HERBIVORY AND PLANT COMMUNITY STRUCTURE IN A SUBARCTIC ALTITUDINAL GRADIENT. Jon Moen Umeå 1993

Transcription

1 HERBIVORY AND PLANT COMMUNITY STRUCTURE IN A SUBARCTIC ALTITUDINAL GRADIENT Jon Moen Umeå 1993 Department of Ecological Botany University of Umeå S Umeå Sweden AKADEMISK AVHANDLING som med vederbörligt tillstånd av rektorsämbetet vid Umeå universitet för erhållande av filosofie doktorsexamen i ekologisk botanik kommer att offentligen försvaras torsdagen den 29 april 1993, kl i Hörsal C, Naturvetarhuset. Examinator: Prof. Lars Ericson, Umeå Opponent: Dr. Nancy Huntly, Pocatello, USA

2 Organization UMEÀ UNIVERSITY Department of Ecological Botany S Umeå, Sweden Document name DOCTORAL DISSERTATION Date of Issue April 1993 Author Jon Moen Titel Herbivory and plant community structure in a subarctic altitudinal gradient Abstract The object of this thesis was to study plant community structure, especially in relation to vertebrate herbivory, in an altitudinal gradient in the Fennoscandian mountain chain. A sowing experiment in a high alpine Ranunculus glacialis population showed that seeds germinated better in cleared microsites than under established individuals. This is contrasted with a hypothesis that predicts positive plant-plant interactions in high alpine environments. It was concluded that plant-plant interactions in die studied population varied from neutral to negative, whereas no indications for positive interactions were found. An exclosure experiment in a snow-bed showed that a lemming population consumed 33 % of the available graminoids and 66 % of the mosses from August to June during a population peak. The results shows that grazing needs to be considered as a structuring factor in snow-bed vegetation. The vegetation in exclosures in another snow-bed changed from a graminoid-dominated to a herb-dominated plant community during a long-term (six years) experiment No changes of the same magnitude were seen in a tall herb meadow on a lower altitude. Survival of transplanted adult shoots from the tall herb meadow was equally high in the snow-bed as on the meadow, and germination was also high on bare ground in the snow-bed. Grazing seemed to be a more important structuring factor in the snow-bed than in the more productive tall herb meadow. Raising the grazing pressure during one growing season by introducing microtine rodents into enclosures did not cause any large short-term effects on plant community structure in a tall hob meadow or in a snow-bed. Marked shoots showed that some preferred plant species had a high shoot mortality, but biomass for pooled categories of plants was not significantly affected. It was predicted that the tall herb meadow would be more grazing sensitive than die snow-bed, but productivity on the meadow seemed to be sufficiently high for the plants to compensate for the grazing during the growing season. A greenhouse experiment showed that voles, when grazing freely, have the potential to deplete productive field layer vegetation contrary to predictions from plant defence theories. A nitrogen-based defence did not prevent heavy shoot mortality for toxic tall herbs. Key words arctic and alpine vegetation, competition, exclosures, lemmings, plant defences, plant-plant interactions, Ranunculus glacialis, snow-bed, structuring factors, tall herb meadow, vertebrate grazing, voles Language ISBN Number of pages English Date December 28,1992 2

3 "Här kan man läsa hur allting är, och hur det står till med både folk och få." "Vad då?" "Låt se..." "Vad menar du med det?" "Jo. Här står hur djuren gör och tänker. Djur har också mänskors känslor, djuren är ju också mänskor. Ja, det är just vad dom är, i Fablernas Värld, i Fablernas Värld. Simma lugnt!" Den Kloka Ugglan 3

4 Cover by Lottie Eriksson The cover depicts symbols from old Lappish sacred drums ("Trolltrummor"). The drums were used by holy men ("nåjder"), and the symbols on the drums represented gods, spirits and other important powers in the lives of the Lappish people. On the front cover, the God of Plants and Fertility stands to the left of the Weather God, who is holding symbols of good weather and rain. On the back cover, there are symbols of the sun, the moon and the stars. The line symbolises both the altitudinal gradient and the edge of the drum. 4

5 CONTENTS LIST OF PAPERS...6 HERB IVORY AND PLANT COMMUNITY STRUCTURE IN A SUBARCTIC ALTITUDINAL GRADIENT... 7 INTRODUCTION...7 AIMS AND OBJECTIVES... 8 STUDY AREA HYPOTHESES ABOUT PLANT COMMUNITY STRUCTURE The role of abiotic factors Plant competition...16 Trophic interactions...17 The importance of bottom-up versus top-down forces WHY DID I DO IT LIKE THIS?...21 SUMMARY OF THE PAPERS...23 High alpine plant populations (I, IH) Effects of excluding mammalian herbivores (II, III) Effects of intensive grazing by microtine rodents (IV, V)...26 DISCUSSION AND CONCLUSIONS...29 ACKNOWLEDGEMENTS REFERENCES

6 LIST OF PAPERS: This thesis is a summary and discussion of the following papers, which will be referred to in the text by their Roman numerals. I. Moen, J. Positive versus negative plant interactions in a high alpine Fennoscandian boulder-field. Submitted manuscript ü. Moen, J., Lundberg, P. A. & Oksanen, L. Lemming grazing on snow-bed vegetation during a population peak. Submitted manuscript III. IV. Moen, J. & Oksanen, L. The relative importance of stress, competition, and grazing for plant community structure in a subarctic altitudinal gradient Submitted manuscript Moen, J Summer grazing by voles and lemmings upon subarctic snow-bed and tall herb meadow vegetation - an enclosure experiment. Holarctic Ecology 3: V. Moen, J., Gardfjell, H., Oksanen, L., Ericson, L. & Ekerholm, P. Grazing by food-limited microtine rodents on a productive experimental plant community: does the "green desert" exist? Submitted manuscript Paper IV is reproduced with due permission from the publisher. 6

7 7 HERBIVORY AND PLANT COMMUNITY STRUCTURE IN A SUBARCTIC ALTITUDINAL GRADIENT INTRODUCTION One of the goals of ecology is to explain patterns of distributions and abundances of organisms in nature (Andrewartha & Birch 1954, Begon et al. 1986, DeAngelis 1992). These patterns are sought on many different spatial scales, from global patterns of biomass and primary productivity (Moen & Oksanen 1992, L. Oksanen et al. 1992) to local distribution patterns (Gurevitch 1986, Reader 1992). Explanations for the patterns come from an equally large array of levels: from physiological adaptations (Crawford 1989) to interactions between organisms on different trophic levels (Fretwell 1987). Due to time trade-offs and personal interests, ecologists tend to specialise on certain levels, and communication between scientists in different fields is sometimes difficult For example, mathematical models of ecological systems are today created at a rate many times faster than the rate at which empirical data to test these models are collected. The lack of testing of general theories may cause ecology to become a "weak" science with little predictive power (Peters 1991). It is thus extremely important that theories and models are scrutinised for generality and testability, and that those models which pass the sieve are tested in the field. Models predicting different mechanisms behind the same pattern in nature are particularly interesting for they permit alternative explanations to be tested in the same system. A specific spatial pattern that have intrigued scientists for a long time is the abrupt changes in the vegetation that occurs when you

