"Forestry, Faculty of"@en . "DSpace"@en . "UBCV"@en . "Shatford, Jeffrey Peter Alan"@en . "2009-03-11T20:21:10Z"@en . "1997"@en . "Master of Science - MSc"@en . "University of British Columbia"@en . "The purpose of this research was to collect detailed demographic information concerning\r\nthe early life stages of antelope bitterbrush Purshia tridentata and elucidate those biotic and\r\nabiotic factors limiting or enhancing seed dispersal, seedling emergence and survival.\r\nAgricultural and urban development have seriously reduced the size and continuity of natural\r\nareas in the Okanagan valley, British Columbia. Antelope bitterbrush has its most northern\r\ndistribution in this setting and this study was initiated to address questions concerning the\r\nability of this shrub to persist in the face of rapid change. The research was undertaken in three\r\nparts represented by the separate chapters of this manuscript; Chapter 1 a survey of juvenile\r\nplants, Chapter 2 a seed removal experiment and Chapter 3 a reciprocal planting experiment.\r\nIn 1994 I estimated the density of juvenile Purshia shrubs at ten separate sites. This\r\nindicated significant variation in seedling establishment between sites and provided the basis for\r\n2 separate lines of research. First, I placed Purshia seeds into the field at feeding stations to\r\ndetermine what class of seed eating animals, insects, birds or rodents, were responsible for seed\r\npredation. Seed predation was mostly due to nocturnal rodents some of which, through caching\r\nbehaviour, also acted as effective agents of seed dispersal.\r\nSecondly, 8,000 seeds were planted individually 2.25 cm beneath the soil surface.\r\nEnvironmental factors explained 40 % of the variation in seedling emergence while variation\r\nbetween maternal families accounted for 16 %. Seedling emergence correlated negatively with\r\nsilt content of soils and positively with moss cover. Shrubs that on average produced seeds with\r\na higher nitrogen content also produced seeds that were less likely to emerge. In silty soils\r\ninitial seedling emergence appeared to be limited by crusting at the soil surface. However, soils\r\nwith higher silt content contained more seedlings at the end of the first year, presumably due to\r\ngreater moisture availability. The correlation between moss cover and seedling emergence\r\nprovides some evidence that cryptogamic crusts confer conditions favourable to seedling\r\nestablishment. Knowledge of factors limiting seedling establishment is required to manage and\r\nwhere appropriate to restore Purshia shrubs in the highly impacted landscape of the south\r\nOkanagan valley."@en . "https://circle.library.ubc.ca/rest/handle/2429/5898?expand=metadata"@en . "3625296 bytes"@en . "application/pdf"@en . "SEED DISPERSAL, SEED ATTRIBUTES A N D EDAPHIC FACTORS: THEIR R O L E A N D IMPACT ON THE REGENERATION OF ANTELOPE BITTERBRUSH (PURSHIA TRIDENTATA, ROSACEAE) by JEFFREY PETER A L A N SHATFORD B.Sc. Queen's University, 1986 A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF THE REQUIREMENTS FOR THE DEGREE OF M A S T E R OF SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES Centre for Applied Conservation Biology Department of Forest Sciences Faculty of Forestry We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A April, 1997 \u00C2\u00A9 J. P. A . Shatford, 1997 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. 1 further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada DE-6 (2/88) Abstract The purpose of this research was to collect detailed demographic information concerning the early life stages of antelope bitterbrush Purshia tridentata and elucidate those biotic and abiotic factors limiting or enhancing seed dispersal, seedling emergence and survival. Agricultural and urban development have seriously reduced the size and continuity of natural areas in the Okanagan valley, British Columbia. Antelope bitterbrush has its most northern distribution in this setting and this study was initiated to address questions concerning the ability of this shrub to persist in the face of rapid change. The research was undertaken in three parts represented by the separate chapters of this manuscript; Chapter 1 a survey of juvenile plants, Chapter 2 a seed removal experiment and Chapter 3 a reciprocal planting experiment. In 1994 I estimated the density of juvenile Purshia shrubs at ten separate sites. This indicated significant variation in seedling establishment between sites and provided the basis for 2 separate lines of research. First, I placed Purshia seeds into the field at feeding stations to determine what class of seed eating animals, insects, birds or rodents, were responsible for seed predation. Seed predation was mostly due to nocturnal rodents some of which, through caching behaviour, also acted as effective agents of seed dispersal. Secondly, 8,000 seeds were planted individually 2.25 cm beneath the soil surface. Environmental factors explained 40 % of the variation in seedling emergence while variation between maternal families accounted for 16 %. Seedling emergence correlated negatively with silt content of soils and positively with moss cover. Shrubs that on average produced seeds with a higher nitrogen content also produced seeds that were less likely to emerge. In silty soils initial seedling emergence appeared to be limited by crusting at the soil surface. However, soils with higher silt content contained more seedlings at the end of the first year, presumably due to i i greater moisture availability. The correlation between moss cover and seedling emergence provides some evidence that cryptogamic crusts confer conditions favourable to seedling establishment. Knowledge of factors limiting seedling establishment is required to manage and where appropriate to restore Purshia shrubs in the highly impacted landscape of the south Okanagan valley. iii Table of Contents Abstract \u00C2\u00BB List of Tables vii List of Figures viii Acknowledgments x CHAPTER 1 GENERAL INTRODUCTION AND BACKGROUND TO THE STUDY 1 Introduction 1 Significance of this work 2 Material and Methods 3 Study Area 3 Juvenile survey 6 Results 7 Discussion 8 CHAPTER 2 ANIMAL IMPACTS ON PLANT POPULATIONS THROUGH SEED DISPERSAL AND PREDATION 10 Introduction 10 Material and Methods 12 Study Area 12 Study species 13 iv Fruit production 14 Fruit harvest: diel and microhabitat effects 14 Fruit harvest: 4 weeks exposure 16 Fruit harvest analysis 16 Natural seedling emergence 17 Results 17 Fruit production 17 Fruit harvest: diel and microhabitat effects 18 Fruit harvest: 4 weeks exposure 19 Granivores 22 Natural Seedling emergence 23 Discussion 26 Fruit harvest 26 Microhabitat 28 Seedling emergence 29 Alternative modes of seed dispersal 32 Management Implications 33 Summary 35 CHAPTER 3 SEEDLING EMERGENCE AND SURVIVAL 37 Introduction 37 Materials and Methods 39 Study species 39 Study area 40 v Seed collection Experimental design Seed planting, relocation and measurement Edaphic measurements Analyses Results Planting date Edaphic factors Site of Growth Populations and maternal families Survival Discussion Edaphic factors Site of Growth Populations Maternal families Planting date Survival Management implications Literature Cited Appendix I List of Tables Table 2.1 Grid locations and elevation of study sites in the South Okanagan Valley, (courtesy of G.G.E. Scudder). 13 Table 2.2. Effect of site and time (day - night) on the removal of Purshia tridentata fruits by seed predators at five study sites in B.C., BO, OS, WT, CWS and K B . 20 Table 2.3. Effect of site and microhabitat on Purshia tridentata seed removal by nocturnal seed predators at four study sites in B.C.; BO, OS, WT and CWS. 21 Table 2.4 Effect of site and microhabitat on seed removal during four weeks of exposure to seed predators and abiotic factors. 22 Table 2.5. Anova for the effect of site and year on the number of Purshia tridentata seedling spots over two years, 1995 and 1996, for four study sites in British Columbia, K B , WT, B O and OS. 25 Table 3.1. Anova table for Purshia tridentata seedling emergence data for seeds reciprocally planted on three sites BO, OS and WT. 49 Table 3.2. Frequency of mortality between April 27th and July 26th, 1995, fox Purshia tridentata seedlings at three study sites in British Columbia. 53 Appendix Table A. 1 Rodents captures on 5 study sites, May-Nov. 1995. 71 vii List of Figures Figure 1.1 Map of i) southern British Columbia and ii) the south Okanagan valley showing five principal and five additional study sites. 4 Figure 1.2. Purshia tridentata juvenile shrub density on ten study sites in the Okanagan valley. B.C., in 1994 (\u00C2\u00B1 1 S.E.M., n = 15). 7 Figure 2.1 Fruit production of Purshia tridentata (mean \u00C2\u00B1 1 S.E.M., n = 5 ) by individual shrubs at four study sites in British Columbia. 18 Figure 2.2. The number of Purshia tridentata seeds removed (mean \u00C2\u00B1 S.E., n= 40) from seed stations during the day and night at each of five study sites in British Columbia. 20 Figure 2.3. The removal of Purshia tridentata seeds in two microhabitats (under shrubs vs. in openings) by nocturnal seed predators at each of five study sites in B.C. (mean \u00C2\u00B1 S.E., n= 20). 21 Figure 2.4. The relationship between the total number of Purshia tridentata seeds removed during a 4 week trial and the average number of rodents captured during 8 trapping weeks at five study sites. 23 Figure 2.5. The number of Purshia tridentata spots containing single seedlings and clumps of seedlings. 25 Figure 2.6 The total number of Purshia tridentata seedling clumps in 1996 at each of five study sites in relationship to the number of pocket mice (P. parvus), deer mice (P. maniculatus) and both species combined. 26 Figure 3.1. Purshia tridentata: longitudinal section of fruit. Adapted from Schopmeyer 1974. 40 Figure 3.2. Schematic view of planting blocks, plots and subplots for two sites of the reciprocal planting experiment. 42 Figure 3.3. Weather data from January - June, 1995 with precipitation compared to 70 year average. Source Environment Canada, Oliver sewage treatment plant station. 46 Figure 3.4. The relationship between; A Purshia tridentata seedling emergence and moss cover and B between seedling emergence and silt content of soils. 48 viii Figure 3.5. Seedling emergence (mean \u00C2\u00B1 S.E.) of four populations of Purshia tridentata reciprocally planted among three sites in British Columbia. 51 Figure 3.6. Average seedling emergence (mean \u00C2\u00B1 1 S.E.) fox Purshia tridentata seeds from four populations reciprocally planted at three sites, BO, OS and WT in British Columbia. 51 Figure 3.7. The relationship between average % nitrogen content (of Purshia tridentata seeds) from each of 20 maternal families and the % seedling emergence at three separate planting sites. 52 Figure 3.8. Moisture loss in loamy and sandy soils between April 27 and June 5th, 1995 for three sites in the Okanagan Valley, British Columbia. 53 Figure 3.9. The relationship between silt content of soils and the percentage of Purshia tridentata seedlings surviving the first year of growth. Data from the reciprocal planting experiment using three locations in British Columbia, BO, OS and WT. 54 ix Acknowledgments I am greatly indebted to Susanne Pa land for encouraging me in the pursuit o f know ledge and for p rov id i ng support i n abundance. I a m also thankfu l to Dr . P a m K r a n n i t z for the opportuni ty to partake i n the interesting and important wo rk i n the south Okanagan and for support ing me i n m y project a long w i t h D r . G e o f f Scudder and the Vancouve r Foundat ion . I a m grateful to m y adv isory commit tee for their t ime and intel l igent comments over the course o f m y research; D r . P a m Krann i t z , D r . P h i l Bur ton , D r . Susan G l e n n , D r . Char les K r e b s and Dr . A n t a l K o z a k . T o m D u r a l i a p rov ided he lp fu l (and humorous) comments on earl ier drafts o f this manuscr ipt . Thanks T o m ! I a m indebted to the many landowners who prov ided access to their propert ies, espec ia l ly the Osoyoos F i rs t Na t ions B a n d and B l a k e and George Kennedy . Severa l people contr ibuted to m y w o r k i n the f i e ld ; D e v o n H a a g , L i s s a For ty , P h y l l i s Gab r i e l , M y r a Bapt is te and Jolene K ruge r . I a m thankfu l to the people at the f ie ld station for their company and cooperat ion espec ia l ly L y n n e A t w o o d , M i k e G i l l and Samantha H i c k s and var ious volunteers w h o showed a keen interest i n m y research; R e m c o T ikkemei je r (Ho l land) , Cat r in Wes tpha l (Germany) and N a n c y Job (South A f r i ca ) . I thank D r . P h i l Bu r ton for opening the door to U . B . C . and D r . Susan G l e n n for keep ing it so. I am also thankfu l for the h igh standard o f teaching exh ib i ted by the facul ty at U . B . C . I am h igh ly indebted to the esprit de corps o f the graduate students, facul ty and staff at the Centre for A p p l i e d Conservat ion B i o l o g y , U . B . C . and at the Bo tan i ca l Institute, Un ivers i ta t B a s e l , Swi tzer land. Las t but not least, I am very thankfu l to the support o f m y fam i l y and fr iends wi thout w h i c h this wo rk w o u l d not have been poss ib le . x Chapter 1 General introduction and background to the study Introduction Antelope bitterbrush Purshia tridentata (Rosaceae) is a shrub species growing throughout the intermountain region of western North America. It grows in association with a variety of vegetation types including; sagebrush (Artemisia tridentata), ponderosa pine (Pinus ponderosa), lodgepole pine (P. contorta), oak (Quercus spp.) and juniper (Juniperus spp.) (Nord, 1965; West, 1968). Purshia is of interest to wildlife and range managers because of the valuable forage it provides to domestic cattle and sheep, as well as mule deer (Odocoileus hemionus) and California bighorn sheep (Ovis canadensis californiana) (Burrell, 1982; Urness 1989). As well, Purshia shrubs provide secondary structure to the plant community and support an exceptional diversity of invertebrate species (Furniss, 1982). Growing up to 3 meters tall, shrubs trap wind-born materials, soil, seeds and snow and shrubs provide shade from the intense heat of the sun. With long tap roots, up to 5 meters, shrubs encourage moisture penetration and bear nodules of nitrogen-fixing bacteria (Frankia spp.) (Dalton and Zobel, 1977). The combination of forage value and structural attributes has fostered research into the potential for using Purshia in range rehabilitation and restoration (Young, 1988). Currently only one cultivar, 'Lassen', is available and is widely recommended throughout the western states (Shaw and Monsen, 1986). Occupying extensive areas from Arizona and California north to Washington and Montana, Purshia's range in Canada is limited to southern British Columbia. This includes the dry valleys of the Rocky Mountain trench, near Cranbrook, and the south Okanagan valley, from Osoyoos north to Penticton and Kelowna. Its occurrence in the Okanagan valley coincides with one of Canada's most endangered landscapes. The area is currently the focal point for 1 conservation efforts because it encompasses a unique assemblage of flora and fauna while rapid development has reduced the available habitat (Hlady, 1993). This overlap of high diversity and rapid loss of natural areas has resulted in a number of provincially red and blue listed animal species; (12 mammals, 20 birds, 4 reptiles and 1 amphibian) and 67 rare, threatened or endangered plant species (Harper et al, 1993). Intensive cattle grazing began to severely impact plant communities in the south Okanagan valley as early as 1880 (Cannings et al, 1987). Agricultural development began in earnest following the construction of an irrigation canal serving the southern part of the valley (circa 1920) (Cannings et al, 1987). Large areas of natural vegetation have been cleared and replaced with fruit orchards, market gardens, hayfields and more recently vineyards (Harper et al, 1993). The conversion of natural communities has reduced the shrub steppe community to a small fraction of its previous distribution in the Okanagan valley (R. Maxwell, B.C. Ministry of Environment). Today many remnant stands of Purshia exist in isolation either surrounded by development or crowded against ponderosa pine (Pinus ponderosa) and Douglas-fir (Pseudotsuga menziesii) forests that skirt drainages and dominate the hillsides at higher elevations. Significance of this work The long term persistence of Purshia in a highly impacted landscape is of concern to land managers, both for its intrinsic value, the role it plays in structuring the landscape and for the wildlife species that depend upon it. To conserve Purshia stands effectively requires an understanding of its requirements over its entire life cycle and how these may vary across the range of habitats that it occupies (Harvey, 1985). Purshia is a perennial plant that reproduces primarily from seed and I have focused primarily on the early life history stages, seeds and 2 seedlings. Following the fate of seeds in the field, while often difficult, does provide valuable information concerning the fate of individuals (Chambers and MacMahon, 1994). An understanding of seed bed characteristics and their relationship to seedling emergence and survival are important considerations in range revegetation (Round and Call, 1988). Theoretical ecologists are interested in the events at the seed and seedling stage as they determine, to a large part, what species occupy a given location and the direction of succession (Grubb, 1977). Consideration is given to these issues in the discussions below. Purshia is considered a seasonal regenerating species whose seeds are dormant at the time of dispersal in summer and which germinate in late winter or early spring (Young and Evans, 1976). Previous research on Purshia has identified the importance of seed dispersal by animals (Vander Wall, 1994b), detailed knowledge of germination in the laboratory (Meyer and Monsen, 1989) and in the field (Young et al, 1993) and environmental impacts on seedling survival (Ferguson, 1972; Vander Wall, 1994b). The research reported here focused on the regeneration of Purshia at various field locations in the Okanagan valley. This would provide baseline demographic information for land managers and identified factors relevant to individuals interested in planting Purshia seeds for restoration. Material and Methods Study Area The South Okanagan Basin is located in the extreme south central part of British Columbia (Figure 1.1). The region constitutes a narrow extension of the Great Basin Desert 3 British Columbia Okanagan Valley * Canada 11 N A Vaseux Lake O l i v e r = 5 km approx. Q C W S O K B \u00E2\u0080\u00A2 K L B W W T R G O O S South Okanagan Valley O B O Osoyoos =\u00E2\u0080\u00A2-Lake Osoyoos O L W A Canada U.S.A. 49\u00C2\u00B015 'N . Q Main study sites A Additional sites 49\u00C2\u00B0 0' N . Figure 1.1 Map of i) southern British Columbia and ii) the south Okanagan valley showing five principal and five additional study sites. A l l sites are labeled with a two or three letter abbreviation, CWS Canadian Wildlife Service, K B Kennedy Bench, K L Kennedy Lowland, B W Bright's Winery, RG Rocky Grazed, WT Water Tower, OS Orchard Site, B O Burrowing Owl, OL Osoyoos Lake and W A Washington. which stretches south to more extensive areas in Washington, Oregon and California. Lying in the rain shadow of the Coast Mountains, the Okanagan valley is one of the driest and hottest regions of the province (mean annual precipitation 310 mm, average temp. 9.9\u00C2\u00B0 C) (Environment Canada, Climate data center). The elevation begins at 270 m (Osoyoos Lake), rises along gentle rolling hills and is punctuated in places by steep sided benchland and rocky cliffs. The area constitutes part of the Ponderosa pine - Bunchgrass biogeoclimatic zone (Meidinger and Pojar, 1991) and is characterized by a shrub steppe plant community. Plant growth is restricted by seasonal extremes, cold winters and dry summers. Two shrub species, big sagebrush (Artemisia tridentata) and antelope bitterbrush (Purshia tridentata) co-dominate areas of the valley bottom and side hills. The distributions of these species are partially separated by soil types with the Artemisia associated with fine textured soils and Purshia with sandy and rocky soils (personal observation). In undisturbed areas widely spaced bunchgrasses, e.g., bluebunch wheatgrass (Agropyron spicatum) and needle-and-thread grass (Stipa comata) are interspersed with annual and perennial herbs and a diverse community of bryophytes and lichens (the cryptogamic crust). The region has been invaded by introduced weeds which thrive in disturbed rangeland, particularly the annual cheatgrass (Bromus tectorum) and diffuse knapweed (Centaurea diffusa) (Harper et al, 1993). Ten study areas were selected for the South Okanagan Grassland Conservation Project (Figure 1.1), a cooperative undertaking between private, government and university interests. The sites were located in the vicinity of Oroville, Washington and along the east side of the Okanagan valley to Vaseux Lake, British Columbia. A l l sites were located at a similar elevation, circa 350-475 meters a.s.l. and were dominated by Purshia and perennial bunchgrasses (see Krannitz in press a, for exact locations of all ten sites). 5 Juvenile survey Juvenile Purshia shrubs were censused on ten sites in the summer of 1994. The juvenile classification included all non-flowering plants. Fifteen circular plots were placed systematically every 20 or 30 meters along a 400 meter transect until a total of 15 plots had been sampled at each site. The radius of each plot was 3.99 meters, describing an area of 50 m 2 . The number of juveniles occupying each plot was recorded. Analysis of variance (Systat, 1992) was carried out to determine if the number of juvenile shrubs was significant differences between sites. Multiple comparisons used the Bonferroni adjustment (Systat, 1992). The base diameter was also recorded for all individuals and age was estimated from the branching pattern (White, 1979) and the appearance of the bark. 6 Results Considerable variation in seedling and juvenile recruitment was evident with some sites having no seedlings and few juveniles while juvenile counts on two sites (KB and OS) were significantly higher (F= 15.99, d.f.= 9, P < 0.001) (Figure 1.2). Juvenile plants were less than 10 mm diameter at the base and this along with branch architecture suggested that most individuals were less than 7 years old. CT\" (0 O O a> a . c \"a c CD > 30 25 20 15 10 5 a a i -a a r * i rf CWS KB Kl BW RG WT OS BO ELO WA LOCATION Figure 1.2. Purshia tridentata juvenile shrub density on ten study sites in the Okanagan valley. B.C. , in 1994 (\u00C2\u00B1 1 S.E.M., n = 15). Sites with the same letter are not significantly different. 7 Discussion Presumably all individuals included in the juvenile survey were < 10 years old because Purshia shrubs do not flower until this age (Nord, 1965). The current number of juveniles found at each site reflected the total amount of seedling establishment over several years. For eight of the ten sites surveyed approximately 0.1 - 1.0 seedling per 100 m were established per year. Given adult densities of 20 - 40 individuals per 100 m (P.G. Krannitz, Environment Canada, unpublished data) and a life span of > 30 years, this rate of recruitment could be sufficient to maintain some of the sites. However, site BO had considerable evidence of adult die-back while recruitment was extremely limited. At sites OS and K B recruitment rates were as much as 10 times higher. Such extensive variation in seedling establishment would suggest that some factor or factors controlling population growth varied between sites in their magnitude and/or direction of influence (Fenner, 1987). Three hypotheses were generated to explain the large variation in seedling establishment between sites. First, the dependence of Purshia on seed dispersal by rodents (Vander Wall, 1994) suggested that between site differences in the size and assemblage of seed predator populations could contribute to differences in recruitment rates. Secondly, between site differences in the quality of seeds and the vigor of seedlings could determine the ability of a stand to regenerate. Thirdly, between site differences in seed bed characteristics would influence seedling emergence and survival and may be a limiting factor in seedling establishment (Roundy and Call, 1988). To approach the answers to these questions I initiated two separate lines of inquiry. The first study was to determine the relative role of different classes of predators on Purshia seeds. Some seed predators may act as seed dispersers (Price and Jenkins, 1986). Therefore I wanted to 8 determine which species may have a positive impact on Purshia populations. To do this I relied on detailed census of rodent populations available from concurrent research (W.F. Klenner, Ministry of Forests, unpublished data), an observational study of seed removal by nocturnal and diurnal seed predators and two spring surveys of natural seedling emergence. Secondly, I initiated a planting experiment to measure biotic and abiotic factors limiting seedling emergence and survival. I characterized planting sites by various environmental factors including soil moisture, texture and competitive environment. By planting seeds in a reciprocal manner, i.e., comparing the performance of seeds from different populations side by side in a number of locations, I was able to determine the relative contribution of seed source to the overall variation in seedling emergence. This technique also identified environment/population interactions that indicated site specific adaptations (Bradshaw, 1984). Due to time and energy constraints I limited the two lines of research to a subset of the original 10 sites used in the juvenile survey. The observational study occurred on five sites (CWS, K B , WT, OS and BO) and the planting experiment on the latter four sites. These were chosen because they represented sites with both high and low juvenile shrub densities and high and low rodent densities as determined by trapping in 1994 (W.F. Klenner, unpublished data). 9 Chapter 2 Animal impacts on plant populations through seed dispersal and predation Introduction Plants have evolved an assortment of dispersal mechanisms to ensure the movement of seeds to microsites favourable to germination and seedling establishment (Harper et al, 1970; Schupp, 1993). In arid grasslands, seed dispersal is frequently by wind or ballistic propulsion (Howe and Smallwood, 1982). Plant ecologists and range managers have, however, become increasingly aware of the role of animals in the dispersal of seeds (Reynolds, 1958; Nord 1965; Janzen, 1971; Reichman, 1979; McAdoo et al, 1983; Price and Jenkins, 1986; Chambers and MacMahon, 1994). For example, rodents and birds frequently store seeds beneath mineral soil (a seed cache) which can serve as a suitable microsite for germination and emergence (McAuliffe, 1990; Vander Wall, 1994b). Plants may depend on a single specialized granivore for seed dispersal or numerous unspecialized predators that vary in the quality of dispersal provided (Smith, 1980; Hutchins et al, 1996). There are also instances where plants depend on two or even three modes of seed dispersal, wind plus rodents (Vander Wall, 1994a) or ballistic plus birds (Willson, 1992). The arid land shrub, Purshia tridentata, dominates the shrub-steppe habitat throughout parts of the intermountain region of North America. Its large seeds contain a high concentration of fat, protein and carbohydrate necessary to sustain seedling establishment in an arid environment (Jenkins, 1988; Baker, 1972). This same concentration of nutrients makes Purshia fruits a valuable food item for a variety of animals, including: nocturnal rodents such as, Great Basin pocket mice (Perognathus parvus) and deer mice (Peromyscus maniculatus) (Everett et al, 1978; Kelrick et al, 1986; Jenkins and Ascanio, 1993) and diurnal rodents such as, yellow-10 pine Chipmunks (Tamias amoenus) (Vander Wall, 1994b), as well as the harvester ants (Pogonomyrmex) and various granivorous birds (Kelrick et al, 1986). Fruit production can exceed 200 kilograms per hectare (Nord 1965) and where granivores exist in sufficient numbers the entire seed crop may be harvested (Evans et al, 1983). The emergence of seedlings of Purshia from seed caches made by animals, presumably rodents, was first described by Hormay (1943) and Nord (1965). Since then, surveys of seedlings and juvenile plants have noted a preponderance of seedlings emerging in tight clumps (West, 1968; Vander Wall, 1994b). Such observations have led a number of authors to speculate that Purshia regeneration is highly associated with and possibly dependent on animal dispersal agents (Nord, 1965; West, 1968; Young and Evans, 1976; Evans et al, 1983; Young, 1988; Vander Wall, 1994b). The various animals that feed on Purshia fruits and seeds undoubtedly differ with regards to their net impact on the plant population and their overall effectiveness as seed dispersers (sensu Schupp, 1993). Those that cache seeds in suitable microsites may increase seedling establishment while those that only prey upon seeds may have a negative impact (Price and Jenkins, 1986). For example, Vander Wall (1994b) showed that yellow-pine chipmunks were responsible for >99% of all Purshia seedling establishment at high elevation sites in Nevada while deer mice were primarily seed predators. The present study was undertaken to distinguish the relative role of different classes of granivores in the predation and dispersal of Purshia seeds at low elevation sites in the south Okanagan valley. Because granivores differ in their foraging behaviour and habitat preferences (Bock, 1984; Kelrick et al, 1986) I was able to distinguish different classes of granivores by monitoring the removal of fruits and seeds from stations during the day and night in two 11 contrasting microhabitats (shrubs and open). Determining the net effect of animals on plant populations was also of interest. The inconspicuous nature of seeds and seed caches makes it extremely difficult to determine the seed shadow created by animal dispersal (Longland and Clements, 1995). As seedling clumps are more easily quantified and provide direct evidence of dispersal by rodents I used them as a measure of the overall effect of rodents on plant populations (Price and Jenkins, 1986). As the size of rodent populations between sites differed by an order of magnitude I anticipated that the number of seedling clumps would increase accordingly. Material and Methods Study Area This study was carried out at five separate study sites along the eastern side of the south Okanagan valley, British Columbia, in shrub steppe habitat dominated by mature stands of Purshia. Two sites, Canadian Wildlife Service (CWS) and Kennedy Bench (KB), were adjacent to the Vaseux Creek Bighorn Sheep Reserve. Orchard site (OS) and Water Tower site (WT), were located on the Inkameep First Nations reserve overlooking the town of Oliver, B.C. Burrowing Owl (BO) site was located 5 kilometers south of Oliver. BO site was the most isolated site being nearest to the valley bottom and surrounded by agricultural development on all sides. A l l sites had similar sandy soils, weather patterns and plant communities. The maximum distance between sites was approximately 17 km (CWS - BO) (see Table 2.1 for precise locations). Not all study sites were included in each phase of the research. For example, CWS was not included in the seed production study and was not surveyed during the first year for emergent seedlings. 