UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The role of disturbance in permanent pastures Parish, Roberta 1987

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

Item Metadata

Download

Media
831-UBC_1987_A1 P37_2.pdf [ 7.62MB ]
Metadata
JSON: 831-1.0097487.json
JSON-LD: 831-1.0097487-ld.json
RDF/XML (Pretty): 831-1.0097487-rdf.xml
RDF/JSON: 831-1.0097487-rdf.json
Turtle: 831-1.0097487-turtle.txt
N-Triples: 831-1.0097487-rdf-ntriples.txt
Original Record: 831-1.0097487-source.json
Full Text
831-1.0097487-fulltext.txt
Citation
831-1.0097487.ris

Full Text

THE ROLE OF DISTURBANCE IN PERMANENT PASTURES by ROBERTA PARISH A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in • THE FACULTY OF GRADUATE STUDIES Department of Botany We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September 1987 ® Roberta Parish, 1987 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. I 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 1956 Main Mall Vancouver, Canada V6T 1Y3 Date °\ OcJr D E - 6 ( 3 / 8 1 ) ABSTRACT This thesis investigates how small disturbances influence community structure in three permanent pastures. Small disturbances play an important role in providing spatial heterogeneity that permits new recruits to enter populations in closed sward communities, thereby promoting diversity and species co-existence. The thesis has four components: the first two are based on observation and measurement of the occurrence of small disturbances, molehills and dung pats, in three pastures. Within pasture seasonal changes in disturbance regime were related to changes in patterns of species abundance by multidimensional contingency table analysis. Dactylis glomerata, Agropyron repens and Taraxacum officinale increased in highly disturbed plots, whereas Holcus lanatus and Trifolium repens decreased. Invasion of molehills and dung pats was usually by rhizomes or stolons from surrounding plants. Seedling recruitment was rare: Trifolium repens was the only species dependent on small gaps for sexual regeneration. Patterns of species replacement on and around the disturbances were non-random. The third part of the thesis investigated the effects of selective removal of Lolium perenne, Holcus lanatus and Trifolium repens from the oldest and youngest pastures. Strong responses to the removal of these species were found only in grasses in the youngest pasture. This is consistent with the hypothesis that competition decreases over time because of niche divergence, but may also reflect biological accommodation to grazing pressure. The fourth part of the thesis investigated changes in species composition in simulated swards in response to different regimes of mowing, fertilizer and small gap creation. Species composition was strongly influenced by mowing and fertilization but was unresponsive to small gap creation. ii ACKNOWLEDGEMENT Many people gave advice and support during the years of research and preparation that led to this thesis. I would like to acknowledge the advice, guidance and patience of my supervisor, Dr. Roy Turkington, and the support of my committee members, Drs. J. Maze, W. Neill and G. Bradfield. I am especially indebted to Dr. Malcolm Greig who provided both advice and patient counsel on statistical methods. John Emmanuel and Ken Pollson wrote computer programs to analyse species replacements. I have been fortunate in being associated with fellow students who were gracious with their time and counsel, in particular, Loyal MehrhofT, Rich Evans and Rob Scagel. I have been doubly fortunate in having excellent technical help, notably, Elena Klein, who worked long hours in all kinds of weather at tedious and repetitive tasks without complaint. Financial support through postgraduate fellowships from the University of British Columbia and the MacMillan Family and field support from the Natural Science and Engineering Research Council of Canada are gratefully acknowledged. I also wish to thank Mr. Bill Chard and Ms. Mary Chard for access to their pastures and for many kindnesses during my field work. Finally, I acknowledge the role of my children, Rochelle and Robert Adams, who have grown, while I attended Graduate School, into supportive, independent people and note that, without them, this thesis would have been finished years ago. iii TABLE OF CONTENTS Abstract ii Acknowledgement iii 1. Introduction 1 1.1. Study area 8 2. Patterns of disturbance and their influence on botanical composition 19 2.1. Introduction 19 2.2. Methods 22 2.2.1. Field methods 22 2.2.2. Analytical methods 25 2.3. Results 29 2.3.1. Vegetation 29 2.3.2. Disturbance 37 2.3.3. Interaction of vegetation and disturbance 44 2.4. Discussion 48 2.4.1. Vegetation 48 2.4.2. Disturbance 50 2.4.3. Interaction of vegetation and disturbance 51 3. Invasion of disturbances 55 3.1. Introduction 55 3.2. Methods 56 3.2.1. Field Methods 56 3.2.2. Analytical methods 57 3.3. Results 58 3.3.1. Species persistence: self-transition 58 3.3.2. Patterns of species replacement 60 3.3.2.1. Monthly transition probabilities 60 3.3.2.2. Transition probabilities over the survey period 62 3.3.3. Colonization of disturbances 66 3.3.3.1. Seedling recruitment 66 3.3.3.2. Colonization patterns 66 3.3.4. Changes in botanical composition 74 3.4. Discussion 83 3.4.1. Species persistence 83 3.4.2. Species replacement and invasion 84 3.4.3. Recruitment and colonization 87 4. Influence of species removal on interspecific interaction 90 4.1. Introduction 90 4.2. Methods 93 4.2.1. Field methods 93 4.2.2. Analytical methods 96 4.3. Results 97 4.3.1. 1939 pasture 97 iv 4.3.2. 1977 pasture 100 4.3.3. Expansion rate 103 4.4. Discussion 116 5. Influence of mowing, fertilization and disturbance regimes 120 5.1. Introduction '. 120 5.2. Methods 122 5.3. Results 124 5.4. Discussion 132 6. General discussion 136 7. Literature cited 146 8. Appendix 1 160 v List of Tables Table 1-1. Species composition of Highland Mix with sowing proportions used in 1977 pasture 16 Table 1-2. Species list for the three pastures in 1985 17 Table 2-1. Mean species cover in the three pastures averaged over eight surveys from July 1982 to October 1984 30 Table 2-2. Mean number of disturbances, molehills and dung per 5x5m plot in the three pastures at each survey period 39 Table 2-3. Significant interaction of species abundance with current and previous disturbance regime, molehills, and dung distributions 45 Table 3-1. Percentage of species that persist throughout the nine month study period in the sward around molehills and dung pats 59 Table 3-2a. Average monthly transition probability of a row species being replaced by a column species for eight common species on and around molehills 61 Table 3-2b. Average monthly transition probability of a row species being replaced by a column species for eight common species on and around dung pats 61 Table 3-3. Deviations from random of replacement patterns of common species, growing on and around molehills and dung pats in the three pastures 63 Table 3-4. Species showing significant differences in the monthly pattern of replacement when self-transitions are included 64 Table 3-5. Species showing significant differences in the monthly pattern of replacement when self-transitions are excluded 65 Table 3-6. Number of seedlings per square metre found on disturbances in the three pastures 67 Table 4-1. Species removed from eight plots in the 1939 and 1977 pastures ..95 Table 4-2. Percentage cover of removed species in four replicates in July 1983 95 Table 4-3. Mean difference (n = 4) in species percentage cover in the 1939 pasture from 1983 to 1984 98 Table 4-4. Mean difference (n = 4) in species percentage cover in the 1939 pasture from 1983 to 1985 99 vi Table 4-5. Mean difference (n = 3) in species percentage cover in the 1977 pasture from 1983 to 1984 101 Table 4-6. Mean difference (n = 4) in species percentage cover in the 1977 pasture from 1983 to 1985 102 Table 4-7. Mean expansion rate for the period 1983-1985 for species found initially in at least three of four replicates 115 Table 5-1. Percentage composition of the original seed mix 124 Table 5-2. Mean and standard deviation of the percentage cover of species and bare ground on South Campus treatment plots in September 1985 125 Table 5-3. Percentage variation in abundance of common species explained by different mowing, fertilization and divot removal regimes .126 vii List of Figures Figure 1-1. Monthly mean temperatures at Aldergrove, B.C. summarized for the period from January 1982 to December 1984 compared with the 21-year average of monthly mean temperatures 10 Figure 1-2. Monthly mean maximum and minimum temperatures at Aldergrove, B.C. summarized for the period from January 1982 to December 1984 compared with the 21-year averages 12 Figure 1-3. Monthly mean precipitation at Aldergrove, B.C. summarized for the period from January 1982 to December 1984 compared with the 17-year average of monthly mean precipitation 14 Figure 2-1. Location of the plots on contour maps of the three pastures 23 Figure 2-2. The. percentage cover of the 10 most common species in the three pastures over the study period from summer 1982 to fall 1984 31 Figure 2-3. Reciprocal averaging ordination of the abundance of all species found in the three pastures in eight surveys 35 Figure 2-4. Maps of the three pastures showing plots with consistently high, intermediate and low levels of total disturbance 40 Figure 2-5. Maps of the three pastures showing plots with consistently high, intermediate and low numbers of molehills 42 Figure 3-1. Average number of Trifolium repens seedlings per molehill or dung pat -germinating in April and surviving until August in each pasture for two consecutive years 68 Figure 3-2. Invasion ratio of species final percentage abundance on disturbed sites divided by initial percentage abundance around the disturbance 70 Figure 3-3. Percentage abundance of species at the final survey on molehills and dung pats, in the surrounding sward, and over the entire quadrat 75 Figure 4-1. Expansion rate from 1983 - 1985 of species remaining in each of the four replicates after the removal of Lolium perenne, Holcus lanatus and Trifolium repens 104 Figure 4-2. Expansion rate from 1983 - 1985 of species remaining after removal treatments against the percentage cover removed in each of four replicates of eight treatments (n = 32) 109 viii Figure 5-1. Percentage abundance of nine species on fertilized and unfertilized plots subject to different mowing regimes 127 Figure 5-2. Number of species per square metre on fertilized and unfertilized plots subject to different mowing regimes 130 ix 1. INTRODUCTION Plant communities are characterized by heterogeneity in both space and time. They are non-uniform systems that continually change. Changes result from the death and regeneration of individual plants and their parts. The view that vegetation is a dynamic system has had a long history. Aristotle imparted the principle of motion to all natural things but held that change was directed towards a goal. This view was maintained for nearly 2000 years. Although Galileo believed all motion to be circular, and hence, not directed towards goals, the observation of change was reiterated in the Dialogue "On earth I continually see herbs, plants, animals generating and decaying .... the appearance of the earth undergoing perpetual change". The Aristotelian view of motion towards a goal was held by one of the most eminent early ecological theorists. Clements (1916) considered the plant community developing toward a climax vegetation state, analogous to an organism undergoing change and development. This view was upheld and promoted by many ecologists but not without opposition. One of the main challenges came from H.A. Gleason (1926) who argued that the botanical composition of an area resulted from the interaction of two 'fortuitous and fluctuating factors', immigration and environment. The view of communities as structured, orderly entities, still persists even though Clements' organism viewpoint has been discredited with the shift in ecology, from 1 Introduction / 2 Aristotelian essentialism to more probabilistic thinking. Structure, i.e. order and organization, may be imposed either from within the community, by the plants themselves, or externally, through forces imposed upon the community. This distinction was originally applied by Tansley (1935) to successional communities. He distinguished autogenic succession in which "changes are brought about by the action of the plants themselves on the habitat" from allogenic succession in which "changes are brought about by external factors". Evidence of community structure, inferred from non-random patterns of species distribution in the community, has been used to challenge Gleason's hypothesis of random association of individual species. Ecologists, however, have taken an increasingly 'Gleasonian' approach as evidence accumulates that "every species of plant is a law unto itself, the distribution of which in space depends upon its pecularities of migration and environmental requirements. Its dissemules migrate everywhere, and grow wherever they find favourable conditions" (Gleason 1926). Structure is believed to result from feedback from established individuals, controlling species composition and abundance of seedlings, and from interactions between these recruits (Harper 1977). Thus, although species have distinct life histories that lead to an individualistic response, pattern in their association suggests rules of assembly rather than random events. Disturbances are a natural component of all communities. Recently, the influence of disturbance regimes on community structure and function has become the focus of much research effort. This interest arose from two impetuses. First, a body of evidence accumulated against classical Clementsian succession and showed that communities were continually reset to earlier successional stages by disturbances. Introduction / 3 This led to an hypothesis that community composition could be controlled by the rate and intensity of disturbance (cf. Connell 1978). Second, the application of Gause's hypothesis of competitive exclusion of species occupying the same niches led to an examination of niche requirements of plant species. The argument has been advanced by Harper (1969) that "the essentially similar requirements of all green plants - solar radiation, water, carbon dioxide, and a basic set of mineral nutrients - provide little opportunity for diversification in relation to food supply. This is in strong contrast with organisms at higher trophic levels where the food net is focused less conspicuously on a few environmental supply factors, and much of the diversity of fauna can be interpreted in terms of differences in food or feeding habit." Grubb (1977), who repeated and expanded this notion, contends that our lack of understanding of plant co-existence is due in part to a failure to take into account the phenomenon of regeneration in plant communities. Furthermore, disturbance could provide a source of temporally and spatially heterogeneous sites for vegetation regeneration that would permit niche separation and species co-existence. A recent definition of disturbance (Sousa 1984a) focuses on gaps providing sites for regeneration. Sousa defines disturbance as "a discrete, punctuated killing, displacement, or damaging of one or more individuals (or colonies) that directly or indirectly creates an opportunity for new individuals (or colonies) to become established". Disturbance can result from both physical processes, e.g., fires, floods, drought, storms, and from biological processes e.g., predation, grazing, diggings of mammals and insects. The impact of both physical and biological sources of disturbance appears similar, removing organisms or their parts and creating opportunities for recruitment (Sousa 1984a). Introduction / 4 Two additional definitions de-emphasize recruitment as a characteristic response to disturbance and emphasize the discrete onset of disturbance and change of resource availability. Bazzaz (1983) defined disturbance as "a sudden change in the resource base of a unit of the landscape that is expressed as a readily detectable change in population response". White and Pickett (1985) defined disturbance as "any relatively discrete event in time that disrupts ecosystem, community, or population structure and changes resources, substrate availability, or the physical environment". Harper (1977) distinguishes between disasters, and catastrophic disturbances. Disasters happen sufficiently frequently in the life span of populations that they exert a selective pressure whereas catastrophies occur so infrequently that populations would lose their "genetic memory" of it by the next time it recurred (Begon et al. 1986) and therefore exert no selective force. In this context the eruption of Mount St Helens is termed catastrophic whereas the eruptions of Kila Laua would be disasterous. Disturbance regimes are characterized by various descriptors. White and Pickett (1985) list: spatial distribution; frequencjr and return cycle; rotation period or turnover rate, which refers to the time needed to disturb a given area;-predictability; size or area of disturbance; magnitude intensity which refers to the force of the disturbance event; severity of the impact on the organism or community; and, synergism, the effect on the subsequent occurrence of other disturbances. Not all of these descriptors are applicable to ever}' type of disturbance. For example, it would be unprofitable to measure the magnitude Introduction / 5 intensity of animal mounds whereas this is an important descriptor of avalanches where the force required to snap tree stems is of interest. Natural disturbances occur in most communities (see reviews by Sousa 1984a; Pickett and White 1985) including, for example, coniferous and deciduous forests (e.g. Henry and Swan 1974; Oliver and Stephens 1977; Runkle 1981, 1982), shrubland (e.g. Christensen 1985), grassland (e.g. Loucks et al. 1985) and intertidal areas (e.g. Dayton 1971; Levin and Paine 1974; Sousa 1979). Disturbance affects various levels of organization: the impact on individual growth, architecture, and dispersal (e.g. Watt 1925; Sousa 1979; Canham and Marks 1985) may result in changes in population genetic structure (Jain 1983; Thompson 1985), age (Tande 1979), size (Veblen 1985) and in community composition, richness (Denslow 1980, 1985), dominance and structure (Brokaw 1985). While the prevalence of disturbance has been documented, there exists a lack of cogent theory concerning the processes and effects of disturbance. Two hypotheses on the effects of disturbance on communities are current: the intermediate disturbance hypothesis (Connell 1978) and the "rate of competitive displacement" hypothesis (Huston 1979). Both are interrelated and suffer from imprecision. The intermediate disturbance hypothesis argues that species richness will be higher in communities experiencing some intermediate level of disturbance frequency, intensity and size, and lower in those experiencing both much lower and much higher frequencies of disturbance. Problems arise in quantifying intermediacy and its relationship to the system under investigation. Other problems arise on Introduction / 6 subdividing disturbance into components such as magnitude, frequency and size in which intermediacy in each may have different impacts on the community (Pickett and White 1985). The underlying assumptions of the intermediate disturbance hypothesis are based on a successional sequence undergoing directional change toward a climax steady-state. Disturbance resets the successional sequence to an earlier stage, increases within community heterogeneity and, over larger geographic areas, establishes a dynamic equilibrium (Connell 1978). Huston's (1979) "rate of competitive displacement" hypothesis may be viewed as an extension of the intermediate disturbance hypothesis (Connell 1978), although it owes an academic debt to both Grime (1973, see Peet et al. 1983) and to theories of predator-mediated co-existence (Caswell 1978). In these theories, disturbance (or management, or predation and herbivory) delays the competitive exclusion of one species by another by periodic reductions in population size. The prolonging of the period to exclusion promotes species co-existence and maintains diversity. Competitive exclusion is accelerated by increased growth rate of the dominant species and diversity is increased by any condition that reduces the rate of growth, including disturbance. For example, low levels of essential nutrients would reduce growth rates and result in higher diversity than in a similar situation with higher nutrient availability (Huston 1979), though, of course, very low nutrient levels restrict the number of species able to establish leading also to low diversity (Grime 1973, 1979). Huston's model, like the predator-mediated co-existence model, focuses on the role of individual or population reduction in promoting species co-existence but fails to take account of changes in resource availability initiated by disturbance. For example, disturbance Introduction / 7 may alter nutrient availability (Bazzaz 1983; Vitousek 1985) or create sites suitable for seedling or other propagule establishment (Grubb 1977; Sousa 1979). There is a rich literature on the impact of disturbance on forest and intertidal communities ( Sousa 1984a). Studies of disturbance in grassland communities have mainly neglected pastures (but see Jones 1933a,b,c,d) and have been in rangeland (see White 1985) and prairie systems. For example, in tall grass prairie, Piatt (1975) studied the role of badger mounds on species associations. Badger mounds increased spatial heterogeneity, led to temporal partitioning in biomass production, and enhanced species co-existence by providing a site for colonization and establishment of a guild of fugitive species that were resticted to these small-scale disturbances. Unlike Piatt's study, Loucks et al.'s (1985) investigation of the response to fire, gopher mounds and small-scale erosion sites in the Wisconsin sand prairies found preferential colonization that led to changes in species abundance rather than specialist colonization of disturbances. They concluded that sand prairie grassland community composition is closely linked to the multiple-scale disturbance regimes. My thesis examines the role of disturbance in a common grassland community, the permanent pasture. Pastures are widespread throughout temperate climates and contain a mix of sown and native species. They persist in many areas only because of periodic disturbance from either mowing or grazing without which they would be rapidly invaded by shrubs and trees. In addition, pastures are subject to numerous small-scale disturbances from small mammals and insects as well as scrapings and droppings from grazers. Introduction / 8 The thesis has four components: the first two are based on observations of the occurrence of small-scale disturbances and species abundance in three pastures. In the third and fourth parts, species abundance was manipulated in field experiments. Field observations were used to study spatial and temporal heterogeneity in pastures. First, within-pasture seasonal changes in disturbance regimes were related to changes in pattern of species abundance. Second, colonization of small gaps recorded at monthly intervals was used to determine the temporal patterns of species establishment. In the third part, the effects of selective removal of dominant grasses and white clover on the abundance of the remaining species was investigated to determine the extent of the contribution of dominants to community structure and species co-existence. The final component of the study involved the setting up of an experiment to investigate species' response to the interaction of larger-scale disturbance, mowing, with small-scale gap formation under different nutrient regimes. 