8 ascend a mountain in the Fennoscandian mountain chain. This altitudinal gradient is characterised by mountain birch forests on lower altitudes, often with luxuriant tall herb vegetation in the field layer, a more or less abrupt termination of the forest at the timberline, and extensive dwarf shrub and graminoid heaths, interspersed with snow-beds, above the timber-line. At the highest altitudes, vast boulder-fields dominate, and the vascular plant cover consists of scattered solitary individuals. Studies of vegetation in the Fennoscandian mountains have a long tradition in Scandinavia (e.g. Vestergren 1902, Fries 1913, Samuelsson 1917, Nordhagen 1928, Rune 1953). The rather harsh environment in tundra areas has perhaps inspired most scientists to focus on adaptations of organisms to the physical environment Recently, many studies have emphasised the role of the biotic environment and especially the effects of herbivores on plant species and on plant community structure. Work on microtine rodents in the Fennoscandian mountains (Oksanen & Oksanen 1981, Andersson & Jonasson 1986), and in North American tundra areas (Batzli 1975, Batzli et al. 1980), together with work on pikas (Lagomorpha) in the Rocky Mountains (Huntly 1987), and on reindeer on the subantarctic island of South Georgia (Leader-Williams 1988) have shown the potential of herbivores to influence and change plant communities in climatically harsh environments. AIMS AND OBJECTIVES This thesis is mainly about plant community structure and the effects of mammalian herbivores, i.e. microtine rodents and reindeer, in the Fennoscandian mountain chain. The rodents are either cyclic, or strongly fluctuating (Oksanen & Oksanen 1992), and thus appear in high numbers certain years. Proposed 8

9 explanations for these fluctuations have included almost anything from sunspots to plant-herbivore interactions to predation (see Resit Akçakaya 1992 for a recent review). However, central to many models and explanations is the interaction between plants and the herbivores, and most authors seem to agree that the rodents should have an effect on the plant community in peak years. One hypothesis which predicts a pattern in the grazing pressure along an altitudinal gradient is the Oksanen hypothesis (which will be summarised and discussed below. The hypothesis is called "the trophic exploitation hypothesis", "the exploitation ecosystems hypothesis" and "the Oksanen hypothesis" in different papers in the thesis). The main questions in the thesis are: 1. What kind of plant-plant interactions are found in a high alpine Ranunculus glacialis population? (I) 2. What determines the distribution of R. glacialis at lower altitudes? (HI) 3. What are the impact of a lemming population during a population peak on snow-bed vegetation? (II) 4. What prevents species common in tall herb meadows from establishing in snow-beds above the timber-line, abiotic stress of grazing? (HI) 5. What are the effects of intensive grazing during the summer on plant communities in a tall herb meadow and a snow-bed?(iv) 6. Are most plants in a productive plant community of sufficiently high quality for herbivores to have a positive energy or nutrient balance, or are they toxic? (V) 9

10 1 f Û in n ^ LL 2 5 ^ < w O ^ Figure 1. Climate data from Finnmarks vidda. A. Mean precipitation and temperature for the period from Solovomi (c. 25 km SW of the research area, 374 m.a.s.1.)* B. Mean precipitation from the research area for the period (Joatka fjellstue, 406 m.a.s.l.)* C. Mean temperature for the period from Halddi (see paper 1,893 m.a.s.1.)* 10

11 STUDY AREA The field work for this thesis has been done on, or close to, Finnmarksvidda in northern Norway. Finnmarksvidda is a large plateau composed of undulating hills (altitude m.a.s.l.) surrounded by higher mountains in the north and west. The highest peaks in the area reaches ~1000 m.a.s.l. The climate is fairly continental with cold winters and warm summers. Snow usually covers the ground between October and May, and the annual precipitation is relatively low ( mm), about half of which falls during summer (Fig. la). The main research area (69 46'N, E) consists of a lower plateau with mainly lichen heaths, a south-facing steep slope with mountain birch forests, tall herb meadows, and willow thickets, and a higher plateau with dwarf shrub heaths and snow-beds (Figs. 2 and 3). The timber-line on the slope runs roughly at 480 m.a.s.l., which is about 100 vertical meters higher than generally in Finnmarksvidda due to the favourable southern aspect of the slope. Vast areas on the lower plateau are tree-less, except in topographically favourable positions. Precipitation data from the period is depicted in Fig. lb. The high alpine Ranunculus glacialis population in paper I is situated about 60 km west of the main research area on the mountain complex Vuorasnjarhaldi (~850 m.a.s.l.). The area is dominated by boulder-fields. Temperature data from the old northern lights observatory, c. 2 km east of the study area (893 m.a.s.1.), are given for the period in Fig. lc. The main vertebrate herbivores in the research area are field voles (Microtus agrestis), root voles (M. oeconomus), grey-sided voles (Clethrionomys rufocanus), red voles (C. rutilus) and lemmings (Lemmus lemmus), together with reindeer (Rangifer tarandus). Other herbivores are two species of grouse (Lagopus lagopus, and L mutus). Predators include mustelids (Mustela erminea and M. 11

12 Figure 2. Map of the research area. Dark areas are lakes, hatched areas are mires, and the stippled areas are mountain birch forests. P stands for the upper snow-bed on the plateau in paper II, T stands for the tall herb meadow, and S for the snow-bed both in paper III. The tall herb meadow and snow-bed studied in paper IV are located c. 1 km W of the T and S. The vertical line starting at P corresponds to the altitudinal profile in Fig

13 ^ 560 ê 540.g 520 å? Distance 2 3km Figure 3. Altitudinal profile along the vertical line in Fig. 2. "T" is the tall herb meadow, "S is the snow-bed, and "P" is the snow-bed in paper II (see below). I 25 Plateau Snow-bed Meadow 86s 86a 87s 87a 88s 88a 89s 89a 90s 90a 91s 91a 92s 92a Year Figure 4. Capture indices for microtine rodents in the area (Data from Ekerholm & Oksanen, unpubl.). 86s stands for spring 1986, 86a for autumn 1986 etc. 13

