1. INTRODUCTION

1.1 BACKGROUND

Lianas are woody, relatively thick-stemmed climbers that begin life as terrestrial seedlings (Gentry, 1991). They are a common feature of many tropical and temperate forests. Although lianas are often found in temperate forests, their contribution to forest abundance, diversity and structure is most substantial in the tropics (Schnitzer et al, 2002). Lianas are viewed as structural parasites as they rely on rigid host plants (usually trees) for mechanical support and this aides their ascension into the forest canopy (Gentry, 1991). Lianas allocate little to structural support, the majority of plant resources being used for reproduction, canopy development, stem and root elongation (Schnitzer et al, 2002). This results in a greater canopy to stem ratio and a resulting greater proportion of photosynthetic biomass than is present in many other woody plants (Schnitzer et al, 2002).

There is wide consensus in the scientific literature over the importance of lianas. Even so, they are a much neglected species. (Clark, et al. 1990; Gentry, 1991; Putz, 1984; Schnitzer, et al 2002). Studies have illustrated the increasingly important roles of lianas in forest regeneration and competition, species diversity and ecosystem-level processes, for example, transpiration and carbon sequestration (Schnitzer et al, 2002). Although climbers constitute approximately one-half of the families of vascular plants (Putz, 1984), our current understanding of lianas and their role in forest dynamics has lagged well behind that of most other vascular plants (Schnitzer et al, 2002). Schnitzer et al further mention that this hiatus could be a result of difficulties in studying the erratic growth patterns, rampant vegetative reproduction and taxonomic uncertainties that are associated with these plants. An example of this is the surveying problems involved with studying lianas because they grow both horizontally in the undergrowth as well as vertically into the canopy (personal observation, 2002). This may include climbing up adjoining trees (personal observation, 2002).

Studies of lianas have mainly focused on host selectivity, their general ecology and role in forests, their possible effects on their host plants and any defence mechanisms used by the hosts against them.

This project aims to study the distributional characteristics of lianas. It is important to understand the factors influencing the distribution of Hedera helix, the study species, as it is a highly prized ornamental landscape plant that has been introduced to Australia, New Zealand, Hawaii, Brazil and North America. It has now established and naturalised in these areas. The rampant vegetative reproduction and adaptability to a wide range of conditions has ensured the continual invasion of this species. In many habitats, native species and forest communities are threatened due to the development of ‘ivy deserts’. The problem has now reached global scale proportions (Okerman 2001) and understanding of the distributional characteristics of H. helix is fundamental for any control measures to be taken for the management of H. helix in these areas.

1.2 PROJECT AIMS

This survey of Hedera helix was conducted to examine its distributional characteristics. The principal objective of the survey was to determine if there was a non-random distribution of H. helix in the wood. Thus the hypothesis may be raised, as to whether H. helix exhibits host preference or not. Several possible factors influencing distribution are studied in this dissertation, namely; host selectivity, host bark texture, host size, ground cover availability, the ‘edge effect’ and the influence of light.

Regarding the host selectivity of Hedera helix, the following questions were posed prior to investigation:

1. Does H. helix display preference for particular tree species?

2. Does host size influence H. helix colonisation?

3. Are some tree species more likely to have mature reproductive vines on them

than others?

Secondly the influence of bark texture on the distribution of Hedera helix was examined. Many authors suggest that bark texture may affect distributional patterns of H. helix. Putz (1984), Stevens (1987) and Talley et al (1996) all found that those trees with rough bark rather than smooth or flaky bark appear to increase liana success. Therefore, the following questions were posed:

1. Does bark texture influence the distribution of H. helix?

2. If so, what constitutes favourable bark texture?

3. Does a combination of host size and bark texture, together influence the distribution of Hedera helix?

Thirdly, the ground cover by Hedera helix was studied. Lianas are not restricted to growing vertically (Putz, 1984) and can often be seen growing as a thick carpet in the under storey. It was considered that the distribution of H. helix ground cover may reflect any patterns of host preference. Therefore, questions asked included:

1. Is there more ground cover under those trees supporting H. helix as

opposed to trees not supporting H. helix?

2. Is H. helix ground cover distributed more abundantly under certain tree

species than others (i.e. is it species specific)?

3. Does host size affect the amount of H. helix ground cover found at the base of the host (i.e. is it size specific)?

Fourthly, location within the forest was examined for influences on H. helix distribution and the question posed;

Is the abundance of H. Helix greater at the forest edge than interior (for both ground cover and trunk cover)?

Finally, the effect of light was examined in relation to H. helix ground cover and trunk cover.

1.3 STUDY SPECIES (Appendix 1a)

Hedera helix (common ivy) belongs to Araliaceae, a family containing 57 genera. It is a hardy evergreen climber, originating in Europe (Key, 1999). Due to its evergreen nature, the plants are very obvious during the winter months (personal observation, 2002). H. helix can grow in extreme soil conditions both basic and acidic. It prefers moist rather than waterlogged soils and is adaptable to different levels of light (Okerman, 2001). However, it is most characteristic of shaded habitats (Grime et al, 1988).

Hedera helix experiences two distinct growth forms known as dimorphism or heteroblasty (Lee et al, 1991). The vines undergo a juvenile and mature phase, with both often evident on one plant (Key, 1999). During the juvenile (non-reproductive) form, H. helix is typically found growing along the ground. Once it reaches an appropriate trellis, it proceeds to grow up into the canopy to reach the sunlight (Key, 1999). The plant ascends the host with the use of adventitious-roots (Appendix 1b). These are clinging rootlets that exude a glue-like substance to support the plant without penetrating the bark of the host (Toman, et al 1990). At this stage, the leaves are characterised by palmate 3-5 lobes which are alternate and simple (Key, 1999; Russell et al, 1997; personal observation, 2002). The stem is a light green colour. (Appendix 1a, 1c).

Transition of Hedera helix to the adult (reproductive) form occurs once the ivy has reached a considerable height (usually after two years and up to 30m). It has been linked with reductions in gibberellic acid caused by the absence of abundant roots (Okerman, 2001). The stems thicken and become arboreal. It develops bushy non-climbing aerial shoots as opposed to adventitious roots. These shoots drape towards the ground and begin to flower and berry. This gives the plant a ‘shrub-like’ appearance. The leaves at this stage lose their shape and become narrow and elliptic. The change from a lobed leaf to a narrow, elliptic one is assumed to give the plant a greater efficiency in catching the sunlight necessary for photosynthesis (Key, 1999), as shown in Appendix 1a and 1c. An unlobed leaf is also better adapted to survive wind exposure than a wide-lobed leaf and allows the flower heads more exposure to the attentions of fertilising flies and late autumn wasps.

