鳥類掠食對水柳(Salix warburgii) 跨食物鍊階層之影響

全文

(1)國立臺灣師範大學生命科學系碩士論文. 鳥類掠食對水柳(Salix warburgii) 跨食物鍊階層之影響 Trophic cascading effects of avian predation on a willow (Salix warburgii). 研究生：吳佩真 Pei-chen Wu 指導教授：李佩珍 Pei-Jen Lee Shaner 中華民國 102 年 12 月.

(2) Table of content. Abstract ......................................................................................................1 摘要.............................................................................................................2 Introduction ...............................................................................................3 Materials and methods .............................................................................6 Study system and experimental design .............................................. 6 Plant growth, reproduction and herbivory damage ......................... 8 Plant defensive phytochemicals.......................................................... 9 Arthropod abundance ...................................................................... 10 Statistical analyses ............................................................................ 11 Results ......................................................................................................12 Plant growth and reproduction ........................................................ 12 Plant leaf phenolics ........................................................................... 13 Correlations among Plant traits ....................................................... 13 Arthropod abundance and herbivory damage ................................ 14 Discussion.................................................................................................15 Defensive phytochemicals in willows ............................................... 15 The cascading effects on the arthropods ......................................... 16 Branch-level resource allocation in willows .................................... 17 Temporal scales of trophic cascades ................................................ 18 Tables and figures ...................................................................................19 References ................................................................................................30 Appendices ...............................................................................................34.

(3) ABSTRACT Trophic cascades, defined as indirect effects of predators on plants via herbivores, play a crucial role in food web functioning. In this study I tested top-down trophic effects of avian predation on plants and arboreal arthropods in a willow (Salix warburgii) food web along a riparian zone of Xindian river in northern Taiwan. Bird predation was excluded by nylon netting around the branches for 20 months. The growth, reproduction and level of defensive phytochemicals of these branches were compared to that of control branches on the same trees; the abundance of arboreal arthropods and level of herbivory were also compared. The bird exclusion caused lower growth and reproduction in the willows but did not affect the abundances of 3 arthropod groups (i.e. herbivores, fungivores and predators) on them. The plant defensive phytochemicals, measured as the amount of phenolic glycosides, were lower for the bird exclusion branches. The level of herbivory was higher in the bird exclusion branches approximately one year after the treatment, but returned to a similar level as the control branches in the second year. This study demonstrated that avian predation has positive cascading effects on willows at branch level by improving their growth and reproduction, as well as reducing their herbivory. Keywords: arthropod, food web, indirect effect, plant-herbivore, top-down control. 1.

(4) 摘要食物網中，掠食者對植物造成的跨食物鏈階層(trophic cascades) 影響受到許多重視，本研究檢測水柳食物網中，鳥類掠食對水柳和樹上的節肢動物（分植食性、真菌食性與掠食者三群）豐度的影響。實驗樣地位於北臺灣新店溪岸，水柳以枝條為單位，使用細網隔絕鳥類 20 個月，並與有鳥類掠食的枝條比較其生長、繁殖、防禦性化學物質的表現以及節肢動物的豐度。無鳥類掠食的枝條生長與繁殖較差，但節肢動物群並未受到影響。防禦性化學物質以酚類配醣(phenolic glycosides)作為代表，無鳥類掠食的枝條其含量較少；葉子所受的植食損傷，無鳥類掠食的枝條在實驗開始約一年後損傷較高，但在第二年則與有鳥類掠食的枝條無異。實驗結果顯示，有鳥類掠食的水柳枝條生長與繁殖表現較好，其植食性損害也較低，證實了掠食者對植物有正向的跨食物鏈階層之影響。. 關鍵字：節肢動物、食物網、間接效應、植物與植食性動物之關係、下行控制. 2.

(5) INTRODUCTION Trophic cascades are defined as the propagation of indirect mutualism between non-adjacent levels in a food chain (Menge 1995, Persson 1999), which have been shown to influence primary productivity (Huryn 1998), herbivore abundance (Marquis and Whelan 1994, Sipura 1999, Ho and Pennings 2008), herbivore species composition (Marquis and Whelan 1994, Jeppesen et al. 1998, Schmitz et al. 2006) and food web dynamics (Jeppesen et al. 1998, Casini et al. 2009). Furthermore, trophic cascades can occur in a variety of ecosystems, including lakes and streams (Huryn 1998, Jeppesen et al. 1998), marine and intertidal zones (Trussell et al. 2002, Casini et al. 2009), grasslands (Spiller and Schoener 1994, Schmitz et al. 2006) and forests (Marquis and Whelan 1994, Sipura 1999, Mooney 2007, Stolter 2008, Böhm et al. 2011). Because of their wide-range effects in food webs and ubiquitous presence across ecosystems, trophic cascades have remained one of the most well-studied subjects in ecology since the idea was first proposed by Hairston, Smith and Slodbokin (1960). In terrestrial environments, herbivorous arthropods often play a crucial role in ecosystem functioning because of their great abundance and large impact on plant growth, reproduction and life history traits (Crawley 1989, Coley and Barone 1996). On the other hand, avian predation has been shown to suppress arthropod abundance and thereby reduce herbivory (Marquis and Whelan 1994, Sipura 1999, Greenberg et al. 2000, Böhm et al. 2011). For example, Marquis and Whelan (1994) 3.

(6) caged white oak (Quercus alba L.) with nylon gill netting in a deciduous forest in Missouri, U.S. for two years to prevent avian predation. As a result, the plants suffered twice as much herbivory as the control plants and consequently produced less aboveground biomass. Interestingly, Sipura (1999) compared the effects of avian predation on two different willow species (Salix phylicifolia and S. myrsinifolia), and found that the birds reduced the densities of leaf-chewing insects and level of leaf damage in S. phylicifolia but not S. myrsinifolia; the latter producing 50-fold more defensive phytochemicals compared to S. phylicifolia. In this study I examined top-down effects of avian predation on arthropods and adult willow (Salix warburgii) trees in a riparian area in northern Taiwan. Plants are sessile organisms, relying on defense, tolerance or trophic cascades to cope with herbivory. As a member of Salicaceae, willow (Salix spp.) is known to develop a wide range of secondary metabolites that could be used as toxins and deterrents to help reduce herbivory (Dudt and Shure 1994, Orians and Fritz 1995, Fritz et al. 2001, Hjältén et al. 2007, Boeckler et al. 2011). Therefore, one may expect to see weak trophic cascades in willows. However, Sipura (1999) demonstrated that level of defense in different willow species affects the strength of trophic cascades, and weakly-defended willows could still exhibit strong, positive trophic cascades. In order to understand patterns of trophic cascades in plants that possess highly evolved chemical defense, more empirical studies on willows are needed. In addition to chemical defense, other bottom-up forces, such as primary productivity (Oksanen et al. 1981, Huryn 1998), are also recognized as important 4.