14 nivalis), foxes (Vulpes vulpes and Alopex lagopus), owls (Sumia ulula and Asio flammeus), and birds-of-prey (Buteo lagopus and Stercorarius longicaudus). The voles are cyclic with a period between peaks of four to five year, whereas lemmings show conspicuous, but erratic, fluctuations (Oksanen & Oksanen 1992; see also Fig. 4, unpublished data from P. Ekerholm & L. Oksanen). The reindeer move through the area mainly during spring and autumn migration, but small, scattered herds may also be seen during the summer. HYPOTHESES ABOUT PLANT COMMUNITY STRUCTURE IN ARCTIC AND ALPINE ALTITUDINAL GRADIENTS The results in this thesis may be interpreted in the light of several hypotheses which concern plant communities in arctic and alpine areas. Factors discussed include influences from the abiotic environment, competitive effects within plant communities, and interactions with other trophic levels. I will discuss the last hypothesis at some length since that has, at least to some extent, been a part of each of the five papers in the thesis. The role of abiotic factors Much work has focused on the direct and indirect effects of the abiotic environment on plants and plant communities. The most important factor influencing primary productivity seems to be the length of the growing season. Even a small increase in the growing 14

15 season (due e.g. to earlier snow-raelt) may give a large increase in carbon gain (Chapin 1987). During the growing season, the plants are influenced by low temperature, both in the soil and in the air, through processes like nutrient uptake, decomposition, and chemical weathering (Billings & Mooney 1968, Jonasson 1983, Billings 1987, Callaghan & Emanuelsson 1985, Callaghan 1989). Most physiological processes in arctic plants seem to be less temperature sensitive, or to reach optimal values at lower temperatures, than in boreal and temperate plants (Crawford 1989). Growth is thus seldom limited by temperatures directly, but more often indirectly by the low rates at which resources become available (Chapin 1983). Snow distribution is rather predictable, even though snow depth may vary considerably, and this may cause typical local patterns in plant distributions (Dahl 1956, Gjærevoll 1956). On exposed ridges, the plant communities consist of plants which can withstand very low temperatures and desiccation (Gjærevoll 1956). The dominating growth forms are cushion plants and prostrate evergreen shrubs which can tolerate both winter frost drought and summer drought (Billings & Mooney 1968). These prostrate growth forms are interpreted as having the ability to exploit the more favourable microclimate near the ground (Callaghan & Collins 1976). In snow-beds, where the snow cover is deeper and lasts longer, there is usually a well developed zonation which relates to the time of snow melt With increasing snow depth there is a series of plant communities consisting of dwarf shrubs, low herbs and grasses, and finally mosses and lichens where the snow is deepest and the growing season is shortest (Dahl 1956, Gjærevoll 1956, Wijk 1986). These plants are adapted to a more or less strongly reduced growing season, but they are not exposed to particularly low temperatures in the winter, and meltwater may supply the plants with moisture for all or most of the growing season. The harshness of the abiotic environment is said to increase with increasing altitude, and the physical environment is seen as 15

16 becoming more and more important for plant population processes at higher altitudes. It is even stated that positive plant-plant interactions, such as commensalisra or mutualism, are particularly important in the most severe environments in the arctic (Callaghan & Emanuelsson 1985, Callaghan 1987, Carlsson & Callaghan 1991). Plant competition The relative importance of positive and negative plant-plant interactions is debated. Tilman (1988, and references therein) assumes that both the intensity and importance of competitive interactions (sensu Weiden & Slauson 1986) are high in all parts of productivity gradients, but that plants compete for different resources in different parts of the gradient. Due to trade-offs in the allocation of resources within a plant, these different factors will select for different growth forms along the gradient. Moving from high to low productivity, stem allocation is predicted to decrease and root allocation to increase (Tilman 1988:111) as light is limiting in high productive areas and nutrients in low productive areas. Some of these ideas were formulated already in the beginning of this century by Cajander (1909, in L. Oksanen 1991a), who saw competition as the main proximate factor determining the distribution and local abundance of plants, and who claimed that changes in the abiotic conditions caused changes in the balance of competitive abilities between plants. These ideas might explain the observed change from tall herbs in subalpine areas to more prostrate plants at higher altitudes. 16

17 Trophic interactions Some authors have emphasised interactions between trophic levels when discussing distribution patterns in plants. Based on the idea of Hairston et al. (HSS; 1960) that herbivores are generally regulated by predators, Fretwell (1977, 1987) extended their argument to situations with different numbers of trophic levels. The main point of HSS was that the world is green because predators keep herbivore populations down. Fretwell (1977) noticed that the world sometimes is not green: some habitats are normally grazed down by naturally occurring herbivores, and he suggested that the number of trophic levels must then be less than the three discussed by HSS. He stated a verbal argument (which was later formally analysed by L. Oksanen et al. 1981) that the number of trophic levels is dependent on primary productivity. The higher the productivity, the more functional trophic levels in the system. The theoretical analyses of L. Oksanen and his colleagues (L. Oksanen et al. 1981, L. Oksanen 1988, T. Oksanen 1990, T. Oksanen et al. 1992) came to the following conclusions: In extremely unproductive areas (primary productivity <30 gnr^yrl; L. Oksanen 1983), plant production is assumed to be too low to sustain herbivorous animals. In undisturbed areas, plants are then predicted to eventually deplete their resources and thus compete. In moderately productive areas, plant production is high enough to sustain herbivores, albeit at low population densities, lower than what is needed for efficient predators to have a positive growth rate. Uncontrolled by predation, these herbivores are predicted to exert a strong grazing pressure on the vegetation. In productive areas (primary productivity >600 g nr^yr*), plant production is high enough to sustain both herbivores and predators. With the herbivores kept down by predators, the plant communities experience a low grazing pressure, and should be structured by competition. The plant communities along the altitudinal (and productivity) gradient should thus be structured chiefly by competition in both productive and extremely unproductive areas with a zone of strong grazing pressure in between (see Fig. 5 for a 17

18 schematic view of the biomass predictions from the model). The abiotic environment enters into this hypothesis as a determinant of primary productivity, which in turn determines the number of trophic levels. Abiotic processes set the stage for the kind of biotic interactions that are possible in a specific area. V) to E o bo - Plants Herbivores Predators Primary productivity Figure 5. Biomass patterns predicted by the Oksanen hypothesis. Note that the different trophic levels do not have the same scale on the y-axis. Later analyses have increased the realism in the original hypothesis by adding spatial heterogeneity (T. Oksanen 1990, T. Oksanen et al. 1992), seasonality (L. Oksanen 1990a), plant defences (L. Oksanen 1990b), and coevolution between herbivores and predators (L. Oksanen 1992) to the basic model. These re-analyses suggest that the above predictions are mainly applicable on a landscape level and on an ecological time-scale. Seasonality may cause cycles in the herbivore populations in areas with a longlasting snow cover where generalist predators cannot be efficient enough to keep herbivores at a constant low level (cf. Hansson & Henttonen 1985). 18

