, 122 °31'W, elevation 1006 m) is within Krajina's (1965) PonderosaPine-Bunchgrass biogeoclimatic zone. This area receives lessthan 200 mm rainfall during the months of May through September,and has a frost-free period (days > 0 C) of about 72 days(Environment Canada, 1982).The mean monthly temperature and precipitation during thestudy period were collected from the local Environment Canadaweather station. Changes in the temperature and precipitationfor the study area from July, 1989 to August, 1991 are shown inFig. 1. The fact that mean monthly temperatures were above zeroin March, 1990 and in April, 1991 suggests that the growingseasons started at different times in these two years. Theaverage total precipitation during the growing season (May to83NCDa)QEFig. 1. The mean temperature and precipitation in Riske creek.Williams lake, B.C. between August 1989 and August 1991 (drawnfrom Environment Canada data).- 10152015140120100fE80 cO-a-56040 a=200--•— TemperatureE-1 Precipitation1 :JAsoND^1989^JFmAnAJJAsoND1990^FMAMJJA1991^84September) is under 200 mm. However, in August of 1989, 124 mmof rain fell on the study area compared to 30 mm and 39 mm inAugust 1990 and 1991, respectively. Precipitation ranged between230 mm to 260 mm in both winters.This study was conducted in a relatively homogenous areawithout any grazing (fenced) or herbicide use. The T. pratensis population was a near monoculture, with a spring seedlingdensity of about 140 plants/m 2 and the plant density of 90plants/m2 in May, 1990.B. Field data collectionThe distribution of T. pratensis seeds in the soil profilewas determined by taking 10-cm diameter soil cores to a depth of6 cm from undisturbed soil and dissecting them in the laboratoryin four layers (0-1, 1-2, 2-4, and 4-6 cm). The four sectionswere dried outdoors under sunlight. Seeds of T. pratensis fromeach layer were separated and counted (Fig. 2).Seeds (i.e., achenes) for burial experiments were collectedfrom plants in infested areas in early August 1989, and in midJuly 1990. The seeds were immediately mixed with sterilized soiland placed in mesh packets (10 x 10 cm, 200 seeds per packet)made from nylon 'No-See-Um Netting' (Mountain Equipment Co-op,Vancouver, B.C.). This facilitated seed recovery and preventedmixing of experimental seeds with seeds naturally present in thesoil. The mesh size was small enough to prevent seed loss butpermitted the movement of water and oxygen to the seeds. Seed1000-1 1-2 2-4 4-693.879.986.810.97.7- 60C')It3 40CDCDCfp 2008 5Soil depths (cm)Fig. 2. The distribution of Tragopogon pratensis seeds in thesoil profile. Each value is the mean of ten replicates.Vertical bars indicate standard error of the mean.86bags were buried at three depths: 0, 2, and 5 cm, based on theinformation of soil seed distribution in Fig. 2. Seed bagsplaced on the soil surface were secured with nails.All experiments were arranged in a randomized complete blockdesign with four blocks (200 seeds per bag) as replicates. Theburied seeds in the first experiment were exhumed in October,1989, and in May, July, and August, 1990. The seeds in thesecond experiment were exhumed in August and October, 1990, andin May, July, and August, 1991. Exhumed seed bags weretransported to the laboratory. Seeds were rinsed with distilledwater and seed germination was assayed at 25 C in darkness.Viability of seeds that failed to germinate was determined usingthe tetrazolium staining procedure (Moore, 1972). To test if thenewly matured seeds were innately dormant, seeds (fourreplicates with 10 seeds of each) collected from plants in 1989and 1990 were put in 9-cm Petri dishes and tested forgermination at 25 C in darkness.C. Data AnalysisFor data analyses, the seed population was classified asfollows (following Sarukhan, 1974): a) seedlings (S), the totalnumber of seedlings from seed samples taken at each collectiondate of the sample; b) enforced-dormancy (ED), the fractionwhich germinated under the laboratory conditions; c)induced-dormancy (ID), seeds that failed to germinate under thelaboratory conditions but were viable as determined by the87tetrazolium staining test; d) non-viable (NV), the dead fractionof the seeds, e) the empty fraction (EF), seeds withoutendosperm and embryo. Microscopic examination of the EFfractions gave no indication of whether they represented dead,germinated, or eaten-by-predator seeds. By adding these fivecategories, the total seed population (TSP) was obtained.TSP (200 seeds) = S + ED + ID + NV + EFAnalysis of variance (ANOVA) was performed to test the effect ofburial depth on population of dormant seeds.The survivorship curves between 1989 and 1990 cohorts werecompared using the Logrank test (Pyke and Thompson, 1986;Hutchings et al., 1991). The logrank value (LR) was calculatedusing the equation: LR = (d 1-E 1 )2/E1 + (d2-E2)2/E2. Where:d1=number of death over the entire observational period incohort 1, E 1=expected number of death over the entireobservational period in cohort 1, d2=number of death over theentire observational period in cohort 2, E 2=expected number ofdeath over the entire observational period in cohort 2. The LRvalue was compared with X 2 (1, 0.05) (Chi-squared critical value)to determine the survival difference between two cohorts. A lifetable on seed mortality risks in two cohorts over the entireobservational periods was generated. Detailed analysis procedurefor constructing the life table will be given in Chapter 5.88ResultsA. Effect of burial depth on seed dormancy1) Seed distribution in soil profilesOver 86% (approximately 1427 seeds/m 2 ) of the T. pratensis seeds were present in the top 1 cm (Fig. 2). The remaining 14%were distributed between 1-4 cm deep. No seeds were found below4 cm depth.2) Seed germination and dormancySeed germination (i.e. seedlings) decreased significantly (P< 0.05) when seeds were buried, and seed dormancy (ED, ID)significantly increased (P < 0.05) with burial depth at 2 monthsof burial in 1989 cohort (Table 1), and at 3 months of burial in1990 cohort (Table 2). Seed populations were nearly completelydepleted after 9 months of burial in 1989 and 10 months in 1990cohorts.While 129 out of 200 seeds germinated on the soil surfaceafter 2 months of after-ripening in 1989 cohort, only 18 and 2seeds out of 200 germinated at 2 and 5 cm depths respectively(Table 1). No seed germinated after 1 month of burial regardlessof the burial depths in 1990 cohort. Approximately 78 out of 200seeds germinated on the soil surface after 3 months of burial in1990 cohort, and only 48 and 14 seeds germinated at 2 and 5 cmdepths (Table 2). The dormant fraction (ED plus ID) in both89Table 1. Effect of burial depth on population dynamics ofTragopogon pratensis seeds in 1989 cohort. Each value is themean of four replicates (±SE). Total seed population = 200 seeds(S + ED + ID + NV + EF).Depths Sa ED^ID^NV^EFafter 2 months of burial (exhumed Oct/1989) O cm^129+10.0** 43+14.5^9+ 2.5^1+0^19+11.02 cm 18+ 0.5^66+14.0 23+ 7.0^1+0.5 92+14.0**5 cm^2+ 2^88+12.5**^61+12.0** 2+0.5^47+ 4.5after 9 months of burial (exhumed May/1990) O cm^90+18.0^7+1.0^4+1.0^4+1.5^97+17.02 cm^43+ 9.0^0^0^0^157+ 9.0**5 cm^88+16.5^1+0.5^1+0.5^0 110+16.5after 11 months of burial (exhumed Jul/1990) O cm^52+6.5^4+0.5^1+0.5^1+0.5^143+6.02 cm 5+0.5 0^0^0^195+1.05 cm^6+0.5^2+0.5^0 0 192+1.5after 12 months of burial (exhumed Aug/1990) O cm^13+1.5^3+1.0^1+0^2+0^184+2.02 cm^0^0^0 0 199+0.55 cm^0 4+1.0^0 1+0.5^195+1.5** The mean is significantly different (Tukey's test at P <0.05)from the mean obtained from the other burial depthswithin each harvesting time.a S: seedlings, ED: enforced dormant seeds, ID: induceddormant seeds, NV: non-viable seeds, EF: empty fraction.90Table 2. Effect of burial depth on population dynamics ofTragopogon pratensis seeds in 1990 cohort. Each value is themean of four replicates(+SE). Total seed population = 200 seeds(S + ED + ID + NV + EF).Depths0 cm2 cm5 cmSa^ED^ID^NV^EFafter 1 month of burial (exhumed Aug/1990)000after^47+5.5^153+5.5^1+0.5^085+7.5 115+7.5^0^068+7.5^130+8.0^2+0.5^03 months of burial (exhumed Oct/1990)0 cm 78+ 7.5** 47+13.5^24+5.0^0^51+ 5.52 cm 48+12.5 41+26.0 2+1.0 0 115+30.5**5 cm 14+ 7.0 174+ 66.5** 13+3.5^0 3+ 1.5after 10 months of burial (exhumed May/1991)0 cm 152+10.0 0^0^0^48±10.02 cm 94+ 5.0 0 0 0 106+ 4.0**5 cm 144+10.0 0 2+0.5^0 56+ 9.0after 12 months of burial (exhumed Jul/1991)0 cm 9+4.0 2+0.5^0^0^190+4.02 cm 7+1.5 0^0 0 193+5.05 cm 1+1.0 3+1.0^0 0 197+3.0after 13 months of burial (exhumed Aug/1991)0 cm 8+1.5 0^0^1+1.0^191+2.02 cm 10+2.5 1+0.5^0 0 189+2.55 cm 0 3+0.5 0 0 197+0.5** The mean is significantly different (Tukey's test at P <0.05) from the mean obtained from the other burial depthswithin each harvesting time.a S: seedlings, ED: enforced dormant seeds, ID: induceddormant seeds, NV: non-viable seeds, EF: empty fraction.91cohorts were very small after 9 or 10 months of burial in thesoil profile. There was no significant change in non-viablefraction among the burial depths in both cohorts. Shallow burialdepth (2 cm) seemed to facilitate the EF fraction. No viableseeds remained after 9 or 10 months burial in both cohorts at 2cm depth. The empty fraction had perfect pericarps withoutintact seeds inside.B. Seasonal change in the seed dormancyFreshly collected T. pratensis seeds showed a high degree ofinnate dormancy in both cohorts. Germination assay showed that28% + 13% of the seeds germinated in 1989 cohort and 25% + 13%of the seeds germinated in 1990 cohort. The seeds which failedto germinate were alive but remained dormant as indicated by thetetrazolium staining test.Seasonal changes in seed dormancy of T. pratensis weredetermined by measuring germination in this study. Over 60% ofthe seed population was depleted (because of germination) on thesoil surface in the late autumn (October) in 1989 cohort, whileonly less than 10% germinated during the same period at 2 and 5cm (Fig. 3). The remaining portion of dormant seeds (ID + ED)was nearly completely depleted in the following spring (May)regardless of the burial depth. Since some of the T. pratensis seeds matured in July in 1990, the burial experiment in 1990 wasconducted in July, unlike in August, 1989. However, there was noseed germinated in August after 1 month of burial (Fig. 4). Only100%80%60%40%20%92Oct^May^ Jul^Aug0 cm-0 80%QD- 80%0 40%........ .. ........... ..;:;::::::::.:.......... .. ........... ... ........... ... ............ ..................„. ......„.....^••.. .. .......---.........—. .... ............... .. ............ .. ......2 cmOct May^ July^ Aug5 cmJulyTimeFig. 3. The fate of Tragopogon pratensis seeds buried in August1989 at 0, 2, and 5 cm depths in the soil. ;;A S: seedlings;ED: enforced dormant seeds;    ID: induced dormant seed;NV: non-viable seeds; [ 1 EF: empty fraction.,2 cm0 cm5 cm-0080%80%040%0)0320%c0)0OL 100%Aug93Fig. 4. The fate of Tragopogon pratensis seeds buried in July1990 at 0, 2, and 5 cm depths in the soil.^S: seedlings;ED: enforced dormant seeds; ^ ID: induced dormant seed;NV: non-viable seeds; I^EF: empty fraction.9439% of seeds on the soil surface, 24% at 2 cm, and 7% at 5 cmdepth germinated in October in 1990 cohort. As for 1989 cohort,the seed population was also depleted in the following spring.Less than 3% of the buried seeds in both cohorts remained afterone year of burial.C. Comparison of survivorship curves between yearsAlthough the T. pratensis seeds were collected in differentyears and the burial experiments were carried out at differenttimes of the years, both cohorts showed similar survivorshipcurves (Fig. 5). The difference between the two curves was notsignificant (P > 0.05) over the two entire observation periodsas indicated by Logrank test. The highest seed mortality rates(k values) in both cohorts were found in October (Table 3). Theseed mortality in this table however, means that the seedpopulations were depleted by germination, predation or seeddecay. The proportion of non-viable seeds in both cohorts of T.pratensis was very small.DiscussionA. Seed distribution in soil profilesTo understand spatial heterogeneity of seed banks, it isessential to determine the distribution of seeds in soilprofiles (Garwood, 1989). Seed distribution in soil profile alsoplays an important ecological role in the successful951000OD00) 100CDC/Da)10E1- 1989 population^1990 population1111111111111111_111 1 11 111 JASONDJFMA\AJJASONDJFMAMJJAS1989^1990^I^1991Calencar yearFig. 