1.1. STUDY AREA The study area comprises three contiguous pastures located at 25704 Fraser Highway, Aldergrove, British Columbia (SW 1/4 sec. 25, Twp. 10) on a farm owned by William and Mary Chard. The farm is situated at 49° 03' 45" N. Lat. and 122° 30' 45" W Long, in the Coastal Douglas fir biogeoclimatic zone (Krajina 1965). The elevation of the pastures varies from 110m to 122m above sea level. Two of the pastures are generally flat and slightly sloping, whereas the third is gently rolling (see contour maps in Aarssen and Turkington 1985a). Introduction / 9 The pastures lie on a parent material of Pleistocene glaciomarine deposits. The soils are Luvisolic Humo-Ferric Podzols and Orthic Humic Gleysols (Canada Soil Survey Committee 1978). They are moderate- to fine-textured clay loams which are moderately to poorly drained. The average top soil depth (A + B horizons) is from 80-100cm. The pH is moderately acid, ranging from 5.2 - 5.9. Results of detailed soil analysis for the pastures can be found in Aarssen and Turkington (1985a). The area has a relatively mild, wet winter and a warm, drier summer. Environment Canada climatic data recorded at Aldergrove, 2.4km east of the farm, provided monthly mean temperatures for the 21 year period from 1960 to 1980 (Figure 1-1). During the study period (1982-1985), winters have been generally milder than average: monthly mean minimum temperatures were somewhat higher than normal whereas monthly mean maximum temperatures were near normal (Figure 1-2). December 1983, however, was a particularly cold month in which precipitation was also lower than normal. Mean annual rainfall for the period 1964-1980 was 1652.3mm (S.D. 232.8), most of which fell in the winter (Figure 1-3). May 1984 was exceptionally wet (265.7mm) but this was well below a seasonal peak of 347.4mm in November 1983. The area was first cleared for farming around 1900 and periodic clearing continued over the next forty years to encompass all of the present three pastures. These were last ploughed and seeded in 1939, 1958 and 1977 respectively, and prior to reseeding, all had been pastures. All pastures were originally sown with a seed mixture, comprising by volume, 5-10% Trifolium Introduction / 10 Figure 1-1. Monthly mean temperatures at Aldergrove, B.C. summarized for the period from January 1982 to December 1984 compared with the 21-year average of monthly mean temperatures. A 1982 • 1983 O 1984 • 21-year average MONTHLY MEAN TEMPERATURE 20-, Introduction / 12 Figure 1-2. Monthly mean maximum and minimum temperatures at Aldergrove, B.C. summarized for the period from January 1982 to December 1984 compared with the 21-year averages. A 1 9 8 2 • 1 9 8 3 O 1 9 8 4 • 21-year average MONTHLY MEAN MAXIMUM TEMPERATURE Introduction / 14 Figure 1-3. Monthly mean precipitation at Aldergrove, B.C. summarized for the period from January 1982 to December 1984 compared with the 17-year average of monthly mean precipitation. A 1982 • 1983 O 1984 • 17-year average Introduction / 16 repens, 15-20% Dactylis glomerata and 70-80% of a mixture known locally as Buckerfields Highland mix. The species proportions in the mix varied and the exact composition seeded is known only for the 1977 pasture (Table 1-1). The mix used in 1977 differed from that used in 1939 and 1958 by the addition of tetraploid L. perenne and the large 'Ladino' T. repens. Various other species not in the original sowing mix have become established in the pastures (Table 1-2). Table 1-1. Species composition of 'Highland Mix' with sowing proportions used in 1977 pasture. Species Percent composition Dactylis glomerata L. 45 Trifolium pratense L. 20 Lolium perenne L. 15 Lolium multiflorum Lam. 10 Phleum pratense L. 5 Trifolium repens L. 2 Trifolium repens var. Ladino 3 Introduction / Table 1-2. Species list for the Grasses Agropyron repens (L.) Beauv. Agrostis alba L. Alopecuris geniculatus L. Alopecuris pratensis L. Anthoxanthum odoratum L. Dactylis glomerata L. Festuca pratensis Huds. Festuca rubra L. Glyceria declinata Breb. Holcus lanatus L. Lolium multiflorum Lam. Lolium perenne L. Lolium multiflorum x Lolium perenne Phleum pratense L. Poa compressa L. Poa trivialis L. Non-grasses Achillea millefolium L. Capsella bursa-pastoris L. Carex sp. Cerastium vulgatum L. Cirsium arvense (L.) Scop. Gnaphalium uliginosum L. Hypochoeris radicata L. Juncus sp. Medicago lupulina L. Plantago lanceolata L. Plantago major L. Ranunculus acris L. Rumex acetosella L. Rumex crispus L. Rumex obtusifolius L. Stellaria media (L.) Vill. Taraxacum officinale Weber Trifolium pratense L. Trifolium repens L. three pastures in 1985. 1939 1958 1977 X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X x - denotes species present in pasture Introduction / 18 A strict management regime is not followed. The farm was managed for dairy production until 1983 and subsequently switched to beef production; however, no changes in management practices in the pastures ensued. The cattle may graze in the pastures between May and November, though frequently, grazing is delayed until after the pastures are cut for hay. In recent years, the pastures have been cut in late June or early July and then strip-grazed by 20-30 cattle once sufficient regrowth has occurred. After a pasture has been strip grazed, the cattle roam freely until the end of the grazing season, usually in November. The pastures have never been treated with chemical pesticides, herbicides or fertilizer, other than barnyard manure which is spread sporadically to maintain soil fertility. 2. PATTERNS OF DISTURBANCE AND THEIR INFLUENCE ON BOTANICAL COMPOSITION 2.1. INTRODUCTION Pastures are regarded as biotic plagio-climax vegetation (Tansley 1939), i.e. vegetation maintained by grazing and mowing. This management regime retains the pasture in a relatively stable state but there is nonetheless wide spatial and temporal variation in botanical composition of pastures (e.g., Spedding 1971; Watkin and Clements 1978; Vickery 1981; Thorhallsdottir 1984; Snaydon 1985). The management regime is also one of the most important ecological features of the pasture because grazing animals not only defoliate, but also create many small-scale disturbances within the pasture - hoof marks, urine patches and dung pats. These disturbances, and additional activities of small mammals and insects, create openings or gaps within the otherwise continuous sward. The rate and manner of colonization of these gaps may have an important impact on the species composition of the pasture, species diversity and population flux. Dung and urine have been found to affect pastures (e.g. Norman and Green 1958; During and McNaught 1961). Urine provides nutrients, mainly nitrogen and potassium, to the sward which may stimulate yield and depress the abundance of clovers (During and McNaught 1961; Mundy 1961). Norman and Green (1958) found that a single application of urine had a negligible effect on species composition, but around dung patches, Dactylis glomerata, Trifolium repens and Festuca rubra all increased in both cover and yield. MacDiarmid and Watkin 19 Patterns of disturbance and their influence on botanical composition / 20 (1971) showed that grasses (mainly Lolium spp.) increased in relative abundance around dung patches while clovers (mainly T. repens) did not. They also found that plants underneath dung decayed rapidly and regrowth was minimal although Hutchinson (1979) noted that a number of species (e.g. Poa trivialis, Agrostis stolonifera, Lolium spp.) were able to penetrate the patch and survive. T. repens was one of the first invaders of areas left bare by decaying dung (Weeda 1967). On average a cow produces about 9kg of urine and 28kg (fresh weight) of dung daily, defecating 11-15 times per day (MacLusky 1960; Marsh and Campling 1970). In the 150 day grazing season typical of the Aldergrove pastures, even if cows were penned at night and only 50% of dung were deposited while grazing, that would amount to about 2100kg of dung per animal and 63,000kg for a small dairy herd of 30 cows. The area covered by a single dung pat ranges 2 2 from 0.05m to 0.12m (Petersen et al. 1956) and the total area affected by dung is estimated to be about six times the actual area covered (Norman and Green 1958; MacLusky 1960). Therefore from 7425 to 17,820m2 (20-50% of the three pastures) could be affected by dung in a 150 day grazing season. Harper (1977) summarized the possible effects of dung in pastures: (i) smothering and exclusion of light from plants; (ii) a local change of nutrient relations; (iii) a local change in the pattern of grazing because animals tend to avoid fouled areas; and (iv) creation of an area for colonization by new or existing individuals. Another source of spatial variation in many pastures is from molehills. Molehills Patterns of disturbance and their influence on botanical composition / 21 in the Fraser Valley are produced by the coast mole (Scapinius orarius True). These moles eat arthropods, annelids and molluscans exclusively. A mole can consume more than 100 worms per day. A study on moles in the Fraser Valley found that the number ranged from 2 - 10 per hectare and was correlated to the number of earthworms (Glendenning 1957). Moles are least active in the summer months. In search of food, a single mole may create 200 to 400 hills between October and March, excavating up to 500m of tunnels usually from 15 to 20cm below the soil surface (Glendenning 1957). Tansley and Adamson (1925) noted that molehills in dune slacks had a characteristic flora differing from the surrounding vegetation. A study of buried seed in molehills in pastures (Jalloq 1975) showed a different buried seed population on molehills than in the sward. Rothwell (1977) investigated the colonization of molehills in Welsh pastures and noted that Holcus lanatus, Agrostis capillaris (syn. A. tenuis) and Anthoxanthum odoratum had significantly lower abundance on molehills than in the adjacent sward, whereas Ranunculus repens increased on molehills. In contrast, also in Welsh pastures, Davies (1966) found molehills to be colonized mainly by A. capillaris and implicated molehills in the degeneration of productive pasture to that dominated by weedy species. Molehills have effects in common with dung in that (i) plants are smothered and excluded from light and (ii) bare areas are created for plant colonization. Temporal variation in plant abundance in pastures may be related to variation in the environment such as seasonal changes, grazing management, or fertilizer application (Snaydon 1985). Temporal variation may also reflect the lifespan of Patterns of disturbance and their influence on botanical composition / 22 plants or parts of plants (Thorhallsdottir 1984) in which the genotype of the plant interacts with abiotic and biotic factors. This chapter examines the influence of two small-scale disturbances, dung and molehills, on the distribution and relative abundance of species in a series of adjacent pastures. The approach taken was to describe and correlate patterns of vegetation and patterns of disturbance within pastures. Three objectives were pursued: (i) to assess temporal variation in species abundance; (ii) to measure the intensity of disturbance caused by cow dung and molehills; and, (iii) to relate the pattern of disturbance to the pattern of distribution of species abundance. 2.2. M E T H O D S 2.2.1. Field methods To measure vegetation patterns and disturbance regimes in the three Aldergrove pastures, each pasture was subdivided into 5 x 5m plots. In order to facilitate comparison with data previously collected from these pastures (Aarssen 1983), plots were systematically located at 10m intervals along established transect lines that were approximately 10m apart. Buffer zones of 10m - 20m from the pastures' boundaries were used in all pastures to avoid edge effects and obviously swampy areas. There were 48 plots in the 1977 pasture, 50 in the 1958 pasture and 47 in the 1939 pasture (Figure 2-1). The plots were surveyed on eight occasions; each spring, summer and fall, from Patterns of disturbance and their influence on botanical composition / 23 Figure 2-1. Location of the plots on contour maps of the three pastures. Elevation in metres. 1 9 5 8 PASTURE Patterns of disturbance and their influence on botanical composition / 25 July 1982 to October 1984. At each survey, the position and diameter of all molehills and all dung pats were plotted. Vegetation was surveyed using four .5 x .5m quadrats per plot placed at 2m intervals along one of the plot diagonals. Each quadrat was subdivided into a grid of 25 points. The species rooted at each point, or bare ground, was recorded. This method was preferred over that used by Aarssen (1983) in these pastures because it is less subject to bias toward larger species and error from variability in wind speed and grass length. 2.2.2. Analytical methods Vegetation data were analysed by reciprocal averaging of the total mean rooted cover for each pasture, using an algorithm outlined by Orloci (1978). Species cover for each sampling period was compared between pastures to detect similar patterns and trends, using the Michigan Interactive Data Analysis System (MIDAS) (Fox and Guire 1976) version of profile analysis. Multidimensional contingency table analysis was used to analyse patterns of disturbance and patterns of species abundance. Contingency table analysis describes the structural relationship between variables (Fienberg 1980) and is especially useful when variables are not independent. Its application to ecological data is discussed by Fienberg (1970) who recommends its use if observations have unequal variances and if the underlying distribution is not normal. Variables are cross-classified into discrete categories referred to as cells. The Patterns of disturbance and their influence on botanical composition / 26 expected value of a cell (e„) is calculated by multipying the row total + ) by the column total and dividing by the total number of observations (N): e.. = n.,n,JN (1-1) y 1 + +J Equation (1-1) may be expressed as a log-linear model of the form loge.j=logn.+ + logn+j - logN (1-2) This model can be reformulated in terms analogous to analysis of variance (Fienberg 1970, 1980) lo g e.. = u + u 1 ( i ) + u (1-3) where: u - grand mean of logarithm of expected cell frequency Ml(i) " deviation from grand mean of expected counts in cell of level i of variable 1 uor. - deviation from grand mean of expected counts in cell of level j of variable 2 The model assumes that the variables are independent and if this is not so, an interaction effect must be incorporated such that \oge.. = u + u 1 ( i ) + u2Q) + u i m (1-4) In this form, a greater number of variables may be incorporated but the number Patterns of disturbance and their influence on botanical composition / 27 of possible models rapidly rises as the number of variables increases. For example, in the case of two variables, only two models are possible, with three variables there are 8 different models and with four variables 113 different models (Fienberg 1980). It is possible that a large number of these models will fit the data and the saturated model with all possible interaction terms always produces expected values equal to the observed values. The statistic used to test goodness of fit of the models is the likelihood ratio chi-square: This statistic may be partitioned into several additive parts and was used in this study to test for the presence of an interaction effect. To examine whether disturbance regime has any relationship to species abundance, a four-way classification was used. This consisted of one response variable, species abundance, and three explanatory variables, disturbance regime, season and year. Species abundance was categorized as high, medium or low for species with greater than 10% cover, and into presence or absence for those with less than 10% cover. Disturbance regime, of a plot was categorized as high, medium or low. Season (spring, summer or fall) and year of the survey were included to account for differences in abundance associated with the timing of vegetative growth and flowering, and the various aspects of climate and management that vary from year to year. The selected model assumed interactions existed between disturbance, year and season, and between abundance, year and season. To test for a relationship between disturbance and abundance, (observed/expected) (1-5) Patterns of disturbance and their influence on botanical composition / 28 the interaction term of abundance and disturbance was added to the model. The difference between the likelihood ratios for the two models was tested against a chi-square distribution to determine if the addition of the interaction term significantly improves the fit of the model. The more specific test for a three-way interaction between abundance, disturbance and season, or between abundance, disturbance and year was accomplished by the subtraction of the likelihood ratio of each model containing one three-way term from the model containing both terms. The two-way interaction of disturbance and abundance is a simpler but more specific case of three-way interaction (Fienberg 1980). The choice of models may be termed the maximal choice in that the model contains as many terms as possible. Effects may be missed in this approach but any significant improvement has no other explanation in terms of the full model used. One potential problem in this type of analysis is the formation of categories from discrete data. There is no best way to accomplish the groupings although the choice of cut-off points for categories does influence the size of the interaction observed (Fienberg 1970). Cut-off points to group disturbances regimes were kept constant over the three pastures. If a total of three or fewer molehills or dung pats per season were found in a plot, that plot was arbitrarily considered to have a low disturbance rate, similarly between three and fifteen was considered an intermediate rate and more than fifteen, a high rate. Cut-off points for each species were varied because of the probability that the carrying capacity of each field differs for some species. A number of cut-off points were tried for each species. Patterns of disturbance and their influence on botanical composition / 29 2.3. RESULTS 2.3.1. Vegetation The species composition of the three pastures differed in mean percent cover (Table 2-1). Percent cover of the ten most abundant species in each pasture at each of the eight surveys is shown in Figure 2-2. Profile analysis revealed no significant parallel trends in the abundance of any species in the three pastures, even though some changes in species cover appeared to be strikingly consistent; e.g. Poa compressa abundance from summer 1983 to fall 1984 in all three pastures; Lolium perenne from fall 1982 to fall 1984; and, Trifolium repens throughout the study period in the 1939 and 1977 pastures. One trend found in Aarssen's (1983) cover data, collected from 1979 to 1981, that continued through 1982 to spring 1984, was the decline in abundance of Dactylis glomerata in the 1977 pasture. There were other major shifts in species abundance; an increase in abundance of T. repens in the 1977 and 1939 pastures, and an increase in L. perenne in the 1939 pasture, but these did not follow from any pattern observed in the previous three years. The abundance of Festuca rubra in the 1958 pasture is related to topographic features. The short-lived, grazing-sensitive Trifolium pratense, present in the Aarssen (1983) study of these pastures, had declined to less than 1% of the cover and consequently has been grouped with 'other species'. The results of reciprocal averaging ordination of species abundances from all eight sampling periods combined (Figure 2-3) showed the 1939 and 1977 pastures, the Patterns of disturbance and their influence on botanical composition / 30 Table 2-1. Mean species cover in the three pastures averaged over eight surveys from July 1982 to October 1984. Species 1977 1958 1939 Lolium perenne 17.87 10.18 23.05 Holcus lanatus 13.79 13.39 12.98 Dactylis glomerata 9.51 15.11 4.98 Agropyron repens 5.76 8.06 2.76 Phleum pratense 2.18 6.11 0.91 Poa compressa 13.36 16.13 20.78 Trifolium repens 17.92 7.08 17.37 Taraxacum officinale 9.04 6.03 7.09 Ranunculus acris 1.77 1.05 2.33 Plantago lanceolata 1.08 1.60 0.21 Festuca rubra 0.02 6.58 0.28 Agrostis alba 1.02 0.55 1.34 Anthoxanthum odoratum 0.93 0.58 0.51 Cerastium vulgatum 0.03 0.13 0.08 Juncus spp. 0.06 0.20 0.00 Bryophytes 2.01 1.49 0.66 Bare ground 2.34 4.42 2.03 Dung 0.59 0.80 1.31 Other 0.69 0.46 0.97 Results / Figure 2-2. The percentage cover of the 10 most common species in the three pastures over the study period from summer 1982 to fall 1984. SPR - spring, SUM - summer • 1977 pasture • 1958 pasture • 1939 pasture as Oi 00 Ci PERCENT COVER _ ^ . ^ - • M M U) O O l O CT O O i O -1 1 1 1 I I Q O O I' C/) I/) Q PERCENT COVER _ -» _» NJ po CO O Ol O Ol O Ol o ' — I I J L C D CD 09 PERCENT COVER _ — _A ISJ ( s j C O O Ol O Ol O Ol o — I 1 1 I L__J "ft O B va r-C D C D PERCENT COVER _. M K> O cn O cn O cn i i i i I co -y -HP o o" c 8" Q —4-c i CO CO PERCENT COVER PERCENT COVER Patterns of disturbance and their influence on botanical composition / 35 Figure 2-3. Reciprocal averaging ordination of the abundance of all species found in the three pastures in eight surveys from July 1982 to October 1984. INITIAL - indicates the first survey in July 1982. AXIS 1 (49% variation) AXIS 1 (49% variation) Patterns of disturbance and their influence on botanical composition / 37 oldest and youngest pastures respectively, grouped near each other on the first axis, distant from the mid-aged 1958 pasture. The first axis represents a gradient based on the average abundance of each species and the second axis reflects the change in abundance of species located on the extremes of underlying physical gradients (Hill 1973). On the second axis, all pastures displayed approximately the same range of variation, although at slightly different positions along the axis. When points were joined sequentially to form a time-line between surveys, species composition in all pastures followed the same direction along the second axis. On the third axis, species composition appeared to oscillate in the 1977 and 1939 pastures. 2.3.2. Disturbance Molehill and dung counts in each plot were summed to form a disturbance measure. The three pastures show different disturbance regimes (Table 2-2). Means for each pasture were tested by using SNK on square root transformed data. The mean number of disturbances in the 1939 and 1958 pastures was not significantly different but dung formed the major component in the 1939 pasture and molehills in the 1958 pasture. The 1977 pasture had a significantly lower number of disturbances than the other two pastures (p<0.05) although there were a larger number of molehills than in the 1939 pasture. The 1977 and 1958 had similar average amounts of dung. The amount of dung and the number of molehills varied both seasonally and yearly in each field. To determine if the distributions of molehills, dung and disturbances were Patterns of disturbance and their influence on botanical composition / 38 random, these were tested against the Poisson distribution, or against the negative binomial distribution if the variance- mean ratio indicated contagion (i.e. > 1). The overall disturbance pattern fitted either the Poisson (11 out of 24) or the negative binomial (11 of 24) distributions and the remaining two were contagious distributions that did not fit any of the simple unimodal distributions (Table 2-2). Dung deposition usually fitted a Poisson distribution (19 out of 24), whereas molehill distribution fitted a negative binomial in 16 of 24 cases. Complex contagious bimodal distributions were not tested because of the computational difficulties surrounding them. The number of disturbances in a plot was examined to determine whether certain areas of the pastures had consistent patterns of disturbance. Categories of disturbance regimes previously defined for contingency table analysis were used to group plots. All three pastures had plots that had consistently high, intermediate and low disturbance rates (Figure 2-4). There were also a number of plots that varied greatly in the amount of disturbance from one sampling period to the next. In the 1977 and 1958 pastures, the consistently low disturbance plots were associated with a lack of molehills. The 1939 pasture also had areas of low mole activity but the higher dung levels in this pasture made no plot consistently low in disturbance. Molehill distribution has been similarly mapped (Figure 2-5). Because dung deposition was almost always random, the amount in each plot varied at each survey although there was a slight tendency for more to be located at the narrow entrance to the 1939 pasture. T a b l e 2 -2 . Mean number of d i s t u r b a n c e s , m o l e h i l l s and dung per 5x5m p l o t i n the t h r e e p a s t u r e s a t each s u r v e y p e r i o d . The da ta f i t t e d a P o i s s o n d i s t r i b u t i o n except those i n d i c a t e d by (*) f i t t e d a n e g a t i v e b i n o m i a l d i s t r i b u t i o n o r (*) f i t t e d no s i m p l e d i s t r i b u t i o n . D i s t u r b a n c e 1939 1958 1977 1939 Summer 1982 6 .00 1 . 