UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Components of regulation of boreal forest understory vegetation : a text of fertilizer and herbivory Dlott, Franklin 1996

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

Item Metadata

Download

Media
831-ubc_1996-0121.pdf [ 3.31MB ]
Metadata
JSON: 831-1.0087064.json
JSON-LD: 831-1.0087064-ld.json
RDF/XML (Pretty): 831-1.0087064-rdf.xml
RDF/JSON: 831-1.0087064-rdf.json
Turtle: 831-1.0087064-turtle.txt
N-Triples: 831-1.0087064-rdf-ntriples.txt
Original Record: 831-1.0087064-source.json
Full Text
831-1.0087064-fulltext.txt
Citation
831-1.0087064.ris

Full Text

C O M P O N E N T S OF R E G U L A T I O N O F BOREAL FOREST UNDERSTORY VEGETATION: A TEST OF FERTILIZERA N D H E R B I V O R Y  by FRANKLIN DLOTT B A . , Biology, University of California at Santa Cruz, 1990 B.A., Linguistics, University of California at Santa Cruz, 1990  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE MASTER OF SCIENCE DEGREE in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF BOTANY  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA February 1996 © Franklin K. Dlott, 1996  In  presenting  degree freely  at  the  available  copying  of  department publication  this  thesis  in  partial  University  of  British  for  this or of  reference  thesis by  this  for  his thesis  and  or  her  for  DE-6 (2/88)  )0  of  F t  k  Columbia, I  study.  I further  representatives.  financial  The University of British Columbia Vancouver, Canada  Date  of  the  ftl(  0  gain  shall  requirements  agree  that  agree  scholarly purposes may  permission.  Department  fulfilment  It not  be is be  that  the  for  an  Library shall  permission for  granted  by  understood allowed  advanced  the that  without  make  it  extensive  head  of  copying my  my or  written  ABSTRACT  This study tests the predictions of two different hypotheses of trophic organization the 'bottom-up' and 'top-down' hypotheses respectively u s i n g plants i n the boreal forest understory. The experiment manipulated plant resource levels by fertilization and consumer levels (vertebrate herbivory rate) using exclosures, and monitored the response of transplanted seedlings and the leaf area of the established vegetation. Survival and growth transplants was poorest at the highest fertilizer levels; a result not predicted by either 'bottomup' or 'top-down'. Herbivore exclosures had no significant effects on survival or g r o w t h at l o w or moderate herbivore densities; fertilizer addition d i d not increase leaf area. These results suggest that the resources added by fertilization and that the herbivores excluded were not limiting at these herbivore densities.  A t h i g h herbivore densities transplant survival and growth was consistently greater inside exclosures w h i c h lends support to the 'top-down' hypothesis for seedling survival and performance, but leaf area d i d not significantly respond to either treatment but was greater inside exclosures especially w h e n fertilized. A model of trophic relations i n the boreal forest between understory plants, their resources and their consumers should only include herbivores as a limiting factor when their densities are abnormally h i g h and even then the 'top-down' hypothesis is supported only from transplant data and not from existing vegetation.  ii  TABLE OF CONTENTS ABSTRACT  ii  TABLE OF CONTENTS  iii  LIST OF TABLES  v  LIST OF FIGURES  vi  ACKNOWLEDGMENTS  xi  INTRODUCTION  .  MATERIAL A N D METHODS  1 .  ,  6  Study area  6  Experimental sites and treatment histories  8  Experimental design  12  Data collection  15  RESULTS  19  A. SEEDLING SURVIVAL  19  1. Fertilizer 1 Grid (low herbivore densities)  19  2. Food 1 Grid (moderate herbivore densities)  21  3. Predator Exclosure + Food Grid (high herbivore densities)  23  B. SEEDLING GROWTH  25  1. Fertilizer 1 Grid (low herbivore densities)  25  2. Food 1 Grid (moderate herbivore densities)  31  3. Predator Exclosure + Food Grid (high herbivore densities)  37  C. LEAF A R E A INDEX  43  1. Fertilizer 1 grid (low herbivore densities)  43  2. Food 1 grid (moderate herbivore densities)  46  3. Predator Exclosure + Food grid (high herbivore densities)...49 D. SUMMARY OF RESULTS  52  iii  DISCUSSION  54  A. FERTILIZER ADDITION  55  B. HERBIVORE EXCLOSURES  60  C. FIELD SITES  64  1. Fertilizer 1 Grid  64  1. Food 1 Grid  65  3. Predator Exclosure + Food Grid  66  D. TROPHIC CONTROL A N D THE SNOWSHOE H A R E CYCLE BIBLIOGRAPHY  67 71  iv  LIST OF TABLES Table 1. Principal species and simplified trophic structure i n the herbaceous understory of the boreal forest  4  Table 2. Predicted changes i n plant biomass resulting from fertilizer addition and herbivore exclosure for the resource regulation hypothesis and the consumer regulation hypothesis  4  Table 3. Fertilizer 1 grid treatment history  9  Table 4. Species lists by groupings from the three study sites i n descending order of frequency  16  Table 5. Summary of A N O V A for transplant survival per subplot on (A) August 6,1992, (B) July 21, 1993, Fertilizer 1 G r i d 20 Table 6. Summary of A N O V A for transplant survival per subplot i n (A) August 1992, (B) July 1993, Food 1 G r i d  22  Table 7. Summary of A N O V A for transplant survival per subplot on (A) August 6,1992, (B) July 21, 1993, Predator Exclosure + Food Grid 24 Table 8. Summary of A N O V A for (A) number of transplanted grass tillers per subplot (square root transformed) and (B) mean height per transplanted grass on August 6, 1992, Fertilizer 1 grid  26  Table 9. Summary of A N O V A for (A) number of transplanted legume leaves per subplot and (B) mean height per legume transplant for August 1992, Fertilizer 1 grid  27  Table 10. Summary of A N O V A for (A) number of transplanted grass tillers per subplot and (B) mean height per grass transplant o n July 21, 1993, Fertilizer 1 grid  29  Table 11. Summary of A N O V A for (A) number of transplanted legume leaves per subplot (square root transformed) and (B) mean height per legume transplant on July 21, 1993, Fertilizer 1 grid  30  Table 12. Summary of A N O V A for (A) number of transplanted grass tillers per subplot (square root transformed) and (B) mean height per transplanted grass on August 6, 1992, Food 1 grid  32  v  Table 13. Summary of A N O V A for (A) number of transplanted legume leaves per subplot and (B) mean height per legume transplant on August 6,1992, Food 1 grid  33  Table 14. Summary of A N O V A for (A) number of transplanted grass tillers per subplot and (B) mean height per grass transplant on July 21,1993, Food 1 grid  35  Table 15. Summary of A N O V A for (A) number of transplanted legume leaves per subplot (square root transformed) and (B) mean height per legume transplant on July 1993, Food 1 grid  36  Table 16. Summary of A N O V A for (A) number of transplanted grass tillers per subplot (square root transformed) and (B) mean height per transplanted grass on August 6, 1992, Predator Exclosure + Food grid  38  Table 17. Summary of A N O V A for (A) number of transplanted legume leaves per subplot (square root transformed) and (B) mean height per legume transplant on August 6, 1992, Predator exclosure + food grid  39  Table 18. Summary of A N O V A for (A) number of transplanted grass tillers (square root transformed) per subplot and (B) mean height per grass transplant (square root transformed) on July 21, 1993, Predator Enclosure + Food grid 41 Table 19. Summary of (A) Kruskal Wallis for number of transplanted legume leaves per subplot and (B) A N O V A of mean height per legume transplant on July 21, 1993, Predator Exclosure + Food grid 42 Table 20. Summary of A N O V A for (A) total, (B) graminoid and (C) dicot (square root transformed) leaf area, July 21, 1993, Fertilizer 1 grid  44  Table 21. Summary of A N O V A on species richness on 21 July 1993, Fertilizer 1 grid  45  Table 22. Summary of A N O V A for (A) total, (B) graminoid (square root transformed) and (C) dicot leaf area, July 21,1993, Food 1 grid  47  Table 23. Summary of A N O V A on species richness on 21 July 1993, Food 1 grid  48  vi  Table 24. Summary of A N O V A for (A) total, (B) graminoid (square root transformed) and, (C) dicot (log transformed) leaf area from July 1993, Predator Exclosure + Food grid 50 Table 25. Summary of A N O V A on species richness on 21 July 1993, Predator Exclosure + Food grid  51  Table 26. Summary of the effect of (A) fertilizer addition and (B) herbivore exclosures on the growth and survival of transplants and the leaf area of established vegetation at three areas of . known herbivore densities 53  vii  LIST OF FIGURES Figure 1. M a p of north-western N o r t h America showing the location of the of study site  6  Figure 2. Location of study areas (Fertilizer 1, Food 1, and Predator Exclosure + Food) along the Alaska highway  9  Figure 3. Population estimates for (A) snowshoe hares, (B) ground squirrels and (C) mice (Clethrionomys and Microtus) at three sites w i t h different treatment histories since 1987  10  Figure 4. Schematic layout of sites, plots and subplots within a grid  12  Figure 5. Percentage survival of transplanted seedlings at the Fertilizer 1 grid  19  Figure 6. The percentage (±SE) of seedlings alive on (A), August 6,1992 and (B), July 21, 1993 at the Fertilizer 1 grid  20  Figure 7. Percentage survival of transplanted seedlings at the Food 1 grid  21  Figure 8. The percentage (+SE) of seedlings alive on (A), A u g u s t 6, 1992 and (B) July 21,1993 at the Food 1 grid  22  Figure 9. Percentage survival of transplanted seedlings at the Fertilizer 1 grid  23  Figure 10. The percentage of seedlings alive (±SE) on (A), August 6, 1992 and (B) July 21,1993 at the Predator Exclosure + Food grid  24  Figure 12. (A) N u m b e r (±SE) of grass tillers per subplot and (B) mean (±SE) height of grass plants on August 6, 1992 at the Fertilizer 1 grid  26  Figure 13. (A) N u m b e r (±SE) of legume leaves per subplot and (B) mean (±SE) height of legume plants on August 6, 1992 at the Fertilizer 1 grid 27 Figure 14. (A) N u m b e r (±SE) of grass tillers per subplot and (B) mean (+SE) height of grass plants o n July 21, 1993 at the Fertilizer 1 grid  viii  29  Figure 15. (A) N u m b e r (+SE) of legume leaves per subplot and (B) mean (+SE) height of legume plants on July 21, 1993 at the Fertilizer 1 grid 30 Figure 16. (A) N u m b e r (±SE) of grass tillers per subplot and (B) mean (±SE) height of grass plants on August 6, 1992 at the Food 1 grid. ..32 Figure 17. (A) N u m b e r (+SE) of legume leaves per subplot and (B) mean (±SE) height of legume plants on August 6, 1992 at the Food 1 grid  33  Figure 18. (A) N u m b e r (±SE) of grass tillers per subplot and (B) mean (+SE) height of grass plants on July 21, 1993 at the Food 1 grid  35  Figure 19. (A) N u m b e r (+SE) of legume leaves per subplot and (B) mean (±SE) height of legume plants on July 21, 1993 at the Food 1 grid  36  Figure 20. (A) N u m b e r (±SE) of grass tillers per subplot and (B) mean (+SE) height of grass plants on August 6, 1992 at the Predator Exclosure + Food grid  38  Figure 21. (A) N u m b e r (±SE) of legume leaves per subplot and (B) mean (+SE) height of legume plants on August 6, 1992 at the Predator Exclosure + Food grid  39  Figure 22. (A) N u m b e r (+SE) of grass tillers per subplot and (B) mean (±SE) height of grass plants on July 21, 1993 at the Predator Exclosure + Food grid  41  Figure 23. (A) N u m b e r (±SE) of legume leaves per subplot and (B) mean (±SE) height of legume plants on July 6, 1993 at the Predator Exclosure + Food grid  42  Figure 24. (A) total, (B) graminoid, and (C) dicot leaf area index o n July 21, 1993, i n the Fertilizer 1 grid Figure 25. Species richness (±SE) at the Fertilizer 1 grid, on 21 July 1993  44 45  Figure 26. (A) total, (B) graminoid, and (C) dicot leaf area index on July 21,1993 i n the Food 1 grid  47  Figure 27. Species richness (±SE) at the Food 1 grid, on July 21,1993  48  Figure 28. (A) total, (B) graminoid, and (C) dicot leaf area index on July 21, 1993, i n the Predator Exclosure + Food grid ix  50  Figure 29. Species richness (±SE) at the Predator Exclosure + Food, on July 21,1993 51 Figure 30. Theoretical changes in plant damage due to herbivory as herbivore density and exclosure efficacy vary  63  Figure 31. Theoretical changes in plant damage due to herbivory as herbivore density, exclosure efficacy and fertilizer addition vary  63  Figure 32. Predictions from the Oksanen et al. (1987) hypothesis showing the direction and magnitude of trophic regulation for ecosystems with from 1 to 3 trophic levels 69  ACKNOWLEDGMENTS  I am very grateful to m y supervisor, D r . Roy Turkington, for his insight, persistence and editorial help. I w o u l d also like to thank m y other committee members, D r . Gary Bradfield and D r . Jack Maze for their support and advice. D r . A . R. E . Sinclair also deserves special thanks for his insights into the inception of the w o r k and for theoretical contributions.  M a n y others have p r o v i d e d useful comments that have led to the completion of this project. I w o u l d like to thank Ji Yong, Stephanie Graham, Jessica Bratty and Jocylyn M c D o w e l l for their help i n the field. Lauch Fraser, A n d r e a B y r o m , and K e n A r i i have all been patient listeners and p r o v i d e d useful comments.  