During the mature phase of Hedera helix, the bark of the stem is light brown, slightly rough and finely scaly with numerous rootlets at the nodes. These are typically only seen on large climbing stems (Seiler, et al 2001). During this study, diameters of H. helix stems reached 155mm and the stems of these vines appeared to be very ‘tree-like’. Older vines are known to reach a foot in diameter (Swearingen, 2000).

Hedera helix flowers from September to November. The flowers are yellow blossoms, with five petals that are arranged in rounded umbels’ (Toman et al, 1990) and are sweetly perfumed (Key, 1999). Following the flowering stage, the H. helix produces berries during the winter months. These ripen in April to June of the following year and are spherical and black with a fleshy outer covering. The berries of H. helix are mildly toxic. This discourages consumption of too many berries by birds during dispersal, resulting in only a few seeds deposited at any one time. The residual time of H. Helix seeds in the stomach of birds is short enough so that seeds are more viable when they are released (Okerman, 2001).

1.4 LIST OF ABBREVIATIONS AND ACRONYMS

CBH Circumference Breast Height (1.37m from ground level)

DBH Diameter Breast Height (1.37m from ground level).

OS Ordonance Survey

2. METHODOLOGY

2.1 STUDY SITE

The study site is located in Drayton Wood, Norwich, England (Grid reference TG 195 128 – OS 1:50,000 sheet 134). Located on the boundary parishes of Drayton and Hellesdon, the wood encompasses an area of 11.1 hectares (County Wildlife Survey). Refer to Appendix 2. The site is listed as a County Wildlife Site (Ref. No: 2022) and is managed by the Norwich County Council. The area is open to public access, with an extensive network of informal footpaths throughout, including a surfaced footpath that leads from Drayton High Road to the residential dwelling at the rear of the wood. (County Wildlife Survey).

The wood is mixed (plantation and secondary) and semi-natural in character. It is predominately high forest and dry in character. The wood comprises of even-aged canopy areas dominated by Acer pseudoplatanus, Quercus robur and Pinus sylvestris. Others common species include Betula pendula, Fraxinus excelsior, Pseudotsuga menziesii and Larix deciduas. The wood is scattered with numerous ornamental/alien shrubs and herbaceous plants which were planted earlier in this century as an extension to a garden, which is adjacent to the wood’s eastern boundary. The ground flora in the woodland is generally poor with much Urtica dioica and Hedera helix (Davies, 2000). Refer to Appendix 3. The wood lies on permeable, acidic sandy soils with low available water reserves and moderately sloping ground. The parent material consists of glaciofluvial sands and gravels (Eldridge, 1980).

The area is presently managed by the Norwich County Council. Current silviculture activities include a thinning programme, replanting of native species, removal of alien tree and shrub species if invasive, glade ride creation and coppicing to promote woodland edge habitats. The above are all prescribed to create an uneven-aged tree structure (Davies, 2000).

2.2 SURVEY METHODOLOGY

The data collection for the distributional survey of Hedera helix, was carried out by the author and assistant Andrew Kythreotis from the 2^nd of August to 3^rd of September, 2002. The survey involved collecting data from 200 randomly sampled trees. For each tree sample, that was equal to or > 10 cm in diameter, the circumference breast height (CBH) was measured using a 30m tape measure.

Distance from the edge of the wood (m) was then calculated using a tape measure where conditions permitted or by calculating the distance off a map. Each tree was identified to species level. The texture of the bark on each tree was then assessed by eye as being smooth or rough. Next, Hedera helix was assessed as being present or absent on the tree trunk. If the tree was infested, the percentage cover on the trunk (calculated as a percentage of the whole trunk) was estimated. The growth phase of the liana trunk cover was then assessed as being of the juvenile or mature phase. This was achieved by distinguishing between the two leaf forms. Where both existed on the same tree, rough measurements of the heights of different growth phases on the trunk were calculated. Finally, the stems of H. helix were counted at the base of the trunk and the diameters at ground level were calculated using a vernier calliper.

Liana sampling methods must take into account the fact that lianas are not restricted to growing vertically (Putz, 1984). Therefore, ground cover characteristics of Hedera helix under the sample tree were assessed. Firstly, the percentage of the ground covered by H. helix surrounding the host tree was assessed using 1 x 1m quadrats. Each grid square with H. helix present within the quadrat corresponded to one percent. Therefore, the number of grid squares were counted and a percentage calculated. Following this, the growth form of the ground cover was recorded.

Finally, a spherical densiometer was used to calculate the forest overstorey density (percentage canopy cover). This was conducted by holding the instrument in front of the body at elbow height. Subsequently, the quarter-square canopy openings in the grid were counted. This was repeated over four readings around the tree, facing north, east, south and west. Once the readings were completed, the total numbers of each were multiplied by 1.04 to obtain percentage not occupied by the canopy. The mean of all four readings was then calculated to arrive at the measurement.

2.3 PROBLEMS OF METHODOLOGY

The main problem encountered with the surveying was correct identification of various tree species in Drayton Wood. When looking at the amount of lianas on the host tree samples, the stems were counted without assuming they were individual vines. It was not possible to distinguish whether individual vines were sexually produced individuals (genets) or genetically identical (ramets).

Due to the management activities in the area, the structure of the wood is not as it would be if left to develop naturally. This in turn affected the outcomes of the results. Furthermore, many of the statistical analyses showed that some results were not significant. This may partly be attributed to the number of trees surveyed, as a larger sample may have produced a different result. However, towards the end of the surveying period, a fire incident in Drayton Wood, prematurely ended data collection.

3. RESULTS Hedera helix

3.1 POPULATION CHARACTERISTICS OF Hedera helix.

Hedera helix is a major feature of Drayton wood and is fairly abundant. Of the 200 trees sampled, 120 (60%) hosted H. helix consisting of 1584 separate vine stems. On 37 trees (18.5%), the mature reproductive phase of H. helix predominated. This usually occurred at least 1m above ground level. On 83 trees (41.5%) the juvenile non-reproductive phase of H. helix predominated. However, both juvenile and mature phases coincided on tree samples in many instances. 115 trees (57.5%) were found to have H. helix ground cover at their base. The largest vine in the study area measured 155mm in diameter and resided in the upper canopy of a Quercus robur tree at 145cm CBH.

3.2 HOST SELECTIVITY

Four factors were used to examine host preference using both juvenile and mature vines; present or absent levels, the total numbers of stems per tree species, the total cross-sectional area of the vine per tree species and the percentage of the host trunk covered by H. helix (Table 1).