98 7.23 1 .87* F a l l 1982 9.26 13.24* 7.75* 4 .62 * S p r i n g 1983 6 .68* 9 .80* 4 .92* 1 .02* Summer 1983 7.26 7.22 3.58 1 .83* F a l l 1983 13 .73 * 17.92 13.56 1 .98* S p r i n g 1984 18 .63 15.50* 12.19* 4 . 4 0 * Summer 1984 * 2.21 5.30* 4 .42* 2 .21* F a l l 1984 11.45* 12.02* 9.73 4 .34* Mean 9 . 4 0 a 1 0 . 3 7 a 7 . 9 2 b 2 . 7 8 a M o l e h i l l Dung 1958 1977 1939 1958 1977 * 1 .84 3 .77* 4. 13 0 . 14* 3 .46 9 .06 * 3 .96* 4.64 4. . 18 3 .79 * 6 . 22 3 .73* 5.66 3 . 58 1 . . 19 4. 18* 3 .58 5 .43* 3. .04 0 .0 11.46* 6 .33* 1 1 .75 6 , .46 7 .23 12.64* 6 .04* 14.23* 2 . 86 6. .15* 5 .30* 4 .42* 0 . 0 0. ,0 0. .0 7 .92* 4 .96* 7.11 4. . 10 4 . 77 7 . 3 3 b 4 . 6 0 ° 6 . 6 2 a 3 . , 0 5 b 3. 3 2 b Means of D i s t u r b a n c e , M o l e h i l l , o r Dung tha t do not sha re a common code ( a , b , c ) a r e s i g n i f i c a n t l y d i f f e r e n t (p<0.05) . .co Patterns of disturbance and their influence on botanical composition / Figure 2-4. Maps of the three pastures showing plots with consistently high, intermediate and low levels of total disturbance. Plots left blank varied in the amount of disturbance from one survey period to the next. high level i intermediate level low level Patterns of disturbance and their influence on botanical composition / 42 Figure 2-5. Maps of the three pastures showing plots with consistently high, intermediate and low numbers of molehills. Plots left blank varied in the number of molehills from one survey period to the next. Si high number intermediate number low number Patterns of disturbance and their influence on botanical composition / 44 2.3.3. Interaction of vegetation and disturbance The interaction of species abundance and the disturbance regime was first determined for two possible response sequences: (i) an immediate response to the current disturbance regime and (ii) a delayed response to the disturbance regime of the previous season. These two sequences were then decomposed into the components of disturbance, molehill distribution and dung distribution. Component interaction with abundance was also determined. Significant interactions between disturbance patterns and the abundance of the nine most common species are shown in Table 2-3. Each species showed an individualistic response pattern: some increased, some decreased and some interacted with season or year, increasing or decreasing over certain time periods. Lolium perenne had no change in abundance associated with the disturbance regime of the current survey. There was, however, a significant (p<0.05) increase in abundance in the next survey in highly disturbed plots in the 1939 and 1977 pastures. The response of abundance to disturbance in the 1939 pasture was dependent on year, that is, management practices and climatic regimes. Although L. perenne increased in two pastures, it should not be assumed that the species always increases; in the 1958 pasture, there was a non-significant negative interaction between abundance and disturbance (0.5>p<0.1). Holcus lanatus had approximately the same abundance in all pastures (about 15%) and the same cut-off points were used in all analyses. H. lanatus Patterns of disturbance and their influence on botanical composition / 45 Table 2-3. Significant interaction of species abundance with current and previous (1) disturbance regime, (2) molehills, and (3) dung distributions: (+) indicates increased abundance and (-) indicates decreased abundance at high disturbance rates. * - p<0.05; ** - p<0.01; *** - p<0.001; s - seasonally dependent interaction; y - yearly dependent interaction. Species Lolium perenne Pasture 1977 1958 1939 Current season disturbance (1) (2) (3) Previous season disturbance + (2) (3). Holcus lanatus 1977 1958 1939 * * * * * * * Trifolium repens 1977 1958 1939 *** * + * * * s * * * * * * * * Dactylis glomerata 1977 1958 1939 + + + s/y + + : Agropyron repens 1977 1958 1939 + + + + Taraxacum officinale 1977 1958 1939 + * s + + ' + ' Poa compressa 1977 1958 1939 + Phleum pratense 1977 1958 1939 + * + s/y + Ranunculus acns 1977 1958 1939 Patterns of disturbance and their influence on botanical composition / 46 abundance decreased in all pastures in response to disturbance in the current season. In the 1977 pasture, however, the response was dependent on season (p<0.05): a decrease in abundance occurred in the summer and fall but there was a slight increase in disturbed areas in the spring. H. lanatus appeared to decrease primarily with increasing number of molehills in the 1977 (p<0.05) and 1958 pastures (p<0.05). This may be because molehills rather than dung made the major contribution to the disturbance regime in these pastures. There was also a significant decrease in plots disturbed in the previous season in the 1977 and 1958 pastures (p<0.001 and p<0.05, respectively). In the 1958 pasture, H. lanatus showed a seasonal response to dung deposited in the previous season (p<0.05). The interaction was negative in the spring and summer but positive by the fall. Trifolium repens decreased in response to current disturbance in the two pastures in which significant interactions were found (p< 0.001 and p<0.05 in the 1977 and 1958 pastures respectively). The overall decrease is suggested to be due to molehills because T. repens increased in abundance with increasing dung deposition in the 1977 pasture (p<0.05) although the increase was restricted to summer in the 1958 pasture (p< 0.001). The overall decrease in abundance was still significant in the survey in the fallowing season (p<0.01 and p<0.05 in the 1977 and 1958 pastures respectively), and this decrease was associated with molehills. Three species, Dactylis glomerata, Agropyron repens and Taraxacum officinale showed significantly increased abundance in highly disturbed plots in all three Patterns of disturbance and their influence on botanical composition / 47 pastures. In addition, all three species increased significantly with molehills suggesting that it was that component of disturbance which most affected abundance. There were some exceptions to the overall pattern (Table 2-3). T. officinale, in particular, had a strong seasonal and yearly component to its interaction with disturbances. The response of Poa compressa to disturbance was highly variable. In the 1958 pasture, P. compressa showed a seasonal interaction with current disturbance (p<0.05), increasing in the summer and fall but decreasing in the spring. The interaction with the previous disturbances was also seasonal but only increased in the fall, that is, the decrease in spring persisted into the summer and the summer increase into the fall, but the fall increase in abundance did not carryover into spring. In the 1939 pasture, P. compressa abundance interacted with disturbance (p<0.05) and was most abundant at medium disturbance rates. There was a seasonal component to the interaction with molehills in which P. compressa only increased in the summer (p<0.01). Phlenm pratense was also inconsistent in its response to disturbance. In the 1977 pasture, P. pratense increased in highly disturbed plots (p<0.05). P. pratense showed an overall decrease with disturbance in the 1958 pasture (p<0.05) and dung in the 1939 pasture (p<0.05) and showed seasonal and yearly variations in its interaction with molehills. P. pratense was absent from molehill sites formed in the spring but present on them in the summer and fall. Because P. pratense generally had peak abundance in the spring (Figure 2-2) and declined in relative abundance over the growing season, the overall effect of disturbance was to Patterns of disturbance and their influence on botanical composition / 48 decrease abundance. P. pratense responded seasonally to the dung previously deposited (p<0.05): presence in spring was positively related to dung of the previous fall and, to lesser extent, summer abundance to spring dung but fall abundance was negatively related to dung deposited in the summer. No overall significant response to disturbance was recorded for Ranunculus acris. This species showed decreased abundance (p<0.05) at high molehill densities in the 1939 pasture. In the 1977 pasture, there was no response to current disturbance, although decreases in abundance were associated with previous disturbances. R. acris was absent from plots with many molehills but it was present in medium density dung sites. These interactions led to no overall significant response to disturbance. 2.4. DISCUSSION 2.4.1. Vegetation Species abundance in the three pastures is characterized by spatial and temporal heterogeneity. Fluctuations in abundance from year to year have been documented in other grassland communities (e.g. Rabotnov 1966, 1974; Kleter 1968; Snaydon 1985) although the chalk grasslands studied b3' Grubb et al. (1982) changed remarkably little over a 12 - 15 year period. Undoubtedly, some temporal variation is associated with seasonal growth patterns such as the spring increase in Phleum pratense abundance and the fall increase of Trifolium repens (Figure 2-2). Whether gains represent vegetative expansion or recruitment into the Patterns of disturbance and their influence on botanical composition / 49 population, or declines result from death of entire plants or from their parts, is unknown. This presents a problem, however, only for demographers interested in the fate of individuals. Despite the lack of significant parallel changes in species abundance among pastures, reciprocal averaging ordination indicated trends in botanical composition followed the same direction in all pastures. The absence of trends associated with pasture age differs from the findings of Aarssen (1983) in these pastures. He found a strong directional trend in community composition and high variability in species composition in the 1977 pasture. Since Aarssen's survey, the 1977 and 1939 pastures have become more similar in species composition. It is possible that topographical variation evident in the 1958 pasture differentiates its botanical composition from the other two relatively flat pastures. The method of data collection may have contributed to differences in results. Rooted cover is a two -dimensional method and as such is less variable than Aarssen's method which allows for a three - dimensional estimate. Information from the vertical dimension is sacrificed in rooted cover estimates for less bias and greater repeatability. Aarssen (1983) attributed differences in the variability in species composition in the 1977 compared to 1939 pasture to the time available for biotic interactions, and hence, for selective forces to accumulate and generate biological accommodation. Biological accommodation between individual plants led to increased stability in the old pasture. Results from this study, however, suggest that accommodation takes place within the first five years after seeding and establishment. This accommodation may be in response to both biotic and abiotic Patterns of disturbance and their influence on botanical composition / 50 factors and involve both genetic change (Snaydon 1978) and acclimation (Berry and Bjorkman 1980; Bradshaw 1965). Subsequently, community composition stabilizes and oscillates in response to such factors as management, grazing and fertilization (e.g. Jones 1933a,b,c,d), and climate (Garwood and Tyson 1979; Kleter 1968). 2.4.2. Disturbance Because the distribution of dung in the pastures is random, usually Fitting a Poisson distribution, all the plants have an equal chance of being affected by dung. Molehills, on the other hand, generally fitted a negative binomial distribution. This distribution can arise from a number of models and the consequences to plants growing near molehills are quite different under these models. The distribution may result from a mixture of several Poisson distributions, each having a different mean, such that the means are a 2 continuous variate with a X distribution (Bliss and Fisher 1953). The negative binomial distribution is also related to the Poisson such that, in some cases, data fit a Poisson at low population densities and a negative binomial at higher densities. This is not simply the case with the dung and molehill densities seen in Table 2-2 because some higher densities fit the Poisson and some lower densities fit the negative binomial. A negative binomial distribution may also arise from non-random contagion. If a distribution is unimodal and the "presence of one individual in a division increases the chance of other individuals falling into that division, a negative Patterns of disturbance and their influence on botanical composition / 51 binomial will fit best" (Student 1919, in Bliss and Fisher 1953). A negative binomial can result then from a mixed or compound Poisson distribution without contagion, or from contagion which changes the probability of the following events (Bliss and Fisher 1953). If the distribution results from a compound Poisson, then the chances of being killed are approximately the same throughout the pasture but if it results from contagion, plants growing near molehills are more likely to be killed than those growing further away. 2.4.3. Interaction of vegetation and disturbance Each species responded differently to the disturbance pattern. Species may be roughly divided into three groups: increasers, decreasers and indifferent, although there are exceptions within groups. The response of some species (e.g. Taraxacum officinale, Poa compressa) to disturbance was to increase but only in certain years, or seasons, or parts of the growth cycle and to decrease in others. External factors such as climate (e.g., a wet spring or summer), or management (e.g., intense grazing in summer or fall) influence species response. Patterns of response have been described but the mechanisms generating the response are still to be investigated. The pattern of response had no correlation to morphology. A robust tillering grass (Dactylis glomerata), a rhizomatous grass which usually exists as single tillers (Agropyron repens), and a clonal dicot that spreads by apomictic seed {Taraxacum officinale) were all found to increase with disturbance. Similarly, Holcus lanatus, a closely tillering, clumped grass which only occasionally spreads Patterns of disturbance and their influence on botanical composition / 52 by stolons and, Trifolium repens a clonal, stoloniferous and capable of rapid expansion and independent existence of ramets were both found to decrease with disturbance. The nature of the disturbance also varies. Dung is not only an agent of mortality but also acts to fertilize, releasing nutrients to surrounding plants. It also affords protection, because cows rarely graze within 30cm of fouled areas (Norman and Green 1958). Many seeds can survive passage through the cow and subsequently germinate (Hutchinson 1979). Dung may serve to disperse seed over long distances from plants both within and outside a pasture. Molehills have a less complex interaction: during formation plants are uprooted and smothered. In addition to freeing areas for colonization, the process of molehill building also brings up a seed source different from that of the surrounding sward (Jalloq 1975). Many of these seeds have been buried and come from a temporally distant source so that not only are different species represented but also different genotypes, containing an "evolutionary memory" (Harper 1977) of the genetic make-up of the pasture. The foregoing discussion of the results has inferred a dynamic response from static patterns, i.e. increased abundance was achieved through colonization or invasion of open space. This is not necessarily the only mechanism. For example, Taraxacum officinale can invade only by seed but its seedings rarely survive on molehills (R. Parish, unpublished data), however, its long tap root frequently survives burial so that it persists while other species are removed. Patterns of disturbance and their influence on botanical composition / 53 Static pattern need not have a dynamic corollary. There may exist a group of correlated patterns of which only two have been related. For example, Holcus lanatus proliferates in areas where water is abundant and is often found where water tables are high (Beddows 1961; Watt and Haggar 1980; R. Parish, pers. obs.). In winter, if the water table in these areas is sufficiently high to exclude moles during their very active period, October to March, high abundance associated with lack of disturbance may not represent exclusion from disturbance sites but a correlation with a third factor, high water levels. RothwelPs (1977) data, however, argues against this interpretation. He found that when H. lanatus was present in the sward it was significantly lower in abundance on adjacent molehills which suggests that it was unable to colonize these disturbances. There are a number of limitations to studies using correlation of patterns. The coincidence of patterns is not sufficent to imply causation. Patterns may be linked by other correlated factors leading to dangers in inferring mechanisms or processes based on changes in patterns. Some of these problems could have been circumvented by an experimental or manipulative approach. One approach would have been to disturb repeatedly certain areas of the pasture. Experimental work such as that of Sousa (1979) in the intertidal showed that perennial species were eliminated by high disturbance rates and ephemerals came to dominate. Another approach would be to remove moles and thereby eliminate a source of disturbance in order to compare which species dominated in their presence or absence. If it had been feasible to create artificial disturbances throughout the pastures and manipulate density, the analysis of subsequent vegetation response would have been unlinked to correlations with groups of other factors. In the Patterns of disturbance and their influence on botanical composition / 54 following chapter, I examine invasion of molehills and dung pats in an effort to resolve some of these issues. 3. INVASION OF DISTURBANCES 3.1. INTRODUCTION Pastures are subject to a variety of disturbances of various magnitudes such as drought, erosion, animal mounds, grazing, scrapings and droppings. Such disturbances create spatial heterogeneity and thus influence species and genetic diversity in the community. Disturbance is currently viewed as providing a dynamic mosaic of small to large, short-lived patches that can be utilized by opportunistic invaders (Piatt 1975; Loucks et al. 1985) or be recolonized in a more or less predicable pattern of succession (Begon et al. 1986). Disturbance may then result in an increase in species diversity by permitting the establishment of opportunistic species that cannot compete in the closed sward and increase genetic diversity by providing 'safe sites' for seed germination and establishment of both colonizing and established species. Grubb (1977) has suggested that the formation of gaps may be essential to the maintenance of species richness in communities. Plants are generally considered to be immobile organisms but the growth form of many pasture species allows movement through space both vertically and horizontally. This mobility permits the individual to sample a spatially variable environment and allows colonization of gaps within the sward. The relative contribution of new individuals compared to the expansion of old individuals will alter spatial heterogeneity and affect community composition. If species differ in their ability to colonize disturbances, the community will reflect those differences. 55 Invasion of disturbances / 56 The pasture environment provides different challenges to component species. Typically, pasture species are perennials which have the ability to persist and expand within the sward. Superimposed upon this are numerous gaps which provide an environment where the ability of a species to colonize and multiply rapidly is a more important element of success. In this study, the colonization of individual gaps created by molehills and dung pats was monitored along with associated changes in the botanical composition of the surrounding sward. The objectives were (i) to determine patterns of species invasion of disturbance; (ii) to determine patterns of species persistence and replacement; and, (iii) to compare the abilities of species to invade both open space and closed sward. 3.2. METHODS 3.2.1. Field Methods The plots described in Chapter 2 were used in this study. Ten plots were randomly chosen in each pasture for each of two survey periods (August 1983 to May 1984, and August 1984 to May 1985). Within each plot, one recently formed molehill and one recently deposited dung pat were selected. A .5 x .5m quadrat, subdivided into 100 systematically arranged points was centered on the molehill or dung pat and the plant species rooted directly under the 100 points (crosswires on the quadrat) were recorded. Depending on the availability of the pasture, surveys could begin in August to October and were repeated monthly Invasion of disturbances / 57 until the molehill or dung pat was completely covered with vegetation or until May when the grass was too long to record accurately by this method. The method required repeated measurements in accurately relocatable quadrats. Because there were cattle and mowing equipment in the pastures, permanent stakes could not be used to mark quadrats. Instead, acrylic paint was applied to the ground at the four corners of the quadrat and this allowed relocation of the quadrat, in subsequent surveys to be accurate within a few millimeters. 3.2.2. Analytical methods Transition probabilities were calculated, using the CROSSTABS procedure of SPSS.*, from the monthly record of species at each point in the quadrat. Data were pooled over the 10 molehills and 10 dung pats per pasture and analysed separately by pasture and by year. A model developed by de Jong and Greig (1984) was used to determine if species replacement over the study period deviated from random. The model describes the behaviour of First-order Markov chains where transitions in which the identity of a species at any one point remained the same (self-transitions) were excluded from the analysis. The removal of self-transitions obviated distinguishing between individuals of the same species occuping the same space at different times. To test the overall pattern of species replacement, a program was written by J. Emmanuel (University of British Columbia, pers. comm.) based on a maximium likelihood fitting algorithm developed by de Jong et al. (1983). A Invasion of disturbances / 58 second program written by K. Pollson (pers. comm.), based on algorithms developed by de Jong and Greig (1985), tested for randomness in pattern of replacement of each species pair. Rare species, less than 2% cover, were eliminated from the analysis because a minimum expected value of 3 was specified for the calculation of chi-square values. 3.3. RESULTS 3.3.1. Species persistence: self-transition Species vary in their ability to hold onto, or persist in, a given space. Persistence, defined as the percentage of a given species, present in the initial survey, that occupied the same point in the sward throughout each nine month survey period, ranged from 0 to 100% (Table 3-1). The variability of a species' ability to persist was large; for example, Ranunculus acris, encompassed the entire range within the same field and year. Species persistence within a pasture, disturbance type and year was ranked and Friedman rank correlation was calculated to determine if average ranks were the same. The pattern of species persistence differed in different years, among pastures, and whether around molehills or dung pats (p<0.001). The resultant groupings from multiple range tests of the rankings suggested that persistence was generally higher in 1984 than 1983, i.e., fewer plant parts died and were replaced by different species in 1984. When the percentage of species persisting were pooled over disturbance, pasture or year, only differences in rankings between years emerged significant (p<0.001) and ability of species to persist showed no significant Table 3-1. Percentage of species that persist throughout the nine month study period in the sward around molehills and dung pats. Included is the percentage of the molehill and dung pat that remains uncolonized. 