19 The hypothesis has received criticism regarding the realism of some of the assumptions, the testability, and the generality. Already HSS was criticised for assuming that all plant was edible (Murdoch 1966). The world could be green because herbivores are unable to deplete the vegetation, not because herbivore populations are kept down by predators. This criticism is examined more in detail in paper IV. Polis (in press) and Power (1992) question whether trophic levels can be clearly delimited, and thus whether trophic level biomass can be measured. However, L. Oksanen (1991b) argues that terrestrial systems do not support more than three trophic levels, and that these three levels are fairly distinct: "it takes entirely different kinds of adaptations to photosynthesise, to consume fibre-rich vegetative plant organs and to capture mobile prey" (L. Oksanen 1991b:58). It remains to be seen whether this will be supported by field studies. Even if trophic levels are abstractions that have measurable properties in nature, Polis (in press and pers. comm.) regards it as unlikely that all species on a particularly trophic level are regulated by the same factor, i.e. that the hypothesis is unrealistically simplistic. However, the variables in the model are total biomass on each trophic level and not biomass of each species, and Slobodkin et al. (1967) in a response to a similar critique said that the statements of HSS apply to "the quantitatively dominant species but not necessarily to the numerical majority of species in any ecosystem" (Slobodkin et al. 1967:109). Thus, the hypothesis does not claim to predict the regulating factors on the species level, but on trophic levels as a whole. Criticism can also be raised against the generality of the hypothesis. Omnivory and cannibalism, i.e. a breakdown of the trophic level concept, have been shown to be common among arthropods even in relatively simple systems like deserts (Polis 1991). Strong (1992) pointed out that most examples of the kind of "trophic cascades" predicted by the hypothesis have been found in aquatic systems or in low-diversity terrestrial systems like the tundra, and is thus to be seen as an exception rather than as a rule. 19

20 His main point is that, in species-rich systems, food webs tend to be more reticulate than chain-like, and that this reticulate structure "buffers" against trophic interactions cascading down the system. L. Oksanen et al. (1981) suggested that the terrestrial grazing chain consists of two fundamentally different branches: the vertebrate and the arthropod. The arthropod branch may be less sensitive to spatial variations in annual primary productivity since arthropods are often able to complete their life cycle during the most favourable season, and the length of the arthropod food chain may thus be uncorrelated to annual primary productivity. The conclusion of L. Oksanen et al. (1981) was that the hypothesis is most applicable to the vertebrate branch. For this reason, invertebrate herbivores have not been studied in this thesis, and "herbivory" refers to "vertebrate herbivory", unless otherwise stated, in all parts of this thesis. A methodological problem with testing the hypothesis in the field is whether the independent and dependent variables of the model (i.e. net primary productivity and trophic level biomass) can be quantified (Power 1992). If the structuring forces in a particular habitat are to be predicted from the model, it is essential to know where in the productivity gradient the habitat lies. Indeed, it is important to know whether or not the biomass patterns predicted by the model exists at all in the productivity gradient of interest. Data from IBP-studies in arctic areas (L. Oksanen 1983), and global patterns in plant and herbivore biomass (Moen & Oksanen 1992, L. Oksanen et al. 1992) indicate that the patterns may exist, but these studies do not make it possible to determine where the transition point from one trophic level to two trophic levels is, or where a system changes from two trophic levels to three. Without knowledge of the biomass pattern in a specific productivity gradient and of the critical productivity values where the transition points should be, the hypothesis cannot be critically tested. 20

21 The importance of bottom-up versus top-down forces: are they exclusive or complementary? The different views outlined above are not mutually exclusive. Abiotic factors will influence populations on all trophic levels (Dunson & Travis 1991, Hunter & Price 1992), and strong competitive effects are predicted both by Tilman (1988) and L. Oksanen et al. (1981). Hunter & Price (1992) make a strong case for the importance of bottom-up effects in ecosystems, i.e. the influence of resources on trophic levels,: "the removal of higher trophic levels leaves lower levels present (if perhaps greatly modified), whereas the removal of primary producers leaves no system at all" (Hunter & Price 1992:725). This is exactly the view taken by L. Oksanen et al. (1981). Abiotic factors in concert determines the primary productivity of a particular area. This energy input determines the number of trophic levels possible and thus the type of trophic interactions: the top-down effects are superimposed on a bottom-up template. WHY DID I DO IT LIKE THIS? or METHODOLOGICAL BACKGROUND I began my research by doing the work in paper IV in 1986, and then setting up the experiment in paper HI in 1987, in order to test various aspects of the Oksanen hypothesis in a gradient in northern Norway. However, due to a lack of knowledge of the actual biomass and productivity patterns mentioned above, the results may be somewhat tricky to interpret and also to communicate to others (the data from IBP-studies in arctic areas summarised in L. Oksanen [1983] suggest that the pattern do exist, but the data is too fuzzy to permit good estimates of the transition points). If one 21

22 assumes that the tall herb meadow is a three-link habitat, i.e. with plants, herbivores and predators, and the snow-bed a two-link habitat, the rodents should be important for plant community structure on the snow-bed but not on the tall herb meadow. However, if I would get a different result, is the hypothesis then falsified? My personal view is that, as long as the actual patterns that the model predicts have not been shown to exist, and the transition points have not been stated with a fair amount of accuracy, it is impossible to predict the specific dynamics in a specific habitat. If I do not get an effect by excluding rodents on the snow-bed, I cannot say whether that is because the hypothesis is wrong or whether the snow-bed is also within the three-link zone where grazing pressure is predicted to be light If I had worked with more plots along the entire altitudinal gradient, and thus with the whole range of productivities, I might have been able to show whether or not there is a zone of intense grazing pressure as predicted by the hypothesis. As it stands, the results refer to differences between habitats that also happen to be positioned in an altitudinal and a productivity gradient. The papers in this thesis are best viewed as inspired by the Oksanen hypothesis, and the results interpreted in the light of the hypothesis, but they do not constitute a critical test. According to Keddy (1989) and Peters (1991), an objective of an ecological theory is to make general predictive statements about the factors or patterns of interest. Tests of these theories ahould also be made as general as possible. There are several ways of increasing generality in field experiments (Keddy 1989): (1) to examine general patterns in the first place (2) to use an increased number of species (3) to use comparative ecology, e.g. functional groups, rather than individual species (4) to use general experimental factors or variables (5) to arrange treatments along natural gradients. 22