A n d y and Carol Williams helped make m y stay at the Arctic Institute more comfortable and enjoyable and M a r k O'Donoghue came through w i t h added personnel and vehicles when desperately needed.  M y family has been a continuing source of support and I am very grateful to all of them. A heartfelt thanks goes to m y wife, Selena, and son, H a v e n , w h o have p r o v i d e d me w i t h support, understanding, and a healthy perspective throughout this project.  xi  1 INTRODUCTION  The relationships between plants, their resources, and the herbivores that feed u p o n them are complex and challenging to explain, and community biologists have long been interested i n h o w they affect one another. A topic of considerable debate is whether herbivores or plant resources are responsible for the regulation of growth and abundance of plants.  There are at least two m a i n viewpoints concerning the ways i n w h i c h trophic levels exert regulatory effects — the so-called 'bottom-up' and 'topd o w n ' hypotheses.  Proponents of 'bottom-up' regulation argue that resource  abundance determines the biomass of the trophic level using those resources. A strict 'bottom-up' position argues that resources alone regulate growth and abundance of the next higher trophic level (White 1978, 1984). This v i e w predicts that plant biomass responds strongly to resources but is not influenced at all b y herbivores. The opposite view is taken by a strict 'top-down' position where the biomass of a trophic level is regulated only by its consumers (Menge and Sutherland 1976). Resources are not predicted to regulate biomass i n the strict 'top-down' hypothesis. In this view, herbivores remove plant material to such a degree that relatively small changes i n plant resources do not change plant biomass. Nutrients alter the productivity of the system under strict predator control and plants grow faster, but there is no increase i n standing crop because the herbivores remove the added plant productivity. Most researchers adopt an intermediate position, recognizing that the hypotheses i n their strictest form account for only a few cases. Hunter and Price (1992) provide a synthetic hypothesis w h i c h allows for both resource regulation and consumer regulation under certain conditions. The debate about whether predators could regulate the trophic level below began w i t h Hairston, Smith and Slobodkin (1960) w h o  2 argued that predators regulate their herbivore prey, and because of that, herbivores cannot regulate their plant prey. Synthetic views allow for more realism than strict views, but synthetic views are very difficult to test by field experiments because of their multi-factorial nature.  The terms 'resource regulation' and 'consumer regulation' will be substituted for 'bottom-up' and 'top-down' respectively because these terms have come to mean different things to different workers (e.g., Carpenter and Kitchell 1987, Northcote 1988, Oksanen 1988, McQueen et al. 1989). The resource regulation and consumer regulation hypotheses will be tested by comparing their predictions with the results obtained from experiments using understory herbaceous plants in the boreal forest.  This work was conducted using some of the facilities of a long term project called the Kluane Boreal Forest Ecosystem Project (KBFEP) headed by Dr. C. J. Krebs and done in the southwestern Yukon, Canada. The KBFEP began in 1987 and has manipulated all trophic levels in order to understand trophic structure and linkage strengths in the boreal forest (Boutin 1990). Various largescale community perturbation experiments (40-100ha) were applied to increase nutrient availability to plants, decrease herbivory, increase herbivore food supply, and quality, and decrease predation. These treatments provide an opportunity to test the predictions of the resource regulation and consumer regulation hypotheses, because they have generated a range of soil nutrient concentrations, herbivore and predator densities.  The experiments done for this work used a factorial design which crossed three fertilizer addition rates with ± herbivore exclosures to provide a test of the resource regulation and consumer regulation hypotheses. N-P-K fertilizer was  3 added to change the concentrations of these potentially limiting resources and herbivore exclosures reduced vertebrate herbivory. The same factorial design was repeated at three areas with different herbivore densities resulting from the KBFEP treatments. The fate of transplanted seedlings in their first and second year of growth provided a fine-scale measure of the relative impact of nutrients and herbivore exclosures. Leaf area index was recorded in the second year to provide a coarser, community viewpoint within the same factorial design. Results taken from the existing vegetation and the transplanted seedlings were compared with emphasis on phenology and scale.. Both fine- and coarse-scales were examined because seedlings are more sensitive to herbivory than established perennials (Cavers and Harper 1967).  Objectives and predictions  The primary objective of this study was to determine how fertilizer additions and mammalian herbivore exclosures would influence transplant survival and growth, and the existing vegetation. The impact of herbivores often varies with their density (Crawley 1989) and so these experiments also link the magnitude of plant response to herbivore density. A simplified trophic pyramid shows the principal species involved (Table 1). In the context of the two trophic arguments outlined above, different predictions can be made about the effects of fertilization and herbivore exclusion (Table 2).  4 Table 1. Principal species and simplified trophic structure in the herbaceous understory of the boreal forest. lynx, coyote, great horned owl, hawk owl, goshawk  PREDATORS  Jsnowshoe hare, arctic ground squirrel, red-backed vole, meadow vole  HERBIVORES  Picea glauca (spruce), Populus tremuloides (poplar), Salix  glauca (shrub willow), Betula nana (dwarf birch)  CANOPY PLANTS  Festuca altaica (fescue), Calamagrostis lapponica (reedgrass), Achillea  millefolium  (yarrow), Linnea boreale (twin flower),  Lupinus arcticus (lupine), Salix myrtillifolia  (ground willow),  Stellaria longipes (starwort), Carex sp. (sedges), Epilobium angustifolium  (fireweed), Anemone parviflora  (Anemone)  UNDERSTORY PLANTS  Table 2. Predicted changes in plant biomass resulting from fertilizer addition and herbivore exclosure for the resource regulation hypothesis and the consumer regulation hypothesis. A (+) predicts greater plant biomass than untreated and a (0) predicts no difference from control. FERTILIZER MANIPULATION untreated untreated added added  HERBIVORE MANIPULATION present excluded present excluded  RESOURCE REGULATION control 0 + +  CONSUMER REGULATION control + 0 +  5  The strict resource regulation hypothesis predicts that herbivore exclosures will have no effect on plant biomass, because this hypothesis argues that herbivores do not limit plant production or standing crop. Plants receiving fertilizer are predicted to have increased productivity and show an increase in standing crop. Transplants in fertilized plots are predicted to have more leaves and grow taller than plants in unfertilized plots. Leaf area is also predicted to increase.  The strict consumer regulation hypothesis predicts that herbivore exclosures result in increased plant biomass, because herbivores limit the standing crop of plants. Fertilizer additions will increase productivity but the increased productivity results in no increase in plant biomass because herbivores remove it. In plots without herbivores, increased productivity is predicted to accumulate as plant biomass. Transplants exposed to herbivores are predicted to be smaller and have poorer survival compared to protected transplants. When herbivores are excluded fertilizer is predicted to increase growth and biomass but if herbivores are present, fertilizer would have no effect.  6 MATERIAL A N D M E T H O D S Study area  The study area is i n a broad glacial valley located near Kluane Lake i n the South-west Y u k o n (60° 5 7 ' N 138° 12'W at an elevation of 800-lOOOm) (Fig. 1). A detailed description of the flora and geology is provided by Douglas (1974). The climate is cold continental; i n Haines Junction 60 k m to the south mean summer temperatures are 10.5°C i n June, 12.4°C i n July and 10.4°C i n A u g u s t (Webber 1973) and mean annual precipitation is 28.2 cm w i t h most falling as snow. The short growing season is from m i d - M a y to late A u g u s t w i t h snow cover from October through M a y (Krebs et al. 1986).  Figure 1. M a p of the pacific north-west of N o r t h America showing the location of the of study site.  7 Soil profiles have three distinct layers: an organic layer, l-3cm deep, a layer m i x i n g the organic and inorganic layers, 5-15cm deep, and an inorganic layer p r i m a r i l y a silty loam. Permafrost depth i n June varies between 50-100cm. Soils overlie and are derived from alluvial gravel moraines and deposits of glacial till and loess, all the result of extensive glaciation. Soils are nutrient poor because the l o w temperatures, l o w precipitation, and short g r o w i n g season limit the biological activity of soil microbes.  Douglas (1974) classed the region as a closed to open spruce (Picea glauca Voss) forest community. O p e n sites can have a well-developed shrub layer i n c l u d i n g dwarf birch (Betula glandulosa Michx), shrub w i l l o w (Salix gluaca L.), and soapberry (Sheperdia canadensis (L.) Nutt.). Understory plants include fescue (Festuca altaica var. scabrella Torr), g r o u n d w i l l o w (Salix myrtillifolia  Anderss.), bear berry (Arctostaphylos uvi-ursa (L.) Spreng.), reedgrass (Calamagrostis lapponica (Whahenb.) Hartm.), y a r r o w (Achillea  millefolium  L.), arctic lupine (Lupinus arcticus Lindl.), sedges (Carex sp.), and twin-flower > (Linnea borealis L.). Soil m i x i n g is slow and so the effects of fire history, glacial deposits, tree fall, and digging by bears and ground squirrels create long lasting soil patchiness.  Soil variation generates variation i n plant species distribution,  composition, and abundance.  The m a i n vertebrate herbivores are snowshoe hares (Lepus americanus) arctic ground squirrels (Spermophilus parryi) and other small rodents (Clethrionomys rultilus and Peromyscus maniculatus).  These herbivores are i n  turn preyed upon by lynx, coyotes and raptors (great horned owls, goshawks).  8 E x p e r i m e n t a l sites a n d t r e a t m e n t h i s t o r i e s  The experiments were conducted w i t h i n three of the treatment grids imposed b y the Kluane Boreal Forest Ecosystem Project (KBFEP) (Boutin 1990). Treatments were applied to these grids (i) to increase nutrient availability to plants (Fertilizer 1), (ii) to increase herbivore food supply, and quality (Food 1; rabbit chow) and (iii) to decrease predation and increase herbivore food supply and quality (Predator Exclosure + Food). These grids used were k n o w n locally as Flint, Gravel Pit and H u n g r y Lake respectively (Fig. 2). Fertilizer 1 and Food 1 were set up i n 1987 and Predator Exclosure + Food i n 1988. Fertilizer 1 had annual applications of fertilizer but the application rates varied (Table 3). Food 1 had rabbit chow (Shur-gain, 16% protein) provided year-round on a 40ha hare trapping grid. The Predator Exclosure + Food grid is surrounded by a 1km x 1km 10-strand electric fence 2m high lined w i t h 5cm mesh wire. The fence allows hares and other small mammals to pass but excludes bears, lynx, coyotes and moose. W i t h i n the predator exclosure, commercial rabbit chow was provided year round on a 40ha hare trapping grid. Population densities of snowshoe hares, arctic ground squirrels, mice, and voles were monitored i n response to large scale treatments (Fig. 3). Hare and squirrel populations were consistently lowest at Fertilizer 1 grid, intermediate at the Food 1, and highest at the Predator exclosure + food grid.  Mice and vole populations had the same  pattern as hares and squirrels i n 1992, but i n 1993 the Fertilizer 1 grid had the highest mouse population.  Figure 2. Location of study areas (Fertilizer 1, Food 1, and Predator Exclosure + Food) along the A l a s k a highway.  Table 3. Fertilizer 1 grid treatment history. In 1987 fertilizer was hand spread onto 40ha, after that fertilizer was applied by air onto lOOha. YEAR 1987  TREATMENT 25 g N . m 2 (NH )2N0  1988  I7.5g N . m 2  (NH )2N0  5g  PO4-  4  P.m  2  2.5g K . m half of 1988 rate 1988 rate half of 1988 rate no application 1988 rate 2  1989 1990 1991 1992 1993  4  K+  3  3  3  10  •  Food 1  HI  Predator Exclosure + food  Figure 3. Population estimates for (A) snowshoe hares (O'Donaguhe, pers. com.) , (B) ground squirrels (Byrom, pers. com.) and (C) mice (Boonstra, pers. com.) (Clethrionomys and Microtus) at three sites with different treatment histories since 1987. Bars are 95% confidence interval; there was no variation estimate for mice and numbers represent minimum number alive.  