5. Survivorship curves of Tragopogon pratensis buried seedsin August 1989 and July 1990 cohorts. Two curves did not differat P > 0.05 by the Logrank test.96Table 3. Life table of Traqopogon pratensis L. seed populationsat Riske creek of B.C. for 1989 and 1990 cohorts. The seedswithin the soil profile were pooled (i.e. 600 seeds from each of0, 2, and 5 cm depths) for analysing the survivorship andmortality rate.*x^ax^1x^dx^qx^logax^kx1989 cohortAug/89 600 1000 515^0.515 2.778 0.314Oct/89 291 485 463 0.955 2.464 1.350May/90 13 22 10^0.455 1.114 0.269Jul/90 7 12 0 0 0.845 0Aug/90 7 12 0.8451990 cohortJul/90 600 1000 3^0.003 2.778 0.001Aug/90 598 997 497 0.498 2.777 0.300Oct/90 300 500 492^0.984 2.477 1.778May/91 5 8 0 0 0.699 0Jul/91 5 8 0^0 0.699 0Aug/91 5 8 0.699* x:^Calendar time from seed burial.ax : Number of seeds observed alive at each observation timex.1x : Standardized number surviving at the start of each timeinterval.dx :^Standardized number dying between x and x+1.qx: Average mortality rate per day during the intervalbeginning with day x.kx :^Killing power, or rate of mortality during the intervalbeginning with day x. kxi = logaxi - logaxi+197establishment of many species. Firstly, deeply buried seeds canavoid surface predation by birds, insects, and rodents.Secondly, asynchronous germination due to seed dormancy withinsoil profiles can maintain the weed populations duringunfavorable environmental conditions. Seedlings from seedsgerminating on soil surface face a greater risk of injury due tothe desiccation or unseasonable frost episodes.Cresswell and Grime (1981) found no buried seeds of T.pratensis within the soil profiles. In this study however, whileabout 86% of T. pratensis seeds were observed on the soilsurface, 14% were found below 1 cm depth (Fig. 2). Thedifference between these two studies could be due to thedifferences of vegetation covering and frequency of animalactivity between the two study sites.How seeds of T. pratensis become buried in the soil is notwell established, but several possibilities arise from work donewith other species. Seeds can be buried in the soil byearthworms (McRill and Sagar, 1973; Grace, 1984), by falling incracks developed during dry seasons (Cavers and Benoit, 1989),or by accumulation of soil particles eroded by water. Inaddition, freezing and thawing cycles can also form cracks inthe soil and seeds can be washed into these cracks by runoffwater from melted snow or by early spring rains (Cavers andBenoit, 1989).Trampling by animals in rangelands enhances seed burial(Thomas, 1960). Since grazing was eliminated by fencing in this98study, a higher percentage of T. pratensis seeds might have beenburied under natural conditions with grazing. The results alsosuggest that germination behaviour of T. pratensis can varywithin the soil profile possibly associated with soilmicro-environment. Thus, the distribution of T. pratensis seedswithin soil profiles may influence seed dormancy, which in turncould distribute germination over time.B. Seed banks in soil profilesSeeds of T. pratensis germinated more rapidly on or near thesoil surface (0 cm) than when they were buried deeper (5 cm) inthe soil (Tables 1 and 2). Dramatic changes in climaticconditions may be responsible for this high seed germination onthe soil surface. High temperature fluctuations near the soilsurface and wetting-drying cycles after seed dispersal are knownto stimulate seed germination by releasing seed dormancy inother species (Stoller and Wax, 1973; Weaver and Cavers, 1979;Baskin and Baskin, 1989b). Buried seeds, on the other hand, takea longer time to germinate due to their relatively more stablesoil environment.Since light penetration is severely reduced at even a 1 cmdepth in the soil (Vincent and Cavers, 1978), burial of seedscan influence germination of light requiring seeds (Wesson andWareing, 1969; Baskin and Baskin, 1989b). Cresswell and Grime(1981) reported that T. pratensis seeds had very low germinationunder the dark condition. The pot culture experiment reported in99Chapter 3 indicated that T. pratensis seeds germinated whileburied at 5 cm depth. In this study, seed populations of T.pratensis were nearly completely depleted after 9 or 10 monthsof burial (Tables 1 and 2). The fact that the seeds of T.pratensis germinated while buried demonstrates that light is notnecessary for seed germination in this species. Thus, T.pratensis seedlings should be able to emerge between grassplants when other germination requirements are met. Thischaracteristic suggests that T. pratensis has the potential tospread to not only the disturbed areas but also to the areaswhere grass canopies are thick.The proportion of empty seeds in the buried seed populationincreased with time (Tables 1 and 2). This suggests that someseeds had germinated and rotted or were eaten by herbivoresbefore the observation was made. Schafer and Chilcote (1970)found a high mortality of Lolium perenne seeds (85%) due to lowseedling survival when seeds were buried at greater depths.C. Seasonal changes in dormancy behaviourMany weeds display some form of innate dormancy andpolymorphism for this trait (Cook, 1980). Germination tests onfreshly collected seeds of weeds in presence of adequate lightand moisture can demonstrate innate dormancy and provide anindication of the potential for immediate seedling recruitment(Roberts, 1986). Germination tests in this study showed thatover 75% of fresh T. pratensis seeds collected from 1989 and1001990 were innately dormant; the remaining 25% germinated readilyunder favorable conditions. This innate dormancy suggests thepotential of T. pratensis to form a persistent seed bank insoil. These results differ from those of Grime et al. (1981)where over 90% of freshly collected T. pratensis seedsgerminated. This difference may be due to genetic difference inthe two seed populations or a difference in climatic conditionsduring seed maturation.The depletion of seed reserves in soil under a particularset of environmental conditions depends on the rate of seedgermination (Weaver and Cavers, 1979). The seasonal patterns ofseedling emergence shown by many species suggest that climaticconditions, in particular temperature, can influence the timingof peak seedling emergence (Baskin and Baskin, 1984;Found-Williams et al., 1984; Bouwmeester and Karssen, 1989). Inthe 1989 study, the soil was very wet from the heavy rain in themonth of August (Fig. 1) and the monthly mean temperature wasabove 5 C in September and October. A flush of seed germinationwas observed on the soil surface in 1989 cohort before the soilfroze (Fig. 3). The dormant seed fractions, stratified duringthe winter, were almost completely depleted by the followingspring when the monthly mean temperature increased to 5 C. After12 months burial, only about 2% of the seeds remained viable onthe surface and at 5 cm depth. A spring flush of emergence ofPlantago ma or has been attributed to rising soil temperatureand subsequent flushes were related to rainfall pattern101(Froud-Williams et al., 1984). In the 1990 study, no germinationoccurred in buried T. pratensis seeds within one month (Table 2)possibly due to the dry weather (Fig. 1). The seeds may haveafter-ripened and become non-dormant in the hot weather, butcould not germinate because of the low precipitation. Athree-month hot and dry spell preceded seed burial, and only asmall fraction of seeds germinated by October, 1990 (Fig. 4).The remaining dormant seeds in the soil were subjected to lowtemperature (stratification) in the winter. The seed bank wasnearly completely depleted by the following spring. Less than 2%of the seeds remained in the soil after 13 months. Germinationof Vulpia fasciculata seeds has been shown to be dependent uponthe level of soil moisture; more than 99% of the seedsgerminated regardless of the burial depth once the soil moisturereached a favorable level (8%) in the autumn (Watkinson, 1978).The timing of within-year seed germination is ecologicallyimportant for successful establishment of plants underfluctuating conditions (Venable, 1989). Seedlings emerging inthe autumn and having some time to develop their root system andaccumulate a large amount of food reserve can survive thewinter. Accordingly, seedlings with low food reserves areexpected to experience high mortality in the winter. Theseedlings emerging in the spring, on the other hand, ensuretheir establishment by avoiding the winter. Thus, seedlingestablishment is closely related to the timing of seedgermination. There are three possible within-year germination102schedules for seeds to yield highest fitness (Fig. 6) (Venable,1989). Seeds of some species germinate predominately in fall(Fig. 6 C), some in spring (Fig. 6 A), and some in both fall andspring (Figs. 6 A and C). Moreover, some species have fairlysynchronous germination (Fig. 6 A or C), while others spreadgermination over a long period of time during the growing season(Fig. 6 B). From the two-year field studies it can be concludedthat T. pratensis seeds germinate in autumn and in spring whenenvironmental conditions are favorable for germination. Thegermination schedules of T. pratensis seeds are of Type A and C.Obviously, the Type B does not apply for T. pratensis seedssince almost its entire seed populations from a single year'scrop was depleted in autumn and spring. Two germinationschedules in T. pratensis seeds may be, therefore, the result ofnatural selection to avoid complete destruction of seedlingsfrom only one germination schedule.Knowledge of seasonal patterns of seedling emergence has adirect practical application in weed control. It can be used indeveloping the best management strategy for a weed. Autumn andspring are two germination windows for T. pratensis seeds.Destruction of T. pratensis seedlings in the spring caneliminate seedlings emerging both in the autumn and the spring.Since this weed does not maintain a large persistent seed bank,seedling control in the spring can be a particularly effectivemanagement strategy to deplete the seed banks of this weed.103A^B^CJAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DECTime of yearFig. 6. Possible within-year germination schedules. A. springwindow; B. a long period of time during the growing season; C.autumn window (from Venable, 1989).104D. Survivorship curves between yearsThe survivorship of a seed population in the soil isdetermined by the loss of seed dormancy over time. The fact thatsurvivorship curves for T. pratensis seeds between years did notdiffer (P > 0.05) suggests that the probability of seed survivalwithin the two observational periods is dependent upon the ageof the seed population and not the experimental years. In otherwords, the climatic differences between the two experimentalyears only affected the total germination percentages in eachgermination window (autumn and spring) but did not influence theoverall survivorship of the two seed populations. Seed mortalityrates within a year are still determined by seasonal climaticconditions indicated by k values, particularly the conditions inautumn (k=1.35 in the 1989 cohort) and spring (k=1.778 in the1990 cohort) (Table 3). The mortality rate here indicates thatseeds either depleted due to germination over time and rotted,or died. Survivorship curves for two years confirms that T.pratensis seed banks are short-lived.E. The formation of a seed bankSeed banks are ecologically and evolutionarily important inthe dynamics of plant populations (Kalisz, 1991). The resultsreported in this study show that T. pratensis seed bank isprimarily short-lived. Seed which did not germinate in theautumn did so in the spring. Thus, the seed bank of T. pratensisfalls into the "transient" category described by Thompson and105Grime (1979). High seed production, therefore, is expected to bean important characteristic for the success of the T. pratensis.Biennial plants grow vegetatively in the first year, andproduce seeds and die in the following year. If T. pratensis isa short-lived perennial, by delaying its reproduction beyond thesecond year, it must have an even higher seed output tocompensate for the additional adult plant mortality and longergeneration time. Even though it takes at least two years for T.pratensis to produce seeds, high seed production can ensure alarge transient seed bank in the soil. This may be the