23 I have tried to apply these criteria in various ways. (1) Examination of global biomass patterns along productivity gradients indicate that the pattern predicted by the Oksanen hypothesis may be a very general one (Moen & Oksanen 1992, L. Oksanen et al. 1992). (2) I have worked with entire plant communities, or at least the entire vascular plant community, rather than manipulating single species, except in paper I and parts of paper HI. (3) I have consistently analysed and reported the results as effects on groups of species rather than as effects on individual species. I have chosen to use fairly coarse groupings (usually herbs, graminoids and woody plants) rather than more refined classifications (e.g. Grime 1979, Barkman 1988, or L. Oksanen & Ranta 1992). This has been done in the hope that the results in this thesis might be more easily compared to other studies, and also that it would be possible to predict responses in other plant communities with a different species composition (see the discussion in paper III). (4) In papers IV and V, biomass has been one of the measured variables. This variable is often measured in studies on plant communities, and the results may thus be compared with other studies. (5) The area of interest is an altitudinal gradient. SUMMARY OF THE PAPERS High alpine plant populations (I, HI) High alpine areas are characterised by a lack of continuous vegetation. The ground is mainly covered with bare rock or gravel, and the vascular vegetation consists of scattered solitary individuals. It has been suggested that positive plant-plant interactions are particularly important in these harsh environments, 23

24 and that aggregations of individuals provides shelter from wind exposure, moisture for germination and seed establishment, organic substrates, and an increased local temperature, which could increase the survival probabilities of individuals. For example, cushion plants have been suggested to provide favourable conditions for germination and seedling establishment of herbs and grasses. However, it is not clear whether high alpine plants with a more erect growth form like Ranunculus glacialis also provides such conditions. To test whether seeds germinate better close to established R. glacialis individuals than in open ground, seeds of Oxyria digyna were sown in autumn in a high alpine habitat in northern Norway (paper I). Seeds sown in 1988 had a higher germination and early survival in cleared (open) microsites than close to established individuals in The experiment was repeated in 1989, but no significant differences in 1990 were found. Measurements of ground surface temperatures gave a significantly lower mean temperature under individuals than on open ground, and this is suggested as a mechanism causing the difference in the seeds sown in Air temperature and precipitation data indicated that the growing season of 1990 was warmer and sunnier than 1989, which could explain why the results from the two years differed. The results thus show that plant-plant interactions in this plant community vary from neutral to negative, whereas no indications of positive interactions were found. Another question regarding high alpine plant populations is why these species are usually not found in lower altitudes (paper III). Climate was considered to be an unlikely explanation since R. glacialis is sometimes, especially along stream beds, found down into taiga areas, and it may also be found on frost heaved ground in intermediate altitudes. It seems more likely that they are either weak competitors in closed vegetation, or that they cannot survive the high grazing pressure in intermediate altitudes predicted by the Oksanen hypothesis. To test this, R. glacialis individuals were transplanted into snow-beds in a factorial design with exclosures and open plots in both cleared and intact vegetation. Survival three and four years after the transplantations showed that survival was 24

25 higher in cleared plots (~45 %) than in closed vegetation (-15 %). Survival in closed vegetation was higher in exclosures than in open plots indicating that both competition and grazing may be important for restricting the altitudinal distribution of the species. Effects of excluding mammalian herbivores (II, HI) The impact of a lemming population on snow-bed vegetation during a population peak was studied in paper II. Snap trap data had indicated an increase in lemmings on the upper plateau of the Joatka area in the autumn of To quantify the grazing effect of the lemmings, twelve small exclosures (c. 50 x 80 cm) and twelve open plots were randomly placed on a Carex bigelowii dominated snow-bed. The relative abundances of vascular plants, bryophytes and lichens, together with litter and faeces were measured from 1988 to The results show a significant decrease in graminoids and bryophytes in the open plots as compared to the exclosures. A conservative estimate indicated that 33 % of the available graminoids and 66 % of the moss cover was consumed between August 1988 and June The lemming population had crashed during the winter of 1988/1989, and the vegetation showed a quick recovery. However, an analyses of the relative abundances in 1992 showed that there were significantly higher abundances of hepatics and Pohlia spp. in the open plots, both of which may be favoured by disturbance. It was concluded that grazing must be considered when the development of snowbed vegetation is discussed. The long-term (six years) effect of excluding herbivores from a tall herb meadow and a snow-bed was studied in paper III. Three exclosures and three open plots (-64 m^ each) were established in each habitat in Plant community structure was measured from 1987 to 1992 with a modified point intercept method. Adult 25

26 plants of fifteen common species (six tall herbs, three woody plants, and six snow-bed plants) were transplanted into intact vegetation in the plots in 1987, and survival and growth were measured from 1987 to Seeds of a subsample of the fifteen species mentioned above (ten species in total: three tall herbs, two woody plants, and five snow-bed plants) were sown on bare soil in 1988 to study germination and early survival of seedlings. The results show that plant community structure changed in the exclosures in the snow-bed from a graminoid-dominated community to a herb-dominated community. No such changes were seen in the open plots or in the tall herb meadow. It also showed that all tall herb species may germinate in the snow-bed if patches of bare soil exist, and that survival and growth for adult shoots from tall herb meadows was as high in the snow-bed as in the meadow. It was concluded that grazing was important for plant community structure on the snow-bed, but not in the more productive tall herb meadow. However, when individual species are examined, other factors were also important, and the relative importance of different structuring factors depends on the level at which the vegetation is studied. Effects of intensive grazing by microtine rodents (IV, V) An alternative way of studying the potential effects of herbivory is to construct fenced areas, introduce herbivores into the enclosures and monitor their effect on the plant community. The created grazing pressure will, for practical reasons, almost always be higher than what is normally found in nature. However, if only one herbivore is introduced in each enclosure, it will provide an opportunity to study the grazing effects in situations where territories or social interactions do not regulate the herbivores, while checking the herbivores regularly or performing the experiment indoors will control for predation. The results can thus 26