At each of the three grids, two sites were chosen to be within 3m of an active ground squirrel burrow system and at least 3m away from the tree canopy. Active burrows were seen at the predator exclosure + food grid, but at Fertilizer 1 and Food 1, squirrels were less common. Consequently, site selection on these grids was based on the location of previously trapped squirrels (O'Donoghue, personal communication). Sites were 10m apart at Fertilizer 1 (KBFEP grid reference points E4 and D3), and Food 1 (KBFEP grid reference points M9 and N8) and 50m apart at the Predator exclosure + food grid (KBFEP grid reference points H4 and J9). The sites were classified using the KBFEP habitat classification protocol (KBFEP handbook of field procedures,  11 1992). The sites at Fertilizer 1 were characterized by open, mature spruce overstory, an open birch shrub cover and a very slight southwest aspect. The sites at Food 1 had a scattered mature spruce overstory, an open birch shrub layer, and were within the bottom of a shallow, approximately 10m deep, eastwest gully. The sites at the Predator Exclosure + Food were different from each other. The site at grid reference H4 had no tree overstory, an open willow shrub layer, some standing dead and a moderate north slope whereas the site at grid reference J9 had an open immature spruce overstory with an open birch shrub layer and no slope.  12 Experimental design  Simultaneous experiments were conducted at the three KBFEP treatment grids described above. At each site, three plots were marked out, and each plot was subdivided to form four subplots, each 0.75m x 0.75m. This permitted a factorial design with three levels of fertilizer addition, and ± herbivore exclosure and four replicates (Fig. 4).  1.5m  Figure 4. Schematic layout of sites, plots and subplots within a grid: (a) KBFEP grid, (b) site, (c) plot, and (d) subplot. Exclosed (fenced) subplots are solid and open subplots are dashed lines.  At each site, three fertilizer treatments were randomly assigned to the three 1.5 x 1.5m plots. I used 35-10-5 fertilizer which was the same as that used by KBFEP in the 1991 fertilization. The low fertilizer treatment was 11.6gN.m-2 as ammonium nitrate ((NH4)2N03), 3.3g.m-2 super phosphate (H2PO4) and 1.6gm-2 potash (K2O) and the high fertilizer treatment was 35g.N.m-2, 10g.m-2  13 super phosphate, and 5g.m 2 potash. The third treatment was a control w i t h no -  added fertilizer. These rates are two-thirds (low), and twice (high) the K B F E P standard rate applied i n 1988. Fertilizer was manually broadcast on the 14th and 29th of June, 1992 and the 3rd and 24th of June, 1993. The fertilizer treatment on the Fertilizer 1 grid was i n addition to the usual K B F E P fertilization i n 1993.  Each 1.5 x 1.5m plot contained four 0.75m x 0.75m subplots w h i c h were each marked w i t h a wooden stake, 4cm x 4cm and 75cm high. In each plot the two opposing corner subplots were randomly assigned to two exclosure treatments, open or exclosed. Fences were erected before 9 June 1992 using 2.5cm mesh chicken w i r e and surrounded the entire plot. The initial exclosure allowed plants to establish i n the absence of herbivores. The fences surrounding the subplots chosen to be i n the open treatments were removed o n 8 July 1992 (Fig. 4).  Seedlings of eight species were g r o w n from seed i n vermiculite based potting soil b y the Arctic A l p i n e Nursery, Whitehorse, Y u k o n Territories, i n single species flats. The eight species were four grasses, Festuca altaica var. scabrella , F. ovina var. ovina L . , F. ovina var. saximontana  (L.) St.-Yves, and  Calamagrostis lapponica Whatnb., and four legumes, Lupinus arcticus L i n d l . , Hedyserum mackensii (Richards.) H i t c h c , Oxytropis campestris (L.) D C , and O .  maydelliana Trautv. By June 10, 1992, seedlings had reached the two-leaf stage and were transported to the Arctic Institute base camp at Kluane Lake.  Between June 18-20, 1992, seedlings were transplanted into the various sites. Subplots were each divided into nine squares i n a three b y three grid w i t h 20cm between rows. The middle block was not planted. In the remaining eight  14 squares a 10cm diameter circle of vegetation was hand weeded. Next, a conical opening was made w i t h a metal spike, 2cm wide x 5cm deep. In each subplot, the eight species were randomly assigned to a transplant site and this location was noted. The bare-rooted seedling was inserted into the hole and the surrounding soil pressed closed around the roots. Each transplant was given 250ml of water immediately after planting. The initial watering was supplemented w i t h 125ml of water every second day for 18 days. After a plot was planted, a clear 4mil plastic sheet was stapled to the top of the wooden fence posts to limit water loss. The plastic was removed and watering ended on 8 July 1992. Individually marked plastic rings were placed around each transplant on 24 June 1992. These rings were later supplemented w i t h individually marked cocktail sticks inserted to ground level 3cm from the transplant on 25 July 1992 because the unattached rings were moved by w i n d and herbivores. Dead transplants were replaced on 25 June and 6 July 1992.  15 Data collection  For each transplanted seedling survival, number of leaves, and plant height were measured on July 15, July 23, July 29, and August 6,1992 and on June 8, June 21, and July 21, 1993. In 1992, leaf counts were made but in 1993 because counts of the very small leaves could not be exact, counts were only made if there were less than 30 leaves. Otherwise the number of leaves was put into three classes as follows: A (30 to 50 leaves), B (51 to 70 leaves), and C (>70 leaves). Only a few of the grasses produced more than thirty leaves. The first survey in 1993 was both counted and put into classes to insure that classifying reflected actual counts.  A non-destructive estimate of cover frequency and species composition, leaf area index, was taken on June 8 and July 21, 1993. A'25cm x 25cm quadrat, with crosswires at 5cm intervals, was placed inside each subplot; this gave 25 sampling points per subplot. A long, 1mm diameter steel pin was positioned vertically so that it touched both the ground and a quadrat crosswire. The number of leaves touching the pin was recorded for each species touched. These data were grouped into graminoids and dicots (Table 4).  16 Table 4. Species lists by groupings from the three study sites i n descending order of frequency. Species w i t h less than 0.4% frequency were not included. Unidentified plants composed less than 1% of the sample. Nomenclature follows H u l t e n (1968). FERTILIZER 1  FOOD 1  G R A S S E S Calamagrostis lapponica Festuca altaica Carex sp  A N D  PREDATOR EXCLOSURE + FOOD S E D G E S  Festuca altaica Calamagrostis lapponica Carex sp.  Festuca altaica Calamagrostis lapponica Carex sp  D I C O T S Achillea millefolium Salix myrtillifolia Epilobium angustifolium Stellaria longipes Epilobium angustifolium Linnea borealis . Epilobium angustifolium Geum sp. Lupinus arcticus Penstemon procerus Achillea millefolium Achillea millefolium Solidago canadensis Penstemon procerus Anemone parviflora Antenaria rosea Potentilla fruiticosa Salix myrtillifolia Arctostaphylos uvi-ursi Rubus chamaemorus Mertensia paniculata Salix myrtillifolia Potentilla diversifolia Solidago canadensis Penstemon procerus Arctostaphylos rubra  17 Transplant survival was measured as the number of transplants, regardless of species, alive per subplot as the dependent variable w i t h fertilizer and exclosure treatments as the independent variables. Transplant growth had two components - the total number of leaves per subplot and the mean height per plant. Gomez and Gomez (1984) argue that measurements taken from dead individuals i n yield experiments should be treated as zeros rather than as missing data when mortality is due to the treatment. The analysis presented here retains dead individuals as zero values for survival, number of leaves or tillers, and height.  Transplant growth tests were done i n two groups - the grasses, (Festuca altaica, F. ovina, F. saximontana, (Lupinus  and Calamagrostis lapponica) and the legumes  arcticus, Hedyserum mackensii, Oxytropis campestris, and O .  magdellena). The three K B F E P grids were analyzed separately to test for withinsite trends but not allowing between-site tests. Survival, number of leaves or tillers, and average height were all tested using the survey taken at the end of the 1992 season (6 August 1992) and the end of the 1993 season (21 July 1993).  Analysis of variance was used to test effects using the factorial A N O V A package i n S Y S T A T (Wilkinson et al. 1992). A l l variables were tested to meet the assumptions of A N O V A and, where necessary, logio, or square root transformed as indicated i n the tables. The number of grass tillers at the Predator Exclosure + Food grid i n the 1993 season failed to meet the assumption of homoscedascity and was analyzed using the Kruskal-Wallis test i n the S Y S T A T package (Wilkinson etal. 1992).  18  Leaf area index was divided into two groups - grasses and sedges in one group and dicots in the other. The two groups, gramminoids and dicots, as well as total leaf area, were analyzed by factorial A N O V A in SYSTAT (Wilkinson et al. 1992) and either square root or logifj transformed when necessary to meet the assumption of homoscedascity.  Other statistical procedures were used but not included here. These included the log-rank test for transplant survival, log-linear tests of transplant growth characters, and multi-dimensional contingency tests. However none of these procedures yielded results different enough from the results of A N O V A to justify their use. These tests are more complicated, sometimes difficult to interpret, or sacrifice power compared to two-way A N O V A (Pyke and Thompson 1986,1987; Hutchings et al. 1991).  19  RESULTS A. SEEDLING SURVIVAL 1. Fertilizer 1 Grid (low herbivore densities) Transplant survival d i d not significantly differ among treatments i n the first or second year (Fig. 6, Table 5). Transplants receiving the treatment of no fertilizer w i t h exclosure had the best first season survival (84%), and transplants receiving either the fertilizer without exclosure treatment, or the l o w fertilizer w i t h exclosure treatment, had the poorest survival (72%). The highest survival for the second season was 50% for transplants i n the l o w fertilizer without exclosure treatment and was lowest (25%) for the unfertilized treatment without exclosure.  Fertilizer 1 (low herbivore densities) 100  r  CO > 3  CO  •M  Qi U  <u D_  10 (XI CD C  Z3 —>  o  CM  (XI CD  CD  (XI CD  "5  "5  "5  LO  CO (XI  CD CXI  —> <—  —>  ~>  C\J CD cn  00 CD c  CO CD c  CO CD  < CD  -) 00  -3  —)  i—  i—  CXI  Figure 5. Percentage survival of transplanted seedlings at the Fertilizer 1 grid. Curves are for open subplots (open symbols), exclosed (fenced) subplots (filled symbols) for three levels of nutrient addition; none ( O ), low ( A ) 11.6gN.m-2 , and high ( • ) 35gN.m-2 of 35:10:5 N P K fertilizer.  20  A. i oo r 80 h  ro >  1992  B.  a  1993  ioo r  J  80 h  60 h  60 V  fy  (/>  +J  c cu o 1_ 0)  40 h  40 h  F~1 //  a. 20 h  20 h  none  low  high  1 fy  none  low  high  Fertilizer treatment Figure 6. The percentage (+SE) of seedlings alive on (A), August 6,1992 and (B), July 21, 1993 at the Fertilizer 1 grid. Open bars represent open subplots, striped bars represent exclosed (fenced) subplots. L o w fertilizer is 11.6gN.m~2 and high is 35gN.m"2 of 35:10:5 N P K fertilizer. Fertilizer treatments that share a common letter are not significantly different (p > 0.05; Scheffe).  Table 5. Summary of A N O V A for transplant survival per subplot on (A) August 6,1992, (B) July 21, 1993, Fertilizer 1 G r i d . A.  1992 S u r v i v a l  Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 0.250 0.375 1.750 52.250  df 2 1 2 18  MS 0.125 0.375 0.875 2.903  F-ratio 0.043 0.129 0.301  P 0.958 0.723 0.743  SS 3.083 0.042 5.083 62.750  df 2 1 2 18  MS 1.542 0.042 2.542 3.386  F-ratio 0.442 0.012 0.729  P 0.649 0.914 0.496  B. 1993 S u r v i v a l Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  21  2. F o o d 1 G r i d ( m o d e r a t e h e r b i v o r e d e n s i t i e s )  Transplants i n the unfertilized exclosed treatment enjoyed 100% survival at the end of the first season and 50% at the end of the study. The poorest survival was i n the high fertilizer, open treatment, w h i c h had 65% survival i n the first  and  3% i n the second season (Fig. 7) . Transplant survival i n the high fertilizer treatment was significantly worse than i n the unfertilized treatment i n 1992 and i n 1993 (Fig. 8, Table 6). Exclosures had no significant effect on survival. Fertilizer significantly lowered transplant survival i n 1992; the unfertilized treatment had significantly higher survival than the high fertilizer treatment. For the 1993 season, transplants i n both fertilized treatments had significantly lower survival compared to those w i t h no fertilizer.  Food 1 (moderate herbivore densities) 100  r-  03 >  w +-> c  io L  CO  CJ  (XI CD  (XI CD  (XI CD  (XI CD  (XI CD  oo CD  oo CD  oo CD  c  —>  Jul  Jul  Jun  Jun  Jul  LO  Aug  CO  00  CD (XI  to  CO  r—• (XI  0\I  o  (XI  Figure 7. Percentage (XI survival of transplanted seedlings at the Food 1 grid. Curves are for open subplots (open symbols), exclosed (fenced) subplots (filled symbols) for three levels of nutrient addition; none ( O ), low ( A ) 11.6gN.m-2 , and high ( • ) 35gN.m-2 of 35:10:5 N P K fertilizer.  