(7) factors modulating trophic cascades. Therefore, relative importance of bottom-up (e.g. plant traits) and top-down processes (e.g. avian predation) in food web dynamics, and how one process may be affected by the other, are still under debates (Hunter and Price 1992, Power 1992, Hunter et al. 1997). Empirical studies on trophic cascades that explicitly consider multiple plant traits should contribute to our overall understanding of bottom-up and top-down processes in food webs. In this study, I investigated multiple plant traits ranging from growth to reproduction and chemical defense. Because I used adult trees, I was able to measure reproduction by the number of flowers produced. Furthermore, adult trees and seedlings often adopt different anti-herbivory strategies (e.g. defense versus compensatory growth; Haukioja et al. 1998, Fritz et al. 2001, Donaldson et al. 2006), yet most of the previous studies had focused on seedlings (Larsson et al. 1986, Dudt and Shure 1994, Marquis and Whelan 1994, Stamp and Bowers 2000, Fine et al. 2004, Griffin and Thaler 2006). Therefore, by using adult trees, I was able to examine plant traits (e.g. reproduction and chemical defense) that could not be investigated by using seedlings or that might have different patterns in seedlings. I used level of phenolic glycosides in willow leaves as a measure of their chemical defense. Phenolic glycosides are some of the most abundant secondary phytochemicals in leaf tissues of willows and poplars (Populus spp.) (Boeckler et al. 2011). Previous studies have found that phenolic glycosides can deter herbivores including mammals (Stolter 2008), insects (Tahvanainen et al. 1985, Larsson et al. 1986, Dudt and 5.

(8) Shure 1994, Hjältén et al. 2007) and molluscs (Fritz et al. 2001). On the other hand, some specialized herbivores, such as willow-feeding beetles Chrysomela popuili and Phratora vitellinae, can utilize these phenolic glycosides for their own defense (Rowell-Rahier and Pasteels 1990, Boeckler et al. 2011). Therefore, phenolic glycosides are more likely used as chemical defense against generalist herbivores. My predictions are that bird exclusion should increase abundance of arthropods and level of leaf damage in the willows, and consequently it should lead to cascading effects on willow growth and reproduction. However, if bird exclusion causes the willows to increase their chemical defense, I expect the cascading effects from avian predation to be minimal.. MATERIALS AND METHODS STUDY SYSTEM AND EXPERIMENTAL DESIGN Salix warburgii is common in riparian zone of low lands throughout Taiwan (Huang et al. 1996). The plant grows as a single tree or a clone; adult trees can grow to 10 m in height (Huang et al. 1996). At the study site, S. warburgii typically starts shedding leaves in late November or early December, which lasts throughout the winter; the amount of leaves left on the tree varies among individuals and between years. In most cases, leaves in the upper canopy and on the distal part of the branches are kept on the trees. The plant is dioecious (Huang et al. 1996); their inflorescences appear from late February to April with new leaves sprouting in March after the inflorescences (personal observation). 6.

(9) The study was conducted from September 2010 to June 2012 in a riparian area along Xindian river in northern Taiwan (25º00’52’’ to 25º01’00’’ N and 121º29’24’’ to 121º29’29’’ E; Figure 1). Annual mean temperature is 23ºC and annual precipitation is 2405 mm between 1981 and 2010 (Taiwan Central Weather Bureau, http://www.cwb.gov.tw, accessed in October 2013). The area is under tidal influence, and some of the willows at the study site are periodically submerged in water during high tides. The mean water level is between 1.14 m and 1.39 m at high tides, and between -0.59 m and -0.88 m at low tides (Taiwan Central Weather Bureau, http://www.cwb.gov.tw, accessed in October 2012). The vegetation is consisted of subtropical broad-leaved trees and shrubs, such as paper mulberry (Broussonetia papyrifera), small-leaved mulberry (Morus australis Poir.), macaranga (Macaranga tanarius), small Philippine Acacia (Acacia confusa), zebra grass (Miscanthus spp.) and water willow (S. warburgii). Salix warburgii is one of the most dominant tree species at the site (personal observation). Common avian predators at the site include tree sparrow (Passer montanus), Chinese bulbul (Pycnonotus sinensis), Japanese white-eye (Zosterops japonica), magpie (Pica pica), crested myna (Acridotheres cristatellus), vinous-throated parrotbill (Paradoxornis webbianus bulomachus) and brown shrike (Lanius cristatus lucionensis) (personal observation). I randomly selected 11 adult trees of S. warburgii (for a detailed description of the 11 trees, see Appendix 1) for this study. On each of the 11 experimental trees, two branches of similar heights (Appendix 1) were selected from lower canopy layer, and randomly assigned to either bird 7.

(10) exclusion or control treatment (no bird exclusion). Birds were excluded by enclosing the whole branch with nylon netting (mesh size 36.36 mm, material thickness 0.12 mm, black), which allowed access of arthropods but not that of the smallest birds at the study site (e.g. Japanese white-eye, body length 11 cm; Chang 1985). The nettings were set up between September and November of 2010; initially on 8 trees in September, and subsequently on 3 additional trees in November. All trees were checked 1-3 times every month to make sure the nettings were working properly. To further deter the birds from bird exclusion branches, I also attached sparkling strips on the nettings. PLANT GROWTH, REPRODUCTION AND HERBIVORY DAMAGE Plant growth was measured with leaf dry weight (mg), leaf area (cm2) and bud count, and plant reproduction was measured with flower count. The leaves were sampled 3 times throughout the experiment: September and November 2010 (pre-treatment sample), October 2011 and June 2012; the buds and flowers were counted twice: March 2011 and February 2012, both were post-treatment samples. For leaf dry weight and leaf area, I selected 10 leaves from each branch; the leaves were at least 20 cm apart from one another. The leaves were cut at the petiole, put in a cooler and transferred back to the lab within 12 hours. In the lab, the leaves were washed off dust and animal traces such as silks, eggs and feces, and dried with paper towels. After that, the leaves were scanned and then dried in the oven at 40ºC ~60ºC for at least 48 hours. The number of leaves on each branch was also recorded, and total leaf dry weight and total leaf area were calculated as 8.

(11) product of the mean value obtained from the 10 leaves and total leaf counts. For herbivory damage, I scanned the images of the 10 leaves and used ImageJ (http://rsbweb.nih.gov/ij/) to determine the leaf area. I estimated whole leaf area by filling the chewed holes on the leaves using Adobe Illustrator and the scanned leaf images (Appendix 2). Damaged leaf area was calculated as the difference between filled leaf area and unfilled leaf area (i.e. damaged leaf area = whole leaf area – unfilled leaf area), and herbivory damage was quantified by the percentage of the damaged leaf area (i.e. damaged leaf area / whole leaf area). PLANT DEFENSIVE PHYTOCHEMICALS I used the amount of phenolic glycosides in the leaves to measure level of chemical defense in S. warburgii (Larsson et al. 1986). I sampled the leaves in October 2011 and June 2012. Ten leaves per branch were collected and mixed for high pressure liquid chromatography (HPLC) analysis to determine the amount of phenolic glycosides in the leaves. In order to control for the effects of leaf age, I collected only the fifth leaf from the distal part of a branch, and avoided galling leaves and immature leaves (Hjältén et al. 2007). The leaves were immediately sealed in plastic bags and transferred back to the lab in a cooler within 12 hours. In the lab, the leaves were rinsed with water, removed of all debris with paper towels and air-dried (approximately 12 hours at room temperature). The veins were removed from the dried leaves, and the remaining leaf tissue was milled and sieved through a 0.25 mm sieve. The leaf samples were then kept at -20ºC until extraction. 9.