TABLE 1: Sample tree species and Hedera helix occupation measurements of hosts, Drayton Wood, Norwich.

		*Hedera helix*
Tree Species	Nos. of Trees	No. Present	No. Absent	Stem Abundance	Total Cross-sectional Area (cm²)	Mean Percentage Trunk Cover
*Acer pseudoplatanus*	83	46	37	372	80.87	15.2
*Pinus sylvestris*	58	39	19	782	626.16	13.0
*Quercus robur*	32	19	13	306	594.54	28.7
*Crataegus monogyna*	10	8	2	70	209.20	31.2
*Betula pendula*	6	4	2	25	4.12	60.5
*Sorbus aucuparia*	2	1	1	18	505.34	NA
*Aesculus hippocastanum*	2	1	1	1	1.04	NA
*Corylus avellana*	1	0	1	0	0	NA
*Fraxinus excelsior*	4	0	4	0	0	NA

NB. – Dead trees are not included in table.

When studying the numbers of trees per species with ivy being present or absent, no significant difference was found between observed and expected frequencies (Chi-square =8.71; df =4; P =0.10). Hence, the vines appear to randomly select a host during the juvenile ground-growing phase. However, it appears that once a tree is colonised, Hedera helix is distributed among tree species in proportion to their relative stem abundance (Chi-square =409.9; df =5; P <0.01; Table 1). Further scrutiny revealed the differences between the species (Table 2). Q. robur (20.29%) and P. sylvestris (69.61%) supported more H. helix stems than expected, whereas A. pseudoplatanus (-43%), C. mongyna (-11..94%), B. Pendula (-47.58%) and those grouped as others (S. aucuparia, A. hippocatanum, F. excelsior, C. avellana) supported less (-73.45%).

Again, there was a significant result when examining the distribution of H. helix over tree species in proportion to the total cross-sectional area of the vine (cm²) (Chi-Square =2950.53; df =5; P <0.01; Table 1). When examined in detail, the results show that P. sylvestris (5.75%), Q. robur (82%) and C. monogyna (104.94%) had a higher abundance than expected, relative to the total cross-sectional area of the vine. On the other hand, A. pseudoplatanus (-90.46%) and B. pendula (-93.27%) supported less than expected (Table 2).

Tree Species

Stem Abundance

% difference

Total Cross-sectional Area (cm²)

% difference

Acer pseudoplatanus

Observed

Expected

Residual

372

659.81

-287.81

-43

Observed

Expected

Residual

80.87

847.3

-766.43

-90.46

Pinus sylvestris

Observed

Expected

Residual

782

461.07

320.93

69.61

Observed

Expected

Residual

626.16

592.09

34.07

5.75

Quercus robur

Observed

Expected

Residual

306

254.38

51.62

20.29

Observed

Expected

Residual

594.54

326.67

267.87

Crataegus monogyna

Observed

Expected

Residual

79.49

-9.49

-11.94

Observed

Expected

Residual

209.2

102.08

107.12

104.94

Betula pendula

Observed

Expected

Residual

47.69

-22.69

-47.58

Observed

Expected

Residual

4.12

61.25

-57.13

-93.27

Sorbus aucuparia

Aesculucs hippocatanum

Fraxinus excelsior

Corylus avellana

Observed

Expected

Residual

71.55

-52.55

-73.45

Observed

Expected

Residual

506.38

91.88

414.5

451.13

TABLE 2: Distribution of Hedera helix among tree species with respect to total numbers of stems per tree species and total cross-sectional area of vine per tree species, Drayton Wood, Norwich. Percentage differences between residual counts and expected counts are also shown.

Studies of the mean percentage trunk cover between the most abundant tree species also showed significant differences (Kruskall-Wallis: Chi-square =23.415; df =7; P =0.001). Mann-Whitney tests revealed that the significant differences lay between A. pseudoplatanus and C. monogyna, with A. pseudoplatanus having a significantly lesser mean percent of H. helix trunk cover than C. monogyna (Mann-Whitney: Z =-4.074; N =93; P <0.001). Similar results also arise between B. pendula and C. monogyna (Mann-Whitney: Z =-2.182; N =16; P =0.031).

When studying juvenile and mature vines separately (on the most abundant tree species), there were significant differences in the distributions of the two (Chi-Square =53.70; df =2; P <0.01). A. pseudoplatanus hosted more juveniles and less adults than expected by chance, whereas Q. robur and P. sylvestris hosted less juveniles and more adults than expected by chance.

3.3 INFLUENCE OF HOST SIZE

Initial findings do not support the hypothesis that those trees with larger CBH (cm), are more likely to support Hedera helix than those with a smaller CBH (t =0.451; df =198; P =0.652). However, when examining the number of H. helix stems against the CBH (cm) of the trees, a significant positive correlation was produced (R² =0.09; N =200; P <0.001; Figure 1)

FIGURE 1: Significant positive correlation between CBH (cm) of trees supporting Hedera helix and the number of H. helix stems residing on them, Drayton Wood, Norwich.

The percentage of the host trunk covered by Hedera helix also showed a significant positive correlation with tree CBH (cm) (R² = 0.042; N = 200; P = 0.003; Figure 2).

FIGURE 2: Significant positive correlation between CBH (cm) of trees supporting Hedera helix and the percentage trunk cover of H. helix, Drayton Wood, Norwich.

Further examination of Hedera helix cover of the host trunk illustrated large differences in the percentage frequencies when broken down into intervals (Figure 3). Figure 3 displays a high frequency of low percentage trunk cover and a moderate frequency of high percentage readings of trunk cover.

FIGURE 3: Frequency distribution of percentage trunk cover (10% intervals) in relation to CBH (cm), Drayton Wood, Norwich.

When considering cross-sectional area of Hedera helix vines and tree CBH (cm), independently of species (Figure 4), significant differences were evident. (Mean cross-sectional area of H. helix ± 1SE; 10-34 cm CBH: 0.14 ±0.09; 35-59cm CBH: 1.89 ±0.99;

FIGURE 4: Mean cross-sectional area of Hedera helix against tree CBH (cm) ±1SE, Drayton Wood, Norwich.