1977 p a s t u r e 1958 p a s t u r e 1939 p a s t u r e m o l e h i l l dung m o l e h i l l dung m o l e h i l l dung S p e c i e s 1983 1984 1983 1984 1983 1984 1983 1984 1983 1984 1983 1984 Dactylis glomerata 5 1 .8 40. .3 68. .6 58, . 1 35 . 7 48 .8 70 .9 78 .0 33 , .3 76 . 7 31 .6 83.0 Holcus lanatus 42 . 5 66 . 1 44 .4 70. .9 9 .3 26 .2 32 .2 29 .8 16 .8 40 .4 5 . 1 51.3 Lol ium perenne 31 . 7 47 . 7 41 . 5 54, .8 3 .9 32 .3 33 .9 20 .0 33, .6 36 .8 25, .6 60. 5 Phleum pratense 64 .3 1 1 . 1 60. .0 50. .0 31 .3 52 . 3 72 .0 55 .6 - - 0 .0 60.0 Poa compressa 34 .0 29. .0 47 . 2 37. , 1 9 . 3 22 .0 53 .5 36 .8 27. .7 41 , .7 14, .4 53.7 Agropyron repens 2 . 7 7 .7 2 .4 6 . 7 9 .8 4 .8 41 .0 21 .6 0, .0 30 .O 4 .3 30.8 Trifolium repens 29 . 1 48 . 1 45, ,2 49, 6 21 .4 61 .9 52 .6 55 .0 40, .7 46, .6 19, .8 55.2 Taraxacum officinale 33 . 3 48. .9 45. 8 55. .6 15 .6 48, . 3 81 . 3 51 .9 37. .7 55 . 2 34, 8 54.5 Ranunculus acris 45. . 1 60, .0 0. .0 57, , 1 0 .0 30, .0 60 •0 50 .0 100 42 , .9 0. .0 71.4 Moss 15. .0 5. ,4 1 1 . 4 1 1 . 6 - 12 . .8 45 .8 15 .9 7. ,7 13. , 3 0. .0 15.4 Molehi1 1 21 .6 27 .5 4. .5 0. .0 30 .9 44 .8 15 .4 23 .6 26 . 8 27 , .2 5. .7 4.8 Dung 0. .0 0. .0 16. .4 5 . 8 0 .0 0. .0 26 . 5 22 .5 4. .0 0, ,0 1 1 . ,4 4.3 Rank 4 7 5 8. 5 2 6 1 1 10 3 8 . ,5 1 12 Friedman Test Statistic = 49.31, p<0.0001 Multiple range test of ranks 12=11=10=8.5=7 / 1 12=11 t 2 1 2 / 3 Invasion of disturbances / 60 differences between pastures nor disturbance types. 3.3.2. Patterns of species replacement 3.3.2.1. Monthly transition probabilities Tables of species transitions were generated (using the CROSSTABS program) from one monthly survey to the following one, and from the initial to the final survey (see Appendix 1). Each observed value was expressed as a percentage of the total row count so that the tables sum to 100% along the rows. Each percentage value can be considered to be the probability of the replacement of the row species by the column species. Of the 22 species found on or around disturbances, not all were sufficiently abundant to be consistently involved in replacement. The values generated were used to calculate the average monthly probability of transition for each species (see Table 3-2 for representative examples). The diagonals are the probability that the same species remains in the same location from one survey period to the next (self-transition) and the off-diagonals are the probability of its replacement by another species. The probability of persisting from one survey to the next was very high for most species, but this probability is dependent on the time interval between surveys. For example, in 1983, Dactylis glomerata, in the sward around molehills in the 1977 pasture, had a 87.9% (Table 3-2) probability of persisting from one month to the next but a 51.8% (Table 3-1) probability of persisting over the nine month survey period. This analysis also estimated monthly colonization rates on disturbances, for example, because dung in 1983 in the 1977 pasture had a Table 3-2. (a)Average monthly transition probability of a row species being replaced by a column species for eight common species on and around molehills, and the average monthly probability of molehill colonization by column species in the 1977 pasture during 1983-1984 survey period. Molehi1 1 Dactyl is Holcus Loli um Phieum Poa Agropyron Trifol ium Taraxacum glomerata 1anatus perenne pratense compressa repens repens officinale Molehi11 70. 6 1 . 5 1 . .0 6, .9 1 , 3 6. . 1 3. . 1 3. .6 2.6 Dactylis glomerata 3 . 7 87 .9 0 .8 2. .0 0. .6 2, .3 0. .4 0. .8 0.4 Holcus lanatus 3. 0 0 .5 83. . 1 2 .5 0. .3 4 .4 1. .0 3. . 2 0.3 Lolium perenne 5. 3 1 .8 1 .8 76. .9 0. .4 7. . 5 1. .4 2. .6 0.8 Phleum pratense 5. 6 0 .0 0 .0 0 .0 89. 9 0 .5 0. .0 2 , .5 0.5 Poa compressa 3 . 3 2 .2 0. 8 3. .9 0. 4 82. .4 2 . . 1 2. .0 1 .4 Agropyron repens 5. 9 3, .4 2 . .0 8. . 3 0. 0 7 , ,4 67. 2 2 . 5 0.5 Trifolium repens 6. 3 1 . . 1 1 . . 1 2 . .7 0. 8 4 . 6 0. 8 79. 1 0.9 Taraxacum officinale 3. 7 1 . .6 0. .0 2 . .5 0. 6 2. ,2 0. ,6 1 . .6 84.8 (b)Average monthly transition probability of a row species being replaced by a column species for eight common species on and around dung pats, and the average monthly probability of dung colonization by column species in the 1977 pasture during 1983-1984 survey period. Dung Dactyl is Hoicus Loli um Phieum Poa Agropyron Trifolium T araxacum glomerat a 1anatus perenne pratense compressa repens repens of f ic i naU Dung 75 .4 1 . 1 2 .0 4 .7 0. .4 3 .2 2 .4 7 . G 2.0 Dactylis glomerata 1 .9 90. .8 1 .5 1 .5 0. .0 1 . 2 0. .0 1 .5 0.4 Holcus lanatus 1 .5 0. .3 85 .3 1 . .7 0. ,5 4 . 2 0. .5 3 . 5 0.7 Lolium perenne 1 .9 1 . .0 1 . 5 82. .8 0. , 1 5. .8 0. 9 2. .9 1 .0 Phleum pratense 0 •0 0. .0 4 .9 0. .0 92. 7 0. .0 0. ,0 2. .4 0.0 Poa compressa 0 .8 1. . 1 1 .9 3. 6 0. 2 85. 6 0. .9 2. .0 0.9 Agropyron repens 2 .5 0. .6 3 .8 6. .3 0. 0 5. .6 58. 8 9 .4 2.5 Trifolium repens 1 .8 0. . 3 2 .5 2, .2 0. 1 3. .8 0. .3 86 .0 0.8 Taraxacum officinale 1 .4 0. .5 0 .0 2 . . 4 0. 0 0. .5 0. ,0 4 . 8 89.5 Invasion of disturbances / 62 75.4% probability of remaining as dung, about 24.6% of dung was colonized each month. 3.3.2.2. Transition probabilities over the survey period Values from the transition probability matrices (Appendix 1) were used to estimate the probability of replacement of one species by another over the survey period from fall to spring in a given year. Tests for randomness of the pattern of replacement showed that on both disturbance types in all three pastures, in both years, species replacement was not random (p<0.0001) and there was a strong pattern to the replacement process (Table 3-3). Test for randomness in the replacement pattern of each pair of species, e.g. the replacement of Lolium perenne by Trifolium repens, or the replacement of Holcus lanatus by Agropyron repens, showed that in almost all cases the pattern was non-random. To determine if the pattern of replacement of one species by other species at each survey was constant over time, contingency table analysis of the monthly pattern of replacement was used. Results from analyses including self-transition and excluding self-transitions are shown in Tables 3-4 and 3-5 respectively. Colonization patterns on molehills and dung pats changed over time throughout the study period. In the sward around disturbances, some species, such as Poa compressa, were replaced by different species at different times of the year whereas others, such as Dactylis glomerata, had the same pattern of replacement each month. In fact, some species, D. glomerata, Taraxacum officinale and Ranunculus acris, in particular, had such strong patterns of persistence that they Invasion of disturbances / 63 Table 3-3. Deviations from random of replacement patterns of common species, growing on and around molehills and dung pats in the three pastures. Y e a r D i s t u r b a n c e *2 S i g n i f i c a n c e 1977 Pasture 1983 1984 Molehill Dung Molehill Dung 214.5 366.3 211.5 203.8 p<0.0001 p<0.0001 p<0.0001 p<0.0001 1958 Pasture 1983 1984 Molehill Dung Molehill Dung 285.9 346.2 179.7 237.4 p<0.0001 p<0.0001 p<0.0001 p<0.0001 1939 Pasture 1983 1984 Molehill Dung Molehill Dung 207.9 390.7 271.1 471.5 p<0.0001 p<0.0001 p<0.0001 p<0.0001 Table 3-4. Species showing differences in the monthly pattern of replacement when self-trans 11ions are included: * - p<0.05; ** - p<0.01; *** - p<0.001; insufficient replacements for analysis. S p e c i e s Dactylis glomerata Holcus lanatus Lolium perenne Phleum pratense Poa compressa Agropyron repens Trifolium repens Taraxacum officinale Ranunculus acris Moss Molehi11 Dung 1977 p a s t u r e m o l e h i l l dung 1983 1984 1983 1984 1958 p a s t u r e * * * ** * * * * * * * * * * * * * * ** * * * * *** m o l e h i l l 1983 1984 *** * * * * * * * * * *** *** * * dung 1983 1984 1939 p a s t u r e molehi11 1983 1984 * * * * * * * * * * dung 1983 1984 * * * * * * *** *** *** * *** * * * * * * * * * * * * Table 3-5. Species showing significant differences in the monthly pattern of replacement when sel f-transi11ons are excluded: * - p<0.05; ** - p<0.01; *** - p<0.001; - insufficient replacements for analysis. S p e c i e s Dactylis glomerata Holcus lanatus Lol ium perenne Poa compressa Agropyron repens Trifolium repens Moss Molehi11 Dung 1977 p a s t u r e m o l e h i l l dung 1983 1984 1983 1984 * * * * * * * * * * * 1958 p a s t u r e 1939 p a s t u r e molehi11 1983 1984 * * * * *** *** * * * dung 1983 1984 molehi11 1983 1984 * * *** *** ** ** dung 1983 1984 * * *** *** cn Invasion of disturbances / 66 had an insufficient number of replacements by other species to allow analysis. 3.3.3. Colonization of disturbances In general, about 12 species invaded disturbed sites and the number of species was not significantly different between dung and molehills, pastures nor year of survey. 3.3.3.1. Seedling recruitment Very few seedlings were recorded on disturbed sites throughout the year and even fewer became established (Table 3-6). Trifolium repens seedlings were found only in late April and May on molehills and decayed dung pats. Seedling have not been observed in the sward in the Aldergrove pastures. Seedlings of T. repens were marked and their survival followed over the summer through mowing and later grazing. Survivorship from germination in spring to the following August ranged from 2% to 45%. In 1984, both a greater number of seeds germinated and seedlings survived (Figure 3-1). 3.3.3.2. Colonization patterns To compare the ability of different species to invade disturbed sites, a ratio was calculated of the final percentage on the disturbed site divided by the initial percentage abundance around it (Figure 3-2). Percentage abundance was used in this ratio to standardize differences in size between area disturbed and that Invasion of disturbances / 67 Table 3-6. Number of seedlings per square metre found on disturbances (molehills and dung pats) in the three pastures. Species Lolium perenne Poa compressa Holcus lanatus Alopecuris pratensis Agrostis alba Taraxacum officinale Ranunculus acris Rumex acetosella Plantago lanceolata Cerastium vulgatum Medicago lupulina 1939 pasture 1983 1984 3.80 3.8 1.92 1958 pasture 1983 1984 1.32 10.0 0.68 1.60 1.32 2.68 2.0 1.32 0.68 2.0 1.08 1977 pasture 1983 1984 7.92 0.92 0.92 0.98 0.48 1.16 0.56 0.56 1.16 0.56 Invasion of disturbances / 68 Figure 3-1. Average number of Trifolium repens seedlings per molehill or dung pat germinating i n A p r i l and s u r v i v i n g u n t i l A u g u s t i n each pasture for two consecutive years. i n i t i a l number per disturbance i n A p r i l ^ f i n a l number per disturbance i n A u g u s t 69 Invasion of disturbances / 70 Figure 3-2. Invasion ratio of species final percentage abundance on disturbed sites divided by initial percentage abundance around the disturbance. If the ratio is >1, species are more abundant on the disturbed site than around it; if the ratio is <1, species are less abundant than in the surrounding sward. NVA90N RATIO M A S O N RAT© NVA90N RATIO NVA90N RATIO M A S O N RATIO Invasion of disturbances / 74 remaining around the disturbance. If the ratio was >1, the species was more abundant on disturbed areas than in the surrounding sward {e.g. Agropyron repens, Poa compressa); if the ratio was < 1, the species was less abundant on disturbed sites than in the surrounding sward {e.g. Dactylis glomerata, Holcus lanatus). 3.3.4. Changes in botanical composition The pastures are characterized by changes in species abundance over time: some species decreased whereas others increased. The change in abundance of a species in this survey was governed by its behaviour in the surrounding sward plus its ability to colonize disturbed areas. Figure 3-3 shows, by pasture and disturbance type, percentage abundance of a species in the final survey on the disturbed area, in the sward around the disturbance and in both, i.e. the area covered by the sampling quadrat. The change in percentage abundance from the initial to final survey has been calculated from the change in number of rooted individuals or tillers for both the area around the disturbance and the entire area sampled (disturbed site plus surrounding sward). Obviously percentage change on disturbed areas was always 100% increase over initial abundance. For some species, such as Trifolium -repens and Agropyron repens, if there were a decline in abundance in the sward, an overall increase could still occur because of proportionally higher invasion onto the disturbed site. Invasion of disturbances / 75 Figure 3-3. Percentage abundance of species at the final survey on molehills and dung pats, in the surrounding sward, and over the entire 0.25m^ quadrat. The percentage change in composition from initial to final survey has been calculated for the surrounding sward and the entire quadrat area. The percentage change in species abundance on the disturbed area is either zero (no invasion) or 100%. disturbance surrounding sward quadrat average Invasion of disturbances / 83 3.4. DISCUSSION 3.4.1. Species persistence In general, the probability that a species would be found at the same point over the nine month study period was low (Table 3.1), suggesting that temporal heterogeneity was high. Persistence serves as an estimate of genet survivorship for species such as Ranunculus acris and of ramet survivorship for clonal species such as Taraxacum officinale which grow as independent individuals. For other species (i.e., tillering grasses, Trifolium repens and Poa compressa), the relationship between persistence and survivorship is obscured. This was not only because release of space may result from death of parts, not individuals, but also because of difficulties distinguishing between different individual plants of the same species occupying the same space in successive time periods. The greatest death risk to established plants appears to occur during the period of most active growth (Sagar 1970; Sarukhan and Harper 1973). The timing of this period varies with the species under consideration: most grasses grow faster in the spring, April to May, at a temperature optimum between 13 and 18°C, whereas Trifolium repens grows faster later in the season at a temperature optimum around 24°C (Spedding 1971). The estimates of persistence reflect the intense growth/death phase for both T. repens and grasses, because the survey usually ran from August, when T. repens was growing rapidly, to April, when the grasses were growing vigorously. Invasion of disturbances / 84 3.4.2. Species replacement and invasion Potentially, any species could be replaced by any of the species in the pasture but only the most abundant were consistently involved in the replacement processes. Replacements were not random, i.e., space capture by a species was not proportional to the abundance of that species in the pastures. The tests for non-randomness were based on a first order Markovian model. This assumes that the system has no memory, i.e., that the current state is determined only by the previous one. Over the study period, the sample size was inadequate to include sufficient multiple transitions to test for higher order effects in which the probability of a species replacement would depend on states prior to the previous one. Higher order processes would not be improbable if some species had temporary seasonal advantages and pre-empted space that was subsequently relinquished to the prior occupant. Thorhallsdottir (1984), using a Markovian model to describe species replacements in a pasture in Wales, found that between-season replacements were non-random but between-year replacements were porportional to species abundance in the pasture. She concluded that temporal heterogeneity was very fine-grained but species became thoroughly mixed over longer periods. In addition, she suggested that the majority of non-random interactions took place during periods of rapid growth and expansion. The considerable variation among the ability of species to invade disturbed areas appeared to be linked to morphological form. This finding differs from the Invasion of disturbances / 85 conclusion of the previous chapter. Pattern detected at one scale of observation 2 2 (0.25m ), was lost at another scale (25m ). In general, rhizomatous {Agropyron repens and Poa compressa) and stoloniferous {Trifolium repens) species invaded disturbed sites in proportionally higher abundance than they had in the surrounding sward. Tillering grasses {Dactylis glomerata, Holcus lanatus and Lolium perenne) were less abundant on disturbed sites than in the surrounding sward. There were exceptions to this: Phleum pratense, a tillering grass, invaded dung although it was a poor invader of molehills. The inability of Phleum pratense to invade disturbed sites in the 1939 pasture, but its ability to invade in the 1977 pasture appears to be based on a genetic difference in the P. pratense seeded in the pastures. Prior to approximately the 1960's in the Fraser Valley, an old land race of P. pratense was sown in pastures but this practice was discontinued. The more robust variety 'Climax' used in the mix after the early 1960's (Richardson Seed Co., pers. comm.) has the ability to invade disturbed areas. The apparent ability of Taraxacum officinale to invade disturbed sites is an artefact because no new individuals established, rather, mature plants emerged from under the disturbance. Emergence was often as long as three months after the disturbance had occurred. This agrees with Roth well's (1977) findings that T. officinale could survive burial. The high abundance of T. officinale associated with disturbance is due to root regeneration on sites on which species less tolerant of burial are removed. Invasion of disturbances / 86 The invasion of disturbed areas is a non-competitive process whereas performance in the sward involves both non-competitive and competitive interactions. Because of the closed nature of the sward in order for an individual to expand horizontally, another must relinquish space. Release of space may be due to density independent forces such as frost or trampling. The non-randomness of replacements, however, suggests that replacement was governed by species interactions not merely proximity. Performance on disturbed sites has been related to species mobility and investment in horizontal growth. There is no a priori reason to expect these characters to be linked to performance in the sward. But some species show similar performances in both situations, e.g., Poa compressa increased in both whereas Holcus lanatus and Lolium perenne, which usually failed to establish well on disturbed areas, also decreased in the sward. In contrast, some species had divergent responses. Phleum pratense increased in the sward but invaded dung only, suggesting that a response to nutrient enrichment facilitated lateral growth, a finding consistent with the fertilizer treatment in Chapter 5. Its lack of response on molehills may have been due to lower moisture availability (M. Pitt, pers. comm.). Dactylis glomerata also increased around dung but it was still a poor invader of disturbed sites. Agropyron repens was the only species showing an ability to invade disturbed sites but generally decreasing in the surrounding closed sward. This suggests that disturbance may provide a refuge serving as the primary agent in determining the pattern of abundance of A. repens in the pastures. 3.4.3. Recruitment and colonization Invasion of disturbances / 87 The role of dung and molehills in providing an entry into the community for new species, or even for sexually produced progeny of established species, was negligible. Colonization by seed occurred consistently only in two species, Ranunculus acris and Trifolium repens. R. acris germinated both on disturbed sites and in the sward but T. repens was restricted to disturbed areas. Similarly, Turkington et al. (1979) observed T. repens seedlings on molehills but not in the sward, and seed sown into the sward failed to establish (0.89% germination) whereas on cleared plots, seed germinated (39% germination). It appears that sexual propagation of T. repens in pastures is obligately linked to disturbances. In contrast to these findings, Piatt (1975) described a guild of fugitive species specialized to exploit badger mounds in tall-grass prairie. Other North American grassland studies have confirmed a role for disturbance in providing suitable sites for germination and survival of annuals (Allen and Knight 1984), or taxa differing from the surrounding area (Polley and Collins 1984; Loucks et al. 1985). Differences between these studies in native prairies and in the Aldergrove pastures could be used to support an argument for evolution of specialists based ort native systems being of sufficient longevity to foster such niche differentiation. Additional support for this interpretation can be derived from Belsky's (1986) observations in the old and relatively pristine Serengeti ecosystem (Whyte 1974) that colonization of disturbed sites was dominated by seedlings and not by vegetative expansion. Disturbances often present different micro-environments which provide opportunities for differential regeneration of species in the community Invasion of disturbances / 88 (Grubb 1977). Badger mounds (Piatt 1975) and buffalo wallows (Polley and Collins 1984) retain higher soil moisture than surrounding areas. Although physical characteristics of molehills were not measured in this study, it is assumed that they have a characteristic pattern of higher soil temperatures and lower moisture typical of mounds in northern latitudes (cf. Spittlehouse 1985). The presence of species that can exploit this heterogeneity probably depends on the frequency of disturbance in addition to the time dimension. Time could be either of evolutionary or ecological length. Evolutionary time would permit the development of mechanisms to exploit heterogeneity and differentiate species, and an ecological time scale would allow influx of species with limited dispersal mechanisms to reach, for example, previously glaciated or formerly non-agricultural sites. One factor tends to obscure comparisons between this study and most on native 2 grasslands. Molehills and dung pats average about 0.07m in area whereas most 2 of the previous studies were of larger disturbances (0.2 - 2.0m ). Size of opening can strongly influence patterns of germination and survival (Davis and Cantlon 1969; Miles 1974; Goldberg and Werner 1983). Miles (1974) found an inverse relationship between germination and survival in gaps of different sizes; 2 proportionally more seeds germinated in 25cm gaps but survivorship was highest 2 in 2500cm gaps. Rapp and Rabinowitz (1985) using artificial disturbances of 2 about 27cm found no difference in seedling survival in disturbed versus control areas. In addition, a plethora of evidence exists to support differentiation among species in pre-germination requirements and conditions for survival and growth. Although Grubb (1977) has argued that these differences in regeneration Invasion of disturbances / 89 requirements maintain species richness in communities, it appears that heterogeneity provided by molehills and dung pats is relatively unimportant to species richness in these pastures. The finding that space made available by disturbances (molehills and dung pats) is generally filled by the expansion of existing individuals is in accord with Sousa's (1984b) observations. Sousa observed that small patches were colonized rapidly by encroachment of adults along the edges and attributed the high rate of closure to the larger perimeter to area ratio of small disturbances. Rapid closure, i.e., within a year, of small disturbances may restrict colonization by species without persistent seed or long distance dispersal mechanisms. Because of the restriction of recruitment of new individuals these disturbances act primarily as agents of mortality removing individuals but providing limited opportunities for new individuals to enter the population. Individuals of a species are eliminated at random but replacement is by individuals (species) with specific characteristics. 