strategyexplaining the dominance of T. pratensis in the rangelands ofB.C. The transient seed bank appears to facilitate seedlingcolonization of gaps created by plant mortality in thevegetation (Grime et al., 1981). In other words, seeds in T.pratensis germinate and establish in autumn when other plantsare dying or in spring when there is less competition with otherplants. Parker and co-workers (1989) also indicated that specieswith transient seed banks may compensate with alternativemechanisms, such as long-lived adult or persistent bulbs orrhizomes.A small number of weed seeds can be sufficient to infest afield in a short time (Sagar and Mortimer, 1976). Less than 3%of dormant T. pratensis seeds remained in soil after one year ofburial. Whether this small number of buried seeds has any impacton persistence of T. pratensis would require a long-terminvestigation.106In conclusion, the study of seed population dynamics showsthat T. pratensis seed banks are short-lived in nature. Theburial depth affects seed germination behaviour of this speciesonly during the first two to three months after initialdispersal. The seed populations become nearly completelydepleted after 9 to 10 months of burial. Less than 3% of buriedseeds remains viable after 13 months of burial. The within yeargermination schedules appeared to be determined by environmentalfactors such as temperature and precipitation pattern, resultingin seed germination flushes in autumn and spring. Thesurvivorship curves between 1989 and 1990 cohorts were similar.107Chapter 5. Population ecology ofTragopogon pratensis L. plantsAbstractTo determine whether Tragopogon pratensis relies onstrategies other than maintaining seed banks for persistence,patterns of survivorship and reproductive characteristics of itspopulations were monitored from May 1990 to October 1991. About50% of the plants marked in 1990 died before winter and anadditional 30% died during winter. There was no correlationbetween plant density and rate of mortality during the growingseason, i.e. the mortality was density-independent. Only 12% ofthe 1990 cohort remained in October 1991; none of the remainingplants flowered in the second growing season. This suggests thatT. pratensis is not a biennial species. Size of age-specificplants varied significantly suggesting variations in growth rateamong individuals.The percentage of flowering increased with increasing rootcrown diameter (RCD). Plants flowered over a range of RCD in twoyears of observation. The minimum RCD for plants that floweredwere 0.2 and 0.6 cm in 1990 and 1991, respectively. Seedproduction ranged between 100 to 800 seeds per plant, increasinglinearly with the plant size. High seed production and longvegetative survival of plants appear to be important strategiesin maintenance of T. pratensis populations.108IntroductionTo develop effective control measures, weed scientists mustunderstand how weeds become and remain abundant in nature.Successful weed control requires information on factorscontrolling weed population densities. The life historycharacteristics of a species (e.g. seedling and adult survival,growth, timing of flowering, and seed production) are componentsof evolutionary fitness of plants (Solbrig, 1980; Davy andSmith, 1988). Studies on dynamics of plant population providedemographic data which give an insight into how populationdensities change. Changes in demographic components of weedpopulations can predict future weed infestation levels byquantifying both the total number of individuals in eachfunctional stage of the life cycle and the rates of flux betweensuccessive stages (Solbrig, 1980; Silvertown, 1984; Navas,1991).Traqopogon pratensis has become abundant in B.C. rangelands.The factors controlling its population density are poorlyunderstood. If we assume that this weed completes its life cyclein two years as reported by Alex et al., (1980), its persistencein rangelands should be mainly determined either by highreproductive output or a persistent seed bank, or both. However,since the seed bank of T. pratensis was found to be transient inthis research (Chapter 3), other stages of its life cycle mustplay an important role in the persistence of this weed. The life109history and reproductive behaviour of T. pratensis should,therefore, be investigated to elucidate the possible causes ofits dominance.The size of individual plants varies significantly withinplant populations (Kachi and Hirose, 1983; Sarukhan et al.,1984; Weiner, 1988). This variation is extremely importantbecause size is correlated with both survivorship and fecundityand, therefore, with plant fitness (Gross and Werner, 1983;Mithen et al., 1984; Weiner and Solbrig, 1984; Weiner, 1988;Weiner and Whigham, 1988). A critical plant size is required forflowering by many biennials, e.g. teasal, wild carrot, ragwortand hound's-tongue (Werner, 1975; Van der Meijden and Van derWaals-Kooi, 1979; De Jong et al., 1989). The prolonged life-spandue to the requirement of critical plant size for flowering inweeds may, therefore, play an important role in theirpersistence. We need to know if T. pratensis plants require toreach a critical size before they flower.Skewness indicates the shape of the distribution of thegrowth parameters, namely long 'tails' to the left (skew < 0) orto the right (skew > 0) or symmetrically bell-shaped (skew = 0).For many years, skewness was interpreted as being a result ofcompetition. Hutchings (1986b) showed that skewness could,however, be generated by size-specific growth and death as wellas by competition. Non-destructive measures of plant size havebeen shown as an appropriate approach to monitor performance ofindividual plants over time (Weiner, 1988). Whether this110skewness can be generated by measuring plant size in T.pratensis over time is not known.The overall objective of this study was to determine ifpersistence of T. pratensis depends on its high seed production,low seedling mortality, and extended life cycle due to acritical size requirement for flowering. The specific objectiveswere to determine: 1) the patterns of survivorship and fecundityof T. pratensis; 2) the changes of size distribution in thepopulation over time; 3) if a critical plant size threshold forflowering occurs; and 4) the correlation between the plant sizeand seed output.Materials and MethodsA. Study siteThe study site was located at Riske Creek, Williams Lake,British Columbia (52 ° 01'N 122 ° 31'W, elevation 1006 m). Thissite was originally dominated by bluebunch wheatgrass (Agropyronspicatum) and Stipa spp. Now it is mainly dominated by T.pratensis. Bluebunch wheatgrass and some short grasses werescarcely scattered in the plant community. The soils aregravelly loam developed on till, and stones occur frequently(Agriculture Canada, 1988). The soil parent material is gravellyclay loam till, which was moderately calcareous. The soils arewell drained and pervious with a subhumid to semiarid moistureregime.111The study site was protected from animal grazing by fencing.Tragopogon pratensis started blooming in the second week of June(Fred Knezvich, 1990, Personal communication). The pasture hadbeen grazed for two weeks in spring (May 15 to June 1) and twoweeks in fall during the study period. The conditions of thestudy area were relatively homogenous.B. DefinitionsTo avoid misunderstanding of the terminology, definitions ofsome terms used in this chapter are given below:Age-specific populations consist of plants which emergeapproximately at the same time and the survival of theseplants is recorded by marking seedlings after emergence.Non-age-specific populations consist of plants which could beat least one year old and the timing of emergence forindividual plants in the population is not known.Flower head, the inflorescence (capitulum) of T. pratensis.Floret, a small flower, such as a disc or ray floret making upthe capitulum (flower head) in T. pratensis.Primary flowering, flowering for the first time.Secondary flowering, flowering following the removal of primaryflowers.C. Life history studyTenlmxlmquadrats in May 1990 and ten 0.5 m x 0.5 m112quadrats in May 1991 were randomly chosen and marked. Allage-specific T. pratensis seedlings (total of 642 in 1990 and481 in 1991) in each quadrat were labelled with numbered labelson tooth picks placed beside each seedling. The mortality andflowering behaviour were observed each month during the studyperiods. At the beginning and the end of each growing season,the plant size was estimated by measuring the root crowndiameter (RCD) with calipers for all surviving individuals. Thisnon-destructive measurement has been shown to be stronglycorrelated with the dry weight of the taproot in other species(De Jong et al., 1989; Powell, 1988).D. Flowering behaviour and seed productionTo determine the flowering behaviour of T. pratensis adultplants, separate quadrats were marked. Ten 0.5 m x 0.5 mquadrats were randomly chosen and permanently marked in a T.pratensis infested area in May, 1990 and 1991. Allnon-age-specific T. pratensis plants (total of 210 in 1990, 182in 1991) in each quadrat were labelled by fastening numberedplastic closures around their bases. The RCD of these plants wasalso measured. Monthly observations of flowering were takenthroughout the growing season.The number of flowering plants was recorded at each census.When seeds were formed, the seeds from individual plants werecollected and the number of seeds per plant was counted. BecauseT. pratensis seeds are formed and disperse asynchronously, the113flower heads produced on individual plants between two adjacentcensus were, therefore, collected and the number of florets ineach flower head was counted. Secondary flowering was alsoinvestigated.E. Data analysisData were analyzed using regression and ANOVA tests in theSYSTAT package (Wilkinson, 1990a). Data for plant size werelog-transformed and subjected to regression analyses. Thecalculations for some parameters such as qx and kx in life tablefollowed the procedures in Begon and Mortimer (1981). Theqx-values are the age-specific mortality rates and are goodmeasures of the intensity of mortality. qx is also equivalent to(1-px), where 1p' refers to the survival-probability, qx = 1-(ax+Vax). ax is population density observed at the time x. Thek-value (k i ) is defined as the difference between the logarithmof densities (N) before and after the measurements:ki= logNi - logN i+ ,in which i is an index of the life stage. The k values areusually defined as the mortality factor, which is largelyresponsible for changes in the total mortality in thepopulation.Mean, variance, and kurtosis also can be used tocharacterize distribution of plant size (Hutchings, 1986b).Kurtosis measures the degree to which the distribution is morepointy (positive values) or more flat-topped (negative values)than a normal distribution (kurtosis = 0).114ResultsA. Pattern of age-specific plant survivorshipa) Survivorship curvesThe mortality of T. pratensis seedlings in 1990 cohort wasabout 50% at the end of the first growing season and additional30% during the first winter (Fig. 1A). Only 12% of the plantsbecame established and survived prior to the second winter. Asimilar pattern of seedling mortality was also observed in the1991 cohort, only about 50% of the seedlings survived prior tothe first winter (Fig. 1B). A cohort life table was generated byfollowing the age-specific population of 1990 cohort for up totwo growing seasons (Table 1). The mortality rates of thepopulation were calculated over time and the k values were foundto be higher in the first year (0.026-0.555) than in the secondyear (0.006-0.052).b) Effect of density on mortalityMortality increased slightly with increasing plant densityfor observations during the periods of May 1990 to October 1991in 1990 cohort (Fig. 2A) and May to October 1991 in 1991 cohort(Fig. 2C). However, there was no significant correlation betweendensity and mortality at P > 0.05 (Figs. 2 A,B, and C) in either1990 or 1991 cohorts. Density had no significant effect onmortality of T. pratensis plants either at the seedling stage10May/199160 120lo '0Mgy/19901000A1990 cohortCO0>JcoOOJ1000c_) 100cr)0)O80^120I^I 180 240Time300 360 420 480 540(day) Vay/1991Time (day)Fig. 1. The survivorship curves ofsurvivorship of 1990 cohort forsurvivorship of 1991 cohort forTragopogon pratensis. A. Thetwo growing seasons; B. Theone