60-84cm CBH: 1.73 ±0.57; 85-109cm CBH: 4.84 ±3.51; 110-134cm CBH: 15.04 ±7.11; 135-159cm CBH: 50.76 ±30.62; 160-184cm CBH: 21.63 ±9.31; 185-209cm CBH: 11.26 ±7.77; 210-234 cm CBH: 7.74 ±4.77; 235-259cm CBH: 16.46 ±10.5; 260-284cm CBH: 0.16 ±0.1. Kruskall-Wallis: Chi-Square =32.04; df =10; P <0.001). Further investigations revealed that the differences lay between the CBH interval of 135-159cm and the first three CBH intervals; 10-34cm, 35-59cm and 60-84cm (Mann-Whitney: Z =-4.11; N =38; P <0.001, Mann-Whitney: Z =-3.57; N =66; P <0.001, Mann-Whitney: Z =-2.38; N =54; P =0.01 respectively).

There was no significant correlation between percentage ground cover of Hedera helix and tree CBH (cm) (R² =0.009; N =200; P =0.171).

When tree CBH was divided into classes, (with adult and juvenile Hedera helix vines considered together) it was found that the vines were not distributed independently of tree CBH (Table 3). When investigating stem abundance against tree abundance within each CBH class, significant results were obtained (Chi-Square =481.91; df =7; P <0.1). The results consistently showed that all tree CBH classes <100cm had a lesser stem abundance than expected by chance (10-39.9cm = -69.10%; 40-69.9cm = -21.51%; 70-99.9 = -26.55%), whereas all those tree CBH classes >100cm had a higher than expected stem abundance (100-129.9 = 8.18%; 130-159.9 = 49.34%; 160-189.9 = 62.82%; 190 – 219.9 = 171.5%; 220+ = 72.19%; Table 4).

Tree CBH (cm)	No. of Trees	Stem Abundance

10-39.9	34	70
40-69.9	52	320
70–99.9	29	167
100–129.9	27	229
130–159.9	24	281
160-189.9	17	217
190–219.9	7	149
220+	10	135

TABLE 3: Tree CBH (cm) classes, number of trees and abundance of Hedera helix, Drayton Wood, Norwich.

TABLE 4: Stem abundance of Hedera helix among tree CBH classes with respect to tree abundance, Drayton Wood, Norwich. Percentage differences between residual counts and expected counts are also shown.

Tree CBH (cm)	Stem Abundance		% difference
10-39.9	Observed Expected Residual	70 226.56 -156.56	-69.10
40-69.9	Observed Expected Residual	320 407.68 -87.68	-21.51
70–99.9	Observed Expected Residual	167 227.36 60.36	-26.55
100–129.9	Observed Expected Residual	229 211.68 17.32	8.18
130–159.9	Observed Expected Residual	281 188.16 92.84	49.34
160–189.9	Observed Expected Residual	217 133.28 83.72	62.82
190–219.9	Observed Expected Residual	149 54.88 94.12	171.5
220+	Observed Expected Residual	135 78.4 56.6	72.19

3.4 INFLUENCE OF BARK TEXTURE

In general, it was found that there is a higher abundance of Hedera helix on rough barked trees than smooth barked trees (Mann-Whitney: Z =-2.04; N =200; P =0.041; Figure 5a).

(5a.)

(5b.)

FIGURE 5: (a) Mean percentages of present/absent levels of Hedera helix on smooth and rough barked trees (b) Mean proportion of Hedera helix trunk cover occupying smooth–barked (light-coloured) and rough-barked (dark-coloured) tree species ±1SE, Drayton Wood, Norwich.

Closer examination between tree species also supports the above statement (Figure 5b). The graph displays the mean percentage of trunk cover by Hedera helix for those tree species supporting a higher abundance of the vine. There are obvious differences between those tree species grouped as rough and those grouped as smooth. Mean percentage of H. helix trunk cover ±1SE: Smooth-barked = A. pseudoplatanus: 15.18% ±0.03; B. pendula: 13.33% ±0.11. Rough-barked = C. monogyna: 63% ±13.59; S. aucuparia: 50% ±0.50; P. sylvestris: 31.90% ±0.06; Q. robur: 29.69% ±0.07; A. hippocastanum: 45% ±0.45 (Kruskall-Wallis: Chi-square =23.415; df =7; P =0.001). Although A. pseudoplatanus had the highest number of trees colonised by Hedera helix, it had a much smaller percent trunk cover as compared with the rough-barked species. Mann-Whitney tests revealed that the significant differences lay between A. pseudoplatanus and C. monogyna, A. pseudoplatanus had a significantly lesser mean percent of H. helix trunk cover than C. monogyna (Mann-Whitney: Z =-4.074; N =93; P <0.001). Similar results also arise between B. pendula and C. monogyna (Mann-Whitney: Z =-2.182; N =16; P =0.031).

When examining trees occupied by Hedera helix according to CBH classes and independently of species (Figure 6a), it was found that there was no significant difference observed between groups (Kruskall-Wallis: Chi-square =1.083; df =3; P =0.781).

(6a) (6b)

FIGURE 6: (a) Mean percentage of trees supporting Hedera helix ±1SE (b) Mean percentage of smooth (light-coloured) and rough-barked (dark-coloured) trees supporting Hedera helix ±1SE, Drayton Wood, Norwich.

However, when studying the percentage of trees supporting Hedera helix among CBH classes, a significant difference was observed between those trees with rough bark and those with smooth bark (Chi-square =73.36; df =3; P <0.1). Chi-squared test was carried out on the numbers of trees not percentages. Further investigations showed that in the 10-39.9cm and 40-69.9cm CBH classes, smooth-barked trees had more trees colonised than expected (107.61%, 49.32%), whereas rough-barked trees had a smaller number of colonised trees than expected (-79.54%, -36.45%). In the 70-99.9cm CBH class, smooth-barked trees had only a slightly higher number of colonised trees than expected by chance (13.54%) and rough-barked trees had slightly less colonised trees than expected (-10.07%). However, in the 100 +cm CBH class, smooth-barked trees had significantly less colonised trees than expected by chance (-77.86%), with rough-barked trees having significantly more than expected (57.53%). Refer to Table 5.

Table 5: Numbers of colonised rough and smooth-barked trees, Drayton Wood, Norwich. Percentage differences between residual counts and expected counts are also shown.

Tree CBH (cm)

No. of Colonised Smooth-Barked trees

% difference

No. of Colonised Rough-Barked trees

% difference

10-39.9

Observed

Expected

Residual

14.45

15.55

107.61

Observed

Expected

Residual

19.55

-15.55

-79.54

40-69.9

Observed

Expected

Residual

22.1

10.9

49.32

Observed

Expected

Residual

29.9

-10.9

-36.45

70–99.9

Observed

Expected

Residual

12.33

1.67

13.54

Observed

Expected

Residual

16.68

-1.68

-10.07

100+

Observed

Expected

Residual

36.13

-28.13

-77.86

Observed

Expected

Residual

48.88

28.12

57.53

3.5 Hedera helix GROUND COVER

Figure 7 demonstrates the higher percentage of Hedera helix ground cover under trees supporting H. helix, as opposed to those trees not supporting H. helix. Analyses show that this difference is significant (Mann-Whitney: Z =-10.023; N =200; P <0.001). Furthermore, all ground cover under tree samples were found to be juvenile.