12 As part of a larger project rodent populations were monitored at each site from May -November 1994 and again in 1995. Estimates of populations from live trapping data indicated a 10 fold range between sites (W.F. Klenner unpublished data, summarized in Appendix Table 1). K B and CWS had the lowest capture rates, while WT had a moderate rodent population and OS and BO had relatively large rodent populations but differed in the relative proportion of pocket mice and deer mice. Table 2.1 Grid locations and elevation of study sites in the South Okanagan Valley, (courtesy of G.G.E. Scudder). Location North Latitude West Longtitude Elevation degrees minutes degrees minutes meters CWS 49 16 119 30 370 KB 49 15 119 30 450 WT 49 10 119 31 475 OS 49 9 119 32 370 BO 49 7 119 33 351 Study species The fruits of Purshia are oblong achenes with a papery husk that is readily removed to reveal a large seed (26.42 \u00C2\u00B1 0.07 mg 1 S.E.M.) (Krannitz, in press a). Purshia flowers in spring and the fruits are ready to drop by mid to late June. The entire crop ripens synchronously and may fall during a single wind storm. Seeds are dormant at the time of dispersal. Dormancy is broken by cool wet conditions during winter and seedlings emerge in spring when growing conditions are most suitable (Young et al, 1993). Seedlings may grow together from a seed cache and what may appear as a single adult shrub may in fact consist of two or more shrubs intertwined (Jeff Shatford, personal observation). 13 Fruit production In June, 1995 fruits were collected from 20 randomly chosen shrubs, five from each of four sites (KB, OS, WT and BO). A large tarpaulin was placed beneath each shrub to catch falling fruits while the branches were shaken and beaten vigorously. Additional seeds were hand picked from branches and the ground beneath the shrub to ensure that a high percentage of the total production was collected. Fruits were stored in paper bags during transport and temporary storage. The collections were cleaned by sieving through wire mesh of various size and then sifted on large cookie sheets to remove insects, leaves and twigs. Care was taken to retain as many fruits as possible including smaller ones that may have been aborted or damaged by insects. For shrubs with relatively few fruits the entire collection was counted and for larger collections (> 1,000), four lots of 100 fruits were weighed and the average used to estimate the total production. Fruit harvest: diel and microhabitat effects Beginning on June 20, 1995 I monitored the removal of fruits at dawn and dusk over several 24 hour periods. This allowed me to distinguish the foraging activity of diurnal and nocturnal granivores. Granivorous birds concentrate their foraging efforts on open ground and exposed mineral soil (Bock et al, 1984). In turn, rodents forage preferentially under shrubs which serve as both cover from predators and in some cases as a food source (Price and Waser, 1985). I segregated the seed stations into these two microhabitats, under shrubs and in the open, to further distinguish the class of granivore responsible for seed removal. At each site a 200 m transect was placed in close proximity to the 7 x 7 live-trapping grid used for the rodent population study. Twenty Purshia shrubs were chosen systematically 14 along each transect. I marked seed stations at the base of each shrub and at a second station > 1.5 m away from the shrub canopy on bare soil. At dusk on the first day of the experiment, 30 fruits were placed at each station and at dawn the following day I recorded the number of fruits remaining. It was assumed that any missing fruits had been harvested without being husked or were dispersed passively by wind or rain. Rodents remove the outer husk of Purshia fruits before either placing the seed in their cheek pouches or eating the seed upon discovery. In the latter case rodents first remove the reddish seed coat exposing the nutritious white embryo and endosperm (Jenkins, 1988). Thus the presence of fruit husks and seed coats indicated seed harvesting by rodents and not birds. Fruits were replenished as required and checked again the following dusk. Stations were considered detected i f at least one fruit or seed was missing. Two sites were monitored in this fashion each day until all sites had been recorded for at least one diurnal period and two nocturnal periods. A l l observations took place between June 20th and June 28th, 1995 during the third-quarter moon phase. Over this period the moon was waning, decreasing in size and brilliance each night, until the new moon on June 27th. This also coincided with the period of highest fruit availability as the fruit crop ripened and fell starting on June 17th. Assuming that the activity of nocturnal mammals is highest under low illumination (cloud cover or new moon), rodent activity would have been at or near maximum (Price et al, 1984). At the time of my observations the only other large-seeded plants readily available at all sites were arrow-leaved balsamroot (Balsamorhiza sagittata) and needle-and-thread grass (Stipa comata). Ponderosa pine (Pinus ponderosa) occurred within sites OS and WT and were at the periphery of sites K B and CWS. 15 Fruit harvest: 4 weeks exposure Following the initial 24 hour observations seed stations were replenished and checked four weeks later on July 26th. At this time the number of fruits and the number of husks remaining at each station were recorded. Fruit harvest analysis The seed removal data were analyzed in three steps. Initially, I determined the effect of site and time (day versus night) on seed removal during the 24 hour observations. Secondly, because very few fruits were harvested during the day the effect of site and microhabitat were analyzed using only seed removal at night in four sites. Data from K B were not included because few seeds were removed during either period. A final analysis included data from all five sites collected after four weeks exposure. This analysis determined the effect of site and microhabitat on seed removal over this extended time period. I used linear models and weighted-least-squares estimation to test all main effects and interactions (PROC C A T M O D ; SAS, 1990). This is a procedure of categorical data modeling recommended for count data which compares response means in a manner analogous to analysis of variance (Agresti, 1984; SAS, 1990; Sokal and Rohlf, 1995). The Wald statistics provided by the program are reported here as Chi square statistics (X ). I used Pearson product-moment correlation ( r ) to measure the strength of the association between the number of rodents captured and seed removal as well as rodent captures and seedling clumps. 16 Natural seedling emergence In May 1995 and April 1996,1 surveyed 180 m at each site for newly emerged Purshia seedlings. Four sites were surveyed during the first year of the study and the fifth site, CWS, was included in the second year as it comprised part of the seed removal experiment. I systematically placed the center of 15 plots at 20 meter intervals along a transect running < 2 through each site. A total of 12 m was surveyed at each plot by a team of two, one worker on either side of a 1 x 1 meter quadrat. I recorded the number of spots from which seedlings emerged as well as the number of seedlings at each spot. To test for differences in the number of seedlings between sites and years I used a repeated measures analysis of variance (F) (SYSTAT, 1992). Results Fruit production Individual shrubs produced an average of 2,841.6 \u00C2\u00B1 458.7 fruits (\u00C2\u00B1 1 S.E., n= 20, range 236 - 9,336). Fruit production was lowest at site K B which was significantly different from the three other sites sampled (F = 8.13 d.f. = 3, p < 0.01, Figure 2.1). Fruit production ranged between 1.6 and 8.8 million fruits per ha. Based on an average fruit weight of 0.0295 grams this represented a total production of 48 to 261 kg per ha. This was comparable to production during a favourable year in California (227 kg per ha, Nord, 1965). 17 6000 5000 4000 \u00E2\u0080\u00A2g 3000 3 TJ O 3 L i . 2000 1000 0 f KB ~i r CWS (* no data available) WT Location BO OS Figure 2.1 Fruit production of Purshia tridentata (mean \u00C2\u00B1 1 S.E.M., n = 5 ) by individual shrubs at four study sites in British Columbia. Fruit harvest: diel and microhabitat effects The majority of fruit removal occurred at night, and from the accumulation of husks and at times seed coats, it was evident that nocturnal rodents were harvesting and eating fruits and seeds. The movement of fruits during the day was attributed to harvesting by chipmunks and birds. The analysis of seed removal data indicated a significant interaction between site and time (Table 2.2, Figure 2.2). This was attributed to the fact that very few seeds were removed from site K B during either period of observation, while at the remaining sites as many as 67 - 830 seeds were removed in a single night. Nocturnal foraging activity was influenced by microhabitat. There were significantly more seeds harvested from under shrubs than from in the open and this pattern was consistent 18 across the four sites considered (Table 2.3, Figure 2.3). Because rodents spend more time under shrubs they would have a higher probability of finding stations in that microhabitat (Price and Waser, 1985). As well, the high density of fruits under shrubs may have acted as an attractant. To remove this bias I repeated the analysis using only those stations that had been detected. The result remained the same, however, as average seed removal was significantly higher from under shrubs than from openings ( X ' = 12.85, d.f. = 1, P < 0.003). Fruit harvest: 4 weeks exposure The pattern of seed removal after four weeks was very similar to that observed during the initial 24 hour period. Sites differed significantly in the number of fruits harvested and the number harvested from under shrubs was significantly greater under shrubs compared to openings (Table 2.4). Of the original 1200 fruits supplied at each study site in June as many as 825 (KB) and as few as 102 (BO) remained four weeks later. Where many fruits were harvested large quantities of seed husks remained behind, indicative of rodent seed foraging. Remnants of seed coats were observed at many seed stations indicating that at least some seeds were eaten without being cached. By the end of the four week trial all stations at sites BO, OS and WT had been detected. At site K B not a single fruit had moved from at least 10 stations during the four weeks. Meanwhile, passive fruit dispersal was evident at CWS where the transect traversed a steep hillside ( > 20% slope). Here the vegetation consisted of discrete bunchgrasses and forbs between which the soil was either bare or covered in a layer of moss and lichen. Where seed stations were surrounded by patches of bare ground fruits were able to disperse downslope. 19 Table 2.2. Effect of site and time (day - night) on the removal of Purshia tridentata fruits by seed predators at five study sites in B.C., BO, OS, WT, CWS and K B . DF Chi-Square Prob Site 4 18.88 0.001 Time 1 12.84 0.000 Site*Time 4 31.88 0.000 2 5 , 20 J T 3 > o E a> i _ in T J a> a> w a> E 20 15 10 \u00E2\u0080\u00A2 Under \u00E2\u0080\u00A2 Open KB CWS WT BO OS Location Figure 2.3. The removal Purshia tridentata seeds in two microhabitats (under shrubs vs. in openings) by nocturnal seed predators at each of five study sites in B.C. (mean \u00C2\u00B1 S.E., n= 20). Site codes as in Figure 1.1. Thirty fruits were presented at each station. 21 Table 2.4 Effect of site and microhabitat on seed removal during four weeks of exposure to seed predators and abiotic factors. DF Chi-Square Prob Site 4 457.4 0.000 Microhabitat 9.97 0.002 Site*Microhabitat 4 5.36 0.252 Granivores Harvester ants do not occur in the south Okanagan valley (Jeff Jarrett, pers. com.) and resident ants were never seen taking Purshia fruits. At least one of several species of granivorous birds were observed in each of the five sites, including Vesper's sparrow (Pooecetes gramineus), Brewer's sparrow (Spizella breweri), chipping sparrow (Spizella passerina), mourning dove (Zenaida macroura) , rufous-sided towhee (Pipilo erythrophthalmus), California quail (Callipepla californica), black-billed magpie (Pica pica) and Clarks nutcracker (Nucifraga columbiana). Brewer's sparrows frequently fed on unripened Purshia fruits but after seeds ripened feeding was no longer observed, presumably because they were dry, hard and no longer palatable (Mike Gil l , pers. com.). Four species of nocturnal rodents, Great Basin pocket mice (Perognathus parvus), deer mice (Peromyscus maniculatus), montane voles (Microtus montanus) and western harvest mice (Reithrodontomys megalotis), were captured during the 16 trapping periods (8 periods per year) between May and October 1994 and 1995 (data courtesy W.F. Klenner). Pocket mice and deer mice comprised 95 % of all captures (Appendix, Table A . l ) . The number of fruits harvested at 22 each site over the four weeks was highly correlated with the average number of rodent captures that same year (r = 0.966, d.f. = 3, p < 0.01, Figure 2.4). The diurnally active Yellow-pine chipmunk (Tamias amoenus) was captured at only one site (OS) where stands of trees, Pinus ponderosa, were intermixed with the shrub-steppe community. 1200 TJ > O E a> in TJ in 1000 J 800 J 600 J 400 200 J 10 20 30 40 Rodent captures 50 60 Figure 2.4. The relationship between the total number of Purshia tridentata seeds removed during a 4 week trial and the average number of rodents captured during 8 trapping weeks at five study sites (r = 0.966, d.f. = 3, p < 0.01). (Capture data courtesy of W.F. Klenner). Natural Seedling emergence I use the term 'seedling spot' to designate the isolated points from which seedlings emerged, whether singly or in groups from seed caches. To further distinguish solitary seedlings from clusters, I refer to the former as singles, and the latter as clumps. Comparing the number of spots between four sites resulted in a significant interaction between site and year (Table 2.5, 23 Figure 2.5). For example, in 1995 site K B contained the most seedlings while the following year it was third in abundance. The occurrence of seedlings at K B , particularly in 1995 (58 singles and 5 clumps) was surprising given the paucity of rodents there. In contrast the site with the most rodents, OS, had only 5 singles and 1 clump over the same period. Combining the data for all sites, I observed a total of 107 and 167 spots in 1995 and 1996 respectively. Although an additional site was included in the 1996 census (CWS) that site only contributed one seedling to the total. The majority of seedling spots contained a single seedling, making up 87.9 % and 70 % of all spots over the two years. The balance comprised clumps of 2-7 individuals growing together with an average clump size of 3.25 \u00C2\u00B1 0.17 (mean \u00C2\u00B1 1S.E.). Seedling densities ranged from 0.01 \u00C2\u00B1 0.01 - 0.64 \u00C2\u00B1 0.10 seedlings m\"2 (mean \u00C2\u00B1 1 S.E.). Using average seed production per hectare I estimated the survival of seeds from production to seedling emergence as 0.001 - 0.028 %. The number of seedling clumps in 1996 was not associated with the number of seeds removed during the seed station experiment in 1995 or with the number of rodent captures, neither Perognathus parvus, Peromyscus maniculatus or both species combined (Figure 2.6). 24 Table 2.5. Anova for the effect of site and year on the number of Purshia tridentata seedling spots over two years, 1995 and 1996, for four study sites in British Columbia, K B , WT, BO and OS. DF SS MS F Prob Site 3 280.6 93.5 12.1 0.000 Year 1 19.7 19.7 2.7 0.107 Site*Year 3 398.5 132.8 18.2 0.000 Error 53 387.8 7.3 o 70 -I 60 50 -I 40 \u00C2\u00A3 30 J 20 . 10 . 0 1995 I I Location Figure 2.5. The number of spots containing single seedlings (shaded bars) and clumps of seedlings, 2-7 individuals respectively (open bars). Locations listed in order of increasing rodent population size (W.F. Klenner unpublished). Results from fifteen plots (total 180 m 2 site\"1) in May 1995 and April 1996. (* no data available for that year). 