(12) Two to three replicates were measured for each branch during each of the 2 sampling periods. Each replicate was consisted of 5 mg of leaf tissue from the mixture of the original 10 leaves. The 5 mg of leaf tissue was added into 0.6 ml of methanol (CH3OH) and homogenized for 30 seconds. After standing on an ice bath for 15 minutes, the sample was re-homogenized and then centrifuged at 16,000 g for 3 minutes at 4ºC. The supernatant was collected into a micro-centrifuge tube. Another 0.6 ml of methanol was added to the residue, and the sample was homogenized and centrifuged again under the same conditions described above. This process was repeatedly applied to the residue from the previous centrifugation for a total of 3 times. All resulting supernatants were combined and vacuum-centrifuged to evaporate the methanol. The sample pellet was then dissolved in 0.1 ml of methanol and filtered with 0.2 μm nylon syringe filter (4 mm, National Scientific, India). The solution after filtration was then injected into the HPLC machine (LaChrom Elite HPLC Systems L-2000 series, HITACHI, Japan). The column used for HPLC analysis is a 250 mm x 4.6 mm i.d. Thermo Hypersil ODS (C18) Column (5μm, Thermo 30105-254630). The two elution solvents were 1.5% tetrahydrofuran plus 0.25% orthophosphoric acid and methanol. The gradient is listed in Appendix 3 (Julkunen-Tiitto and Sorsa 2001). The flow rate was 0.4 ml/min and the injection volume was 15 μl. Photodiode array detector (HITACHI 890-0443) was used to obtain UV absorbance at 277 nm. ARTHROPOD ABUNDANCE Arboreal arthropods were sampled 3 times throughout the 10.

(13) experiment: September and November 2010 (pre-treatment sample), October 2011 and June 2012. The sampling was performed between 7 am and 5 pm on sunny days. Every leaf and twig on each branch were inspected by two technicians, and all living arthropods found were removed, stored in plastic bags and transferred back to the lab in a cooler within 12 hours. The arthropods in each sample were counted and then oven dried at 40ºC ~60ºC for at least 48 hours. The arthropods were divided into 3 categories according to their feeding guilds: 1) herbivores, i.e. Lepidoptera larvae and sap suckers which include Tingidae (Hemiptera) and Icerya seychellaurm (Hemiptera: Coccoidea); 2) fungivores i.e. larvae and adults of ladybug Illeis koebelei (Coleoptera: Coccinellidae); and 3) predators i.e. spiders (Araneae). The abundance of each arthropod group was standardized by dividing their dry weight with mean original leaf area (Marquis and Whelan 1994, Sipura 1999). In attached appendix I also calculated number of individuals in each arthropod group, and used total leaf area of the branch to standardize the values. STATISTICAL ANALYSES For leaf dry weight, leaf area and herbivory damage, I used the mean of the 10 leaves as a single observation for a branch. The level of phenolic glycosides was measured on mixture of 10 leaves; therefore, the mean of the 2 to 3 replicates was used as a single observation for a branch. Generalized linear mixed model was used to test the effects of bird exclusion, time and their interaction on the leaf dry weight, leaf area, herbivory damage and level of phenolic glycosides, with plant identity 11.

(14) included as a random factor. Although each plant had a pair of bird exclusion and control branch, I did not apply a block design because of missing data from some of the branches (e.g. on a given sampling date, one branch might not have any leaves on it while the other branch did). The time effect was different between different plant response variables. For leaf dry weight, leaf area and herbivory damage, the time effect included 1 pre-treatment period (September-November 2010) and 2 post-treatment periods (October 2011, June 2012). For leaf phenolics, the time effect also included 2 post-treatment periods (October 2011, June 2012). Due to the small sample size and highly skewed data distribution of bud count, flower count and arthropod abundance, I applied non-parametric Kruskal-Wallis tests to compare the differences in these response variables between the bird exclusion and control branches for each of the sampling periods separately. Because many of the plant traits used in this study are likely correlated, I also performed Spearman correlations on leaf dry weight, leaf area and amount of leaf phenolic glycosides for October 2011 and June 2012 respectively, as well as on bud count and flower count for March 2011 and February 2012 respectively.. RESULTS PLANT GROWTH AND REPRODUCTION The bird exclusion had negative impacts on S. warburgii growth and reproduction (Tables 1&2 and Appendices 4&5). For plant growth traits, the bud count was lower in the bird exclusion branches than control 12.

(15) branches during February 2012 (Table 2 & Figure 3c). The mean leaf dry weight was lower in the bird exclusion branches than control branches during October 2011, approximately one year after the treatment (post-hoc comparison, control versus bird exclusion during October 2011, P = 0.034; Figure 3a); similar pattern was found for the mean leaf area (post-hoc comparison, control versus bird exclusion during October 2011, P = 0.059; Figure 3b). Both the total leaf dry weight and leaf area were lower in the bird exclusion branches during October 2011 and June 2012 (Appendices 4&5). For plant reproduction trait, the flower count was lower in the bird exclusion branches than control branches during February 2012 (Table 2 & Figure 4). The effects of bird exclusion on plant growth and reproduction were more evident in October 2011 and February 2012, approximately one year after the bird exclusion but did not persist into the following summer. In addition, there were significant decreasing trends in all growth and reproduction traits from 2011 to 2012 (Figures 3&4). PLANT LEAF PHENOLICS The amount of phenolic glycosides in S. warburgii leaves was consistently lower in the bird exclusion branches than control branches from October 2011 to June 2012 (Table 3 & Figure 5a). This is different from plant growth and reproduction traits for which the effects of bird exclusion were mainly seen in 2011. CORRELATIONS AMONG PLANT TRAITS As expected, the leaf dry weight was positively correlated with leaf area in both October 2011 (rs = 0.92, P < 0.0001, n = 22) and June 2012 13.