22 A.  1992  B.  100  I c cu o  i_ CD  100  ab b.  80h  1993  80 k  601  60f-  40 r-  40h  20 U  20 h  1  I  kLL high high Fertilizer treatment Figure 8. The percentage (±SE) of seedlings alive on (A), August 6, 1992 and (B) July 21,1993 at the Food 1 grid. Open bars represent open subplots, striped bars represent exclosed (fenced) subplots. L o w fertilizer is 11.6gN.m"2 and high is 35gN.m~2 of 35:10:5 N P K fertilizer. Fertilizer treatments that do not share a common letter are significantly different (p < 0.05; Scheffe).  oi  none  0l  low  none  low  Table 6. Summary of A N O V A for transplant survival per subplot i n (A) August 1992, (B) July 1993, Food 1 G r i d . A.  1992 Survival  Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 16.083 3.375 0.250 24.250  df 2 1 2 18  MS 8.042 3.375 0.125 1.347  F-ratio 5.969 2.505 0.093  P 0.010 0.131 0.912  SS 42.750 3.375 0.250 33.250  df 2 1 2 18  MS 21.375 3.375 0.125 1.847  F-ratio 11.571 1.827 0.068  P 0.001 0.193 0.935  B. 1993 S u r v i v a l Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  23 3. P r e d a t o r E x c l o s u r e + F o o d G r i d ( h i g h h e r b i v o r e d e n s i t i e s )  Transplants i n the unfertilized exclosed treatment had the highest survival both seasons, 90% and 72% respectively. Transplants i n the high fertilizer open treatment had the poorest survival both years, 18% and 3% (Fig. 9). Transplant survival inside exclosures was significantly higher both seasons (Fig. 10, Table 7) and significantly lower w i t h i n the high fertilizer treatment.  Figure 9. Percentage survival of transplanted seedlings at the Fertilizer 1 grid. Curves are for open subplots (open symbols), exclosed (fenced) subplots (filled symbols) for three levels of nutrient addition; none ( O ), l o w ( A ) 11.6gN.m-2 , and h i g h ( • ) 35gN.m-2 of 35:10:5 N P K fertilizer.  24 A. 1992 100  B. 1993 100  ab  r  b.  80 h  I  03  |  601if)  4->  c 40 U co o j_  CO CL  20 r-  80 h 60 40 20 h  I  low hig high Fertilizer treatment Figure 10. The percentage of seedlings alive (±SE) on (A), August 6, 1992 and (B) July 21, 1993 at the Predator Exclosure + Food grid. Open bars represent open subplots, striped bars represent exclosed (fenced) subplots. Low fertilizer is 11.6gN.m- and high is 35gN.m" of 35:10:5 NPK fertilizer. Fertilizer treatments that do not share a common letter are significantly different (p < 0.05; Scheffe). none  low  2  none  2  Table 7. Summary of A N O V A for transplant survival per subplot on (A) August 6,1992, (B) July 21, 1993, Predator Exclosure + Food Grid. A.  1992 Survival  Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 38.583 60.167 0.583 50.000  df 2 1 2 18  -MS 19.292 60.167 0.292 2.778  F-ratio 6.945 21.660 0.105  P 0.006 0.000 0.901  SS 21.583 70.042 1.083 51.250  df 2 1 2 18  MS • 10.792 70.042 0.542 2.847  F-ratio 3.790 24.600 0.190  P 0.042 0.000 0.828  B. 1993 Survival Source of variation Fertilizer Exclosure Fertilizer x exclosure Error  25 B.  SEEDLING GROWTH  1. F e r t i l i z e r 1 G r i d ( l o w h e r b i v o r e d e n s i t i e s )  The total number of grass tillers, and legume leaves per subplot and the mean height of grasses and legumes d i d not significantly differ at the end of the first season (Figs. 12,13, Tables 8, 9).  26 B.  60 r w 50 ^  E o  40  1-2  C/5  CO CC  o> 30 M—  O CD  20  -Q  E "3  10  1  U3  4h  2h  CJ)  c  1  03  Ih  none low high high Fertilizer treatment Figure 12. (A) N u m b e r (±SE) of grass tillers per subplot and (B) mean (±SE) height of grass plants on August 6, 1992 at the Fertilizer 1 grid. O p e n bars represent open subplots, striped bars represent exclosed (fenced) subplots. L o w fertilizer is 11.6gN.m"2 and high is 35gN.m~2 of 35:10:5 N P K fertilizer. Fertilizer treatments that share a common letter are not significantly different (p > 0.05; Scheffe). none  low  Table 8. Summary of A N O V A for (A) number of transplanted grass tillers per subplot (square root transformed) and (B) mean height per transplanted grass on A u g u s t 6,1992, Fertilizer 1 grid. A . Grass tillers per subplot, 1992 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 12.036 4.128 0.631 105.897  df 2 1 2 18  MS 6.018 4.128 0.316 5.883  F-ratio 1.023 0.702 0.054  P 0.380 0.413 0.948  df 2 1 2 18  MS 0.100 1.729 0.977 1.173  F-ratio 0.085 1.473 0.832  P 0.919 0.241 0.451  B. M e a n grass height, 1992 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 0.199 1.729 1.954 21.123  27  15 r  A.  5r  CO  cu >  CO JD  I  1  <ulO E ZJ O) CO  CO -Q  3 CD  JC CO  i  z  JfU  c  CO CO  0  4  '53 3  CD  5h  B.  none  low  i  high  none  low  high  Fertilizer treatment Figure 13. (A) Number (±SE) of legume leaves per subplot and (B) mean (±SE) height of legume plants on August 6, 1992 at the Fertilizer 1 grid. Open bars represent open subplots, striped bars represent exclosed (fenced) subplots. Low fertilizer is ll.6gN.rn" and high is 35gN.m" of 35:10:5 NPK fertilizer. Fertilizer treatments that share a common letter are not significantly different (p > 0.05; Scheffe). 2  2  Table 9. Summary of A N O V A for (A) number of transplanted legume leaves per subplot and (B) mean height per legume transplant for August 1992, Fertilizer 1 grid. A. Legume leaves per subplot, 1992 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 27.083 0.375 7.750 390.750  df 2 1 2 18  MS 13.542 0.375 3.875 21.708  F-ratio 0.624 0.017 0.179  P 0.547 0.897 0.838  df 2 1 2 18  MS 0.384 0.810 0.457 0.723  F-ratio 0.531 1.120 0.633  P 0.597 0.304 0.542  B. Mean legume height, 1992 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 0.768 0.810 0.915 13.010  28 A t the end of the study a different pattern was observed than for the previous season. Transplanted grasses d i d not significantly differ i n total number of tillers, but were significantly taller i n the open subplots (Fig. 14, Table 10). Legume leaf number or height d i d not respond significantly to either treatment (Fig. 15, Table 11).  29  15 r  A.  B.  CO CD  E u  > ro  _CD  I  <D!0  E -3  ~1  r  D) CD  Z  5f-  'CD  _c E  CD  C  E  CD  zz  0  2h  _CD  none  low  lh  0  high  none  low  high  Fertilizer treatment Figure 14. (A) N u m b e r (±SE) of grass tillers per subplot and (B) mean (±SE) height of grass plants on July 21, 1993 at the Fertilizer 1 grid. Open bars represent open subplots, striped bars represent exclosed (fenced) subplots. L o w fertilizer is 11.6gN.m~2 and high is 35gN.m~2 of 35:10:5 N P K fertilizer. Fertilizer treatments that share a common letter are not significantly different (p > 0.05; Scheffe).  Table 10. Summary of A N O V A for (A) number of transplanted grass tillers per subplot and (B) mean height per grass transplant on July 21, 1993, Fertilizer 1 grid. A . N u m b e r of tillers per subplot, 1993 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 3007.000 1247.042 152.333 24426.250  df 2 1 2 18  MS 1503.500 1247.042 76.167 1357.014  F-ratio 1.108 0.919 0.056  P 0.352 0.350 0.946  df 2 1 2 18  MS 52.465 118.296 42.150 25.901  F-ratio 2.026 4.567 1.627  P 0.161 0.047 0.224  B. M e a n grass height, 1993 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 104.929 118.296 84.300 466.222  30  12  7r  r  in  co 1 0 >  h  03 _fJJ  CO  E CD CO  CO  8h  E u  6h  JZ  5h  JZ  4h  CD CO  B.  CO  6h  1  4 h  2 h  low  none  E ri  3h  CD _C0  2h  c 03 CO  Ih  high  1I  none  low  I  high  Fertilizer treatment Figure 15. (A) N u m b e r (±SE) of legume leaves per subplot and (B) mean (+SE) height of legume plants on July 21, 1993 at the Fertilizer 1 grid. O p e n bars represent open subplots, striped bars represent exclosed (fenced) subplots. L o w fertilizer is 11.6gN.m 2 and high is 3 5 g N . m " of 35:10:5 N P K fertilizer. Fertilizer treatments that share a common letter are not significantly different (p > 0.05; Scheffe). _  2  Table 11. Summary of A N O V A for (A) number of transplanted legume leaves per subplot (square root transformed) and (B) mean height per legume transplant on July 21, 1993, Fertilizer 1 grid. A . Legume leaves per subplot, 1993 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 5.491 1.099 0.408 27.222  df 2 1 2 18  MS 2.746 1.099 0.204 1.512  F-ratio 1.816 0.727 0.135  P 0.191 0.405 0.875  df 2 1 2 18  MS 12.743 2.847 0.072 8.221  F-ratio 1.550 0.346 0.009  P 0.239 0.563 0.991  B. M e a n height of legumes, 1993 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 25.485 2.847 0.144 147.975  31 2. Food 1 Grid (moderate herbivore densities) A t the end of the first season the number of grass tillers responded significantly to fertilizer (Table 12), but this was not an increasing trend (Fig. 16). M e a n grass height d i d not respond to the exclosure treatment (Table 12), however grasses were taller (p=0.086) i n the fertilized treatments. The number of legume leaves responded significantly to both treatments in 1992 (Fig. 17, Table 13) w i t h fewer leaves i n the h i g h fertilizer treatment and more leaves inside exclosures. Legume height d i d not respond significantly to exclosures, however there was a significant fertilizer effect (Fig. 17, Table 13) w i t h the tallest plants i n the l o w fertilizer treatment.  32  125 r  to  w  A.  8  E o  100  B.  7 6 5  75r  CD  x: to to  50  CO i_  CD JZ  i  25 h 0  none  low  ro CD  4 3  1  2 1 0'  none high Fertilizer treatment  low  high  F i g u r e 16. ( A ) N u m b e r (±SE) of grass tillers p e r s u b p l o t a n d (B) m e a n (±SE) h e i g h t of grass p l a n t s o n A u g u s t 6, 1992 at the F o o d 1 g r i d . O p e n bars represent o p e n s u b p l o t s , s t r i p e d bars represent exclosed (fenced) s u b p l o t s . L o w f e r t i l i z e r is l l . 6 g N . n r a n d h i g h is 3 5 g N . m " of 35:10:5 N P K fertilizer. F e r t i l i z e r t r e a t m e n t s that d o n o t share a c o m m o n letter are s i g n i f i c a n t l y different (p < 0.05; Scheffe). 2  2  T a b l e 12. S u m m a r y of A N O V A for ( A ) n u m b e r of t r a n s p l a n t e d grass tillers p e r s u b p l o t (square root transformed) a n d (B) m e a n h e i g h t p e r t r a n s p l a n t e d grass o n A u g u s t 6,1992, F o o d 1 g r i d . A . G r a s s tillers p e r s u b p l o t , 1992 S o u r c e of v a r i a t i o n Fertilizer Exclosure Fertilizer * Exclosure Error  SS  df  MS  F-ratio  P  8226.750 1120.667 3293.083 7947.500  2 1 2 18  4113.375 1120.667 1646.542 441.528  9.316 2.538 3.729  0.002 0.129 0.044  df 2 1 2 18  MS 7.358 0.049 2.796 2.605  F-ratio 2.824 0.019 1.073  •  B . M e a n grass height, 1992 S o u r c e of v a r i a t i o n Fertilizer Exclosure Fertilizer * Exclosure Error  SS 14.715 0.049 5.593 46.898  P 0.086 0.893 0.363  33  none  low  high none low high Fertilizer treatment Figure 17. (A) Number (±SE) of legume leaves per subplot and (B) mean (±SE) height of legume plants on August 6, 1992 at the Food 1 grid. Open bars represent open subplots, striped bars represent exclosed (fenced) subplots. Low fertilizer is l l . 6 g N . r n and high is 35gN.m of 35:10:5 NPK fertilizer. Fertilizer treatments that do not share a common letter are significantly different (p < 0.05; Scheffe). -2  -2  Table 13. Summary of A N O V A for (A) number of transplanted legume leaves per subplot and (B) mean height per legume transplant on August 6, 1992, Food 1 grid. A. Legume leaves per subplot, 1992 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 191.083 51.042 6.083 188.750  df 2 1 2 18  MS 95.542 51.042 3.042 10.486  F-ratio 9.111 4.868 0.290  P 0.002 0.041 0.752  df 2 1 2 18  MS 4.579 0.940 0.018 0.745  F-ratio 6.142 1.261 0.024  P 0.009 0.276 0.976  B. Mean legume height, 1992 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 9.158 0.940 0.036 13.418  34 The number of grass tillers and grass height d i d not show significant treatment effects for the 1993 season (Fig. 18, Table 15).  The number of legume  leaves and mean legume height was significantly lowered by fertilizer addition (Fig. 19, Table 16), however, there was no exclosure effect.  35  70 ID  CD  =  to in  03 i_  30  B.  60 h u  50  20  40  'CD  CJ)  »+-  o 1_  CD -Q  30 h  M  03  cn 10 c  20 h  I  03 CD  10h  JIt2 none  low  high  none  F"1 low  high  Fertilizer treatment Figure 18. (A) Number (±SE) of grass tillers per subplot and (B) mean (+SE) height of grass plants on July 21, 1993 at the Food 1 grid. Open bars represent open subplots, striped bars represent exclosed (fenced) subplots. L o w fertilizer is l l . 6 g N . r n ' and high is 3 5 g N . m " of 35:10:5 N P K fertilizer. Fertilizer treatments that share a common letter are not significantly different (p > 0.05; Scheffe). 2  2  Table 14. Summary of A N O V A for (A) number of transplanted grass tillers per subplot and (B) mean height per grass transplant on July 21, 1993, Food 1 grid. A . N u m b e r of tillers per subplot, 1993 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 1570.333 2795.042 316.333 22021.250  df 2 1 2 18  MS 785.167 2795.042 158.167 1223.403  F-ratio 0.642 2.285 0.129  P 0.538 0.148 0.880  df 2 1 2 18  MS 108.186 131.056 153.631 80.053  F-ratio 1.351 1.637 1.919  P 0.284 0.217 0.176  B. M e a n grass height, 1993 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 216.372 131.056 307.262 1440.955  36  A.  4r  5r  B.  to > CO JD CD E  4h  3h CD '<D  Pi  Z3  CD CD  £ 2h 2h  zs  CD JD C CO CD 2  CD  lh  1h  0  0  none  low  high  none  low  high  Fertilizer treatment Figure 19. (A) N u m b e r (±SE) of legume leaves per subplot and (B) mean (±SE) height of legume plants on July 21, 1993 at the Food 1 grid. Open bars represent open subplots, striped bars represent exclosed (fenced) subplots. L o w fertilizer is l l . 