25 , , ^ 0 20 40 60 Rodent Captures Figure 2.6 The total number of Purshia tridentata seedling clumps in 1996 at each of five study sites in relationship to the number of pocket mice (P. parvus), deer mice (P. maniculatus) and both species combined. Discussion Fruit harvest The relative role of nocturnal and diurnal predators The majority of seed harvesting and predation was accredited to nocturnal rodents, deer mice and Great Basin pocket mice. This is supported by i) the observation that most fruits were removed at night from beneath shrubs and ii) the strength of the correlation between fruit harvest and the number of rodents captured. This is consistent with our knowledge of food preferences shown by these rodents (Everett et al, 1978; Jenkins and Ascanio, 1993). The presence of husks and seed coats scattered in and around seed stations was consistent with the handling of Purshia fruits by these rodents (Jenkins and Ascanio, 1993; Vander Wall, 1993). In general, these results correspond to a number of other grassland studies that classified rodents as CD E 3 50 40 30 20 10 0 0 20 40 60 Perognathus parvus 0 20 40 60 Peromyscus maniculatus 26 the primary seed predators (Borchert and Jain 1978; Reichman 1979; McAullife, 1990; Hulme, 1994). Given the abundance of seeds, rodents would have harvested more than they ate and cached the surplus (Longland and Clements, 1995). Pocket mice, with their extensive cheek pouches, are capable of carrying large quantities of seeds and have been credited with the dispersal of numerous plant species (Morton et al, 1980; McAuliffe, 1990; Longland and Clements, 1995). This designates pocket mice as the principal animal-dispersal mechanism of Purshia at these sites in the Okanagan valley. The foraging behavior of these rodents dictates that seeds wil l be dispersed to a maximum of 50 meters from the source shrub, but usually much less (Longland and Clements, 1995). Furthermore, I predicted that the quantity of seeds cached would be reflected in the number of seedling clumps emerging in the spring. Despite the presence of granivorous birds at all sites and the diurnally active yellow-pine chipmunk at site OS, the removal of fruits during the day was extremely limited. This is in stark contrast to a seed station experiment using Purshia and five other native species, in which day time removal of seeds was equal to or greater than that which occurred at night (Kelrick et al, 1986). In that study the use of elevated Petri dishes may have attracted the attention of visually oriented seed predators (birds, chipmunks) more than the inconspicuous seed stations used in my study. As well, birds may shift their foraging activity with seasonal changes in food availability and preference (Kelrick et al, 1986). Pulliam and Brand (1975) recognized that birds, sparrows in particular, foraged on insects during the summer months and fed on seeds when insects were no longer abundant (Sept. - March). Because birds are capable of transporting seeds a much greater distance than rodents their role in dispersal can be highly significant (Hutchins et al, 1996) especially in fragmented habitats (Auclair and Cottam 1971). Future 27 research should determine whether birds actively forage for Purshia fruits during the fall and whether they pass Purshia seeds undigested in their faeces. Vander Wall (1994b) observed that most Purshia fruits were removed during the day supporting his conclusion that chipmunks were responsible. Although I anticipated that chipmunks would play a similar role in the Okanagan I found no evidence to support this. The availability of another food source, Pinus ponderosa, may have diverted the attention of chipmunks. In another study chipmunks preferred seeds of the congener Pinus Jeffreyi, which provided 3.8 times higher energy intake compared to Purshia (Vander Wall 1995a). Perhaps the cone bearing P. ponderosa at site OS attracted chipmunks more than the fruit bearing Purshia. This could explain the lack of diurnal foraging on Purshia over the course of the seed station experiment. As Purshia is distributed from the valley bottom up into the hillsides the shift from grassland to forest would involve a dramatic shift in the rodent community as well. Thus, the role of yellow-pine chipmunks and pocket mice as agents of seed dispersal may change along this gradient. Microhabitat Animals harvested significantly more seeds from stations beneath shrubs than from those placed in the open. This was consistent with the observation that quadrupedal rodents, such as pocket mice and deer mice, forage preferentially under cover to increase predator avoidance and foraging efficiency (Brown et al, 1979; Price et al, 1984; Price and Waser, 1985). As a result, seeds in covered microhabitats have a greater chance of being eaten than those in the open, or inversely, once seeds are dispersed into the open they have a greater chance of survival. Herrara (1984) and Hulme (1994) found a similar pattern of predation as seeds placed under cover of vegetation were more likely to be predated by rodents. I repeated the analysis of seed 28 removal over nightly exposures including only those seed stations that had been detected. This would determine i f microhabitats differed in seed removal once the rodents had discovered the station. The differences in seed removal between microhabitats remained significant, further supporting the conclusion that fruits under shrubs were at a greater risk of predation than those in the open. Seedling emergence As a measure of disperser effectiveness The seed station experiment showed that the sites varied enormously in the quantity of seeds removed by rodents. In turn, counting emergent seedlings provided a measure of the quantity and quality of seed dispersal as only seeds in suitable microsites would emerge as seedlings (Young et al, 1993). I anticipated that if rodents were the principal mode of seed dispersal, and in the absence of other factors, seedling emergence should increase with the size of the rodent population. This did not bear true, however, as the number of seedling clumps did not follow any apparent pattern. Clearly other factors did affect seedling emergence and acted to obscure any pattern in cache density afforded by the different rodent densities. Whelan et al (1991) suggested that the recruitment of vertebrate-dispersed plant species is subject to several random and unpredictable factors including soil disturbance, weather patterns, and great fluctuations in seed predation due to migratory birds and insect population irruption. This leads to the conclusion that finding patterns in seedling establishment caused by rodent seed dispersal may require longer term study and techniques that can determine the seed shadow created by rodent seed caching (e.g., Longland and Clements, 1995; Vander Wall, 1994b). Despite the absence of any obvious relationship between the size of the rodent populations and seedling emergence two trends in the data warrant mention. First, the overall 29 survival of seeds from production to seedling emergence was 0.02 - 0.001 %. This was two to three orders of magnitude less than the estimate of 1.1 - 2.4 % provided by Vander Wall (1994b). This could indicate differences in quantity of seeds dispersed by animals or differences in climatic and edaphic conditions between these separate study areas, or both. Secondly, based on the abundance of seedling clumps dispersal effectiveness was highest at WT over both years. This site had a moderate sized population of pocket mice during this period, supporting the conclusion that this species was an effective disperser of Purshia in the Okanagan. The next most common species, deer mice, are known to hoard food (Barry 1976) but their role in seed dispersal is not well documented. Vander Wall (1994b) concluded that deer mice were not effective seed dispersers as they tended to eat seeds rather than cache them. The lack of seedling emergence at site OS, which contained the largest population of deer mice, may reflect their appetite for Purshia seeds (Everett et al, 1978). However, the closely related white-footed mouse (Peromyscus leucopus) was found to be a highly effective disperser of white pine (Abbott and Quink 1970). The abundance and ubiquitous distribution of deer mice warrants further research into their overall impact on Purshia populations. Seedling emergence compared to other regions Seedling spots in the Okanagan were characterized by a high percentage of single seedlings (> 70%). In other regions the percentage of single seedlings was as low as 14.2 % (Vander Wall, 1994b). Although not conclusive evidence, the large number of single seedlings suggests that passive dispersal of Purshia fruits by wind and rain could be more important here than compared to other regions (e.g., Nord, 1965; West, 1968; Evans et al, 1983; Vander Wall, 1994b). Rodents remove the husk before placing them in their cheek pouch (Vander Wall, 1995b) and it has been suggested that such handling is necessary for seeds to germinate (Young, 30 1988). The occurrence of at least two seedlings from seeds with the husk still intact indicates that Purshia is capable of germinating without the assistance of rodents. Vander Wall (1994b) also found numerous seedlings emerging from seeds with the husk attached or in close proximity. Possible modes of passive seed dispersal are discussed below. The emergence of seedling clumps in the Okanagan was also distinguished by the relatively low number of individuals per clump. Average clump size for two years data combined was 3.25 \u00C2\u00B10.17 (1 S.E.) with a range of 2-7. This was considerably smaller than the average clump size and range reported in the literature: Oregon, 10.8 seedlings per clump, range 2-55 (Stanton in West, 1968); 11.6 range 2 -50 (West, 1968); California 12.5, range 7 -31 (Evans et al, 1983) and Nevada 12.1, range 2-104 (Vander Wall 1994b). Two factors could contribute to differences in the sizes of seedling clumps observed in the Okanagan compared to elsewhere. First, differences in the caching behaviour of species could lead to the differences in seedling clump size (Vander Wall, 1990). Diurnal rodents such as the Yellow-pine chipmunk and golden-mantled ground squirrel (Spermophilus lateralis), both of which occur in other parts of Purshia's distribution, are larger than any of the nocturnal rodents in the Okanagan. Body size will affect the number of seeds an animal is capable of carrying (Morton et al, 1980) and could affect the size of seed caches. Secondly, climatic differences will contribute to the differences in seed survival and the size of seedling clumps observed. In the Okanagan Purshia fruit ripen in early summer while at the higher elevations it ripens 1-2 months later (Vander Wall, 1994b). This represents an additional period of time over which seeds in the Okanagan are exposed to seed predators. Over such an extended period seed caches could be visited repeatedly until few or no seeds remain. As well, microsite conditions over the winter months determine whether seeds will survive and 31 germinate (Young et al, 1993). Conditions under the snow pack are particularly suitable for germination and seedling emergence (S.E. Meyers, pers. comm.; Young et al, 1993). At low elevations in the Okanagan valley snowfall is typically limited to < 60 cm a year and rarely stays on the ground for long periods of time. Even where seed caches contained large numbers of seeds few seedlings would emerge i f stratification requirements were not met over the winter months. At least two of the locations cited above (West, 1968; Vander Wall, 1994b) occurred at high elevation (circa 2,000 m) where the snow pack is typically greater than in the Okanagan valley. Alternative modes of seed dispersal Undoubtedly, rodents are effective at moving seeds beyond the shrub canopy and placing them in mineral soil at a depth suitable for germination and emergence (Evans et al, 1983). Vander Wall (1994b) showed that chipmunks cache seeds away from the shrub canopy in 50 % of all cases. However, they need not be the only means of dispersal, especially where rodent populations are low and where forces exist that can effectively redistribute plant material, including seeds (Orndorff and Lang, 1981). I observed events that suggest alternative modes of seed dispersal may operate in Purshia regeneration. At two neighbouring sites, K B and CWS, large numbers of fruits remained unharvested, even after four weeks of exposure. At site K B where the ground was level, fruits remained beneath the shrub canopy where they germinated and emerged as single seedlings. At site CWS the uneven terrain favored the downslope dispersal of fruits presumably assisted by wind and rain. Few seedlings were observed here, however, due to a lack of microsites suitable for germination (Fenner, 1987). From these observations I conclude that alternative modes of dispersal are possible and should be considered in addition to the role of animals as factors contributing to Purshia regeneration. 32 Further evidence of passive seed dispersal was observed during intense rain showers in the Okanagan. On at least one occasion, surface run-off resulted in the movement of large quantities of sediment and leaf litter (pers. obs.). The physical properties of seeds are similar to that of leaf litter. This material accumulates around obstacles, in depressions and frequently forms conspicuous features of the landscape (Noy-Meir, 1985; Facelli and Pickett, 1991). Where the movement of seeds is accompanied by the movement of soil, seeds may become buried in sites suitable for germination and seedling establishment (Westelaken and Maun, 1985). The relative role of these factors will be determined by both the local terrain and weather patterns (Orndorff and Lang, 1981). Predictably, this type of dispersal will be highly directional, that is downslope, and fairly limited in terms of distance in comparison to seed dispersal afforded by rodents. Management Implications Brown and Heske (1990) provided conclusive evidence that rodent populations play an important part in structuring grassland habitats. Reynolds (1958) observed that kangaroo rats (Dipodomys merriami) have a positive impact on plant populations during years of high seed production and a negative impact during years of poor production. Attempting to put such knowledge to use Chambers and MacMahon (1994) suggested that where high densities of seed-dispersing rodents occur it could be possible to encourage the regeneration of native plant species. The results of the current study emphasize how difficult it may be to implement such tactics. M y observations occurred over two years of high fruit production, yet an increase in seedling numbers corresponding to the variation in rodent population size did not occur. Even at sites with few rodents some seeds were effectively dispersed while at sites with large rodent populations the effect was both positive and negative. Site specific factors appear to have 33 interacted with seed dispersal to determine the overall pattern of seedling emergence. The pathway that seeds follow from production to recruitment is perhaps too complex (Price and Jenkins, 1986) and subject to too many stochastic factors (Whelan et al, 1991) to allow us to predict the effect of rodents on plant populations. Although managing rodent populations may be of limited gain in the short term (years) it is important to maintain rodent populations over the long term (decades or centuries). If pocket mice are in fact the primary animal dispersal agent in the lower valley the persistence of Purshia may be dependent on the persistence of pocket mice. This is particularly true of locations where the terrain does not allow alternative modes of seed dispersal as described above. The role of animal dispersal will be particularly important following large scale disturbances. However, as the Okanagan landscape becomes more fragmented the dispersal distance achieved by rodents may become insufficient. Following a fire the distance between available seed sources could be extensive. If an isolated stand were completely destroyed by fire seed dispersal by rodents could be unable to bridge a wide expanse particularly across agricultural land (Hobbs, 1987). If pocket mice are eliminated or stands are widely separated, direct seeding or the planting of seedlings may be the only means of reintroducing or enhancing shrub populations. The variation in seed removal between microhabitats as well as the correlation between seed removal and rodent population size illustrates the utility of this simple technique. Monitoring seed removal experiments could provide useful information regarding differential microhabitat use, temporal variation in food requirements, population size and animal catchability. Hulme (1994) was able to show variation at the site level by correlating seed removal with number of rodents captured in the vicinity of separate seed stations. Brown et al (1988) quantified seed removal rates to determine the microhabitat use of various nocturnal 34 rodents and their response to increased risks of predation by barn owls. The use of Purshia seeds is particularly well suited for monitoring rodent populations in the Okanagan because it is both native and highly sought after. This technique could be easily modified for a variety of research purposes and could be used to monitor rodent populations across areas that would normally require monumental efforts (e.g., MacMahon et al, 1989). Summary Monitoring seed removal was a simple yet effective means of distinguishing various classes of granivores and their role in seed predation and dispersal. Nocturnal rodents appear to be the most influential class in this system, at least in June and July, the time of maximum fruit abundance. In comparison, the role of diurnal granivores, birds and chipmunks, was of minor significance. The survey of emergent seedlings documented the effective dispersal of Purshia seeds by animals in the Okanagan shrub-steppe ecosystem. Pocket mice were implicated as the most important agent of seed dispersal among the sites in this study. Although seed harvesting followed a predictable pattern correlated with a tenfold range in rodent population size the pattern of seedling emergence was complicated by other factors. The large number of single seedlings and relatively small size of seedling clumps was in contrast to patterns of seedling emergence reported from other regions. Whether this is due to differences in caching behaviour of the rodents or climatic differences is not presently known. Alternatively, single seedlings can result from passive seed dispersal in which case this mode of dispersal may be more important than has been previously noted. Evidently patterns of seedling emergence and the animal species responsible for dispersal vary across the range of habitats occupied by Purshia. 