FIGURE 7: Mean percentage of ground cover under those trees not supporting ivy (light coloured) and those supporting ivy (dark coloured) ±1SE, Drayton Wood, Norwich.

Figure 8a shows the wide variation of Hedera helix ground cover between those abundant tree species not hosting the vine. The differences between the means are not significant (Kruskall-Wallis: Chi-Square =3.69; df =4; P =0.45).

(8a)

(8b)

FIGURE 8: Mean percentage of H. helix ground cover under (a) Trees not hosting H. helix ±1SE and (b) those trees hosting H. helix ±1SE., Drayton Wood, Norwich

There were no significant differences in the mean percentage ground cover under those more abundant tree species supporting Hedera helix (one-way ANOVA F = 1.109; P = 0.356; Figure 8b). However, a paired t-test revealed that there was a significant difference in mean percentage ground cover between those trees hosting H. helix and with those trees not hosting H. helix (t =-3.85; df =4; P =0.018).

Looking at ground cover of Hedera helix under tree CBH (cm) classes irrespective of species, for those trees not supporting H. helix, no significant difference was found between the size classes (ANOVA: F =0.340; P =0.933; Figure 9a). Mean percentage ground cover under colonised trees does reveal significant differences (Kruskall-Wallis: Chi-square =17.16; df =7; P =0.016; Figure 9b). Further inspection reveals that significant differences lay between CBH classes 40-69.9cm and 220+cm (Mann-Whitney: Z =-2.782; N =40; P =0.004). Paired t-tests once again reveal that there are significant differences in the mean percentage ground cover between those trees with colonised H. helix and those without (t = -8.003; df =7; P <0.001).

(9a)

(9b)

FIGURE 9: Mean percentage Hedera helix ground cover under tree CBH classes (a) trees without ivy (b) trees with ivy ±1SE, Drayton Wood, Norwich

A positive significant correlation was found between percentage of Hedera helix ground cover and distance from the edge in metres (R² =0.036; N =200; P =0.007). Looking at ground cover under those trees not supporting ivy separately, there were no significant differences (one-way ANOVA: F=0.398 P=0.673; Figure 10a). Again when studying H. helix ground cover with distance from edge below trees supporting H. helix, there was no significant difference between the means (one-way ANOVA F =1.538; P =0.208; Figure 10b).

(a) (b)

FIGURE 10: Mean percentage Hedera helix ground cover with distance from edge (a) trees supporting Hedera helix (b) trees not supporting Hedera helix ±1SE, Drayton Wood, Norwich.

A paired t-test again revealed significant differences in the mean percent Hedera. helix ground cover at distance intervals when comparing all those trees not supporting H. helix and all those trees supporting H. helix. A significantly higher mean percentage of H. helix ground cover was observed under trees supporting ivy, than those trees not supporting Hedera helix (t =-8.524; df =3; P =0.003).

3.6 INFLUENCE OF LOCATION

Figure 11 shows the distance from the edge divided into intervals of 40m and the mean percentage of ground cover for each. Investigations looking at the number of colonised trees, displayed that there was no significant difference with distance from the edge (Chi-square =2.96; df =3; P =0.50).

FIGURE 11: Mean percentage of tree colonised with Hedera helix in relation to distance from the edge (m) ±1SE, Drayton Wood, Norwich.

Results reveal that there is a significant positive correlation of distance from the edge (m) and the percentage trunk cover of Hedera helix (R² =0.077; N =200; P <0.001; Figure 11).

FIGURE 12: Positive correlation between distance from edge (m) and percentages of H. helix trunk cover, Drayton Wood, Norwich.

Investigations of mean percent Hedera helix trunk cover and distance intervals from the edge resulted in a significant difference (Kruskall-Wallis: Chi-square =23.204; df =3; P <0.001; Figure 13). Further examination revealed a significantly higher mean percent ivy trunk cover when considering intervals 81~120m and 121~160m against the first distance interval 1~40m (Mann-Whitney: Z=-3.644; N=149; P<0.001), (Mann-Whitney Z = -3.747; N=134; P<0.001). The distance interval 41~80m was found to have a higher mean percent trunk cover than interval 81~120m and a significantly smaller mean percentage than interval 121~160m (Mann-Whitney: Z = -2.030; N=66; P=0.042, Mann-Whitney; Z =-2.4; N =51; P =0.016).

FIGURE 13: Mean percentage H. helix trunk cover with distance from edge (m) ±1SE, Drayton Wood, Norwich.

3.7 INFLUENCE OF LIGHT

The effect of light does not significantly affect the distribution of Hedera helix ground cover (R² =0.017; N =200; P =0.063). Even when just examining areas where H. helix ground cover was present (and excluding those sample spots without H. helix), there was no significant correlation between the two (R² =0.007; N =200; P =0.369).

Results revealed that there was no significant correlation between canopy cover and trunk cover (R² =0.002, N =200 P =0.576). When excluding those sample spots with no Hedera helix ground cover reached (even when just examining areas where H. helix ground cover was present and excluding those sample spots without H. helix), no significant correlation was found (R² =0 .012; N =200; P =0.224).

4. DISCUSSION

4.1 HOST SELECTIVITY

Results reveal that Hedera helix does not display host preference at the present/absent level on host trees. This suggests that during the juvenile ground growing phase, H. helix is not selective of available support (trees) and randomly targets potential hosts. However, once climbing, other factors determine the ‘success’ of the vine development (i.e. vine abundance). Therefore, H. helix is at its greatest abundance and most likely to reach reproductive maturity if the conditions of the host tree favour it. Occupation measurements demonstrate that the amount of H. helix on a host is not distributed equally between species. H. helix is distributed non-randomly among tree species in proportion to relative vine stem abundance, total cross-sectional area of the vine and percentage of the host trunk covered by the vine. Quercus robur, Crataegus monogyna and Pinus sylvestris species are found to have a significantly higher abundance of H. helix than Acer pseudoplatanus and Betula pendula in general over the three occupation measurements.