growing season.180 240115116Table 1. Life table of Tragopogon pratensis L. at Riske creek,B.C. from 1990 to 1991.x ax lx dx qx logax kx0 642 1000 58 0.058 2.808 0.02630 605 942 78 0.083 2.782 0.03860 555 864 225 0.260 2.744 0.13190 410 639 86 0.135 2.613 0.063150 355 553 399 0.722 2.550 0.555365 99 154 17 0.110 1.996 0.052407 88 137 9 0.066 1.944 0.030433 82 128 2 0.016 1.914 0.006469 81 126 8 0.063 1.908 0.027518 76 118 1.881* x:^Age in days after emergence.ax : Number of plants alive at the beginning of each ageinterval.lx :^Standardized number of plants surviving at the startof each age interval.dx :^Standardized number dying between x and x+1.Average mortality rate per day during the intervalbeginning with day x.logax Log transformation of ax .kx :^Killing power, or rate of mortality during theinterval beginning with day x.qx:1000^20^40^60^80Density (plants/m2 )Fig. 2. The relationship between the plant density and seedlingmortality of Tragopogon pratensis populations. A. May 1990 toOctober 1991; B. May 1990 to October 1990; C. May 1991 toOctober 1991.120808040020020^40^60^80^100^120Density (plants/d)12010080bq— 60o 4020806020^40^60^80^100Density (plants/m 2 ):= 40020B118(Figs. 2 B,C) or the young adult stage (Fig. 2A).c) Changes of mortality rates over timeTwo peaks of T. pratensis mortality were observed during thestudy period (Fig. 3A). The first mortality peak occurred underthe conditions of high temperatures and low precipitation duringthe first growing season (Fig. 3B); the second mortality peakoccurred during the first winter when the temperature reachedbelow 0 C and the precipitation was high. The mortality ofestablished T. pratensis plants were very low during the secondgrowing season and did not appear to be influenced by changes inclimate (Fig. 3A).B. Size distribution of plant populationsAge-specific T. pratensis plants marked in the spring of1990 showed different size hierarchies (Figs. 4A,B). The newlyemerged plants from 1990 and 1991 cohorts were too small tomeasure at the end of the first growing season (< 0.1 cm).Therefore, the size measurements of age-specific plants in the1990 cohort were made in the second growing season. RCD ofplants in 1990 cohort ranged between 0.1 cm to 0.5 cm one yearafter emergence; over 40% of the plants had RCD less than 0.2 cm(Fig. 4A). The range of RCD expanded to between 0.1 cm to 0.7 cmprior to the second winter (Fig. 4B), the proportion of thesmallest plants in the population (RCD < 0.2 cm), however,remained approximately the same. Skewness increased with time0 80 0 6CODCO>x 0.40May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul AugI^1990^I^1991^iTime of observation^_ 15 ^1^1^I^1^1^I^1^t^1^L.^1^I^1^1 ^0May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug^I^1990^I^1991^IFig. 3. Age-specific mortality rates (q ) and killing power (ks )of Tragopocron pratensis population aNd temperature andprecipitation data during the study period. A. q and k xvalues; B. the mean temperature and precipitatiofi.119406Q>. 300c0)0- 20AB05012050100.1^0.2^0.3^0.4^0.5RCD (cm)406Q>, 300C0)0-2020)LL100 0.1^0.2^0.3^0.4^0.5RCD (cm)Fig. 4. Size (root crown diameter-RCD) distribution ofage-specific population over two growing seasons. A. 1990cohort measured in May 1991; B. 1990 cohort measured inOctober 1991.0.6^0.7121before the onset of flowering because the variation of plantsize also increased with time (Table 2).The size (RCD) distribution of T. pratensis plants innon-age-specific populations is shown in Fig. 5. Sizedistributions in 1990 and 1991 populations were found to be verysimilar (Figs. 5 A,C). Most plants (approximately 85% in bothyears) had RCD between 0.10 to 0.59 cm. A small proportion hadRCD greater than 0.6 cm. At the end of the second growingseason, the frequency distributions of both cohorts were lessskewed than those at the beginning because the size variationdecreased due to reproduction of larger plants (Table 3). Whenthe large plants of the populations flowered and died at the endof each growing season, the size distributions became muchflattened (Fig. 5B). Kurtosis decreased with time and skewnessalso went down (Table 3). However, the size distribution in 1991cohort was not so much flattened (Fig. 5D) because theproportions of mortality and flowering in this cohort were smallduring the observation period (Fig. 6).C. Reproduction of T. pratensis plants1) Effect of plant size on floweringThe proportion of flowered, non-flowered and dead plants ofnon-age-specific populations in 1990 and 1991 cohorts are shownin Fig. 6. The proportion of flowering plants in T. pratensis populations varied in both years. About 65 (31%) out of 210122Table 2. Summary statisticsage-specific plant populationcohort.of size (RCD) distribution ofof Tragopogon pratensis in 1990Parameters May/1991 Oct/1991N 99 78Mean 0.207 0.251Variance 0.012 0.029Skewness 0.868 0.922Kurtosis 0.074 -0.050GO 30A25 26bQ 20C 150)0-0 10LL66000.1 1.00.4^0.6^0.0^0.7^0.8^0.9RCD (cm)0.2^0.3 0.1 0.2 0.3 0.4 0.6 0.8 0.7 0.8 0.9RCD (cm)0.90.7 0.8ID02 0.3^0.4^0.5^0.6^0.7^0.8^0.9RCD (cm)B30OO-M 10LL0 ^0.10.2^0.3^0.4^0.5^0.6RCD (cm)00.130262P 20C 150)Cr0) 10LL6CFig. 5. Size (measured as RCD) distribution of non-age-specificplant populations in two years. A. 1990 cohort measured in May1990; B. 1990 cohort measured in October 1991; C. 1991 cohortmeasured in May 1991; D. 1991 cohort measured in October 1991.124Table 3. Summary statistics of size (ROD) distribution ofnon-age-specific plant populations of Tragopogon pratensis in1990 and 1991 cohorts.Parameters 1990 a 1990 b 1991 a 1991bN 211 78 181 153Mean 0.404 0.463 0.356 0.311Variance 0.041 0.045 0.051 0.035Skewness 1.026 0.404 1.284 1.042Kurtosis 0.796 -0.572 1.620 1.0171990 a: adult plants were first labled and measured in May 1990.1990 b: the same plants from May 1990 were measured in Oct.1991.1991 a: adult plants were first labled and measured in May 1991.1991 b: the same plants from May 1991 were measured in Oct.1991.1251990^1991Cohor t CohortFig. 6. Proportion of flowered, non-flowered, and dead plants in1990 and 1991 population cohorts of Traqopogon pratensis inOctober 1990 and 1991 respectively.126plants in non-age-specific population flowered in the field in1990 (Fig. 6); an additional 7.1% of remaining plants floweredin 1991. However, only 14 (7.7%) out of 182 plants flowered in1991 cohort (Fig. 6). Plants which died mostly had RCD < 0.2 cm.Approximately 30% of the non-age-specific plants labelled in1990 cohort remained vegetative after two growing seasons (Fig.6). Plants labelled in May did not all flower at the same timeand the flowering period lasted for over one month (data notshown).While plants in the non-age-specific populations floweredover a range of RCDs, there was a positive correlation betweenthe RCD and the frequency of flowering (Fig. 7). The minimum RCDfor T. pratensis plants that flowered was 0.2 cm in 1990 cohort(Fig. 7A) and flowering percentage for plants in this size rangewas only 3.9%. The flowering percentage then increased withincreasing RCD. While about 61% of plants with RCD of 0.5 cmflowered, all plants with RCD > 0.6 cm flowered (Fig. 7A). Asimilar relationship between RCD and flowering was also observedin the 1991 population (Fig. 7B), but the curve shifted to theright side of the distribution. Plants that flowered in the 1991cohort had RCD of 0.6 cm or higher. However, less than 20% ofplants flowered at RCD of 0.6 cm; almost 80% of the plantsflowered at RCD of 0.9 cm (Fig. 7B). Thus, the minimum RCDs forT. pratensis plants that flowered in 1990 and 1991 were 0.2 to0.6 cm, respectively.Removal of primary flowering heads for the estimation ofA1006.Q0)CCY)0w806040200) 60C0 40LL2012710000.1^0.2^0.3^0.4^0.5^0.6^0.7RCD (cm)B 0.1^0.2^0.3^0.4^0.5^0.6^0.7^0.8^0.9^1.1^1.2RCD (cm)Fig. 7. The relationship between RCD and the percentage ofTragopogon pratensis plants that flowered. A. The relationshipfor 1990 population; B. the relationship for 1991 population.800.8^0.8^1. 0128seed production induced secondary flowering in 1990 (Fig. 8).Similar results also were found in 1991 (data not shown).However, there was no correlation between the percentage ofsecondary flowering and the RCD. The secondary flowered T.pratensis plants did not produce any seeds prior to winter inthis study.b) Effect of RCD on seed productionSince no flowering plant was observed after monitoring anage-specific T. pratensis population for two growing seasons, afecundity table could not be generated.Since the seeds of T. pratensis do not mature and dispersesynchronously in the field, it was not possible to collect allseeds at the time of maturation. Therefore, total floret numberwas used as a measure of seed production potential. Thefrequency distribution of total florets per plant and totalflower heads per plant are shown in Fig. 9. Both were relativelynormally distributed.Total flower heads and florets per plant correlatedsignificantly with the RCD (Figs. 10 A,B). A significant linearcorrelation between the total floret number per plant and numberof flower heads per plant was also observed (Fig. 10C). Similarresults were observed in 1991 cohort (Fig. 11), except the totalnumber of flowering plants was much lower in the 1991 cohortthan in the 1990 cohort.129Primary flowering^Secondary flowering0.10.20.3"0 0.50.60CC 80 60 40^20^0^20^40Flowering (%)Fig. 8. Secondary flowering in Traqopogon pratensis upon removalof primary flowers.60^80 100112^3^4^5^6^7^8^9^10Number of flower heads/plant1302015C 100(D-OLL 5100 150 200 250 300 350 400 450 500 550 500 650 700 750 800Number of florets/plantFig. 9. The frequency distributions of flower heads per plant andflorets per plant in Tragopogon pratensis. A. Frequencydistribution of flower heads per plant; B. frequencydistribution of florets per plants.00•ioU-302520151310.4^0.5^0.8^0.7^0.8^0.e^1ROD (cm)0.3^0.4^0.5^0.8^0.7RCD (cm)1000800Zas-a--... 800co0= 40073I-2000^2^4^8^8Flower heads/plantFig. 10. Relationship between RCD and florets (A) and flower head(B) in Tragopogon pratensis in 1990; C. Correlation betweenfloret number and flower heads per plant.0io^'2132500400E'(U0..300?Da0-4-., 200SD_0LL10000^0.2^0.4^0.8^0.8^1^1.2^1.4RCD (cm)Fig. 11. Relationship between RCD and floret number in Tragonogonpratensis in 1991.133DiscussionA. Pattern of age-specific survivorshipThe survivorship curves show survival of a populationthrough time in nature. The survivorship of plants is affectedby mortality at different life stages, particularly during theseedling establishment, in many species (Kelly, 1989a). In thisstudy, the seedlings of T. pratensis were found to suffer a highmortality risk. The survival rates in 1990 cohort declined withage in the first year and then remained relatively stable in thesecond year (Fig. 1A). Similar pattern of seedling survivalrates was also found in 1991 cohort (Fig. 1B).Survivorship curves can be used to compare the lifehistories of different species (Hutchings, 1986a). Deevey (1947)established three models describing survivorship curves forannual, biennial and perennial plants. The Deevey type IIIsurvivorship pattern appears to be common in many herbaceousperennials (Sarukhan and Harper, 1973; Baskin and Baskin, 1979;Kelly, 1989a). The type III model indicates that young plantssuffer a heavy mortality risk; the mortality risk declines withage, and a few individuals will normally achieve longlife-spans. The survivorship curve in T. pratensis plantsappears to match the Deevey type III model. How long theremaining T. pratensis plants take to flower and finish theirlife-cycle could not be determined in this study. However, thefact that some plants remained without seed production beyond134two growing seasons indicates that these plants are short-livedperennials. It is, therefore, concluded that T. pratensis is nota true biennial species.Plant density plays an important role in regulation ofpopulation size in many species (Silvertown, 1984; Law, 1981).However, no significant correlation between the density andmortality of T. pratensis plants was found in this study (Fig.2), i.e., the mortality was density-independent. The resultsfrom this study suggest that the abundance of T. pratensis inrangeland may be determined by the combination of factors suchas environmental conditions, high seedling density, and highsurvival rate of the adult plants.Demographic data in the form of life tables can provideimportant