6 g N . r n and high is 3 5 g N . m of 35:10:5 N P K fertilizer. Fertilizer treatments that do not share a common letter are significantly different (p < 0.05; Scheffe). - 2  - 2  Table 15. Summary of A N O V A for (A) number of transplanted legume leaves per subplot (square root transformed) and (B) mean height per legume transplant on July 1993, Food 1 grid. A . Legume leaves per subplot, 1993 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 29.083 0.375 1.750 37.750  df 2 1 2 18  MS 14.542' 0.375 0.875 2.097  F-ratio 6.934 0.179 0.417  P 0.006 0.677 0.665  df 2 1 2 18  MS 10.447 0.135 0.045 1.233  F-ratio 8.472 0.109 0.036  P 0.003 0.745 • 0.964  B. M e a n height of legumes, 1993 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 20.893 0.135 0.090 22.195  37 3. P r e d a t o r E x c l o s u r e + F o o d G r i d ( h i g h h e r b i v o r e d e n s i t i e s )  The number of grass tillers and mean grass height was significantly greater inside exclosures (Fig. 20, Table 16). The number of legume leaves responded significantly to both fertilizer and exclosure (Fig. 21, Table 17). There were significantly more leaves inside exclosures and when the exclosure effect is factored out, l o w fertilizer addition had significantly more leaves than the control and the high fertilizer addition. Legumes were taller inside exclosures.  38  125  A.  B.  CO 05100 CO CO CO  1— CD  u CD  75  '<D  71  "V  4—  CO CO CO  o 50 h  s_ CD -Q  E 25 h 3  3h  I—  CD  1  c  ft  none  CO CD  2h 1h  11  i.  low  high none low high Fertilizer treatment Figure 20. (A) Number (+SE) of grass tillers per subplot and (B) mean (±SE) height of grass plants on August 6, 1992 at the Predator Exclosure + Food grid. Open bars represent open subplots, striped bars represent exclosed (fenced) subplots. Low fertilizer is 11.6gN.nr and high is 35gN.m' of 35:10:5 NPK fertilizer. Fertilizer treatments that share a common letter are not significantly different (p > 0.05; Scheffe). 2  2  Table 16. Summary of A N O V A for (A) number of transplanted grass tillers per subplot (square root transformed) and (B) mean height per transplanted grass on August 6,1992, Predator Exclosure + Food grid. A. Grass tillers per subplot, 1992 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 19.955 74.576 16.675 117.758  df 2 1 2 18  MS 9.978 74.576 8.338 6.542  F-ratio 1.525 11.399 1.274  P 0.244 0.003 0.304  df 2 1 2 18  MS 0.767 18.580 2.222 1.993  F-ratio 0.385 9.323 1.115  P 0.686 0.007 0.349  B. Mean grass height, 1992 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 1.535 18.580 4.445 35.872  39 A. 4r  15  B.  ab  CD > CO  I  _co CL)  E  10  CJ) 'CD  x:  3  C D CD  CD  E  13  CD X 5  C D  5H  _QJ C  ca  fy  o  CD  j*0  1  none  low  high none low high Fertilizer treatment Figure 21. (A) Number (±SE) of legume leaves per subplot and (B) mean (±SE) height of legume plants on August 6, 1992 at the Predator Exclosure + Food grid. Open bars represent open subplots, striped bars represent exclosed (fenced) subplots. Low fertilizer is 11.6gN.m 2 and high is 35gN.m'2 of 35:10:5 NPK fertilizer. Fertilizer treatments that do not share a common letter are significantly different (p < 0.05; Scheffe). _  Table 17. Summary of A N O V A for (A) number of transplanted legume leaves per subplot (square root transformed) and (B) mean height per legume transplant on August 6, 1992, Predator exclosure + food grid. A. Legume leaves per subplot, 1992 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 3.552 11.231 2.025 7.865  df 2 1 2 18  MS 1.776 11.231 1.013 0.437  F-ratio 4.065 25.704 2.317  P 0.035 0.000 0.127  df 2 1 2 18  MS 1.002 9.957 1.567 1.04  F-ratio 0.961 9.549 1.503  P 0.401 0.006 0.249  B. Mean legume height, 1992 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 2.003 9.957 3.134 18.769  40  Trends for the 1992 season continued in 1993. The number of grass tillers was significantly higher inside exclosures (Fig. 22, Table 18) but did not respond to fertilizer. Grasses were significantly taller inside the exclosure (Fig. 22, Table 18) but did not respond to fertilizer. The number of legume leaves and legume height was significantly greater inside exclosures (Fig. 23, Table 19).  41  125  to OJ  25  r  B.  20 100  75 h  .9?  J  50 h  15  CO  ^z  to  ro  10  iC_  25  CD CO co  h  none  low  5h  high  I  none  low  j high  Fertilizer treatment  Figure 22. (A) Number (±SE) of grass tillers per subplot and (B) mean (+SE) height of grass plants on July 21, 1993 at the Predator Exclosure + Food grid. Open bars represent open subplots, striped bars represent exclosed (fenced) subplots. Low fertilizer is 11.6gN.m- and high is 35gN.m" of 35:10:5 N P K fertilizer. Fertilizer treatments that share a common letter are not significantly different (p > 0.05; Scheffe). 2  2  Table 18. Summary of A N O V A for (A) number of transplanted grass tillers (square root transformed) per subplot and (B) mean height per grass transplant (square root transformed) on July 21, 1993, Predator Enclosure + Food grid. A . Number of tillers per subplot, 1993 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 37.609 114.998 12.430 179.053  df 2 1 2 18  MS 18.805 114.998 6.215 9.947  F-ratio 1.890 11.561 0.625  P 0.180 0.003 0.547  df 2 1 2 18  MS 1.579 13.957 1.786 0.951  F-ratio 1.661 14.675 1.878  P 0.218 0.001 0.182  B. Mean grass height, 1993 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 3.159 13.957 3.572 17.120  42  CD >  ro  12  r  io  r  B. 4 r  3 t  _CD  CL)  E  "3 O) CD  CD JD  E  8 h 7  6 h  '/ '/  4 h 2  13  0  h  'CD J= CD 2  z  A  none  i i low  E zz cn  h  _CD  c 1 h ro CD  high  1i '/ fy  none  fy  I ''z  low  high  Figure 23. (A) Number (+SE) of legume leaves per subplot and (B) mean (±SE) height of legume plants on July 6, 1993 at the Predator Exclosure + Food grid. Open bars represent open subplots, striped bars represent exclosed (fenced) subplots. Low fertilizer is 11.6gN.nr and high is 35gN.m of 35:10:5 N P K fertilizer. Fertilizer treatments that share a common letter are not significantly different (p > 0.05; (A) Scheffe, (B) Kruskal Wallis). 2  -2  Table 19. Summary of (A) Kruskal Wallis for number of transplanted legume leaves per subplot and (B) A N O V A of mean height per legume transplant on July 21,1993, Predator Exclosure + Food grid. A . Legume leaves per subplot, 1993 Treatment Fertilizer & Exclosure Fertilizer Exclosure  Kruskal Wallis statistic 15.182 3.310 11.182  df 5 2 1  P 0.010 0.191 0.001  B. Mean legume height, 1993 Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 5.412 17.213 0.147 29.417  df 2 1 2 18  MS 2.706 17.213 0.074 1.634  F-ratio 1.656 10.532 0.045  P 0.219 0.004 0.956  43 C. L E A F A R E A I N D E X 1. F e r t i l i z e r 1 g r i d ( l o w h e r b i v o r e d e n s i t i e s )  Total and graminoid leaf area d i d not respond significantly to fertilizer or exclosure treatments but dicot leaf area responded significantly to both treatments (Fig. 24, Table 20). Graminoid leaf area (p=0.096) increased weakly w i t h fertilizer addition. Dicot leaf area significantly decreased w i t h fertilizer addition; leaf area was half as great i n the high fertilizer compared to the unfertilized treatment. Dicot leaf area was significantly greater inside exclosures.  Species richness d i d not respond to the exclosure treatment but decreased (p=0.056) i n response to fertilization (Fig. 25, Table 21).  44 A. Total  B. Graminoid  C. Dicot 2.0  Q. O  ro "° Sf . = re Q .  1.5  ij  1.0  ro w  0.5  o  none low  high  none  ow  high  •0.0  none  low  J  high  Fertilizer Treatment Figure 24. (A) total, (B) graminoid, and (C) dicot leaf area index on July 21, 1993, in the Fertilizer 1 grid. Open bars represent open subplots, striped bars represent exclosed (fenced) subplots. Low fertilizer is 11.6gN.m"2 and high is 35gN.m of 35:10:5 NPK fertilizer. Fertilizer treatments that do not share a common letter are significantly different (P < 0.05; Scheffe). Bars are ±SE. -2  Table 20. Summary of A N O V A for (A) total, (B) graminoid and (C) dicot (square root transformed) leaf area, July 21, 1993, Fertilizer 1 grid. A. Total leaf area Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 4281.33 40.04 500.33 21346.2  df 2 1 2 18  MS 2140.66 40.04 250.16 1185.90  F-ratio 1.805 0.034 0.211  P 0.193 0.856 0.812  SS 7881.08 770.66 361.58 26532.0  df 2 1 2 18  MS 3940.54 770.66 180.79 1474.00  F-ratio 2.673 0.523 0.123  P 0.096 0.479 0.885  SS 19.769 9.018 1.168 35.913  df 2 1 2 18  MS 9.884 9.018 0.584 1.995  F-ratio 4.954 4.520 0.293  P 0.019 0.048 0.750  B. Graminoid leaf area Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error C. Dicot leaf area Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  45  9 co  8  a a  a a  <o to  co c  uy h:  CO CO CJ CO COLO  6  a a  1  5  4 none  low  high  Fertilizer Treatment  Figure. 25. Species richness (±SE) at the Fertilizer 1 grid, on 21 July 1993. Open bars represent open subplots, striped bars represent exclosed (fenced) subplots. Low fertilizer is 11.6gN.m- and high is 35gN.m" of 35:10:5 N P K fertilizer. Exclosure treatments that share a common letter are not significantly different (p<0.05; Scheffe). 2  2  Table 21. Summary of A N O V A on species richness on 21 July 1993, Fertilizer 1 grid. Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 19.750 0.042 2.583 52.250  df 2 1 2 18  MS 9.875 0.042 1.292 2.903  F-ratio 3.402 0.014 0.445  P 0.056 0.906 0.648  46 2. Food 1 grid (moderate herbivore densities)  Neither fertilizer nor exclosures had significant effects on total, graminoid, or dicot leaf area (Fig. 26, Table 22). Fertilizer increased total leaf area (p=0.083) due to the response of the graminoids (p=0.080). The interaction effect (fertilizer * exclosure: p=0.051) for dicot leaf area shows that the unfertilized open treatment had the greatest leaf area. This result is different from the general trend for leaf area to increase with fertilizer addition or with an exclosure.  Species richness did not respond to the exclosure or the fertilization treatments (Fig. 27, Table 23).  47 A. Total  B. Graminoid  10  ° -a St  cu  10  4  8  3  8  o  ^"D c  ro Q. CD .ti  C. Dicot  6  2  R  6  4  4  1-  2 none  low  high  " 0  none low  0  high  0  iiB  none  low  L  high  Fertilizer Treatment  Figure 26. (A) total, (B) graminoid, and (C) dicot leaf area index on July 21,1993 in the Food 1 grid. Open bars represent open subplots, striped bars represent exclosed (fenced) subplots. Low fertilizer is l l . 6 g N . r n and high is 35gN.m of 35:10:5 N P K fertilizer. Fertilizer treatments that share a common letter are not significantly different (p > 0.05; Scheffe). Bars are +SE. -2  -2  Table 22. Summary of A N O V A for (A) total, (B) graminoid (square root transformed) and (C) dicot leaf area, July 21, 1993, Food 1 grid. A . Total leaf area Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 16322.5 1785.3 820.7 51346.2  df 2 1 2 18  MS 8161.29 1785.37 410.375 2852.56  F-ratio 2.861 0.626 0.144  P 0.083 0.439 0.867  SS 31.180 3.926 4.908 96.160  df 2 1 2 18  MS 15.590 3.926 2.454 5.342  F-ratio 2.918 0.735 0.459  P 0.080 0.403 0.639  SS 2079.08 30.37 2790.75 7113.75  df 2 1 2 18  MS 1039.54 30.37 1395.37 395.20  F-ratio 2.630 0.077 3.531  P 0.100 0.785 0.051  B. Graminoid leaf area Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error C. Dicot leaf area Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  48  a a  <A  a a  8  a a  to cn CD  c  o y  6  cn CD  'u  CD Q. CO  none  low  high  Fertilizer Treatment  Figure 27. Species richness (+SE) at the Food 1 grid, on July 21, 1993. Open bars represent open subplots, striped bars represent exclosed (fenced) subplots. Lowfertilizer is 11.6gN.m" and high is 35gN.m of 35:10:5 N P K fertilizer. Exclosure treatments that share a common letter are not significantly different (P < 0.05; Scheffe). 2  -2  Table 23. Summary of A N O V A on species richness on 21 July 1993, Food 1 grid. Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 6.583 0.167 3.083 36.000  df 2 1 2 18  MS 3.292 0.167 1.542 2.000  F-ratio 1.646 0.083 0.771  P 0.221 0.776 0.477  49 3. P r e d a t o r E x c l o s u r e + F o o d g r i d ( h i g h h e r b i v o r e d e n s i t i e s )  N e i t h e r total, g r a m i n o i d n o r d i c o t leaf area r e s p o n d e d s i g n i f i c a n t l y to fertilizer or e x c l o s u r e treatments ( F i g . 28, Table 24). T o t a l a n d g r a m i n o i d leaf area d o u b l e d w i t h fertilizer a d d i t i o n (p=0.124, p=0.251 respectively) b u t o n l y w h e n herbivores were excluded.  D i c o t leaf area i n c r e a s e d w i t h fertilizer  (p=0.106) a n d w a s greater i n s i d e exclosures (p=0.097) except i n the h i g h fertilizer treatment.  Species r i c h n e s s d i d not r e s p o n d to f e r t i l i z a t i o n treatments b u t i n c r e a s e d i n s i d e h e r b i v o r e exclosures ( F i g . 29, Table 25).  50 B. Graminoid  A. Total  C. Dicot  2.0 CL  1.5  o ro  T3  £ .