27 be seen as the potential effects of the herbivores that may or may not be realised in natural conditions. The Oksanen hypothesis predicts a low grazing pressure in productive habitats, e.g. tall herb meadows, and a high grazing pressure in less productive habitats, e.g. snow-beds. This implies that plants on the snow-bed should be more grazing-tolerant than plants from the tall herb meadow. If a high grazing pressure was created, this should have a larger effect on plant community structure in a productive meadow than on a less productive snowbed. This was tested in paper IV by establishing three enclosures and three open plots (-64 m^ each) in each of the two habitats, and permitting lemmings and grey-sided voles to graze in the enclosures for the major part of a growing season. The three enclosures were divided into two grey-sided vole enclosures and one lemming enclosure in each habitat. Initially one individual was introduced in each plot, and later two to four new individuals were introduced into each plot to simulate reproduction (see Tab. 2 in paper IV). Shoots of common species were marked and censused, plant community structure was measured, and biomass was sampled at the end of the experiment. Known preferred species had a significantly higher shoot mortality in the grazed treatments than in the ungrazed controls, but no major changes were seen in the plant community structure during the summer. Pooled aboveground biomass was not significantly lower in the grazed plots than in the controls in any of the two habitats at the end of the experiment The only significant reduction in biomass was in woody plants on the snow-bed in both grey-sided vole and lemming plots. The vegetation was thus little influenced by the grazing even though the grazing pressure was higher than in peak years. It seems that the productivity in the tall herb meadow was high enough to compensate for the grazing during the 34 to 54 days of the experiment If anything, the grazing effects were larger on the snow-bed than in the meadow, even though the snow-bed experiment lasted only 18 days, which is quite contrary to the predictions. This might be explained by the different growth forms in the two habitats. The tall herb meadow was dominated by 27

28 herbaceous species while the snow-bed was dominated by a woody plant, Salix herbacea. It is possible that the regrowth capacity of S. herbacea is lower than for the meadow species, and that it thus did not have the potential for short-term responses to the grazing during the experiment An essential assumption in HSS and the Oksanen hypothesis, is that the majority of the green tissue in plant species are of sufficiently high quality for naturally co-occurring herbivores to survive. The world is "green, i.e. herbivores do not normally deplete the vegetation, because the herbivores are limited by something else than food, e.g. predation or disease. Other authors have pointed out that the world may be green because plants are either toxic or of such low quality that the herbivores are foodlimited even though there seems to be an abundance of forage. This was tested using microtine rodents and field layer vegetation from a productive boreal spruce forest in paper V. The productivity and structure of the plant community were comparable to highly productive subalpine tall herb meadows. Ten plant communities of 7.5 m^ were constructed in a greenhouse by excavating 25 x 25 cm squares of vegetation and randomly allocating them to ten tables. The vegetation types from which the plant material were Geranium-Oxalis-Myrtillus-type and Geranium-Oxalis-Maianthemum-iype according to the Finnish forest type scheme. The squares were chosen to contain at least one individual of either Geranium sylvaticum (tall herb, non-toxic), Solidago virgaurea (rosette plant, non-toxic), or Actaea spicata (tall herb, toxic). The randomly created plant communities retained a large part of the small-scale spatial pattern found in natural communities, and they were also highly comparable to each other. One individual of either Microtus agrestis, Clethrionomys rufocanus, C. glareolus, or Lemmus lemmus were allowed to graze on a table until the weight development showed that the voles were starving. The total number of replicates in the experiment was three each for M. agrestis and C. rufocanus, one each for C. glareolus and L. lemmus, and two ungrazed control tables. The results for the three vole species show that the 28

29 Vegetation became severely depleted before the animals started to lose weight: an average of 10 % of the biomass remained as compared to the controls. Shoot mortality for toxic tall herbs was almost as high as for non-toxic tall herbs, and the toxic species also showed a lower ability to recover after grazing than the non-toxic species. The lemming, which is the only species that is not normally found in this type of vegetation, lost weight rapidly after depleting the grasses. We concluded that this type of vegetation is edible to the voles, and that they thus have the potential for depleting the vegetation. DISCUSSION AND CONCLUSIONS Most of the results in this thesis support the predictions from the Oksanen hypothesis, or, to put it differently, the Oksanen hypothesis captures some aspects of what goes on in this altitudinal gradient. Furthermore, the results in paper V shows that the assumption of the general edibility of plants in the model is a realistic one. Some of the results, however, are a bit ambiguous in relation to the hypothesis. Callaghan and coworkers have argued that the selective forces in tundra areas differ fundamentally from those in temperate and tropical areas (e.g. Callaghan & Emanuelsson 1985, Callaghan 1987). They suggest that negative biotic interactions, like plant competition and herbivory, are less important, and that these processes are replaced by positive plant-plant interactions as a protection against the harsh abiotic environment Paper I makes it quite clear that this is not a general pattern in all high alpine populations: it is not always an advantage to germinate in close proximity to another individual, and negative interactions may be found. Griggs (1956) also considered the interactions between cushion plants and the species that germinate in them as 29

30 competitive. The Oksanen hypothesis predicts competition in undisturbed, extremely unproductive environments since plant production is too low to sustain resident herbivores. Competition for either light or nutrients is somewhat hard to envisage in plant communities with scattered solitary individuals, and L. Oksanen (1980) and L. Oksanen & Ranta (1992) interpreted this situation as competition for favourable microsites, and the plant strategies as abilities to monopolise these microsites in a generally inhospitable "matrix" of stones and gravel. To test this, it would be necessary to see whether most or all potential microsites suitable for plant establishment are filled or not. This would be extremely difficult since safe sites can only be defined after establishment has taken place. The existence of negative interactions is supportive of the hypothesis, but it is not a critical test. The pattern of scattered solitary individuals could be due to a lack of propagules. The similarities in growth forms between some high alpine plants, like Ranunculus glacialis and Oxyria digyna, and herbs from tall herb meadows could be caused by localized nutrient-rich conditions due to frost action rather than adaptations to competition for microsites (see Jonasson & Sköld 1983, Jonasson 1986 for a discussion on the effect of frost heaving on soil nutrients). A slightly different scenario is described in Tilman (1988), who has shown that a higher loss rate (caused e.g. by disturbance) would select for a higher allocation to leaves, and thus to higher relative growth rate. A high loss rate would lead to a higher availability of both nutrients and light which, in turn, would decrease the selection pressure on allocation to roots and stems. This could also explain why plants with competitive characteristics sensu Grime (1979) is found both at low and high altitudes in the gradient (L. Oksanen 1990b, L. Oksanen & Ranta 1992). A central prediction of the Oksanen hypothesis is the strong grazing pressure at intermediate levels of the gradient. Significant effects of grazers were indeed found: lemmings did consume large amounts of the available forage on a snow-bed during a population peak (II), and a long-term exclosure experiment showed considerable changes in plant community structure over six years 30