4. INFLUENCE OF SPECIES REMOVAL ON INTERSPECIFIC INTERACTION 4.1. INTRODUCTION Over the past twenty-five years, perturbation experiments have been used increasingly to determine relationships between species and analyse the structure of communities, (e.g., Sagar and Harper 1961; Pinder 1975; Gross 1980; Fowler 1981; Hils and Vankat 1982). One type of perturbation experiment involves selective removal of one or more species from the community and measurement of the magnitude of the response by the remaining associated species. Early removal experiments measured the response of target species to the mass removal of associated groups of species. For example, Sagar and Harper (1961) found that vegetative expansion and seedling establishment of Plantago spp., and Putwain and Harper (1970) for Rumex spp., was restricted by grasses rather than by other dicotyledenous species. Later studies aggregated the target species to examine community structure. For example, Pinder (1975) removed dominant grasses in old-fields and detected a diffuse response of increasing biomass in the remaining subordinate species. Also in old-fields, Allen and Forman (1976) found removal of tall species caused the greatest changes in the response of remaining species. The response of species to removal was non-reciprocal; for example, few species could invade patches from which Potentilla simplex had been removed but P. simplex was able to invade areas made vacant by the removal of other species. 90 Influence of species removal on interspecific interaction / 91 Removal experiments were used to determine pairwise interspecific interactions by Fowler (1981) in grasslands and by Silander and Antonovics (1982) in coastal marsh-dune communities. Species interaction in both these communities were rarely reciprocal and could be specific or diffuse. In addition, interactions varied seasonally and changed when the same species pairs interacted in different environments (Silander and Antonovics 1982). The ability of removal experiments to determine competitive interaction between species has been challenged by Bender et al. (1984). They argued that except in a two-species community, the presence of a net gain in one species following the removal of another is not a sufficient or necessary condition to demonstrate competition. The gain reaction may actually result from interaction with a third species. Similarly, the absence of response to removal may not necessarily demonstrate absence of competition. Because of a network of interactions, response to removal may be weak or non-existent. To avoid this ambiguity, Bender et al. (1984) suggest the separate removal of each species in the community to create a complete interaction matrix. This approach is conceded to be impractical in most multispecies communities, not to mention prohibitively expensive in time and cost. Another consideration raised by Bender et al. (1984) is that of overlooked species. In defining a community such as a pasture in terms of botanical composition, some obvious and some inconspicuous members of the community are overlooked. Pastures, by definition, always include grazing species, whose effect on the pasture is enormous. Even if the effects of the overlooked species on target Influence of species removal on interspecific interaction / 92 species are very large, the interaction is negligible if the effects of the target species on the ignored species are very small (Bender et al. 1984). It is reasonable to assume that in pastures, although the cumulative effect of all herbaceous species on grazers is very large, the effect of each component species on the grazer is very small. Thus, the interaction between plant species can be validly studied in isolation from the consumer effect. Ecological theory argues that coexisting plants diverge in their use of resources and thereby reduce or avoid competition; Harper (1967) called this "ecological combining ability". Likewise, Fowler (1981), working in an 30-year-old grassland in North Carolina stated that "this is a community characterized by relatively weak and approximately equal competitive relationships among all of its component species " (p. 851). Aarssen (1983), however, suggests an alternative possibility: he argues for strong competitive interactions between individual species that become more equal over time. Furthermore, he argues that if competition is an organizing force in determining community structure, there should be strong, unequal interaction in recent pastures and strong, equal interactions in older ones. The predictions of both viewpoints are the same - strong interactions will be found in recently established communities. In older communities, interactions should be detectable if they are strong but not if they are weak or reduced through niche divergence. Unfortunately, it is beyond the scope of removal experiments to differentiate between the mechanisms that lead to equality in interactions among species over time. The physical proximity and similar sowing and management histories of the Influence of species removal on interspecific interaction / 93 Aldergrove pastures make them suitable to test for changes in species interactions over time. This chapter describes a removal experiment carried out using three of the common species: Lolium perenne, one of the originally seeded grasses, Trifolium repens a seeded legume, and the weedy grass Holcus lanatus. The design allowed an examination of the species interactions with each other and with associated species and the effects of their combined removal. If one of the three dominant species is removed, the remaining dominant species could control the resultant interactions. Moreover, removal of species combinations as well as single species allows any substructuring of the community to be observed. 4.2. METHODS 4.2.1. Field methods Four sites were selected in the 1939 and 1977 pastures. The sites were chosen to conform to the following criteria: (i) absence of any mole activity; (ii) no visually obvious discontinuities in vegetation; (iii) slopes of less than 5%; and (iv) absence of any evidence of waterlogging. The 1958 pasture was excluded from the study because of the difficulty in finding sufficient sites which met" these criteria. The sites were staked into 2.8 x 6m blocks and further subdivided into eight one metre square plots. Around each plot was a 10cm boundary from which all vegetation was removed, and an uncleared zone 20cm wide was left between Influence of species removal on interspecific interaction / 94 plots. Plots within each site were assigned randomly to one of eight removal treatments shown in Table 4-1. The four sites in each pasture were initially surveyed in July 1983. Rooted frequency in each plot was measured by placing a 50 point, 0.5x0.5m square 2 grid, four times, to give a total 200 points per lm plot. Although the sites selected for the experiment were visually uniform, considerable variation existed in rooted frequency of the three removal species and consequently, in the amount of bare ground exposed after removal (Table 4-2). Removal of the grasses and clover began in late July and continued throughout August and September. All plots were cleared of regrowth in September and October. A screwdriver or large iron nail was used to dig out the grasses and lift the Trifolium repens stolon and roots. Care was taken to avoid disturbing the remaining vegetation and loosened soil was pressed back into place. As far as possible, plots were kept dung-free during the duration of the experiment by removing any depositions. In July 1984, one year after the initial survey, the plots were re-surveyed in the same manner as in 1983. Subsequently, the plots were cleared of regrowth of all three species through fall 1984 and again in spring 1985. The final survey was made in July 1985. In an effort to minimize seasonal and management-induced variations, each survey was made as soon as possible after haymaking, although a timelag was necessary to allow regrowth of species to an identifiable size. Influence of species removal on interspecific interaction / 95 Table 4-1. Species removed from eight plots in the 1939 and 1977 pastures. Treatment Species removed C control plot, no species removed T Trifolium repens removed H Holcus lanatus removed HT Holcus lanatus and Trifolium repens removed L Lolium perenne removed LT Lolium perenne and Trifolium repens removed LH Lolium perenne and Holcus lanatus removed LHT Lolium perenne, Holcus lanatus and Trifolium repens removed Table 4-2. Percentage cover of removed species in four replicates in July 1983. Treatment Percentage cover removed 1939 pasture J_ T 6.5 H 9 HT 16 L 26 LT 54 LH 47 LHT 48.5 _2 12 14 17 25 41 41 47. _3 9.5 23 43 24.5 32.5 54.5 70 _4 11 29 31 18 38.5 54.5 70.5 mean 9.8 18.8 26.8 23.4 41.5 49.3 59.1 1977  pasture T H HT L LT LH LHT 15.5 11 30 9.5 35 26 37 8.5 8 26 14.5 28.5 22.5 46 6.5 18.5 38 28 39 32.5 43.5 19 27.5 59 34 45.5 54.5 75 12.4 16.3 38.5 21.5 37.0 33.9 50.4 Influence of species removal on interspecific interaction / 96 4.2.2. A n a l y t i c a l m e t h o d s To determine the effects of Lolium perenne, Holcus lanatus and Trifolium repens removal on the abundance of remaining species, the difference in rooted frequency between the initial 1983 and final 1985 survey was used. The model chosen is univariate, that is, effects of all removals were tested on one respondent species at a time. Some species, for example Rumex acetosella and Alopecuris geniculatus, occurred too infrequently for statistical analysis. Analysis of variance computer program 2V of BMDP (Dixon 1983), analysis with three fully crossed factors and no groupings, was used. The identification of three factors, removal of L. perenne, H. lanatus and T. repens from each relevant treatment whether as single species or combination examined the effects of single removals as well as the two- and three-way interactions. Because the removed species and the respondent species have different initial abundances, an expansion rate, (R..) was calculated to standardize the response of U species i to removal of species j. R.. is measured by using the change in rooted y frequency of species i from 1983 to 1985, divided by the original 1983 frequency, i.e. R.. = (N.. - N.)/N. (4-1) ij ij i i where N . - original abundance of species i (in 1983) N.. - abundance of species i (in 1985) when removed y species j is zero If the original abundance of a species in a plot was zero, the expansion rate Influence of species removal on interspecific interaction / 97 was not calculated for that treatment plot to avoid division by zero. Expansion rate (R„) was examined in response to the abundance of the removed species and as a response to amount of area available for colonization. 4.3. RESULTS 4.3.1. 1939 pasture There was considerable regeneration of Lolium perenne after initial removal but little of Holcus lanatus and Trifolium repens in July 1984; however, in August and September, T. repens invaded the removal plots by stolon extension from adjacent areas. In July 1985, regeneration of removed species was essentially zero. In the 1939 pasture, differences between controls and treatment plots were not significant (i.e. p>0.05) in most cases (Tables 4-3 and 4-4) despite the removal of large amounts of cover and the opening up of bare ground for colonization. Bare ground increased overall in the experimental sites but was significantly less in those from which H. lanatus had been removed. It does not necessarily follow that H. lanatus removal sites were more readily colonized. Rather, species had an extra growing season to invade bare ground created in 1983 because H. lanatus did not regenerate after removal and did not colonize from seed in 1984, so that little further treatment was required and no new bare spots were made in the second season. Neither Lolium perenne nor Holcus lanatus responded to the removal of the other nor to the removal of Trifolium repens. T. repens showed a significant increase T a b l e 4-3. Mean d i f f e r e n c e (n=4) i n s p e c i e s percentage c o v e r i n the 1939 p a s t u r e from 1983 t o 1984. Year e f f e c t i n d i c a t e s d i f f e r e n c e s i n abundance averaged over a l l t r e a t m e n t s . Removal e f f e c t i n d i c a t e s d i f f e r e n c e s i n abundance a t t r i b u t e d t o d i f f e r e n t t r e a t m e n t s . * - p<0.05; ** - p<0.01; *** - p<0.001. y e a r Spec i es c o n t r o l T H HT L LT LH LHT e f f e c t Agropyron repens 2 . 3 -2. .0 4. .8 -0 .5 -3 .3 4 .8 1 .5 -o . 2 0. . 59 Dactylis glomerata -10. . 2 0. .4 4. .8 6 .7 0 .0 -0 .5 0 .9 6 .9 0. . 59 Holcus lanatus -26. . 1 -26 . . 4 - - -28 .0 2 . 8 - - 0. .01** Lolium perenne -7. 4 5. .3 -4 . 6 2 .8 - - - 0. .31 Poa compressa -1 . 9 19. .3 -5 .3 13 .8 -5 .3 18 .5 9 . 7 23 .7 0. .34 Phleum pratense 0. . 3 1 . 8 0. .3 3 .6 -1 .5 0 .0 1 .3 -0 .4 0. . 22 Anthoxanthum odoratum 1 . 8 2. .2 1 . .3 1 .8 3. .8 4 .0 2 .4 1 . .8 0. ,05 Agrostis alba 0. 0 -3. 8 -1 . 3 0 .0 -2. .5 0 .0 0 . 5 5. .0 0. .73 T r i f o l i u m repens 39. 2 - 23 , .7 - 30, 8 - 39 .9 - 0. ,30 Ranunculus a c r i s 0. . 3 2. .9 1 . .8 2 .9 7. .8 3 .0 0 .8 5. .3 0. .02* Taraxacum of f i c i n a l e -1 . ,8 -2. 9 -5. . 3 -2 .4 0. .0 -5 .8 -11 . 2 1 , .0 0. ,02* removal e f f e c t H , LT LTH co T a b l e 4-4. Mean d i f f e r e n c e (n=4) i n s p e c i e s p e r c e n t a g e cover i n e f f e c t i n d i c a t e s d i f f e r e n c e s i n abundance averaged o v e r d i f f e r e n c e s i n abundance a t t r i b u t e d t o d i f f e r e n t t r e a t m e n t s . * the 1939 p a s t u r e from 1983 t o 1985. Year a l l t r e a t m e n t s . Removal e f f e c t i n d i c a t e s - p<0.05; ** - p<0.01; *** - p<0.001. y e a r removal S p e c i e s c o n t r o l T H HT L LT LH LHT e f f e c t e f f e c t Agropyron repens 12. .3 -1 .5 16 .5 3 .0 5. 8 13 .0 14 .5 1 .8 0. .23 Dactylis glomerata -6. .0 2 .5 4 .3 10 .8 3, ,5 5 .5 5 .8 15 .5 0. .24 Holcus lanatus -18. .0 -17 .5 - - -19, ,5 5 .0 - - 0. .06 Loli um perenne -18, . 3 -3, .5 -7. .5 -4 .8 - - - 0. ,38 Poa compressa 6 .3 22 .3 -1 .3 24 .0 10. ,3 30 .5 18 .5 47 , .8 0. .07 HT* Phleum pratense 0. .8 0 .0 1 .8 4 .5 0. .0 9 .3 3 .3 3 .0 0. ,04* Ant hoxant hum odoratum 2. . 3 4. .8 1 .0 7 .5 1 . 5 4 . 5 1 .5 8. .0 0. . 19 Agrostis alba 0. .0 -3. 8 -4. .5 -1 .0 -2. ,8 O .0 0 .0 6, .0 0. , 25 T r i f o l i u m repens 23. 8 - 23, .5 - 32. 5 - 48 .8 - 0. ,02* Ranunculus a c r i s 2, .3 4 .0 6 .0 5 .0 1 1 . .0 9 . 5 7 .5 10. . 3 0. ,04* Taraxacum o f f i c i n a l e -4 . . 5 -4. .0 -9, .8 -6 .3 -3. ,8 -5 . 3 -13 . 5 2, .8 0. ,01** HT* vo VO Influence of species removal on interspecific interaction / 100 (p = 0.017) in abundance in the two year period but did not expand preferentially in either the L. perenne or H. lanatus removal plots. Both Poa compressa and Taraxacum officinale responded to the joint removal of Trifolium repens and Holcus lanatus (p<0.05); but whereas, P. compressa abundance increased under this treatment, the decline of T. officinale was accelerated. Abundance of T. officinale in the 1939 field declined gradually over the experimental period (p = 0.006, see also Table 4-4). Ranunculus acris abundance increased (p = 0.036) although it was not significantly increased by any treatment. 4.3.2. 1977 pasture In July 1984, there was little regeneration of the removed species in three of the four replicates in the 1977 pasture, although Trifolium repens invaded the removal plots in August and September. The fourth replicate was recleared and, along with the other recleared replicates, showed little regeneration of removed species by July 1985. Several grasses responded to the removal of Trifolium repens and the combined removal of T. repens and Holcus lanatus (Tables 4-5 and 4-6). Dactylis glomerata increased with the removal of T. repens{p<0.01), H. lanatus (p<0.05) and with the LHT removal (p<0.01), and Phleum pratense increased with the combined removal of T. repens and H. lanatus (p<0.05). Both Poa compressa and Anthoxanthum odoratum increased in overall abundance during the experimental period (p = 0.020 and p = 0.004 respectively): both increased significantly with T. repens removal (p<0.01 and p< 0.001 respectively) and, T a b l e 4-5. Mean d i f f e r e n c e (n=3) i n s p e c i e s p e r c e n t a g e c o v e r i n the 1977 p a s t u r e from 1983 t o 1984. Year e f f e c t i n d i c a t e s d i f f e r e n c e s i n abundance averaged o v e r a l l t r e a t m e n t s . Removal e f f e c t I n d i c a t e s d i f f e r e n c e s i n abundance a t t r i b u t e d to d i f f e r e n t t r e a t m e n t s . * - p<0.05; ** - p<0.01; *** - p<0.001. y e a r removal S p e c i e s c o n t r o l r H HT L LT LH LHT e f f e c t e f f e c t Agropyron repens -16, .3 -0. .7 1 .6 6 .7 -1 .0 12.0 -o .7 7.7 0.65 D a c t y l i s glomerata -2 . 3 4 . , 7 0 .0 10 .0 0 . 7 5.3 8 . 3 2.3 0. 24 Holcus lanatus -5. .0 0. .3 - - 5 .0 13.0 - - 0.13 Lolium perenne -10. . 3 -21 . .3 -13 .3 -9 . 3 - - - 0.07 Poa compressa -3 .  3 10. 0 2 .7 5 . 3 3. . 3 7.0 0 . 7 39.7 0.07 r*,LHT Phleum pratense 2 . 7 -1 . 3 2 .7 3 .3 1 . 3 2.7 0 .0 -1.7 0. 53 r** Agrostis alba 7 . 7 0. .7 1 .3 2 .7 0, .7 0.7 1 .7 2 . 3 0.40 T r i f o l i u m repens 27. , 3 - 28 .7 - 24. .3 - 26 .3 - 0.89 Ranunculus a c r i s 0. 7 -0. 3 -1 .7 3 .3 1 . .7 1 .7 1 .3 2.0 0.30 Taraxacum o f f i c i n a l e -8 .  3 2 . 3 -7 .3 -0 .7 -8. .7 -1 .33 -8 .0 -8.33 0.03* T a b l e 4-6. Mean d i f f e r e n c e (n=4) i n s p e c i e s p e r c e n t a g e c o v e r i n t h e 1977 p a s t u r e from 1983 t o 1985. Year e f f e c t i n d i c a t e s d i f f e r e n c e s i n abundance averaged o v e r a l l t r e a t m e n t s . Removal e f f e c t i n d i c a t e s d i f f e r e n c e s i n abundance a t t r i b u t e d t o d i f f e r e n t t r e a t m e n t s . * - p<0.05; ** - p<0.01; *** - p<0.001. Spec 1es c o n t r o l r H HT Agropyron repens -10. .8 -21 .5 -15 .5 -9.0 Dactylis glomerata 3 . 8 14 .5 10 . 3 26.0 Holcus lanatus 13. .0 10 .8 - -Lolium perenne -25. .0 -20 .0 -14 .0 -4.5 Poa compressa -8 . 3 17 . 3 15 .8 22.5 Phleum pratense 4 . 3 0 .3 1 .0 5.5 Anthoxanthum odoratum 2 . 5 14 .3 5 .3 5.5 Agrostis alba 3. ,5 0 .0 0 .3 0.8 T r i f o l i u m repens 17. .3 31 .0 -Ranunculus a c r i s 3. ,5 3 .0 0 . 3 6.3 Taraxacum o f f i c i n a l e -10. 5 -7 .75 -8 . 25 -4 . 25 y e a r removal 1 LT LH LHT e f f e c t e f f e c t -10. .8 -9 .8 -16 .0 -8 .5 0. . 10 5. .3 18 . 3 17 .0 16 .5 0. .07 w*,r** LHT** 18. . 3 21 .8 _ _ 0. 0. .06 . 12 H* 4. .8 17 . 3 8 .3 42 .5 0. * .02 T** 1 . .5 2 .8 1 .3 0 .8 0. . 14 HT* 5. .8 10 . 3 8 . 5 13 .8 0, .004* T***,HT* -3. .0 0 .3 1 .0 0 .5 0. .45 24. ,0 - 39 .5 - 0. .0004* 3. .5 1 . 75 3. 75 7 .5 0. 094 -12. .0 -9 . 75 -7 .0 -11.25 o. .09 to ro Influence of species removal on interspecific interaction / 103 additionally, A. odoratum responded to the combined removal of T. repens and H. lanatus (p<0.05). Of the removed species, only Lolium perenne showed a significant (p<0.05) response to the removal of another species: L. perenne declined in abundance over the experimental period but its decline was less where Holcus lanatus had beeen removed. This relationship was not reciprocal: the removal of L. perenne did not affect the abundance of H. lanatus. Among the dicots, the trend of declining Taraxacum officinale and increasing Trifolium repens and Ranunculus acris abundances was the same as in the 1939 field (see also Table 4-6). No dicot responded significantly to any of the removal treatments. 4.3.3. Expansion rate Expansion rate (R„) was calculated using equation 4-1 and plotted as a function of the amount of Lolium perenne, Holcus lanatus and Trifolium repens removed (Figure 4-1). The response of R.. to the amount of each species removed used only single species removals to avoid the complication of differing or synergistic responses to two or more species removals. Figure 4-2 illustrates the relationship between expansion rate and the area available for colonization and uses cover removed from all four replications of all removal treatments as the independent variable. Expansion rate of a species was normalized using a Box-Cox transformation. The coefficient of determination of the regression of normalized Influence of species removal on interspecific interaction / 104 Figure 4-1. Expansion rate from 1983 - 1985 of species remaining in each of the four replicates after the removal of Lolium perenne, Holcus lanatus and Trifolium repens. Expansion rate at 0% cover removed is the average change in cover in the control plots. • Lolium perenne removed A Holcus lanatus removed • Trifolium repens removed 1939 PASTURE Lolium perenne 5 10 16 20 25 PERCENTAGE COVER REMOVED 30 1977 PASTURE 0.5-1 Lolium perenne 0 5 10 15 2 0 2 5 3 0 3 5 PERCENTAGE COVER REMOVED 1939 PASTURE Trifolium repens PERCENTAGE COVER REMOVED 1977 PASTURE Trifolium repena PERCENTAGE COVER REMOVED o 1939 PASTURE Taraxacum officinale PERCENTAGE COVER REMOVED 1977 PASTURE Taraxacum officinale -i r 8 K> 15 20 26 30 PERCENTAGE COVER REMOVED i 38 Influence of species removal on interspecific interaction / 109 Figure 4-2. Expansion rate from 1983 - 1985 of species remaining after removal treatments against the percentage cover removed in each of four replicates of eight treatments (n=32). Coefficient of determination and significance of the regression were calculated from normalized values of the expansion rate. EXPANSION RATE EXPANSION RATE 0 - 4 1 > > c n -s-- ^  > cn trjg fig 3 ft gs 2 c n B s t> > r o _ J _ ^ o _ 5 1 1 1 to CM-CR s 3 o ** CD K> z o on <o _ o i . gtt s° B g NI CO _1 1_ I 81 00 , s: o < m O 0 - » o > en -s-—* c n -K> O c • 8-CO 0 0 1 > CO _ 01 s-s-g -c n . 01 > 8-en _ c n 3-c n EXPANSION RATE EXPANSION RATE * c n a «J OB J I I 1 1 > fe > 5 A S CO <A CD 0> m o c n 0 5! • o o 8-c o o 8! i s B g s 3 1 8 £ P 1 - r» 01 o c n — 01 N 01 ' I 1 1 ' ' 1 > c > > > o n EXPANSION RATE EXPANSION RATE 8-6 <S o eb o o bp is * k> o io Xk —I 1 1 1 I I I t> 5 3D m g 1 3 a a o> o cn I O cn I CO co CO CO 30 m A | EXPANSION RATE o cn _ l _ to _JL_ fo co cn u cn _ l I 1_ 8H x> 3 A 3= I 2 6 cn EXPANSION RATE cn > t> > > co co co 5 33 m "I 8 i i s CO III EXPANSION RATE EXPANSION RATE _ I_ _ L _ cn _ l i O l -a-<*-8-8-8-a-s-$-8-a-8-8-as: > > > i 5 cn -2 °i o 9 s 3 3 EXPANSION RATE EXPANSION RATE 6 o Ol _ cn cn 1 1 1 B » > 1 i 1 1 o > > > o > > > -Ml I 9. - 8 § 5-s c? 5 6 cn O cn 3 z o 8 ° \J at m " 3 a 9 8 a a-o - > > > > 3-8 211 EXPANSION RATE EXPANSION RATE 0 6 6 6 0 0 0 0 J. te e»; * k* O M * to to i 1 1 1 1 I I I L _ > > > 7 - s 5 6 0 6 6 0 0 _ J I I I I I I I I • o-8 3 > EXPANSION RATE EXPANSION RATE i O _ M CO _ l I L _ o» a «J OB as _ i 1 1 ' ' > > > 5 a i m 9 P 8 3 CP -si CO _ l I ' t> £11 Influence of species removal on interspecific interaction / 114 scores on percentage cover removed is included in Figure 4-2. Poa compressa which showed a significant response to Trifolium repens removal in the 1977 pasture and to HT treatment in the 1939 pasture (Tables 4-6 and 4-4), showed a concommitant increased expansion rate with increased Holcus lanatus removal in both pastures. Moreover, expansion rate increased with increased T. repens removal in the 1939 pasture but fell off sharply when large amounts of T. repens were removed in the 1977 pasture (Figure 4-1). Expansion rates of Dactylis glomerata and Lolium perenne showed no consistent pattern of response to removal (Figure 4-1); however, both D. glomerata and L. perenne were able to take advantage of open space and respond with increased expansion rate (Figure 4-2). Ranunculus acris appeared to have a high expansion rate; this is anomalous because it changed from zero presence in some plots. Both the rate of R. acris expansion and the rate of T. officinale decline were independent of treatment. Expansion rates of species were averaged over replications of each single species removal and over the change in abundance in the controls for 1983 - 1985 period (Table 4-7). Values are only for those species originally present in three or more of the four replicates. Expansion rates in controls were very similar to those in removal treatments. Some species, such as Trifolium repens showed consistently high expansion rates while others, such as Dactylis glomerata and Agropyron repens, showed rapid spread only on occasion. Analysis of variance on Box-Cox transformed expansion rates and a non-parametric test (Kruskal-Wallis) Influence of species removal on interspecific interaction / 115 Table 4-7. Mean expansion rate for the period 1983-1985 for species found initially in at least three of four replicates. R I C is the response in control plots, Rj/j in Holcus lanatus removal plots, R^ in Lolium perenne removal plots and Rj ( in Trifolium repens removal plots. Species Expansion rate 1939 1977 Agropyron Ric 1.767 -0.715 repens Rit -0.410 -0.745 Rih X -0.797 Rfl 0.783 -0.730 Dactylis Rz'c -0.567 0.100 glomerata Rit -0.040 1.117 Rih 0.450 0.755 Ru 1.148 0.055 Holcus lanatus Rfc -0.562 0.507 Rit -0.330 0.325 R« -0.355 0.772 Lolium perenne Ric -0.347 -0.535 Rit -0.015 -0.385 Rih -0.110 -0.227 Poa compressa R ic 0.485 -0.227 Rit 0.805 1.220 Rih 0.307 1.560 Ril 0.457 0.980 Trifolium RW 1.412 1.070 repens Rih 2.767 1.022 Ril 1.815 1.322 Taraxacum Ric -0.275 -0.625 officinale Rit -0.207 -0.535 Rih -0.437 -0.685 Ril -0.150 -0.722 x- species present in less than three of four replicates in 1983. Influence of species removal on interspecific interaction / 116 on untransformed scores showed the rates were not significantly associated with treatment, and consequently, significant responses to removal discussed in the previous section cannot be explained by differential expansion rates of the responding species. 