Similar studies have found that lianas are found growing more abundantly on some tree species than others. Kernott (2002), found that the abundance of Hedera helix was significantly different between tree species in Monte Palanzana forest, central Italy. Campbell et al (1993) found that lianas are generally not distributed at random on their potential hosts trees due to certain host characteristics which confer resistance or susceptibility to them (Danum Valley Conservation Area, Sabah, East Malaysia). Likewise, Putz (1984) noted that some trees seemed particularly prone to liana infestation with higher abundance than other trees (Barro Colorado Island, Panama). This was explained through host size, shape and location in the forest. Talley et al (1996) found that ‘distribution over the rainforest trees was not homogenous, but exhibited host associations’ in Mount Spec State forest, North Queensland.

Furthermore, juvenile and mature vines were not distributed equally. Acer pseudoplatanus hosted more juveniles and less adults than expected whereas Quercus robur and Pinus sylvestris hosted less juveniles and more mature vines than expected. As mentioned earlier, this could be due to the random host targeting of H. helix ground cover. Quercus robur and Pinus sylvestris must possess important characteristics that ensure the development and abundance of reproductively mature H. helix upon them.

So what explains this apparent non-random distribution among tree species? Lianas are dependent on trees for mechanical support and are generally detrimental to their hosts (Clark et al, 1990; Okerman, 2001; Stevens, 1987; Teramura et al, 1991). They affect trees in several ways; (1) reducing stem diameter growth rates of their hosts, (2) mechanical abrasion, (3) passive strangulation, (4) increasing host susceptibility to ice and wind damage, (5) increasing the probability of the host tree falling (Putz, 1984), (6) competing with the host trees for light, nutrients and water (Putz, 1980), (7) increasing tree mortality rates by weighing down tree crowns and increasing mechanical strain (torque) on the stem and roots, (8) increasing the size of trees pulled down when liana-laden trees fall and (9) slowing rates of tree sapling height growth in treefall gaps through the combined effects of shading and mechanical damage (Chittibabu et al, 2001; Putz, 1984). According to Schnitzer et al (2002), lianas decrease the growth, fecundity and survivorship of trees even at low abundances.

Therefore, it is to a trees advantage to avoid or shed lianas (Hegerty, 1991; Putz 1980). It may be assumed that trees have evolved defence mechanisms to prevent infestation of lianas (Balfour et al, 1993). This will obviously affect the distribution and abundance of H. helix.

There are several factors along with host defence that could affect the distribution and abundance of H. helix, these include, bark texture, host size, the availability of ground cover, location in relation to the edge of the wood and the influence of light (these will be discussed later). Not accounted for in this study, were factors such as altitudinal gradient, level of base nutrients, nitrogen concentration in the soils, aspect, trellis availability and effects of disturbance. These have all been shown to have a significant effect on climber abundance and distribution (Balfour et al, 1993). However, the distribution of climbers is more likely to be influenced by biotic factors such as host architecture, rather than by climatic or soil factors (Balfour et al, 1993).

4.2 INFLUENCE OF HOST SIZE

In this study, vines were not distributed independently of host size. The number of Hedera helix stems and the percentage of host trunk covered by H. helix were significantly positively correlated with host size. A detailed study of percentage trunk cover reveals that there is a higher frequency of tree cover in the lower percentages and higher percentages. This may suggest that once trees are colonised, H. helix will grow rapidly, covering the trunk. According to Putz (1980), Clark et al (1990) and Campbell et al (1993), once a tree has been breached by one liana, others often follow as the first lianas provide a trellis that increases the tree’s accessibility for other lianas.

Mean cross-sectional area of the vine, assumes a near normal distribution where it peaks at CBH classes 135–159cm. This may be explained by the following. Firstly, those trees with smaller CBH generally have a smaller surface area than larger CBH trees and tend to be fast-growing pioneer species. This is not ideal for Hedera helix growth as fast-growing trees have, on average, a better chance of avoiding lianas than slow-growing trees (Putz, 1980). Larger CBH trees may have a smaller cross-sectional area of vine as there is a larger surface area for H. helix to colonise, before reaching the canopy and therefore appears as a smaller cross-sectional area. Those trees with a CBH that falls in the middle are between the two extremes with a dampening of effects and therefore a higher cross-sectional area of the vine. This is supported by findings of Chittibabu et al (2001) when studying diversity and host relationships in a tropical evergreen forest in the Indian Eastern Ghats. They found that lianas frequently infested trees of the 120-150cm CBH.

When divided into size classes, and looking at stem abundance against tree abundance, those trees <100cm CBH had a lesser stem abundance than expected by chance, and all those >100cm CBH had a higher stem abundance than expected. Other studies also support the distribution of lianas according to host size. Clark et al (1990) found that liana loads were positively correlated with tree diameter (La Selva Biological Station, Costa Rica). Furthermore, they found that most non-pioneer trees ≥70cm diameter (CBH = 109.96cm) were colonised. Kernott (2002) found that when studying CBH independently of species, there was a lesser abundance than expected on trees <30cm CBH and correspondingly more than expected on trees >30cm CBH (Monte Palanza, Viterbo, Italy). Talley et al (1996) also found that smaller diameter trees had lower infestation factors than trees with DBH >30cm (CBH = 94.25cm). Talley et al (1996) also found a threshold diameter of about 35-40cm (CBH: 109.96cm – 125.66cm) where infestation increases significantly (Mt. Spet State Forest, North Queensland). This is also evident in this study.

When considering the abundance of lianas according to host size, it is important to note the climbing mechanisms employed by the liana. Hedera helix, like other adventitious root-climbers, is unique in its climbing mechanisms. Stem twiners and vines with modified leaves and tendrils are restricted to climbing small-medium diameter hosts usually <10cm (Teramura et al, 1991). However, vines that climb with roots and tendrils with adhesive disks can climb trees regardless of their diameter (Putz, 1984; Talley et al, 1996), though are most effective at ascending trunks of large trees (Teramura et al, 1991). During this study, it was found that H. helix can colonise a range of different sized hosts both plant and non-plant (personal observation, 2002). Therefore, H. helix is a robust plant adaptable to a variety of conditions.

4.3 INFLUENCE OF BARK TEXTURE

Hedera helix exhibits host preference for those trees with rough-textured bark than smooth-textured bark. When looking at a species level, this was also the case, where smooth-barked species A. pseudoplatanus and B. pendula had a lower mean percentage trunk cover than their rough-barked species counterparts. Although A. pseudoplatanus had the highest number of trees colonised by H. helix, it had a low mean percentage of trunk occupied by H. helix. Why this difference? Adventitious-root-climbing vines, like H. helix require a stable surface for root attachment to the sides of the host tree. Rough-barked trees provide crevices and cavities in which H. helix may more firmly attach, compared to smooth-barked trees. Furthermore, rough-bark trees may trap a good deal of moisture in the crevices which can be exploited by the adventitious roots of H. helix.