= ro 4—  T.0-  Q_ \  ro w cu *± _ J -s= o  rrf  0.5 Y none  low  high  none  low  high  -0.0  none  low  high  Fertilizer Treatment  Figure 28. (A) total, (B) graminoid, and (C) dicot leaf area index on July 21, 1993, in the Predator Exclosure + Food grid. Open bars represent open subplots, striped bars represent exclosed (fenced) subplots. Low fertilizer is 11.6gN.m-2 and high is 35gN.m-2 of 35:10:5 N P K fertilizer. Fertilizer treatments that share a common letter are not significantly different (p > 0.05; Scheffe). Bars are ±SE. Table 24. Summary of A N O V A for (A) total, (B) graminoid (square root transformed) and, (C) dicot (log transformed) leaf area from July 1993, Predator Exclosure + Food grid. A . Total leaf area Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  SS 0.226 1.278 0.675 8.813  df 2 1 2 18  MS 0.113 1.278 0.338 0.490  F-ratio 0.231 2.609 0.689  P 0.796 0.124 0.515  SS 4.800 19.292 18.261 246.486  df 2 1 2 18  MS 2.400 19.292 9.131 13.694  F-ratio 0.175 1.409 0.667  P 0.841 0.251 0.526  SS 0.951 0.573 0.791 3.365  df 2 1 2 18  MS 0.476 0.573 0.395 0.187  F-ratio 2.545 3.065 2.115  P 0.106 0.097 0.150  B. Graminoid leaf area Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error C. Dicot leaf area Source of variation Fertilizer Exclosure Fertilizer * Exclosure Error  51  V).  8  tn  co c JZ  4  y  'al  in CD CJ  g. 41  co  11  none  low  1  high  Fertilizer Treatment  Figure. 29. Species richness (+SE) at the Predator Exclosure + Food, on July 21, 1993. O p e n bars represent open subplots, striped bars represent exclosed (fenced) subplots. L o w fertilizer is 11.6gN.m- and high is 3 5 g N . m " of 35:10:5 N P K fertilizer. Exclosure treatments that do not share a common letter are significantly different (P < 0.05; Scheffe). 2  Table 25. Summary of A N O V A on Exclosure + Food grid. Source of variation SS Fertilizer 1.083 16.667 Exclosure Fertilizer * Exclosure 1.583 Error 44.500  2  species richness on 21 July 1993, Predator df 2 1 2 18  MS 0.542 16.667 0.792 2.472  F-ratio 0.219 6.742 0.320  P 0.805 0.018 0.730  52 D. S U M M A R Y O F R E S U L T S  The study produced two results that were not predicted b y either the resource regulation hypothesis or the consumer regulation hypothesis;  a  decrease i n plant growth or survival i n response to fertilization and a decrease i n g r o w t h i n response to an herbivore exclosure. A summary of results from all the grids shows that the negative response to fertilizer occurred at a l l three sites (Table 26 A ) , but was only consistently negative for seedling growth at the Predator Exclosure + Food grid. A t the Food 1 grid there was a mixed effect of fertilizer addition. In 1992, grasses had more tillers, and legumes had more leaves and were taller. In 1993, fertilizer addition had no effect on the grasses but reduced legume leaf number and mean height. A t the Fertilizer 1 grid, fertilizer had a significant negative effect on the leaf area of established dicots but there was no effect on the growth or survival of transplants. Herbivore exclosures either had no effect, or increased transplant survival or growth, for all measurements except for the Fertilizer 1 grid i n 1992 for grass height (Table 26B).  53  Table 26. Summary of the effect of (A) fertilizer addition and (B) herbivore exclosures o n the growth and survival of transplants and the leaf area of established vegetation at three areas of k n o w n herbivore densities (ns: p>0.05). GRID(and herbivore density) Dependent Fertilizer 1 Food 1 Predator variable Excl. + Food (low) (high) (moderate) A. Fertilizer addition Survival 1992- ns • 1992- decreased 1992- decreased 1993- ns 1993- decreased 1993- decreased Seedling growth; 1992- ns grasses 1993- ns  1992- increased tillers 1993- ns  Seedling growth; 1992- ns legumes 1993- ns  1992- increased leaves 1992- decreased leaves increased height 1993- decreased leaves 1993- ns decreased height  Graminoid leaf area  ns  ns  ns  Dicot leaf area  decreased  ns  ns  Total leaf area  ns  ns  ns  Species richness  ns  ns  ns  1992- ns 1993- ns  1992- increased 1993- increased  Seedling growth; 1992- ns grasses 1993- decreased height  1992-ns  1992- increased tillers increased height 1993- increased tillers increased height  Seedling growth; 1992- ns legumes 1993- ns  1992- increased leaves 1992- increased leaves increased height 1993- ns 1993- increased leaves increased height  Graminoid leaf area  ns  ns  ns  Dicot leaf area  increased  ns  ns  Total leaf area  ns  ns  ns  Species richness  ns  ns  increased  B. Herbivore Exclosure Survival 1992- ns 1993- ns  1993- ns  1992- ns 1993- ns  54  DISCUSSION  The primary objective of this study was to evaluate the predictions made by two different hypotheses of trophic regulation based on plant responses to fertilizer additions, and to herbivore exclosures. These treatments were designed to change soil nutrient levels and herbivory rates. The fertilizer pellets incorporated into the soil two to three weeks after application and there were no heavy rains to cause excessive run-off. Plants i n the fertilized plots took on a characteristic bright green color about four weeks after the initial treatment suggesting that fertilizer was getting into plants. The resource regulation hypothesis is evaluated for the soil resources - nitrogen, phosphorous and potassium. The herbivore exclosures were effective at excluding adult hares and adult ground squirrels, but juvenile ground squirrels, mice, voles and insects could gain entry. A d u l t ground squirrels and snowshoe hares comprise most of the herbivore biomass i n this ecosystem. These herbivores eat understory plants as a summer diet. If any form of consumer regulation of plant biomass occurs i n summer-time, these vertebrate species are the most likely cause.  Both the resource regulation and the consumer regulation hypotheses group all of the species within a trophic level and treat them as a single entity. However, Pace and Funke (1991) showed that not all prey were equally affected by predators.  A factorial manipulation of predators (Daphnia)  and herbivore  resources (algae) showed that Daphnia significantly reduced ciliates but not flagellates. Goldberg and M i l l e r (1990) found that increasing soil nutrients lowered species richness, lowered plant stem density but increased plant biomass. The strict versions of both resource regulation and consumer regulation hypothesis make predictions across an entire trophic level, and cannot account for differences w i t h i n a trophic level. This experiment examined the responses  55 of i n d i v i d u a l s w i t h i n a trophic level. If either the resource regulation hypothesis or the consumer regulation hypothesis is acceptable, the mechanisms that are proposed to operate across an entire trophic level must also operate u p o n populations and individuals. In order for consumers or resources to shape plant community structure or biomass, the effects of resources or consumers w o u l d first be noticeable among seedlings because seedlings are more sensitive than established plants to herbivory (Dirzo 1984), and to l o w nutrients (Bazzaz and Harper 1977, Harper 1977).  A . FERTILIZER A D D I T I O N Plant responses to fertilization were dependent u p o n plant species, plant stage, and herbivore densities. According to both the resource regulation hypothesis and the consumer regulation hypothesis only initial soil nutrient concentration or herbivore density should change the results of fertilizer addition. The prediction that different plant stages and species respond to fertilization i n the same way is an obvious oversimplification. The results from each plant type and stage can only be taken as evidence for regulation w i t h i n that stage or plant type, but cannot be generalized to the entire trophic level. Crawley (1988) argues that w i t h i n an environment, different plants may experience different forms of regulation; some plants may be consumer regulated, and others may be resource regulated. Plant species, and stage results, are interpreted separately i n this study to allow for the identification and discussion of different sources of regulation.  Fertilizer addition resulted i n no increase i n total leaf area of established plants and reduced survival and growth of transplants. Neither the resource regulation hypothesis nor the consumer regulation hypothesis predict that  56  plants will have a negative response to fertilizer and this result differs from the expectations of most fertilization experiments. DiTommaso and Aarssen (1989) point out in their review that nutrients additions typically increase plant community productivity, decrease diversity, change species composition and accelerate succession. Species composition changed with gramminoids increasing their dominance over dicots. Species richness decreased with fertilizer addition at the Fertilizer 1 and Food 1 grids, but not at the Predator Exclosure + Food grid.  There are at least three possible explanations for the negative response to fertilizer addition.  (1) Nitrogen toxicity: Although the design of the experiment cannot exclude this possibility, there are several points that argue against it. Numerous studies done within 20km of my site have used fertilization rates similar and have not reported toxicity. For example, Nams et al. (1993) used fertilizer addition rates from zero to 125gN.m"2 and recorded increases in grass biomass, stalk weight of Epilobium angustifolium  and stalk density and leaf weight of Achillea millefolium  at  fertilized sites. A parallel study within fifteen kilometers of the sites used in this experiment and a fertilizer addition range of zero to 50gN.m , showed no -2  evidence of toxicity among the established plants but a shift in species composition (Dlott, unpublished data). Graham (1994) reported that Lupinus arcticus did not respond to two years of fertilization, 9gN.m~2yr~l and 18gN.m~2yr l respectively. Arii (unpublished data) found that in response to 12 _  to 2 5 g N . m , Achillea millefolium and Festuca altaica increased in biomass, -2  Lupinus arcticus showed no response and Anenome parviflora  showed  increased mortality with increased fertilizer addition, from no mortality with no  57 added fertilizer, to 30% mortality at 12gN.m" and 80% mortality at 25gN.m~ . 2  2  A r i i ' s results show that for some species, even relatively l o w fertilizer addition rates can increase mortality rates for some species. However, i n this study, Anemone was not used i n the transplant experiment and accounted for less than 1% of the leaf area of the established vegetation. Further/the species found i n my experiment are the same species reported to increase i n response to fertilizer by John and Turkington (1995), Nams et al. (1993) and Turkington (unpublished data). There was no evidence of chlorosis i n the field w h i c h provides another argument against a toxic effect of the added fertilizer .  (2) Shading mortality: The shading mortality hypothesis argues that as soil resources increase, the increased growth w i l l lead to increased biomass thereby reducing light at ground level, consequently there w i l l be an increase i n plant mortality. Shading may account for a decrease i n transplant performance i n response to fertilization. Plants typically grow taller i n response to fertilization and plants subordinate i n the canopy may fall below their compensation point, grow more slowly, and eventually die (Tilman 1987; Goldberg and M i l l e r 1990). Self-thinning mortality i n monocultures is faster when soil nutrients are abundant (Harper 1977), presumably due to increased competition for light. Goldberg and M i l l e r (1990) dubbed this the 'light-mortality' hypothesis and this hypothesis may be supported by the results from the study.  T i l m a n (1988) based his model, A L L O C A T E , on the idea that there is a trade-off between competitive ability for light and soil resources. In his model, plants that allocate comparatively little biomass to stems are poorer competitors for light and w i l l be excluded from an environment w i t h h i g h soil resources. T i l m a n (1993) showed that above ground competition increased and light at  58 ground level decreased as soil resources increased along an experimental gradient. Grime's (1977) theory of life-history strategies makes the same predictions as Tilman, but using different mechanisms. The 'importance' of competition along a resource gradient is an extremely controversial topic i n plant ecology and increased mortality under high soil fertilities has added to the debate. G r i m e predicts that soil resource addition causes a shift i n plant lifehistory strategy from nutrient 'stress tolerators' to 'competitors'. A c c o r d i n g to Grime, the plants dominating an area w i l l shift from slow g r o w i n g rosettes to faster growing, tall plants.  Resource additions increase the selective advantage  of competitive ability i n general, as some plants die because the environment favors competitive ability (Grime 1977). This differs from Tilman's v i e w where competition is equally important along a soil resource gradient. The resources change. W h e n soil resources are added, shoot competition for light increases, and root competition for soil resources decreases. T i l m a n argues that competitive ability of plants cannot be ranked for all environments, while G r i m e argues that competitive ability is a plant trait, independent of the environment. G r i m e argues that plant mortality increases as resources increase and fewer plants are able to survive i n the more competitive environment w h e n soil resources are abundant. O n the other hand, Tilman argues that as light at ground level decreases those plants able to withstand l o w light intensity w i l l survive at the expense of plants unable to withstand l o w light.  