means for quantifying the impact of biotic and abioticstress factors (Solbrig, 1980). The qx is the age-specificmortality rate. The k-factor analysis also indicates the stagesduring the life cycle at which mortality rates are the highest(Hutchings, 1986b). The k values are log transformed, i.e. theyare standardized figures and can be used to compare this studywith others. In this study, the observation that q and k valueswere the highest during the seedling stage suggests that thisstage is the most vulnerable period in life cycle of T.pratensis.Seasonal changes in temperature and precipitation candetermine the survivorship pattern of some plant populations(Law, 1981; Kelly, 1989a). Powell (1988) reported that seedling135mortality rates in Centaurea spp. were generally greater than90%. The seedlings of this species on dry rangeland devoted mostof their early growth to extending their taproots; in contrast,the mortality rates of established rosettes were less than 5%per year. The mortality rate in T. pratensis plants appears tobe affected by the environmental changes. In rangelands,fluctuating temperatures could be the key factor in influencingplant mortality. Dry, hot summers (Fig. 3B) may enhance themortality rates of newly emerged T. pratensis seedlings.Freezing temperature in the following winters could accelerateseedling mortality rates before these plants could adapt to theenvironmental conditions (Fig. 3). Once seedlings areestablished, seasonal changes would have little effect on plantmortality. Therefore, seedling establishment is crucial for thesuccess of T. pratensis.B. Size distribution and flowering behaviour1) Variation of size distributionSize variation within plant populations is extremelyimportant because it is correlated with both survivorship andfecundity (Weiner, 1988). Small plants have higher mortalityrisk than the large plants. Plant size could vary ratherindependently from chronological age due to variation of growthrates among individual plants (Sarukhan et al., 1984). Floweringtime within a plant population of some species can be predicted136from the size distribution when there is a critical sizerequirement for flowering (Gross and Werner, 1983; Sarukhan etal., 1984). Therefore, studies of size distribution ofindividual plants in a population can help understand thereproductive characteristics of some weeds.While seedling populations of some species have symmetricalsize distribution, a variety of factors such as competition,predation, and environmental conditions could combine to bringabout an asymmetrical positively skewed distribution of adultplant size within age-specific populations (Benjamin andHardwick, 1986; Hutchings, 1986b). Plants in age-specific T.pratensis populations showed a wide range of size distribution(RCD) and this distribution expanded over time (Fig. 4). Theplant size hierarchies in age-specific T. pratensis populationssuggest differential growth rates among individuals. Thisvariation in size also indicates that not all plants in thepopulation will finish their life-cycle in the same year if theplant size is a factor in determining the timing ofreproduction. The size distribution of Zizania aquatica has beenshown to change from high inequality and positive skewness tomore uniform and equal distribution (Weiner and Whigham, 1988).How size variation in T. pratensis changes when plantreproduction starts could not be determined in age-specificplants in this study.A wide range of sizes also appeared in the non-age-specificpopulations of T. pratensis, i.e. in plants which were137established at least one year before they were sampled (Fig. 5).As the plants flowered and died, mortality was concentrated inthe smaller size classes, so that the skewness decreased oncethe plants had begun to die (Table 3). This is because deaths ofthe smallest individuals truncated the left-hand end of thedistribution of plant sizes, whereas the rapid growth of thelargest plants extended the right-hand end of the distribution(Figs. 5 B,D).2) Critical size for floweringA critical size requirement for flowering means that a plantonly flowers when it reaches a certain size. Recent studies haveshown that some species classified as 'biennial' may take three,four, or even five years to flower in natural habitats (Holt,1972; Werner, 1975; Harper, 1977; Gross, 1981; De Jong et al.,1986; Klinkhamer et al., 1987; Powell, 1988) because theircritical size requirement for flowering. The distinction betweenthe 'biennials' and perennials has, therefore, become blurred(Silvertown, 1984). There are some strict biennial plants suchas Linum catharticum and Gentianella amarella which do alwaysflower in the second year (Kelly, 1989a). Why some speciesrequire a critical size before they can flower is notunderstood.Although a critical plant size requirement for flowering hasbeen suggested for many 'biennial' species: e.g. teasal, wildcarrot, ragwort and hound's-tongue (Werner, 1975; Van der138Meijden and Van der Waals-Kooi, 1979; De Jong et al., 1989), thecriteria used for critical size in these studies were notclearly defined. Some researchers define critical size as thesize at which plants start to flower. The probability offlowering at the critical size in these cases were, however,very low. On the other hand, the probability of flowering wassignificantly high at the critical size defined in otherstudies. For example, the rosette diameter of Senecio iacobaea between 2-4 cm was considered to be the critical size, althoughthe probability of flowering in this size range was less than0.05; the probability of flowering increased with plant size andall plants flowered at rosette diameter of > 10 cm (Van derMeijden and Van der Waals-Kooi, 1979). Kachi and Hirose (1983)reported that the rosette size of 10 cm was critical when theprobability of bolting for Oenothera erythrosepala plants wasless than 0.2; the probability reached > 0.8 at rosette diameterof 20 cm. Werner (1975) reported that the probability offlowering in teasal plants was > 0.8 when rosettes reached acritical size of 30 cm diameter; some small teasal plants,however, also flowered, but the probability of flowering wasbetween 0.01 and 0.32.The evidence described above suggests that critical size forflowering varies between species. Where to draw the line for thecritical size in the distribution of plant size depends on howsignificant the plant size is in determining flowering (i.e. theprobability of flowering). The probability of flowering at139critical size should be significantly high. Rosette sizes of 2-4cm for Senecio iacobaea and 10 cm for Oenothera erythroseiala are, therefore, not acceptable critical sizes because of the lowprobability of flowering. There is little discussion inliterature on the effects of flowering of small plants on theevolutionary fitness of plants, and of seed production by smallplants on persistence of weeds.3) Is there a critical size requirement for flowering in T.Pratensis?The T. pratensis plant size at which significant floweringoccurred differed in the two years of this study. The plantflowered over a range of plant size (RCD) and the sizepositively affected the percentage of flowering in T. pratensis(Fig. 7). In 1990, some plants started to flower at RCD of 0.2cm and all plants flowered when RCD reached > 0.6 cm. In 1991,however, plants started to flower at RCD of 0.6 cm and almost80% of the plants flowered at 0.9 cm RCD. These results suggestthat T. pratensis does not have a critical size requirement forflowering.What regulates the timing of flowering in T. pratensis plants seems to be a complex issue. In a separate greenhousestudy, some Tragopogon plants which had RCD > 2 cm did notproduce flowers (M.Q. Qi, unpublished results). This furthersuggests that the plant size is not the only factor affectingflowering of Tragopogon species. Environmental factors and140genetic variability could be important in regulation offlowering in this species.Many species require vernalization and/or a long photoperiodfor induction of flowering (Schwabe, 1954; Gross, 1981; Kachiand Hirose, 1983). For example, in a study involving fourbiennial species, Gross (1981) reported that only few Daucus carota rosettes flowered in the greenhouse. Oenothera, Verbascumand Tragopogon dubius rosettes, which were above the criticalsize for flowering, did not bolt under greenhouse conditionssuggesting a vernalization requirement for bolting. A highpercentage of flowering among individuals in Aster pulosus wasfound in nutrient-rich field soil; plants growing on a sandysoil exhibited a lower percentage of bolting (Peterson andBazzaz, 1978). Lacey (1986) indicated that the critical size forflowering in some biennial species may be environmentallydetermined since it differs with nutrient treatments. Baskin andBaskin (1979) also explained that the variation of nutrientsupply in soils could be the factor determining flowering over awide range of sizes in some species. Whether the timing offlowering in T. pratensis plants is determined by vernalization,photoperiod, and nutrient supply in soil, and whether plant sizeinteracts with one or more of these factors need furtherinvestigation.4) Secondary flowering in T. pratensisSecondary flowering may play an important role in141persistence of a species by allowing it to adapt to frequentdisturbance caused by grazing in rangelands. T. pratensis plantswere found to produce secondary flowers upon removal of theprimary flowers (Fig. 8). The secondary flowers, however, didnot produce any seed prior to winter in this study. The timingof removal of primary flowers may be critical for the secondaryflowers to have enough time to produce seeds (De Jong et al.,1990; Van Mierlo and Van Groenendael, 1991). For example, seedproduction of C. officinale was significantly reduced by acomplete removal of the flowering stems at the end of May (DeJong et al., 1990). On the other hand, the seed production ofAnthriscus sylvestris L. was promoted by removal of primaryflowering stems when they had just formed (Van Mierlo and VanGroenendael, 1991). Apparently, primary flowers were not removedearly enough in this study to give secondary flowers enough timeto form seeds prior to winter which killed secondary flowers.Whether seed production by secondary flowers in T. pratensis isinfluenced by the timing of removal of flowers in B.C.conditions needs to be determined.C. Effect of plant size on seed productionThe relative fitness of a plant species depends on itsreproductive ability, which is closely linked to plant size(Kelly, 1984, 1989b; Lee, 1988). Positive relationships betweenplant size, survivorship and fecundity have been reported. Thenumber of flower heads per plant and number of seeds per flower142head are two important components in determining seed productionin members of Compositae (Kelly, 1984; Schmitt, 1983). Both thenumber of florets per plant and number of flower heads per plantincreased with plant size, and number of florets correlatedpositively with number of flower heads in T. Dratensis in thisstudy (Figs. 10, 11). The small individuals seemingly have lessreserves to survive and less energy to flower. The resultsindicate that flowering of larger individuals may beadvantageous and may at least in part compensate for the highjuvenile mortality and longer generation time in this species.D. Significance of delayed seed productionDelay of flowering may play an important role in dominanceof a plant population because it results in a higher seedproduction for some species (Van der Meijden and Van derWaals-Kooi, 1979). Presence of a bud bank has the same impact asthe seed bank on the population persistence (De Jong andKlinkhamer, 1988), i.e. plants that remain vegetative over anumber of years can buffer against little or no seedreproduction in some years in some species. This study indicatesthat the flowering percentage of T. pratensis in 1990 cohort(31%) was higher than that in 1991 cohort (7.7%). The facts thatseed production in T. pratensis varies between years and thatseed bank of this weed is "transient" suggest that analternative strategy may be required to maintain the populationdensity in some years. It is, therefore, concluded that delayed143seed production in T. pratensis could be beneficial in thedominance of this weed by increasing the size of the bud bank.In conclusion, the facts that no seeds were produced priorto