35 The numerous factors that interact with seed dispersal make it difficult to assess and predict the overall impact of rodents on plant populations. The size of the seed crop and of the resident rodent population are not necessarily good predictors of seedling emergence. Despite these limitations seed removal experiments can be adapted to various research questions, including large scale projects not amenable to the labor intensive methods currently in use. 36 Chapter 3 Seedling emergence and survival Introduction Once seeds are released from the parent their fate is coupled to the particular microenvironment into which they disperse (Keddy, 1981). The various microsites that they occupy will differ in moisture, temperature, soil texture, light level and other parameters that determine their chances of germination and survival. However, not all seeds are created equal and it has long been recognized that the environmental conditions required to initiate germination varies among species (Harper et al, 1965; Oomes and Elberse, 1976; Bewley and Black, 1985). Even seeds of the same species may differ in their response to a given environment and ecologists have noted marked differences between populations and between cohorts from the same population with regard to germination and seedling emergence (Schmidt and Levin, 1985; Gutterman, 1992). This is because factors intrinsic to the seed itself, attributes such as size (Wulff, 1986), nutrient content (Parrish and Bazzaz, 1985) and genotype (Schaal, 1984) will determine its ability to cope with a given set of conditions. Although the relative role of the environment and the attributes inherent to different seed provenances have long been recognized by foresters, the population genetics of rangeland plants has only begun to be appreciated (Meyer and Kitchen, 1994). The existence of population differentiation for phenotypic traits has been shown by comparing the performance of plants from two or more sources grown in a common environment (Clausen et al, 1940). Taking this a step further, comparing the fitness of different populations when grown in each others as well as their native environment (a reciprocal transplant) researchers have shown that this differentiation is in fact adaptive (Bradshaw, 1984). This has lead to a debate over how this variation is maintained and how is it structured into the population, i.e., is the variation greatest 37 between populations, between sibships or within sibships? Answers to these questions relate to an even larger issue, how will this variation affect a species ability to invade and reproduce in a given habitat and persist in that habitat as it changes over time? (Bradshaw, 1984; Antonovics and Via, 1988; Linhart and Grant, 1996). Purshia tridentata is an important browse plant in rangelands making it a valuable component of many habitats throughout western North America. In order to foster the regeneration of Purshia by natural and artificial means, researchers have focused on the environmental conditions required to break seed dormancy. This has been undertaken through laboratory tests (Young and Evans, 1976; Evans and Young, 1977; Meyer and Monsen, 1989) and more recently in the field (Young et al, 1993). While the role of the environment has received considerable attention it is equally pertinent to ask how variation between the seed sources interacts with the environment. Meyer and Monsen (1989) tested several populations but found no differences in their response to various temperature and chemical treatments. I was interested in whether differences between local populations could be observed at the seedling stage under field conditions, and i f so, could they contribute to the tenfold difference in regeneration that was measured between stands of Purshia in the Okanagan (Chapter 1). I relied upon the artificial introduction of seeds as this has proven to provide explicit and interpretable information regarding the demography of plants in the field (Keddy 1981). To investigate the effect of seed source on seedling emergence I incorporated a reciprocal transplant to determine the amount of variation in seedling emergence within and between four local populations in the Okanagan valley. From each of the populations I randomly selected 20 maternal families (seeds originating from a single shrub). Each was characterized by the average mineral content (% magnesium and nitrogen by weight) as determined by Krannitz (in press b). 38 I also investigated environmental constraints to seedling establishment. Previous research suggested that soil texture (Mustart and Cowling, 1994) and moss cover (Gold and Bliss, 1995) affected seedling emergence through their influence on soil moisture. Therefore, I characterized each planting location (i.e., block) according to these edaphic factors. Materials and Methods Study species Purshia tridentata occupies well drained soils throughout arid valleys in the intermountain region of western North America. Individuals may live to be more than 100 years old (Nord, 1965). Plants are commonly killed by fire although some individuals or populations are prone to resprouting. Regeneration is most commonly from seed. Purshia fruit consist of an oblong achene, a single seed surrounded by a papery husk or pericarp (Figure 3.1).Seeds are relatively large with an average weight of 27 mg (Schopmeyer, 1974). Approximately 30% of this consists of the outer seed coat, which protects an embryo high in lipids, proteins and carbohydrates (Jenkins, 1988). Seeds are dormant at the time of dispersal in mid to late June. Dormancy is broken when seeds imbibe water and are exposed to cool temperatures for a period of >4 weeks (Meyer and Monsen, 1989). This mechanism serves to link seed germination and seedling emergence to environmental cues signaling the coming of spring (Young et al, 1993). 39 7mm Pericarp Seed coat Cotyledon Hypocotyl Radicle 0 [ actual size L J ] Figure 3.1. Purshia tridentata: longitudinal section of fruit. Adapted from Schopmeyer 1974. The study was conducted on four sites in the Okanagan valley, near the town of Oliver, British Columbia. Mean annual precipitation is 30.8 cm/yr. and mean monthly temperature ranges from -2\u00C2\u00B0 C (January) to 22\u00C2\u00B0 C (July). Four sites, separated from one another by 1-17 km, were included in the planting experiment. A l l were located in the shrub-steppe community dominated by mature stands of Purshia and perennial bunchgrasses; blue-bunch wheatgrass (Agropyron spicatum), red three-awn (Aristida longiseta), sand dropseed (Sporobolus cryptandrus) and needle-and-thread grass (Stipa comata). A previous survey of juvenile Purshia (non-flowering individuals) had identified a tenfold difference in current regeneration between sites. Burrowing Owl site (BO - 49\u00C2\u00B007' N , 119\u00C2\u00B033' W) and Orchard Site (OS - 49\u00C2\u00B009' N , 119\u00C2\u00B032' W) were on the outskirts of Oliver as was Water Tower (WT - 49\u00C2\u00B0 10' N , 119\u00C2\u00B031' W). Kennedy Study area 40 Bench (KB 49u15' N , 119u30' W) was the furthest north being approximately 5 kilometers north of Oliver. Seed collection In July, 1994 several hundred fruits were collected from 5 shrubs at each of the four study sites. The pericarp was removed by hand rubbing the fruits on rubber mats and seeds were stored in a freezer from September, 1994 until February, 1995. The collection consisted of a minimum of 400 full seeds with no visible defects, i.e., insect damage. The identity of all seeds and subsequent seedlings was maintained throughout the experiment. Seeds originating from the four sites are referred to here as separate populations and seeds from the 20 individual shrubs as maternal families. Experimental design Seeds were planted reciprocally among the four study sites, that is, seeds from each site were planted at the site of collection as well as at the other three sites (sensu Clausen et al, 1940). Planting took place between February 22-25th, 1995 at BO, OS and WT. Freezing conditions delayed the planting of K B until March 12th. The planting comprised a randomized complete block design with five replicate blocks per site (Figure 3.2). Each block was divided into four 1 x 1 m plots, one per population. Blocks were randomly placed along a 400 meter transect established through each site with a minimum distance between adjacent blocks of 20 meters (max. approx. 200 meters). A minimum distance of 1 meter was maintained between the border of the planting block and the drip line of adult shrubs. 41 Seed source Block 1 x 1 m plot Figure 3.2. Schematic view of planting blocks, plots and subplots for two sites of the reciprocal planting experiment. Each subplot received 20 seeds from one maternal family. Each plot received 100 seed from a single population. 42 Seed planting, relocation and measurement The corners of each 1 x lm plot were marked by a 10 inch galvanized spike. At the time of planting a wooden frame was positioned over the plot such that its corners aligned with the four spikes. A cross beam with pointers at 10 cm spacing was aligned with each planting row and seeds planted at the location of each pointer. Seeds were planted 10 cm apart at a depth of 2.25 cm. Planting was achieved by pushing a 5 mm diameter tube and plunger to this depth, removing the plunger, dropping a single seed down the tube and replacing the plunger before extracting the entire apparatus. Seeds and seedlings were relocated by repositioning the quadrat over the spikes and aligning the cross beam over each planting row. With this apparatus planting position could be relocated within 2 cm. At the time of planting soils were moist and easily penetrated by the planting tube. Plots were checked the last week of April, the first week of May and again at three week intervals until July 27th, 1995. Seedling emergence and survival were recorded on each date. As seedling emergence at site K B was extremely limited a sub-sample of seeds that had been planted on the same date (but for a separate experiment) was removed in May 1995 and tested for germination and viability (Tetrazolium test). In spring 1996, 14 months after sowing, plots were monitored once again for seedling survival as well as newly emerged seedlings. Edaphic measurements I visually estimated percent cover of moss, lichen, grasses, annual and perennial herbs, litter and bare ground at each planting plot. Cover estimation was assisted by the use of a 1 x 1 meter frame dissected by string into 25 cells. Soil texture and percent moisture content were determined for each block. Sampling used a 1.25 cm (diameter) core taking only the portion 43 between 15 - 20 cm deep, the rooting depth of 6 week old seedlings (pers. obs.). Soil cores were taken on April 27th, May 8th and June 1st, 1995. Soil texture was determined by Norwest Labs, Vancouver and soil moisture was determined gravimetrically (Brady, 1990). Analyses A n analysis of variance was conducted to determine the effect of planting location (site and block) and seed source (population and maternal family) on seedling emergence. A small percentage (1-1.5%) of seeds planted in 1995 emerged 14 months later in April 1996. This was included in an initial analysis to determine if there was significant variation between families for seed banking (seeds not capable of germinating in 1995) hence justifying the use of covariate analysis. This variation did not have a significant effect on the treatments hence it was not included in subsequent analyses. To test for evidence of differential responses among populations or amongst maternal families I included the interaction terms, site of growth x population and site of growth x maternal family. A significant interaction would indicate that seedling emergence was determined by the joint effect of planting location and seed source, providing evidence of local adaptation among populations (Schemske, 1984). The analysis of variance was performed using PROC G L M (SAS, 1990) with a mixed model including four main factors and two interaction terms. Fixed factors included Site and Population and random factors included block and maternal family. As the proportion of the variation accounted for by each treatment was the primary interest (rather than indicating which treatments were significantly different) I calculated the percent of the total variance explained by each factor. (Treatment Sum of Square x 100 / Total Sum of Square = Percent variance explained). 44 The analysis of soil moisture data determined both the effect of soil texture (loamy versus sandy) and their change over three sampling periods (April, May, June) using a repeated measures analysis of variance (Systat, 1992). Environmental factors were investigated by correlating total emergence in blocks with their corresponding edaphic measurements (Pearson's product-moment correlation coefficient, r; Zar, 1984). In order to normalize the distribution of percent values, these were arcsine square root transformed prior to analyses (Zar, 1984). Significant differences are reported using P< 0.05 unless stated otherwise. The relationship between seed attributes (% Mg and % N content) was tested using the Spearman rank correlation coefficient rs. Results Planting date Percent emergence was considerably higher at planting sites seeded in February 26.02 % \u00C2\u00B1 2.88 (mean \u00C2\u00B1 1 S.E., n=15) than for those seeded in March 1.35 % \u00C2\u00B1 0.36 (n=5). Low emergence at site K B was attributed to the later planting date and a sub-sample of seeds was extracted and tested for viability. A l l seeds remained viable though germination in the dark at fluctuating temperatures was poor (10-15 %). Data from K B were not included in subsequent analyses. Weather data from Environment Canada (Oliver station) indicated that precipitation was above average in January and March, 1995 while temperatures were typical of that time of year (Figure 3.3). 