This relationship is widely acknowledged in the scientific literature. Putz (1984) found that trees with rough (but not flakey) bark appear more easily climbed than smooth-barked trees by twining lianas (Barro Colorado Island, Panama). Flakey, sloughable bark was proposed by Stevens (1987) to reduce liana success (San Emilio Forest, Costa Rica). Bark characteristics were also proposed as the main factor for the absence of adventitious root-climbing vines on trees of Syzygium papyracelum and Austronmyrtus shepard (Talley et al, 1996).

Studies of mean percentages of trees with ivy show that in smaller CBH classes it is the smooth-barked trees that support the majority of H. helix (usually juvenile). Whereas in the higher CBH classes (>100cm CBH) it is the rough-barked trees (usually mature vines). As mentioned previously, this can be explained by the random selection of host trees during the juvenile, ground-growing phase of H. helix. Those trees that fall in the smaller CBH classes are often fast-growing pioneer trees that will easily shed the vines. They are also usually smoother and contain less moisture and nutrients than larger trees (Talley et al, 1996). Therefore, on these trees, the conditions for H. helix development are far from ideal and the chance of reaching reproductive maturity is slim. Therefore, on these CBH classes, usually more juvenile than adult vines are found. The inverse is true of the larger CBH classes where a high mean percentage of rough-barked trees support H. helix allowing the vine to reach maturity. Talley et al (1996) also suggests that besides having more surface area available for colonisation, larger trees are longer lived, exposing a surface area for longer. This appears to suggest that H. helix displays host preference for large, rough-barked host trees as these conditions provide a well-developed tall trellis, stable root attachment opportunities (Kernott, 2002) and an additional water source.

4.4 Hedera helix GROUND COVER

Colonised trees were closely associated with higher levels of ground cover at their base (almost eighteen times as much) as those that weren’t. It is important to note that during the juvenile ground-growing phase, H. helix can form dense, thick mats of cover. Perhaps, once a certain threshold is crossed, trees in the area have a significant chance of becoming colonised by H. helix (Kernott, 2002). However, tree size and tree species do not appear to play a role in the amount of ground cover at the base. This reinforces the idea of random host targeting of H. helix during its juvenile ground-growing phase.

When considering distance from edge, a significantly higher mean percentage of ivy ground cover was found under trees supporting ivy, than those trees not supporting ivy (with increasing distance from the edge). As mentioned previously, tendril climbers, twiners and scramblers are restricted to small diameter hosts/trellises like shrubs and small-diameter trees. These are usually associated with high irradiant environments such as forest edges or gap edges. Therefore, it can be assumed that there is a high level of competition at forest edges. H. helix like other root-climbers is not restricted by host size. Therefore the vine can exploit a wide range of habitats including those within the dark understorey of late-successional forests, with limited trellis structures (Teramura et al, 1991). Competition with lianas in this area is limited allowing H. helix to thrive. Significant differences were also present when considering those trees with ivy and those without against host size and tree species. This reiterates the idea that in fact the major role in the amount of ground cover at the base of a tree is dependent on whether or not the tree already hosts ivy.

4.5 INFLUENCE OF LOCATION

Results once again show that the mean percent of trees colonised showed no significant difference (with distance from edge). This is due to the non-random host selection of H. helix during its juvenile ground-growing phase. However, Hedera helix trunk cover shows positive correlation with increasing distance from edge. This is also evident when distance is divided up into size classes. The inverse relationship is true of many lianas. Gap and forest edges are most likely to fulfil the trellis and light requirements of lianas (Putz, 1984). During studies of Barro Colorado Island, Panama, Putz (1984) found that 90% of seedlings at the gap edge found trellis supports, but in the forest interior, only 30% of the seedlings found any support whatsoever. Due to the climbing mechanism of H. helix and its adaptability to different irradiance levels, it can thrive in the dark interior of forests where limited trellis structures are available. The lack of competition in the interior also ensures the success of the vine. Other explanations put forward by Kernott (2002), include a lack of potential hosts (i.e. less stems) at edge locations compared to interior of forests and competition between fast-growing, light preferring herb/shrub species with H. helix at edge locations.

4.6 INFLUENCE OF LIGHT

Canopy cover does not seem to directly affect the distribution of either Hedera helix ground cover or trunk cover. Most lianas are suited to high irradiance edge/gap environments. However H. helix is suited to different light environments and therefore, light does not seem to significantly affect the distribution of H. helix. Vine species such as H. helix do well within heterogenous light environments and display a high degree of physiological plasticity to such type conditions (Castellanos, 1991). An example is the process of diomorphism/heteroblasty with the production of two physiologically distinct leaf types. This effectively extends the capability of H. helix to acclimate to the wide seasonal range of irradiances found in temperate forest understories (Teramura et al, 1991).

Canopy cover is a measure of vertical irradiance and not accounted for in this study was the effect of horizontal irradiance. This may play a larger role than vertical irradiance in temperate woodlands. It has already been found that there is a higher abundance of H. helix with distance from the edge of the wood. This relationship may also be due to available light. H. helix may prefer those areas in the interior with low irradiance as opposed to edge areas with high irradiance. However, during the period of this study H. helix was found to occur under many different light regimes.

This study was conducted during summer months when there was low irradiance in the understorey. However, during winter months, prior to foliage production by overstorey species, the light available to understorey plants, including evergreen lianas, is greatly enhanced (Teramura et al, 1991). Perhaps a more pronounced relationship between abundance of H. helix and light would be obvious during winter months.

The lack of influence of light on the distribution of H. helix is conducive to a study by Balfour et al (1993), which found that the distribution of climbers is more likely to be influenced by biotic factors such as host architecture, rather than by climatic or soil factors.

5. CONCLUSION

This study has shown that Hedera helix does exhibit host preference. During the juvenile ground-growing phase, H. helix randomly targets potential hosts. However, once a tree is colonised, other factors will determine the abundance, distribution and development of H. helix upon the tree host. Results have shown that tree species Quercus robur, Crataegus mongyna and Pinus sylvestris had a significantly higher abundance of H. helix than tree species Acer pseudoplatanus and Betula pendula. Therefore, these tree species must have characteristics that ensure development and abundance of reproductively mature H. helix upon them, or otherwise those characteristics that inhibit colonisation and growth.

It was found that H. helix develops and grows most successfully on large, rough-barked hosts. These two factors ensure a well-developed tall trellis for climbing, stable root attachment opportunities and an additional water source for the roots. H. helix was found to be less successful on small, smooth-barked hosts as these trees tend to be fast-growing and have a better chance of shedding lianas, furthermore, the height of these trees is not significant for H. helix development and the bark does not allow for stable root attachment.