(16) (rs = 0.99, P < 0.0001, n = 22). Interestingly, heavier leaves also had higher levels of phenolic glycosides (October 2011: rs = 0.67, P = 0.0087, n = 14; June 2012: rs = 0.70, P = 0.036, n = 9); and to a lesser degree, larger leaves also tended to have higher levels of phenolic glycosides (October 2011: rs = 0.37, P = 0.20, n = 14; June 2012: rs = 0.73, P = 0.025, n = 9). This suggests that the braches with heavier and larger leaves might be more capable of investing in phenolic glycosides, and/or the plants were preferentially allocating phenolic glycosides to branches with heavier and larger leaves. During March 2011 and February 2012, the branches that produced more flowers also produced more buds (March 2011: rs = 0.70, P = 0.0003, n = 22; February 2012: rs = 0.93, P < 0.0001, n = 22), suggesting there were no trade-offs between flowering and subsequent budding. ARTHROPOD ABUNDANCE AND HERBIVORY DAMAGE Although the herbivores showed higher abundance in the bird exclusion branches during October 2011 (Figure 6b), it was not statistically significant (Table 4 and Appendix 6). All 3 groups of arthropods (i.e. herbivores, fungivores and predators) showed an opposite trend of higher abundance in the control branches during June 2012 (Figure 6c, 6f and 6i), but the effects of bird exclusion were generally weak (Table 4). Given the small sample size for arthropod abundances, the power of these statistical tests is likely limited. The bird exclusion had marginal effects on herbivory damage (Table 1), which was higher in the bird exclusion branches during October 2011 (Figure 5b). However, this effect of increased herbivory with bird 14.

(17) exclusion did not last through the second year (Figure 5b).. DISCUSSION The bird exclusion reduced willow growth, reproduction, and increased herbivory. Because the bud count was positively correlated with the flower count, a willow branch with reduced growth was more likely to have lower reproduction and vice versa. Therefore, the overall cascading effects of bird predation on willows could be substantial. Furthermore, the willows decreased chemical defense when the birds were excluded, suggesting they could not compensate for the loss of bird predation by increasing chemical defense. Although I did not find that bird exclusion reduced the abundance of herbivores or other arthropods, the statistical power of the tests on arthropod abundances is likely low due to small sample size. DEFENSIVE PHYTOCHEMICALS IN WILLOWS Despite that production of defensive phytochemicals in willows is quite flexible and can be induced after encountering herbivores (Stolter 2008, Yoneya et al. 2012), chemical defense is generally believed to be costly because it competes with growth and reproduction for the same type of resources (Strauss and Agrawal 1999, Agrawal et al. 2006, Neilson et al. 2013). In a case study of trophic cascades in a willow (S. phylicifolia) food web, the strength of trophic cascades was much more pronounced for the willows without fertilization compared with the fertilized willows (Sipura 1999), suggesting chemical defense in willows could be resource-limited and trophic cascades from predators could be 15.

(18) an important anti-herbivory tool for willows. Therefore, the lower level of phenolics in the bird exclusion branches might be a result of direct resource limitation (i.e. these branches simply did not have the resource to produce phenolics), branch-level trade-offs (i.e. the plants allocated resources to other branches), or a combination of both. In fact, I found the amount of phenolic glycosides to be positively correlated with leaf dry weight and leaf area, suggesting the production of these phytochemicals had substantial costs to the willows. If the willows had increased chemical defense in response to bird exclusion, I might not have seen the negative effects of bird exclusion on willow growth and reproduction. However, my results suggest that the willows were not able to mitigate the negative impacts from bird exclusion by increasing their chemical defense. One limitation of this study is that I was not able to identify specific compounds of the phenolic glycosides in S. warburgii. During the preliminary HPLC analyses, I attempted salicin (D-(-)-salicin, Sigma-Aldrich, Chhina), dimethoxybenzene (1,3-dimethoxybenzene, Sigma-Aldrich, Chhina) and hydroxybenzol alcohol (2-hydroxybenzyl alcohol, Sigma-Aldrich, Chhina) but they did not match any of the compounds extracted from the willow leaves. Because herbivore deterrent effects vary among different phenolic glycosides (Smiley et al. 1985, Tahvanainen et al. 1985, Hjältén et al. 2007), it is difficult to attribute the amount of phenolic glycosides in the willows entirely to increased level of chemical defense. THE CASCADING EFFECTS ON THE ARTHROPODS 16.

(19) The herbivores showed a trend of increased abundance in the bird exclusion branches during October 2011, coinciding with the only time period that herbivory damage was higher in the bird exclusion branches. Intermediate predators could reverse the cascading effects on plants (Casini et al. 2009) or posing only trivial effects on the strength of trophic cascades (Letourneau and Dyer 1998). In this study, I also monitored the abundance of arthropod predators (i.e. spiders), which was the intermediate predators that can feed on herbivores but are themselves prey to the birds. However, the abundance of these intermediate predators was not different between the bird exclusion and control branches, suggesting the spiders were not likely to reverse the cascading effects of the birds found in this study. During June 2012, all 3 arthropod feeding guilds (i.e. herbivores, fungivores and predators) showed the same trend with higher abundance in the control branches. This is not a byproduct of the lower leaf area in the control branches (arthropod abundance was quantified as arthropod dry weight per unit leaf area), because during this time, the leaf area was actually similar between the control and bird exclusion branches. It is plausible, however, that the consistently lower growth of the bird exclusion branches into the second year had made them unattractive to all arthropods. BRANCH-LEVEL RESOURCE ALLOCATION IN WILLOWS In the study by Stolter (2008), poplar trees increased chemical defense and reduced palatability of new grown twigs, but at the same time, they decreased chemical defense and increased palatability of 17.

(20) heavily browsed twigs. This suggests plants can use branch-level trade-offs to manage overall herbivory. Although my experimental design did not allow partitioning between branch-level and tree-level effects of bird exclusion, it is more likely that my findings reflected the plants’ resource allocation between the 2 branches experiencing different level of pressures. There are two reasons for this speculation: 1) the willows used in this study were adult trees having many branches at different levels of the canopy; and 2) there were branch-level differences in sprouting and shedding patterns (e.g. the branches in the lower canopy were more likely to shed leaves, and less likely to sprout). TEMPORAL SCALES OF TROPHIC CASCADES There were substantial time effects on most of the plant and arthropod response variables in this study. Furthermore, the effects of bird exclusion were not always consistent between the two post-treatment sampling periods. For instances, herbivory damage was higher in the bird exclusion branches in October 2011 but not in June 2012; flower counts were lower in the bird exclusion branches in February 2012 but not in March 2011. As pointed out by Borer et al. (2005) in a meta-analysis of trophic cascades, the length of the study period can affect the outcome of trophic cascades. My results also support the idea that trophic cascading effects could be sensitive to the length of the study. More longer-term studies, especially on adult trees, may reveal surprising complexity in the patterns of trophic cascades.. 18.