4.4. DISCUSSION Grass species had a greater number of significant responses to removal in the 1977 pasture than in the older 1939 pasture. Other authors have found that removal experiments in old fields and pastures demonstrate weak or diffuse interspecific interactions. For example, Fowler (1981) found no evidence of differences in magnitude of response of one species to the removal of others and concluded that relationships among species were relatively weak and diffuse. Veresoglou (1983) found that removal of subdominants, both in an old pasture and under more controlled glasshouse conditions, had little impact on remaining species. That members of the same species showed a significant response to removal in the newer but not in the older pasture supports the hypothesis that competition between neighbouring species abates over time, either through niche divergence (Harper 1964), or by increased equality of resource acquisition (Aarssen 1983). Support, however, depends on acceptance of the 1977 and 1939 pastures representing a chronosequence. Like all chronosequences based on spatial arrangements of differing ages, this can be challenged. There is no doubt that the genetic composition of the sown species differs from 1939 to 1977. Lolium Influence of species removal on interspecific interaction / 117 perenne is the diploid variety in the 1939 pasture and tetraploid in the 1977 pasture. Nevertheless, L. perenne collected from both pastures, despite initial phenotypic differences, had complete overlap of morphological characters after growing for two years in a common garden (R. Turkington, unpubl. data). Similar results were observed for Trifolium repens and Holcus lanatus. Abiotic factors are unlikely to be substantially different in the two pastures because of close proximity and similar topography. Although soil differences exist between the two pastures (Aarssen and Turkington 1985a), many of these difference may be due to soil development over time, i.e. the multiplicity of processes triggered by plant growth and decay. This is not to say that the 1977 pasture will mirror the present condition of the 1939 pasture 38 years in the future because past history influences further development. Rather, it suggests that relationships between species will follow a similar course of development. If strong competitive interactions occur in the newer pasture, what species compete? The increased abundance of the grasses, Dactylis glomerata, Poa compressa and Anthoxanthum odoratum, upon removal of Trifolium repens, may have resulted from their release from suppression by T. repens, although this seems unlikely. An alternative and testable hypothesis is that the removal of T. repens provides a source of nitrogen for which other species compete. Nitrogen products left behind by decaying T. repens roots and nodules can facilitate growth and expansion. If the grasses are able to take advantage of this nitrogen source, competition is not between T. repens and the grasses but among grasses for the nitrogen made available by T. repens decay. Influence of species removal on interspecific interaction / 118 That the increased abundance of grasses on Holcus lanatus removal is because of release from competitive suppression is supported by other studies. Evidence of competition was found in replacement series studies (Aarssen 1983; Turkington and Aarssen 1984) and pairwise competition studies (Aarssen and Turkington 1985b), in which H. lanatus demonstrated a strong competitive advantage over Lolium perenne which, however, decreased with pasture age. In Dactylis glomerata - H. lanatus interactions, D. glomerata had a competitive advantage, but demonstrated a similar trend of increasing equality of yield with increasing pasture age (Aarssen and Turkington 1985b). These trends were postulated to result from two different mechanisms: niche differentiation between D. glomerata and H. lanatus, and balancing of the competitive abilities of L. perenne and H. lanatus. Thus, H. lanatus may act to control the expansion of other grasses in the initial stages of community formation. In removal experiments, it is possible that species respond directly to the presence of open space. Dactylis glomerata in the 1977 pasture showed a trend of increased expansion rate in response to increased available space and increased significantly in the LHT combined removal which always left large amounts of open area. A lack of response to L. perenne removal, however, indicates that space alone is not the only consideration. The presentation or geometry of that space may be involved. Jayasingam (1985) found that the removal of densely rhizomatous, clumped plants had a greater impact on the abundance of remaining species than removal of loosely rhizomatous species. In removing the various species, we observed that clearing Holcus lanatus, a densely tillering, clumped species, left larger, more discrete gaps than the other removals. These Influence of species removal on interspecific interaction / 119 observations do not explain the differences in response of the same species to the same treatment in two different pastures. The applicability and utility of removal experiments to study community structure and determine competitive interactions has already been challenged (Bender et al. 1984). Additional limitations of the approach became evident during the study I conducted. Removal experiments in pastures may inadvertently mimic grazing, albeit selective grazing, and therefore, may not represent a novel perturbation. , Indeed, at the end of the grazing season, it was visually impossible to distinguish removed from grazed areas. Grazing, however, leaves roots and rhizomes that later regenerate. Grazing provides an intermittent but directional pressure to which plants in the older pasture have been subject for a longer time period. I have argued previously (Chapter 2) that biological accommodation seems to take place rapidly, i.e. within five years from initial establishment, and will show in Chapter 5 that species composition responds rapidly to management practices. In this chapter, the experiment may be uncovering fine-scale accommodation that continues for many subsequent years. Some portion of the measured response may be attributed to dimunition of competition but some also reflects convergence in response to grazing pressure. 5. INFLUENCE OF MOWING, FERTILIZATION AND DISTURBANCE REGIMES 5.1. INTRODUCTION Pastures are maintained by grazing or mowing. These activities remove biomass, eliminate individual plants and influence the distribution, abundance and competitive balance between species. Darwin (1859) recognized that grazing and mowing prevented vigorous species from eliminating the less vigorous in a lawn experiment. It has long been recognized that grazing management of pastures has a profound influence on botanical composition (e.g., Jones 1933a,b,c; Baker 1937; Milton 1940). Evidence from both grazing trials and manipulations of cutting frequency and intensity indicate that species' response to changes in the management regime can be quite rapid, taking place in a time frame of within a few months to a year or more (Jones 1933d; Gervais 1960). In addition, the experimental exclosure of grazers such as rabbits (Tansley and Adamson 1925; Watt 1957) and sheep (Welch and Rawes 1964) have demonstrated a decrease in the number of plant species and, the addition of grazers in biological control experiments (Dodd 1940 in Harper 1969; Huffaker and Kennet 1959) has shown an increase in species diversity. The results of these latter additive experiments are unfortunately confounded by disruption of the existing vegetation prior to experimental manipulation (Harper 1969). Grime (1973) has proposed that forms of management such as grazing or mowing prevent competitive species from attaining maximum size, thereby 120 Influence of mowing, fertilization and disturbance regimes / 121 reducing their competitive edge and increasing species diversity. Caswell (1978) viewed the grazer as a predator, and argued that grazed species are maintained through a form of predator-mediated co-existence which delays competitive exclusion. In the absence of grazing, competitive exclusion dominates; at high grazing densities, grazers would be non-selective and exert a uniformly high predation pressure resulting in a decrease in species co-existence; and, at low grazing densities, grazers would be selective and exert an intermediate, patchy pressure resulting in an increase in co-existence (Caswell 1978). Low nutrient levels constrain the rate of dry matter production of vegetation (Grime 1973, 1977) while fertilization usually increases size and growth rate (Spedding 1971). In pure stands, accelerated growth rate increases the rate of mortality (Sukatschew 1928, in Harper 1977) and in mixtures, can alter the competitive outcome (Stern and Donald 1962a,b). Mahmoud and Grime (1976) found no change in the hierarchy of dominance when fertilizer was added to grass mixtures. Arrhenatherum elatius remained the dominant grass in all mixtures and completely eliminated the subdominant grass, Festuca ovina, at higher fertilizer levels. The addition of fertilizer to nutrient-deficient vegetation usually causes large-statured species to expand and reduces the number of species in a given area (Grime 1979). Ecological and agronomic studies confirm the importance of grazing and nutrient levels in determining plant abundance and influencing species co-existence but little attention has been given to their interaction with the small gaps that are inherent in animal grazing e.g., rabbit scrapings and cattle hoof skids. These Influence of mowing, fertilization and disturbance regimes / 122 small gaps may function to provide micro-sites for seed germination and seedling establishment (Silvertown 1981), serve as a refuge for non-aggressive species, or facilitate vegetative expansion of existing plants. Both Grubb (1977) and Grime (1979) have stressed the importance of gaps in facilitating the ingress of, and maintaining, species in the community. To properly understand the influence of grazing and fertilization on pasture composition requires controlled field experiments. This facility was not available at the Aldergrove pastures so that to estimate this, a pasture-like area was seeded at the University of British Columbia South Campus Field Station. An experiment was designed to measure the effects on pasture species of the simultaneous manipulation of levels of mowing, fertilization and removal of small divots to create small gaps and simulate discrete disturbances. Although mowing differs from grazing in many ways (e.g. it is homogeneous, lacks the biting and pulling action of the grazer, and is non-selective) it is a best approximation of grazing in a field station facility. 5.2. METHODS Preparation for the experiment commenced in May 1982 in a section of the 2 South Campus Field Station. Seed was sown at a rate of 4g/m ; the seed mix (Table 5-1) approximated the proportions of Buckerfields Highland Mix over the past few years. This is the same mix used in the youngest pasture at Aldergrove in 1977. It was fertilized in June 1982 with half normal dilution of 20-20-20 fertilizer and weeded as required. By May 1983, the sown species Influence of mowing, fertilization and disturbance regimes / 123 dominated the plots and no further weeding was necessary. The experimental procedure began in May 1983. The seeded area was subdivided 2 into 8 blocks. Four of the blocks received 150g/m of 4-12-8 dry fertilizer every 3 weeks and the remaining four received no fertilizer. Each pair of fertilized -unfertilized blocks were further subdivided into 4 mowing treatments: every week, once every 3 weeks, once every 6 weeks and no mowing. Within each of the 2 eight combinations of mowing and fertilizer treatments were four lm plots. 2 From these lm plots, different 'disturbance regimes' were imposed by removing circular divots 10cm in diameter. Originally I had planned four small gap intensities: no divots, 1 divot every week, 2 divots every 2 weeks and 4 divots every 2 weeks but this proved to be too intense. Subsequent to July 1983, the regime was modified to: no divots, 1 divot every 2 weeks, 2 divots every 4 weeks and 4 divots every 4 weeks. The divots were made by randomly locating the appropriate number of spots and pounding a metal plug into the ground to remove the upper l-2cm of soil and rooted material. The treatments ran from May to October of each year from 1983 to 1985 and are planned to continue beyond this date to assess longer-term effects. All plots were surveyed in September 1985, using four .5 x .5m quadrats per 2 lm plot. Each quadrat was subdivided into a grid of 25 points and the rooted cover at each point was recorded (100 points per plot). Lolium perenne and L. multiflorum were not distinguished because of difficulties in differentiating them in the vegetative state. Influence of mowing, fertilization and disturbance regimes / 124 Table 5-1. Percentage composition of the original seed mix. Species Percentage Dactylis glomerata Lolium perenne Lolium multiflorum Phleum pratense Trifolium pratense Trifolium repens 46 12 8 4 15 15 5.3. RESULTS In September 1985, all of the seeded species plus several volunteers were present in the plots (Table 5-2). The origin of Festuca pratensis is uncertain although it can occur in commercial seed sources of Lolium spp. Of the seeded species, only Dactylis glomerata was found in all treatments and its mean abundance was similar to the input frequency of 46%. The abundance of ryegrasses (Lolium spp.) and clovers (Trifolium repens and T. pratense) was considerably lower than had originally been seeded. Analysis of the abundance data was univariate by species using a BMDP (Dixon 1983) fully crossed factorial analysis of variance. Only the nine most common species (>2% cover) plus bare ground were analysed. Mowing had a highly significant effect on the percentage cover of all species except Phleum pratense; fertilization had a significant effect on six species; but, divot removal had a significant effect only on Dactylis glomerata (p = 0.0509) (Table 5-3). Influence of mowing, fertilization and disturbance regimes / 125 Table 5-2. Mean and standard deviation of the percentage cover of species and bare ground on South Campus treatment plots in September 1985. Species Mean S.D. Dactylis glomerata 45.5 16.4 Lolium spp. 8.89 7.01 Agrostis sp. 6.73 12.1 Hypochoeris radicata 5.04 5.83 Festuca pratensis 4.91 4.60 Trifolium repens 4.48 6.20 Rumex acetosella 3.44 5.21 Trifolium pratense 2.89 4.58 Phleum pratense 2.34 2.51 Poa sp. 0.43 1.63 Holcus lanatus 0.23 1.12 Taraxacum officinale 0.21 0.77 Cirsium arvense 0.20 0.62 Plantago major 0.13 0.61 Medicago lupulina 0.13 0.69 Cerastium vulgatum 0.07 0.36 Plantago lanceolata 0.05 0.21 Alopecuris pratensis 0.02 0.12 Agropyron repens 0.02 0.12 Moss 0.03 0.17 Bare ground 14.3 11.9 Because divot removal rarely had a significant effect, these treatments were combined to show species response to fertilization and mowing (Figure 5-1). The percentage cover of Dactylis glomerata was reduced by frequent mowing but other species responded positively to some frequency of cutting. The fine weedy grass, Agrostis sp., was abundant only in the frequently mown blocks, whereas Lolium spp. and Festuca pratensis were more abundant at intermediate mowing frequencies. Trifolium repens was reduced less by more frequent mowing than the taller, more upright T. pratense. Weedy dicots, Rumex acetosella and Hypochoeris radicata were most abundant at intermediate mowing frequencies. Influence of mowing, fertilization and disturbance regimes / 126 Table 5-3. Percentage variation in abundance of common species and bare ground explained by different mowing (m), fertilization (f) and divot removal (d) regimes. *** - p<0.001; * * p<0.01; * - p<0.05. Species m f m-f d m-d f-d mfd #Dactylis glomerata *** 31 ** 37 8.2 * 1.7 1.2 0.4 3.1 #Lolium spp. * * * 60 • * 7.1 ** 5.5 1.2 1.9 0.9 * 3.9 #Phleum pratense 3.0 * 44 1.6 3.4 4.5 1.4 3.2 #Trifolium repens * * * 30 ** 32 18 1.3 1.5 * 1.9 1.9 #Trifolium pratense * * * 23 ** 31 * * 16 0.3 4.6 0.2 4.1 Festuca pratensis * * * 20 0.4 2.7 4.9 * * 12 1.0 7.2 Agrostis sp. *** 53 1.2 *** 6.2 0.4 1.8 0.2 1.5 Hypochoeris radicata * * 24 * * 27 *** 11 1.3 2.6 0.6 2.1 Rumex acetosella ** 28 2.0 9.8 2.5 3.5 0.3 1.8 bare ground * * * 66 2.1 ** 6.2 2.0 2.4 0.9 2.1 Species labelled with an # were constituents of the original seed mix. Influence of mowing, fertilization and disturbance regimes / 127 Figure 5-1. Percentage abundance of nine species on fertilized and unfertilized plots subject to different mowing regimes. CS Agresf I* tp. B HyfwxJiooii i«»urtu CD Fnrhicn pffltwwto • Irlfatum r«p«f» S I (bjmax oeatoaale C3 VtfoAufn prfftanM 70 - i 6 0 -5 0 -4 0 -3 0 -2 0 -1 0 -Fertilized plots CD Q_ Unfertilized plots 6 0 - i O C o T5 C =3 O <D o 5 0 4 0 3 0 -10 Influence of mowing, fertilization and disturbance regimes / 129 Fertilization increased the abundance of all the responding grasses and decreased the abundance of clovers, a result typical of grass-clover mixtures (Herriott and Wells 1960; Stern and Donald 1962a,b; Dennis and Woledge 1985). Many species showed a strong interaction effect between fertilization and mowing and some species responded to an interaction with divot removal. Mowing and low intensity gap formation increased the abundance of Festuca pratensis whereas lack of fertilizer and medium intensity gaps increased the abundance of Trifolium repens. The amount of bare ground decreased with the addition of fertilizer and with increased mowing frequency. At the highest mowing frequency, however, there was a greater proportion of bare ground in the most intense divot removal treatment leading to an interaction effect. 2 The number of species per lm plot ranged from 2 - 14, averaging eight species per plot. The lowest number of species was associated with fertilized, least frequently mowed plots and the greatest number with unfertilized, most frequently mowed plots (Figure 5-2). The distribution of the variance of number of species per plot was unknown, and therefore, analysis of variance was not strictly applicable to test for treatment effects. Nonetheless, factorial analysis of variance was used in an exploratory manner, because the large number of treatment cells constrained the use of alternate non-parametric techniques. Results were similar for both untransformed and square-root transformed values: highly significant mowing and fertilizer effects (p< 0.001) and strong interaction effects Influence of mowing, fertilization and disturbance regimes / 130 Figure 5-2. Number of species per square metre on fertilized and unfertilized plots subject to different mowing regimes. fertilized unfertilized tex Influence of mowing, fertilization and disturbance regimes / 132 (p<0.05) but no effect from divot removal on the number of species in the plot. 5.4. DISCUSSION Mowing and fertilizing were found to strongly influence species abundance and plot composition. The response of species abundance to mowing and fertilizer regimes has been associated with tolerance of shading and many large-statured species such as Dactylis glomerata, responding primarily to nitrogen, shade smaller-statured species (Stern and Donald 1962a,b; Snaydon 1971; Rhodes and Stern 1978). Shading is further exacerbated by infrequent cutting. Fertilization of grasses frequently suppresses low growing clovers, because grasses are able to grow taller in response to nitrogen application. Mowing can also provide nitrogen to grasses, because cutting clovers causes leaf and nodule drop, thereby increasing the amount of nitrogen transferred to grasses and favouring their growth (Bland 1967). Unlike fertilization alone, the repeated cutting of grasses and clover prevents shading, thereby ameliorating suppression of clover. The cultivars used in this experiment are typical of those sown in pastures. Most cultivars sown in pastures are selected for herbage production, i.e., rapid, upright growth and to a lesser extent, seed production (Snaydon 1984). These cultivars require a nutrient-rich environment to maintain them as the dominant species (Grime 1979). Grime relates the reduction in species density over the past thirty years in European meadows and pastures to increased "competitive dominance brought about by stimulating the yield of more robust and productive species and genotypes through application of high rates of mineral fertilizer" (p. Influence of mowing, fertilization and disturbance regimes / 133 125). In the absence of continued fertilizer input, however, botanical composition of pastures changes to a mixture of sown, naturalized and weedy species. In addition, grazing selects against many of the characters for which cultivars are bred so that naturalized pasture populations are more likely to contain more low growing and laterally spreading forms than the original seed source. Snaydon (1984) also points out that the traits of cultivars may well carry the penalty of poorer persistence so that established pastures will contain a greater range of variation in plant populations as well as higher species density. In pastures, grazing removes biomass that would shade and suppress the growth of less vigorous or small-statured species, thereby delaying competitive exclusion of these species. Competition for light, water and nutrients interacts with grazer - predator frequency, intensity and selectivity to determine botanical composition and species co-existence. Grime's (1973) proposed model, in which maximum potential species density is found at intermediate levels of nutrient availability, predicts that maximum species density occurs in relatively infertile pastures. In infertile pastures, the interaction of the five processes (Grime 1979) that control species density are such that (i) dominance of large species is reduced; (ii) nutrient "stress" and (iii) disturbance are increased which promotes (iv) ingress of suitable species and (v) niche differentiation. Accordingly, the continued decline in abundance of Dactylis glomerata in the Aldergrove pastures, could be attributed to failure to maintain high fertility levels coupled with reduction because of its sensitivity to grazing. The experiment in this chapter confirms that mowing and fertilization influences Influence of mowing, fertilization and disturbance regimes / 134 species composition and co-existence but negates a role for small disturbances. The created gaps appear to neither facilitate nor impede species abundance and spread, and to have no effect on species richness. These gaps provide no additional opportunities for new species to establish in the plots. These results agree with the observations of Rapp and Rabinowitz (1985) who created 2 disturbances of about 27cm in tall grass prairie and found that seedling survival was no greater on disturbances than in undisturbed controls. This is probably related to a number of factors. First, constituent species of the community may have either no need for gaps or require a larger area to trigger germination (Grime 1979; see Hilton et al. 1984 for variation within a species). A second factor is the influence of surrounding plants on survival and growth of seedlings. In general, more space leads to higher seedling survival and growth (Gross 1980; Snaydon and Howe 1986) although intolerance of heat and desiccation may limit establishment of some species in very large openings (Harper 1977). A third consideration is that even if gaps were favourable for seedling establishment, the amount of divot removal was inadequate and did not significantly increase the amount of bare areas available. Rather, the results indicated that plant spacing in response to cutting and fertilizer treatment had a greater effect on creating bare ground than actual divot removal. The results of this experiment invoke a comment on certain experimental practices to determine competitive outcome on interactions between species. It is apparent that the outcome of species interactions can be manipulated, to a greater or lesser extent, by the cutting and fertilization regime. Numerous examples exist such as the increased production of Lolium perenne relative to Influence of mowing, fertilization and disturbance regimes / 135 Dactylis glomerata with increased frequency of cutting or earliness of defoliation (England 1968) or the reduction in suppression of Agrostis capillaris (syn. A. tenuis) by L. perenne with severe defoliation (Harris and Thomas 1972). Replacement series, additive, and Nelder designs used to test competitive interactions, all require management regimes which often include fertilization and clipping for biomass estimation. Such management can strongly influence the outcome of species interactions and the results of such experiments can be generalized only with extreme caution. 