45 J F M A M J Date Figure 3.3. Weather data from January - June, 1995 with precipitation compared to 70 year average. Source Environment Canada, Oliver sewage treatment plant station. Edaphic factors Within each site planting blocks varied considerably in soil texture and percent moss cover. Both characteristics were found to correlate with seedling emergence. Moss cover varied from widely spaced clumps to contiguous carpets that covered the soil surface between perennial plants. Percent moss cover of planting blocks ranged from 3 - 50% and was positively correlated with seedling emergence (r = 0.713, d.f. = 13, P < 0.001, Figure 3.4A). Typical mosses included; Tortula ruralis, Ceratodon purpureus, Polytrichum piliferum and Bryum caespiticium (Lynne Atwood, pers com.). Following the classification of Wittneben (1986) soils in all planting blocks were coarse to moderately coarse in texture. Maximum clay content was 2 % while silt content ranged from 5-45 %. BO soils consisted of sand to loamy sand with a high coarse fragment of small cobbles 46 2-5 mm in diameter. OS soils were the most variable ranging from Sand to Sandy Loam while soils at WT were consistently high in silt content. The high affinity of soil moisture for fine textured soils and the dependence of germination on adequate moisture suggested that seedling emergence would be positively associated with silt content (Mustart and Cowling, 1993). However, the relationship that I observed was the opposite as seedling emergence and silt content of soils were negatively correlated ( r = - 0.65, d.f. = 13, P <0.01, Figure 3.4B). 47 T3 d> 200 5 150 . Q E 3 a> u \u00C2\u00A7 100 E? a> E 100 D) L _ Q) E 0) 50 10 20 30 40 50 % Silt (arcsin sqrt transformed) Figure 3.4. The relationship between; A Purshia tridentata seedling emergence and moss cover across 15 planting blocks. (r = 0.71, d.f.= 13, P < 0.001) and B between seedling emergence and silt content of soils (r = -.065, d.f. = 13, P < 0.01). A total of 400 seeds were planted at each block in February, 1995 and emerged in April and early May 1995. 48 Table 3.1. Anova table for Purshia tridentata seedling emergence data for seeds reciprocally planted on three sites BO, OS and WT. Model R 2 = 0.62. Column F indicates the Mean square values used in the F test for each treatment. Source of variation DF Sum of Squares Mean square F P Variance (%) A Site of Growth 2 419.6 209.8 A/B 0.129 11.7 B Block (Site of Growth) 12 1030.6 85.9 B/G 0.003 28.8 C Population 3 87.6 29.2 C/D 0.424 2.5 D Family(Population) 16 473.8 29.6 D/F 0.000 13.3 E Site of Growth*Population 6 28.4 4.7 E/F 0.531 0.8 F Site of Growth*Family(Pop) 32 172.6 5.4 F/G 0.621 4.8 G Error 227 1361.4 5.97 38.1 Site of Growth Considerably more variation in seedling emergence was due to differences within sites than to differences between sites. The between site variation accounted for only 11.7% of the total variation while within site variation (blocks) accounted for 28.8%, the highest amount of the variation explained by any single factor (Table 3.1). No significant differences between study sites were detected (F = 2.44, d.f.= 2, P = 0.129, Figure 3.5). Populations and maternal families The variation within populations (between maternal families) was greater than that between populations - 13.3 % versus 2.5% (Table 3.1). Seedling emergence was not significantly different between populations (F= 0.99, d.f. = 3, P > 0.42, Figure 3.5). Differences between families were, however, very highly significant, (F = 4.96, d.f. = 16, P < 0.0001, 49 Figure 3.6). For example, families from site WT ranged in percent emergence from 15.7 % \u00C2\u00B1 2.92 (mean \u00C2\u00B1 S.E.) to 37.3 % \u00C2\u00B15.15. The interactions, Site of Growth x Population and Site of Growth x Maternal Family, were not significant indicating that there was no change in the relative performance of populations or maternal families across planting locations. Unlike measuring seed weight, it is not possible to characterize individual seeds for their nutrient content before planting. Hence I used the magnesium (Mg) and nitrogen (N) (expressed as % N of total seed weight) for each maternal family determined by Krannitz (in press b). This was determined from an analysis of 4-6 seeds per maternal family. No relationship was found between Mg content and the percentage of seeds that emerged. However, there was a negative correlation between average N content and seedling emergence between families. This relationship was significant at site BO (rs = -0.48, d.f. = 20, P < 0.05,) and OS (rs = -0.53, d.f.= 20, P < 0.05) but not at site WT (rs = -0.36, d.f.= 20, P > 0.10) where emergence was generally lower (Figure 3.7). 50 10 9 : OS BO WT Site of Growth Figure 3.5. Seedling emergence (mean \u00C2\u00B1 S.E.) of four populations of Purshia tridentata reciprocally planted among three sites in British Columbia. Site codes as in Figure 1.1. Average emergence from 20 seeds planted at each subplot, n = 25. 10 j 9 I B K O W O B S T Population Figure 3.6. Average seedling emergence (mean \u00C2\u00B1 1 S.E.M.) for Purshia tridentata from four populations reciprocally planted at three sites, BO, OS and WT in British Columbia. Average emergence from 20 seeds planted at each subplot, n = 15. 51 O S B O WT 4 5 \u00C2\u00AB . 3 5 \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 4 5 3 5 \u00E2\u0080\u00A2D CD tt> m 2 5 1 5 \u00E2\u0080\u00A2 2 5 1 5 4 5 3 5 2 5 1 5 9 7 9 7 % Nitrogen (arcsin sqrt transformed) Figure 3.7. The relationship between average % nitrogen content (of Purshia tridentata seeds) from each of 20 maternal families and the % seedling emergence at three separate planting sites. Survival The investigation of seedling survival was limited to environmental factors as the high mortality resulted in a sample size too small to test for differences between populations or maternal families. Less than 30 % of the 6,000 seeds planted survived to produced seedlings and less than 1% survived to establishment (1 year old). Survival of seedlings over the first weeks and months was very low with > 60% of all seedlings dying by early June, 1995. The largest proportion of mortality was attributed to water stress (desiccation) although grazing by rodents and trampling by horses was evident (Table 3.2). Below average rainfall during April and May (Figure 3.3) could have contributed to the high mortality over this time period. Moisture levels in the soil layer 15-20cm deep decreased significantly between April 27th and June 5th (F= 174.23, d.f.= 2, P < 0.001). However, the water level in loamy soils was consistently higher than that in sandy soils (F= 95.76, d.f.= 1, P < 0.01) (Figure 3.8). This variation between soil types was presumed to be responsible for the 52 positive correlation between survival and the silt content of soils (r = 0.676, d.f.= 13, P < 0.02, Figure 3.9). This was in contrast to the inverse relationship between soil texture and seedling emergence. The net result in terms of seedling establishment, i.e., 1 year old seedlings, was 1.89 \u00C2\u00B11.1 (mean \u00C2\u00B1 1 S.E., n= 9) in sandy soils and 6.0 \u00C2\u00B1 1.7 (n= 6) in loamy soils and this difference was highly significant (t = 4.85, P < 0.01). Table 3.2. Frequency of mortality between April 27th and July 26th, 1995, for Purshia tridentata seedlings at three study sites in British Columbia. Site codes as in Figure 1.1. Site of Water Unknown Grazed Trampled Growth stress BO 278 124 17 0 OS 404 89 81 2 WT 239 37 6 10 10 n April May June Date Figure 3.8. Moisture loss in loamy ( \u00E2\u0080\u00A2 > 20%) and sandy ( \u00E2\u0080\u00A2 0-20%) soils between April 27 and June 5th, 1995 for three sites in the Okanagan Valley, British Columbia. 53 Figure 3.9. The relationship between silt content of soils and the percentage of Purshia tridentata seedlings surviving the first year of growth (r = 0.68, D.F. = 13, P < 0.01 ). Data from the reciprocal planting experiment using three locations in British Columbia, BO, OS and WT. Discussion Purshia seedlings emerged over a variety of site conditions with patterns generated more by date of planting and edaphic factors than by differences between sites of growth or populations. However, differences within populations were highly significant suggesting that there is considerable variation between shrubs in the quality of seeds that they produced or their germination requirements. 54 Edaphic factors Seedling emergence was inversely related to the silt content of the soil. This is in contrast to the results of Mustart and Cowling (1993) investigating factors limiting germination of seeds from two Protea and two Leucadendron species in an arid region. They observed greater seedling emergence on fine textured soils (high silt and clay content) which maintained higher levels of soil moisture. Likewise, Keddy and Constable (1986) observed finer grained soils maintained higher soil moisture and provided a favorable seed bed for germination. However, both of these studies observed the emergence of seeds sown at the soil surface. For Purshia seeds sown at a depth of 2.25 cm soil resistance may have had a significant effect on emergence as has been shown for other species (Triticum, Sorghum and Glycine, Hanks and Thorp 1957; Zea, Weaich et al, 1992). As this represents a serious agronomic problem in arid regions around the world, soil capping or hardening has received considerable attention from soil scientists (Bengough and Mullins, 1991). Where the organic content of the soil is low, soil aggregates or crumbs are poorly developed (Brady, 1990). In this case, loamy soils are susceptible to crusting at the surface where rapid drying increases soil strength (Brady, 1990; Weaich et al, 1992). In the Okanagan the hardening of loamy soils was readily observed in July when moisture levels in the upper 20 cm of soil were extremely low. At this time loamy soils were virtually impenetrable to the soil sampling device while sandy soils remained loose and accessible. This led to the conclusion that Purshia seedlings growing in loamy soils were more restricted than those growing in sandy soils. The positive correlation between Purshia seedling emergence and percent moss 55 cover suggests that moss cover has some influence on soil moisture, soil temperature or both. In a polar desert a cryptogamic crust, the non-vascular plant community at the soil surface, was found to increase seed germination and seedling survival (Gold and Bliss, 1995). These authors reported that soil moisture levels were consistently higher under crusted compared to uncrusted soils. In arid woodlands in Arizona infiltration rates were enhanced by moss cover compared to areas where moss cover had been removed (Brotherson and Rushforth, 1983). Research in the intermountain region provide conflicting reports on the impact of moss crusts on seedling emergence and survival (see Harper and Marble, 1988 for a review). Seeds planted through the crust appear to benefit (Harper and Marble, 1988) while the crust itself may act as a barrier to those seeds falling onto it (Harris et al, 1987). In my research, no clear relationship was observed between moss cover and soil moisture as the association was complicated by varying soil textures within sites. However, current research in the Okanagan indicates that cryptogamic crusts positively affect soil moisture (L. Atwood, pers. comm.). This and the observations of the researchers described above indicate that moss may affect seeds and seedlings either directly or indirectly. Firstly, by conserving soil moisture moss cover could have a direct effect on the water level in the microsite surrounding the seed. Secondly, by slowing the drying rate moss cover could thereby preventing soil capping or hardening (Weaich et al, 1992). Soil capping will also be limited where soil aggregates are sufficiently developed (Brady 1990). Algae and moss have been shown to improve soil aggregation (Bailey et al, 1973; Schulten, 1985) which again would reduce capping. Through these various influences on soil structure and moisture, moss and other components of the cryptogamic crust could ameliorate the harsh conditions found in this arid environment. 56 Site of Growth The three sites used in this study showed no significant difference in seedling emergence for seeds planted at a common depth. In other reciprocal planting studies significant site effects were found where environmental characters, such as soil moisture, light levels and litter cover were consistently different between sites. Van Groenendael (1985) observed differences in seed germination in Plantago between a wet meadow and a dry grassland as the wetter site promoted earlier germination. Winn (1985) characterized sites by tree cover, vegetation and litter cover and showed marked differences between sites in seedling emergence for Prunella. In both these cases environmental factors were consistent enough to produce differential selection pressure. By reciprocally transplanting individuals between the two locations these researchers were able to show how the populations responded to the environmental variation. In one instance differential selection led to significant genotypic variation between populations (van Groenendael, 1985) while in the other seed attributes were shown to be phenotypically plastic, meaning that the populations were able to adapt their growth form to suit the environment in which they were growing (Winn, 1985). Apparently, environmental features at the three sites used in the current study did not differ in any consistent manner. This along with the proximity of their locations would suggest that selection pressures operating at each site are similar. The large within site variation in seedling emergence indicates the patchy distribution of edaphic factors affecting germination and seedling growth. Bell et al, (1991) measured spatial variation in edaphic factors (soil nutrients) that influenced seedling establishment of Impatiens pallida at a scale of <10 m and as little as 1 m. Soil texture throughout the Okanagan valley varies in a complex manner due to the numerous factors influencing soil formation, mainly, glaciofluvial, glaciolacustrine and eolian deposition and the variety of parent materials that there 57 (Wittneben 1986; R. Maxwell, pers com.). These processes have left a mosaic of soil types which were reflected in the pattern of seedling emergence. At the 3 sites used in this study spatial variation in soil texture appears to be important at a scale of 20-50 meters, the distance between replicate blocks, although in may be much less. In general the sites responded in a similar fashion to the addition of seeds, that is the number of seedlings increased significantly when seeds were planted into the soil. This would infer that establishment is limited at other stages of the life cycle, i.e., seed dispersal or seedling survival (Keddy, 1981). Populations Although environmental factors affected seedling emergence the direction and strength of their influence was similar on all populations, that is, a given environment had no more effect on one population than on another. This was indicated by the lack of an interaction between site of growth and population and supports the conclusion that populations showed no differences in traits related to germination or early growth (Schemske, 1984). Few studies have identified among population differences for traits related to germination and seedling emergence except under artificial conditions (Schaal, 1984; McArthur et al 1987). Field studies have been able to measure differences in germination response between populations as well, but only where strong selection pressures have lead to population differentiation (van Groenendael, 1985; Schmidt and Levin, 1985). Meyer and Monsen (1989) found no difference in the germination requirements of Purshia seeds from a wider geographic range than used in the current study. Such wide spread similarities among populations indicate that they may be responding to similar selection pressures, i.e., timing seedling emergence to correspond to periods of high soil moisture. This is consistent with the idea that seasonal variation in water availability is the driving pressure in 58 arid and semi-arid environments (Leishman and Westoby, 1994). In other words, the climate restricts seedling establishment of Purshia to the favorable growing conditions that occur in the spring. Unless other selective pressures override this predominant factor, germination and seedling emergence will be similar for all populations. Maternal families The bulk of the variation explained by seed source was derived