31 on another snow-bed (III). Large effects by grazers in similar areas have also been found by Andersson & Jonasson (1986) and L. Oksanen (1988, 1990), and Jonasson (1992) found that plant competition was not a major limiting factor for co-occurring plants on a tundra area above the timber-line in northern Sweden. Taken together, this makes a strong case for the importance of herbivores in tundra areas. However, the data does not make it possible to delimit the altitudinal (or productivity) zone where grazing is important. To summarise, the main messages from this thesis are: (1) Plant-plant interactions in high alpine areas are not different from other areas, i.e. plants may show negative interactions even in climatically harsh areas (2) Grazers are important for plant community structure in snowbeds, but not subalpine tall herb meadows (3) Most field layer plants in productive boreal plant communities seems to be of sufficiently high quality for voles to survive, i.e voles have the potential to deplete the vegetation. This opens up several avenues for future research. The structuring factors, specifically the importance of disturbance, in high alpine plant communities should be studied more in detail. Another important task is to try to delimit a zone in an altitudinal gradient where grazers are important. The criticism against the hypothesis discussed in this summary makes it clear that it is very important for theoretical models which try to predict patterns in nature to be constructed in such a way that the critical variables are quantifiable in the field. Testing of models should also be made easier by stating assumptions and predictions as clearly as possible, preferably both in mathematical form and in plain English, so that critical tests might be designed as easily as possible. 31

32 ACKNOWLEDGEMENTS I would first like to thank Jacques Cousteau, because it was his films that made me realise as a ten-year-old that biology could be exciting. I would also like to thank my mother for always putting books in my hands, and for teaching me that reading is fun. I want to thank my supervisor, Lauri Oksanen, for believing in me in the beginning of my "career" (I hope he still does). He has forced me to think for myself, especially on practical matters concerning field work. It is a privilege for a graduate student to have a supervisor with an unending supply of ideas. I just hope that I have managed to sift out the good ones. Lars Ericson has also provided inspiration and much needed detailed knowledge of plants, and his comments on the manuscripts have greatly improved them. Working on this thesis would not have been half as fun without the contributions of many friends. Thanks to: Hans for many hours of discussions on anything from computer games to finer statistical points and ecological theories. Thank you for always having an opinion on everything, My two room mates for putting up with me (I guess they did not have much choice, and anyway I had to put up with them also). Thank you, Jonas, for having your things in such good order, and you, Peter, for not having them in order..., Pere, Doris, Michael, Peter and Taija for making field work a memorable event each year, and for many good discussions over the years, Anders for having the same Angst over his thesis as I had over mine at the same time, Katarina for drawing nice things (including the map in the summary), 32

33 Ulf Sperens for enjoying owls and rowans, Stickan for rearing delicious sheep, Âsa for not bringing lunch to the department either, Stefan for having books on everything, BeGe for knowing bryophytes (imagine that!) and sharing his knowledge freely, Jon Å for reading and commenting on the summary, and also for having such a nice name, Lottie for designing and drawing the cover of the thesis. I am honored!, Colin for creating a stimulating atmosphere amidst chaos, Chris for knowing statistics and being a nice guy at the same time (!), The paleo-group for something, I am sure, The floor ball gang for providing a welcome chance to get rid of some aggressions. I cannot imagine what Tuesdays (and the thesis) would have been like without them, Second Hand Band for "sex, drugs and rock & roll" (I think we must have missed the first part, then we had tea, and then we played some good old music ), If I have forgotten anybody who, of course, should have been thanked, I blame it on the state of exhaustion that characterises people at this stage of the thesis. However, to correct for my embarrasing omission, please fill in your name on the dotted line: I especially thank... I would also like to extend my gratitude to the Romsdal family in Norway who have created a warm and friendly atmosphere during long field seasons, and of course to all those field assistants who have contributed to this thesis over the years. Finally, Elisabeth, Hanna & Sofia. What would life have been without you? Thank you for your never ending love, support and understanding. This thesis was financially supported by The Swedish Natural Science Research Council (through grants to Lauri Oksanen), The 33

34 Royal Swedish Academy of Sciences, J. C. Kempe Memorial Fund, and R. & G. Björkman's Fund for Botanical Research. REFERENCES Andersson, M. & Jonasson, S Rodent cycles in relation to food resources on an alpine heath. Oikos 46: Andrewartha, H. G. & Birch, L. C The Distribution and Abundance of Animals. Univ. of Chicago Press, Chicago. Barkman, J. J New systems of plant growth forms and phenological plant types. In: Werger, M. J. A., van der Aart, P. J. M., During, H. J. & Verhoeven, J. T. A. (eds.) Plant Form and Vegetation Structure. SPB Academic Publishing bv, The Hague, pp Batzli, G The influence of grazers on tundra vegetation and soils. In: Proceedings of Cicumpolar Conference on Northern Ecology, Ottawa. National Research Council of Canada, Ottawa, pp Batzli, G. O., White, R. G., MacLean, S. F., Jr., Pitelka, F. A. & Collier, B. D The herbivore-based trophic system. In: Brown, J., Miller, P. C., Tieszen, L. L. & Bunnell, F. L. (eds.) An Arctic Ecosystem. The Coastal Tundra at Barrow, Alaska. Dowden, Hutchinson & Ross, Stroudsburg, pp Begon, M., Harper, J. L. & Townsend, C. R Ecology. Blackwell, Oxford. Billings, W. D Constraints to plant growth, reproduction, and establishment in arctic environments. ArcL Alp. Res. 19: Billings, W. D. & Mooney, H. A The ecology of arctic and alpine plants. Biol. Rev. 43: Cajander, A. K Über Waldtypen. Acta For. Fenn. 1: Callaghan, T. V Plant population processes in arctic and boreal regions. Ecol. Bull. 38: Callaghan, T. V Physiological and demographic implications of modular construction in cold environments. In: Davy, A. J., Hutchings, M. J. & Watkinson, A. R. (eds.). Plant Population Ecology. Blackwell, Oxford, pp Callaghan, T. V. & Collins, N. J Strategies of growth and population dynamics of tundra plants. 1. Introduction. Oikos 27: Callaghan, T. V. & Emanuelsson, U Population structure and processes of tundra plants and vegetation. In: White, J. (ed.). The Population Structure of Vegetation. Junk, Dordrecht, pp. 34