the second winter in age-specific T. pratensis populationsand that the plant died after seed production innon-age-specific populations suggest that T. pratensis is amonocarpic perennial, and not a biennial weed. This weed doesnot require a critical plant size for flowering. The resultsalso suggest that plant size and environmental conditions may beinteracting in determining the timing of flowering. The highseed production and high vegetative survivorship could be thekey strategies in maintaining the dominance of T. pratensis innature.144General conclusionsThe objectives outlined in the general introduction wereachieved by systematic studies carried out in this thesis. Theinformation obtained reveals the mechanism regulating seedgermination in C. officinale. Ecophysiological characteristicsof Tragopogon spp. seeds and ecological characteristics of T.pratensis plants established in this research provideexplanations for persistence of these weeds in rangelands.A. Cynoglossum officinaleCynoglossum officinale reproduces only by seeds and theseeds show a deep innate dormancy. Removal of the seed coatresults in nearly complete germination of innately dormantseeds. The seed coat does not prevent water uptake nor does itcontain any water soluble germination inhibitor. However, theseed coat inhibits 02 uptake. The seed coat also prevents theleaching of phenolic substances, but this leaching is notnecessary for seed germination. Seeds of C. officinale containhigh levels of phenolic substances and seed extracts show highpolyphenol oxidase activity. Rosmarinic acid, the most prominentphenolic in C. officinale seeds, does not inhibit thegermination of the decoated seeds at concentration up to 3 mM.No significant quantitative or qualitative correlation could beestablished between changes in specific phenolic compounds andseed germination induced by stratification or seed coat removal.145It is concluded from this study that C. officinale seed coatsregulate seed germination by controlling 0 2 availability to theembryo.This work does not rule out the possibilities that seed coatmay prevent seed germination by restricting the embryo growthmechanically or by containing germination inhibitors on sideadjacent to the embryo; washing of seeds will not remove theseinhibitors. The possibility that seed coat may inhibit embryogermination in C. officinale seeds by preventing red light fromreaching the embryo also cannot be ruled out. Seed germinationrate and root growth of Agropyron spicatum, a dominant grass inB.C. rangelands, were inhibited by leachates of C. officinale seeds (Upadhyaya and Furness, unpublished data). More studiesare needed to investigate the allelopathic effect of C.officinale during various developmental stages on other plants.B. Tragopogon pratensis and T. dubius Several strategies were found to be involved in persistenceof Tragopogon species. Development of an effective weed controlmeasure must consider these strategies.Tragopogon pratensis and T. dubius exhibited secondarydormancy when subjected to anaerobiosis treatment. They may relyon this induced dormancy for their persistence. The induceddormancy is released by air drying or stratification (at 5 C).This stratification requirement would prevent germination ofdormant T. pratensis seeds in the fall thereby avoiding high146seedling mortality in the winter. Complete germination ofnon-dormant seeds of these weeds buried in soil and nolight-requirement for seed germination suggest that Tragopogonseed populations in soil would become completely depleted underconditions conducive for seed germination regardless of theirseed distribution in soil profile.Tragopogon pratensis does not maintain a large persistentseed bank. The survivorship curves between 1989 and 1990 cohortsare not significantly different as indicated by the Logranktest. Two germination peaks, one in autumn after burial and theother in the following spring, suggest the timing of seedgermination is affected by the climatic changes within eachyear. With a short-lived seed bank, the persistence of T.pratensis in rangelands may depend on its high seed production.Tragopogon pratensis is not a biennial species. Highseedling mortality is observed during the seedling stage and themortality is density-independent. The remaining plants from anage-specific population showed no flowering for two growingseasons. Size of the age-specific plants varied significantlysuggesting variations in growth rate among individuals. Thepercentage of flowering increased with the root crown diameter(RCD). Plants in the non-age-specific populations flowered overa range of RCD in two years of observation. The minimum RCD forplants that flowered were 0.2 and 0.6 cm in 1990 and 1991,respectively. This argues against a critical size requirementfor flowering in T. pratensis. Seed production increased147linearly with the plant size. High seed production and longvegetative survival of plants appear to be important strategiesin maintenance of T. pratensis populations.Multiple persistence strategies, characteristics of manysuccessful weedy species, can enlarge the range of growingconditions (Grime, 1979). The success of T. pratensis may dependon its secondary seed dormancy, high reproductive output, anddelayed life-cycle. Based on the theory of r and K selectionproposed by MacArthur and Wilson (1967), Grime (1979) indicatedthat plants have evolved three primary strategies (ruderal,stress tolerant, and competitive) in response to variousenvironmental conditions. Although the ruderal strategy is oftenselected by weedy species due to the characteristic of severelydisturbed but potentially productive habitat, T. pratensis appears to be able to establish in both disturbed and productivehabitats. A long-term study to compare persistence strategies ofTragopogon species between different habitats (e.g. from baredto thick grass cover) is necessary to explore the possibilitiesof their future infestation in rangelands.Development of an effective control method usually takes along time and a lot of energy. A biological control program forknapweeds (Centaurea spp.) in B.C. is a good example. Diffuseknapweed is another common weed in southern B.C. rangelands. Itproduces numerous seeds per plant and the dormant seeds form apersistent seed bank (Nolan, 1989; Davis, 1990). The life-cycleof this weed takes either two or more years to complete because148its plants require a critical rosette size for flowering(Powell, 1988). Two tephritid flies, Urophora affinis and U.quadrifasciata, and buprestid beetle (Sphenoptera jugoslavica)were introduced in 1970's. The three biocontrol agents controlthe knapweeds effectively by reducing the seed production aswell as reducing the survivorship of seedlings and rosettes(Harris and Myers, 1984; Powell, 1988). In comparison withdiffuse knapweed, the seed bank of T. pratensis is short-lived.High seed production and long vegetative survival of plantsappear to be important strategies in maintenance of T. pratensispopulations. The ideal biocontrol agents to control T. pratensis should be, therefore, selected to reduce the survivorship ofplants and finally reduce seed production.Tragopogon dubius is a favored dietary item for pocketgophers; the rodents may consume 20 to 80% of the primary roots(Reichman and Smith, 1991). However, soil disturbance caused bythe rodents is of concern. Aboveground parts of the plants mayalso be subjected to herbivory by several vertebrates (deer,squirrels, and rabbits) which consume the flowering stalks.Innovative utilization of selective herbivory combined withcarefully monitored grazing systems on Tragopogon spp. infestedsites may enhance successful management of these commonrangeland weeds.Although development of biological control by usingphytophogeous organisms could be an effective long-term controlmeasure, an integrated chemical and cultural control on149Tragopogon species will reduce their further infestation. Mowingor controlled fire after bolting but before seed setting caneffectively be used to reduce Tragopogon seed production.Seedling destruction should be followed after emergence in bothfall and spring by using some residual herbicides to reducepopulation density of these weeds. Picloram has been found toprovide good to excellent control of T. pratensis plants on B.C.rangelands over a four-year period (Cranston et al., 1986.).150ReferencesAbdel-Samad, I.M. and Pearce, R.S. 1978. Leaching of ions,organic molecules, and enzymes from seeds of peanuts (Arachishypogea L.) imbibing without testas or with intact testas. J.Exp. Bot. 29: 1471-1478.Agriculture Canada, 1988. Soils of the Williams Lake-Alexis Creekarea, British Columbia. Report No. 53, British Columbia soilsurvey.Alex, J.F., Cayouette, R., and Mulligan, G.A. 1980. Common andbotanical names of weeds in Canada. Agric. Can., Ottawa, Ont.Publ. 1397, 132pp.Alex, J.F. and Switzer, C.M. 1976. Ontario weeds. Ont. Min.Agric. and Food, Toronto, Ont. Publ. 505, 200pp.Ballard, L.A. 1973. Physical barriers to germination. Seed Sci.and Technol. 1: 285-303.Bartik, M. and Pistac, A. 1981. Veterinary toxicology. ElsevierScientific Publ. Co., Amesterdam.Baskin, J.M. and Baskin, C.C. 1979. Studies on the autecologyand population biology of the monocarpic perennial Grindelialanceolata. Am. Midl. Nat. 102: 290-299.Baskin, J.M. and Baskin, C.C. 1983. Seasonal changes in thegermination responses of seeds of Veronica peregrina duringburial, and ecological implications. Can. J. Bot. 61:3332-3336.Baskin, J.M. and Baskin, C.C. 1984. Effect of temperature duringburial on dormant and non-dormant seeds of Lamiumamplexicaule L. and ecological implications. Weed Res. 24:333-339.Baskin, J.M. and Baskin, C.C. 1985. The annual dormancy cycle inburied weed seeds: A continuum. BioSci. 35: 492-498.Baskin, J.M. and Baskin, C.C. 1987. Temperature requirements forafter-ripening in buried seeds of four summer annual weeds.Weed Res. 27: 385-389.Baskin, J.M. and Baskin, C.C. 1989a. Physiology of dormancy andgermination in relation to seed bank ecology. In Ecology ofsoil seed banks, eds. Leck, M.A., Parker, V.T., and Simpson,R.L. Academic Press, INC.Baskin, J.M. and Baskin, C.C. 1989b. Germination responses of151buried seeds of Capsella bursa-pastoris exposed to seasonaltemperature changes. Weed Res. 29: 205-212.Baskin, J.M. and Baskin, C.C. 1990. Seed germination ecology ofpoison hemlock, Conium maculatum. Can. J. Bot. 68: 2018-2024.Begon, M. and Mortimer, M. 1981. Population ecology. A unifiedstudy of animals and plants. pp. 9-15. Sinauer Ass. Inc.Publ., Sunderland, Massachusetts.Benjamin, C.R. and Hardwick, R.C. 1986. Sources of variation andmeasures of variability in even-aged stands of plants. Ann.Bot. 58: 757-778.Bewley, J.D. and Black, M. 1982. Physiology and biochemistry ofseeds (in relation to germination). Vol. II. Viability,dormancy, and environmental control. Springer-Verlag. BerlinHeidelburg. New York.Bewley, J.D. and Black, M. 1985. Seeds: Physiology of developmentand germination. Plenum Press, New York & London.Blackshaw, R.E. 1990. Influence of temperature, soil moisture,and seed buried depth on the emergence of round-leaved mallow(Malva pusila). Weed Sci. 38:518-521.Boorman, L.A. and Fuller, R.M. 1984. The comparative ecology oftwo sand dune biennials: Lactuca virosa L. and Cynoglossumofficinale L. New Phytol. 96: 609-629.Bouwmeester, H.J. and Karssen, C.M. 1989. Environmental factorsinfluencing the expression of dormancy patterns in weed seedsAnn. Bot. 63: 113-120.Bradford, M.M. 1976. A rapid and sensitive method for thequantitation of microgram quantities of protein utilizing theprinciple of protein-dye binding. Anal. Biochem. 72: 248-254.Cavers, P.B. 1983. Seed demography. Can J. Bot. 61: 3578-3590.Cavers, P.B. and Benoit, D.L. 1989. Seed banks in arable land. InEcology of soil seed banks. eds. Leck, M.A., Parker, V.T.,and Simpson, R.L. pp. 309-328. Academic Press, INC.Chancellor, R.J. 1982. Dormancy in weed seeds. Outlook Agric. 