The 'light-mortality' hypothesis is supported by the increase i n transplant mortality at the Food 1 grid and Predator Exclosure + Food grid. A n y change i n light intensity at ground level was due solely to the understory plants themselves because tree cover was minimal i n these sites and d i d not change during the study. There was not a significant increase i n total leaf area i n response to fertilizer addition, but at all sites leaf area was greatest i n the plots  59 receiving the highest fertilizer rate. Light intensity was not measured i n this study, but leaf area provides an estimate of light at ground level because as leaf area increases the amount of light reaching ground level decreases. Species richness also showed no significant response to fertilizer addition. A t the Fertilizer 1 and the Food 1 site species richness was lowest when fertilizer addition was highest w h i c h is consistent w i t h an hypothesis of increased competitive exclusion as fertilizer addition increased. A t the Predator Exclosure + Food grid there was a large reduction i n species richness outside exclosures and a negative response to fertilizer. The light-mortality hypothesis can not account for the negative response to fertilizer addition at this site because herbivore exclosures caused a significant effect on species richness and the light mortality hypothesis alone can not account for w h y herbivore exclosures could produce this effect.  3) Increased palatability: Fertilized plants are often more palatable to herbivores than unfertilized plants. If fertilizer addition stimulates herbivore activity, then the negative response to fertilizer seen i n this experiment can be explained by increased herbivory and the corresponding increase i n plant damage. Eucalyptus tree amino acid concentration increased w h e n given balanced N P K fertilizer (Fox and M o r r o w 1992). Plant defensive secondary compounds decreased i n Salix twigs receiving fertilizer and boreal w o o d y plants i n areas where fertilizer h a d been added were more likely to be eaten by hares (Bryant et al. 1987). Red grouse, hares and rabbits fed preferentially on fertilized heather (Miller 1968). Pocket gopher activity and foraging effort increased on fertilized plots (Huntly and Inouye 1988). Herbivores appear to detect the improved food quality of fertilized plants and eat them preferentially.  60  The herbivore exclosure treatment at the Fertilizer 1 grid and the Food 1 grid d i d not increase transplant survival or growth, therefore, increased plant palatability alone cannot explain the negative plant response to fertilizer. A t the Predator Exclosure + Food grid, however, increased palatability may account for the negative response to fertilizer, because herbivore exclosures increased transplant survival. Further, fertilizer addition increased leaf area only w h e n herbivores were excluded and exclosures increased species richness at the Predator Exclosure + Food grid, but not at the other grids.  B. H E R B I V O R E E X C L O S U R E S The impact of herbivores on plants varies w i t h herbivore density (Crawley 1989, N o y M e i r 1990). Herbivore exclosures were successful i n increasing transplant survival, number of leaves and height only at the h i g h herbivore density found at the Predator Exclosure + Food grid, but exclosures d i d not influence the total leaf area of the established vegetation. Leibold (1989) points out that plant biomass is less likely to change due to grazing than species composition is because a less palatable species w i l l likely take over following intense grazing. The overall effect of herbivory then, if plants vary i n their palatability, is for some plant species to decrease and others to increase i n abundance (Crawley 1988) but for total vegetation biomass to stay the same.  A t l o w and moderate densities, herbivore exclosures d i d not influence seedling survival or growth, nor d i d exclosures limit the abundance of established plants. Therefore, d u r i n g l o w hare densities, herbivores do not limit plants i n the herbaceous understory. O n l y when herbivore densities are artificially high, such as achieved i n this experiment using an area that had supplemental food and no terrestrial predators, were young plants limited b y  61 herbivores. Herbivore densities at the Predator Exclosure + Food grid were four times greater than the normal peak herbivore densities (Krebs et al. 1995). The consumer regulation hypothesis is not supported by this experiment for naturally occurring densities of snowshoe hares and juvenile ground squirrels and their food plants.  M i c e , voles and juvenile ground squirrels were able to pass through the exclosure, and w h e n the increased palatability of fertilized plants is considered increased herbivory from these unexcluded herbivores could account for the negative response to fertilization and the lack of an exclosure response. If the fertilized plants were heavily grazed b y the mice, voles and juvenille ground squirrels then there w o u l d be a decrease i n plant growth and survival i n response to fertilizer and no exclosure effect since these animals can pass through the fence. A t the Food 1 grid, small mammals on occasion were seen w i t h i n exclosures as were their turds, chewed plants, missing, and chewed tags. M i s s i n g and moved tags were also common at the Predator Exclosure + Food grid. The impact of these small rodents can only be estimated. Pyke (1986) s h o w e d that grazing by Microtus montanus and Peromyscus maniculatus  l i m i t i n g seedling survival and biomass of the grass Agropyron spicatum, but not Bromus tectorum. Pyke's study had small m a m m a l densities m u c h greater than those observed here, 750 animals/ha compared to 6 animals/ha at Kluane.  The herbivore exclosures instead of stopping all vertebrate herbivores were semi-permeable - stopping adult ground squirrels and snowshoe hares but not stopping juvenile ground squirrels, mice, voles and insects. Juvenile ground squirrels however, became too large to pass through the 2.5cm-mesh exclosures three weeks after they emerged i n the spring. After this time, there was less evidence of a breach i n the exclosures at the Predator Exclosure + Food grid.  62 M i c e and voles were never excluded and evidence of their activity was seen throughout the first season at the Food 1 grid. A n imperfect exclosure may differ from a perfect exclosure by reducing the frequency of attack, regardless of herbivore density (Fig. 30). Increased palatability may be modeled as an increase i n the effective density of herbivores. Superimposing a fertilizer-induced increase i n palatability w i t h the effect of an imperfect exclosure, shows that herbivores can be more damaging when plants are fertilized (Fig. 31). Increased herbivore impact on plants could increase mortality i n fertilized plots. The 'leaky fence hypothesis', w o u l d appear as a simple negative reaction to fertilization but w o u l d mask the true effect of herbivory - the increased palatability of fertilized plants stimulates herbivore activity, and an imperfect exclosure w o u l d allow some herbivores to eat plants thought to be unavailable.  63  cu ZZ •D  CD  £  ...no exclosure  O  >  E-S  _ imperfect exclosure  CC OJ T3 4-> O _C0  — p e r f e c t exclosure  D_  Herbivore Density Figure 30. Theoretical changes i n plant damage due to herbivory as herbivore density and exclosure efficacy vary.  no exclosure • fertilized  imperfect exclosure , 'fertilized  CD  ZZ TCD  £ O  S* ^ co •--  E -2  CO CD T 3 -~~ +-> O C +-» _CO  no exclosure ..-•unfertilized imperfect exclosure — unfertilized  Q_  perfect exclosure  Herbivore Density Figure 31. Theoretical changes i n plant damage due to herbivory as herbivore density, exclosure efficacy and fertilizer addition vary.  64 C. F I E L D SITES The three K B F E P sites are evaluated separately because sites varied i n initial plant species composition, light intensity at ground level, water availability, soil nutrient concentration, and herbivore density. After each site is considered alone, sites are compared to gain a broader understanding of the trophic relations.  1. Fertilizer 1 G r i d Results from this site d i d not support the suite of predictions from either the resource regulation hypothesis or the consumer regulation hypothesis. Transplant growth and survival of both grasses and legumes d i d not increase i n response to herbivore exclosures, contrary to the predictions of the consumer control hypothesis but i n accordance w i t h the resource regulation hypothesis.  This is an extremely dry environment (mean annual precipitation 28.2 cm) and other resources such as water may be limiting. Water was not manipulated experimentally i n this study, so the resource regulation hypothesis must be rejected, at least for the resources manipulated (nitrogen, phosphorous, and potassium). If another resource is regulating vegetation biomass then a different version of the resource regulation hypothesis may be constructed proposing a resource other than nitrogen, phosphorous, or potassium. If water or light was proposed as the regulating resource, then the negative response to fertilizer could be accounted for. The repeated fertilizer additions since 1987, w o u l d have already caused short term changes i n species composition. The decrease i n dicot leaf area shows a negative response to fertilizer w h i c h may be accounted for b y the 'light-mortality' hypothesis.  65 1. Food 1 G r i d A s for the Fertilizer 1 grid, predictions from neither the resource regulation hypothesis nor the consumer regulation hypothesis matched results. Herbivore exclosures had no effect on growth of transplanted grasses or leaf area of established plants i n conflict w i t h the predictions of the consumer regulation hypothesis. The number of transplanted legumes was greater inside exclosures than outside w h i c h argues for partial consumer regulation. The contrast between the negative effect of fertilizer addition on transplant survival, legume height and the number of legume leaves and the positive effect on total leaf area may be accounted for by either the 'light-mortality' hypothesis or the 'leaky fence' hypothesis. Seedlings are more sensitive to both herbivory and competition (Cavers and Harper 1967; Crawley 1988) than larger plants. This experiment was not designed to test either of these hypotheses and therefore their relative impacts can only be evaluated post hoc.  Shading appears to have had more of an impact than increased palatability for three reasons. First, leaf area was high, averaging 5.78 leaves, i n the unfertilized plots and the thick, dense swards of grass shaded the transplants by the end of the first season, and severely shaded transplants by the end of the second season. Second, mice and vole densities were higher i n the first season than i n second season but the difference between fertilized and unfertilized transplant survival was greatest after the second season. If the mice and voles were decreasing transplant survival, the greatest difference w o u l d be expected w h e n densities were highest. Third, the decrease i n the growth of transplanted legumes was m u c h more dramatic than for the grasses. Grasses were m u c h taller and are better suited for competition for light. The palatability argument w o u l d require that the legumes were more palatable than the grasses. Other than lupine, legumes are not preferred foods i n cafeteria trials (D. H i k , pers. com.).  66  The grasses Festuca and Calamagrostis are both w i t h i n the top ten species of preferred summer forage for hares.  The resource regulation hypothesis must be rejected for the soil resources added - nitrogen, phosphorous and potassium, but can be modified to account for the data by altering the resources hypothesized to be limiting. The lack of a response to herbivore exclosures supports resource based regulation as opposed to consumer regulation. Water d i d not appear to be limiting because the sites were located at the bottom of a damp gully. The dense herbaceous cover suggested that light was the limiting resource, and therefore the resource regulation hypothesis using light as the regulating resource could account for these data.  3. Predator Exclosure + Food G r i d Results from the transplant experiment closely matched the predictions of the consumer regulation hypothesis at this grid. Survival, height and number of leaves were consistently higher inside exclosures.  The consumer regulation hypothesis, however, is not supported by the leaf area results, because herbivore exclosures had no effect on leaf area.  Fertilizer addition increased mortality and decreased growth of transplants, but had no effect on the leaf area of established plants. Neither the resource regulation nor consumer regulation hypothesis predict this result. A resource regulation hypothesis using a resource other than those added may account for the negative response to fertilization.  