6. GENERAL DISCUSSION The sun rises and the sun sets and hurries back to where it rises. The wind blows to the south and turns to the north round and round it goes, ever returning on its course. All streams flow to the sea, yet the sea is never full. To the place the streams come from, there they return again. Ecclesiastes, Chapter 1, vs 5-7 The chapters of this thesis focus on aspects of the role of disturbance in a pasture community. Disturbance by grazing or mowing is intrinsic to the maintenance of pastures in non-grassland climates. In addition, on a local scale "grazing animals frequently sit, lie, scratch and paw on the pasture in addition to walking, running and jumping on it" (Spedding 1971) and they also deposit dung and urine. Add to these activities, mounds of small animals, insect grazing and mechanical disturbance from haymaking and harrowing and the view of pastures as dynamic and disrupted is confirmed. It is not surprising that one feature of the Aldergrove pastures is the variability associated with plant abundance because fluctuations in abundance in grasslands and pastures are well documented (e.g. Rabotnov 1966, 1974; Spedding 1971; Vickery 1981; Thorhallsdottir 1984; Snaydon 1985). They are characteristically reversible, and associated with variations in meteorological and (or) hydrological conditions on a seasonal, yearly or short-term (e.g. 1-3, 5-10 years) basis 136 General discussion / 137 (Rabotnov 1974). Biotic interactions, such as the nitrogen-mediated feedback loop that is posited for Lolium/Trifolium fluctuations (Blaser and Brady 1950), also lead to changes in relative abundance. Moreover, management decisions on timing and extent of mowing, grazing and fertilizing, effect species composition and abundance. Yet, despite the flux of species composition, the botanical composition of the pastures is remarkably stable. Over the eight years that records have been kept for these pastures, only the 1977 field showed directional change in species composition and abundance and this ceased 4-5 years subsequent to ploughing and seeding. No species present in the initial seed mix has been completely eliminated from these pastures although the abundance of Trifolium pratense is severely reduced. Rather, it appears that the initial species composition determined the subsequent composition although proportions have changed (cf. Egler 1954). Ingress of weedy species occurs rapidly - witness the abundance of Holcus lanatus, Taraxacum officinale, Agropyron repens and Ranunculus acris. The composition of the pastures appears, nevertheless, to be constrained by the major biomass/abundance contributors which vary in relative proportions from year to year. Less abundant species are interspersed in various patterns which apparently have little impact on the total dynamics. One can imagine visiting the pastures in future years and seeing the same species maintaining various proportions and distributions of abundance within some set of limits for which neither boundary conditions nor rules of assembly are yet defined or understood. Maintenance of the limits and rules appears to be dependent on the continuation General discussion / 138 of the management practices in the pastures. Holling (1973) notes that in 'real world examples' systems may persist indefinitely until a novel introduction changes the driving variables and flips the system outside the boundary conditions. For example, permanent or long term removal of herbivores or repeated harvesting for maximum yield are likely to strain the resilience of the system and push it beyond current boundaries. But, within the current management practices, the pastures appear to exhibit considerable variation within stable limits. The three pastures are not entirely closed communities, existing as they do within a matrix of semi-rural copses, roads and rights-of-way that constrains species and gene influx. Both plant and animals are able, however, to disperse to a limited extent between the pastures. Observations suggest that the continued presence of Trifolium pratense in the older pastures is derived from recruitment from adjacent right-of-ways whereas the occasional local outbreak of Gnathalium uliginosum is from a persistent seed bank. These species are minor constituents of the community, but major constituents, such as Holcus lanatus, have dispersed into the pastures, not only from adjacent areas but also, via a seed bank accumlated over time (see Grime 1979). Small disturbances such as molehills and dung pats have been shown to play little or no role in providing access for weedy species to establish neither in the Aldergrove pastures nor on South Campus. In order for seedlings to grow, however, a space of some size must be available. The size required will be dependent on species (e.g. Denslow 1980) or genotype (Hilton et al. 1984). Many General discussion / 139 species have a requirement for light, or high or fluctuating temperatures in order to germinate. Hilton et al. (1984) found that Poa trivialis seed from plants growing in closed grassland swards germinated under low red:far-red ratios similar to dense shade conditions whereas those from open, arable land required higher red:far red ratios. Other studies have shown that the rate of seedling emergence is lower when vegetative cover is high (e.g. Fenner 1978; Silvertown 1981; Gross 1984) and growth and survival is typically reduced (Ross & Harper 1972). On average, between 1.2% (in the 1939 pasture) and 3.4% (in the 1958 pasture) was bare from various causes during the study period. Grazing and associated activities removed considerable biomass and left numerous small gaps because plants, some more readily than others (e.g. Dactylis glomerata) were more readily uprooted. Very small gaps may be sufficiently large for many pastures species to regenerate. Indeed, root death after loss of shoot mass may even provide, after complex cycling through soil fauna, an enriched nutrient environment for seedling establishment (Newbery 1979). Several authors (e.g. King 1971; Snaydon & Howe 1986) have shown that nutrient limitation by established plants is a major deterrent of seedling establishment although the extent to which this affects survival is species • dependent. In the pastures, molehills and dung pats provided sites for establishment of Trifolium repens, and to a lesser extent, Ranunculus acris seedlings. Neither species required these gaps to maintain their presence in the community. Ranunculus acris was found to germinate and survive in a variety of gap sizes General discussion / 140 whereas T. repens spread by stolons throughout the pastures. Cahn and Harper (1976) suggested that frequent inputs of new genotypes of clonal species is essential to avoid local dominance by a few either competitively dominant or long-lived clones. Occasional seedling establishment (3% of the ramet population per year) is sufficient to maintain considerable genet variation within a population (Soane and Watkinson 1979). In good seed years, e.g. 1984, seedling establishment of this magnitude was calculated to occur in the pastures (R. Parish, unpublished) but the frequency of 'good' seed years has yet to be established. The variability in seed germination recorded for Trifolium repens, cautions against generalizations that molehills and dung pats never have a role in sexual regeneration of pasture species. In the fall of 1982, a large flush of monocot seedlings was observed in the three pastures both on and off molehills. It is a pity that no crystal ball or seer provided the foresight to measure seedling survivorship that first year because no other monocot flush was observed in subsequent years. It appears that seed germination is highly episodic in these pastures. Gaps such as molehills and dung pats did provide opportunities for vegetative expansion. These gaps appeared to function a in manner similar to small gaps in forests (Runkle 1982) allowing competitive release of neighbouring individuals. Pasture species that are able to take advantage of these openings are usually those capable of rapid lateral spread either by rhizomes (e.g. Agropyron repens, Poa compressa) or stolons (e.g. Trifolium repens). General discussion / 141 Formation of molehills and dung pats killed most plants that were buried under them although those capable of regeneration from fleshly roots (e.g. Taraxacum officinale) survived temporary burial. In general, the probability of death for a genet depends on its size. The contribution of small disturbance to the removal of genets of such species as Trifolium repens and Poa compressa, in which ramets may be spread over large areas was probably minimal, but genet death could result for closely tillering grasses such as Holcus lanatus and Lolium perenne. McNeilly & Roose (1984) found that the number of genotypes of L. perenne per unit area of an old pasture was considerbly lower than in younger pastures. The loss of genotypes during sward establishment has been related to the loss of smallest individuals, often resulting from small differences in timing of seedling emergence (Ross & Harper 1972). Subsequent losses have been attributed to competitive dominance of a few superior genotypes which are adapted to sward management practices (McNeilly & Roose 1984). Strong competitive interactions have yet to be demonstrated in pastures (Strong 1983) despite generally accepted statements that they do exist (Donald 1963; Snaydon 1978). Small disturbances serve to randomly remove individuals yet provide no consistent opportunity for new recruits to enter. Depletion of genotypes may, in part, result from random removal without replacement, rather than selective processes. Species respond differently to disturbance of various sizes. Trifolium repens could utilize molehills and dung pats for both sexual propagation and vegetative colonization. Dactylis glomerata was frequently found to occupy the same point from fall to spring. In addition, D. glomerata generally increased in the sward around dung but not around molehills, proliferated when fertilizer was added to General discussion / 142 artificial swards, and responsed positively when T. repens was removed. These observations suggest that D. glomerata was able to take advantage of nutrient release, either by pre-empting nitrates released by decaying nodules or by capturing nutrients leached from dung. D. glomerata had, however, little ability to invade molehills and dung pats. The response of Lolium perenne was very variable. In general, it did not invade small disturbances. In a year of declining abundance, L. perenne declined less in areas from which Holcus lanatus had been removed. Removal of H. lanatus led to increases in several species although H. lanatus did not increase with removal of other species nor could it readily invade molehills and dung. These findings agree with those of Thorhallsdottir (1984) who found H. lanatus effectively held space against invasion by other species but showed little tendency to pre-empt gaps in the sward. The response pattern of individual species to molehills and dung pats recorded at 2 2 0.25 m scale of observation was not evident at 25 m . The direction of association of species and disturbance, in general, was not predictable from one scale of observation to the other. There were exceptions: Agropyron repens readily invaded small disturbances and its abundance increased in areas of high disturbance .and Holcus lanatus, unable to exploit disturbances, was not abundant in disturbed areas. For other species, invasion pattern was not a factor significantly influencing distribution of abundance within the pasture at anything beyond the local scale. From these discrepancies, I must conclude that despite significant interaction between pattern of species abundance and disturbance, General discussion / 143 disturbances per se have little organizing or causal force. The most obvious explanation of the pattern within the pastures is that because most significant interactions were associated with molehills, plant species were tracking resources linked to those that moles were also exploiting. Because moles predominantly consume earthworms, species positively associated with increased density of molehills may be those exploiting areas of increased soil fertility. Aarssen (1983) and Aarssen and Turkington (1985a) have proposed a model of community processes derived from observations and experiments on species from these pastures. They suggested that the community evolved over time as less fit genotypes were eliminated by competition and grazing. The resultant community is characterized by strongly competitive individuals that co-exist because of a competitive stalemate in which none can achieve an advantage. Recruits are filtered through interactions with existing individiuals so that only those of equivalent abilities in that time and location persist. Evidence to support or refute this model was not collected in this study in the same pastures. Yet, a discussion of some of the ideas seems appropriate. In order to have selection, there must be genetic variation, which is generated by mutation and/or recombination. In higher plant populations, genetic variation is usually achieved by sexually produced seeds. This study of the Aldergrove pastures provided little information on rates of recruitment for most species but because of the focus on small disturbance, some idea of the 'life history' of Trifolium repens has emerged. Seed germination is apparently highly episodic. 2 Probably it is restricted to gaps of 500cm or greater; none were observed in General discussion / 144 any of the smaller gaps. Seedling density varied and mortality was generally high. However, early mortality resulted from, in the main, such density-independent factors as grazing by slugs and small mammals, burial by re-opening of mole tunnels, or up-rooting by haymaking. Initial interactions with other individual plants were most certainly intraspecific but occurred after the plant became established on the disturbance. Without specific demographic studies following the fates of these survivors as other species spread into the gaps, it is impossible to refute Aarssen's hypothesis of competitive combining ability. I will suggest, however, the following scheme: once established in the relatively competition-free environment of the molehill, T. repens begins to spread laterally into the sward. In areas where conditions are favourable, internodes are short and individual plants concentrate biomass in these areas but where condition are unfavourable, internodes are long and individuals do not establish abundantly (Solangaarachchi 1985). What constitutes a 'favourable' site? Axiomatically, it is adequate resource levels. This can result from numerous factors; (1) resources may not be limiting, i.e. abundant beyond the capability of the plant to deplete in the time frame available; (2) resources may be available because of partitioning by niche differentiation (ecological combining ability); (3) resources may be available because ability to exploit is equivalent (competitive combining ability). The relative importance of these factors was not tested but the role of selection appears to be much less than previously postulated: recruits are rare, random events predominate, yet the pastures remain stable.. If pastures are characterized by random event, it does not follow that they are also characterized by random processes. Events may trigger an orderly series of General discussion / 145 processes. The replacement of one species by another is not random. The complex, interweaving of scale and pattern observed in the pasture suggests a combination of process and forces whose interaction await further study. 7. LITERATURE CITED Aarssen, L.W. 1983. Interactions and coexistence of species in pasture community evolution. Ph.D. thesis, University of British Columbia. Aarssen, L.W. and R. Turkington. 1985(a). Vegetation dynamics and neighbour associations in pasture community evolution. J. Ecol. 73:585-603. Aarssen, L.W. and R. Turkington. 1985(b). Competitive relations among species from pastures of different ages. Can. J. Bot. 63:2319-2325. Allen, E.B. and R.T.T. Forman. 1976. Plant species removals and old-field community structure and stability. J. Ecol. 57:1233-1243. Allen, E.B. and D.H. Knight. 1984. The effects of introduced annuals on secondary succession in sagebrush - grassland, Wyoming. Southwestern Naturalist 29:407-422. Baker, H. 1937. Alluvial meadows: a comparative study of grazed and mown meadows. J. Ecol. 25:408-420. Bazzaz, F.A. 1983. Characteristics of populations in relation to disturbance in natural and man-modified ecosystems. In Disturbance and Ecosystems, H.A. Mooney & M. Godron (eds.), pp. 259-277. Springer-Verlag, Berlin. Beddows, A.R. 1961. Biological flora of the British Uses. Holcus lanatus L. J. Ecol. 49:421-430. Begon, M., J.L. Harper and C.R. Townsend. 1986. Ecology: Individuals, Populations and Communities. Sinauer Assoc., Inc. Sunderland, Mass. Belsky, A.J. 1986. Revegetation of artificial disturbances in grasslands of the Serengeti National Park, Tanzania. I. Colonization of grazed and ungrazed plots. J. Ecol. 74:419-437. Bender, E.A., T.J. Case and M.E. Gilpin. 1984. Perturbation experiments in community ecology: theory and practice. Ecology 65:1-13. 146 Literature cited / 147 Berry, J . and O. Bjorkman. 1980. Photosynthetic response and adaptation to temperature in higher plants. Ann. Rev. Plant Physiol. 31:491-543. Bland, B.F. 1967. The effect of cutting frequency and root segregation on the yield from perennial ryegrass - white clover associations. Agric. Sci. Camb. 69:391-397. Blaser, R.E. and N.C. Brady. 1950. Nutrient competition in plant associations. Agron. J. 42:128-135. Bliss, C.I. and Fisher, I. 1953. Fitting the negative binomial distribution to biological data. Biometrics 9:176-200. Bradshaw, A.D. 1965. Evolutionary significance of phenotypic plasticity in plants. Adv. Genet. 13:115-155. Brokaw, N.V.L. 1985. Treefalls, regrowth, and community structure in tropical forests. In The Ecology of Natural Disturbances and Patch Dynamics. S.T.A. Pickett and P.S. White (eds.) pp. 53-69. Academic Press, Inc. Orlando. Cahn, M.G. and J.L. Harper. 1976. The biology of the leaf mark polymorphism in Trifolium repens L. Distribution of phenotypes at a local scale. Heredity 37:309-325. Canada Soil Survey Committee. 1978. The Canadian system of soil classification. Can. Dept. Agric. Publ. 1646. Supply and Services, Ottawa. Canham, C D . and P.L. Marks 1985. The response of woody plants to disturbance: patterns of establishment and growth. In The Ecology of Natural Disturbance and Patch Dynamics. S.T.A. Pickett and P.S. White (eds.) pp 197-216. Academic Press, Inc. Orlando. Caswell, H. 1978. Predator-mediated coexistence, a nonequilibrium model. Amer. Nat. 112:127-154. Literature cited / 148 Christensen, N.L. 1985. Shrubland fire regimes and their evolutionary consequences. In The Ecology of Natural Disturbance and Patch Dynamics. S.T.A. Pickett and P.S. White (eds.) pp. 86-100. Academic Press, Inc. Orlando. Clements, F.E. 1916. Plant Succession. An Analysis of the Development of Vegetation. Carnegie Inst., Washington. Connell, J.H. 1978. Diversity in tropical rain forests and coral reefs. Science 199:1302-1310. Darwin, C. 1859. The Origin of Species. Penguin Books, New York, 1968 edition. Davies, H. 1966. Molehills and pasture reversions. J. British Grassld. Soc. 21:148-149. Davis, R.M. and J.E. Cantlon. 1969. Effect of size area open to colonization on species composition in early old-field succession. Bull. Torrey Bot. Club 96:660-673. Dayton, P.K. 1971. Competition, disturbance and community organization: the provision and subsequent utilization of space in a rocky intertidal community. Ecol. Monogr. 41:351-389. Dennis, W.D. and J . Woledge. 1985. The effect of nitrogenous fertilizer on the photosynthesis and growth of white clover/perennial ryegrass swards. Ann. Bot. 55:171-178. Denslow, J.S. 1980. Gap partitioning among tropical rainforest trees. Biotropica 12; Suppl., 47-55. Denslow, J.S. 1985. Disturbance-mediated coexistence of species. In The Ecology of Natural Disturbances and Patch Dynamics. S.T.A. Pickett and P.S. White (eds.) pp.307-323. Academic press, Inc., Orlando. Dixon, W.J. 1983. BMDP Statistical Software. University of California Press, Berkeley. Literature cited / 149 Donald, C M . 1963. Competition among crop and pasture plants. Adv. Agron. 15:1-118. During, C. and K.J. McNaught. 1961. Effects of cow urine on growth of pasture species and uptake of nutrients. N.Z.J. Agric. Res. 4:591-605. Egler, F.E. 1954. Vegetation science concepts 1. Initial floristic composition, a factor in old-field vegetation development. Vegetatio 4:412-417. England, F. 1968. Competition in mixtures of herbage grasses. J. Appl. Ecol. 5:227-242. Fenner, M. 1978. A comparison of the abilities of colonizers and closed-turf species to establish from seed in artificial swards. J. Ecol. 66:953-963. Fienberg, S.E. 1970. The analysis of multidimensional contingency tables. Ecology 51:419-433. Fienberg, S.E. 1980. The Analysis of Cross-classified Categorical Data. 2nd ed. MIT Press, Cambridge, Mass. Fowler, N. 1981. Competition and coexistence in a north Carolina grassland. II. Effects of the experimental removal of species. J . Ecol. 69:843-854. Fox, D.J. and K.E. Guire. 1976. Documentation for MIDAS. Statistical Research Laboratory, University of Michigan, Ann Arbor. Garwood, E.A. and K.C. Tyson. 1979. Productivity and botanical composition of a grazed ryegrass/white clover sward over 24 years as affected by soil conditions and weather. In Changes in Sward Composition and Productivity. A.H. Charles and R.J. Haggar (eds.), Brit. Grassld. Soc. Hurley, pp. 41-46. Gervais, P. 1960. Effects of cutting treatments on Ladino clover grown alone and in mixtures with grasses. I. Productivity and botanical composition of forage. Can. J . Plant Sci. 40:317-327. Literature cited / 150 Gleason, H.A. 1926. The individualistic concept of the plant association. Bull. Torrey Bot. Club 53:7-26. Glendenning, R. 1957. Biology and control of the coast mole Scapanus ovarius True, in British Columbia. Can. J . Anim. Sci. 39:34-44. Goldberg, D.E. and P.A. Werner. 1983. The effects of size of opening in vegetation and litter cover on seedling establishment of goldenrod {Solidago spp.). Oecologia (Berlin) 60:149-155. Grime, J.P. 1973. Control of species density in herbaceous vegetation. J . Envir. Man. 1:151-167. Grime, J.P. 1977. Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. Amer. Nat. 777:1169-1194. Grime, J.P. 1979. Plant Strategies and Vegetation Processes. John Wiley and Sons, Chichester. Gross, K.L. 1980. Colonization by Verbascum thapsus (Mullein) of an old-field in Michigan: experiments on the effects of vegetation. J . Ecol. 68:919-927. Gross, K.L. 1984. Effects of seed size and growth form on seedling establishment of six monocarpic perennial plants. J . Ecol. 72:369-387. Grubb, P.J. 1977. The maintenance of species-richness in plant communities: the importance of the regeneration niche. Biol. Rev. 52:107-145. Grubb, P.J., D. Kelly and J. Mitchley. 1982. The control of relative abundance in communities of herbaceous plants. In The Plant Community as a Working Mechanism. E.I. Newman (ed.) pp. 79-97. Special Publication No. 1 of British Ecological Society. Blackwell Scientific Publications, Oxford. Harper, J.L. 1964. The nature and consequence of interference amongst plants. In Genetics today, XI International Congress of Genetics 2:465-482. Literature cited / 151 Harper, J.L. 1967. A Darwinian approach to plant ecology. J . Ecol. 55:247-270. Harper, J .L. 1969. The role of predation in vegetational diversity. In Diversity and Stability in Ecological Systems, Brookhaven Symp. in Biol. 22, 48-62. Harper, J.L. 1977. Population Biology of Plants. Academic Press Inc., London. Harris, W. and V.J. Thomas. 