11:87-90.Colosi, J.C., Cavers, P.B., and Bough, M.A. 1988. Dormancy andsurvival in buried seeds of proso millet (Panicum miliaceum).Can. J. Bot. 66: 161-168.152Come, D. and Tissaour, T. 1973. Interrelated effects ofimbibition, temperature and oxygen on seed germination. InSeed physiology. ed. Heydecker, W. pp. 157-168. Butterworth,London.Cook, R.E. 1979. Patterns of juvenile mortality and recruitmentin plants. In Topics in plant population biology. eds.Solbrig, O.T., Jain, S., Johnson, G.B., and Raven, P.H. pp.207-231, Columbia University Press, New York.Cook, R.E. 1980. The biology of seeds in soil. In Demography andevolution in plant populations. ed. Solbrig, O.T. pp.107-130.Blackwell Sci. Publ., Oxford, London.Corbineau, F., Lecat, S. and Come, D. 1986. Dormancy of threecultivars of oat seeds (Avena sativa L.). Seed Sci. andTechnol. 14, 725-735.Coumans, M., Come, D. and Gaspar, T. 1976. Stabilized dormancy insugarbeet fruits. I. Seed coats as a physiological barrier tooxygen. Bot. Gaz. 137: 274-278.Cranston, R.S. and Pethybrige, J.L. 1986. Report on hound's-tongue (Cynoglossum officinale) in British Columbia. B.C.Ministry of Agriculture and Forestry, Internal Rept. 8pp.Cranston, R.S., Woods, J.A., and Blumenauer, D. 1986. Meadowgoat's-beard control on rangeland. Expert Committee on Weeds-Western Canada, 1986 Research Report. 3: 69.Cresswell, E.G. and Grime, J.P. 1981. Interaction of a lightrequirement during seed development and its ecologicalconsequences. Nat. 291: 583-585.Darlington, H.G. 1931. The 50-year period for Dr. Beal's seedviability experiment. Amer. J. Bot. 18: 262-265.Davis, E.S. 1990. Spotted knapweed (Centaurea maculosa L.) seedlongevity, chemical control and seed morphology. MSc thesis,Montana State University, Bozeman, Montana, USA.Davy, A.J. and Smith, H. 1988. Life-history variation andenvironment. In Plant population ecology. pp. 1-22. eds.Davy, A.J., Hutchings, M.J. and Watkinson, A.R. BlackwellSci. Publ. Oxford.Deevey, E.S. 1947. Life tables for natural populations ofanimals. Q. Rev. Biol. 22: 283-314.De Jong, T.J. and Klinkhamer, P.G.L. 1988. Population ecology ofthe biennials Circium vulgare and Cynoglossum officinale in acoastal sand-dune area. J. Ecol. 76: 366-382.153De Jong, T.J. and Klinkhamer, P.G.L. 1989. Limiting factors forseed production in Cynoglossum officinale. Oecol. 80:167-172.De Jong, T.J., Klinkhamer, P.G.L., and Prins, A.H. 1986.Flowering behaviour of the monocarpic perennial Cynoglossumofficinale L. New Phytol. 103: 219-229.De Jong, T.J., Klinkhamer, P.G.L., Nell, H.W., and Troelstra,S.R. 1987. Growth and nutrient accumulation of the biennialsCirsium vulgare and Cynoglossum officinale undernutrient-rich conditions. Oikos 48: 62-72.De Jong, T.J., Klinkhamer, P.G.L., Geritz, S.A.H., and Van derMeijden, E. 1989. Why biennials delay flowering: anoptimization model and field data on Cirsium vulgare andCynoglossum officinale. Acta Bot. Neerl. 38: 41-55.De Jong, T.J., Klinkhamer, P.G.L., and Boorman, L.A. 1990.Biological flora of the British Isles. No. 170. J. Ecol. 78:1123-1144.Demos, E.K., Woolwine, M., Wilson, R.H., and McMillan, C. 1975.The effects of ten phenolic compounds on hypocotyl growth andmitochondrial metabolism of mung bean. Amer. J. Bot. 62:97-102.Dickerson, J.R. and Fay, P.K. 1982. Biology and control ofhound's-tongue (Cynoglossum officinale). Proc. Western Soc.Weed Sci. 35: 83-85.Duke, S.H. and Kakefuda, G. 1981. Role of testa in preventingcellular rupture during imbibition of legume seeds. PlantPhysiol. 67: 449-456.Duvel, J.W.T. 1902. Seeds buried in soil. Sci. (New York) 17:872-873.Eberlein, C.V. 1987. Germination of Sorghum almum seeds andlongevity in soil. Weed Sci. 35: 796-801.Edwards, M.M. 1968. Oxygen and coat effects in Sinapis arvensis.J. Exp. Bot. 20: 874-894.Egley, G.H. 1986. Stimulation of weed seed germination in soil.Rev. Weed Sci. 2: 67-89.Egley, G.H. and Chandler, J.M. 1983. Longevity of weed seedsafter 5.5 years in the Stoneville 50-year buried-seed study.Weed Sci. 31: 264-270.154Egley, G.H. and Duke, S.O. 1985. Physiology and weed seeddormancy and germination. In Weed Physiology. Vol. I. Reproduction and ecophysiologv. ed. Duke, S.O. pp. 27-64. CRCpress, Inc. Boca Raton, Florida.Einhellig, F.A. and Rasmussen, J.A. 1979. Effects of threephenolic acids on chlorophyll content and growth of soybeanand grain sorghum seedlings. J. Chem. Ecol. 5: 815-824.Enu-Kwesi, L. and Dumbroff, E.B. 1980. Changes in phenolicinhibitors in seeds of Acer saccharum during stratification.J. Exp. Bot. 31: 425-436.Environment Canada, 1982. Canadian climate normal 1951-1980.British Columbia temperature and precipitation.Fenner, M. 1985. Seed ecology. London, New York, Chapman andHall.Ferraris, L., Abbattista, G., and Matta, A. 1987. Variations ofphenols concentration as a consequence of stress that induceresistance to Fusarium wilt of tomato. Z. PflanzenkrPflanzenschutz 94: 624-629.Flentje, N.T. and Saksena, H.K. 1964. Pre-emergence rotting ofpeas in South Australia. III. Host pathogen interaction.Aust. J. Biol. Sci. 17: 665-675.Frankland, B. and Wareing, P.F. 1962. Changes in endogenousgibberellins in relation to chilling of dormant seeds. Nat.194: 313-314.Frankton, C. and Mulligan, G.A. 1970. Weeds of Canada. Can. Dept.Agric., Ottawa, Ont., Publ. 948., 217pp.Froud-Williams, R.J., Chancellor, R.J., and Drennan, D.S.H. 1984.The effects of seed burial and soil disturbance on emergenceand survival of arable weeds in relation to minimalcultivation. J. Appl. Ecol. 21: 629-641.Garwood, N.C. 1989. Tropical soil seed banks: A review. InEcology of soil seed banks. ed. Leck, M.A., Parker, V.T., andSimpson, R.L. pp. 149-210. Academic Press, INC.Grace, J.B. 1984. Effects of tubificid worms on the germinationand establishment of Tvpha latifolia. Ecol. 75: 321-331.Greatorex, J.C. 1966. Some unusual cases of plant poisoning inanimals. Vet. Rec. 78: 725-727.Grime, J.P. 1979. Plant strategies and vegetative processes.Wileg, Chichester. New York.155Grime, J.P., Mason, G., Curtis, A.V., Rosman, J., Band, S.R.,Mowforth, M.A.G., Neal, A.M., and Shaw, S. 1981. Acomparative study of germination characteristics in a localflora. J. Ecol. 69: 1017-1059.Gross, K.L. 1981. Prediction of fate from rosette size in four"biennial" plant species: Verbacum thapsus, Oenothera biennis, Daucus carota, and Tragobogon dubius. Oecol.(Berlin) 48: 209-213.Gross, K.L. 1984. Effects of seed size and growth form onseedling establishment of six monocarpic perennial plants. J.Ecol. 72: 369-387.Gross, K.L. and Werner, P.A. 1982. Colonizing abilities of"biennial" plant specis in relation to ground cover:implication for their distributions in a successional sere.Ecol. 63: 921-931.Gross, R.S. and Werner, P.A. 1983. Probabilities of survival andreproduction relative to rosette size in the common burdock.(Arctium minus: Compositae). Ame. Midl. Nat. 109: 185-193.Hamilton, D.F. and Carpenter, P.L. 1975. Regulation of seeddormancy in Elaeagnus umbellata by endogenous growthsubstances. Can. J. Bot. 53: 2303-2311.Hamilton, D.F. and Carpenter, P.L. 1976. Regulation of seeddormancy in Elaeagnus angustifolia by endogenous growthsubstances. Can. J. Bot. 54: 1068-1073.Harper, J.L. 1959. The ecological significance of dormancy andits importance in weed control. Proc. of the 6th Int. Cong.of Crop Protection, Hamburg 1957, Vol.I pp. 415-420.Harper, J.L. 1977. Population biology of plants. Academic Press,London.Harper, J.L. and Ogden J. 1970. The reproductive strategy ofhigher plants. I. The concept of strategy with specialreference to Senecio vulgaris L. J. Ecol. 59: 681-698.Harper, J.L. and White, J. 1970. The dynamics of plantpopulations. Proc. Adv. Study Inst. Dynamics Numbers Popul.pp. 41-63.Harper, J.L. and White 1974. The demography of plants. A. Rev.Ecol. Syst. 5: 419-463.Harris, P. and Myers, J.H. 1984. Centaurea diffusa L. and C.maculosa L. s. lat., diffuse and spotted knapweed156(Compositae). In Biological Control Programmes Against Insects and Weeds in Canada 1969-1980. eds. Kelleher, J.S.and Hulme, M.A., Commonwealth Agricultural Bureaux, pp.127-138.Hegarty, T.W. 1973. Temperature relations of germination. In Seedecology. ed. Heydecker, W. pp. 411-432. London: Butterworth.Hermanutz, L.A. and Weaver, S.E. 1991. Variability intemperature-dependent germination in eastern black nightshade(Solanum ptycanthum). Can. J. Bot. 68:1463-1470.Holt, B.R. 1972. Effect of arrival time on recruitment,mortality, and reproduction in successional plantpopulations. Ecol. 53:668-673.Hutchings, M.J. 1986a. Plant population biology. In Methods inplant ecology. eds. Moore, P.D. and Chapman, S.B. pp.377-435. Blackwell Sci. Publ. Oxford.Hutchings, M.J. 1986b. The structure of plant populations. InPlant ecology. ed. Crawley, M.J. pp. 97-136. Blackwell Sci.Publ. Oxford.Hutchings, M.J., Booth, K.D., and Waite, S. 1991. Comparison ofsurvivorship by the Logrank test: criticisms andalternatives. Ecol. 72: 2290-2293.Kachi, N. and Hirose, T. 1983. Bolting induction in Oenotheraerythrosepala Borbas in relation to rosette size,vernalization, and photoperiod. Oecol. (Berlin) 60: 6-9.Kalisz, S. 1991. Experimental determination of seed bank agestructure in the winter annual Collinsia verna. Ecol. 72:575-585.Karssen, C.M. 1982. Seasonal patterns of dormancy in weed seed.In The physiology and biochemistry of seed development. dormancy and germination. ed. Khan, A.A. pp. 243-267.Elservier, New York.Kelly, D. 1984. Seeds per fruit as a function of fruits per plantin "depauperate" annuals and biennials. New Phytol.96:103-114.Kelly, D. 1985. On strict and facultative biennials. Oecol. 67:292-294.Kelly, D. 1989a. Demography of short-lived plants in chalkgrassland. I. Life cycle variation in annuals and strictbiennials. J. Ecol. 77: 747-769.157Kelly, D. 1989b. Demography of short-lived plants in chalkgrassland. II. Control of mortality and fecundity. J. Ecol.77: 770-784.Khan, M.A. and Ungar, I.A. 1986. Inhibition of germination inAtriplex triangularis seeds by applications of phenols andreversal of inhibition by growth regulators. Bot. Gaz. 147:148-151.Kim, Y-S. and Kil, B-S. 1989. Identification and growthinhibition of phytotoxic substances from tomato plant. KoreanJ. Bot. 32: 41-49.Kivilaan, A. and Bandurski, R.S. 1973. The ninety year period orDr. W.J. Beal's seed viability experiment. Am. J. Bot. 68:1290-1292.Klinkhamer, P.G.L., De Jong, T.J., and Meelis, E. 1987. Delay offlowering in the "biennials" Cirsium vulgare: size effectsand devernalization. Oikos 49: 303-308.Knight, A.P., Kimerling, C.V., Stermitz, F.R., and Roby, M.R.1984. Cvnoglossum officinale (hound's-tongue) - A cause ofpyrrolizidine alkaloid poisoning in horses. J. Am. Vet. Med.Assoc. 184: 647-650.Koller, D. 1964. The survival value of germination-regulatingmechanisms in the field. Herbage Abstracts 34:1-7.Kozolowski, T.T. 1972. Seed biology. Vol. II. Germinationcontrol, metabolism, and pathology. Academic Press. New Yorkand London.Krajina, V.J. 1965. Biogeoclimatic zones and biogeocoenoses ofBritish Columbia. In Ecology of western north America. Vol. I. ed. Krajina, V.J. pp. 1-17.Lacey, E.P. 1986. The genetic and environmental control ofreproductive timing in a short-lived monocarpic speciesDaucua carota (Umbelliferae). J. Ecol. 74:73-86.Larson, L.A. 1968. The effect soaking pea seeds with or withoutseed coats has on seedling growth. Plant Physiol. 43:255-269.Law, R. 1979. Ecological determinants in the evolution of lifehistories. In Population dynamics. eds. Anderson, R.M.,Turner, B.D. and Taylor, L.R. pp. 81-103, Symp. Brit. Ecol.Soc. 20 Blackwell Scientific Publ. , Oxford.Law, R. 1981. The dynamics of a colonizing population of Poaannua. Ecol. 62: 1267-1277.158Lee, T.D. 1988. Patterns of fruit and seed production. In Plant reproductive ecology. patterns and strategies. eds. LovettDoust, J. and Lovett Doust, L. pp. 179-202. Oxford universitypress, Oxford and New York.Lenoir, C., Corbineau, F. and Come, D. 1986. Barley (Hordeumvulgare) seed dormancy as related to glumellacharacteristics. Physiol. Plant. 68: 301-307.Lhostka, M. 1982. Germination ecology and diasporology of theCzechoslovak representatives of the genus Cynoglossum. FoliaGeobotanica et Phytotaxonomica 17:269-293.MacArthur, R.H. and Wilson, E.O. 1967. The theory of islandbiogeography. Princeton Press University, Princeton, N.J.Mandryka, I.I. 1979. Cynoglossum officinale (hound's-tongue) as apoisonous plant. Vetrin. 9: 69-70.Marbach, I. and Mayer, M.A. 1974. Permeability of seed coats towater as related to drying conditions and metablism ofphenolics. Plant physiol. 54: 817-820.Marbach, I. and Mayer, M.A. 1975. Changes in catechol oxidasesand permeability to water in seed coats of Pisum elatius during seed development and maturation. Plant Physiol. 