from differences between individual shrubs. This indicated that shrubs differed significantly in the quality of seeds that they produced. These qualitative differences could reflect genetic variation between individuals but may also reflect differences in other attributes such as seed weight (Schaal, 1984) or nutrient content (Parrish and Bazzaz, 1985). We must consider that these qualitative differences need not imply better or worse in terms of fitness, only that under the conditions imposed by the experiment some maternal families produced more seedlings than others. This reflects their response to numerous factors including being collected, threshed and exposed to unnatural conditions during storage, and so on. Once seeds germinate they 'bet their life' that environmental conditions are suitable for seedling establishment (Angevine and Chabot, 1979). However, if all the seeds from a single species, population or cohort were to react in a uniform fashion to the same environmental cues, a single devastating event (i.e. drought, frost, fire) could eliminate the entire seedling population. By varying the level of dormancy and the conditions required to break dormancy between individual seeds or maternal families, individuals and populations effectively spread there reproductive effort in space and time. Through artificial selection plant breeders, intentionally or otherwise, eliminate this type of variation to achieve complete and uniform seedling emergence (Gutterman, 1992). For wild plant populations variation in the germination 59 response appears to be an adaptation to the highly variable and intense selective pressure operating at the seed and seedling stage (Angevine and Chabot, 1979). It is not clear how nitrogen content would negatively affect seedling emergence. Parrish and Bazzaz (1985) found that fitness increased with the nitrogen content of the seed families. Purshia seeds are very rich in nitrogenous compounds as indicated by the crude protein content of the embryo, roughly 40% (Jenkins, 1988). These compounds have three major roles in seeds; i) as proteins they are hydrolyzed and used as building blocks by the rapidly growing embryo, ii) in enzymes that control seed dormancy and iii) as secondary compounds to protect seeds from predators or pathogenic attack (Bewley and Black, 1985). In the first case, the growth response to nitrogen would suggest that seedling emergence should be positively associated with nitrogen content (Parrish and Bazzaz. 1985). If, however, the higher nitrogen content of seeds is related to compounds, i.e. enzymes, associated with seed dormancy or defense compounds (Bewley and Black, 1985) seeds with greater N content may have poorer germination. The correlative nature of this relationship makes any causal explanation highly speculative at this stage and further research is required in order to determine the nature of the relationship. Planting date The differences in emergence between seeds planted in February vs. March suggests that seeds used for restoration purposes should be planted as early as possible. Young et al (1993) placed Purshia seeds in the field in October and seeds germinated as early as January, but was highest in March and April. They observed as much as 75 % germination for seeds buried at a 60 depth of 2.5 cm, which is three times as much as the highest rate of emergence in the present study. Of those seeds that did not germinate in 1995, some survived in the soil until the following spring. This was indicated by the small percentage of seedlings that emerged in 1996, 14 months after planting. Purshia seeds rarely survive in the seed bank, either germinating in their first year, becoming rodent food or decomposing (Crist and Friese, 1993). The survival of seeds over this extended period, although extremely artificial due to the nature of the experiment, does indicate that solitary seeds buried beneath mineral soil are safe from seed predators for considerable lengths of time. Likewise, Evans et al (1983) found that single Purshia seeds buried 1 cm were never found by rodents while clumps of 2 or more seeds experienced 75-100% mortality. Survival Seedlings are highly vulnerable during the early stages of emergence and growth (Fenner, 1987). Evidence of this came from the large percentage of Purshia seedlings that were dead within 6 weeks of emergence. Factors contributing to seedling mortality during the early stages included grazing by rodents, trampling and desiccation. The influence of each of these factors changed through time. Newly emerged Purshia seedlings represent an important source of mineral nutrients to rodents which eat cotyledons (Clements and Young, 1996). This type of grazing was observed, especially at site OS where the density of rodents was high (Chapter 2). Once seedlings have entered the true leaf stage this was no longer a factor. Purshia from some regions are susceptible to grazing by grasshoppers (Edgerton et al, 1983). Although there was some evidence of grazing by invertebrates it did not appear to have a large impact on seedling survival. 61 Desiccation was the most important factor influencing seedling survival. This was specified as the primary cause of death for numerous experiments in arid environments (Sharitz and McCormick, 1975; Mack, 1976, Sacchi and Price, 1992). Although seedling emergence was negatively associated with the silt content of soil, seedling survival increased with silt content. This suggests that seedlings were provided a more reliable supply of water in loamy soils compared to sandy soils (Salter and Williams, 1965). The water holding capacity of sandy soils is extremely low, especially in the absence of organic material (Brady, 1990). Loamy soils have a high water holding capacity yet at a tension that makes it readily available to plants (Salter and Williams, 1965). This could explain the positive association between survival and soil texture. The net result was that there was very limited seedling establishment on sandy soils compared to loamy soils. This was particularly surprising considering that the adult plants are frequently associated with coarse textured and rocky soils (Nord, 1965). Although trampling by horses and deer affected only a small number of individuals it is important to note that this occurred in late summer. By this time seedlings had already survived the more vulnerable phase immediately following emergence. Thus, despite the low number of seedlings killed by trampling the role of this factor is evidently important. Other reports on the impacts of large mammals on Purshia seedling establishment suggest animals should be removed for up to four years following direct seeding (Monsen and Shaw, 1983). Management implications When restoring or enhancing shrub populations by direct seeding land managers should consider; i) the variability of seed quality, ii) the variability of the soil at a given site and iii) the time of planting. Seed collection should focus on collecting a moderate amount from a number of individual shrubs rather than on large quantities from one or a few sources (Primack, 1993). 62 When planting the goal should be to ensure that seeds experience conditions capable of breaking the dormancy mechanism and timing seedling emergence with optimal conditions for growth and survival. Soil texture, particularly the potential for seedling emergence to be hindered by soil capping, should be taken into consideration. The importance of this factor will depend on the size and topography of a given landscape unit and the variability of the soil texture (Main, 1987). By adjusting the depth of seed planting and by utilizing mulches to protect the soil from excessive drying soil capping may be minimized. Management plans should attempt to keep soil disturbances to a minimum as they tend to remove the soil structure that has built up over millennia (Brady, 1990) and the cryptogamic crust that has built up over decades or even centuries (Harper and Marble, 1988). Exposing mineral soil to direct exposure of rainfall may also promote erosion and problems related to soil crusting (Al-Durrah and Bradford, 1982). If time of seeding does not allow for natural conditions to break dormancy (6 weeks of cool moist conditions) artificial means of breaking dormancy may be required. These include various chemical and physical treatments that reduce dormancy and making the time of planting more flexible (Meyer and Monsen, 1989). Planting seeds late in the fall may be most appropriate as seeds will have a high chance of escaping predators but sufficient time to mature and break seed dormancy, coincident with natural patterns of seed and seedling development. The high mortality of Purshia seedlings is indicative of the difficulties involved in attempting to enhance plant populations in arid habitats by direct seeding. As irrigation is rarely an option it may be necessary to plan for more than one year of planting to achieve sufficient seedling establishment (O'Keefe, 1996). It is also recommended that domestic animals be kept to a minimum over the first several years to reduce seedling losses due to trampling and grazing. 63 Conclusions In the opening chapter I presented several hypotheses to explain the significant variation in juvenile recruitment between sites. Briefly, these were 1) differences in the quality of seeds produced, 2) differences in edaphic factors influencing seedling establishment and 3) differences in the quantity and quality of seed dispersal afforded by animals. Through the course of my research I was able to show that each of these factors had a significant impact on seedling establishment at study sites in the Okanagan valley. However, none of the hypotheses adequately explained the variation in recruitment rates observed in 1994. I present here a brief summary of the evidence and the conclusions to be drawn from these. The reciprocal planting experiment provided the most conclusive information. Comparing the performance of seeds from five shrubs from each site, there was greater within site variation in seed quality than between site variation. Therefore, there is no reason to accept hypothesis 1 that sites differed in the quality of seeds produced. This conclusion is however, based on a rather limited sample size (five shrubs per site). More conclusive evidence would require a larger sample size (see Krannitz, in press a, b). Edaphic factors, such as soil texture and soil moisture, were highly correlated with seedling emergence and survival, respectively. However, significant variation was observed at the microscale (between replicates) while there was no significant variation at the macroscale (between sites). This indicated that sites were heterogeneous for these factors, and therefore, there was no evidence to support hypothesis 2. The role of dispersal by rodents remains particularly difficult to interpret with regards to the observed variation in the density of juvenile Purshia, and hypothesis 3. I was not able to detect any obvious relationship between rodent density and seedling emergence. This may be 64 because of the limited number of locations at which the study was conducted, or, as I have argued in chapter 2, because of the stochastic factors impacting seed survival and seedling emergence. This suggests that seedling emergence cannot be predicted by measuring differences in the quantity of seeds collected by rodents. Measuring the quality of dispersal using more intensive methods (Vander Wall, 1994; Longland and Clements, 1995) will be required to elicit a more definitive answer regarding this hypothesis. As well, rodent populations are known to fluctuate temporally as much as 10 fold (Krebs, 1985). It is highly possible that the rodent assemblage currently impacting seed populations at the five locations I observed have changed significantly over the past decade. This emphasizes the dangers of relying on present day trends to explain patterns create by processes and events that occurred in the past. 65 Literature Cited Abbott, H.G. and T.F., Quink. 1970. Ecology of eastern white pine seed caches made by small forest mammals. Ecology 51:271-278. Agresti, A. 1984. Analysis of Ordinal Categorical Data. John Wiley and Sons, Inc., New York. Al-Durrah, M.M. and J.M. Bradford. 1982. The mechanism of raindrop splash on soil surfaces. Soil Science Soc. Am. J. 46:1086-1090. Angevine, M.W. and Chabot, B.F. 1979. Seed germination syndromes in higher plants. In topics in plant population biology. O.T. Solbrig, O.T., S. Jain, G.B. Johnson and P.H. Raven (eds.) Columbia University Press. New York. 188-206. Antonovics, J. and S. Via. 1988. Genetic influences on the distribution and abundance of plants. In Plant Population Ecology. A.J. Davy, M.J. Hutchings and A.R. Watkinson (eds.), Blackwell Scientific Publications, Oxford. Auclair, A.N. and Cottam, G. 1971. Dynamics of black cherry (Prunus serotina Ehrh.) in southern Wisconsin oak forests. Ecological Monographs 41:153-177.z Bailey, D., A.P. Mazurak and J.R. Rosowski. 1973. Aggregations of soil particles by algae. J. Phycol. 9:99-101. Baker, H.G. 1972. Seed weight in relation to environmental conditions in California. Ecology 53:997-1010. Barry, W.J. 1976 Environmental effects on food hoarding in deer mice (Peromyscus). J. Mammal. 57:731-746. Bell, G., M.J. Lechowicz and D.J. Schoen. 1991. The ecology and genetics of fitness in forest plant. Ill Environmental variance in natural populations of Impatiens pallida. Journal of Ecology 79:697-713. Bengough, A.G. and C.E. Mullins 1991. Penetometer resistance, root penetration resistance and root elongation rate in two sandy loam soils. Plant and Soil 131: 59-66. Bewley, J.D. and M . Black. 1985. Seeds: Physiology of development and germination. Plenum Press. New York. Bock, C.E. , J.H. Bock, W.R. Kenney and V.M. Hawthorne. 1984. Responses of birds rodents and vegetation to livestock exclosure in a semidesert grassland. JRM 37:239-242. Bradshaw, A.D. 1984. Ecological significance of genetic variation between populations. In Perspectives on Plant Population Ecology, Dirzo, R. and J. Sarukhan (eds.), Sinauer Assoc. Sunderland MA. Brady, N.C. 1990 The nature and properties of soils. 10th ed. MacMillan, New York. Brochert, M.J. and S.K. Jain. 1978. The effect of rodent seed predation on four species of California annual grasses. Oecologia. 33:101-113. Brotherson, J.D. and S.R. Rushforth. 1983. Influence of cryptogamic crusts on moisture relationships of soils in Navajo National Monument, Arizona. Great Basin Naturalist 43: 73-78. Brown, J.H., O.J. Reichman and D.W. Davidson. 1979. Granivory in desert ecosystems. Ann. Rev. Ecol. Syst. 10:201-227. Brown, J.H. and E.J. Heske. 1990. Control of a desert-grassland transition by a keystone rodent guild. Science 250: 1705-1707. Brown, J.S., B.P. Kotler, R.J. Smith and W.O. Wirtz II. 1988. The effects of owl predation on the foraging behavior of heteromyid rodents. Oecologia 76:408-415. Burrell, G.C. 1982. Winter Diets of Mule Deer in relation to bitterbrush. Journal of Range Management 35: 508-510. 66 Cannings, R.A., R.J. Cannings and S.G. Cannings. 1987. Birds of the Okanagan Valley, British Columbia. Royal B.C. Museum, Victoria. Chambers, J.C., J.A. MacMahon. 1994. A day in the life of a seed: Movements and fates of seeds and their implications for natural and managed systems. Ann. Rev. Ecol. Syst. 25:263-292. Clausen, J . , D.D. Keck and W.M. Hiesey. 1940. Experimental studies on the nature of species I. Effect of varied environments on western North American plants. Carnegie Insitute Wash. Publ. 520. Washington, D.C. Clements, C D . and J.A. Young. 1996. Influence of rodent predation on antelope bitterbrush seedlings. Journal of Range Management 49: 31-34. Crist, C.F. and T.O. Friese. 1993. The impact of fungi on soil seeds: implications for plants and granivores in a semiarid shrub-steppe. Ecology 74:2231-2239. Dalton, D.A. and D.B. Zobel. 1977. Ecological aspects of nitrogen fixation by Purshia tridentata. Plant and Soil 57-80. Edgerton, P.J., J.M. Geist and W.G. Williams. 1983. Survival and growth of apache-plume, stansbury cliffrose and selected sources of antelope bitterbrush in northeast Oregon. In USDA, Forest Service, General Technical Reports INT 152: 195-202. Evans, R.A., J.A. Young. 