35 Carlsson, B. Å. & Callaghan, T. V Positive plant interactions in tundra vegetation and the importance of shelter. J. Ecol. 79: Chapin, F. S., III Direct and indirect effects of temperature on arctic plants. Polar Biol. 2: Chapin, F. S., III Environmental controls over growth of tundra plants. Ecol. Bull. 38: Crawford, R. M. M Studies in Plant Survival. Blackwell, Oxford. 296 pp. Dahl, E Rondane. Det Norske Videnskaps-akademi i Oslo. I. Mat.- Naturv. Klasse. No pp. DeAngelis, D. L Dynamics of Nutrient Cycling and Food Webs. Chapman & Hall, London. 270 pp. Dunson, W. A. & Travis, J The role of abiotic factors in community organization. Am. Nat. 138: Fretwell, S. D The regulation of plant communities by the food chains exploiting them. Persp. Biol. Med. 20: Fretwell, S. D Food chain dynamics: the central theory of ecology? Oikos 50: Fries, T. C. E Botanische Untersuchungen im nördlichsten Schweden. Vetensk. o. prakt, undersökn. i Lappi. Flora & Fauna 2. Gjærevoll, O The Plant Communities of the Scandinavian Alpine Snow-beds. Det Kgl. Norske Videnskabers Selskabs Skrifter. Nr pp. Griggs, R. F Competition and succession on a Rocky Mountain fellfield. Ecol. 37:8-20. Grime, J. P Plant Strategies and Vegetation Processes. Wiley, Chichester. 222 pp. Gurevitch, J Competition and the local distribution of the grass Stipa neomexicana. Ecol. 67: Hairston, N. G., Smith, F. E. & Slobodkin, L. B Community structure, population control, and competition. Am. Nat. 94: Hansson, L. & Henttonen, H Gradients in density variations of small rodents: the importance of latitude and snow cover. Oecol. 67: Hunter, M. D. & Price, P. W Playing chutes and ladders: heterogeneity and the relative roles of bottom-up and top-down forces in natural communities. Ecol. 73: Huntly, N Influence of refuging consumers (pikas: Ochotona princeps) on subalpine vegetation. Ecol. 68: Jonasson, S Nutrient content and dynamics in north Swedish shrub tundra. Holarct. Ecol. 6: Jonasson, S Influence of frost heaving on soil chemistry and on the distribution of plant growth forms. Geogr. Annaler 68A:

36 Jonasson, S Plant responses to fertilization and species removal in tundra related to community structure and clonality. Oikos 63: Jonasson, S. & Sköld, S. E Influences of frost-heaving on vegetation and nutrient regime of polygon-patterned ground ground. Vegetation 53: Keddy, P. A Competition. Chapman & Hall, London. 202 pp. Leader-Williams, N Reindeer on South Georgia. Cambridge Univ. Press, Cambridge. Moen, J. & Oksanen, L Ecosystem trends. Nature 353:510. Murdoch, W. W "Community structure, population control, and competition" - a critique. Am. Nat. 100: Nordhagen, R Die Vegetation und Flora des Sylenegebietes I, II. Skr. utg. av D. Norske Vid.-akad. I. Matem.-naturv. Klasse. Oksanen, L Abundance relationships between competitive and grazingtolerant plants in productivity gradients on Fennoscandian mountains. Ann. Bot. Fenn. 17: Oksanen, L Trophic exploitation and arctic phytomass patterns. Am. Nat 122: Oksanen, L Ecosystem organization: mutualism and cybernetics or plain Darwinian struggle for existence? Am. Nat. 131: Oksanen, L. 1990a. Exploitation ecosystems in seasonal environments. Oikos 57: Oksanen, L. 1990b. Predation, herbivory, and plant strategies along gradients of primary productivity. In: Grace, J. B. & Tilman, D. (eds.). Perspectives on Plant Competition. Acad. Press, San Diego, pp Oksanen, L. 1991a. A century of community ecology: how much progress? Trends Ecol. Evol. 6: Oksanen, L. 1991b. Trophic levels and trophic dynamics: a consensus emerging? Trends Ecol. Evol. 6: Oksanen, L Evolution of exploitation ecosystems I. Predation, foraging ecology and population dynamics in herbivores. Evol. Ecol. 6: Oksanen, L. & Oksanen, T Lemmings (Lemmus lemmus) and greysided voles (Clethrionomys rufocanus) in interaction with their resources and predators on Finnmarksvidda, northern Norway. Rep. Kevo SubarcL Res. Stat. 17:7-31. Oksanen, L. & Oksanen, T Long-term microtine dynamics in north Fennoscandian tundra: the vole cycle and the lemming chaos. Ecography 15: Oksanen, L. & Ranta, E Plant strategies along mountain vegetation gradients: a test of two theories. J. Veg. Sci. 3:

37 Oksanen, L., Moen, J. & Lundberg, P. A The time-scale problem in exploiter-victim models: does the solution lie in ratio-dependent exploitation? Am. Nat 140: Oksanen, L., Fretwell, S. D., Arruda, J. & Niemelä, P Exploitation ecosystems in gradients of primary productivity. Am. Nat 118: Oksanen, T Exploitation ecosystems in heterogeneous habitat complexes. Evol. Ecol. 4: Oksanen, T., Oksanen, L. & Gyllenberg, M Exploitation ecosystems in heterogeneous habitat complexes II: impact of small-scale heterogeneity on predator-prey dynamics. Evol. Ecol. 6: Peters, R. H A Critique for Ecology. Cambridge Univ. Press, Cambridge. 366 pp. Polis, G. A. Food webs, trophic cascades and community structure. Aust. J. Ecol. (in press). Polis, G. A Complex trophic interactions in deserts: an empirical critique of food web theory. Am. Nat 138: Power, M. E Top-down and bottom-up forces in food webs: do plants have primacy? Ecol. 73: Reader, R. J Herbivory, competition, plant mortality and reproduction on a topographic gradient in an abandoned pasture. Oikos 65: Resit Akçakaya, H Population cycles of mammals: evidence for a ratiodependent predation hypothesis. Ecol. Monogr. 62: Rune, O Plant Life on Serpentine and Related Rocks in the North of Sweden. Acta Phytogeogr. Suec. 31. Samuelsson, G Studien über die Vegetation der Hochgebirgsgegenden von Dalame. Nova Acta Reg. Soc. Scient Ups. Ser. 4. no. 8. Slobodkin, L. B., Smith, F. E. & Hairston, N. G Regulation in terrestrial ecosystems and the implied balance of nature. Am. Nat 101: Strong, D. R Are trophic cascades all wet? Differentiation and donorcontrol in speciose ecosystems. Ecol. 73: Tilman, D Plant Strategies and the Dynamics and Structure of Plant Communities. Princeton Univ. Press, Princeton. 360 pp. Vestergren, T Om den olikformiga snöbetäckningens inflytande på vegetationen i Sarekfjällen. Bot Not Weiden, C. W. & Slauson, W. L The intensity of competition versus its importance: an overlooked distinction and some implications. Quart Rev. Biol. 61: Wijk, S Salix herbacea and the alpine snow-bed environment. Ph. D. thesis, University of Lund, Sweden. 37