1972. Competition among pasture plants. II. Effects of frequency and height of cutting on competition between Agrostis tenuis and two ryegrass cultivars. N.Z. J. Agric. Res. 15:19-32. Henry, J.D. and J.M.A. Swan. 1974. Reconstructing forest history from live and dead plant material - an approach to the study of forest succession in southwest New Hampshire. Ecology 55:772-783. Herriott, J.B.D. and D. Wells. 1960. Clover nitrogen and sward productivity. J . Br. Grassld. Soc. 15:63-69. Hill, M.O. 1973. Reciprocal averaging: an eigenvector method of ordination. J . Ecol. 61:237-249. Hils, M.H. and J.L. Vankat. 1982. Species removals from a first-year old-field plant community. Ecology 63:705-711. Hilton, J.R., R.J. Frond-Williams and J. Dixon. 1984. A relationship between phytochrome photoequilibrium and germination of seeds of Poa trivialis L. from contrasting habitats. New Phytol. 97:375-379. Holling, C S . 1973. Resilience and stability of ecological systems. Ann. Rev. Ecol. Syst. 4:1-23. Huffaker, C.B. and C.E. Kennet. 1959. A ten-year study of vegetational changes associated with biological control of Klamath Weed. J. Range Mgmt. 12:69-82. Huston, M. 1979. A general model of species diversity. Amer. Nat. 113:81-101. Literature cited / 152 Hutchinson, C.S. 1979. Some effects of cow dung on pasture composition. M.Sc. thesis, University College of North Wales, Bangor, Wales. Jain, S.K. 1983. Genetic characteristics of populations In Disturbance and Ecosystems. H.A. Mooney and M. Gordon (eds.) pp. 240-258. Springer-Verlag, Berlin. Jalloq, M.C. 1975. The invasion of molehills by weeds as a possible factor in the degeneration of reseeded pasture. I. The buried viable seed population of molehills from four reseeded pastures in West Wales. J . Appl. Ecol. 12:643-657. Jayasingam, T. 1985. Comparative demography of four Composites. Ph.D. thesis, University College of North Wales, Bangor, Wales. Jones, M.G. 1933a. Grassland management and its influence on the sward. I. Factors influencing the growth of pasture plants. Emp. J . exp. Agric. 1:43-47. Jones, M.G. 1933b. Grassland management and its influence on the sward. II. The management of a clovery sward and its effects. Emp. J . exp. Agric. 1:122-128. Jones, M.G. 1933c. Grassland management and its influence on the sward. III. The management of a 'grassy' sward and its effects. Emp. J . exp. Agric. 1:223-234. Jones, M.G. 1933d. Grassland management and its influence on the sward. IV. The management of poor pastures. V. Edaphic and biotic influences on pastures. Emp. J. exp. Agric. 1:361-367. de Jong, P. and M. Greig. 1984. First-order Markov chains with a zero diagonal transition matrix. Biometrics 40:101-107. de Jong, P. and M. Greig. 1985. Models and methods for pairing data. Can. J. Stat. 13:233-241. de Jong, P., M. Greig and D. Madan. 1983. Testing for random pairing. J . Amer. Stat. Assoc. 78:332-336. Literature cited / 153 King, J. 1971. Competition between established and newly sown grass species. J. Gr. Grassld. Soc. 26:221-229. Kleter, H.J. 1968. Influence of weather and nitrogen fertilization on white clover percentage of permanent grassland. Neth. J . Agric. Sci. 16:43-52. Krajina, V.J. 1965. Biogeoclimatic zones and classification of British Columbia. Ecol. Western N. America 1:1-17. Levin, S.A. and R.T. Paine. 1974. Disturbance, patch formation and community structure. Proc. Nat. Acad. Sci. (USA) 71:2744-2747. Loucks, O.L., M.L. Plumb-Mentjes and D. Rogers. 1985. Gap processes and large-scale disturbances in sand prairies. In The Ecology of Natural Disturbance and Patch Dynamics. S.T.A. Pickett and P.S. White (eds.) pp. 72-83. Academic Press, Inc. Orlando. MacDiarmid, B.N. and B.R. Watkin. 1971. The cattle dung patch. I. Effect of dung patches on yield and botanical composition of surrounding and underlying pature. J . Br. Grassld. Soc. 26:239-245. MacLusky, D.S. 1960. Some estimates of the areas of pasture fouled by the excretion of dairy cows. J. Brit. Grassld. Soc. 15:181-188. Mahmoud, A. and J.P. Grime. 1976. An analysis of competitive ability in three perennial grasses. New Phytol. 77:431-435. Marsh, R. and R.C. Campling. 1970. Fouling of patures by dung. Herbage Abstracts 40:123-130. McNeilly, T. and M.L. Roose. 1984. The distribution of perennial ryegrass genotypes in swards. New Phytol. 98:503-514. Miles, J . 1974. Effects of experimental interference with stand structure on establishment of seedlings in Callenetum. J . Ecol. 64:745-755. Milton, W.E.J. 1940. The effect of manuring, grazing and cutting on the yield, botanical and chemical composition of natural hill pastures. J. Ecol. 28:326-356. Literature cited / 154 Mundy, E.J. 1961. The effect of urine and its components on the botanical composition and production of a grass/clover sward. J. Brit. Grassld. Soc. 16:100-105. Newbery, D. McC. 1979. The effects of decomposing roots on the growth of grassland plants. J . Appl. Ecol. 16:613-622. Norman, M.J.T. and J.O. Green. 1958. The local influence of cattle dung and urine upon the yield and botanical composition of permanent pasture. J . Br. Grassld. Soc. 13:39-45. Oliver, C D . and E.P. Stephens. 1977. Reconstruction of a mixed-species forest in central New England. Ecology 58:562-572. Orloci, L. 1978. Multivariate Analysis in Vegetation Research, 2nd ed. Junk, The Hague. Peet, R.K., D.C. Glenn-Lewin and J . Walker Wolf. 1983. Predictions of man's impact on plant species diversity. A challenge for vegetation science. In Man's Impact on Vegetation. W. Holzner, M.J.A. Werger and I. Ikusima (eds.) Junk, The Hague. Petersen, R.G., H.L. Lucas and W.W. Woodhouse, Jr. 1956. The distribution of excreta by freely grazing cattle and its effect on pature fertility: I. Excretal distribution. Agron. J. 48:440-444. Pickett, S.T.A. and P.S. White. 1985. The Ecology of Natural Disturbance and Patch Dynamics. Academic Press, Inc. Orlando Pinder, J.E. III. 1975. Effects of species removal on an old-field plant community. Ecology 56:747-751. Piatt, W.J. 1975. The colonization and formation of equilibrium plant species associations on badger disturbances in a tall-grass prairie. Ecol. Monogr. 45:285-305. Polley, H.W. and S.L. Collins. 1984. Relationship of vegetation and environment in buffalo wallows. Amer. Midi. Nat. 112:178-186. Literature cited / 155 Putwain, P.D. and J.L. Harper. 1970. Studies in the dynamics of plant populations. III. Influence of associated species on populations of Rumex acetosa L. and R. acetosella L. in grassland. J. Ecol. 58:251-264. Rabotnov, T.A. 1966. Pecularities of the structure of polydominant meadow communities. Vegetatio 13:109-116. Rabotnov, T.A. 1974. Differences between fluctuations and successions. In Handbook of Vegetation Science Part 8. Vegetation dynamics. R. Knapp (ed.) pp. 19-24. Junk, The Hague. Rapp, J.K. and D. Rabinowitz. 1985. Colonization and establishment of Missouri prairie plants on artificial soil disturbances. I. Dynamics of forb and graminoid seedlings and shoots. Amer. J . Bot. 72:1618-1628. Rhodes, I. and W.R. Stern. 1978. Competition for light. In 'Plant Relations in Pastures'. J.R. Wilson (ed.), pp. 175-189. CSIRO, Melbourne. Ross, M.A. and J.L. Harper. 1972. Occupation of biological space during seedling establishment. J . Ecol. 60:77-88. Rothwell, P.I. 1977. The plant ecology of molehills. M.Sc. thesis, University College of North Wales. Bangor, Wales. Runkle, J.R. 1981. Gap regeneration in some old-growth forests of the eastern United States. Ecology 62:1041-1051. Runkle, J.R. 1982. Patterns of disturbance in some old growth mesic forests of eastern North America. Ecology 63:1533-1546. Sagar, G.R. 1970. Factors controlling the size of plant populations. Proc. X. Brit. Weed Control Conference pp. 965-979. Sagar, G.R. and J.L. Harper. 1961. Controlled interference with natural populations of Plantago lanceolata, P. major and P. media. Weed Res. 1:163-176. Literature cited / 156 Sarukhan, J. and J.L. Harper. 1973. Studies on plant demography: Ranunculus repens L. , R. bulbosus L., and R. acris L. I. Population flux and survivorship. J. Ecol. 61:675-716. Silander, J.A. and J. Antonovics. 1982. Analysis of interspecific interactions in a coastal plant community - a perturbation approach. Nature 298:557-560. Silvertown, J.W. 1981. Micro-spatial heterogeneity and seedling demography in species-rich grassland. New Phytol. 88:117-128. Snaydon, R.W. 1971. An analysis of competition between plants of Trifolium repens L. populations collected from contrasting soils. J. Appl. Ecol. 8:687-697. Snaydon, R.W. 1978. Genetic change in pasture populations. In Plant Relations in Pastures. J.R. Wilson (ed.) pp. 253-269. CSIRO, Melbourne. Snaydon, R.W. 1984. Plant demography in an agricultural context. In Perspectives on Plant Population Ecology. R. Dirzo and J. Sarukhan (eds.). pp. 389-407, Sinauer Assoc. Inc. Sunderland, Mass. Snaydon, R.W. 1985. Aspects of the ecological genetics of pasture species. In Structure and Functioning of Plant Populations 2. Phenotypic and Genotypic Variation in Plant Populations. J. Haech and J.W. Woldendorp (eds.). pp. 131-152. North-Holland Pub. Co., Amsterdam. Snaydon, R.W. and C D . Howe 1986. Root and shoot competition between established ryegrass and invading grass seedlings. J . Appl. Ecol. 23:667-674. Soane, I.D. and A.R. Watkinson. 1979. Clonal variation in populations of Ranunculus repens. New Phytol. 82:557-573. Solangaarachchi, S.M. 1985. The nature and control of branching pattern in white clover (Trifolium repens L.) Ph.D. thesis, University College of North Wales, Bangor, Wales. Literature cited / 157 Sousa, W.P. 1979. Disturbance in marine intertidal boulder fields: the non-equilibrium maintenance of species diversity. Ecology 60:1225-1239. Sousa, W.P. 1984a. The role of disturbance in natural communities. Ann. Rev. Ecol. Syst. 15:353-391. Sousa, W.P. 1984b. Intertidal mosaics: patch size, propagule availability, and spatially variable patterns of succession. Ecology 65:1918-1935. Spedding, C.R.W. 1971. Grassland Ecology. Clarendon Press, Oxford, pp. 221. Spittlehouse, D. 1985. Components of the seedling environment. In Interior Spruce Seedling Performance: State of the Art. Northern Silviculture Committee, Prince George, B.C. SPSS.* (1986) SPSS.* User's Guide, 2nd. edition. SPSS Inc., Chicago. Stern, W.R. and C.H. Donald 1962a. Light relationships in grass-clover swards. Aust. J . Agr. Res. 13:599-614. Stern, W.R. and C.H. Donald 1962b. The influence of leaf area and radiation on the growth of clover in swards. Aust. J. Agr. Res. 13:615-623. Strong, D.R. Jr. 1983. Natural variability and the manifold mechanisms of ecological communities. Amer. Nat. 122:636-660. Tande, G.F. 1979. Fire history and vegetation pattern of coniferous forests in Jasper National Park, Alberta. Can. J . Bot. 57:1912-1931. Tansley, A.G. 1935. The use and abuse of vegetational concepts and terms. Ecology 16:284-307. Tansley, A.G. 1939. The British Islands and Their Vegetation. Cambridge University Press, Cambridge. Literature cited / 158 Tansley, A.G. and R.S. Adamson. 1925. Studies of the vegetation of the English Chalk. 111. The chalk grasslands of the Hampshire-Sussex Border. J . Ecol. 13:177-223. Thompson, J.N. 1985. Within patch dynamics of life histories, populations, and interactions: selection over time in small spaces. In The Ecology of Natural Disturbance and Patch Dynamics. S.T.A. Pickett and P.S. White (eds.) pp. 253-264. Academic Press, Inc. Orlando. Thorhallsdottir, T.E. 1984. The dynamics of a grassland community with special reference to five grasses and white clover. Ph.D. thesis, University College of North Wales, Bangor, Wales. Turkington, R. and L.W. Aarssen. 1984. Local scale differentiation as a result of competition. In Perspectives in Plant Population Ecology. J . Sarukhan and R. Dirzo (eds.) pp. 107-127. Sinauer Assoc., Cambridge, Mass. Turkington, R., M.A. Cahn, A. Vardy and J.L. Harper. 1979. The growth, distribution and neighbour relarionships of Trifolium repens in a permanent pasture. Ill The establishment and growth of Trifolium repens in natural and perturbed sites. J. Ecol. 67:231-243. Veblen, T.T. 1985. Stand dynamics in Chilean Nothofagus forests. In The Ecology of Natural Disturbance and Patch Dynamics. S.T.A. Pickett and P.S. White (eds.) pp. 35-51. Academic Press, Inc., Orlando. Veresoglou, D.S. 1983. Partitioning of nutrients between co-existing grassland plant species. D. Phil, thesis. University of York. Vickery, P.J. 1981. Pasture growth under grazing. In Grazing Animals, F.H.W. Morley (ed.). pp. 55-77. Elsevier, Amsterdam. Vitousek, P.M. 1985. Community turnover and ecosystem nutrient dynamics. In The Ecology of Natural Disturbance and Patch Dynamics. S.T.A. Pickett and P.S. White (eds.) pp.325-333. Academic Press, Inc. Orlando. Watkin, B.R. and R.J. Clements. 1978. The effects of grazing animals on patures. In Plant Relations in Pastures. J.R. Wilson (ed.) pp. 273-289. CSIRO, Melbourne. Literature cited / 159 Watt, A.S. 1925. On the ecology of the British beechwoods with special reference to their regeneration. Part II, Section II and III. The development and structure of beech communities on the Sussex Downs. J. Ecol. 13:27-73. Watt, A.S. 1957. The effect of excluding rabbits from grassland B (Mesobrometum) in Breckland. J. Ecol. 45:861-878. Watt, T.A. and R.J. Haggar. 1980. The effect of height of water table on the growth of Holcus lanatus with reference to Lolium perenne. J . Appl. Ecol. 17:423-430. Weeda, W.C. 1967. The effects of cattle dung patches on pasture growth, botanical composition, and pasture utilization. N.Z. J. Agric. Res. 10:150-157. Welch, D. and M. Rawes. 1964. The early effects of excluding sheep from high-level grasslands in the north Penniones. J. Appl. Ecol. 1:281-300. White, J. 1985. The census of plants in vegetation. In The Population Structure of Vegetation. J. White (ed.). Junk, The Hague. White, P.S. and S.T.A. Pickett. 1985. Natural disturbance and patch dynamics: an introduction. In The Ecology of Natural Disturbance and Patch Dynamics. S.T.A. Pickett and P.S. White (eds.) pp. 3-13. Academic Press, Inc. Orlando. Whyte, O.R. 1974. Tropical Grazing Lands: Communities and Constituent Species. Junk, The Hague. 8. APPENDIX 1 160 T r a n s i t i o n p r o b a b i l i t y o f row s p e c i e s b e i n g r e p l a c e d b y a c o l u m n s p e c i e s , o n a n d a r o u n d m o l e h i l l s i n t h e 1 9 3 9 p a s t u r e f o r t h e p e r i o d f r o m S e p t e m b e r 1984 t o A p r i l 1 9 8 5 . M o l e h i l l Dactylis Holcus Lolium Phleum PoaAgropyronTrifolium Taraxacum Ranunculus glomerat a /anatus perenne pratensecompress a r e p e n s r e p e n s officinale acr i s M o l e h i 1 1 33 .9 3 .5 0 . .4 8 . 1 0 . .4 18 .4 2 . 8 18. .4 3 . .9 1 .4 Dactylis glomerata 0 . ,0 91 .7 0 , .0 0 . 0 0 , ,0 5. . 0 1 , .7 1 . . 7 0 . . 0 0 . 0 Holcus lanatus 0 . . 0 8 .5 4 2 , 6 12 .8 0 , ,0 19. . 1 0 . .0 10. .6 2. . 1 0 . 0 Lolium perenne 4 . 1 2 .9 1 .2 43 .9 0 , .6 17 .5 0 . .6 19 .9 2. .9 2: 3 Phleum pratense 0, . 0 10 .0 10 .0 10 .0 0 . .0 0 . 0 0 . .0 4 0 .0 10. . 0 0 . 0 Poa compressa 1 . .2 2 .4 2, .4 6 . 0 0 . ,0 59 . 5 1 . 2 16. .7 3 . .6 3 . 6 Agropyron repens 10, O 10 .0 0 , ,0 10 .O 0 , 0 10 .O 3 0 . .0 10. .0 0 . . 0 0 . 0 Trifolium repens 4 , . 5 4. .9 0 , .4 6 .9 0 , 8 15. .8 2. .4 5 7 . . 1 1 . .6 1 . 2 Taraxacum officinale 3. ,4 0 , .0 3, .4 6. .9 0 . ,0 13. 8 6. .0 3. .4 62 . 1 6 . 9 Ranunculus acris 0 , .0 0 , ,0 0 . .0 0 . .0 0 . 0 0 . .0 14 . 3 14. .3 0 . 0 71 .4 T r a n s i t i o n p r o b a b i l i t y o f row s p e c i e s b e i n g r e p l a c e d b y a c o l u m n s p e c i e s , p a s t u r e f o r t h e p e r i o d f r o m A u g u s t 1983 t o A p r i l 1 9 8 4 . o n a n d a r o u n d m o l e h i l l s i n t h e 1958 M o l e h i l l Dactylis Holcus Lolium Phleum PoaAgropyronT ri fo! ium glomerata /anatus perenne pratensecompressa repens repens officinale M o l e h i 1 1 37 . .8 2, .0 6 .4 1 . . 2 0 , .0 12 .9 2 0 . .5 7 . 2 4 . 0 Dactylis glomerata 5. , 2 56 , .5 1 , . 7 3 ,5 5. . 2 15. . 7 6 . , 1 2 .6 0 . 9 Holcus lanatus 8 . , 1 2 , .3 2 5 , .6 4 . , 7 5. .8 19 . 8 10. . 5 5. .8 9 . 3 Lolium perenne 10 .8 10, 8 2 . 0 9. .8 8 . 8 2 2 , .5 12. .7 8 .8 2 . 9 Phleum pratense 12 .5 12 . .5 0 . 0 0 , .0 5 0 . .0 12 , , 5 0 , . 0 0 . . 0 6 . 3 Poa compressa 14 .4 4, .2 3 .4 4, .2 5. . 1 3 3 , .9 12, .7 7 .6 7 . 6 Agropyron repens 10. 6 5 . .7 3, .3 3, , 3 3. . 3 23 . .6 38 . . 2 1 . .6 2 . 4 Trifolium repens 0 . .0 7 . . 1 10, .4 3, ,6 0 . .0 10. . 7 0 . .0 4 6 , .4 7 . 1 Taraxacum officinale 13, , 3 2. ,2 4. ,4 4 , ,4 2. .2 15. .6 2 2 . .2 6. .7 2 6 . 7 Ranunculus acris 0 , ,0 0 . .0 0 . .0 5 0 , ,0 0 . 0 0 . .0 2 5 . .0 0 , .0 25 . 0 Taraxacum Ranunculus acr i s O . O 0 . 0 1 . 2 4 . 9 0 . 0 0 . 0 1 .6 0 . 0 0 . 0 O . O CT) ro T r a n s i t i o n p r o b a b i l i t y o f row s p e c i e s b e i n g r e p l a c e d b y a c o l u m n s p e c i e s , o n a n d a r o u n d m o l e h i l l s i n t h e 1977 p a s t u r e f o r t h e p e r i o d f r o m A u g u s t 1984 t o A p r i l 1 9 8 5 . M o l e h i 1 1 Dactylis Ho 1cus Lolium Ph1eum PoaAgropyronT rifolium Taraxacum Ranuncu1 us g1omerata 1anatus perenne pratensecompress a r e p e n s repens officinale acr i s M o l e h i 1 1 27 .5 4 . 1 1 1 . .9 6. .9 5 . 5 1 1 .9 2. .3 15 . 1 5. . 0 3 . 3 Dactylis glomerata 0 . 0 44 .7 7 . .9 5. .3 13 . 2 13 .2 0 . 0 5 . 3 5, . 3 0 . 0 Holcus lanatus 0 . 0 4, .4 77 . .8 4 . .4 0 . 0 4 . 4 0 . . 0 6. .7 0 , . 0 0 . 0 Lolium perenne 3 .4 1 .7 19. .0 41 . .4 3. .4 13 .8 0 . .0 6. .9 1 . . 7 0 . 0 Phleum pratense 0 . 0 14 .3 0 . .0 57 . 1 28 .6 0 . 0 0 . . 0 0 . 0 0 . . 0 0 . 0 Poa compressa 5 . 0 10 .0 10. .0 5, .0 2, . 5 4 0 . 0 2 .5 7. . 5 7. .5 0 . 0 Agropyron repens 2 0 . 0 0 . 0 0 . .0 2 0 . .0 0 . 0 2 0 .0 0 . . 0 4 0 . . 0 0 . . 0 0 . 0 Trifolium repens 0 . 0 3 .9 9 . .0 1 . .3 2 .6 5. . 1 1 . .3 6 9 . .2 1 . , 3 1 . 3 Taraxacum officinale 14 .3 4. .8 0 . .0 4. 8 9 . .5 19 .0 0 . .0 4 . .8 4 2 . .9 0 . 0 Ranunculus acris 0 .O 0 , .0 O. 0 0 . 0 0 . . 0 0 . ,0 0 . O 0 . ,0 0 . O 1 0 0 . 0 CO T r a n s i t i o n p r o b a b i l i t y o f row s p e c i e s b e i n g r e p l a c e d b y a c o l u m n s p e c i e s , o n a n d a r o u n d d u n g i n t h e 1 9 3 9 p a s t u r e f o r t h e p e r i o d f r o m A u g u s t 1983 t o A p r i l 1 9 8 4 . Dactyl is Holcus Lolium Phleum Poa Agropyron!rifolium Taraxacum Ranunculus D u n g g1omerata /anatus perenne pratense c o m p r e s s a repens repens of f i c i nale acr i s Dactylis glomerata 35 .0 0 . .0 15 .0 0 .0 25 . 0 0 . 0 10 .0 5. .0 10. . 0 0 . . 0 Holcus lanatus 2 . 1 12. .8 14. .9 2 . 1 36 . 2 0 . 0 21 , . 3 6, .4 4 , .3 0 . , 0 Lolium perenne 1 .7 0 . .8 4 0 , ,9 0 , .0 34 .6 0 . .4 10, . 1 8 . .4 1 . .3 0 . O Phleum pratense 0 .0 0 . .0 0 . .0 100, .0 0 . 0 0 . . 0 0 , .0 0 , ,0 0 , .0 0 . O Poa compressa 2 .0 1 . .0 3 0 . .6 0 , .0 3 5 , .7 2, . 0 14 , , 3 4 . . 1 4 . , 1 0 . ,0 Agropyron repens 0 . .0 0 . .0 35 .3 0 . ,0 2 9 .4 11 , 8 5. .9 1 1 . 8 5. .9 0 . 0 Trifolium repens 1 .4 0 . .0 2 3 , .2 0 , .0 33 .3 1 .4 3 0 , .4 4 . .3 2, .9 0 . . 0 Taraxacum officinale 0 .O 0 . ,0 2 2 . . 2 0 , .0 25 . 9 0 , . 0 3 . .7 44 , .4 O. O 0 . O Ranunculus acris 16. .7 0 . .0 5 0 , .0 0 , ,0 33 .3 0 . . 0 0 , ,0 0 . .0 0 . .0 0 . ,0 D u n g 0 . . 7 0 . 0 2 6 . .9 0 . ,0 34 , . 3 3 . ,0 17 . .2 6 . .7 1 . ,5 8 . 2 T r a n s i t i o n p r o b a b i l i t y o f r o w s p e c i e s b e i n g r e p l a c e d b y a c o l u m n s p e c i e s , o n a n d a r o u n d d u n g i n t h e 1958 p a s t u r e f o r t h e p e r i o d f r o m S e p t e m b e r 1983 t o A p r i l 1 9 8 4 . Dactylis Holcus Lolium Phleum Poa AgropyronTrifolium Taraxacum Ranunculus D u n g g1omerata /anatus perenne pratense compressa r e p e n s r e p e n s officinale acr i s Dactylis glomerata 7 5 , ,8 0 . .0 1 . 6 0 , .0 12. ,9 1 . .6 3. ,2 1 , .6 0 . ,0 0 , . 0 Holcus lanatus 1 , .4 3 2 . .9 1 . .4 1 . .4 43 , 8 4 . . 1 8. .2 1 . .4 0 . 0 0 . . 0 Lol ium perenne 4 .9 2 . ,4 41 .5 2. .4 3 4 . . 1 7. .3 4, ,9 0 . 0 0 . . 0 0 , . 0 Phleum pratense 0 .0 0 . .0 0 . 0 8 8 . 9 0 . .O O. . 0 11 . . 1 0 . . 0 O. .0 0 . 0 Poa compressa 3, . 1 0 . ,0 5. .4 4 , .7 68 . .2 10. . 1 3. . 1 2, .3 0 . .0 0 . . 0 Agropyron repens 5, .9 2. .9 5. ,9 8. .8 14 . , 7 52 . ,9 0 . .0 0 . .0 0 . ,0 0 . . 0 Trifolium repens 0 , .0 O. .0 5. . 3 0 . .0 21 . , 1 5 . . 3 6 3 . .2 O. ,0 0 . O 0 . . 0 Taraxacum officinale 0 .0 0 . .0 9. . 1 0 , ,0 0 . . 0 0 , .0 9 . . 1 81 .8 0 . .0 0 . . 0 Ranunculus acris O, .0 0 . .0 0 . .0 0 , .0 0 . .0 0 . .0 2 0 . .0 0 . . 0 6 0 . .0 0 . . 0 D u n g 8. .3 2. 8 2. 8 2. . 1 26 . .2 1 1 . ,7 4. . 1 6. .2 1 . 4 3 2 . .4 CTl CJ1 T r a n s i t i o n p r o b a b i l i t y o f r o w s p e c i e s b e i n g r e p l a c e d b y a c o l u m n s p e c i e s . o n a n d a r o u n d d u n g i n t h e 1977 p a s t u r e f o r t h e p e r i o d f r o m S e p t e m b e r 1984 t o A p r i l 1 9 8 5 . Dactylis Holcus Lolium Phleum Poa AgropyronTrifol ium Taraxacum Ranuncu1 us D u n g g1omerat a 1anatus perenne pratense compressa repens r e p e n s officinale acri s Dactylis glomerata 7 5 , .6 0. .0 4, .4 2 .2 4 .4 2 .2 6. . 7 2 . .2 0 , . 0 2 . 2 Holcus lanatus 7 , . 5 7 3 , , 1 1 . . 5 1 .5 3 . 0 1 . 5 10. . 2 O. .0 O. .0 0 . 0 Lol ium perenne 3, . 1 12. . 5 5 3 . . 1 0 . 0 6 .3 0 . 0 21 . .9 3. . 1 0 . . 0 0 . 0 Phleum pratense 16 . . 7 0 . .0 0 . ,0 50 . 0 0 . 0 0 . . 0 16 . , 7 0 . .0 0 . .0 0 . 0 Poa compressa 1 . . 9 5 . .6 13 . ,0 0 .0 42 .6 0 . . 0 2 0 . .4 3 . ,7 0 . .0 1 .9 Agropyron repens 0 . .0 0 . .0 22 .2 22 .2 0 . 0 11 . 1 2 2 . .2 0 . ,0 11 . . 1 0 . 0 Trifolium repens 9 . . 8 9 . 8 5 . . 2 3 . 3 11 .8 0 . . 7 5 4 . .9 2. .6 0 . . 7 0 . 7 Taraxacum officinale 4 . .5 0 . .0 9 . . 1 4. .5 9 . 1 0 . . 0 13. .6 5 4 . .5 4 . ,5 0 . 0 Ranunculus acris 0 . .0 0 . 0 0 . .0 0 . .0 0 . . 0 0 . ,0 3 3 . ,3 16 . 7 5 0 . .0 0 . 0 D u n g 8 . .2 5. 4 16. ,3 6 . . 1 1 1 . .6 4 . , 1 3 2 . .7 5. 4 1 . ,4 3 . 4 6S cn T r a n s i t i o n p r o b a b i l i t y o f row s p e c i e s b e i n g r e p l a c e d b y a c o l u m n p a s t u r e f o r t h e p e r i o d f r o m S e p t e m b e r 1983 t o O c t o b e r 1 9 8 3 . s p e c i e s , o n a n d a r o u n d d u n g i n t h e 1 9 3 9 Dactylis Ho 1cus Loli um Ph1eum Poa AgropyronTrifol ium T a r a x a c u m R a n i / n c u f u s D u n g glomerat a 1anatus perenne pratense compressa repens r e p e n s of f ic inale a c r i s Dactylis glomerata 9 7 .9 0 . . 0 0, .0 2 . 1 0 . 0 0 . . 0 0 . 0 0 . 0 0 . . 0 0 . 0 Holcus lanatus 0 .0 87 . 2 2. 6 0 . . 0 5 . 1 0 . .0 2 .6 0 . 0 0 . . 0 2 . 6 Lolium perenne 0 .0 0 . 7 9 0 . 1 0 . 0 4 .6 0 . 7 2 . 0 0 . 0 0 . . 7 0 . 7 Phleum pratense 0 .0 0 . .0 0 . .0 100 .0 0 . 0 0 . .0 0 . 0 0 . 0 0 , . 0 0 . 0 Poa compressa 0 . , 0 1 . . 7 5. .0 0 . .8 8 5 . . 1 2. .5 1 . 7 0 . . 0 0 . 8 0 . 8 Agropyron repens 0 . .0 15. .4 15 . .4 0 . .0 0 . . 0 6 9 , ,2 0 . . 0 0 . .0 0 . .0 0 . 0 Trifolium repens 0 . .0 0 . .0 1 . .9 0 . .0 1 . .4 0 . ,0 94 . .3 0 . .0 1. .0 0 . 0 Taraxacum officinale 3 . ,0 0 . .0 0 . .0 0 . ,0 0 . .0 0 . ,0 3. .0 9 0 . .9 0 . .0 0 . 0 Ranunculus acris 0 . .0 0 . O 0 . O 0 . .0 0 . . 0 0 . .0 O, O 0 . ,0 1 0 0 . .0 0 . 0 D u n g 0 . . 7 0 . .0 10. .9 0 . .4 10. .9 2. ,5 7, ,6 2. ,9 1. 4 6 0 . 5 CTl T r a n s i t i o n p r o b a b i l i t y of row s p e c i e s b e i n g r e p l a c e d by a column s p e c i e s , p a s t u r e f o r the p e r i o d from September 1983 to October 1983. and around m o l e l i l l s i n the 1939 M o l e h i l l Dactylis Holcus Lolium Phleum glomerat a /anatus perenne pratensecompress a repens repens officinale Molehi11 75 .6 0, . 7 0 .0 5. .3 0 .0 8 . 1 0 .4 5. . 3 1 .8 Dactylis glomerata 5 .0 95 , .0 0. .0 0. .0 0. .0 0 .0 0 .0 0. .0 0 .0 Holcus lanatus 4, , 3 2. . 1 70 .2 10. ,6 0, ,o 6 .4 0 .0 2. . 1 0 .0 Lolium perenne 10 . 5 0. .0 0 .6 79, .5 0, .6 4 . 1 0, .6 2 .9 1 .2 Phleum pratense 10 .0 0, .0 0. .0 0. .0 60, .0 20 .0 0. .0 0. .0 0 .0 Poa compressa 8 .3 0. .0 0, ,0 3 . 6 0. .0 76 . 2 4 , .8 4. 8 0 .0 Agropyron repens 0. ,0 10. .0 0. .0 0. ,0 0, ,0 0 .0 70. .0 1 . .0 O.O Trifolium repens 9, .3 O, .4 0 .4 2, .0 0, .0 2 .4 0 .4 84. .2 O.O Taraxacum officinale 6, .9 0. .0 0, .0 3. .4 0, .0 6, .9 0. .0 3. .4 • 79.3 Ranunculus acris 14, , 3 0. 0 0. .0 0. ,0 0. .0 0. .0 0. .0 O. .0 0 . 0 PoaAgropyronTrifol ium Taraxacum Ranunculus acr i s 0 .0 0 .0 0 .0 0 .0 0 .0 0 .0 0 .0 0 . 4 0 .0 85.7 CTl CO 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0097487/manifest

Comment

Related Items