(21) TABLES AND FIGURES Table 1. The effects of bird exclusion, sampling time and their interaction on mean leaf dry weight, mean leaf area and herbivory damage in Salix warburgii. Plant identity was included as a random factor. In October 2011 and June 2012, there were 8 and 13 branches that did not have any leaves on them. I excluded these data points from this analysis. The herbivory damage was square-root transformed prior to the analysis.. Effect. Numerator. Denominator. DF. DF. bird exclusion. 1. time bird exclusion x time. Mean leaf dry weight. Mean leaf area. Herbivory damage. F. P. F. P. F. P. 29. 5.26. 0.029. 4.89. 0.035. 3.35. 0.077. 2. 29. 17.48. < 0.0001. 8.94. 0.0009. 0.26. 0.77. 2. 29. 0.69. 0.51. 0.27. 0.76. 0.97. 0.39. 19.

(22) Table 2. The effects of bird exclusion on bud and flower counts in Salix warburgii.. Sampling time Ncontrol Nbird-exclusion DF. Bud count. Flower count. χ2. P. χ2. P. 0.25. 1.40. 0.24. March 2011. 11. 11. 1. 1.32. February 2012. 11. 11. 1. 6.46 0.011 6.91 0.0086. 20.

(23) Table 3. The effects of bird exclusion, sampling time and their interaction on the amount of phenolic glycosides in Salix warburgii. Plant identity was included as a random factor. In October 2011 and June 2012, there were 8 and 13 branches that did not have any leaves on them; therefore, phenolics data were not available. The amount of phenolic glycosides was measured as the mean peak area of 2-3 replicate runs (Appendix 2).. Effect. Numerator DF Denominator DF. F. P. bird exclusion. 1. 12. 7.48 0.018. time. 1. 12. 0.06. 0.81. bird exclusion x time. 1. 12. 0.46. 0.51. 21.

(24) Table 4. The effects of bird exclusion on the abundances of herbivorous, fungivorous and predatory arthropods. The abundance of each of 3 arthropod groups was standardized by dividing their dry weight with the leaf area (mg/100 cm2). In October 2011 and June 2012, there were 8 and 13 branches that did not have any leaves on them. I excluded these data points from this analysis because they could not provide leaf area data.. Sampling time. Ncontrol Nbird-exclusion. DF. Herbivore. Fungivore. Predator. χ2. P. χ2. P. χ2. P. pre-treatment: September-November 2010. 11. 11. 1. 0.34. 0.56. 0.25. 0.62. 0.45. 0.50. post-treatment: October 2011. 8. 6. 1. 0.48. 0.49. 0.37. 0.55. 0.16. 0.69. post-treatment: June 2012. 7. 2. 1. 0.78. 0.38. 0.34. 0.56. 0.34. 0.56. 22.

(25) Figure 1. The satellite image of the study site. The study site was located in a riparian zone along Xindian river in northern Taiwan (25º00’52’’ N to 25º01’00’’ N and 121º29’24’’ E to 121º29’29’’ E). The text labels show the IDs and locations of the 11 experimental trees (see Appendix 1).. 23.

(26) Figure 2. The sampling design and Salix warburgii phenology based on the 11 experimental trees. The ID of each experimental tree is labeled with a letter “A” followed by a two-digit number (e.g. A01). The horizontal, thick-black bars denote leafing time periods whereas the thin-black bars denote leaf-shedding periods. The diamond symbols denote the start of the flowering season, which lasted for approximately 2 months. The columns with lighter shades were the samplings of leaf dry weight, leaf area, herbivory damage and arthropod abundance; the columns with darker shades were the samplings of bud and flower counts.. The bird exclusion was applied to 8 of the 11 trees in September 2010 (i.e. A01-A07, A09) and to the remaining 3 in November 2010 (i.e. A08, A10, A12). The unfilled portion of the horizontal bars for tree A01, A06, A09 and A10 denotes the time period when both branches on these trees had 24.

(27) zero leaf count.. 25.

(28) Figure 3. The effects of bird exclusion on Salix warburgii growth. a) leaf dry weight, b) leaf area, and c) bud count. The grey columns in a) and b) denote pre-treatment values sampled during September and November 2010; the error bars denote standard errors and the asterisks denote significant differences between the control and bird exclusion branches for a given time period at P < 0.05. The boxes and whiskers in c) denote 25-75th and 5-95th quantiles respectively.. 26.

(29) Figure 4. The effects of bird exclusion on Salix warburgii reproduction. The boxes and whiskers denote 25-75th and 5-95th quantiles respectively.. 27.

(30) Figure 5. The effects of bird exclusion on chemical defense and herbivory in Salix warburgii. a) amount of phenolic glycosides and b) percent herbivory damage. The grey column denotes pre-treatment values sampled during September and November 2010. The error bars denote standard errors and the asterisks denote significant differences between the control and bird exclusion branches for a given time period at P < 0.05.. 28.

(31) Figure 6. The arthropod abundance on Salix warburgii between the control and bird exclusion branches across sampling times. The Sep.-Nov. 2010 values were pre-treatment values. The boxes denote 25-75th quantiles and the whiskers denote 5-95th quantiles.. 29.

(32) REFERENCES Agrawal, A. A., J. A. Lau, and P. A. Hambäck. 2006. Community heterogeneity and the evolution of interactions between plants and insect herbivores. Quarterly Review of Biology 81:349-376. Böhm, S. M., K. Wells, and E. K. Kalko. 2011. Top-down control of herbivory by birds and bats in the canopy of temperate broad-leaved oaks (Quercus robur). Plos One 6:e17857. Boeckler, G. A., J. Gershenzon, and S. B. Unsicker. 2011. Phenolic glycosides of the Salicaceae and their role as anti-herbivore defenses. Phytochemistry 72:1497-1509. Borer, E. T., E. W. Seabloom, J. B. Shurin, K. E. Anderson, C. A. Blanchette, B. Broitman, S. D. Cooper and B. S. Halpern. 2005. What determines the strength of a trophic cascade? Ecology 86:528-537. Casini, M., J. Hjelm, J.-C. Molinero, J. Lövgren, M. Cardinale, V. Bartolino, A. Belgrano, and G. Kornilovs. 2009. Trophic cascades promote threshold-like shifts in pelagic marine ecosystems. Proceedings of the National Academy of Sciences 106:197-202. Chang, W. F. 1985. A field guide to the birds of Taiwan. Taiwan: Bird Image Publication (in Chinese). Coley, P. D., and J. A. Barone. 1996. Herbivory and plant defenses in tropical forests. Annual Review of Ecology and Systematics 27:305-335. Crawley, M. J. 1989. Insect herbivores and plant population dynamics. Annual Review of Entomology 34:531-562. Donaldson, J. R., M. T. Stevens, H. R. Barnhill, and R. L. Lindroth. 2006. Age-related shifts in leaf chemistry of clonal aspen (Populus tremuloides). Journal of Chemical Ecology 32:1415-1429. Dudt, J. F., and D. J. Shure. 1994. The influence of light and nutrients on foliar phenolics and insect herbivory. Ecology 75:86-98. Fine, P. V., I. Mesones, and P. D. Coley. 2004. Herbivores promote habitat specialization by trees in Amazonian forests. Science 305:663-665. Fritz, R. S., C. G. Hochwender, D. A. Lewkiewicz, S. Bothwell, and C. M. Orians. 2001. Seedling herbivory by slugs in a willow hybrid system: developmental changes in damage, chemical defense, and plant performance. Oecologia 129:87-97. Greenberg, R., P. Bichier, A. C. Angon, C. MacVean, R. Perez, and E. Cano. 2000. 30.