56:93-96.Mayer, A.M. and Poljakoff-Mayber, A. 1989. The germination ofseeds. Pergamon press, Oxford.McDonough, W.T. and Chadwick, D.L. 1970. Pre-emergence leachinglosses from seeds. Plant & Soil 32: 327-334.McRill, M. and Sagar, G.R. 1973. Earthworms and seeds. Nat.243: 482.Mithen, R., Herper, J.L., and Weiner, J. 1984. Growth andmortality of individual plants as a function of "availablearea". Oecol. (Berlin) 62: 57-60.Moore, R.P. 1972. Tetrazolium staining for assessing seedquality. In Seed ecology: Proceedings of the nineteentheaster school in agriculture science, ed. Heydecker, W., pp.347-366, Butterworths, London.Naqvi, H.H. and Hanson, G.P. 1982. Germination and growthinhibitors in guayule (Parthenium argentatum Gray) chaff andtheir possible influence in seed dormancy. Amer. J. Bot. 69:985-989.159Navas, M.L. 1991. Using plant population biology in weedresearch: a strategy to improve weed management. Weed Res.31: 171-179.Nolan, D.G. 1989. Seed germination characteristics of Centaureadiffusa and C. maculosa. M.Sc. thesis, Department of PlantScience, University of British Columbia, Vancouver, B.C.Nolan, D.G. and Upadhyaya, M.K. 1988. Primary seed dormancy indiffuse and spotted knapweed. Can. J. Plant Sci. 68: 775-783.Odum, S. 1965. Germination of ancient seeds; floristicalobservation and experiments with archaeologically dated soilsamples. Dansk. Botanisk Arkiv. 24: 1-70. (see Harper, J.L.1977. Population biology of plants. pp.79).Oxley, E.R.H. 1977. The population dynamics of the foxglove,Digitalis purpurea L. In Population biology of plants. ed.Harper, J.L. pp. 266. Acadamic Press, London.Parker, V.T., Simpson, R.L., and Leck, M.A. 1989. Pattern andprocess in the dynamics of seed banks. In Ecology of soil seed banks. eds. Leck, M.A., Parker, V.T., and Simpson, R.L.pp. 367-384. Academic Press, INC.Patterson, D.T. 1981. Effects of allelopathic chemicals on growthand physiological responses of soybean (Glycine max). WeedSci. 29: 53-59.Pesis, E. and Ng, T.J. 1986. The effect of seed coat removal ongermination and respiration of maskmelon seeds. Seed Sci. andTechnol. 14: 117-125.Peterson, D.L. and Bazzaz, F.A. 1978. Life cycle characteristicsof Aster pilosus in early successional habitat. Ecol. 59:1005-1013.Pianka, E.R. and Parker, W.S. 1975. Age specific reproductivetactics. Am. Nat. 109: 453-464.Pons, T.L. 1991. Dormancy, germination and mortality of seeds ina chalk grassland flora. J. Ecol. 79: 765-780.Powell, A.A. and Matthews, S. 1978. The damaging effect of wateron dry pea embryos during imbibition. J. Exp. Bot. 29:1215-1229.Powell, R.D. 1988. Diffuse knapweed (Centaurea diffusa):reproductive thresholds, population ecology and responses tothe interaction of the buprestid beetle SphenoptermJugoslavia. Ph.D. thesis. The University of British Columbia,Canada.160Pyke, D.A. and Thompson, J.N. 1986. Statistical analysis ofsurvival and removal rate experiments. Ecol. 67: 240-245.Radosevich, S.R. and Holt, J.S. 1984. Weed ecology: implicationfor vegetation management. Wiley-Intescience Publication.John Wiley Sons.Reichman, O.J. and Smith, S.C. 1991. Responses to simulated leafand root herbivory by a biennial, Tragopogon dubius. Ecol.72: 116-124.Reisman-Berman, 0., Kigel, J. and Rubin, B. 1989. Short soakingin water inhibits germination of Datura ferox L. and D.stramonium L. seeds. Weed Res. 29: 357-362.Reisman-Berman, 0., Kigel, J. and Rubin, B. 1991. Dormancypatterns in buried seeds of Datura ferox L. and D. stramoniumL. Can. J. Bot. 69: 173-179.Richard, G., Raymond, P., Corbineau, F. and Pradet, A. 1989.Effect of the pericarp on sugar beet (Beta vulgaris L.) seedgermination: Study of the energy metabolism. Seed Sci. andTechnol. 17: 485-497.Roberts, E.H. 1988. Temperature and seed germination. In Plantsand temperature. eds. Song, S.P. and Woodward, F.I. pp.109-132. The company of Biologists Ltd., London.Roberts, H.A. 1970. Viable weed seeds in cultivated soils. Rep.Natl. Veg. Res. Stn., 1969. pp. 25-38.Roberts, H.A. 1981. Seed banks in soils. Adv. Appl. Biol. 6:1-55.Roberts, H.A. 1986. Seed persistence in soil and seasonalemergence in plant species from different habitats. J. Appl.Ecol. 23: 639-656.Roberts, H.A. and Boddrell, J.E. 1983. Seed survival andperiodicity of seedling emergence in eight species ofCruciferae. Ann. Appl. Biol. 103: 301-309.Roberts, H.A. and Boddrell, J.E. 1984. Seed survival and seasonalemergence of seedlings of some ruderal plants. J. Appl. Ecol.21: 617-628.Roberts, H.A. and Chancellor, R.J. 1979. Periodicity of seedlingemergence and achene survival in some species of Carduus,Cirsium and Onopordium. J. Appl. Ecol. 16: 641-647.161Roberts, H.A. and Nelson, J.E. 1981. Seed survival andperiodicity of seedling emergence in twelve weedy species ofCompositae. Ann. Appl. Biol. 97: 325-334.Sagar, G.R. and Mortimer, A.M. 1976. An approach to the study ofthe population dynamics of plants with special reference toweeds. Appl. Biol. 1: 1-47.Sarukhan, J. 1974. Studies on plant demography: Ranunculus repens L., R. bulbosus L., and R. acris L. II. Reproductivestrategies and seed population dynamics. J. Ecol. 62:151-177.Sarukhan, J. and Harper, J.L. 1973. Studies on plant demography:Ranunculus repens L., R. bulbosus L., and R. acris L. I.Population flux and survivorship. J. Ecol. 61: 675-716.Sarukhan, J., Martinez-Ramos, M., and Pinero, D. 1984. Theanalysis of demographic variability at the individual leveland its population consequences. In Perspectives on plant population ecology. eds. Dirzo, R. and Sarukhan, J. pp.83-140. Sinauer Associates Inc., Sunderland, Massachusetts.Schaal, B.A. 1984. Life-history variation, natural selection, andmaternal effects in plant populations. In Perspectives on plant population ecology. eds. Dirzo, R. and Sarukhan, J. pp.188-206. Sinauer Associates Inc, Sunderland, Massachusetts.Schafer, D.E. and Chilcote, D.O. 1970. Factors influencingpersistence and depletion in buried seed populations. II. Theeffects of soil temperature and moisture. Crop Sci. 10:342-345.Schwabe, W.W. 1954. Factors controlling flowering in theChrysanthemum. IV. The site of vernalization andtranslocation of the stimulus. J. Exp. Bot. 5: 389-400.Schmitt, J. 1983. Individual flowering phenology, plant size, andreproductive success in Linanthus androsaceus, a Californiaannual. Oecol. (Berlin) 59: 135-140.Scroth, M.N. and Cook, R.J. 1964. Seed exudation and itsinfluence on pre-emergence damping-off of beans. Phytopathol.54: 670-673.Sheldon, J.C. 1974. The behaviour of seeds in soil. III. Theinfluence of seed morphology and behaviour of seedlings onthe establishment of plants from surface-lying seeds. J.Ecol. 62: 47-66.Silvertown, J.W. 1983. Why are biennials sometimes not so few?Amer. Nat. 121: 448-453.162Silvertown, J.W. 1984. Introduction to plant population biology.Longman Sci. and Techn., London and New York.Simon, E.W. 1978. Membranes in dry and imbibing seeds. In DryBiological Systems. eds. Crowe, J.H. and Clegg, J.S., pp.1215-1229. Academic Press, N.Y.Simpson, R.L., Leck, M.A., and Parker, V.T. 1989. Seed banks:general concepts and methodological issues. In Ecology of soil seed banks. eds. Leck, M.A., Parker, V.T., and Simpson,R.L. pp 3-8. Academic Press, INC.Solbrig, O.T. 1980. Demography and natural selection. InDemography and evolution in plant populations. ed. Solbrig,O.T. pp. 1-20. Blackwell Sci. Publ. Oxford.Sreeramulu, N. 1983. Auxins, inhibitors and phenolics inbambarranut seeds (Voandzeia subterranea Thouars) in relationto loss of viability during storage. Ann. Bot. 51: 209-216Steinbauer, G.P. and Grigsby, B. 1957. Interaction oftemperature, light and moistening agent in the germination ofweed seeds. Weeds 5: 175-182.Stoller, E.W. and Wax, L. M. 1973. Temperature variations in thesurface layers of an agricultural soil. Weed Res. 13:273-282.Taylorson, R.B. 1972. Phytochrome controlled changes in dormancyand germination of buried weed seeds. Weed Sci. 20: 417-422.Taylorson, R.B. 1987. Environmental and chemical manipulation ofweed seed dormancy. Rev. Weed Sci. 3: 135-154.Taylorson, R.B. and Hendricks, S.B. 1977. Dormancy in seeds. Ann.Rev. Plant Physiol. 28: 331-354.Thapliyal, P. and Nautiyal, A.R. 1989. Inhibition of seedgermination by pericarp in Fraxinus micrantha Lang. Seed Sci.and Technol. 17: 125-130.Thill, D.C., Zamora, D.L., and Kambitsch, D.L. 1985. Germinationand viability of common crupina (Crupina vulgarus) achenesburied in field. Weed Sci. 33: 344-348.Thomas, A.S. 1960. The trampling animal. J. Br. Grassl. Soc.15: 89-93.Thompson, K. and Grime, J.P. 1979. Seasonal variation in the seedbanks of herbaceous species in ten contrasting habitats. J.Ecol. 67: 893-921.163Tilsner, H.R. and Upadhyaya, M.K. 1985. Induction and release ofsecondary seed dormancy in genetically pure lines of Avena fatua. Physiol. Plant. 64: 377-382.Toole, V.K. 1973. Effects of light, temperature and theirinteractions on the germination of seeds. Seed Sci. andTechnol. 1: 339-396.Upadhyaya, M.K., Tilsner, H.R., and Pitt, M.D. 1988. The biologyof Canadian weeds. 87. Cynoglossum officinale L. Can J. PlantSci. 68: 763-774.Van Baalen, A. 1982. Germination ecology and seed populationdynamics of Digitalis purpurea. Oecol.(Berl.) 53: 61-67.Van Breeman, A.M.M. 1984. Comparative germination ecology ofthree short-lived monocarpic Boraginaceae. Acta Bot. Neerl.33: 283-306.Van Breeman, A.M.M. and Van Leeuwen, B.H. 1983. The seed bank ofthree short-lived monocarpic species, Crisium vulgare(Compositae), Echium vulgare and Cynoglossum officinale(Boraginaceae). Acta Bot. Neerl. 32: 245-246.Van Leeuwen, B.H. and Van Breeman, A.M.M. 1980. Similarities anddifferences in some biennials. Acta Bot. Neerl. 29: 209-210.Van der Meijden, E. and Van der Waals-Kooi, R.E. 1979. Thepopulation ecology of Senecio lacobaea in a sand dune system.J. Ecol. 67: 131-153.Van Mierlo, J.E.M. and Van Groenendael, J.M. 1991. A populationdynamics approach to the control of Anthriscus sylvestris (L.) Hoffm. J. Appl. Ecol. 28: 128-139.Vegis, A. 1964. Dormancy in higher plants. Ann. Rev. PlantPhysiol. 15: 185-224.Venable, D.L. 1984. Using intraspecific variation to study theecological significance and evolution of plantlife-histories. In Perspectives on plant population ecology.eds. Dirzo, R. and Sarukhan, J. pp. 166-187. SinauerAssociates Inc, Sunderland, Massachusetts.Venable, D.L. 1989. Modeling the evolutionary ecology of seedbanks. In Ecology of soil seed banks. eds. Leck, M.A.,Parker, V.T., Simpson, R.L. pp. 67-87. Academic Press, INC.Vincent, E.M. and Cavers, P.B. 1978. The effects of wetting anddrying on the subsequent germination of Rumex crispus L. Can.J. Bot. 56: 2207-2217.164Watkinson, A.R. 1978. The demography of a sand dune annual:Vulpia fasciculata. II. The dynamics of seed populations. J.Ecol. 66: 35-44.Weaver, S.E. and Cavers, P.B. 1979. Dynamics of seed populationsof Remex crispus and R. obtusifolius (Polygonaceae) indisturbed and undisturbed soil. J. Appl. Ecol. 16: 909-917.Webb, D.P. and Agnihotri, V.P. (1970) Presence of fungalinhibitors in the pericarps of Acer saccharum fruits. Can. J.Bot. 48: 2109-2116.Webb, D.P. and Van Staden, J., and Wareing, P.F. 1973. Seeddormancy in Acer. J. Exp. Bot. 24: 105-116.Weiner, J. 1988. Variation in the performance of individuals inplant populations. In Plant population ecology. eds. Davy,A.J., Hutchings, M.J., and Watkinson, A.R. pp. 59-82.Blackwell Sci. Publ. Oxford.Weiner, J. and Solbrig, 0. T. 1984. The meaning and measurementof size hierarchies in plant population. Oecol. (Berlin) 61:334-336.Weiner, J. and Whigham, D.F. 1988. Size inequality andself-thinning in wild rice (Zizania aquatica). Amer. J. Bot.73: 445-448.Werker, E. (1980/81) Seed dormancy as explained by the anatomy ofthe embryo envelops. Isrl. J. Bot. 29: 22-24.Werner, P.A. 1975. Prediction of fate from rosette size in teasel(Dipsacus fullonum L.). Oecol. (Berlin) 20: 197-201.Wesson, G. and Wareing, P.F. 1969. The interaction of lightsensitivity in weed seeds by burial. J. Exp. Bot. 20:413-425.Wilkinson, L. 1990a. SYSTAT: The system for statistics. Evanston,IL, Systat Inc.Wilkinson, L. 1990b. SYGRAPHY: The system for graphics. Evanston,IL, Systat Inc.Williams, R.D. and Hoagland, R.E. 1982. The effects of naturallyoccurring phenolic compounds on seed germination. Weed Sci.30: 206-212.Wright, G.C., McWilliam, J.R., and Whalley, R.D.B. 1980. Effectsof light and leaching on germination of saffron thistle(Carthamus lanatus L.). Aust. J. Plant Physiol. 7: 587-594.


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