H. helix trunk cover and ground cover was found to be more abundant with increasing distance from the edge. Most lianas are suited to the high irradiance environments of gap and forest edges. Therefore, it can be assumed that there is a high level of competition in these sites. However, due to the climbing mechanism of H. helix and its adaptability to different irradiance levels, it can thrive in the dark interior of forests where limited trellis structures are available. The lack of competition in the interior also ensures the success of the vine.

There also appears to be a threshold level when considering ground cover of H. helix. When this is exceeded trees in the area have a significant chance of becoming colonised by H. helix.

Finally the abiotic factor of light did not seen to directly affect the distribution or abundance of ground cover or trunk covered by H. helix. This is attributed to the adaptability of H. helix to different light environments. Furthermore, it has been proven that the distribution of climbers is in fact influenced by biotic factors such as host architecture, rather than by climatic factors.

Although these factors limiting the abundance and development of H. helix appear to be consequential, many studies have shown that trees may in fact have evolved defence mechanisms to avoid or shed lianas. Lianas are generally detrimental to a host and those tree species with less than expected H. helix may have actually evolved defence mechanisms to hinder the growth of H. helix.

FUTURE DIRECTIONS

Information from this dissertation looking at factors influencing host preference can be used in other environments where H. helix is posing a major threat to native forest communities. Currently there are two control methods used in the United States; herbicide application and physical removal (Okerman, 2001). However, these have not proven to be completely successful as Hedera helix is a highly robust and adaptable plant species. Studies such as this that examine the underlying factors such as the distribution characteristics of H. helix may help to create effective management schemes.

However, this dissertation just scratches the surface with respect to the factors influencing Hedera helix distribution, abundance and development. Other factors in the literature that significantly affect the distribution of lianas include: defence mechanisms of trees. For example phytotoxic allellochemical interactions (Talley et al, 1996), buttressing, rapid diameter growth, symbionts and trees with large compound leaves that regularly drop branches (Putz, 1980). As well as this, studies have been conducted on the effect of aspect (Kernott, 2002), altitudinal gradient, bole heights of host trees (Balfour et al, 1993), light, water and nitrogen (Dillenburg et al, 1993).

An important point to take away is that it is often not only one factor influencing the distribution, abundance and success of lianas but rather several factors in combination.

ACKNOWLEDGEMENTS

TABLE OF CONTENTS

LIST OF TABLES

LIST OF FIGURES

LIST OF APPENDICES

1. INTRODUCTION

1.1 BACKGROUND

1.2 PROJECT AIMS

1.3 STUDY SPECIES (Appendix 1a)

1.4 LIST OF ABBREVIATIONS AND ACRONYMS

2. METHODOLOGY

2.1 STUDY SITE

2.2 SURVEY METHODOLOGY

2.3 PROBLEMS OF METHODOLOGY

3. RESULTS Hedera helix

3.2 HOST SELECTIVITY

3.3 INFLUENCE OF HOST SIZE

3.4 INFLUENCE OF BARK TEXTURE

3.6 INFLUENCE OF LOCATION

3.7 INFLUENCE OF LIGHT

4. DISCUSSION

4.1 HOST SELECTIVITY

4.2 INFLUENCE OF HOST SIZE

REFERENCES

OKERMAN, A., (2001), Combating the “Ivy Desert”: The Invasion of Hedera helix (English Ivy) in the Pacific Northwest United States.

RUSSELL, A.B., HARDIN, J.W., GRAND, L. and FRASER, A., (1997), Poisonous Plants: Hedera helix, North Carolina State University, North Carolina. http://www.ces.ncsu.edu/depts/hort/consumer/poison/Hederhe.htm

SEILER, J., JENSEN, E.C., and PETERSON, J., (2001), Hedera Helix Fact Sheet, English Ivy, Araliaceae., Department of Forestry, Virginia Tech, Virginia.

SWEARINGEN, J.M., (2000), PCA Alien Plant Working Group – English Ivy (Hedera helix), U.S. National Park Service, Washington, DC.
http://www.nps.gov/plants/alien/fact/hehe1.htm

APPENDIX 1

Source: OKERMAN, A., (2001), Combating the “Ivy Desert”: The Invasion of Hedera helix (English Ivy) in the Pacific Northwest United States.

ACKNOWLEDGEMENTS

TABLE OF CONTENTS

LIST OF TABLES

LIST OF FIGURES

LIST OF APPENDICES

1. INTRODUCTION

1.1 BACKGROUND

1.2 PROJECT AIMS

1.3 STUDY SPECIES (Appendix 1a)

1.4 LIST OF ABBREVIATIONS AND ACRONYMS

2. METHODOLOGY

2.1 STUDY SITE

2.2 SURVEY METHODOLOGY

2.3 PROBLEMS OF METHODOLOGY

3. RESULTS Hedera helix

3.2 HOST SELECTIVITY

3.3 INFLUENCE OF HOST SIZE

3.4 INFLUENCE OF BARK TEXTURE

3.6 INFLUENCE OF LOCATION

3.7 INFLUENCE OF LIGHT

4. DISCUSSION

4.1 HOST SELECTIVITY

4.2 INFLUENCE OF HOST SIZE

REFERENCES

OKERMAN, A., (2001), Combating the “Ivy Desert”: The Invasion of Hedera helix (English Ivy) in the Pacific Northwest United States.

RUSSELL, A.B., HARDIN, J.W., GRAND, L. and FRASER, A., (1997), Poisonous Plants: Hedera helix, North Carolina State University, North Carolina. http://www.ces.ncsu.edu/depts/hort/consumer/poison/Hederhe.htm

SEILER, J., JENSEN, E.C., and PETERSON, J., (2001), Hedera Helix Fact Sheet, English Ivy, Araliaceae., Department of Forestry, Virginia Tech, Virginia.

SWEARINGEN, J.M., (2000), PCA Alien Plant Working Group – English Ivy (Hedera helix), U.S. National Park Service, Washington, DC. http://www.nps.gov/plants/alien/fact/hehe1.htm

APPENDIX 1

Source: OKERMAN, A., (2001), Combating the “Ivy Desert”: The Invasion of Hedera helix (English Ivy) in the Pacific Northwest United States.

SWEARINGEN, J.M., (2000), PCA Alien Plant Working Group – English Ivy (Hedera helix), U.S. National Park Service, Washington, DC.
http://www.nps.gov/plants/alien/fact/hehe1.htm