(33) The impact of avian insectivory on arthropods and leaf damage in some Guatemalan coffee plantations. Ecology 81:1750-1755. Griffin, C. A., and J. S. Thaler. 2006. Insect predators affect plant resistance via density‐ and trait‐mediated indirect interactions. Ecology Letters 9:335-343. Hairston, N. G., F. E. Smith, and L. B. Slobodkin. 1960. Community structure, population control, and competition. The American Naturalist 94:421-425. Haukioja, E., V. Ossipov, J. Koricheva, T. Honkanen, S. Larsson, and K. Lempa. 1998. Biosynthetic origin of carbon-based secondary compounds: cause of variable responses of woody plants to fertilization? Chemoecology 8:133-139. Hjältén, J., L. Niemi, A. Wennström, L. Ericson, H. Roininen, and R. Julkunen-Tiitto. 2007. Variable responses of natural enemies to Salix triandra phenotypes with different secondary chemistry. Oikos 116:751-758. Ho, C.-K., and S. C. Pennings. 2008. Consequences of omnivory for trophic interactions on a salt marsh shrub. Ecology 89:1714-1722. Hunter, M. D., and P. W. Price. 1992. Playing chutes and ladders: heterogeneity and the relative roles of bottom-up and top-down forces in natural communities. Ecology 73:724-732. Hunter, M. D., G. C. Varley, and G. R. Gradwell. 1997. Estimating the relative roles of top-down and bottom-up forces on insect herbivore populations: a classic study revisited. Proceedings of the National Academy of Sciences 94:9176-9181. Huryn, A. D. 1998. Ecosystem-level evidence for top-down and bottom-up control of production in a grassland stream system. Oecologia 115:173-183. Jeppesen, E., M. Søndergaard, J. P. Jensen, E. Mortensen, A.-M. Hansen, and T. Jørgensen. 1998. Cascading trophic interactions from fish to bacteria and nutrients after reduced sewage loading: An 18-year study of a shallow hypertrophic lake. Ecosystems 1:250-267. Julkunen-Tiitto, R., and S. Sorsa. 2001. Testing the effects of drying methods on willow flavonoids, tannins, and salicylates. Journal of Chemical Ecology 27:779-789. Larsson, S., A. Wirén, L. Lundgren, and T. Ericsson. 1986. Effects of light and nutrient stress on leaf phenolic chemistry in Salix dasyclados and susceptibility to Galerucella lineola (Coleoptera). Oikos 47:205-210. Letourneau, D. K., and L. A. Dyer. 1998. Density patterns of Piper ant-plants and associated arthropods: top-predator trophic cascades in a terrestrial system? Biotropica, 30:162-169. Marquis, R. J., and C. J. Whelan. 1994. Insectivorous birds increase growth of white oak through consumption of leaf-chewing insects. Ecology 75:2007-2014. 31.

(34) Menge, B. A. 1995. Indirect effects in marine rocky intertidal interaction websL patterns and importance. Ecological Monographs 65:21-74. Mooney, K. A. 2007. Tritrophic effects of birds and ants on a canopy food web, tree growth, and phytochemistry. Ecology 88:2005-2014. Neilson, E. H., J. Q. Goodger, I. E. Woodrow, and B. L. Møller. 2013. Plant chemical defense: at what cost? Trends in Plant Science 18:250-258. Oksanen, L., S. D. Fretwell, J. Arruda, and P. Niemelä. 1981. Exploitation ecosystems in gradients of primary productivity. The American Naturalist 118:240-261. Orians, C. M., and R. S. Fritz. 1995. Secondary chemistry of hybrid and parental willows: Phenolic glycosides and condensed tannins inSalix sericea, S. eriocephala, and their hybrids. Journal of Chemical Ecology 21:1245-1253. Persson, L. 1999. Trophic cascades: abiding heterogeneity and the trophic level concept at the end of the road. Oikos 85:385-397. Power, M. E. 1992. Top-down and bottom-up forces in food webs: do plants have primacy. Ecology 73:733-746. Rowell-Rahier, M., and J. M. Pasteels. 1990. Phenolglucosides and interactions at three trophic levels: Salicaceae-herbivores-predators. Insect-Plant Interactions, Vol. 2 (eds E. A. Bernay), pp. 75-94. CRC Press, Boca Raton, Florida. Schmitz, O. J., E. L. Kalies, and M. G. Booth. 2006. Alternative dynamic regimes and trophic control of plant succession. Ecosystems 9:659-672. Shurin, J. B., E. T. Borer, E. W. Seabloom, K. Anderson, C. A. Blanchette, B. Broitman, S. D. Cooper, and B. S. Halpern. 2002. A cross-ecosystem comparison of the strength of trophic cascades. Ecology Letters 5:785-791. Sipura, M. 1999. Tritrophic interactions: willows, herbivorous insects and insectivorous birds. Oecologia 121:537-545. Smiley J. T., J. M. Horn and N. E. Rank. 1985. Ecological effects of salicin at three trophic levels: new problems from old adaptations. Science(Washinton) 229:649-651. Spiller, D. A., and T. W. Schoener. 1994. Effects of top and intermediate predators in a terrestrial food web. Ecology 75:182-196. Stamp, N. E., and M. D. Bowers. 2000. Do enemies of herbivores influence plant growth and chemistry? Evidence from a seminatural experiment. Journal of Chemical Ecology 26:2367-2386. Stolter, C. 2008. Intra-individual plant response to moose browsing: feedback loops and impacts on multiple consumers. Ecological Monographs 78:167-183. Strauss, S. Y., and A. A. Agrawal. 1999. The ecology and evolution of plant tolerance to herbivory. Trends in Ecology and Evolution 14:179-185. Tahvanainen, J., R. Julkunen-Tiitto, and J. Kettunen. 1985. Phenolic glycosides 32.

(35) govern the food selection pattern of willow feeding leaf beetles. Oecologia 67:52-56. Trussell, G. C., P. J. Ewanchuk, and M. D. Bertness. 2002. Field evidence of trait‐ mediated indirect interactions in a rocky intertidal food web. Ecology Letters 5:241-245. Huang T.C., D.E. Boufford, H. Ohashi, Y.P. Yang and S.Y. Lu. 1996. Flora of Taiwan, Secon Edition. Editorial Committee of the Flora of Taiwan, Secon Edition, Taipei, Taiwan, ROC. Yoneya, K., M. Uefune, and J. Takabayashi. 2012. An apparent trade-off between direct and signal-based induced indirect defence against herbivores in willow trees. Plos One 7:e51505.. 33.