1977. Bitterbrush germination at constant and alternating temperatures. Journal of Range Management 30: 30-32/ Evans, R.A., J.A. Young, G.J. Cluff and J.K. McAdoo. 1983. Dynamics of antelope bitterbrush seed caches. In USDA, Forest Service, General Technical Reports INT 152: 195-202. Everett, R.L., R.0 Meeuwig and R. Stevens. 1978. Deer mouse preference for seed of commonly planted species, indigenous weed seed and sacrifice foods. Journal of Range Management 31:70 - 73. Facelli, J.M. and S.T.A. Pickett. 1991. Plant litter: its dynamics and effects on plant community structure. Bot. Rev. 57:1-32. Fenner, M . 1987 Seedlings. New Phytologist 106: 35-47. Ferguson, R.B. 1972. Bitterbrush seedling establishment as influenced by soil moisture and soil surface temperature. Journal of Range Management 25: 47-49. Furniss, M.M. 1982. Entomology of bitterbrush. In Proceedings Research and management of bitterbrush and cliffrose in western North America. USDA Forest Service, INT-152. Gold, W.G. and L.C. Bliss 1995. Water limitations and plant community development in a polar desert. Ecology 76: 1558-1568. Gutterman, Y. 1992. Maternal effects on seeds during development. In Seeds: the ecology and regeneration of plant communities (ed.) M . Fenner C.A.B. International, Wallingford, U.K. p 27-60. Grubb, P.J. 1977. The maintenance of species richness in plant communities: the importance of the regeneration niche. Biological Reviews 52:107-145. Hanks, R.J and F.C. Thorp, 1957. Seedling emergence of wheat, grain sorghum and soybeans as influenced by soil crust strength and moisture content. Soil Science Soc. America Vol. 21:357-359. Harper, J.L., J.T. Williams and G.R. Sagar 1965 The behaviour of seeds in soil I. Heterogeneity of soil surfaces and its role in determining the establishment of plants from seed. Journal of Ecology 53: 273-286. Harper, J.L. P.H. Lovell and K.G. Moore. 1970 The shapes and sizes of seeds. Ann. Rev. Ecol. Syst. 327-353. Harper, K.T. and J.R. Marble. 1988. A role for nonvascular plants in the management of arid and semiarid rangelands. In Vegetation science applications for rangeland analysis and management. P.T. Tueller (ed). Kluwer Academic Publ. Boston. Harper, W.L., E.C. Lea and R.E. Maxwell. 1993. Biodiversity Inventory in the South Okanagan. In Our Living Legacy: Proceedings of a symposium on biological diversity. M.A. Fenger, E.H. Miller, J.A. Johnson and E.J.R. Williams (eds.). Royal B.C. Museum, Victoria. 67 Harris, E. R.N. Mack, and M.S.B. Ku. 1987. Death of steppe cryptogams under the ash of Mount. St. Helens. Am. J. Bot. 74:1249-1253. Harvey, H.J. 1985. Population biology and the conservation of rare plants. In Studies on plant demography: A festschrift for John L. Harper, (ed) J. White. Academic Press. 111-125. Herrara, C M . 1984. Seed dispersal and fitness determinants in wild rose: Combined effects of hawthorn, birds, mice and browsing ungulates. Oecologia 63:386-393. Hlady, D.A. 1993. South Okanagan Conservation Strategy. In Our Living Legacy: Proceedings of a symposium on biological diversity. M.A. Fenger, E.H. Miller, J.A. Johnson and E.J.R. Williams (eds.). Royal B.C. Museum, Victoria. Hobbs, R.J. 1987. Disturbance regimes in remnants of vegetation. In Nature Conservation: the role of remnants of native vegetation. D.A. Saunders, G.W. Arnold, A.A. Burbridge and A.J.M. Hopkins (eds). Surrey Beatty and Sons. Australia. Hormay, A.L. 1943. Bitterbrush in California. USDA, Forest Service. Forest Research Note 34:1-13. Howe, H.F. and J. Smallwood. 1982. Ecology of seed dispersal. Ann. Rev. Ecol. Syst. 1982. 13:201-28. Hulme, P.E. 1994. Post-dispersal seed predation in grasslands: its magnitude and sources of variation. Journal of Ecology 82:645-652. Hutchins, S.E. S.A. Hutchins and B.W. Liu. 1996. The role of birds and mammals in Korean Pine (Pinus koraiensis) regeneration dynamics. Oecologia 107:120-130. Janzen, D.H. 1971. Seed Predation. Ann. Rev. Ecol. Syst. 2 :465-492. Jenkins, S.H. 1988. Comments on relationships between native seed preferences of shrub-steppe granivores and seed nutritional characteristics. Oecologia 75:481-482. Jenkins, S.H. and R. Ascanio. 1993. A potential nutritional basis for resource partitioning by desert rodents. Amer. Midi. Nat. 130: 164- 172. Keddy, P.A. 1981 Experimental demography of a sand-dune annual, Cakile edentula, growing along an environmental gradient in Nova Scotia. Journal of Ecology 9: 615-630 Keddy, P.A. and P. Constabel 1986. Germination of ten shoreline plants in relation to seed size, soil particle size and water level: an experimental study. Journal of Ecology 74: 133-141. Kelrick, M.I., J.A. MacMahon, R.R. Parmenter and D.V. Sisson. 1986. Native seed preferences of shrub-steppe rodents, birds and ants: the relationship of seed attributes and seed use. Oecologia 68:327-337. Krannitz, P.G. in press a. Seed weight variability of antelope bitterbrush (Purshia tridentata: Rosaceae) Am. Midi. Nat. Krannitz, P.G. in press b. Variation in magnesium and nitrogen content in seeds of antelope bitterbrush (Purshia tridentata (Pursh) DC). Am. J. Bot. Krebs, C.J. 1985. The experimental analysis of distribution and abundance. 3rd edition. Harper and Row, New York, (p.295). Leishman, M.R. and M . Westoby 1994. The role of seed size in seedling establishment in dry soil conditions - experimental evidence from semi-arid species. Journal of Ecology 82:249-258. Linhart, Y.B. and M.C. Grant. 1996. Evolutionary significance of local genetic differentiation in plants. Ann. Rev. Ecol. Syst. 27:237-277. Longland, W.S. and C. Clements. 1995. Use of fluoresecnt pigments in studies of seed caching by rodents. J. Mammalogy 76: 1260-1266. Mack, R.N. 1976. Survivorship of Cerastium atrovirens at Aberffraw.' Anglesey.' Journal of Ecology 64:309-312. McAdoo, J.K., C.C. Evans, B.A. Roundy, J.A. Young and R.A. Evans. 1983. Influence of Heteromyid rodents on Oryzopsis hymenoides germination. Journal of Range Management 36: 61-64. 68 McArthur, E.D., Meyer S.E. and D.J. Weber 1987. Germination rate at low temperature: rubber rabbitbush population differences. Journal of Range Management 40 : 530-533 . McAuliffe, J.R. 1990. Paloverdes, pocket mice and bruchid beetles: interrelationships of seeds, dispersers and seed predators. The Southwestern Naturalist 35:329-337. MacMahon, J.A., R.R. Parmenter, K.A. Johnson and C M Crisafulli. 1989. Small Mammal Recolonization on the Mount St. Helens Volcano: 1980-1987. Am. Midi. Nat. 122:365-387. Main, A.R. 1987. Management of remnants of Native vegetation - A review of the problems and the development of an approach with reference to the Wheatbelt of Western Australia. In Nature Conservation: the role of remnants of native vegetation. D.A. Saunders, G.W. Arnold, A.A. Burbridge and A J . M . Hopkins (eds). Surrey Beatty and Sons. Australia. Meyer, S.E. and S.B. Monsen 1989. Seed germination biology of antelope bitterbrush Purshia tridentata. in Proceedings- symposium on shrub ecophysiology and biotechnology USDA General Technical Report INT-256. Meyer, S.E. and S.G. Kitchen 1994. Habitat correlated variation in seed germination response to chilling in Penstemon section Glabri (Scrophulariaceae). Am. Midi. Nat. 132: 349-365. Monsen, S.B. and N.L. Shaw. 1983. Seeding Antelope Bitterbrush with grasses on South-Central Idaho rangelands: a 39 year response. In Proceedings: Research and management of bitterbrush and cliffrose in western North America. USDA, Forest Service, General Technical Reports INT 152: 126-136. Morton, S.R., D.S. Hinds and R.E. MacMillen. 1980. Cheek pouch capacity of heteromyid rodents. Oecologia 46:143-146. Mustart, P.J. and R.M. Cowling. 1993. The role of regeneration stages in the distribution of edaphically restricted Fynbos Proteaceae. Ecology 74: 1490-1499. Nord, E. C. 1965. Autecology of bitterbrush in California. Ecological Monographs 35: 307-334. Noy-Meir, I. 1985. Desert ecosystems structure and function. In Hot deserts and Arid shrublands Evenari, M . and I. Noy-Meir (eds.) Elseviar, New York. O'Keefe, M.A. 1996. Rainfall and prairie establishment. Restoration and management notes 14:1. Oomes, M.J.M. and W.T. Elberse 1976. Germination of 6 grassland herbs in microsites with different water contents. Journal of Ecology 64: 745-755. Orndorf, K.A. and G.E. Lang. 1981. Leaf litter redistribution in a west Virginia hardwood forest. Journal of Ecology 69:225-235. Parrish, J.A.D. and F.A. Bazzaz. 1985. Nutrient content of Abutilon theophrasti seeds and the competitive ability of the resulting plant. Oecologia 65:247-251. Price, M.V., N.M. Waser and T.A. Bass. 1984. Effects of moonlight on microhabitat use by desert rodents. J. Mammalogy 65:353-356. Price, M.V. and N.M. Waser. 1985. Microhabitat use by heteromyid rodents: effects of artificial seed patches. Ecology 66:211-219. Price, M.V. and S.H. Jenkins. 1986. Rodents as seed consumers and dispersers. In Seed Dispersal, D.R. Murray (ed.) Academic Press, Toronto, p 191-235. Primack, R.B. 1993 Essentials of Conservation Biology. Sinauer Assoc. Inc. Sunderland, Mass. Pulliam, H.R. and M.R. Brand. 1975. The production and utilization of seeds in plains grasslands of southeastern Arizona. Ecology 56: 1158-1166. Reichman, O.J. 1979. Desert granivore foraging and its impact on seed densities and distributions. Ecology 60:1085-92. Reynolds, H.G. 1958. The ecology of the merriam kangaroo rat (Dipodomys merriami Mearns) on the grazing lands of southern Arizona. Ecological Monographs 28: 111-127. Roach, D.A. and Wulff, R.D. 1987. Maternal effects in plants. Ann. Rev. Ecol. Syst. 18: 209-235. 69 Roundy, B.A. and C.A. Call. 1988. Revegetation of arid and semiarid rangelands. In Vegetation science applications for rangeland analysis and management. P.T. Tueller (ed). Kluwer Academic Publ. Boston, pp. 607-635. Sacchi, C.F. and P.W. Price 1992 The relative roles of abiotic factors in seedling demography of Arroyo Willow Salix lasiolepis. Am. J. Bot. 79: 395-405. Salter, P.J. and J.B. Williams 1965. The influence of texture on the moisture characteristics of soils II: available-water capacity and moisture release characteristics. Journal of Soil Science 16: 310-317. SAS, 1990. Volume 1 User's Guide Version 6, Fourth edition. Cary, N.C. Schaal, B.A. Life-history variation, natural selection and maternal effects in plant populations. In Perspectives on Plant Population Ecology, Dirzo, R. and J. Sarukhan (eds.), Sinauer Assoc. Sunderland MA. 188-206. Schemske, D.W. 1984. Population structure and local selection in Impatiens pallida (Balsaminaceae), a selfing annual Evolution 38: 817-832. Schmidt, K.P. and D.A. Levin. 1985. The comparative demography of reciprocally sown populations of Phlox drummondii Hook. I. Survivorships, Fecundities and Finite Rates of Increase. Evolution 39: 396-404. Schopmeyer, C S . [ed.] 1974. Purshia DC. in Seeds of woody plants in the United States. USDA Forest Service Handbook 450. Washington. p686-688. Schulten, J.A. 1985. Soil aggregation by cryptogams of a sand prairie. Am. J. Bot. 72:1657-1661. Schupp, E.W. 1993. Quantity, quality and the effectiveness of seed dispersal by animals. Vegatatio 107/108: 15-29. Sharitz, R.R. and J.F. McCormick. 1973. Population dynamics of two competing annual plant species. Ecology 54:723-740. Shaw, N . and S.B. Monsen. 1986. 'Lassen' antelope bitterbrush: a browse plant for game and livestock ranges. Rangelands 8:122-124. Smith, C.C. 1980. The coevolution of plants and seed predators. In Coevolution of animals and plants, revised edition. L.E. Gilbert and P.H. Raven (eds.) p53-77. Austin, Texas. Sokal, R.R. and F.J. Rohlf 1995 Biometry: The principles and practice of statistics in biological research 3rd edition. W.H. Freeman and Co. New York, page 213. Systat for Windows Version 5 1992. Evanston, Illinois. Urness, P.J. 1989. Shrubs as habitat for wildlife. In 'The biology and utilization of shrubs'. Academic Press. p441-458. Vander Wall, S.B. 1990. Food Hoarding in Animals. University of Chicago Press. Vander Wall, S.B. 1993. Seed water content and the vulnerability of buried seeds to foraging rodents. Am. Midi. Nat. 129:272-281. Vander Wall, S.B. 1994a. Removal of wind-dispersed pine seeds by ground foraging vertebrates. Oikos 69:125-132. Vander Wall, S.B. 1994b. Seed fate pathways of antelope bitterbrush: dispersal by seed-caching yellow pine chipmunks. Ecology 75: 1911-1926 Vander Wall, S.B. 1995a. The effects of seed value on caching behaviour in yellow pine chipmunks. Oikos 74:533-537. Vander Wall, S.B. 1995b. Salivary water loss during seed husking in deer mice and Great Basin pocket mice. Physiological Zoology 68: 878-886. van Groenendael, J.M. 1985. Differences in the life histories of two ecotypes of Plantago lanceolata L. In Studies on Plant Demography A festschrift for John L. Harper, (ed) J. White. Academic Press 51-68. 70 Weaich, K., K.L. Bristow and A. Cass 1992. Preemergent shoot growth of maize under different drying conditions. Soil Science Society of America 56: 1272-1276. West, N.E. 1968. Rodent-influenced establishment of ponderosa pine and bitterbrush seedlings in central Oregon. Ecology 49:1009-1011. Westelaken, I.L. and M.A. Maun. 1985. Spatial pattern and seed dispersal of Lithospermum caroliniense on Lake Huron sand dunes. Can. J. Bot. 63:125-132. Whelan, C.J., M.F. Willson, C A . Tuma and I.Souza-Pinto. 1991. Spatial and temporal patterns of postdispersal seed predation. Can. J. Bot. 69:428-436. White, J. 1979. The plant as a metapopulation. Ann. Rev. Ecol. Syst.10: 109-145. Willson, M.F. 1992. The ecology of seed dispersal. In seeds: the ecology of regeneration in plant communities. M . Fenner (ed). CAB International, UK. Willson, M.F. and C.J. Whelan. 1990. Variation in postdispersal survival of vertebrate-dispersed seeds: effects of density, habitat, season and species. Oikos 57: 191-198. Winn, 1985. Effects of seed size and microsite on seedling emergence of Prunella vulgaris in four habitats. Journal of Ecology 73:831-840. Wittneben, U. 1986. Soils of the Okanagan and Similkameen Valleys. Report no 52 B.C. Soil Survey, Ministry of Environment Technical report 18, Victoria, B.C. Wulff, R.D. 1986 Seed size variation in Desmodium paniculatum II. Effects on seedling growth and physiological performance. Journal of Ecology 74: 99-114. Young, J.A. 1988. Seedbed as selective factors in the species composition of rangeland communities. In Vegetation science applications for rangeland analysis and management. P.T. Tueller (ed). Kluwer Academic Publ. Boston. Young, J.A., R. Evans 1976. Stratification of bitterbrush seeds. Journal of Range Management 29: 421-425. Young, J.A., R. Evans, D. Palmquist 1990. Soil surface characteristics and emergence of big sagebrush seedlings. Journal of Range Management 43: 358-367. Young, J.A., J.R. Wight and J.E. Mowbray 1993 Field stratification of antelope bitterbrush seeds. Journal of Range Management 46:325-330. Zar, J.H. 1984. Biostatistical Analysis (2nd edition). Prentice-Hall, Inc. New Jersey. (Chap. 19) 71 Appendix I Table A. 1. Average number of rodents captured ( \u00C2\u00B1 Stan. Dev.) over two consecutive trapping nights over eight separate weeks between May and November 1995. SITE P. parvus P. maniculatus M. montanu R. megalotis All spp. KB 1.88 \u00C2\u00B1 1.36 2.50 \u00C2\u00B1 4.34 0.00 \u00C2\u00B1 0.00 1.00 \u00C2\u00B1 1.85 5.38 \u00C2\u00B1 6.46 CWS 5.50 \u00C2\u00B1 2.45 1.88 \u00C2\u00B1 3.72 0.13 \u00C2\u00B1 0.35 3.25 \u00C2\u00B1 5.20 10.75 \u00C2\u00B1 10.08 WT 26.00 \u00C2\u00B1 12.24 12.00 \u00C2\u00B1 4.21 0.13 \u00C2\u00B1 0.35 0.38 \u00C2\u00B1 0.74 38.50 \u00C2\u00B1 12.67 BO 44.75 \u00C2\u00B1 19.24 7.38 \u00C2\u00B1 6.39 2.50 \u00C2\u00B1 2.98 0.75 \u00C2\u00B1 1.39 55.38 \u00C2\u00B1 17.35 OS 19.63 \u00C2\u00B1 10.49 36.38 \u00C2\u00B1 17.22 0.38 \u00C2\u00B1 1.06 0.00 \u00C2\u00B1 0.00 56.38 \u00C2\u00B1 20.78 72 "@en . "Thesis/Dissertation"@en . "1997-05"@en . "10.14288/1.0075269"@en . "eng"@en . "Forestry"@en . "Vancouver : University of British Columbia Library"@en . "University of British Columbia"@en . "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en . "Graduate"@en . "Seed dispersal, seed attributes and edaphic factors : their role and impact on the regeneration of antelope bitterbrush (purshia tridentata, rosaceae)"@en . "Text"@en . "http://hdl.handle.net/2429/5898"@en .