67 A t this grid, there is support for the 'leaky fence' hypothesis. Fertilizer had no effect on leaf area except when the area was inside an exclosure. Leaf area appears to have been reduced by herbivores i n open plots. The difference i n transplant mortality between fertilized and unfertilized plots occurred early i n the first season, presumably before the understory plants could shade one another and precisely w h e n juvenile ground squirrel activity was greatest.  Regulation was different for the two plant stages examined, consumer regulation for seedlings, and resource regulation for the established plants. Pyke (1986) showed that the impact of small mammals varied depending on plant size and species. Karban and Strauss (1993) observed 83% mortality of seedlings due to gophers over a six-year period but over a ten-year period only a single established clonal adult died i n the study and its death was attributable to the overgrowth of an introduced grass, not gopher predation. Seedling mortality can be dramatically different than adult mortality and indeed population dynamics may be driven by rare w i n d o w s of seedling establishment (Crawley 1988).  D. TROPHIC C O N T R O L A N D THE SNOWSHOE H A R E C Y C L E These results can be integrated into the overall boreal forest ecosystem using results from the work of Oksanen, Fretwell, A r r u d a , and N i e m e l a (1981) and Oksanen (1988; 1990). The Oksanen hypothesis uses both resource regulation and consumer regulation to account for the distribution of plants, herbivores and predators along a wide range of plant productivity. Oksanen et al. (1987) predict that the type of regulation is dictated by the number of functional trophic levels. The Oksanen hypothesis w o u l d predict that plants w o u l d be the sole functional trophic level w h e n snowshoe hare density is low. A t the Fertilizer 1 grid and the Food 1 grid, because plant biomass appears to be resource regulated rather than consumer regulated, plants are the sole functional trophic level and  68 plant dynamics w o u l d be driven by competition (Fig. 32 A ) . A t the Predator Exclosure + Food grid, there w o u l d be two functional trophic levels, both plants and herbivores and plant dynamics are consumer regulated (Fig. 32 B).  Extending the view of Oksanen et al. (1987) to account for the unmanipulated snowshoe hare cycle, during the early stage of a hare peak plants w o u l d be regulated by snowshoe hares and ground squirrels and these herbivores w o u l d not yet be predator regulated, a two trophic level system (Fig. 32-B). A t the hare peak and during the decline, predators become a functional trophic level and regulate herbivores, creating a three functional trophic level system (Fig. 32 C). D u r i n g a hare low, however, herbivores w o u l d no longer regulate plants and the whole system w o u l d revert to a one-trophic level, plantonly system (Fig. 32 A ) . The Fertilizer 1 and Food 1 grids are both at the postdecline stage and support this model. The Predator Exclosure + Food grid appears to have sufficient snowshoe hare and ground squirrel numbers to limit transplant survival and growth but not the biomass of the established understory plants. Consumer regulation w o u l d be partial, controlling seedling reproduction only. The predator exclosure at this grid means that herbivores cannot be predator regulated and runaway plant consumption is predicted d u r i n g a hare peak. The herbivore-plant link from the Predator Exclosure + Food grid supports the Oksanen et al. (1987) hypothesis.  69  B  • N  • V  N  • 1*  • N  Productivity Figure 32. Predictions from the Oksanen et al. (1981) hypothesis showing the direction and magnitude of trophic regulation for ecosystems w i t h from 1 to 3 trophic levels ( A - C respectively). A r r o w s indicate direction and strength of interaction. (N= nutrients (abiotic factors), V= vegetation, H = herbivores, and P= predators). Bold boxes indicate self regulation, by competition.  A c c o r d i n g to the Oksanen hypothesis, plant productivity must change i n order for trophic organization to change w h i c h is a problematic issue because productivity appears to remain relatively constant throughout the snowshoe hare cycle. Nevertheless, Bryant et al. (1987) showed that the quality of winter browse decreased following the severe grazing of a hare peak. Thus, while overall plant productivity may remain constant, the portion available to herbivores dramatically decreases just after a peak, w h i c h is what is required by Oksanen et al. (1981). Edible plant productivity peaks as the system becomes a three-functional trophic level system. The change i n edible plant productivity could fall w i t h i n the range predicted by the Oksanen et al. (1987) hypothesis.  70  The predictions from the consumer regulation hypothesis d i d not match the results from experiments from l o w and moderate herbivore densities. A t h i g h herbivore densities, the consumer regulation hypothesis predicted the observed transplant responses to herbivore exclosures, but d i d not predict the results from the established vegetation. Generalizing, the understory plants i n the boreal forest do not appear to be consumer regulated under natural herbivore densities d u r i n g the snowshoe hare low. The predictions from the resource regulation hypothesis d i d not match results from the fertilization experiments at all herbivore densities. In fact transplants that were i n plots w i t h fertilizer added had higher mortality and less growth than those i n unfertilized plots. However, these results provide some evidence for resource regulation, i n that increased competition for light may account for the decrease i n plant performance i n response to fertilizer addition.  71 BIBLIOGRAPHY  Bazzaz, F. A . , & Harper, J. L . (1977). Demographic analysis of the growth of Linum usitatissimum. L . N e w Phytologist, 78, 193-208. Boutin, S. (1990). Food supplementation experiments w i t h terrestrial vertebrates: patterns, problems and the future. Canadian Journal of Zoology, 68, 203-220. Bryant, J. P., C h a p i n III, F. S., Reichardt, P. B., & Clausen, T. P. (1987). Response of winter chemical defense i n Alaska paper birch and green alder to manipulation of plant carbon/nutrient balance. Oecologia (Berlin), 72, 510-514. Carpenter, S. R., & Kitchell, J. F. (1987). The temporal scale of variance i n lake productivity. A m e r i c a n Naturalist, 129, 417-433. Cavers, P. B., & Harper, J. L . (1967). Studies i n the dynamics of plant populations. I. The fate of seed and transplants introduced into various habitats. Journal of Ecology, 55,59-71. Crawley, M . J. (1988). Herbivores and Plant Population Dynamics. In A . J. Davy, M . J. Crawley, & A . R. Watkinson (Eds.), Plant Population Ecology: The 28th Symposium of the British Ecological Society (pp. 367-392). Oxford: Blackwell Scientific Publications. Crawley, M . J. (1989). Rabbit grazing, plant competition and seedling recruitment in acid grassland. Journal of A p p l i e d Ecology, 27, 803-820. D i r z o , R. (1984). Herbivory: a phytocentric overview. In R. D i r z o & J. Sarukhan (Eds.), Perspectives on Plant Population Ecology (pp. 141-165). Sunderland, M A : Sinauer Associates Inc. DiTommaso, A . , & Aarssen, L . W . (1989). Resource manipulations i n natural vegetation: a review. Vegetatio, 84, 9-29. Douglas, G . W . (1974). Montane zone vegetation of the Alsek River region, southwest Y u k o n . Canadian Journal of Botany, 52, 2505-2532. Fox, L . R., & M o r r o w , P. A . (1992). Eucalypt responses to fertilization and reduced herbivory. Oecologia, 89(2), 214-222. Goldberg, D . E., & Miller, T. E. (1990). Effects of different resource additions on species diversity i n an annual plant community. Ecology, 71(1), 213-225.  72 Gomez, K . A . , & Gomez, A . A . (1984). Statistical Procedures for Agricultural Research (Second ed.). N e w York Chichester Brisbane Toronto Singapore: John W i l e y and Sons. Graham, S. A . (1994) The relative effect of clipping, neighbours, and fertilization on the population dynamics of Lupinus arcticus (family Fabaceae). M . S c , The University of British C o l u m b i a . Grime, J. P. (1977). Evidence for the existence of three primary strategies i n plants and its relevance to ecological and evolutionary theory. A m e r i c a n Naturalist, 111, 1169-1194. Hairston, N . G . , Smith, F. E., & Slobodkin, L . B. (1960). C o m m u n i t y structure, population control, and competiton. American Naturalist, 94, 421-424. Harper, J. L . (1977). Population Biology of Plants. London: Academic Press. Hulten, E . (1968). Flora of Alaska and Neighboring Terretories: A manual of Vascular Plants. Stanford: Stanford University Press. Hunter, M . D . , & Price, P. W . (1992). Playing chutes and ladders: heterogeneity and the relative roles of bottom-up and top-down forces i n natural communities. Ecology, 73(3), 724-732. H u n t l y , N . J., & Inouye, R. S. (1988). Pocket gophers i n ecosystems: patterns and mechanisms. BioScience, 38(1), 786-793. Hutchings, M . J., Booth, K . D . , & S., W . (1991). Comparison of survivorship by the logrank test: criticisms and alternatives. Ecology, 72(6), 2290-2293. John, E . , & Turkington, R. A . (1995). Herbaceous vegetation i n the understorey of the boreal forest; does nutrient supply or snowshoe hare herbivory regulate species composition and abundance? Journal of Ecology, 83, 581-590. Karban, R., & Strauss, S. Y . (1993). Effect of herbivores on growth and reproduction of their perrenial host, Erigeron glaucus. Ecology, 74(1), 39-46. Krebs, C . J., Boutin, S., Boontra, R., Sinclair, A . R. E., Smith, J. N . M . , Dale, M . R. T., M a r t i n , K . , & Turkington, R. (1995). Impact of food and predation on the snowshoe hare cycle. Science, 269,1112-1115. Krebs, C . J., Gilbert, B. S., Boutin, S., Sinclair, A . R. E., & Smith, J. N . (1986). Population biology of snowshoe hares. I. Demography of food supplemented populations of the southern Y u k o n . Journal of A n i m a l Ecology, 55, 963-982.  73 Leibold, M . A. (1989). Resource edibility and the effects of predators and productivity on the outcome of trophic interactions. American Naturalist, 134, 922-949. McQueen, D. J., Johannes, M . R. S., Post, J. R., Stewart, T. J., & Lean, D. R. S. (1989). Bottom-up and top-down impacts on fresh-water pelagic community structure. Ecological Monographs, 59, 289-309. Menge, B. A., & Sutherland, J. P. (1976). Species diversity gradients: synthesis of the roles of predation, competition, and temporal heterogeneity. American Naturalist, 110, 351-369. Miller, G. R. (1968). Evidence for selective feeding on fertilized plots by red grouse, hares, and rabbits. Journal of Wildlife Management, 32, 949-853. Nams, V. O., Folkard, N . F. G., & Smith, J. N . M . (1993). Effects of nitrogen fertilization on several woody and nonwoody boreal forest species. Canadian Journal of Botany, 71, 93-97. Northcote, T. G. (1988). Fish in the structure and function of freshwater ecosystem: a "top-down" view. Canadian Journal of Fisheries and Aquatic Sciences, 45, 361-379. Noy-Meir, I. (1990). The effect of grazing on the abundance of wild wheat, barley and oat in Israel. Biological Conservation, 51, 299-310. Oksanen, L. (1988). Ecosystem organization: mutualism and cybernetics or plain Darwinian struggle for existance. American Naturalist, 131, 424-444. Oksanen, L. (1990). Predation, herbivory, and plant strategies along gradients of primary productivity. In D. Tilman & J. Grace (Eds.), Perspectives on Plant Consumption (pp. 445-474). New York: Acamemic Press. Oksanen, L., Fretwell, S. D., Arruda, J., & Niemela, P. (1981). Exploitation ecosystems in gradients of primary productivity. American Naturalist, 118, 240261. Pace, M . L., & Funke, E. (1991). Regulation of planktonic microbial communities by nutrients and herbivores. Ecology, 72(3), 904-914. Pyke, D. A . (1986). Demographic responses of Bromus tectorum and seedlings of Agropyron spicatum to grazing by small mammals: occurrence and severity of grazing. Journal of Ecology, 74, 739-754. Pyke, D. A., & Thompson, J. N . (1986). Statistical analysis of survival and removal rate experiments. Ecology, 67(1), 240-245.  74 Pyke, D . A . , & Thompson, J. N . (1987). Erratum. Ecology, 68, 232. Tilman, D . (1987). Secondary succession and the pattern of plant dominance along experimental nitrogen gradients. Ecological Monographs, 57(3), 189-214. Tilman, D . (1988). Plant Strategies and the Dynamics and Structure of Plant C o m m u n i t i e s . Princeton, N e w Jersey: Princeton University Press. Tilman, D . (1993). Species richness of experimental productivity gradients: h o w important is colonization limitation? Ecology, 74(8), 2179-2191. Webber, B. L . (1973). Climactic inventory of Kluane National Park. D o w n s v i e w , O N Canada: White, T. C . R. (1978). The importance of a relative shortage of food i n animal ecology. Oecologia, 33, 71-86. White, T. C . R. (1984). The abundance of invertebrate hervbivores i n relation to the availability of nitrogen i n stressed food plants. Oecologia, 63, 90-105. W i l k i n s o n , L . , H i l l , M . A . , & Vang, E. (1992). SYSTAT: statistics (version 5.2 for Macintosh ed.). Evanston, IL U S A : SYSTAT, Inc.  

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.831.1-0087064/manifest

Comment

Related Items