(36) APPENDICES Appendix 1. The Salix warburgii used in this study. The heights of the trees were estimated by visual inspection. The letters C and B in the parenthesis following the branch height denote the control and bird exclusion branches respectively.. Plant ID. Treatment date. Sex. Tree height (m) DBH (cm). Longitude/ latitude. Branch height (cm). A01. 2010/9/1. M. 7. 114. N25 00.989 E121 29.424. 203 (C), 201 (B). A02. 2010/9/3. M. 4. 40.5. N25 00.986 E121 29.415. 77 (C), 72 (B). A03. 2010/9/3. M. 4. 84.5. N25 00.942 E121 29.421. 200 (C), 200 (B). A04. 2010/9/3. F. 3. 30. N25 00.950 E121 29.442. 110 (C), 130 (B). A05. 2010/9/3. F. 6. 79.5. N25 00.932 E121 29.460. 208 (C), 210 (B). A06. 2010/9/4. M. 3.5. 23.2. N25 00.975 E121 29.406. 120 (C), 132 (B). A07. 2010/9/4. M. 6. 186.9. N25 00.982 E121 29.427. 196 (C), 217 (B). A08. 2010/11/11. M. 5. 251.1. N25 01.004 E121 29.425. 121 (C), 203 (B). A09. 2010/9/27. M. 2.5. 81.8. N25 00.957 E121 29.407. 179 (C), 164 (B). A10. 2010/11/11. M. 7. 127.1. N25 00.947 E121 29.414. 187 (C), 196 (B). A12. 2010/11/29. F. 5. 267.4. N25 01.003 E121 29.417. 183 (C), 201 (B). 34.

(37) Appendix 2. Examples of the leaf images used to estimate herbivory damage in Salix warburgii. The leaf images were from the control branches (a,b) and bird exclusion branches (c,d) on the same experimental tree (plant ID A05) sampled in October 2011. All images were calibrated with the same ruler scanned with leaves. (a) and (c) are intact leaf images used for leaf area measurements, and (b) and (d) are the leaf images of (a) and (c) respectively with their chew holes manually filled.. 35.

(38) Appendix 3. The gradients of two elution solvents used in the HPLC analysis of leaf phenolic glycosides. Elution solvent A contains 1.5% tetrahydrofuran plus 0.25% orthophosphoric acid, and elution solvent B is 100% methanol. The injection volume was 15 μl and the flow rate was 0.4 ml/min for the whole process.. Time (min) 0 5 10 20 30 40 45 45.1 50 50.1 55 65. Solvent A (%) 100 100 80 70 65 50 50 0 0 100 100 Ending. 36. Solvent B (%) 0 0 20 30 35 50 50 100 100 0 0 Ending.

(39) Appendix 4. The effects of bird exclusion, sampling time and their interaction on total leaf dry weight and total leaf area in Salix warburgii. Plant identity was included as a random factor. In October 2011 and June 2012, there were 8 and 13 branches that did not have any leaves on them. I excluded these data points from this analysis.. Effect. Numerator. Denominator. DF. DF. bird exclusion. 1. time bird exclusion x time. Leaf dry weight. Leaf area. F. P. F. P. 29. 7.52. 0.010. 8.69. 0.0063. 2. 29. 1.16. 0.33. 0.64. 0.53. 2. 29. 2.33. 0.12. 2.26. 0.12. 37.

(40) Appendix 5. The effects of bird exclusion on Salix warburgii total leaf dry weight and leaf area. a) total leaf dry weight and b) total leaf area. The grey columns in a) and b) denote pre-treatment values sampled during September and November 2010; the error bars denote standard errors and the asterisks denote significant differences between the control and bird exclusion branches for a given time period at P < 0.05.. 38.

(41) Appendix 6. The effects of bird exclusion on the abundances of herbivorous, fungivorous and predatory arthropods. Three arthropod abundance indices were used: a) total arthropod dry weight divided by total leaf area, b) total number of individual arthropods divided by mean leaf area, and c) total number of individual arthropods divided by total leaf area. In October 2011 and June 2012, there were 8 and 13 branches that did not have any leaves on them. I excluded these data points from this analysis.. a) Total dry weight / total leaf area Sampling time. Ncontrol Nbird-exclusion. DF. Herbivore. Fungivore. Predator. χ2. P. χ2. P. χ2. P. pre-treatment: September-November 2010. 11. 11. 1. 0.16. 0.69. 0.25. 0.62. 0.45. 0.50. post-treatment: October 2011. 8. 6. 1. 0.48. 0.49. 0.0045. 0.95. 0.44. 0.51. post-treatment: June 2012. 7. 2. 1. 0.78. 0.38. 0.00. 1.00. 0.086. 0.77. 39.

(42) b) Number of individuals / mean leaf area Sampling time. Ncontrol Nbird-exclusion. DF. Herbivore. Fungivore. Predator. χ2. P. χ2. P. χ2. P. pre-treatment: September-November 2010. 11. 11. 1. 0.45. 0.50. 0.25. 0.62. 0.34. 0.56. post-treatment: October 2011. 8. 6. 1. 0.076. 0.78. 0.55. 0.46. 0.44. 0.51. post-treatment: June 2012. 7. 2. 1. 0.78. 0.38. 0.086. 0.77. 0.086. 0.77. c) Number of individuals / total leaf area Sampling time. pre-treatment: September-November 2010. Ncontrol Nbird-exclusion. 11. 11. DF. 1. 40. Herbivore. Fungivore. Predator. χ2. P. χ2. P. χ2. P. 0.24. 0.62. 0.25. 0.62. 0.24. 0.62.

(43) post-treatment: October 2011. 8. 6. 1. 0.076. 0.78. 0.0045. 0.95. 1.12. 0.29. post-treatment: June 2012. 7. 2. 1. 0.00. 1.00. 0.086. 0.77. 0.34. 0.56. 41.

(44) Appendix 7. The means, standard deviations (SD) and coefficients of variation (CV) in peak areas among replicate runs in the HPLC analysis of leaf phenolic glycosides. The replicate runs comprised 2-3 subsamples of the leaves taken from the same branch on the same date; the replicates were analyzed in the same run of HPLC.. Treatment Control Bird exclusion Control Bird exclusion Control Bird exclusion Control Bird exclusion Control Control Bird exclusion Control Control Bird exclusion Mean CV. Plant ID Sampling time A02C October 2011 June 2012 A02T October 2011 A03C October 2011 June 2012 A03T October 2011 June 2012 A04C October 2011 June 2012 A04T October 2011 A05C October 2011 June 2012 A05T October 2011 A07C October 2011 June 2012 A08C October 2011 June 2012 A08T October 2011 June 2012 A09C October 2011 A12C October 2011 June 2012 A12T October 2011. 42. N 2 3 2 2 3 2 2 2 3 2 2 3 2 2 3 2 3 2 3 2 2 3 2. Mean 13250196 21988986 6421031 8421293 7558796 8253328 3809644 9547288 8981247 6621099 17801886 18384096 20095076 20565722 11400180 18979404 19328593 11625652 11990053 12118129 13924206 22745930 7414309. SD 773533 158198 580880 499211 352589 1698645 284733 830174 42036 72661 63353 681518 455611 216384 368057 982097 716676 32707 218369 1304662 679408 234704 17256. CV% 5.84 0.72 9.05 5.93 4.66 20.58 7.47 8.70 0.47 1.10 0.36 3.71 2.27 1.05 3.23 5.17 3.71 0.28 1.82 10.77 4.88 1.03 0.23 4.48.

(45) Appendix 8. The leaf traits, herbivory damage and arthropod dry weights of the 11 experimental trees (Salix warburgii) used in this study. The letters C and B in the first column denote the control and bird exclusion branches respectively. Leaf phenolic glycosides were not sampled during September and November 2010; a zero leaf count was marked with a dash, for which leaf dry weight, leaf area, herbivory damage and arthropod abundances were not estimated.. Treatment Plant ID. Leaf dry Whole Peak area of Herbivory Herbivore Fungivore Predator Leaf area weight leaf area leaf phenolic damage dry weight dry weight dry weight (cm2) (mg) (cm2) glycosides (%) (mg) (mg) (mg). September to November 2010 A01 52.6. 12.26. 12.49. NA. 2. 0. 0. 0. A02 A03 A04 A05 A06 A07 A08 A09 A10. 111 116.9 70.8 112.2 115.1 120.7 149.4 93.8 133.6. 16.61 14.44 10.02 19.06 15.56 16.48 16.09 16.21 16.01. 17.15 15.57 10.06 19.25 15.78 16.97 18.24 16.4 17.25. NA NA NA NA NA NA NA NA NA. 3 7 0 1 1 3 12 1 7. 0 0 0 0 0 0 0.27 1.1 0. 0 0 0 0 0 0 0 0 0. 0 0 0 0 0 0 0 7.97 9.04. A12. 121.5. 13.51. 13.74. NA. 2. 0. 30.4. 0.4. A01 A02 A03 A04 A05 A06 A07 A08. 62.4 78.4 105 68.6 154.9 110.6 113.7 151.1. 12.89 14.84 13.15 9.53 13.06 15.65 17.92 15.28. 13.82 15.79 13.22 10.14 14.66 15.74 18.03 15.57. NA NA NA NA NA NA NA NA. 7 6 1 6 11 1 1 2. 0 0 0 0 0 0 0 0.38. 0 0 0 0 0 0 0 1.72. 0 0 0 0 0 0 0 2.34. A09. 88.7. 13.29. 14.03. NA. 5. 0.38. 0. 0. C. B. 43.

(46) Treatment Plant ID. Leaf dry Whole Peak area of Herbivory Herbivore Fungivore Predator Leaf area weight leaf area leaf phenolic damage dry weight dry weight dry weight (cm2) (mg) (cm2) glycosides (%) (mg) (mg) (mg). A10 A12. 88.9 102.7. 15.56 12.99. 16.57 13.25. NA NA. 6 2. 22.3 0. 3.71 0. 0.15 0. October 2011 A01 A02. 131.3. 20.19. 20.78. 12049840. 3. 0 0. 0 1.16. 0.25 0.9. A03 A04 A05 A06 A07. 99.5 66.3 100.6 106.8. 14.48 8.66 14.74 14.42. 14.37 9.02 16.05 14.71. 7892449 9547288 17469969 19555446. 0 4 8 2. 0.11 0.2 0 0 0. 0 2.51 0 0 0. 0 0 2.52 0 0. A08 A09 A10 A12. 150.6 101.8 108.6. 19.79 17.48 19.25. 19.88 17.67 20.03. 17688956 11959454 13986735. 0 1 4. 0.621 3.68 0 0. 0.61 3.48 0 16.41. 0.38 6.34 0 0.83. A01 A02 A03 A04 A05 A06 A07 A08 A09. 87 61.4 73 114.4 134 -. 14.71 10.9 10.57 15 15.58 -. 17.65 11.48 10.98 15.24 16.02 -. 5858798 8021510 6218784 19050121 10185440 -. 17 5 4 2 3 -. 0 0 17.15 0 0.72 0 0 14.7 0. 0 0 0 0 1.32 0 0 1.59 0. 7.83 16.12 1.68 0 0.95 12.69 0 0.49 0. A10 A12. 70.1. 15.96. 18.68. 7750161. 15. 0 0. 0 1.86. 1.69 0. A01 A02 A03 A04 A05. 77.1 47.7 36.9 54.6. 13 10.06 6.02 8.5. 13.13 10.28 6.65 8.58. 21988986 7558796 8981247 18384096. 1 2 9 1. 0 24.2 0 14.4 0.36. 0 3.46 1.22 3.22 1.3. 0 9.1 0.35 1.61 1.3. A06. -. -. -. -. -. 0. 0. 0. C. B. June 2012. C. 44.

(47) Treatment Plant ID. B. Leaf dry Whole Peak area of Herbivory Herbivore Fungivore Predator Leaf area weight leaf area leaf phenolic damage dry weight dry weight dry weight (cm2) (mg) (cm2) glycosides (%) (mg) (mg) (mg). A07 A08 A09 A10 A12. 80.6 118.8 86.6. 12.98 16.16 15.66. 14.87 16.62 15.91. 11400180 19328593 22745930. 13 3 2. 3.17 12.92 0 0 0. 11.7 14.22 0 0 2.24. 1.85 11.08 0 0 0.37. A01 A02 A03 A04 A05 A06 A07 A08 A09. 34.2 107.1 -. 6.74 14.58 -. 7.11 15.35 -. 3672326 11990053 -. 5 5 -. 0 0 0.16 0 0 0 0 0.36 7.1. 0 0 2.23 0 0 0 0 1.94 0. 0 0 0.69 0 0 0 0 2.15 0. A10 A12. -. -. -. -. -. 0 0. 0 0. 0 8.8. 45.

(48) Appendix 9. The bud and flower counts of the 11 experimental trees (Salix warburgii) used in this study. Plant ID. March 2011 A01. Bud count Control Bird exclusion. Flower count Control Bird exclusion. 165. 161. 50. 63. A02 A03 A04 A05 A06. 588 582 898 899 1266. 1376 536 755 822 1495. 113 441 773 694 1185. 5 436 579 587 1164. A07 A08 A09 A10. 1812 1264 1082 423. 486 500 700 322. 1069 539 528 111. 386 457 266 65. A12. 1327. 677. 841. 539. February 2012 A01 0 A02 290 A03 145 A04 290 A05 135 A06 0. 0 0 105 0 0 0. 0 30 10 137 74 0. 0 0 0 0 0 0. A07. 610. 0. 344. 0. A08 A09 A10 A12. 395 0 0 869. 65 0 0 0. 442 0 0 224. 48 0 0 0. 46.

(49)